{"collection":"philosophy-ethics","full_text_available":true,"graph_url":"https://unchartedknowledge.5gfusion.net/v2/api/page/gutenberg_43572/graph","html_url":"https://unchartedknowledge.5gfusion.net/v2/page/gutenberg_43572","id":"gutenberg_43572","license":"Public Domain (US)","rights_basis":"gutenberg_metadata_copyright_false","source":"gutenberg","source_url":"https://www.gutenberg.org/ebooks/43572","summary":null,"tags":["Category: Philosophy & Ethics","Kant, Immanuel, 1724-1804. Kritik der reinen Vernunft"],"text":"Produced by Meredith Bach, lizardcry, Chuck Greif and the\nOnline Distributed Proofreading Team at http://www.pgdp.net\n(This file was produced from images generously made\navailable by The Internet Archive/American Libraries.)\n\n\n\n\n\n\n\n   Every attempt has been made to replicate the original as printed.\nSome typographical errors have been corrected; a list follows the text.\n                     The footnotes follow the text.\n               Index Contents (etext transcriber's note)\n\n\n\n\n                              A COMMENTARY\n                                   TO\n                          KANT\u2019S \u2018CRITIQUE OF\n                              PURE REASON\u2019\n\n                                   BY\n\n                       NORMAN KEMP SMITH, D.PHIL.\n\n          McCOSH PROFESSOR OF PHILOSOPHY, PRINCETON UNIVERSITY\n                               AUTHOR OF\n                 \u2018STUDIES IN THE CARTESIAN PHILOSOPHY\u2019\n\n                       MACMILLAN AND CO., LIMITED\n                      ST. MARTIN\u2019S STREET, LONDON\n                                  1918\n\n                               COPYRIGHT\n\n                             TO THE MEMORY\n\n                                   OF\n\n                             ROBERT ADAMSON\n\n                WISE IN COUNSEL, IN FRIENDSHIP UNFAILING\n\n                          GRATEFULLY DEDICATED\n\n\n\n\nPREFACE\n\n\nThe _Critique of Pure Reason_ is more obscure and difficult than even a\nmetaphysical treatise has any right to be. The difficulties are not\nmerely due to defects of exposition; they multiply rather than diminish\nupon detailed study; and, as I shall endeavour to show in this\n_Commentary_, are traceable to two main causes, the composite nature of\nthe text, written at various dates throughout the period 1772-1780, and\nthe conflicting tendencies of Kant\u2019s own thinking.\n\nThe _Commentary_ is both expository and critical; and in exposition no\nless than in criticism I have sought to subordinate the treatment of\ntextual questions and of minor issues to the systematic discussion of\nthe central problems. Full use is made of the various selections from\nKant\u2019s private papers that have appeared, at intervals, since the\npublication of his _Lectures on Metaphysics_ in 1821. Their significance\nhas not hitherto been generally recognised in English books upon Kant.\nThey seem to me to be of capital importance for the right understanding\nof the _Critique_.\n\nSome apology is perhaps required for publishing a work of this character\nat the present moment. It was completed, and arrangements made for its\npublication, shortly before the outbreak of war. The printers have, I\nunderstand, found in it a useful stop-gap to occupy them in the\nintervals of more pressing work; and now that the type must be released,\nI trust that in spite of, or even because of, the overwhelming\npreoccupations of the war, there may be some few readers to whom the\nvolume may be not unwelcome. That even amidst the distractions of actual\ncampaigning metaphysical speculation can serve as a refuge and a solace\nis shown by the memorable example of General Smuts. He has himself told\nus that on his raid into Cape Colony in the South African War he carried\nwith him for evening reading the _Critique of Pure Reason_. Is it\nsurprising that our British generals, pitted against so unconventional\nan opponent, should have been worsted in the battle of wits?\n\nThe _Critique of Pure Reason_ is a philosophical classic that marks a\nturning-point in the history of philosophy, and no interpretation, even\nthough now attempted after the lapse of a hundred years, can hope to be\nadequate or final. Some things are clearer to us than they were to\nKant\u2019s contemporaries; in other essential ways our point of view has\nreceded from his, and the historical record, that should determine our\njudgments, is far from complete. But there is a further difficulty of an\neven more serious character. The _Critique_ deals with issues that are\nstill controversial, and their interpretation is possible only from a\ndefinite standpoint. The limitations of this standpoint and of the\nphilosophical _milieu_ in which it has been acquired unavoidably\nintervene to distort or obscure our apprehension of the text. Arbitrary\nand merely personal judgments I have, however, endeavoured to avoid. My\nsole aim has been to reach, as far as may prove feasible, an unbiassed\nunderstanding of Kant\u2019s great work.\n\nAmong German commentators I owe most to Vaihinger, Adickes, B. Erdmann,\nCohen, and Riehl, especially to the first named. The chief English\nwriters upon Kant are Green, Caird, and Adamson. In so far as Green and\nCaird treat the Critical philosophy as a half-way stage to the Hegelian\nstandpoint I find myself frequently in disagreement with them; but my\nindebtedness to their writings is much greater than my occasional\ncriticisms of their views may seem to imply. With Robert Adamson I\nenjoyed the privilege of personal discussions at a time when his earlier\nview of Kant\u2019s teaching was undergoing revision in a more radical manner\nthan is apparent even in his posthumously published University lectures.\nTo the stimulus of his suggestions the writing of this _Commentary_ is\nlargely due.\n\nMy first study of the _Critique_ was under the genial and inspiring\nguidance of Sir Henry Jones. With characteristic kindliness he has read\nthrough my manuscript and has disclosed to me many defects of\nexposition and argument. The same service has been rendered me by\nProfessor G. Dawes Hicks, whose criticisms have been very valuable,\nparticularly since they come from a student of Kant who on many\nfundamental points takes an opposite view from my own.\n\nI have also to thank my colleague, Professor Oswald Veblen, for much\nhelpful discussion of Kant\u2019s doctrines of space and time, and of\nmathematical reasoning.\n\nMr. H. H. Joachim has read the entire proofs, and I have made frequent\nmodifications to meet his very searching criticisms. I have also\ngratefully adopted his revisions of my translations from the _Critique_.\nSimilar acknowledgments are due to my colleague, Professor A. A. Bowman,\nand to my friend Dr. C. W. Hendel.\n\nI have in preparation a translation of the _Critique of Pure Reason_,\nand am responsible for the translations of all passages given in the\npresent work. In quoting from Kant\u2019s other writings, I have made use of\nthe renderings of Abbott, Bernard, and Mahaffy; but have occasionally\nallowed myself the liberty of introducing alterations.\n\nShould readers who are already well acquainted with the _Critique_\ndesire to use my _Commentary_ for its systematic discussions of Kant\u2019s\nteaching, rather than as an accompaniment to their study of the text, I\nmay refer them to those sections which receive italicised headings in\nthe table of contents.\n\n                                              NORMAN KEMP SMITH.\n\nLONDON, _January 1918_.\n\n\n\n\nCONTENTS\n\n\nINTRODUCTION\n\n                                                                    PAGE\n\n  I. TEXTUAL--\n\n     Kant's Method of composing the _Critique of Pure Reason_        xix\n\n II. HISTORICAL--\n\n     Kant's Relation to Hume and to Leibniz                          xxv\n\nIII. GENERAL--\n\n     1. The Nature of the _a priori_                              xxxiii\n     2. Kant's Contribution to the Science of Logic                xxxvi\n     3. The Nature of Consciousness                                xxxix\n     4. Phenomenalism, Kant's Substitute for Subjectivism            xlv\n     5. The Distinction between Human and Animal Intelligence      xlvii\n     6. The Nature and Conditions of Self-Consciousness                l\n     7. Kant's threefold Distinction between Sensibility,\n        Understanding, and Reason                                    lii\n     8. The place of the _Critique of Pure Reason_ in\n        Kant's Philosophical System                                   lv\n\n\nTHE CRITIQUE OF PURE REASON[1]\n\n_Title_                                                                1\n_Motto_                                                                4\n_Dedication to Freiherr von Zedlitz_                                   6\n\nPREFACE TO THE FIRST EDITION                                           8\n\n    _Comment on Preface_                                              10\n    _Dogmatism_, _Scepticism_, _Criticism_                            13\n\nPREFACE TO THE SECOND EDITION                                         17\n\n    _The Copernican Hypothesis_                                       22\n\nINTRODUCTION                                                          26\n\n    _Comment upon the Argument of Kant's Introduction_                33\n    _How are Synthetic_ a priori _Judgments possible?_                43\n    _The Analytic and Synthetic Methods_                              44\n    _Purpose and Scope of the Critique_                               56\n    _Kant's relation to Hume_                                         61\n    _Meaning of the term Transcendental_                              73\n\n\nTHE TRANSCENDENTAL DOCTRINE OF ELEMENTS\n\nPart I. THE TRANSCENDENTAL AESTHETIC                              79-166\n\n    _Definition of Terms_                                             79\n    _Kant's conflicting Views of Space_                               88\n\n  Section I. SPACE                                                    99\n\n    _Kant's Attitude to the Problems of Modern Geometry_             117\n\n  Section II. TIME                                                   123\n\n    _Kant's Views regarding the Nature of Arithmetical Science_      128\n    _Kant's conflicting Views of Time_                               134\n    General Observations on the Transcendental Aesthetic             143\n    _The Distinction between Appearance and Illusion_                148\n    _Kant's Relation to Berkeley_                                    155\n    _The Paradox of Incongruous Counterparts_                        161\n\nPart II. THE TRANSCENDENTAL LOGIC                                    167\n\n  Introduction                                                       167\n\n         I. Logic in General                                         167\n        II. Transcendental Logic                                     170\n       III. The Division of General Logic into Analytic\n            and Dialectic                                            172\n\n  Division I. THE TRANSCENDENTAL ANALYTIC                            174\n\n    Book I. THE ANALYTIC OF CONCEPTS                                 175\n\n      Chapter I. THE CLUE TO THE DISCOVERY OF ALL PURE CONCEPTS\n                 OF THE UNDERSTANDING                                175\n\n      Section I. The Logical Use of the Understanding                176\n             _Comment on Kant's Argument_                            176\n             _Stages in the Development of Kant's\n              Metaphysical Deduction_                                186\n\n      Section II. The Logical Function of the Understanding\n                  in Judgment                                        192\n\n      Section III. The Categories on Pure Concepts of the\n                   Understanding                                     194\n\n               _Distinction between Logical Forms and Categories_    195\n\nPart II. THE TRANSCENDENTAL LOGIC--_Continued._\n\n    Chapter II. DEDUCTION OF THE PURE CONCEPTS OF THE UNDERSTANDING  202\n\n        _Analysis of the Text: the Four Stages in the\n         Development of Kant's Views_                            202-234\n\n          I. _Enumeration of the Four Stages_                        203\n\n         II. _Detailed Analysis of the Four Stages_                  204\n             _Kant's Doctrine of the Transcendental Object_          204\n\n        III. _Evidence yielded by the \"Reflexionen\" and \"Lose\n             Bl\u00e4tter\" in Support of the Analysis of the\n             Text_                                                   231\n\n         IV. _Connected Statement and Discussion of Kant's\n             Subjective and Objective Deductions in the\n             First Edition_                                          234\n\n        _Distinction between the Subjective and the Objective\n        Deductions_                                                  235\n\n        _The Subjective Deduction in its initial empirical\n        Stages_                                                      245\n\n        _Objective Deduction as given in the First Edition_          248\n\n        _The later Stages of the Subjective Deduction_               263\n\n        _The Distinction between Phenomenalism and Subjectivism_     270\n\n        _Transcendental Deduction of the Categories in the\n        Second Edition_                                              284\n\n        _The Doctrine of Inner Sense_                                291\n\n        _Kant's Refutations of Idealism_                             298\n\n        _Inner Sense and Apperception_                               321\n\n  Book II. THE ANALYTIC OF PRINCIPLES                                332\n\n    Chapter I. THE SCHEMATISM OF PURE CONCEPTS OF\n               THE UNDERSTANDING                                     334\n\n    Chapter II. SYSTEM OF ALL PRINCIPLES OF PURE UNDERSTANDING       342\n\n      1. The Axioms of Intuition                                     347\n\n      2. The Anticipations of Perception                             349\n\n      3. The Analogies of Experience                                 355\n\n        _A._ First Analogy                                           358\n\n        _B._ Second Analogy                                          363\n\n          _Schopenhauer's Criticism of Kant's Argument_              365\n\n          _Kant's Subjectivist and Phenomenalist Views\n          of the Causal Relation_                                    373\n\n          _Reply to Further Criticisms of Kant's Argument_           377\n\nPart II. THE TRANSCENDENTAL LOGIC--_Continued._\n\n          _C._ Third Analogy                                         381\n\n            _Schopenhauer's Criticism of Kant's Argument_            387\n\n        4. The Postulates of Empirical Thought in General            391\n\n      Chapter III. ON THE GROUND OF THE DISTINCTION\n                   OF ALL OBJECTS WHATEVER INTO\n                   PHENOMENA AND NOUMENA                             404\n\n        _Relevant Passages in the Section on Amphiboly_              410\n\n        _Alterations in the Second Edition_                          412\n\n        _Comment on Kant's Argument_                                 414\n\n      Appendix. The Amphiboly of the Concepts of Reflection          418\n\n  Division II. THE TRANSCENDENTAL DIALECTIC                          424\n\n        _Introductory Comment upon the composite Origin and\n        conflicting Tendencies of the Dialectic_                     425\n\n        _The History and Development of Kant's Views in\n        regard to the Problems of the Dialectic_                     431\n\n    Introduction                                                     441\n\n       I. Transcendental Illusion                                    441\n\n      II. Pure Reason as the Seat of Transcendental Illusion         442\n\n    Book I. THE CONCEPTS OF PURE REASON                              446\n\n      Section I. Ideas in General                                    447\n\n      Section II. The Transcendental Ideas                           450\n\n      Section III. System of the Transcendental Ideas                453\n\n    Book II. THE DIALECTICAL INFERENCES OF PURE REASON               455\n\n      Chapter I. THE PARALOGISMS OF PURE REASON                      455\n\n          First Paralogism: of Substantiality                        457\n\n          Second Paralogism: of Simplicity                           458\n\n          Third Paralogism: of Personality                           461\n\n          Fourth Paralogism: of Ideality                             462\n\n        _Second Edition Statement of the Paralogisms_                466\n\n        _Is the Notion of the Self a necessary Idea of Reason?_      473\n\n      Chapter II. THE ANTINOMY OF PURE REASON                        478\n\n        Section I. System of the Cosmological Ideas                  478\n\n        Section II. Antithetic of Pure Reason                        480\n\n          _Comment on Kant's Method of Argument_                     481\n\n          First Antinomy                                             483\n\n          Second Antinomy                                            488\n\n          Third Antinomy                                             492\n\n          Fourth Antinomy                                            495\n\nPart II. THE TRANSCENDENTAL LOGIC--_Continued._\n\nSection III. The Interest of Reason in this Self-Conflict            498\n\nSection IV. Of the Transcendental Problems of Pure\nReason in so far as they absolutely must\nbe capable of Solution                                               499\n\nSection V. Sceptical Representation of the Cosmological\nQuestions                                                            501\n\nSection VI. Transcendental Idealism as the Key to the\nSolution of the Cosmological Dialectic                               503\n\nSection VII. Critical Decision of the Cosmological\nConflict of Reason with itself                                       504\n\nSection VIII. The Regulative Principle of Pure Reason\nin regard to the Cosmological Ideas                                  506\n\nSection IX. The Empirical Employment of the Regulative\nPrinciples of Reason in regard\nto all Cosmological Ideas                                            508\n\nSolution of the First and Second Antinomies                          508\n\nRemarks on the Distinction between the\nMathematical-Transcendental and the\nDynamical-Transcendental Ideas                                       510\n\n_Comment on Kant's Method of Argument_                               510\n\nSolution of the Third Antinomy                                       512\n\nPossibility of harmonising Causality through\nFreedom with the Universal Law of\nNatural Necessity                                                    513\n\nExplanation of the Relation of Freedom to\nNecessity of Nature                                                  514\n\n_Comment on Kant's Method of Argument_                               517\n\nSolution of the Fourth Antinomy                                      518\n\nConcluding Note on the whole Antinomy\nof Pure Reason                                                       519\n\n_Concluding Comment on Kant's Doctrine\nof the Antinomies_                                                   519\n\nChapter III. THE IDEAL OF PURE REASON                                522\n\nSections I. and II. The Transcendental Ideal                         522\n\n_Comment on Kant's Method of\nArgument_                                                            524\n\nSection III. The Speculative Arguments in Proof of the\nExistence of a Supreme Being                                         525\n\nSection IV. The Impossibility of an Ontological Proof                527\n\n_Comment on Kant's Method of Argument_                               528\n\nPart II. THE TRANSCENDENTAL LOGIC--_Continued._\n\nSection V. The Impossibility of a Cosmological Proof\nof the Existence of God                                              531\n\n_Comment on Kant's Method of Argument_                               533\n\nDiscovery and Explanation of the Transcendental\nIllusion in all Transcendental\nProof of the Existence of a necessary\nBeing                                                                534\n\n_Comment on Kant's Method of Argument_                               535\n\nSection VI. The Impossibility of the Physico-Theological\nProof                                                                538\n\nSection VII. Criticism of all Theology based on speculative\nPrinciples of Reason                                                 541\n\n_Concluding Comment_                                                 541\n\nAPPENDIX TO THE TRANSCENDENTAL DIALECTIC                             543\n\nThe Regulative Employment of the Ideas of Pure Reason                543\n\n_Hypotheses not permissible in Philosophy_                           543\n\nOn the Final Purpose of the Natural Dialectic of Human\nReason                                                               552\n\n_Concluding Comment on the Dialectic_                                558\n\n_APPENDIX A._\n\nTHE TRANSCENDENTAL DOCTRINE OF METHODS                               563\n\nChapter I. THE DISCIPLINE OF PURE REASON                             563\n\nSection I. The Discipline of Pure Reason in its Dogmatic\nEmployment                                                           563\n\nSection II. The Discipline of Pure Reason in its\nPolemical Employment                                                 567\n\nSection III. The Discipline of Pure Reason in regard\nto Hypotheses                                                        568\n\nSection IV. The Discipline of Pure Reason in regard\nto its Proofs                                                        568\n\nChapter II. THE CANON OF PURE REASON                                 569\n\nSection I. The Ultimate End of the Pure Use of our\nReason                                                               569\n\nSection II. The Ideal of the Highest Good, as a Determining\nGround of the Ultimate End of\nPure Reason                                                          570\n\nSection III. Opining, Knowing, and Believing                         576\n\nChapter III. THE ARCHITECTONIC OF PURE REASON                        579\n\nChapter IV. THE HISTORY OF PURE REASON                               582\n\n_APPENDIX B._\n\n_A more detailed Statement of Kant's Relations to his Philosophical\nPredecessors_                                                        583\n\nINDEX                                                                607\n\n\n\n\nNOTE\n\nIn all references to the _Kritik der Reinen Vernunft_ I have given the\noriginal pagings of both the first and second editions. References to\nKant\u2019s other works are, whenever possible, to the volumes thus far\nissued in the new Berlin edition. As the _Reflexionen Kants zur Kritik\nder reinen Vernunft_ had not been published in this edition at the time\nwhen the _Commentary_ was completed, the numbering given is that of B.\nErdmann\u2019s edition of 1884.\n\n\n\n\nABBREVIATIONS\n\nBerlin edition of Kant\u2019s works                                    _W_\n\nPagings in the first edition of the _Kritik der reinen Vernunft_   A\n\nPagings in the second edition                                      B\n\nAdickes\u2019 edition of the _Kritik der reinen Vernunft_ (1889)       _K_\n\n\n\n\nINTRODUCTION\n\n\n\n\nI. TEXTUAL\n\n\nKANT\u2019S METHOD OF COMPOSING THE \u2018CRITIQUE OF PURE REASON\u2019\n\nSeldom, in the history of literature, has a work been more\nconscientiously and deliberately thought out, or more hastily thrown\ntogether, than the _Critique of Pure Reason_. The following is the\naccount which Kant in a letter to Moses Mendelssohn (August 16, 1783)\nhas given of its composition:\n\n     \u201d[Though the _Critique_ is] the outcome of reflection which had\n     occupied me for a period of at least twelve years, I brought it to\n     completion in the greatest haste within some four to five months,\n     giving the closest attention to the content, but with little\n     thought of the exposition or of rendering it easy of comprehension\n     by the reader--a decision which I have never regretted, since\n     otherwise, had I any longer delayed, and sought to give it a more\n     popular form, the work would probably never have been completed at\n     all. This defect can, however, be gradually removed, now that the\n     work exists in a rough form.\u201d[2]\n\nThese statements must be allowed the greater weight as Kant, in another\nletter (to Garve, August 7, 1783), has given them in almost the same\nwords:\n\n     \u201cI freely admit that I have not expected that my book should meet\n     with an immediate favourable reception. The exposition of the\n     materials which for more than twelve successive years I had been\n     carefully maturing, was not composed in a sufficiently suitable\n     manner for general comprehension. For the perfecting of its\n     exposition several years would have been required, whereas I\n     brought it to completion in some four to five months, in the fear\n     that, on longer delay, so prolonged a labour might finally become\n     burdensome, and that my increasing years (I am already in my\n     sixtieth year) would perhaps incapacitate me, while I am still the\n     sole possessor of my complete system.\u201d[3]\n\nThe twelve years here referred to are 1769-1780; the phrase \u201cat least\ntwelve years\u201d indicates Kant\u2019s appreciation of the continuity of his\nmental development. Hume\u2019s first influence upon Kant is probably to be\ndated prior to 1760. The choice, however, of the year 1769 is not\narbitrary; it is the year of Kant\u2019s adoption of the semi-Critical\nposition recorded in the _Inaugural Dissertation_ (1770).[4] The \u201cfour\nto five months\u201d may be dated in the latter half of 1780. The printing of\nthe _Critique_ was probably commenced in December or January 1780-1781.\n\nBut the _Critique_ is not merely defective in clearness or popularity of\nexposition. That is a common failing of metaphysical treatises,\nespecially when they are in the German language, and might pass without\nspecial remark. What is much more serious, is that Kant flatly\ncontradicts himself in almost every chapter; and that there is hardly a\ntechnical term which is not employed by him in a variety of different\nand conflicting senses. As a writer, he is the least exact of all the\ngreat thinkers.\n\nSo obvious are these inconsistencies that every commentator has felt\nconstrained to offer some explanation of their occurrence. Thus Caird\nhas asserted that Kant opens his exposition from the non-Critical\nstandpoint of ordinary consciousness, and that he discloses the final\nposition, towards which he has all along been working, only through\nrepeated modifications of his preliminary statements. Such a view,\nhowever, cannot account either for the specific manner of occurrence or\nfor the actual character of the contradictions of which the _Critique_\naffords so many examples. These are by no means limited to the opening\nsections of its main divisions; and careful examination of the text\nshows that they have no such merely expository origin. The publication\nof Kant\u2019s _Reflexionen_ and _Lose Bl\u00e4tter_, and the devoted labours of\nBenno Erdmann, Vaihinger, Adickes, Reicke and others, have, indeed,\nplaced the issue upon an entirely new plane. It can now be proved that\nthe _Critique_ is not a unitary work, and that in the five months in\nwhich, as Kant tells us, it was \u201cbrought to completion\u201d (_zu Stande\ngebracht_), it was not actually written, but was pieced together by the\ncombining of manuscripts written at various dates throughout the period\n1772-1780.\n\nKant\u2019s correspondence in these years contains the repeated assertion\nthat he expected to be able to complete the work within some three or\nsix months. This implies that it was already, at least as early as 1777,\nin great part committed to writing. In 1780 Kant must therefore have\nhad a large body of manuscript at his disposal. The recently published\n_Lose Bl\u00e4tter_ are, indeed, part of it. And as we shall have constant\noccasion to observe, the _Critique_ affords ample evidence of having\nbeen more or less mechanically constructed through the piecing together\nof older manuscript, supplemented, no doubt, by the insertion of\nconnecting links, and modified by occasional alterations to suit the new\ncontext. Kant, it would almost seem, objected to nothing so much as the\nsacrifice of an argument once consecrated by committal to paper. If it\ncould be inserted, no matter at what cost of repetition, or even\nconfusion, he insisted upon its insertion. Thus the _Subjective_ and\n_Objective Deductions_ of the first edition can, as we shall find, be\nbroken up into at least four distinct layers, which, like geological\nstrata, remain to the bewilderment of the reader who naturally expects a\nunified system, but to the enlightenment of the student, once the clues\nthat serve to identify and to date them have been detected. To cite\nanother example: in the _Second Analogy_, as given in the first edition,\nthe main thesis is demonstrated in no less than five distinct proofs,\nsome of which are repetitions; and when Kant restated the argument in\nthe second edition, he allowed the five proofs to remain, but\nsuperimposed still another upon them. Kant does, indeed, in the second\nedition omit some few passages from various parts of the _Critique_; but\nthis is in the main owing to his desire to protect himself against\nserious misunderstanding to which, as he found, he had very unguardedly\nlaid himself open. The alterations of the second edition are chiefly of\nthe nature of additions.\n\nAdickes\u2019 theory[5] that Kant in the \u201cfour to five months\u201d composed a\nbrief outline of his entire argument, and that it was upon the framework\nof this outline that the _Critique_ was elaborated out of the older\nmanuscript, may possibly be correct. It has certainly enabled Adickes to\ncast much light upon many textual problems. But his own supplementary\nhypothesis in regard to the section on the _Antinomies_, namely, that it\nformed an older and separate treatise, may very profitably be further\nextended. Surely it is unlikely that with the expectation, continued\nover many years, of completion within a few months, Kant did not\npossess, at least for the _Aesthetic_, _Dialectic_, and _Methodology_, a\ngeneral outline, that dated further back than 1780. And doubtless this\noutline was itself altered, patched, and recast, in proportion as\ninsight into the problems of the _Analytic_, the problems, that is to\nsay, which caused publication to be so long deferred, deepened and took\nfinal form.\n\nThe composite character of the _Critique_ is largely concealed by the\nhighly elaborate, and extremely artificial, arrangement of its parts. To\nthe general plan, based upon professedly logical principles, Kant has\nhimself given the title, architectonic; and he carries it out with a\nthoroughness to which all other considerations, and even at times those\nof sound reasoning, are made to give way. Indeed, he clings to it with\nthe unreasoning affection which not infrequently attaches to a favourite\nhobby. He lovingly elaborates even its minor detail, and is rewarded by\na framework so extremely complicated that the most heterogeneous\ncontents can be tidily arranged, side by side, in its many compartments.\nBy its uniformity and rigour it gives the appearance of systematic order\neven when such order is wholly absent.\n\nBut we have still to consider the chief reason for the contradictory\ncharacter of the contents of the _Critique_. It is inseparably bound up\nwith what may perhaps be regarded as Kant\u2019s supreme merit as a\nphilosophical thinker, especially as shown in the first\n_Critique_,--namely, his open-minded recognition of the complexity of\nhis problems, and of the many difficulties which lie in the way of any\nsolution which he is himself able to propound. Kant\u2019s method of working\nseems to have consisted in alternating between the various possible\nsolutions, developing each in turn, in the hope that some midway\nposition, which would share in the merits of all, might finally disclose\nitself. When, as frequently happened, such a midway solution could not\nbe found, he developed his thought along the parallel lines of the\nalternative views.\n\n     \u201cYou know that I do not approach reasonable objections with the\n     intention merely of refuting them, but that in thinking them over I\n     always weave them into my judgments, and afford them the\n     opportunity of overturning all my most cherished beliefs. I\n     entertain the hope that by thus viewing my judgments impartially\n     from the standpoint of others some third view that will improve\n     upon my previous insight may be obtainable.... Long experience has\n     taught me that insight into a subject which I am seeking to master\n     is not to be forced, or even hastened, by sheer effort, but demands\n     a fairly prolonged period during which I return again and again to\n     the same concepts, viewing them in all their aspects and in their\n     widest possible connections, while in the intervals the sceptical\n     spirit awakens, and makes trial whether my conclusions can\n     withstand a searching criticism.\u201d[6] \u201cIn mental labour of so\n     delicate a character nothing is more harmful than preoccupation\n     with extraneous matters. The mind, though not constantly on the\n     stretch, must still, alike in its idle and in its favourable\n     moments, lie uninterruptedly open to any chance suggestion which\n     may present itself. Relaxations and diversions must maintain its\n     powers in freedom and mobility, so that it may be enabled to view\n     the object afresh from every side, and so to enlarge its point of\n     view from a microscopic to a universal outlook that it adopts in\n     turn every conceivable standpoint, verifying the observations of\n     each by means of all the others.\u201d[7] \u201cI am not of the opinion of\n     the well-meaning writer who has recommended us never to allow\n     doubts in regard to a matter upon which we have once made up our\n     minds. In pure philosophy that is not feasible. Indeed the\n     understanding has in itself a natural objection to any such\n     procedure. We must consider propositions in all their various\n     applications; even when they may not seem to require a special\n     proof, we must make trial of their opposites, and in this way fight\n     for delay, until the truth becomes in all respects evident.\u201d[8]\n\nThat these are no mere pious expressions of good intention, but\nrepresent Kant\u2019s actual method of working, is amply proved by the\ncontents of the _Critique_. We find Kant constantly alternating between\nopposed standpoints, to no one of which he quite definitely commits\nhimself, and constantly restating his principles in the effort to remove\nthe objections to which, as he recognises, they continue to lie open.\nThe _Critique_, as already stated, is not the exposition of a single\nunified system, but is the record of Kant\u2019s manifold attempts to\nformulate and to solve his many-sided problems. Even those portions of\nthe _Critique_ which embody his latest views show that Kant is still\nunwilling to sacrifice insight to consistency. When he is guilty of\nspecial pleading--for he cannot be altogether absolved even from that\ncharge--it is in the interests of his logical architectonic, for which,\nas I have said, he cherishes a quite unreasoning affection, and not of\nhis central principles. So far from concealing difficulties, or unduly\ndwelling upon the favouring considerations, Kant himself emphasises the\noutstanding objections to which his conclusions remain subject. If his\nteaching is on certain points very definite, it is in other hardly less\nimportant respects largely tentative.\n\nThe value of Kant\u2019s _Critique_ as an introduction to modern philosophy\nis greatly enhanced by this method of procedure. The student who has\nsteeped himself in the atmosphere of the _Critique_, however\ndissatisfied he may perhaps be with many of its doctrines, has become\nfamiliar with the main requirements which a really adequate metaphysics\nmust fulfil, or at least will have acquired a due sense of the\ncomplexity of the problems with which it deals.\n\nRecognition of the composite nature of the text will safeguard us in two\nways. In the first place, citation of single passages is quite\ninconclusive. Not only must all the relevant passages be collated; they\nmust be interpreted in the light of an historical understanding of the\nvarious stages in Kant\u2019s development. We must also be prepared to find\nthat on certain main questions Kant hesitates between opposed positions,\nand that he nowhere definitively commits himself to any quite final\nexpression of view.\n\nSecondly, we cannot proceed on the assumption that Kant\u2019s maturest\nteaching comes where, had the _Critique_ been a unitary work, composed\nupon a definite and previously thought out plan, we should naturally\nexpect to find it, namely, in its concluding portions. The teaching of\nmuch of the _Dialectic_, especially in its account of the nature of the\nphenomenal world and of its relation to the knowing mind, is only\nsemi-Critical. This is also true of Kant\u2019s _Introduction_ to the\n_Critique_. Introductions are usually written last; and probably Kant\u2019s\n_Introduction_ was written after the completion of the _Aesthetic_, of\nthe _Dialectic_, and of the _Analytic_ in its earlier forms. But it\nbears all the signs of having been composed prior to the working out of\nseveral of his most characteristic doctrines in the central parts of the\n_Analytic_.\n\nThus both Kant\u2019s introductory statements of the aims and purposes of the\n_Critique_, and his application of his results in the solution of\nmetaphysical problems, fail to represent in any adequate fashion the new\nand revolutionary principles to which he very gradually but successfully\nworked his way. The key to the _Critique_ is given in the central\nportions of the _Analytic_, especially in the _Deduction of the\nCategories_. The other parts of the _Critique_ reveal the Critical\ndoctrines only as gradually emerging from the entangling influence of\npre-Critical assumptions. Their teaching has to be radically remodelled\nbefore they can be made to harmonise with what, in view both of their\nintrinsic character and of the corresponding alterations in the second\nedition, must be regarded as Kant\u2019s maturest utterances.\n\nThis was a task which Kant never himself attempted. For no sooner had he\nattained to comparative clearness in regard to his new Critical\nprinciples and briefly expounded them in the _Analytic_ of the first\nedition, than he hastened to apply them in the spheres of morality,\naesthetics, and teleology. When the _Critique_ appeared in 1781 he was\nfifty-seven years of age; and he seems to have feared that if he\nallowed these purely theoretical problems, which had already occupied\nhis main attention for \u201cat least twelve years,\u201d to detain him longer, he\nwould be debarred from developing and placing on permanent record the\nnew metaphysics of ethics which, as the references in the first\n_Critique_ show, had already begun to shape itself in his mind. To have\nexpended further energy upon the perfecting of his theoretical\nphilosophy would have endangered its own best fruits. Even the\nopportunity in 1787 of a second edition of the _Critique_ he used very\nsparingly, altering or adding only where occasional current\ncriticism--his puzzled contemporaries having still for the most part\nmaintained a discreet silence--had clearly shown that his modes of\nexposition were incomplete or misleading.\n\n\n\n\nII. HISTORICAL\n\n\nKANT\u2019S RELATION TO HUME AND TO LEIBNIZ\n\nKant\u2019s manner of formulating his fundamental problem--How are synthetic\n_a priori_ judgments possible?--may well seem to the modern reader to\nimply an unduly scholastic and extremely rationalistic method of\napproach. Kant\u2019s reasons for adopting it have, unfortunately, been\nlargely obscured, owing to the mistaken interpretation which has usually\nbeen given to certain of his personal utterances. They have been\nsupposed to prove that the immediate occasion of the above formula was\nHume\u2019s discussion of the problem of causality in the _Enquiry into the\nHuman Understanding_. Kant, it is argued, could not have been acquainted\nwith Hume\u2019s earlier and more elaborate _Treatise on Human Nature_, of\nwhich there was then no translation; and his references to Hume must\ntherefore concern only the later work.\n\nVaihinger has done valuable service in disputing this reading of Kant\u2019s\nautobiographical statements. Kant does not himself make direct mention\nof the _Enquiry_, and the passages in the _Critique_ and in the\n_Prolegomena_[9] in which Hume\u2019s teaching is under consideration seem\nrather to point to the wider argument of the _Treatise_. This is a\nmatter of no small importance; for if Vaihinger\u2019s view can be\nestablished, it will enable us to appreciate, in a manner otherwise\nimpossible, how Kant should have come to regard the problem of _a priori\nsynthesis_ as being the most pressing question in the entire field of\nspeculative philosophy.\n\nThe essential difference between the _Treatise_ and the _Enquiry_, from\nthe standpoint of their bearing upon Critical issues, lies in the wider\nscope and more radical character of the earlier work. The _Enquiry_\ndiscusses the problem of causality only in the form in which it emerges\nin _particular_ causal judgments, _i.e._ as to our grounds for asserting\nthat this or that effect is due to this or that cause. In the\n_Treatise_, Hume raises the broader question as to our right to\npostulate that events must always be causally determined. In other\nwords, he there questions the validity of the _universal_ causal\nprinciple, that whatever begins to exist must have a cause of existence;\nand he does so on the explicit ground that it demands as necessary the\nconnecting of two concepts, that of an event and that of an antecedent\ncause, between which _no connection of any kind_ can be detected by the\nmind. The principle, that is to say, is not self-evident; it is\nsynthetic. The concept of an event and the concept of a cause are quite\nseparate and distinct ideas. Events can be conceived without our\nrequiring to think antecedent events upon which they are dependent. Nor\nis the principle capable of demonstration. For if it be objected that in\nquestioning its validity we are committing ourselves to the impossible\nassertion that events arise out of nothing, such argument is only\napplicable if the principle be previously granted. If events do not\nrequire a cause, it is as little necessary to seek their source in a\ngeneration out of nothing as in anything positive. Similarly, when it is\nargued that as all the parts of time and space are uniform, there must\nbe a cause determining an event to happen at one moment and in one place\nrather than at some other time or place, the principle is again assumed.\nThere is no greater difficulty in supposing the time and place to be\nfixed without a cause than in supposing the existence to be so\ndetermined. The principle, Hume concludes, is non-rational in character.\nIt is an instrument useful for the organisation of experience; and for\nthat reason nature has determined us to its formation and acceptance.\nProperly viewed, it expresses a merely instinctive belief, and is\nexplicable only in the naturalistic manner of our other propensities, as\nnecessary to the fulfilling of some practical need. \u201cNature has\ndetermined us to judge as well as to breathe and feel.\u201d\n\nFrom this naturalistic position Hume makes a no less vigorous attack\nupon the empirical philosophies which profess to establish general\nprinciples by inductive inference from the facts of experience. If the\nprinciples which lie at the basis of our experience are non-rational in\ncharacter, the same must be true of our empirical judgments. They may\ncorrectly describe the uniformities that have hitherto occurred in the\nsequences of our sensations, and may express the natural expectations to\nwhich they spontaneously give rise; but they must never be regarded as\ncapable of serving as a basis for inference. In eliminating _a priori_\nprinciples, and appealing exclusively to sense-experience, the\nempiricist removes all grounds of distinction between inductive\ninference and custom-bred expectation. And since from this standpoint\nthe possibility of universal or abstract concepts--so Hume argues--must\nalso be denied, deductive inference must likewise be eliminated from\namong the possible instruments at the disposal of the mind. So-called\ninference is never the source of our beliefs; it is our fundamental\nnatural beliefs, as determined by the constitution of our nature in its\nreaction upon external influences, that generate those expectations\nwhich, however they may masquerade in logical costume, have as purely\nnatural a source as our sensations and feelings. Such, briefly and\ndogmatically stated, is the sum and substance of Hume\u2019s teaching.[10]\n\nNow it was these considerations that, as it would seem, awakened Kant to\nthe problem of _a priori synthesis_. He was, and to the very last\nremained, in entire agreement with Hume\u2019s contention that the principle\nof causality is neither self-evident nor capable of logical\ndemonstration, and he at once realised that what is true of this\nprinciple must also hold of all the other principles fundamental to\nscience and philosophy. Kant further agreed that inductive inference\nfrom the data of experience is only possible upon the prior acceptance\nof rational principles independently established; and that we may not,\ntherefore, look to experience for proof of their validity. Thus with the\nrejection of self-evidence as a feature of the _a priori_, and with the\nconsequent admission of its synthetic character, Kant is compelled to\nacquiesce in the inevitableness of the dilemma which Hume propounds.\nEither Hume\u2019s sceptical conclusions must be accepted, or we must be able\nto point to some criterion which is not subject to the defects of the\nrationalist and empirical methods of proof, and which is adequate to\ndetermine the validity or invalidity of general principles. Is there any\nsuch alternative? Such is Kant\u2019s problem as expressed in the formula:\nHow are synthetic _a priori_ judgments possible?\n\nIt is a very remarkable historical fact that notwithstanding the\nclearness and cogency of Hume\u2019s argument, and the appearance of such\ncompetent thinkers as Thomas Reid in Scotland, Lambert and Crusius in\nGermany, no less than thirty years should have elapsed before Hume found\na single reader capable of appreciating the teaching of the _Treatise_\nat its true value.[11] Even Kant himself was not able from his reading\nof the _Enquiry_ in 1756-1762 to realise the importance and bearing of\nthe main problem.[12] Though in the _Enquiry_ the wider issue regarding\nthe general principle of causality is not raised, the bearing of Hume\u2019s\ndiscussion, when interpreted in the light of Kant\u2019s own teaching, is\nsufficiently clear; and accordingly we cannot be absolutely certain that\nit was not a re-reading of the _Enquiry_ or a recalling of its\nargument[13] that suggested to Kant the central problem of his Critical\nphilosophy. The probability, however, is rather that this awakening took\nplace only indirectly through his becoming acquainted with the wider\nargument of the _Treatise_ as revealed in James Beattie\u2019s extremely\ncrude and unsympathetic criticism of Hume\u2019s philosophy.[14] Beattie had\ngreat natural ability, and considerable literary power. His prose\nwritings have a lucidity, a crispness, and a felicity of illustration\nwhich go far to explain their widespread popularity in the latter half\nof the eighteenth century. Their literary quality is, however, more than\ncounterbalanced by the absence of any genuine appreciation of the\ndeeper, speculative implications and consequences of the problems\ndiscussed. And this being so, he is naturally at his worst in\ncriticising Hume. In insisting, as he does, upon the absurd practical\nresults[15] that would follow from the adoption of Hume\u2019s sceptical\nconclusions, he is merely exploiting popular prejudice in the\nphilosophical arena. That, however, may be forgiven him, if, as would\nseem to be the case, the quotations which he gives verbatim from Hume\u2019s\n_Treatise_ really first revealed to Kant the scope and innermost meaning\nof Hume\u2019s analysis of the causal problem.\n\nThe evidence in support of this contention is entirely circumstantial.\nThe German translation of Beattie\u2019s _Essay on the Nature and\nImmutability of Truth_ was published at Easter 1772, _i.e._ in the year\nin which Kant, in the process of his own independent development, came,\nas is shown by his famous letter to Herz,[16] to realise the mysterious,\nproblematic character of _a priori_ knowledge _of the independently\nreal_. He was then, however, still entirely unconscious of the deeper\nproblem which at once emerges upon recognition that _a priori_\nprinciples, quite apart from all question of their objective validity,\nare synthetic in form. We know that Kant was acquainted with Beattie\u2019s\nwork; for he twice refers to Beattie\u2019s criticism of Hume.[17] What more\nprobable than that he read the translation in the year of its\npublication, or at least at some time not very long subsequent to the\ndate of the letter to Herz? The passages which Beattie quotes from the\n_Treatise_ are exactly those that were necessary to reveal the full\nscope of Hume\u2019s revolutionary teaching in respect to the general\nprinciple of causality. There seems, indeed, little doubt that this must\nhave been the channel through which Hume\u2019s influence chiefly acted. Thus\nat last, by a circuitous path, through the quotations of an adversary,\nHume awakened philosophy from its dogmatic slumber,[18] and won for his\nargument that appreciation which despite its cogency it had for thirty\nyears so vainly demanded.\n\nLet us now turn our attention to the rationalist philosophy in which\nKant was educated. Hume\u2019s contention that experience cannot by itself\njustify any inductive inference, forms the natural bridge over which we\ncan best pass to the contrasting standpoint of Leibniz. Hume and Leibniz\nfind common ground in denouncing empiricism. Both agree in regarding it\nas the mongrel offspring of conflicting principles. If rationalism\ncannot hold its own, the alternative is not the finding of firm foothold\nin concrete experience, but only such consolation as a sceptical\nphilosophy may afford.[19] The overthrow of rationalism means the\ndestruction of metaphysics in every form. Even mathematics and the\nnatural sciences will have to be viewed as fulfilling a practical end,\nnot as satisfying a theoretical need. But though Leibniz\u2019s criticism of\nempiricism is, in its main contention, identical with that of Hume, it\nis profoundly different both in its orientation and in the conclusions\nto which it leads. While Hume maintains that induction must be regarded\nas a non-rational process of merely instinctive anticipation, Leibniz\nargues to the self-legislative character of pure thought.\nSense-experience reveals reality only in proportion as it embodies\nprinciples derived from the inherent character of thought itself.\nExperience conforms to _a priori_ principles, and so can afford an\nadequate basis for scientific induction.\n\nThere is a passage in Hume\u2019s _Enquiry_ which may be employed to\nillustrate the boldly speculative character of Leibniz\u2019s interpretation\nof the nature and function of human thought. \u201cNothing ... [seems] more\nunbounded than the thought of man, which not only escapes all human\npower and authority, but is not even restrained within the limits of\nnature and reality.... While the body is confined to one planet, along\nwhich it creeps with pain and difficulty, the thought can in an instant\ntransport us into the most distant regions of the universe.... What\nnever was seen, or heard of, may yet be conceived; nor is anything\nbeyond the power of thought, except what implies an absolute\ncontradiction.\u201d This passage in which Hume means to depict a false\nbelief, already sufficiently condemned by the absurdity of its claims,\nexpresses for Leibniz the wonderful but literal truth. Thought is the\nrevealer of an eternal unchanging reality, and its validity is in no\nway dependent upon its verification through sense. When Voltaire in his\n_Ignorant Philosopher_ remarks that \u201cit would be very singular that all\nnature, all the planets, should obey eternal laws, and that there should\nbe a little animal, five feet high, who, in contempt of these laws,\ncould act as he pleased, solely according to his caprice,\u201d[20] he is\nforgetting that this same animal of five feet can contain the stellar\nuniverse in thought within himself, and has therefore a dignity which is\nnot expressible in any such terms as his size may seem, for vulgar\nestimation, to imply. Man, though dependent upon the body and confined\nto one planet, has the sun and stars as the playthings of his mind.\nThough finite in his mortal conditions, he is divinely infinite in his\npowers.\n\nLeibniz thus boldly challenges the sceptical view of the function of\nreason. Instead of limiting thought to the translating of sense-data\ninto conceptual forms, he claims for it a creative power which enables\nit out of its own resources to discover for itself, not only the actual\nconstitution of the material world, but also the immensely wider realm\nof possible entities. The real, he maintains, is only one of the many\nkingdoms which thought discovers for itself in the universe of truth. It\nis the most comprehensive and the most perfect, but still only one out\nof innumerable others which unfold themselves to the mind in pure\nthought. Truth is not the abstracting of the universal aspects in\nthings, not a copy of reality, dependent upon it for meaning and\nsignificance. Truth is wider than reality, is logically prior to it, and\ninstead of being dependent upon the actual, legislates for it. Leibniz\nthus starts from the possible, as discovered by pure thought, to\ndetermine in an _a priori_ manner the nature of the real.\n\nThis Leibnizian view of thought may seem, at first sight, to be merely\nthe re-emergence of the romantic, rationalistic ideal of Descartes and\nMalebranche. So to regard it would, however, be a serious injustice. It\nwas held with full consciousness of its grounds and implications, and\nreality was metaphysically reinterpreted so as to afford it a genuine\nbasis. There was nothing merely mystical and nothing undefined in its\nmain tenets. Leibniz differs from Malebranche in being himself a\nprofound mathematician, the co-discoverer with Newton of the\ndifferential calculus. He also differs from Descartes in possessing an\nabsorbing interest in the purely logical aspects of the problem of\nmethod; and was therefore equipped in a supreme degree for determining\nin genuinely scientific fashion the philosophical significance and value\nof the mathematical disciplines.\n\nHume and Leibniz are thus the two protagonists that dwarf all others.\nThey realised as neither Malebranche, Locke, nor Berkeley, neither Reid,\nLambert, Crusius, nor Mendelssohn ever did, the really crucial issues\nwhich must ultimately decide between the competing possibilities. Each\nmaintained, in the manner prescribed by his general philosophy, one of\nwhat then appeared to be the only two possible views of the function of\nthought. The alternatives were these: (_a_) Thought is merely a\npractical instrument for the convenient interpretation of our human\nexperience; it has no objective or metaphysical validity of any kind;\n(_b_) Thought legislates universally; it reveals the wider universe of\nthe eternally possible; and prior to all experience can determine the\nfundamental conditions to which that experience must conform. Or to\ninterpret this opposition in logical terms: (_a_) The fundamental\nprinciples of experience are synthetic judgments in which no relation is\ndiscoverable between subject and predicate, and which for that reason\ncan be justified neither _a priori_ nor by experience; (_b_) all\nprinciples are analytic, and can therefore be justified by pure thought.\n\nThe problem of Kant\u2019s _Critique_, broadly stated, consists in the\nexamination and critical estimate of these two opposed views. There is\nno problem, scientific, moral, or religious, which is not vitally\naffected by the decision which of these alternatives we are to adopt, or\nwhat reconciliation of their conflicting claims we hope to achieve.\nSince Kant\u2019s day, largely owing to the establishment of the evolution\ntheory, this problem has become only the more pressing. The\nnaturalistic, instrumental view of thought seems to be immensely\nreinforced by biological authority. Thought would seem to be reduced to\nthe level of sense-affection, and to be an instrument developed through\nnatural processes for the practical purposes of adaptation. Yet the\ncounter-view has been no less powerfully strengthened by the victorious\nmarch of the mathematical sciences. They have advanced beyond the limits\nof Euclidean space, defining possibilities such as no experience reveals\nto us. The Leibnizian view has also been reinforced by the successes of\nphysical science in determining what would seem to be the actual,\nobjective character of the independently real. Kant was a rationalist by\neducation, temperament, and conviction. Consequently his problem was to\nreconcile Leibniz\u2019s view of the function of thought with Hume\u2019s proof of\nthe synthetic character of the causal principle. He strives to\ndetermine how much of Leibniz\u2019s belief in the legislative power of pure\nreason can be retained after full justice has been done to Hume\u2019s\ndamaging criticisms. The fundamental principles upon which all\nexperience and all knowledge ultimately rest are _synthetic_ in nature:\nhow is it possible that they should also be _a priori_? Such is the\nproblem that was Kant\u2019s troublous inheritance from his philosophical\nprogenitors, Hume and Leibniz.[21]\n\n\n\n\nIII. GENERAL\n\n\nIn indicating some of the main features of Kant\u2019s general teaching, I\nshall limit myself to those points which seem most helpful in\npreliminary orientation, or which are necessary for guarding against the\nmisunderstandings likely to result from the very radical changes in\nterminology and in outlook that have occurred in the hundred and thirty\nyears since the publication of the _Critique_. Statements which thus\nattempt to present in outline, and in modern terms, the more general\nfeatures of Kant\u2019s philosophical teaching will doubtless seem to many of\nmy readers dogmatic in form and highly questionable in content. They\nmust stand or fall by the results obtained through detailed examination\nof Kant\u2019s _ipsissima verba_. Such justification as I can give for them\nwill be found in the body of the _Commentary_.\n\n\nI. THE NATURE OF THE _A PRIORI_\n\nThe fundamental presupposition upon which Kant\u2019s argument rests--a\npresupposition never itself investigated but always assumed--is that\nuniversality and necessity cannot be reached by any process that is\nempirical in character. By way of this initial assumption Kant arrives\nat the conclusion that the _a priori_, the distinguishing\ncharacteristics of which are universality and necessity, is not given in\nsense but is imposed by the mind; or in other less ambiguous terms, is\nnot part of the matter of experience but constitutes its form. The\nmatter of experience is here taken as equivalent to sensation; while\nsensation, in turn, is regarded as being the non-relational.\n\nThe explanation of Kant\u2019s failure either to investigate or to prove this\nassumption has already been indicated. Leibniz proceeds upon the\nassumption of its truth no less confidently than Hume, and as Kant\u2019s\nmain task consisted in reconciling what he regarded as being the\nelements of truth in their opposed philosophies, he very naturally felt\nsecure in rearing his system upon the one fundamental presupposition on\nwhich they were able to agree. It lay outside the field of controversy,\nand possessed for Kant, as it had possessed for Hume and for Leibniz,\nthat authoritative and axiomatic character which an unchallenged\npreconception tends always to acquire.\n\nThe general thesis, that the universal and necessary elements in\nexperience constitute its form, Kant specifies in the following\ndeterminate manner. The form is fixed for all experience, that is to\nsay, it is one and the same in each and every experience, however simple\nor however complex. It is to be detected in consciousness of duration no\nless than in consciousness of objects or in consciousness of self. For,\nas Kant argues, consciousness of duration involves the capacity to\ndistinguish between subjective and objective succession, and likewise\ninvolves recognition[22] with its necessary component\nself-consciousness. Or to state the same point of view in another way,\nhuman experience is a temporal process and yet is always a consciousness\nof meaning. As temporal, its states are ordered successively, that is,\nexternally to one another; but the consciousness which they constitute\nis at each and every moment the awareness of some single unitary meaning\nby reference to which the contents of the successive experiences are\norganised. The problem of knowledge may therefore be described as being\nthe analysis of the consciousness of duration, of objectivity, and of\nself-consciousness, or alternatively as the analysis of our awareness of\nmeaning. Kant arrives at the conclusion that the conditions of all four\nare one and the same.[23]\n\nKant thus teaches that experience in all its embodiments and in each of\nits momentary states can be analysed into an endlessly variable material\nand a fixed set of relational elements. And as no one of the relational\nfactors can be absent without at once nullifying all the others, they\ntogether constitute what must be regarded as the determining form and\nstructure of every mental process that is cognitive in character.\nAwareness, that is to say, is identical with the act of judgment, and\ntherefore involves everything that a judgment, in its distinction from\nany mere association of ideas, demands for its possibility.\n\nKant\u2019s position, when thus stated, differs from that of Leibniz only in\nits clearer grasp of the issues and difficulties involved, and\nconsequently in the more subtle, pertinacious, and thoroughgoing\ncharacter of the argument by which it is established. Its revolutionary\ncharacter first appears when Kant further argues, in extension of the\nteaching of Hume, that the formal, relational elements are of a\n_synthetic_ nature. The significance and scope of this conclusion can\nhardly be exaggerated. No other Kantian tenet is of more fundamental\nimportance.[24] With it the main consequences of Kant\u2019s Critical\nteaching are indissolubly bound up. _As the principles which lie at the\nbasis of our knowledge are synthetic, they have no intrinsic necessity,\nand cannot possess the absolute authority ascribed to them by the\nrationalists._ They are prescribed to human reason, but cannot be shown\nto be inherently rational in any usual sense of that highly ambiguous\nterm. They can be established only as brute conditions, verifiable in\nfact though not demonstrable in pure theory (if there be any such\nthing), of our actual experience. They are conditions of\n_sense_-experience, and that means of our knowledge of appearances,\nnever legitimately applicable in the deciphering of ultimate reality.\nThey are valid within the realm of experience, useless for the\nconstruction of a metaphysical theory of things in themselves. This\nconclusion is reinforced when we recognise that human experience, even\nin its fundamental features (_e.g._ the temporal and the spatial), might\nconceivably be altogether different from what it actually is, and that\nits presuppositions are always, therefore, of the same contingent\ncharacter. Even the universality and necessity which Kant claims to have\nestablished for his _a priori_ principles are of this nature. Their\nnecessity is always for us extrinsic; they can be postulated only if,\nand so long as, we are assuming the occurrence of human\nsense-experience.\n\nThus Kant is a rationalist of a new and unique type. He believes in, and\nemphasises the importance of, the _a priori_. With it alone, he\ncontends, is the _Critique_ competent to deal. But it is an _a priori_\nwhich cannot be shown to be more than relative. It does, indeed, enable\nus to conceive the known as relative, and to entertain in thought the\npossibility of an Absolute; but this it can do without itself possessing\nindependent validity. For though the proof of the _a priori_ is not\nempirical in the sense of being inductive, neither is it logical in the\nsense of being deduced from necessities of thought. Its \u201ctranscendental\u201d\nproof can be executed only so long as experience is granted as actual;\nand so long as the fundamental characteristics of this experience are\nkept in view.\n\nLastly, the _a priori_ factors are purely relational. They have no\ninherent content from which clues bearing on the supersensible can be\nobtained. Their sole function is to serve in the interpretation of\ncontents otherwise supplied.\n\nThe _a priori_, then, is merely relational, without inherent content; it\nis synthetic, and therefore incapable of independent or metaphysical\nproof; it is relative to an experience which is only capable of yielding\nappearances. The _a priori_ is as merely factual as the experience which\nit conditions.\n\nEven in the field of morality Kant held fast to this conviction.\nMorality, no less than knowledge, presupposes _a priori_ principles.\nThese, however, are never self-evident, and cannot be established by any\nmere appeal to intuition. They have authority only to the extent to\nwhich they can be shown to be the indispensable presuppositions of a\nmoral consciousness that is undeniably actual.[25]\n\nThat the _a priori_ is of this character must be clearly understood.\nOtherwise the reader will be pursued by a feeling of the unreality, of\nthe merely historical or antiquarian significance, of the entire\ndiscussion. He may, if he pleases, substitute the term formal or\nrelational for _a priori_. And if he bears in mind that by the\nrelational Kant is here intending those elements in knowledge which\nrender possible the relations constitutive of _meaning_, he will\nrecognise that the Critical discussion is by no means antiquated, but\nstill remains one of the most important issues in the entire field of\nphilosophical enquiry.\n\n\n2. KANT\u2019S CONTRIBUTION TO THE SCIENCE OF LOGIC\n\nThe above conclusions have an important bearing upon logical doctrine.\nJust as modern geometry originates in a sceptical treatment of the axiom\nof parallels, so modern, idealist logic rests upon Kant\u2019s demonstration\nof the revolutionary consequences of Hume\u2019s sceptical teaching. If\nprinciples are never self-evident, and yet are not arrived at by\ninduction from experience, by what alternative method can they be\nestablished? In answer to this question, Kant outlines the position\nwhich is now usually entitled the _Coherence_ theory of truth.[26] That\ntheory, though frequently ascribed to Hegel, has its real sources in the\n_Critique of Pure Reason_. It expresses that modification in the\nLeibnizian rationalism which is demanded by Hume\u2019s discovery of the\nsynthetic character of the causal axiom. Neither the deductive methods\nof the Cartesian systems nor the inductive methods of the English\nphilosophies can any longer be regarded as correctly describing the\nactual processes of scientific proof.\n\nGeneral principles are either presuppositions or postulates. If _a\npriori_, they are presupposed in all conscious awareness; as above\nindicated, they have a _de facto_ validity within the experience which\nthey thus make possible. If more special in nature, they are the\npostulates to which we find ourselves committed in the process of\nsolving specific problems; and they are therefore discovered by the\nmethod of trial and failure.[27] They are valid in proportion as they\nenable us to harmonise appearances, and to adjudicate to each a kind of\nreality consistent with that assigned to every other.\n\nProof of fact is similar in general character. The term fact is\neulogistic, not merely descriptive; it marks the possession of cognitive\nsignificance in regard to some body of knowledge, actual or possible. It\ncan be applied to particular appearances only in so far as we can\ndetermine their conditions, and can show that as thus conditioned the\nmode of their existence is relevant to the enquiry that is being\npursued. The convergence of parallel lines is fact from the standpoint\nof psychological investigation; from the point of view of their physical\nexistence it is merely appearance. Ultimately, of course, everything is\nreal, including what we entitle appearance;[28] but in the articulation\nof human experience such distinctions are indispensable, and the\ncriteria that define them are prescribed by the context in which they\nare being employed.\n\nThus facts cannot be established apart from principles, nor principles\napart from facts. The proof of a principle is its adequacy to the\ninterpretation of all those appearances that can be shown to be in any\nrespect relevant to it, while the test of an asserted fact, _i.e._ of\nour description of a given appearance, is its conformity to the\nprinciples that make insight possible.\n\nThough the method employed in the _Critique_ is entitled by Kant the\n\u201ctranscendental method,\u201d it is really identical in general character\nwith the hypothetical method of the natural sciences. It proceeds by\nenquiring what conditions must be postulated in order that the\nadmittedly given may be explained and accounted for.[29] Starting from\nthe given, it also submits its conclusions to confirmation by the given.\nConsidered as a method, there is nothing metaphysical or high-flying\nabout it save the name. None the less, Kant is in some degree justified\nin adopting the special title. In view of the unique character of the\nproblem to be dealt with, the method calls for very careful statement,\nand has to be defended against the charge of inapplicability in the\nphilosophical field.\n\nThe fundamental thesis of the Coherence theory finds explicit\nformulation in Kant\u2019s doctrine of the judgment: the doctrine, that\nawareness is identical with the act of judging, and that judgment is\nalways complex, involving both factual and interpretative elements.\nSynthetic, relational factors are present in _all_ knowledge, even in\nknowledge that may seem, on superficial study, to be purely analytic or\nto consist merely of sense-impressions. Not contents alone, but contents\ninterpreted in terms of some specific setting, are the sole possible\nobjects of human thought. Even when, by forced abstraction, particulars\nand universals are held mentally apart, they are still being apprehended\nthrough judgments, and therefore through mental processes that involve\nboth. They stand in relations of mutual implication within a _de facto_\nsystem; and together they constitute it.\n\nThis is the reason why in modern logic, as in Kant\u2019s _Critique_, the\ntheory of the judgment receives so much more attention than the theory\nof reasoning. For once the above view of the judgment has been\nestablished, all the main points in the doctrine of reasoning follow of\nthemselves as so many corollaries. Knowledge starts neither from\nsense-data nor from general principles, but from the complex situation\nin which the human race finds itself at the dawn of self-consciousness.\nThat situation is organised in terms of our mental equipment; and this\nalready existing, rudimentary system is what has made practicable\nfurther advance; to create a system _ab initio_ is altogether\nimpossible. The starting-point does not, however, by itself alone\ndetermine our conclusions. Owing to the creative activities of the mind,\nregulative principles are active in all consciousness; and under their\nguidance the experienced order, largely practical in satisfaction of the\ninstinctive desires, is transformed into a comprehended order,\ncontrolled in view of Ideal ends. Logic is the science of the processes\nwhereby this transformation is brought about. An essentially\nmetaphysical discipline, it cannot be isolated from the general body of\nphilosophical teaching; it is not formal, but transcendental; in\ndefining the factors and processes that constitute knowledge, its chief\npreoccupation is with ultimate issues.\n\nIn calling his new logic \u201ctranscendental\u201d Kant, it is true, also intends\nto signify that it is supplementary to, not a substitute for, the older\nlogic, which he professes to accept.[30] Moreover his intuitional theory\nof mathematical science, his doctrine of the \u201cpure concept,\u201d his\nattributive view of the judgment--all of them survivals from his\npre-Critical period[31]--frequently set him at cross-purposes with\nhimself. His preoccupation, too, with the problem of the _a priori_\nleads him to underestimate the part played in knowledge by the merely\nempirical. But despite all inconsistencies, and notwithstanding his\nperverse preference for outlandish modes of expression, he succeeds in\nenforcing with sufficient clearness the really fundamental tenets of the\nCoherence view.\n\n\n3. THE NATURE OF CONSCIOUSNESS\n\nI shall now approach Kant\u2019s central position from another direction,\nnamely, as an answer to the problem of the nature of consciousness. We\nare justified, I think, in saying that Kant was the first in modern\ntimes to raise the problem of the nature of awareness, and of the\nconditions of its possibility. Though Descartes is constantly speaking\nof consciousness, he defines it in merely negative terms, through its\nopposition to matter; and when he propounds the question how material\nbodies can be known by the immaterial mind, his mode of dealing with it\nshows that his real interest lies not in the nature of consciousness but\nin the character of the existences which it reveals. His answer,\nformulated in terms of the doctrine of representative perception, and\nbased on the supposed teaching of physics and physiology, is that\nmaterial bodies through their action on the sense-organs and brain\ngenerate images or duplicates of themselves. These images, existing not\nin outer space but only in consciousness, are, he asserts, mental in\nnature; and being mental they are, he would seem to conclude,\nimmediately and necessarily apprehended by the mind. Thus Descartes\ngives us, not an analysis of the knowing process, but only a\nsubjectivist interpretation of the nature of the _objects_ upon which it\nis directed.\n\nQuite apart, then, from the question as to whether Descartes\u2019 doctrine\nof representative perception rests on a correct interpretation of the\nteaching of the natural sciences--Kant was ultimately led to reject the\ndoctrine--it is obvious that the main epistemological problem, _i.e._\nthe problem how awareness is possible, and in what it consists, has so\nfar not so much as even been raised. Descartes and his successors\nvirtually assume that consciousness is an ultimate, unanalysable form of\nawareness, and that all that can reasonably be demanded of the\nphilosopher is that he explain what objects are actually presented to\nit, and under what conditions their presentation can occur. On\nDescartes\u2019 view they are conditioned by antecedent physical and\nphysiological processes; according to Berkeley they are due to the\ncreative activity of a Divine Being; according to Hume nothing\nwhatsoever can be determined as to their originating causes. But all\nthree fail to recognise that even granting the objects to be of the\ncharacter asserted, namely, mental, the further problem still remains\nfor consideration, how they come to be consciously apprehended, and in\nwhat such awareness consists.\n\nCertain interpretations of the nature of the knowing process are, of\ncourse, to be found in the writings of Descartes and his successors. But\nthey are so much a matter of unexamined presupposition that they never\nreceive exact formulation, and alternate with one another in quite a\nhaphazard fashion. We may consider three typical views.\n\n1. There is, Descartes frequently seems to imply--the same assumption is\nevident throughout Locke\u2019s _Essay_--a self that stands behind all mental\nstates, observing and apprehending them. Consciousness is the power\nwhich this self has of contemplating both itself and its ideas.\nObviously this is a mere ignoring of the issue. If we assume an\nobserver, we _ipso facto_ postulate a process of observation, but we\nhave not explained or even defined it.\n\n2. There is also in Descartes a second, very different, view of\nconsciousness, namely, as a diaphanous medium analogous to light. Just\nas light is popularly conceived as revealing the objects upon which it\nfalls, so consciousness is regarded as revealing to us our inner states.\nThis view of consciousness, for reasons which I shall indicate shortly,\nis entirely inadequate to the facts for which we have to account. It is\nno more tenable than the corresponding view of light.\n\n3. In Hume we find this latter theory propounded in what may at first\nsight seem a more satisfactory form, but is even less satisfactory.\nSensations, images, feelings, he argues, are _states_ of consciousness,\none might almost say _pieces_ of consciousness, _i.e._ they are\nconceived as carrying their own consciousness with them. Red, for\ninstance, is spoken of as a sensation, and is consequently viewed both\nas being a sense-content, _i.e._ something sensed or apprehended, and\nalso at the same time as the sensing or awareness of it. This view is\nunable to withstand criticism. There is really no more ground for\nasserting that red colour carries with it consciousness of itself than\nfor saying that a table does. The illegitimacy of the assertion is\nconcealed from us by the fact that tables appear to exist when there is\nno consciousness present, whereas redness cannot be proved to exist\nindependently of consciousness--it may or may not do so. Many\npresent-day thinkers, continuing the tradition of the English\nassociationists, hold to this pre-Kantian view. Sensations, feelings,\netc., are, it is implied, pieces of consciousness, forms of awareness;\nthrough their varying combinations they constitute the complex\nexperiences of the animal and human mind.\n\nKant\u2019s teaching is developed in direct opposition to all such views. If\nwe discard his antiquated terminology, and state his position in current\nterms, we find that it amounts to the assertion that _consciousness is\nin all cases awareness of meaning_. There is no awareness, however\nrudimentary or primitive, that does not involve the apprehension of\nmeaning. Meaning and awareness are correlative terms; each must be\nstudied in its relation to the other. And inasmuch as meaning is a\nhighly complex object of apprehension, awareness cannot be regarded as\nultimate or as unanalysable. It can be shown to rest upon a complexity\nof generative conditions and to involve a variety of distinct factors.\n\nThere are thus, from the Kantian standpoint, two all-sufficient reasons\nwhy the diaphanous view of consciousness, _i.e._ any view which treats\nconsciousness merely as a medium whereby the existent gets itself\nreported, must be regarded as untenable. In the first place, as already\nremarked, it is based on the false assumption that consciousness is an\nultimate, and that we are therefore dispensed from all further\ninvestigation of its nature. Kant claims to have distinguished\nsuccessfully the many components which go to constitute it; and he also\nprofesses to have shown that until such analysis has been made, there\ncan be no sufficient basis for a philosophical treatment either of the\nproblems of sense-perception or of the logical problems of judgment and\ninference. The diaphanous view, with its mirror-like mode of\nrepresentation, might allow of the side-by-sideness of associated\ncontents; it can never account for the processes whereby the associated\ncontents come to be apprehended.\n\nSecondly, the diaphanous view ignores the fundamental distinction\nbetween meaning and existence. Existences rest, so to speak, on their\nown bottom; they are self-centred even at the very moment of their\nreaction to external influences. Meaning, on the other hand, always\ninvolves the interpretation of what is given in the light of wider\nconsiderations that lend it significance. In the awareness of meaning\nthe given, the actually presented, is in some way transcended, and this\ntranscendence is what has chiefly to be reckoned with in any attempt to\nexplain the conscious process. Kant is giving expression to this thesis\nwhen he contends that all awareness, no matter how rudimentary or\napparently simple, is an act of judgment, and therefore involves the\nrelational categories. _Not passive contemplation but active judgment,\nnot mere conception but inferential interpretation, is the fundamental\nform, and the only form, in which our consciousness exists._ This, of\ncourse, commits Kant to the assertion that there is no mode of cognition\nthat can be described as immediate or unreflective. There is an\nimmediate _element_ in all knowledge, but our consciousness of it is\nalways conditioned and accompanied by interpretative processes, and in\ntheir absence there can be no awareness of any kind.\n\nBy way of this primary distinction between existence and meaning Kant\nadvances to all those other distinctions which characterise our human\nexperience, between appearance and reality, between the real and the\nIdeal, between that which is judged and the criteria which control and\ndirect the judging process. Just because all awareness is awareness of\nmeaning, our human experience becomes intelligible as a purposive\nactivity that directs itself according to Ideal standards.\n\nThe contrast between the Kantian and the Cartesian views of\nconsciousness can be defined in reference to another important issue.\nThe diaphanous view commits its adherents to a very definite\ninterpretation of the nature of relations. Since they regard\nconsciousness as passive and receptive, they have to maintain that\nrelations can be known only in so far as they are apprehended in a\nmanner analogous to the contents themselves. I do not, of course, wish\nto imply that this view of relational knowledge is in all cases and in\nall respects illegitimate. Kant, as we shall find, has carried the\nopposite view to an impossible extreme, assuming without further\nargument that what has been shown to be true of certain types of\nrelation (for instance, of the causal and substance-attribute\nrelations) must be true of all relations, even of those that constitute\nspace and time. It cannot be denied that, as William James and others\nhave very rightly insisted, such relations as the space-relations are\n_in some degree or manner_ presentational. This does not, however,\njustify James in concluding, as he at times seems inclined to do, that\nall relations are directly experienced. Such procedure lays him open to\nthe same charge of illegitimate reasoning. But even if we could grant\nJames\u2019s thesis in its widest form, the all-important Critical question\nwould still remain: in what does awareness, whether of presented\ncontents or of presented relations, consist, and how is it possible? In\nanswering this question Kant is led to the conclusion that consciousness\nmust be regarded as an activity, and as supplying certain of the\nconditions of its own possibility. Its contribution is of a uniform and\nconstant nature; it consists, as already noted, of certain relational\nfactors whose presence can be detected in each and every act of\nawareness.\n\nThere is one other respect in which Kant\u2019s view of consciousness differs\nfrom that of his Cartesian predecessors.[32] Consciousness, he\nmaintains, does not reveal itself, but only its objects. In other words,\nthere is no awareness of awareness. So far as our mental states and\nprocesses can be known at all, they are known in the same objective\nmanner in which we apprehend existences in space.[33] Now if that be so,\na very important consequence follows. If there is no awareness of\nawareness, but only of meanings all of which are objective, there can be\nno consciousness of the generative, synthetic processes that constitute\nconsciousness _on its subjective side_. For consciousness, being an\n_act_ of awareness in which _meaning_ is apprehended, has a twofold\nnature, and must be very differently described according to the aspect\nwhich at any one time we may have in view. When we regard it on its\n_objective_ side as awareness of _meaning_, we are chiefly concerned\nwith the various factors that are necessary to meaning and that enter\ninto its constitution. That is to say, our analysis is essentially\nlogical. When, on the other hand, we consider consciousness as an _act_\nof awareness, our problem is ontological or as it may be entitled\n(though the term is in this reference somewhat misleading, since the\nenquiry as defined by Kant is essentially metaphysical) psychological in\ncharacter. Between these two aspects there is this very important\ndifference. The logical factors constitutive of meaning can be\nexhaustively known; they are elements in the meanings which\nconsciousness reveals; whereas the synthetic processes are postulated\nsolely in view of these constituent factors, and in order to account for\nthem. The processes, that is to say, are known only through that which\nthey condition, and on Kant\u2019s teaching we are entirely ruled out from\nattempting to comprehend even their possibility.[34] They must be\nthought as occurring, but they cannot be known, _i.e._ their nature\ncannot be definitely specified. The postulating of them marks a gap in\nour knowledge, and extends our insight only in the degree that it\ndiscloses our ignorance. As consciousness rests upon, and is made\npossible by, these processes, it can never be explained in terms of the\nobjective world to which our sense-experience, and therefore, as Kant\nargues, our specific knowledge, is exclusively limited. The mind can\nunfold its contents in the sunshine of consciousness, only because its\nroots strike deep into a soil that the light does not penetrate. These\nprocesses, thus postulated, Kant regards as the source of the _a priori_\nelements, and as the agency through which the synthetic connections\nnecessary to all consciousness are brought about.\n\nAccording to Kant\u2019s Critical teaching, therefore, consciousness, though\nanalysable, is not such as can ever be rendered completely\ncomprehensible. When all is said, it remains for us a merely _de facto_\nform of existence, and has to be taken just for what it presents itself\nas being. It is actually such as to make possible the logical processes\nof judgment and inference. It is actually such as to render possible a\nsatisfactory proof of the scientific validity, within the field of\nsense-experience, of the principle of causality, and of such other\nprinciples as are required in the development of the positive sciences.\nIt is also such as to render comprehensible the controlling influence of\nIdeal standards. But when we come to the question, how is consciousness\nof this type and form possible, that is, to the question of its\nmetaphysical significance and of the generative conditions upon which it\nrests, we find, Kant maintains, that we have no data sufficient to\njustify any decisive answer.\n\nThe ontological, creative, or dynamical aspect of consciousness, I may\nfurther insist, must be constantly borne in mind if the Critical\nstandpoint is to be properly viewed. The logical analysis is, indeed,\nfor the purposes of the central portions of the _Critique_ much the more\nimportant, and alone allows of detailed, exhaustive development; but the\nother is no less essential for an appreciation of Kant\u2019s attitude\ntowards the more strictly metaphysical problems of the _Dialectic_.\n\nHegel and his disciples have been the chief culprits in subordinating,\nor rather in entirely eliminating, this aspect of Kant\u2019s teaching. Many\nof the inconsistencies of which they accuse Kant exist only if Kant\u2019s\nteaching be first reduced to a part of itself. To eliminate the\nontological implications of his theory of consciousness is, by\nanticipation, to render many of his main conclusions entirely untenable,\nand in particular to destroy the force of his fundamental distinction\nbetween appearance and reality. If consciousness knows itself in its\nultimate nature--and such is Hegel\u2019s contention--one half of reality is\ntaken out of the obscurity in which, on Kant\u2019s reading of the situation,\nit is condemned to lie hidden. Man is more knowable than nature, and is\nthe key to nature; such is Hegel\u2019s position, crudely stated. Contrast\ntherewith the teaching of Kant. We can know nature more completely\n(though still very incompletely) than we can ever hope to comprehend the\nconditions that make possible and actual man\u2019s spiritual life. The moral\nconsciousness is an autonomously acting source of independent values,\nand though a standing miracle, must be taken for all that on independent\nand separate enquiry it is found to be. Hegel, in his endeavour to\nestablish an intellectual monism, does violence to some of the highest\ninterests which he professes to be safeguarding. Kant, while outlining\nin Idea a Kingdom of Ends, remains satisfied with a pluralistic\ndistinction between the intellectual and the moral categories. The\nantithesis of the two philosophies is in some degree the ancient\nopposition between Aristotle and Plato, restated in modern terms.\n\n\n4. PHENOMENALISM, KANT\u2019S SUBSTITUTE FOR SUBJECTIVISM\n\nThe revolutionary character of the above conclusions is shown by the\ndifficulty which Kant himself found in breaking away from many of the\npresuppositions that underlie the views which he was renouncing; and\nthis is nowhere more evident than in his constant alternation throughout\nthe _Critique_ between a _subjectivism_[35] that is thoroughly\nCartesian--we might almost, allowing for his rationalism, say\nBerkeleian--in character, and a radically different position which may\nbe entitled _phenomenalism_. The latter is alone genuinely Critical, and\npresents Kant\u2019s teaching in its maturest form. For though first\nformulated only in those portions of the _Analytic_ that are late in\ndate of writing, and in those passages of the second edition which\nsupplement them, it would seem to be the only logical outcome of Kant\u2019s\nother main doctrines.\n\nI have especially in mind Kant\u2019s fundamental distinction between\nappearance and reality; it has an all-important bearing upon the\nCartesian opposition between the mental and the material, and especially\nupon the question as to what view ought to be taken of our so-called\n_subjective_ experiences. The objective is for the Cartesians the\nindependently real; the subjective is asserted to have an altogether\ndifferent kind of existence in what is named the field of consciousness.\nKant\u2019s phenomenalist restatement of this distinction is too complex and\nsubtle to be made intelligible in the brief space available in this\n_Introduction_--it is expounded in the body of the _Commentary_[36]--but\nits general character I may indicate in a few sentences. All\nsubjectivist modes of stating the problem of knowledge, such as we find\nin Hume and in Leibniz no less than in Descartes, Locke, and Berkeley,\nare, Kant finally concluded, illegitimate and question-begging. Our\nso-called subjective states, whether they be sensations, feelings, or\ndesires, are _objective_ in the sense that they are _objects_ for\nconsciousness.[37] Our mental states do not run parallel with the system\nof natural existences; nor are they additional to it. They do not\nconstitute our consciousness of nature; they are themselves part of the\nnatural order which consciousness reveals. They compose the empirical\nself which is an objective existence, integrally connected with the\nmaterial environment in terms of which alone it can be understood. The\nsubjective is not opposite in nature to the objective, but a sub-species\nwithin it. While, however, the psychical is thus to be regarded as a\nclass of known appearances, and as forming together with the physical a\nsingle system of nature, this entire order is, in Kant\u2019s view,\nconditioned by an underlying realm of noumenal existence; and when the\nquestion of the possibility of the _knowing_, that is, of the\n_experiencing_ of such a comprehensive natural system, is raised, it is\nto this noumenal sphere that we are referred. Everything experienced,\neven a sensation or feeling, is an event, but the experiencing of it is\nan act of awareness, and calls for an explanation of an altogether\ndifferent kind.\n\nThus the problem of knowledge, stated in adequate Critical terms, is not\nhow we can advance from the merely subjective to knowledge of the\nindependently real,[38] but how, if everything known forms part of a\ncomprehensive natural system, consciousness and the complex factors\nwhich contribute to its possibility are to be interpreted. On this\nlatter question, as already indicated, Kant, though debarring both\nsubjectivism and materialism, otherwise adopts a non-committal attitude.\nSo long as we continue within the purely theoretical domain, there are a\nnumber of alternatives between which there are no sufficient data for\ndeciding. To debar subjectivism is not to maintain the illusory or\nphenomenal character of the individual self; and to rule out materialism\nis not to assert that the unconscious may not generate and account for\nthe conscious. In other words, they are ruled out not for any ulterior\nreasons derived from their supposed metaphysical consequences, but\nsolely because they are based on palpable misinterpretations of the\ncognitive situation that generates those very problems to which they\nprofess to be an answer.\n\n\n5. THE DISTINCTION BETWEEN HUMAN AND ANIMAL INTELLIGENCE\n\nThe inwardness of Kant\u2019s Critical standpoint may perhaps be made clearer\nby a brief consideration of his view of animal intelligence. We are\naccustomed nowadays to test a psychology of human consciousness by its\ncapacity to render conceivable an evolution from lower forms. How does\nKant\u2019s teaching emerge from such a test?\n\nIt may at once be admitted that Kant has made no special study of animal\nbehaviour, and was by no means competent to speak with authority in\nregard to its conditions. Indeed it is evident that anything which he\nmay have to say upon this question is entirely of the nature of a\ndeduction from results obtained in the human sphere. But when this has\nbeen admitted, and we are therefore prepared to find the problems\napproached from the point of view of the difference rather than of the\nkinship between man and the animals, we can recognise that, so far as\nthe independent study of human consciousness is concerned, there is a\ncertain compensating advantage in Kant\u2019s pre-Darwinian standpoint. For\nit leaves him free from that desire which exercises so constant, and\nfrequently so deleterious an influence, upon many workers in the field\nof psychology, namely, to maintain at all costs, in anticipation of\nconclusions not yet by any means established, the fundamental identity\nof animal and human intelligence. This besetting desire all too easily\ntends to the minimising of differences that may perhaps with fuller\ninsight be found to involve no breach of continuity, but which in the\npresent state of our knowledge cannot profitably be interpreted save in\nterms of their differentiating peculiarities.\n\nThe current controversy between mechanism and vitalism enforces the\npoint which I desire to make. Biological problems, as many biologists\nare now urging, can be most profitably discussed in comparative\nindependence of ultimate issues, entirely in view of their own domestic\ncircumstances. For only when the actual constitution of organic\ncompounds has been more completely determined than has hitherto been\npossible can the broader questions be adequately dealt with. In other\nwords, the differences must be known before the exact nature and degree\nof the continuity can be defined. They cannot be anticipated by any mere\ndeduction from general principles.\n\nThe value of Kant\u2019s analysis of human consciousness is thus closely\nbound up with his frank recognition of its inherent complexity. Not\nsimplification, but specification, down to the bedrock of an irreducible\nminimum of correlated factors, is the governing motive of his Critical\nenquiries. His results have therefore the great advantage of being\ninspired by no considerations save such as are prescribed by the actual\nsubject-matter under investigation. As already noted, Kant maintains\nthat human consciousness is always an awareness of meaning, and that\nconsequently it can find expression only in judgments which involve\ntogether with their other factors the element of recognition or\nself-consciousness.\n\nThis decides for Kant the character of the distinction to be drawn\nbetween animal and human intelligence. As animals, in his view, cannot\nbe regarded as possessing a capacity of self-consciousness, they must\nalso be denied all awareness of meaning. However complicated the\nassociative organisation of their ideas may be, it never rises to the\nhigher level of logical judgment. For the same reason, though their\nideas may be schematic in outline, and in their bearing on behaviour may\ntherefore have the same efficiency as general concepts, they cannot\nbecome universal in the logical sense. \u201cAnimals have apprehensions, but\nnot apperceptions, and cannot, therefore, make their representations\nuniversal.\u201d[39] In support of this position Kant might have pointed to\nthe significant fact that animals are so teachable up to a certain\npoint, and so unteachable beyond it. They can be carried as far as\nassociative suggestion will allow, but not a step further. To this day\nit remains true--at least I venture the assertion--that no animal has\never been conclusively shown to be capable of apprehending a sign as a\nsign. Animals may seem to do so owing to the influence of associated\nideas, but are, as it would appear, debarred from crossing the boundary\nline which so sharply distinguishes associative suggestion from\nreflective knowledge.\n\nBut Kant is committed to a further assertion. If animals are devoid of\nall awareness of meaning, they must also be denied anything analogous to\nwhat we must signify by the term consciousness. Their experience must\nfall apart into events, that may, perhaps, be described as mental, but\ncannot be taken as equivalent to an act of awareness. \u201c_Apprehensio\nbruta_ without consciousness,\u201d[40] such is Kant\u2019s view of the animal\nmind. Its mental states, like all other natural existences, are events\nin time, explicable in the same naturalistic fashion as the bodily\nprocesses by which they are conditioned; they can not be equated with\nthat human consciousness which enables us to reflect upon them, and to\ndetermine the conditions of their temporal happening.\n\nThe distinction which Kant desires to draw is ultimately that between\nevents and consciousness of events. Even if events are psychical in\ncharacter, consisting of sensations and feelings, there will still\nremain as fundamental the distinction between what is simply a member of\nthe causal series of natural events and the consciousness through which\nthe series is apprehended. Kant\u2019s most explicit statements occur in a\nletter to Herz.[41] He is referring to data of the senses which cannot\nbe self-consciously apprehended:\n\n     \u201cI should not be able to know that I have them, and they would\n     therefore be for me, as a cognitive being, absolutely nothing.\n     They might still (if I conceive myself as an animal) exist in me (a\n     being unconscious of my own existence) as representations ...,\n     connected according to an empirical law of association, exercising\n     influence upon feeling and desire, and so always disporting\n     themselves with regularity, without my thereby acquiring the least\n     cognition of anything, not even of these my own states.\u201d[42]\n\nAs to whether Kant is justified in maintaining that the distinction\nbetween animal and human consciousness coincides with the distinction\nbetween associative and logical or reflective thinking, I am not\nconcerned to maintain. This digression has been introduced solely for\nthe purpose of defining more precisely the central tenets of Kant\u2019s\nCritical teaching.\n\n\n6. THE NATURE AND CONDITIONS OF SELF-CONSCIOUSNESS\n\nWe have still to consider what is perhaps the most serious of all the\nmisunderstandings to which Kant has laid himself open, and which is in\nlarge part responsible for the widespread belief that his Critical\nprinciples, when consistently developed, must finally eventuate in some\nsuch metaphysics as that of Fichte and Hegel. I refer to the view that\nKant in postulating synthetic processes as conditioning consciousness is\npostulating a noumenal self as exercising these activities, and is\ntherefore propounding a _metaphysical explanation_ of the synthetic, _a\npriori_ factors in human experience.[43]\n\nKant\u2019s language is frequently ambiguous. The Leibnizian spiritualism, to\nwhich in his pre-Critical period he had unquestioningly held, continued\nto influence his terminology, and so to prevent his Critical principles\nfrom obtaining consistent expression. This much can be said in support\nof the above interpretation of Kant\u2019s position. But in all other\nrespects such a reading of his philosophy is little better than a parody\nof his actual teaching. For Kant is very well aware that the problem of\nknowledge is not to be solved in any such easy and high-handed fashion.\nIn the _Critique_ he teaches quite explicitly that to profess to explain\nthe presence of _a priori_ factors in human experience by means of a\nself assumed for that very purpose would be a flagrant violation, not\nonly of Critical principles, but even of the elementary maxims of\nscientific reasoning. In the first place, explanation by reference to\nthe activities of such a self would be explanation by faculties, by the\nunknown; it is a cause that will explain anything and everything equally\nwell or badly.[44] Self-consciousness has, indeed, to be admitted as a\nfact;[45] and from its occurrence Kant draws important conclusions in\nregard to the conditions which make experience possible. But, in so\ndoing, Kant never intends to maintain that we are justified in\npostulating as part of those conditions, or as condition of those\nconditions, a noumenal self. The conditions which make experience\npossible, whatever they may be, are also the conditions which make\nself-consciousness possible. Since the self is known only as appearance,\nit cannot be asserted to be the conditioning ground of appearance.\n\nThis first objection is not explicitly stated by Kant, but it is implied\nin a second argument which finds expression both in the _Deduction of\nthe Categories_ and in the chapter on the _Paralogisms_. The only self\nthat we know to exist is the _conscious_ self. Now, as Kant claims to\nhave proved, the self can be thus conscious, even of itself, only in so\nfar as it is conscious of objects. Consequently we have no right to\nassume that the self can precede such consciousness as its generating\ncause. That would be to regard the self as existing prior to its own\nconditions, working in darkness to create itself as a source of light.\n\nBut there is also a third reason why Kant\u2019s Critical solution of the\nproblem of knowledge must not be stated in spiritualist terms.\nSelf-consciousness, as he shows, is itself _relational_ in character. It\nis a fundamental factor in human experience, not because the self can\nbe shown to be the agency to which relations are due, but solely\nbecause, itself a case of recognition, it is at the same time a\nnecessary condition of recognition, and recognition is indispensably\npresupposed in all consciousness of meaning.[46] Awareness of meaning is\nthe fundamental mystery, and retains its profoundly mysterious character\neven when self-consciousness has been thus detected as an essential\nconstituent. For self-consciousness does not explain the possibility of\nmeaning; it is itself, as I have just remarked, only one case of\nrecognition, and so is itself only an instance, though indeed the\nsupreme and most important instance, of what we must intend by the term\nmeaning. All awareness, not excepting that of the knowing self, rests\nupon noumenal conditions whose specific nature it does not itself\nreveal. Only on moral grounds, never through any purely theoretical\nanalysis of cognitive experience, can it be proved that the self is an\nabiding personality, and that in conscious, personal form it belongs to\nthe order of noumenal reality.\n\n\n7. KANT\u2019S THREEFOLD DISTINCTION BETWEEN SENSIBILITY, UNDERSTANDING, AND\nREASON\n\nEven so summary a statement of Critical teaching as I am attempting in\nthis _Introduction_ would be very incomplete without some reference to\nKant\u2019s threefold distinction between the forms of sensibility, the\ncategories of the understanding, and the Ideas of Reason.\n\nOn investigating space and time Kant discovers that they cannot be\nclassed either with the data of the bodily senses or with the concepts\nof the understanding. They are sensuous (_i.e._ are not abstract but\nconcrete, not ways of thinking but modes of existence), yet at the same\ntime are _a priori_. They thus stand apart by themselves. Each is unique\nin its kind, is single, and is an infinite existence. To describe them\nis to combine predicates seemingly contradictory. In Kant\u2019s own phrase,\nthey are monstrosities (_Undinge_), none the less incomprehensible that\nthey are undeniably actual. To them, primarily, are due those problems\nwhich have been a standing challenge to philosophy since the time of\nZeno the Eleatic, and which Kant has entitled \u201cantinomies of Reason.\u201d\n\nIn contrast of sensibility Kant sets the intellectual faculties,\nunderstanding and Reason. In the understanding originate certain pure\nconcepts, or as he more usually names them, categories. The chief of\nthese are the categories of \u201crelation\u201d--substance, causality and\nreciprocity. They combine with the forms of sensibility and the manifold\nof sense to yield the consciousness of an empirical order, interpretable\nin accordance with universal laws.\n\nTo the faculty of Reason Kant ascribes what he entitles Ideas. The Ideas\ndiffer from space, time, and the categories in being not \u201cconstitutive\u201d\nbut \u201cregulative.\u201d They demand an _unconditionedness_ of existence and a\n_completeness_ of explanation which can never be found in actual\nexperience. Their function is threefold. In the first place, they render\nthe mind dissatisfied with the haphazard collocations of ordinary\nexperience, and define the goal for its scientific endeavours. Secondly,\nthey determine for us the criteria that distinguish between truth and\nfalsity.[47] And thirdly, in so doing, they likewise make possible the\ndistinction between appearance and reality, revealing to us an\nirreconcilable conflict between the ultimate aims of science and the\nhuman conditions, especially the spatial and temporal conditions under\nwhich these aims are realised. The Ideas of Reason are the second main\nfactor in the \u201cantinomies.\u201d\n\nThe problem of the _Critique_, the analysis of our awareness of meaning,\nis a single problem, and each of the above elements involves all the\nothers. Kant, however, for reasons into which I need not here enter, has\nassigned part of the problem to what he entitles the _Transcendental\nAesthetic_, and another part to the _Transcendental Dialectic_. Only\nwhat remains is dealt with in what is really the most important of the\nthree divisions, the _Transcendental Analytic_. But as the problem is\none and indivisible, the discussions in all three sections are condemned\nto incompleteness save in so far as Kant, by happy inconsistency,\ntransgresses the limits imposed by his method of treatment. The\n_Aesthetic_ really does no more than prepare the ground for the more\nadequate analysis of space and time given in the _Analytic_ and\n_Dialectic_, while the problem of the _Analytic_ is itself incompletely\nstated until the more comprehensive argument of the _Dialectic_ is taken\ninto account.[48] Thus the statement in the _Aesthetic_ that space and\ntime are _given_ to the mind by the sensuous faculty of receptivity is\nmodified in the _Analytic_ through recognition of the part which the\nsyntheses and concepts of the understanding must play in the\nconstruction of these forms; and in the _Dialectic_ their apprehension\nis further found to involve an Idea of Reason. Similarly, in the\nconcluding chapter of the _Analytic_, in discussing the grounds for\ndistinguishing between appearance and reality, Kant omits all reference\nto certain important considerations which first emerge into view in the\ncourse of the _Dialectic_. Yet, though no question is more vital to\nCritical teaching, the reader is left under the impression that the\ntreatment given in the _Analytic_ is complete and final.\n\nPartly as a consequence of this, partly owing to Kant\u2019s inconsistent\nretention of earlier modes of thinking, there are traceable throughout\nthe _Critique_ two opposed views of the nature of the distinction\nbetween appearance and reality. On the one view, this distinction is\nmediated by the relational categories of the understanding, especially\nby that of causality; on the other view, it is grounded in the Ideas of\nReason. The former sets appearance in opposition to reality; the latter\nregards the distinction in a more tenable fashion, as being between\nrealities less and more comprehensively conceived.[49]\n\nA similar defect is caused by Kant\u2019s isolation of immanent from\ntranscendent metaphysics.[50] The former is dealt with only in the\n_Analytic_, the latter only in the _Dialectic_. The former, Kant\nasserts, is made possible by the forms of sensibility and the categories\nof the understanding; the latter he traces to an illegitimate employment\nof the Ideas of Reason. Such a mode of statement itself reveals the\nimpossibility of any sharp distinction between the immanent and the\ntranscendent. If science is conditioned by Ideals which arouse the mind\nto further acquisitions, and at the same time reveal the limitations to\nwhich our knowledge is for ever condemned to remain subject; if, in\nother words, everything known, in being correctly known, must be\napprehended as appearance (_i.e._ as a subordinate existence within a\nmore comprehensive reality), the distinction between the immanent and\nthe transcendent falls within and not beyond the domain of our total\nexperience. The meaning which our consciousness discloses in each of its\njudgments is an essentially metaphysical one. It involves the thought,\nthough not the knowledge, of something more than what the experienced\ncan ever itself be found to be. The metaphysical is immanent in our\nknowledge; the transcendent is merely a name for this immanent factor\nwhen it is falsely viewed as capable of isolation and of independent\ntreatment. By Kant\u2019s own showing, the task of the _Dialectic_ is not\nmerely to refute the pretensions of transcendent metaphysics, but to\ndevelop the above general thesis, in confirmation of the positive\nconclusions established in the _Analytic_. The _Critique_ will then\nsupply the remedy for certain evils to which the human mind has hitherto\nbeen subject.\n\n     \u201c_The Critique of Pure Reason_ is a preservative against a malady\n     which has its source in our rational nature. This malady is the\n     opposite of the love of home (the home-sickness) which binds us to\n     our fatherland. It is a longing to pass out beyond our immediate\n     confines and to relate ourselves to other worlds.\u201d[51]\n\n\n8. THE PLACE OF THE _CRITIQUE OF PURE REASON_ IN KANT\u2019S PHILOSOPHICAL\nSYSTEM\n\nThe positive character of Kant\u2019s conclusions cannot be properly\nappreciated save in the wider perspectives that open to view in the\n_Critique of Practical Reason_ and in the _Critique of Judgment_. Though\nin the _Critique of Pure Reason_ a distinction is drawn between\ntheoretical and moral belief, it is introduced in a somewhat casual\nmanner, and there is no clear indication of the far-reaching\nconsequences that follow in its train. Unfortunately also, even in his\nlater writings, Kant is very unfair to himself in his methods of\nformulating the distinction. His real intention is to show that\nscientific knowledge is not coextensive with human insight; but he\nemploys a misleading terminology, contrasting knowledge with faith,\nscientific demonstration with practical belief.\n\nAs already indicated, the term knowledge has, in the Critical\nphilosophy, a much narrower connotation than in current speech. It is\nlimited to _sense_-experience, and to such inferences therefrom as can\nbe obtained by the only methods that Kant is willing to recognise,\nnamely, the mathematico-physical. Aesthetic, moral and religious\nexperience, and even organic phenomena, are excluded from the field of\npossible knowledge.\n\nIn holding to this position, Kant is, of course, the child of his time.\nThe absolute sufficiency of the Newtonian physics is a presupposition of\nall his utterances on this theme. Newton, he believes, has determined in\na quite final manner the principles, methods and limits of scientific\ninvestigation. For though Kant himself imposes upon science a further\nlimitation, namely, to appearances, he conceives himself, in so doing,\nnot as weakening Newton\u2019s natural philosophy, but as securing it against\nall possible objections. And to balance the narrow connotation thus\nassigned to the term knowledge, he has to give a correspondingly wide\nmeaning to the terms faith, moral belief, subjective principles of\ninterpretation. If this be not kept constantly in mind, the reader is\ncertain to misconstrue the character and tendencies of Kant\u2019s actual\nteaching.\n\nBut though the advances made by the sciences since Kant\u2019s time have\nrendered this mode of delimiting the field of knowledge altogether\nuntenable, his method of defining the sources of _philosophical insight_\nhas proved very fruitful, and has many adherents at the present day.\nWhat Kant does--stated in broad outline--is to distinguish between the\nproblems of _existence_ and the problems of _value_, assigning the\nformer to science and the latter to philosophy.[52] Theoretical\nphilosophy, represented in his system by the _Critique of Pure Reason_,\ntakes as its province the logical values, that is, the distinction of\ntruth and falsity, and defining their criteria determines the nature and\nlimits of our theoretical insight. Kant finds that these criteria enable\nus to distinguish between truth and falsity only on the empirical plane.\nBeyond making possible a distinction between appearance and reality,\nthey have no applicability in the metaphysical sphere.\n\nThe _Critique of Practical Reason_ deals with values of a very different\ncharacter. The faculty of Reason, which, as already noted,[53] renders\nour consciousness a purposive agency controlled by Ideal standards, is\nalso, Kant maintains, the source of the moral sanctions. But whereas in\nthe theoretical field it subdues our minds to the discipline of\nexperience, and restrains our intellectual ambitions within the limits\nof the empirical order, it here summons us to sacrifice every natural\nimpulse and every secular advantage to the furtherance of an end that\nhas absolute value. In imposing duties, it raises our life from the\n\u201cpragmatic\u201d[54] level of a calculating expediency to the higher plane of\na categorical imperative.\n\nThe categorical imperative at once humbles and exalts; it discloses our\nlimitations, but does so through the greatness of the vocation to which\nit calls us.\n\n     \u201cThis principle of morality, just on account of the universality of\n     the legislation which makes it the formal supreme determining\n     principle of our will, without regard to any subjective\n     differences, is declared by the Reason to be a law for all\n     rational beings.... It is, therefore, not limited to men only, but\n     applies to all finite beings that possess Reason and Will; nay, it\n     even includes the Infinite Being as the Supreme Intelligence.\u201d[55]\n\nConsequently, in employing moral ends in the interpretation of the\nUniverse, we are not picturing the Divine under human limitations, but\nare discounting these limitations in the light of the one form of value\nthat is known to us as absolute.\n\n     \u201c_Duty!_ ... What origin is worthy of thee and where is to be found\n     the root of thy noble descent ... a root to be derived from which\n     is the indispensable condition of the only worth that men can give\n     themselves.\u201d[56]\n\nIn his earlier years Kant had accepted the current, Leibnizian view that\nhuman excellence consists in intellectual enlightenment, and that it is\ntherefore reserved for an _\u00e9lite_, privileged with the leisure and\nendowed with the special abilities required for its enjoyment. From this\narid intellectualism he was delivered through the influence of Rousseau.\n\n     \u201cI am by disposition an enquirer. I feel the consuming thirst for\n     knowledge, the eager unrest to advance ever further, and the\n     delights of discovery. There was a time when I believed that this\n     is what confers real dignity upon human life, and I despised the\n     common people who know nothing. Rousseau has set me right. This\n     imagined advantage vanishes. I learn to honour men, and should\n     regard myself as of much less use than the common labourer, if I\n     did not believe that my philosophy will restore to all men the\n     common rights of humanity.\u201d[57]\n\nThese common rights Kant formulates in a purely individualist manner.\nFor here also, in his lack of historic sense and in his distrust alike\nof priests and of statesmen, he is the child of his time. In the\neducation and discipline of the soul he looks to nothing so artificial\nand humanly limited--Kant so regards them--as religious tradition and\nsocial institutions. Human rights, he believes, do not vary with time\nand place; and for their enjoyment man requires no initiation and no\nequipment beyond what is supplied by Nature herself. It is from this\nstandpoint that Kant adduces, as the twofold and sufficient inspiration\nto the rigours and sublimities of the spiritual life, the starry heavens\nabove us and the moral law within. They are ever-present influences on\nthe life of man. The naked eye reveals the former; of the latter all men\nare immediately aware. In their universal appeal they are of the very\nsubstance of human existence. Philosophy may avail to counteract certain\nof the hindrances which prevent them from exercising their native\ninfluence; it cannot be a substitute for the inspiration which they\nalone can yield.\n\nThus the categorical imperative, in endowing the human soul with an\nintrinsic value, singles it out from all other natural existences, and\nstrengthens it to face, with equanimity, the cold immensities of the\ncosmic system. For though the heavens arouse in us a painful feeling of\nour insignificance as animal existences, they intensify our\nconsciousness of a sublime destiny, as bearers of a rival, and indeed a\nsuperior, dignity.\n\nIn one fundamental respect Kant broke with the teaching of Rousseau,\nnamely, in questioning his doctrine of the natural goodness and\nindefinite perfectibility of human nature.[58] Nothing, Kant maintains,\nis good without qualification except the good will; and even that,\nperhaps, is never completely attained in any single instance. The\nexercise of duty demands a perpetual vigilance, under the ever-present\nconsciousness of continuing demerit.\n\n     \u201cI am willing to admit out of love of humanity that most of our\n     actions are indeed correct, but if we examine them more closely we\n     everywhere come upon the dear self which is always\n     prominent....\u201d[59] \u201cNothing but moral fanaticism and exaggerated\n     self-conceit is infused into the mind by exhortation to actions as\n     noble, sublime and magnanimous. Thereby men are led into the\n     delusion that it is not duty, that is, respect for the law, whose\n     yoke ... they must bear, whether they like it or not, that\n     constitutes the determining principle of their actions, and which\n     always humbles them while they _obey_ it. They then fancy that\n     those actions are expected from them, not from duty, but as pure\n     merit.... In this way they engender a vain high-flying fantastic\n     way of thinking, flattering themselves with a spontaneous goodness\n     of heart that needs neither spur nor bridle, nor any\n     command....\u201d[60]\n\nIn asserting the goodness and self-sufficiency of our natural impulses\nRousseau is the spokesman of a philosophy which has dominated social and\npolitical theory since his day, and which is still prevalent. This\nphilosophy, in Kant\u2019s view, is disastrous in its consequences. As a\nreading of human nature and of our moral vocation, it is hardly less\nfalse than the Epicurean teaching, which finds in the pursuit of\npleasure the motive of all our actions. A naturalistic ethics, in either\nform, is incapacitated, by the very nature of its controlling\nassumptions, from appreciating the distinguishing features of the moral\nconsciousness. Neither the successes nor the failures of man\u2019s spiritual\nendeavour can be rightly understood from any such standpoint. The human\nrace, in its endurance and tenacity, in its dauntless courage and in its\nsoaring spirit, reveals the presence of a _prevenient_ influence,\n_non-natural_ in character; and only if human nature be taken as\nincluding this higher, directive power, can it assume to itself the\neulogy which Rousseau so mistakenly passes upon the natural and\nundisciplined tendencies of the human heart. For as history\ndemonstrates, while _men_ are weak, _humanity_ is marvellous.\n\n     \u201cThere is one thing in our soul which, when we take a right view of\n     it, we cannot cease to regard with the highest astonishment, and in\n     regard to which admiration is right and indeed elevating, and that\n     is our original moral capacity in general.... Even the\n     incomprehensibility of this capacity,[61] a capacity which\n     proclaims a Divine origin, must rouse man\u2019s spirit to enthusiasm\n     and strengthen it for any sacrifices which respect for his duty may\n     impose on him.\u201d[62]\n\nWe are not here concerned with the detail of Kant\u2019s ethical teaching, or\nwith the manner in which he establishes the freedom of the will, and\njustifies belief in the existence of God and the immortality of the\nsoul. In many respects his argument lies open to criticism. There is an\nunhappy contrast between the largeness of his fundamental thesis and\nthe formal, doctrinaire manner in which it is developed. Indeed, in the\n_Critique of Practical Reason_ the individualist, deistic, rationalistic\nmodes of thinking of his time are much more in evidence than in any\nother of his chief writings; and incidentally he also displays a curious\ninsensibility--again characteristic of his period--to all that is\nspecific in the religious attitude. But when due allowances have been\nmade, we can still maintain that in resting his constructive views upon\nthe supreme value of the moral personality Kant has influenced\nsubsequent philosophy in hardly less degree than by his teaching in the\n_Critique of Pure Reason_.[63]\n\nThe two _Critiques_, in method of exposition and argument, in general\noutcome, and indeed in the total impression they leave upon the mind,\nare extraordinarily different. In the _Critique of Pure Reason_ Kant is\nmeticulously scrupulous in testing the validity of each link in his\nargument. Constantly he retraces his steps; and in many of his chief\nproblems he halts between competing solutions. Kant\u2019s sceptical spirit\nis awake, and it refuses to cease from its questionings. In the\n_Critique of Practical Reason_, on the other hand, there is an austere\nsimplicity of argument, which advances, without looking to right or\nleft, from a few simple principles direct to their ultimate\nconsequences. The impressiveness of the first _Critique_ consists in its\nappreciation of the _complexity_ of the problems, and in the care with\nwhich their various, conflicting aspects are separately dealt with. The\nsecond _Critique_ derives its force from the fundamental conviction upon\nwhich it is based.\n\nSuch, then, stated in the most general terms, is the manner in which\nKant conceives the _Critique of Pure Reason_ as contributing to the\nestablishment of a humanistic philosophy. It clears the ground for the\npractical Reason, and secures it in the autonomous control of its own\ndomain. While preserving to the intellect and to science certain\ndefinitely prescribed rights, Kant places in the forefront of his system\nthe moral values; and he does so under the conviction that in living up\nto the opportunities, in whatever rank of life, of our common heritage,\nwe obtain a truer and deeper insight into ultimate issues than can be\nacquired through the abstruse subtleties of metaphysical speculation.\n\nI may again draw attention to the consequences which follow from Kant\u2019s\nhabitual method of isolating his problems. Truth is a value of universal\njurisdiction, and from its criteria the judgments of moral and other\nvalues can claim no exemption. Existences and values do not constitute\nindependent orders. They interpenetrate, and neither can be adequately\ndealt with apart from the considerations appropriate to the other. In\nfailing to co-ordinate his problems, Kant has over-emphasised the\nnegative aspects of his logical enquiries and has formulated his ethical\ndoctrines in a needlessly dogmatic form.\n\nThese defects are, however, in some degree remedied in the last of his\nchief works, the _Critique of Judgment_. In certain respects it is the\nmost interesting of all Kant\u2019s writings. The qualities of both the\nearlier _Critiques_ here appear in happy combination, while in addition\nhis concrete interests are more in evidence, to the great enrichment of\nhis abstract argument. Many of the doctrines of the _Critique of Pure\nReason_, especially those that bear on the problems of teleology, are\nrestated in a less negative manner, and in their connection with the\nkindred problems of natural beauty and the fine arts. For though the\nfinal decision in all metaphysical questions is still reserved to moral\nconsiderations, Kant now takes a more catholic view of the field of\nphilosophy. He allows, though with characteristic reservations, that the\n_empirical_ evidence obtainable through examination of the broader\nfeatures of our total experience is of genuinely philosophical value,\nand that it can safely be employed to amplify and confirm the\nindependent convictions of the moral consciousness. The embargo which in\nthe _Critique of Pure Reason_, in matters metaphysical, is placed upon\nall tentative and probable reasoning is thus tacitly removed; and the\nterm knowledge again acquires the wider meaning very properly ascribed\nto it in ordinary speech.\n\n\n\n\nA COMMENTARY TO KANT\u2019S \u201cCRITIQUE OF PURE REASON\u201d\n\n\nTITLE: KRITIK DER REINEN VERNUNFT\n\n\nThe term critique or criticism, as employed by Kant, is of English\norigin. It appears in seventeenth and eighteenth century English,\nchiefly in adjectival form, as a literary and artistic term--for\ninstance, in the works of Pope, who was Kant\u2019s favourite English poet.\nKant was the first to employ it in German, extending it from the field\nof aesthetics to that of general philosophy. A reference in Kant\u2019s\n_Logic_[64] to Home\u2019s _Elements of Criticism_[65] would seem to indicate\nthat it was Home\u2019s use of the term which suggested to him its wider\nemployment. \u201cCritique of pure reason,\u201d in its primary meaning, signifies\nthe passing of critical judgments upon pure reason. In this sense Kant\nspeaks of his time as \u201cthe age of criticism (_Zeitalter der Kritik_).\u201d\nFrequently, however, he takes the term more specifically as meaning a\ncritical investigation leading to positive as well as to negative\nresults. Occasionally, especially in the _Dialectic_, it also signifies\na discipline applied to pure reason, limiting it within due bounds. The\nfirst appearance of the word in Kant\u2019s writings is in 1765 in the\n_Nachricht_[66] of his lectures for the winter term 1765-1766. Kant\nseldom employs the corresponding adjective, critical (_kritisch_). His\nusual substitute for it is the term transcendental.\n\n_Pure_ (_rein_) has here a very definite meaning. It is the absolutely\n_a priori_. Negatively it signifies that which is independent of\nexperience. Positively it signifies that which originates from reason\nitself, and which is characterised by universality and necessity.[67] By\n\u201cpure reason\u201d Kant therefore means reason in so far as it supplies out\nof itself, independently of experience, _a priori_ elements that as such\nare characterised by universality and necessity.\n\n_Reason_ (_Vernunft_) is used in the _Critique_ in three different\nmeanings. In the above title it is employed in its widest sense, as the\nsource of all _a priori_ elements. It includes what is _a priori_ in\nsensibility as well as in understanding (_Verstand_). In its narrowest\nsense it is distinct even from understanding, and signifies that faculty\nwhich renders the mind dissatisfied with its ordinary and scientific\nknowledge, and which leads it to demand a completeness and\nunconditionedness which can never be found in the empirical sphere.\nUnderstanding conditions science; reason generates metaphysic.\nUnderstanding has categories; reason has its Ideas. Thirdly, Kant\nfrequently employs understanding and reason as synonymous terms,\ndividing the mind only into the two faculties, sensibility and\nspontaneity. Thus in A 1-2, understanding and reason are used\npromiscuously, and in place of _reine Vernunft_ we find _reiner\nVerstand_. As already stated, the term reason, as employed in Kant\u2019s\ntitle, ought properly to be taken in its widest sense. Sensibility falls\nwithin reason in virtue of the _a priori_ forms which it contains. Kant\ndoes not himself, however, always interpret the title in this strict\nsense. The triple use of the term is an excellent example of the\nlooseness and carelessness with which he employs even the most important\nand fundamental of his technical terms. Only the context can reveal the\nparticular meaning to be assigned in each case.\n\nThe phrase \u201c_of_ pure reason\u201d (_der reinen Vernunft_) has, as Vaihinger\npoints out,[68] a threefold ambiguity. (1) Sometimes it is a genitive\nobjective. The critical enquiry is directed upon pure reason as its\nobject. This corresponds to the view of the _Critique_ as merely a\ntreatise on method. (2) Sometimes it is a genitive subjective. The\ncritical enquiry is undertaken by and executed through pure reason. This\nexpresses the view of the _Critique_ as itself a system of pure rational\nknowledge. (3) At other times it has a reflexive meaning. Pure reason is\nsubject and object at once. It is both subject-matter and method or\ninstrument. Through the _Critique_ it attains to self-knowledge. The\n_Critique_ is the critical examination of pure reason by itself. The\nfirst view would seem to be the original and primary meaning of the\ntitle. The second view very early took its place alongside it, and\nappears in many passages. The third view must be taken as representing\nKant\u2019s final interpretation of the title; it is on the whole the most\nadequate to the actual content and scope of the _Critique_. For the\n_Critique_ is not merely a treatise on method; it is also a system of\npure rational knowledge. It professes to establish, in an exhaustive and\nfinal manner, the _a priori_ principles which determine the possibility,\nconditions, and limits of pure rational knowledge.[69]\n\n\n\n\nMOTTO\n\n     De nobis ipsis silemus: De re autem, quae agitur, petimus: ut\n     homines eam non opinionem, sed opus esse cogitent; ac pro certo\n     habeant, non sectae nos alicuius, aut placiti, sed utilitatis et\n     amplitudinis humanae fundamenta moliri. Deinde ut suis commodis\n     aequi ... in commune consulant ... et ipsi in partem veniant.\n     Praeterea ut bene sperent, neque instaurationem nostram ut quiddam\n     infinitum et ultra mortale fingant, et animo concipiant; quum\n     revera sit infiniti erroris finis et terminus legitimus.\n\n\nThis motto, which was added in the second edition, is taken from the\npreface to Bacon\u2019s _Instauratio Magna_, of which the _Novum Organum_ is\nthe second part. As the first part of the _Instauratio_ is represented\nonly by the later, separately published, _De Augmentis Scientiarum_,\nthis preface originally appeared, and is still usually given, as\nintroductory to the _Novum Organum_.\n\nThe complete passage (in which I have indicated Kant\u2019s omissions) is\nrendered as follows in the translation of Ellis and Spedding:[70]\n\n     \u201cOf myself I say nothing; but in behalf of the business which is in\n     hand I entreat men to believe that it is not an opinion to be held,\n     but a work to be done; and to be well assured that I am labouring\n     to lay the foundation, not of any sect or doctrine, but of human\n     utility and power. Next, I ask them to deal fairly by their own\n     interests [and laying aside all emulations and prejudices in favour\n     of this or that opinion], to join in consultation for the common\n     good; and [being now freed and guarded by the securities and helps\n     which I offer from the errors and impediments of the way] to come\n     forward themselves and take part [in that which remains to be\n     done]. Moreover, to be of good hope, nor to imagine that this\n     Instauration of mine is a thing infinite and beyond the power of\n     man, when it is in fact the true end and termination of infinite\n     error.\u201d\n\nThe opening sentence of Bacon\u2019s preface might also have served as a\nfitting motto to the _Critique_:\n\n     \u201cIt seems to me that men do not rightly understand either their\n     store or their strength, but overrate the one and underrate the\n     other.\u201d\n\nOr again the following:\n\n     \u201cI have not sought nor do I seek either to enforce or to ensnare\n     men\u2019s judgments, but I lead them to things themselves and the\n     concordances of things, that they may see for themselves what they\n     have, what they can dispute, what they can add and contribute to\n     the common stock.... And by these means I suppose that I have\n     established for ever a true and lawful marriage between the\n     empirical and the rational faculty, the unkind and ill-starred\n     divorce and separation of which has thrown into confusion all the\n     affairs of the human family.\u201d\n\n\nDEDICATION\n\nTO\n\nFREIHERR VON ZEDLITZ\n\nKarl Abraham, Freiherr von Zedlitz had been entrusted, as Minister\n(1771-1788) to Frederick the Great, with the oversight and direction of\nthe Prussian system of education. He held Kant in the highest\nesteem.[71] In February 1778 we find him writing to thank Kant for the\npleasure he had found in perusing notes of his lectures on physical\ngeography, and requesting the favour of a complete copy.[72] A week\nlater he invited Kant to accept a professorship of philosophy in\nHalle,[73] which was then much the most important university centre in\nGermany. Upon Kant\u2019s refusal he repeated the offer, with added\ninducements, including the title of Hofrat.[74] Again, in August of the\nsame year, he writes that he is attending, upon Mendelssohn\u2019s\nrecommendation (and doubtless also in the hope of receiving from this\nindirect source further light upon Kant\u2019s own teaching in a favourite\nfield), the lectures on anthropology of Kant\u2019s disciple and friend,\nMarcus Herz. The letter concludes with a passage which may perhaps have\nsuggested to Kant the appropriateness of dedicating his _Critique_ to so\nwise and discerning a patron of true philosophy.\n\n     \u201cShould your inventive power extend so far, suggest to me the means\n     of holding back the students in the universities from the bread and\n     butter studies, and of making them understand that their modicum of\n     law, even their theology and medicine, will be immensely more\n     easily acquired and safely applied, if they are in possession of\n     more philosophical knowledge. They can be judges, advocates,\n     preachers and physicians only for a few hours each day; but in\n     these and all the remainder of the day they are men, and have need\n     of other sciences. In short, you must instruct me how this is to be\n     brought home to students. Printed injunctions, laws,\n     regulations--these are even worse than bread and butter study\n     itself.\u201d[75]\n\nA Minister of Education who thus ranks philosophy above professional\nstudies, and both as more important than all academic machinery, holds\nhis office by divine right.\n\n\n\n\nPREFACE TO THE FIRST EDITION\n\n\nDetailed discussion of the _Prefaces_ is not advisable. The problems\nwhich they raise can best be treated in the order in which they come up\nin the _Critique_ itself. I shall dwell only on the minor incidental\ndifficulties of the text, and on those features in Kant\u2019s exposition\nwhich are peculiar to the _Prefaces_, or which seem helpful in the way\nof preliminary orientation. I shall first briefly restate the argument\nof the _Preface_ to the first edition, and then add the necessary\ncomment.\n\nHuman reason is ineradicably metaphysical. It is haunted by questions\nwhich, though springing from its very nature, none the less transcend\nits powers. Such a principle, for instance, as that of causality, in\ncarrying us to more and more remote conditions, forces us to realise\nthat by such regress our questions can never be answered. However far we\nrecede in time, and however far we proceed in space, we are still no\nnearer to a final answer to our initial problems, and are therefore\ncompelled to take refuge in postulates of a different kind, such, for\ninstance, as that there must be a first unconditioned cause from which\nthe empirical series of causes and effects starts, or that space is\ncapable of existing as a completed whole. But these assumptions plunge\nreason in darkness and involve it in contradictions. They are the\nsources of all the troubles of the warring schools. Error lies somewhere\nconcealed in them--the more thoroughly concealed that they surpass the\nlimits of possible experience. Until such error has been detected and\nlaid bare, metaphysical speculation must remain the idlest of all tasks.\n\nIn the latter part of the eighteenth century metaphysics had fallen, as\nKant here states, into disrepute. The wonderful success with which the\nmathematical and natural sciences were being developed served only to\nemphasise by contrast the ineffectiveness of the metaphysical\ndisciplines. Indifference to philosophy was the inevitable outcome, and\nwas due, not to levity, but to the matured judgment of the age, which\nrefused to be any longer put off with such pretended knowledge. But\nsince the philosophical sciences aim at that knowledge which, if\nattainable, we should be least willing to dispense with, the failure of\nphilosophy is really a summons to reason to take up anew the most\ndifficult of all its tasks. It must once and for all determine either\nthe possibility or the impossibility of metaphysics. It must establish\n\n     \u201c...a tribunal which will assure to reason its lawful claims, and\n     which will also be able to dismiss all groundless pretensions, not\n     by despotic decrees, but in accordance with its own eternal and\n     unalterable laws. This tribunal is no other than the _Critique of\n     Pure Reason_.\u201d[76] \u201cOur age is, in especial degree, the age of\n     criticism (_Kritik_), and to such criticism everything must submit.\n     Religion, through its sanctity, and law-giving, through its\n     majesty, may seek to exempt themselves from it. But they then\n     awaken just suspicion, and cannot claim the sincere respect which\n     reason accords only to that which has been able to sustain the test\n     of free and open examination.\u201d[77]\n\nAs has already been emphasised in the preceding historical sketch, Kant\nhad learnt to trust the use of reason, and was a rationalist by\neducation, temperament, and conviction. He here classifies philosophies\nas dogmatic and sceptical; and under the latter rubric he includes all\nempirical systems. \u2018Empiricism\u2019 and \u2018scepticism\u2019 he interprets as\npractically synonymous terms. The defect of the dogmatists is that they\nhave not critically examined their methods of procedure, and in the\nabsence of an adequate distinction between appearance and reality have\ninterpreted the latter in terms of the former. The defect of the\nempiricists and sceptics is that they have misrepresented the nature of\nthe faculty of reason, ignoring its claims and misreading its functions,\nand accordingly have gone even further astray than their dogmatic\nopponents. All knowledge worthy of the name is _a priori_ knowledge. It\npossesses universality and necessity, and as such must rest on pure\nreason. Wherever there is science, there is an element of pure reason.\nWhether or not pure reason can also extend to the unconditioned is the\nquestion which decides the possibility of constructive metaphysics. This\nis what Kant means when he declares that the _Critique_ is a criticism\nof the power of reason, in respect of all knowledge after which it may\nstrive independently of experience. Pure reason is the subject-matter of\nthe enquiry; it is also the instrument through which the enquiry is\nmade.[78] Nothing empirical or merely hypothetical has any place in it,\neither as subject-matter or as method of argument.\n\nFrom this position Kant draws several important consequences. First,\nsince pure reason means that faculty whereby we gain knowledge\nindependently of all experience, it can be isolated and its whole nature\nexhaustively determined. Indeed pure reason (Kant seeks to prove) is so\nperfect a unity that if \u201cits principle\u201d should be found insufficient to\nthe solution of a single one of all the questions which are presented to\nit by its own nature, we should be justified in forthwith rejecting it\nas also incompetent to answer with complete certainty any one of the\nother questions. In metaphysics it must be either all or nothing,[79]\neither final and complete certainty or else absolute failure.\n\n     \u201cWhile I am saying this I can fancy that I detect in the face of\n     the reader an expression of indignation mingled with contempt at\n     pretensions seemingly so arrogant and vainglorious; and yet they\n     are incomparably more moderate than the claims of all those writers\n     who on the lines of the usual programme profess to prove the simple\n     nature of the soul or the necessity of a first beginning of the\n     world.\u201d[80]\n\nIn so doing they pretend to define realities which lie beyond the limits\nof possible experience; the _Critique_ seeks only to deal with that\nfaculty of reason which manifests itself to us within our own minds.\nFormal logic shows how completely and systematically the simple acts of\nreason can be enumerated. Aristotle created this science of logic\ncomplete at a stroke. Kant professes to have established an equally\nfinal metaphysics; and as logic is not a science proper, but rather a\npropaedeutic to all science, metaphysics, thus interpreted, is the only\none of all the sciences which can immediately attain to such\ncompleteness.\n\n     \u201cFor it is nothing but the inventory of all our possessions through\n     pure reason, systematically arranged. In this field nothing can\n     escape us. What reason produces entirely out of itself cannot lie\n     concealed, but is brought to light by reason itself immediately the\n     common principle has been discovered.\u201d[81]\n\nSecondly, the _Critique_ also claims certainty. With the removal of\neverything empirical, and the reduction of its subject-matter to pure\nreason, all mere opinion or hypothesis is likewise eliminated.\nProbabilities or hypotheses can have no place in a _Critique of Pure\nReason_.[82] Everything must be derived according to _a priori_\nprinciples from pure conceptions in which there is no intermixture of\nexperience or any special intuition.\n\n       *       *       *       *       *\n\nThis _Preface_ to the first edition, considered as introductory to the\n_Critique_, is misleading for two reasons. First, because in it Kant is\npreoccupied almost exclusively with the problems of metaphysics in the\nstrict ontological sense, that is to say, with the problems of the\n_Dialectic_. The problems of the _Analytic_, which is the very heart of\nthe _Critique_, are almost entirely ignored. They are, it is true,\nreferred to in A x-xi, but the citation is quite externally\nintercalated; it receives no support or extension from the other parts\nof the _Preface_. This results in a second defect, namely, that Kant\nfails to indicate the more empirical features of his new Critical\nstandpoint. Since ultimate reality is supersensuous, metaphysics, as\nabove conceived, can have no instrument save pure reason. The subjects\nof its enquiry, God, freedom, and immortality, if they are to be known\nat all, can be determined only through _a priori_ speculation. This\nfact, fundamental and all-important for Kant, was completely ignored in\nthe popular eclectic philosophies of the time. They professed to derive\nmetaphysical conclusions from empirical evidence. They substituted, as\nKant has pointed out,[83] \u201ca physiology of the human understanding\u201d for\nthe Critical investigation of the claims of reason, and anthropology for\nethics. They were blind to the dogmatism of which they are thereby\nguilty. They assumed those very points which most call for proof,\nnamely, that reason is adequate to the solution of metaphysical\nproblems, and that all existence is so fundamentally of one type that we\ncan argue from the sensuous to the supersensuous, from appearance to\nreality. When they fell into difficulties, they pleaded the\ninsufficiency of human reason, and yet were all the while\nunquestioningly relying upon it in the drawing of the most tremendous\ninferences. Such, for instance, are the assumptions which underlie Moses\nMendelssohn\u2019s contention that since animals as well as men agree in the\napprehension of space, it must be believed to be absolutely real.[84]\nThese assumptions also determine Priestley\u2019s assertion that though every\nevent has its cause, there is one causeless happening, namely, the\ncreative act to which the existence of the world is due.[85] On such\nterms, metaphysics is too patently easy to be even plausible.\n\u201cIndifference, doubt, and, in final issue, severe criticism, are truer\nsigns of a profound habit of thought.\u201d[86] The matter of experience\naffords no data for metaphysical inference. In the _a priori_ forms of\nexperience, and there alone, can metaphysics hope to find a basis, if\nany basis is really discoverable.\n\nThis is Kant\u2019s reason for so emphatically insisting that the problem of\nthe _Critique_ is to determine \u201chow much we can hope to achieve by\nreason, when all the material and assistance of experience is taken\naway.\u201d[87] But in keeping only this one point in view Kant greatly\nmisrepresents the problems and scope of the _Critique_. Throughout the\n_Preface_ he speaks the language of the _Aufkl\u00e4rung_. Even in the very\nact of limiting the scope of reason, he overstresses its powers, and\nomits reference to its empirical conditions. It is well to contrast this\nteaching with such a passage as the following:\n\n     \u201cThe position of all genuine idealists from the Eleatics to\n     Berkeley is contained in this formula: \u2018All cognition through the\n     senses and experience is nothing but mere illusion, and only in the\n     ideas of pure understanding and Reason is there truth.\u2019 The\n     fundamental principle ruling all my idealism, on the contrary, is\n     this: \u2018All cognition of things solely from pure understanding or\n     pure Reason is nothing but mere illusion, and only in experience is\n     there truth.\u2019\u201d[88]\n\nBut that passage is equally inadequate as a complete expression of\nKant\u2019s Critical philosophy. The truth lies midway between it and the\nteaching of the _Preface_ to the first edition. Pure reason is as\ndefective an instrument of knowledge as is factual experience. Though\nthe primary aim of metaphysics is to determine our relation to the\nabsolutely real, and though that can only be done by first determining\nthe nature and possible scope of _a priori_ principles, such principles\nare found on investigation to possess only empirical validity. The\ncentral question of the _Critique_ thus becomes the problem of the\nvalidity of their empirical employment. The interrelation of these two\nproblems, that of the _a priori_ and that of experience, and Kant\u2019s\nattitude towards them, cannot be considered till later. The defects of\nthe _Preface_ to the first edition are in part corrected by the\nextremely valuable _Preface_ substituted in the second edition. But some\nfurther points in this first _Preface_ must be considered.\n\n=Prescribed by the very nature of reason itself.=[89]--Metaphysics exists\nas a \u201cnatural disposition,\u201d and its questions are not therefore merely\nartificial.\n\n     \u201cAs natural disposition (_Naturanlage_) ... metaphysics is real.\n     For human reason, without being moved merely by the idle desire for\n     extent and variety of knowledge, proceeds impetuously, driven on\n     by an inward need, to questions such as cannot be answered by any\n     empirical employment of reason, or by principles thence derived.\n     Thus in all men, as soon as their reason has become ripe for\n     speculation, there has always existed and will always continue to\n     exist some kind of metaphysics.\u201d[90]\n\nHence results what Kant entitles _transcendental illusion_.\n\n     \u201cThe cause of this transcendental illusion is that there are\n     fundamental rules and maxims for the employment of Reason,\n     subjectively regarded as a faculty of human knowledge, and that\n     these rules and maxims have all the appearance of being objective\n     principles. We take the subjective necessity of a connection of our\n     concepts, _i.e._ a connection necessitated for the advantage of the\n     understanding, for an objective necessity in the determination of\n     things in themselves. This is an illusion which can no more be\n     prevented than we can prevent the sea from appearing higher at the\n     horizon than at the shore, since we see it through higher light\n     rays; or to cite a still better example, than the astronomer can\n     prevent the moon from appearing larger at its rising, although he\n     is not deceived by this illusion.... There exists, then, a natural\n     and unavoidable dialectic of pure Reason, not one in which a\n     bungler might entangle himself through lack of knowledge, or one\n     which some sophist has artificially invented to confuse thinking\n     people, but one which is inseparable from human Reason, and which,\n     even after its deceiving power has been exposed, will not cease to\n     play tricks with it and continually to entrap it into momentary\n     aberrations that will ever and again call for correction.\u201d[91]\n\n\n=Dogmatism.=[92]--According to Kant there are three possible standpoints\nin philosophy--the dogmatic, the sceptical, and the critical. All\npreceding thinkers come under the first two heads. A dogmatist is one\nwho assumes that human reason can comprehend ultimate reality, and who\nproceeds upon this assumption. He does not, before proceeding to\nconstruct a metaphysics, enquire whether it is possible. Dogmatism\nexpresses itself (to borrow Vaihinger\u2019s convenient mode of\ndefinition[93]) through three factors--_rationalism_, _realism_, and\n_transcendence_. Descartes and Leibniz are typical dogmatists. As\nrationalists they hold that it is possible to determine from pure _a\npriori_ principles the ultimate nature of God, of the soul, and of the\nmaterial universe. They are realists in that they assert that by human\nthought the complete nature of objective reality can be determined. They\nalso adopt the attitude of transcendence. Through pure thought they go\nout beyond the sensible and determine the supersensuous. Scepticism\n(Kant, as above stated,[94] regards it as being in effect equivalent to\nempiricism) may similarly be defined through the three terms,\n_empiricism_, _subjectivism_, _immanence_. A sceptic can never be a\nrationalist. He must reduce knowledge to sense-experience. For this\nreason also his knowledge is infected by subjective conditions; through\nsensation we cannot hope to determine the nature of the objectively\nreal. This attitude is also that of immanence; knowledge is limited to\nthe sphere of sense-experience. Criticism has similarly its three\nconstitutive factors, _rationalism_, _subjectivism_, _immanence_. It\nagrees with dogmatism in maintaining that only through _a priori_\nprinciples can true knowledge be obtained. Such knowledge is, however,\nsubjective[95] in its origin, and for that reason it is also only of\nimmanent application; knowledge is possible only in the sphere of\nsense-experience. Dogmatism claims that knowledge arises independently\nof experience and extends beyond it. Empiricism holds that knowledge\narises out of sense-experience and is valid only within it. Criticism\nteaches that knowledge arises independently of particular experience but\nis valid only for experience.\n\nThe following passages in the _Methodology_ give Kant\u2019s view of the\nhistorical and relative values of the two false methods:\n\n     \u201cThe sceptic is the taskmaster who constrains the dogmatic reasoner\n     to develop a sound critique of the understanding and reason. When\n     the latter has been made to advance thus far, he need fear no\n     further challenge, since he has learned to distinguish his real\n     possessions from that which lies entirely beyond them, and to which\n     he can therefore lay no claim.... Thus the sceptical procedure\n     cannot of itself yield any satisfying answer to the questions of\n     reason, but none the less it prepares the way by awakening its\n     circumspection, and by indicating the radical measures which are\n     adequate to secure it in its legitimate possessions.\u201d[96] \u201cThe\n     first step in matters of pure reason, marking its infancy, is\n     _dogmatic_. The second step is _sceptical_, and indicates that\n     experience has rendered our judgment wiser and more circumspect.\n     But a third step, such as can be taken only by fully matured\n     judgment, is now necessary.... This is not the censorship but the\n     critique of reason, whereby not its present bounds but its\n     determinate [and necessary] limits, not its ignorance on this or\n     that point, but in regard to all possible questions of a certain\n     kind, are demonstrated from principles, and not merely arrived at\n     by way of conjecture. Scepticism is thus a resting-place for human\n     reason, where it can reflect upon its dogmatic wanderings and make\n     survey of the region in which it finds itself, so that for the\n     future it may be able to choose its path with more certainty. But\n     it is no dwelling-place for permanent settlement. That can be\n     obtained only through perfect certainty in our knowledge, alike of\n     the objects themselves and of the limits within which all our\n     knowledge of objects is enclosed.\u201d[97]\n\n=Locke.=[98]--Cf. A 86 = B 119; A 270 = B 327; B 127.\n\n=On the unfavourable contrast between mathematics and\nmetaphysics.=[99]--Cf. _Ueber die Deutlichkeit der Grunds\u00e4tze_ (1764),\n_erste Betrachtung_, and below, pp. 40, 563 ff.\n\n=The age of criticism.=[100]--Kant considered himself as contributing to\nthe further advance of the eighteenth century Enlightenment.[101] In\nview, however, of the contrast between eighteenth and nineteenth century\nthought, and of the real affiliations and ultimate consequences of\nKant\u2019s teaching, it seems truer to regard the Critical philosophy as at\nonce completing and transcending the _Aufkl\u00e4rung_. Kant breaks with many\nof its most fundamental assumptions.\n\n=The Critique of Pure Reason.=[102]--Kant here defines the _Critique_ as\ndirected upon pure reason.[103] Further, it is a criticism of knowledge\nwhich is \u201cindependent of all experience,\u201d or, as Kant adds \u201cfree from\nall experience.\u201d Such phrases, in this context, really mean\n_transcendent_. The _Critique_ is here taken as being a Critical\ninvestigation of transcendent metaphysics, of its sources, scope, and\nlimits.[104]\n\n=Opinion or hypothesis not permissible.=[105]--Cf. below, p. 543 ff.\n\n=I know no enquiries, etc.=[106]--The important questions raised by this\nparagraph are discussed below, p. 235 ff.\n\n=Jean Terrasson= (1670-1750).[107]--The quotation is from his work\nposthumously published (1754), and translated from the French by Frau\nGottsched under the title _Philosophie nach ihrem allgemeinen Einflusse\nauf alle Gegenst\u00e4nde des Geistes und der Sitten_ (1762). Terrasson is\nalso referred to by Kant in his _Anthropologie_, \u00a7\u00a7 44 and 77. Terrasson\nwould seem to be the author of the _Trait\u00e9 de l\u2019infini cr\u00e9\u00e9_ which has\nbeen falsely ascribed to Malebranche. I have translated this latter\ntreatise in the _Philosophical Review_ (July 1905).\n\n=Such a system of pure speculative reason.=[108]--The relation in which\nthis system would stand to the _Critique_ is discussed below, pp. 71-2.\nSpeculative does not with Kant mean transcendent, but merely theoretical\nas opposed to practical. Cf. B 25, A 15 = B 29, A 845 = B 873.\n\n=Under the title: Metaphysics of Nature.=[109]--No such work, at least\nunder this title, was ever completed by Kant. In the Kantian terminology\n\u201cnature\u201d signifies \u201call that is.\u201d Cf. below, p. 580.\n\n\n\n\nPREFACE TO THE SECOND EDITION\n\n\nI shall again give a brief explanatory paraphrase, before proceeding to\ndetailed comment. The main points of the preface of the first edition\nare repeated. \u201cMetaphysics soars above all teaching of experience, and\nrests on concepts only. In it reason has to be her own pupil.\u201d[110] But\nKant immediately proceeds to a further point. That logic should have\nattained the secure method of science is due to its limitation to the\nmere _a priori_ form of knowledge. For metaphysics this is far more\ndifficult, since it \u201c_has to deal not with itself alone, but also with\nobjects_.\u201d[111]\n\nThe words which I have italicised form a very necessary correction of\nthe first edition preface, according to which the _Critique_ would seem\nto \u201ctreat only of reason and its pure thinking.\u201d A further difference\nfollows. The second edition preface, in thus emphasising the objective\naspect of the problem, is led to characterise in a more complete manner\nthe method to be followed in the Critical enquiry. How can the\n_Critique_, if it is concerned, as both editions agree in insisting,\nonly with the _a priori_ which originates in human reason, solve the\nspecifically metaphysical problem, viz. that of determining the\nindependently real? How can an idea in us refer to, and constitute\nknowledge of, an object? The larger part of the preface to the second\nedition is devoted to the Critical solution of this problem. The\nargument of the _Dialectic_ is no longer emphasised at the expense of\nthe _Analytic_.\n\nKant points out that as a matter of historical fact each of the two\nrational sciences, mathematics and physics, first entered upon the\nassured path of knowledge by a sudden revolution, and by the adoption of\na method which in its general characteristics is common to both. This\nmethod consists, not in being led by nature as in leading-strings, but\nin interrogating nature in accordance with what reason produces on its\nown plan. The method of the geometrician does not consist in the study\nof figures presented to the senses. That would be an empirical (in\nKant\u2019s view, sceptical) method. Geometrical propositions could not then\nbe regarded as possessing universality and necessity. Nor does the\ngeometrician employ a dogmatic method, that of studying the mere\nconception of a figure. By that means no new knowledge could ever be\nattained. The actual method consists in interpreting the sensible\nfigures through conceptions that have been rigorously defined, and in\naccordance with which the figures have been constructively generated.\nThe first discovery of this method, by Thales or some other Greek, was\n\u201cfar more important than the discovery of the passage round the\ncelebrated Cape of Good Hope.\u201d[112]\n\nSome two thousand years elapsed before Galileo formulated a\ncorresponding method for physical science. He relied neither on mere\nobservation nor on his own conceptions. He determined the principles\naccording to which alone concordant phenomena can be admitted as laws of\nnature, and then by experiment compelled nature to answer the questions\nwhich these principles suggest. Here again the method is neither merely\nempirical nor purely dogmatic. It possesses the advantages of both.\n\nMetaphysics is ripe for a similar advance. It must be promoted to the\nrank of positive science by the transforming power of an analogous\nmethod. The fundamental and distinguishing characteristic of\nmathematical and physical procedure is the legislative power to which\nreason lays claim. Such procedure, if generalised and extended, will\nsupply the required method of the new philosophy. Reason must be\nregarded as self-legislative in all the domains of our possible\nknowledge. _Objects must be viewed as conforming to human thought, not\nhuman thought to the independently real._ This is the \u201chypothesis\u201d to\nwhich Kant has given the somewhat misleading title, \u201cCopernican.\u201d[113]\nThe _method of procedure_ which it prescribes is, he declares, analogous\nto that which was followed by Copernicus, and will be found to be as\nrevolutionary in its consequences. In terms of this hypothesis a\ncomplete and absolutely certain metaphysics, valid now and for all time,\ncan be created at a stroke. The earliest and oldest enterprise of the\nhuman mind will achieve a new beginning. Metaphysics, the mother of all\nthe sciences, will renew her youth, and will equal in assurance, as she\nsurpasses in dignity, the offspring of her womb.\n\nFrom this new standpoint Kant develops phenomenalism on rationalist\nlines. He professes to prove that though our knowledge is only of\nappearances, it is conditioned by _a priori_ principles. His \u201cCopernican\nhypothesis,\u201d so far from destroying positive science, is, he claims,\nmerely a philosophical extension of the method which it has long been\npractising. Since all science worthy of the name involves _a priori_\nelements, it can be accounted for only in terms of the new hypothesis.\nOnly if objects are regarded as conforming to our forms of intuition,\nand to our modes of conception, can they be anticipated by _a priori_\nreasoning. Science can be _a priori_ just because, properly understood,\nit is not a rival of metaphysics, and does not attempt to define the\nabsolutely real.\n\nBut such a statement at once suggests what may at first seem a most\nfatal objection. Though the new standpoint may account for the _a\npriori_ in experience and science, it can be of no avail in metaphysics.\nIf the _a priori_ concepts have a mental origin, they can have no\nvalidity for the independently real. If we can know only what we\nourselves originate, things in themselves must be unknown, and\nmetaphysics must be impossible. But in this very consequence the new\nhypothesis first reveals its full advantages. It leads to an\ninterpretation of metaphysics which is as new and as revolutionary[114]\nas that which it gives to natural science. Transcendent metaphysics is\nindeed impossible, but in harmony with man\u2019s practical and moral\nvocation, its place is more efficiently taken by an immanent metaphysics\non the one hand, and by a metaphysics of ethics on the other. Together\nthese constitute the new and final philosophy which Kant claims to have\nestablished by his Critical method. Its chief task is to continue \u201cthat\nnoblest enterprise of antiquity,\u201d[115] the distinguishing of appearances\nfrom things in themselves. The unconditioned is that which alone will\nsatisfy speculative reason; its determination is the ultimate\npresupposition of metaphysical enquiry. But so long as the empirical\nworld is regarded as true reality, totality or unconditionedness cannot\npossibly be conceived--is, indeed, inherently self-contradictory. On the\nnew hypothesis there is no such difficulty. By the proof that things in\nthemselves are unknowable, a sphere is left open within which the\nunconditioned can be sought. For though this sphere is closed to\nspeculative reason, the unconditioned can be determined from data\nyielded by reason in its practical activity. The hypothesis which at\nfirst seems to destroy metaphysics proves on examination to be its\nnecessary presupposition. The \u201cCopernican hypothesis\u201d which conditions\nscience will also account for metaphysics properly conceived.\n\nUpon this important point Kant dwells at some length. Even the negative\nresults of the _Critique_ are, he emphasises, truly positive in their\nultimate consequences. The dogmatic extension of speculative reason\nreally leads to the narrowing of its employment, for the principles of\nwhich it then makes use involve the subjecting of things in themselves\nto the limiting conditions of sensibility. All attempts to construe the\nunconditioned in terms that will satisfy reason are by such procedure\nruled out from the very start. To demonstrate this is the fundamental\npurpose and chief aim of the _Critique_. Space and time are merely forms\nof sensuous intuition; the concepts of understanding can yield knowledge\nonly in their connection with them. Though the concepts in their purity\npossess a quite general meaning, this is not sufficient to constitute\nknowledge. The conception of causality, for instance, necessarily\ninvolves the notion of time-sequence; apart from time it is the bare,\nempty, and entirely unspecified conception of a sufficient ground.\nSimilarly, the category of substance signifies the permanent in time and\nspace; as a form of pure reason it has a quite indefinite meaning\nsignifying merely that which is always a subject and never a predicate.\nIn the absence of further specification, it remains entirely problematic\nin its reference. The fact, however, that the categories of the\nunderstanding possess, in independence of sensibility, even this quite\ngeneral significance is all-important. Originating in pure reason they\nhave a wider scope than the forms of sense, and enable us to conceive,\nthough not to gain knowledge of, things in themselves.[116] Our dual\nnature, as being at once sensuous and supersensuous, opens out to us the\napprehension of both.\n\nKant illustrates his position by reference to the problem of the freedom\nof the will. As thought is wider than sense, and reveals to us the\nexistence of a noumenal realm, we are enabled to reconcile belief in the\nfreedom of the will with the mechanism of nature. We can recognise that\nwithin the phenomenal sphere everything without exception is causally\ndetermined, and yet at the same time maintain that the whole order of\nnature is grounded in noumenal conditions. We can assert of one and the\nsame being that its will is subject to the necessity of nature and that\nit is free--mechanically determined in its visible actions, free in its\nreal supersensible existence. We have, indeed, no knowledge of the soul,\nand therefore cannot assert on theoretical grounds that it possesses any\nsuch freedom. The very possibility of freedom transcends our powers of\ncomprehension. The proof that it can at least be conceived without\ncontradiction is, however, all-important. For otherwise no arguments\nfrom the nature of the moral consciousness could be of the least avail;\nbefore a palpable contradiction every argument is bound to give way.\nNow, for the first time, the doctrine of morals and the doctrine of\nnature can be independently developed, without conflict, each in\naccordance with its own laws. The same is true in regard to the\nexistence of God and the immortality of the soul. By means of the\nCritical distinction between the empirical and the supersensible worlds,\n_these conceptions are now for the first time rendered possible of\nbelief_. \u201cI had to remove _knowledge_, in order to make room for\n_belief_.\u201d[117] \u201cThis loss affects only the _monopoly of the schools_,\nin no respect the _interests of humanity_.\u201d[118]\n\nLastly, Kant emphasises the fact that the method of the _Critique_ must\nbe akin to that of dogmatism. It must be rational _a priori_. To adopt\nany other method of procedure is \u201cto shake off the fetters of _science_\naltogether, and thus to change work into play, certainty into opinion,\nphilosophy into philodoxy.\u201d[119] And Kant repeats the claims of the\npreface of the first edition as to the completeness and finality of his\nsystem. \u201cThis system will, as I hope, maintain through the future this\nsame unchangeableness.\u201d[120]\n\n=Logic.=[121]--For Kant\u2019s view of the logic of Aristotle as complete and\nperfect, cf. below, pp. 184-5. Kant compares metaphysics to mathematics\nand physics on the one hand, and to formal logic on the other. The\nformer show the possibility of attaining to the secure path of science\nby a sudden and single revolution; the latter demonstrates the\npossibility of creating a science complete and entire at a stroke.\nThanks to the new Critical method, metaphysics may be enabled, Kant\nclaims, to parallel both achievements at once.\n\n=Theoretical and practical reason.=[122]--Such comment as is necessary\nupon this distinction is given below. Cf. p. 569 ff.\n\n=Hitherto it has been supposed that all knowledge must conform to the\nobjects.=[123]--This statement is historically correct. That assumption\ndid actually underlie one and all of the pre-Kantian philosophies. At\nthe same time, it is true that Kant\u2019s phenomenalist standpoint is\npartially anticipated by Hume, by Malebranche and by Leibniz, especially\nby the first named. Hume argues that to condemn knowledge on the ground\nthat it can never copy or truly reveal any external reality is to\nmisunderstand its true function. Our sense perceptions and our general\nprinciples are so determined by nature as to render feasible only a\npractical organisation of life. When we attempt to derive from them a\nconsistent body of knowledge, failure is the inevitable result.[124]\nMalebranche, while retaining the absolutist view of conceptual\nknowledge, propounds a similar theory of sense-perception.[125] Our\nperceptions are, as he shows, permeated through and through, from end to\nend, with illusion. Such illusions justify themselves by their practical\nusefulness, but they likewise prove that theoretical insight is not the\npurpose of our sense-experience. Kant\u2019s Copernican hypothesis consists\nin great part of an extension of this view to our conceptual, scientific\nknowledge. But he differs both from Malebranche and from Hume in that he\ndevelops his phenomenalism on rationalist lines. He professes to show\nthat though our knowledge is only of the phenomenal, it is conditioned\nby _a priori_ principles. The resulting view of the distinction between\nappearance and reality has kinship with that of Leibniz.[126] The\nphenomena of science, though only appearances, are none the less _bene\nfundata_. Our scientific knowledge, though not equivalent to\nmetaphysical apprehension of the ultimately real, can be progressively\ndeveloped by scientific methods.\n\n=The two \u201cparts\u201d of metaphysics.=[127]--Kant is here drawing the important\ndistinction, which is one result of his new standpoint, between\n_immanent_ and _transcendent_ metaphysics. It is unfortunate that he\ndoes not do so in a more explicit manner, with full recognition of its\nnovelty and of its far-reaching significance. Many ambiguities in his\nexposition here and elsewhere would then have been obviated.[128]\n\n=The unconditioned which Reason postulates in all things by themselves,\nby necessity and by right.=[129]--Points are here raised the discussion\nof which must be deferred. Cf. below, pp. 429-31, 433-4, 558-61.\n\n=The Critique is a treatise on method, not a system of the science\nitself.=[130]--Cf. A xv.; B xxxvi.; and especially A 11 = B 24, below pp.\n71-2.\n\n=The Copernican hypothesis.=[131]--Kant\u2019s comparison of his new hypothesis\nto that of Copernicus has generally been misunderstood. The reader very\nnaturally conceives the Copernican revolution in terms of its main\nultimate consequence, the reduction of the earth from its proud position\nof central pre-eminence. But that does not bear the least analogy to\nthe intended consequences of the Critical philosophy. The direct\nopposite is indeed true. Kant\u2019s hypothesis is inspired by the avowed\npurpose of neutralising the naturalistic implications of the Copernican\nastronomy. His aim is nothing less than the firm establishment of what\nmay perhaps be described as a Ptolemaic, anthropocentric metaphysics.\nSuch naturalistic philosophy as that of Hume may perhaps be described as\nCopernican, but the Critical philosophy, as humanistic, has genuine\nkinship with the Greek standpoint.\n\nEven some of Kant\u2019s best commentators have interpreted the analogy in\nthe above manner.[132] It is so interpreted by T. H. Green[133] and by\nJ. Hutchison Stirling.[134] Caird in his _Critical Philosophy of Kant_\nmakes not the least mention of the analogy, probably for the reason that\nwhile reading it in the same fashion as Green, he recognised the\ninappropriateness of the comparison as thus taken. The analogy is stated\nin typically ambiguous fashion by Lange[135] and by H\u00f6ffding.[136] S.\nAlexander, while very forcibly insisting upon the Ptolemaic character of\nthe Kantian philosophy, also endorses this interpretation in the\nfollowing terms:\n\n     \u201cIt is very ironical that Kant himself signalised the revolution\n     which he believed himself to be effecting as a Copernican\n     revolution. But there is nothing Copernican in it except that he\n     believed it to be a revolution. If every change is Copernican which\n     reverses the order of the terms with which it deals, which declares\n     A to depend on B when B had before been declared to depend on A,\n     then Kant--who believed that he had reversed the order of\n     dependence of mind and things--was right in saying that he effected\n     a Copernican revolution. But he was not right in any other sense.\n     For his revolution, so far as it was one, was accurately\n     anti-Copernican.\u201d[137]\n\nAs the second edition preface is not covered by the published volumes of\nVaihinger\u2019s _Commentary_, the point has not been taken up by him.\n\nNow Kant\u2019s own statements are entirely unambiguous and do not justify\nany such interpretation as that of Green and Alexander. As it seems to\nme, they have missed the real point of the analogy. The misunderstanding\nwould never have been possible save for our neglect of the scientific\nclassics. Kant must have had first-hand acquaintance with Copernicus\u2019\n_De Revolutionibus_, and the comparison which he draws assumes similar\nknowledge on the part of his readers. Copernicus by his proof of the\n\u201chypothesis\u201d (his own term) of the earth\u2019s motion sought only to achieve\na more harmonious ordering of the Ptolemaic universe. And as thus merely\na simplification of the traditional cosmology, his treatise could\nfittingly be dedicated to the reigning Pope. The sun upon which our\nterrestrial life depends was still regarded as uniquely distinct from\nthe fixed stars; and our earth was still located in the central region\nof a universe that was conceived in the traditional manner as being\nsingle and spherical. Giordano Bruno was the first, a generation later,\nto realise the revolutionary consequences to which the new teaching,\nconsistently developed, must inevitably lead. It was he who first taught\nwhat we have now come to regard as an integral part of Copernicus\u2019\nrevolution, the doctrine of innumerable planetary systems side by side\nwith one another in infinite space.\n\nCopernicus\u2019 argument starts from the Aristotelian principle of relative\nmotion. To quote Copernicus\u2019 exact words:\n\n     \u201cAll apprehended change of place is due to movement either of the\n     observed object or of the observer, or to differences in movements\n     that are occurring simultaneously in both. For if the observed\n     object and the observer are moving in the same direction with equal\n     velocity, no motion can be detected. Now it is from the earth that\n     we visually apprehend the revolution of the heavens. If, then, any\n     movement is ascribed to the earth, that motion will generate the\n     appearance of itself in all things which are external to it, though\n     as occurring in the opposite direction, as if everything were\n     passing across the earth. This will be especially true of the daily\n     revolution. For it seems to seize upon the whole world, and indeed\n     upon everything that is around the earth, though not upon the earth\n     itself.... As the heavens, which contain and cover everything, are\n     the common locus of things, it is not at all evident why it should\n     be to the containing rather than to the contained, to the located\n     rather than to the locating, that a motion is to be ascribed.\u201d[138]\n\nThe apparently objective movements of the fixed stars and of the sun are\nmere appearances, due to the projection of our own motion into the\nheavens.\n\n     \u201cThe first and highest of all the spheres is that of the fixed\n     stars, self-containing and all-containing, and consequently\n     immobile, in short the locus of the universe, by relation to which\n     the motion and position of all the other heavenly bodies have to be\n     reckoned.\u201d[139]\n\nNow it is this doctrine, and this doctrine alone, to which Kant is\nreferring in the passages before us, namely, Copernicus\u2019 hypothesis of a\nsubjective explanation of apparently objective motions. And further, in\nthus comparing his Critical procedure to that of Copernicus, he is\nconcerned more with the positive than with the negative consequences of\ntheir common hypothesis. For it is chiefly from the point of view of the\n_constructive_ parts of the _Aesthetic_, _Analytic_, and _Dialectic_\nthat the comparison is formulated. By means of the Critical hypothesis\nKant professes on the one hand to account for our scientific knowledge,\nand on the other to safeguard our legitimate metaphysical aspirations.\nThe spectator projects his own motion into the heavens; human reason\nlegislates for the domain of natural science. The sphere of the fixed\nstars is proved to be motionless; things in themselves are freed from\nthe limitations of space and time. \u201cCopernicus dared, in a manner\ncontradictory of the senses but yet true, to seek the observed\nmovements, not in the heavenly bodies, but in the spectator.\u201d[140]\n\nIn view of Kant\u2019s explicit elimination of all _hypotheses_ from the\n_Critique_[141] the employment of that term would seem to be\nillegitimate. He accordingly here states that though in the _Preface_\nhis Critical theory is formulated as an hypothesis only, in the\n_Critique_ itself its truth is demonstrated _a priori_.\n\n       *       *       *       *       *\n\n=Distinction between knowing and thinking.=[142]--Since according to\nCritical teaching the limits of sense-experience are the limits of\nknowledge, the term knowledge has for Kant a very limited denotation,\nand leaves open a proportionately wide field for what he entitles\nthought. Though things in themselves are unknowable, their existence may\nstill be recognised in thought.\n\n\n\n\nINTRODUCTION\n\n\nI shall first[143] give a restatement, partly historical and partly\nexplanatory, of Kant\u2019s main argument as contained in the enlarged\n_Introduction_ of the second edition.\n\nThere were two stages in the process by which Kant came to full\nrealisation of the Critical problem. There is first the problem as\nformulated in his letter of 1772 to Herz: how the _a priori_ can yield\nknowledge of the independently real.[144] This, as he there states it,\nis an essentially metaphysical problem. It is the problem of the\npossibility of transcendent metaphysics. He became aware of it when\nreflecting upon the function which he had ascribed to intellect in the\n_Dissertation_. Then, secondly, this problem was immeasurably deepened,\nand at the same time the proper line for its treatment was discovered,\nthrough the renewed influence which Hume at some date subsequent to\nFebruary 1772 exercised upon Kant\u2019s thought.[145] Hume awakened Kant to\nwhat may be called the _immanent_ problem involved in the very\nconception of _a priori_ knowledge as such. The primary problem to be\nsolved is not how we advance by means of _a priori_ ideas to the\nindependently real, but how we are able to advance beyond a subject term\nto a predicate which it does not appear to contain. The problem is\nindeed capable of solution, just because it takes this _logical_ form.\nHere as elsewhere, ontological questions are viewed by Kant as soluble\nonly to the extent to which they can be restated in logical terms. Now\nalso the enquiry becomes twofold: how and in what degree are _a priori_\nsynthetic judgments possible, first in their employment within the\nempirical sphere (the problem of immanent metaphysics) and secondly in\ntheir application to things in themselves (the problem of transcendent\nmetaphysics). The outcome of the Critical enquiry is to establish the\nlegitimacy of immanent metaphysics and the impossibility of all\ntranscendent speculation.\n\nThe argument of Kant\u2019s _Introduction_ follows the above sequence. It\nstarts by defining the problem of metaphysical knowledge _a priori_, and\nthrough it leads up to the logical problem of the _a priori_ synthetic\njudgment. In respect of time all knowledge begins _with_ experience. But\nit does not therefore follow that it all arises _from_ experience. Our\nexperience may be a compound of that which we receive through\nimpressions, and of that which pure reason supplies from itself.[146]\nThe question as to whether or not any such _a priori_ actually exists,\nis one that can be answered only after further enquiry. The two\ninseparable criteria of the _a priori_ are necessity and universality.\nThat neither can be imparted to a proposition by experience was Kant\u2019s\nconfirmed and unquestioned belief. He inherited this view both from\nLeibniz and from Hume. It is one of the presuppositions of his argument.\nExperience can reveal only co-existence or sequence. It enables us only\nto assert that so far as we have hitherto observed, there is no\nexception to this or that rule. A generalisation, based on observation,\ncan never possess a wider universality than the limited experience for\nwhich it stands. If, therefore, necessary and universal judgments can\nanywhere be found in our knowledge, the existence of an _a priori_ that\noriginates independently of experience is _ipso facto_\ndemonstrated.[147]\n\nThe contrast between empirical and _a priori_ judgments, as formulated\nfrom the dogmatic standpoint, is the most significant and striking fact\nin the whole range of human knowledge. _A priori_ judgments claim\nabsolute necessity. They allow of no possible exception. They are valid\nnot only for us, but also for all conceivable beings, however different\nthe specific conditions of their existence, whether they live on the\nplanet Mars or in some infinitely remote region of stellar space, and no\nmatter how diversely their bodily senses may be organised. Through these\njudgments a creature five feet high, and correspondingly limited by\ntemporal conditions, legislates for all existence and for all time.\nEmpirical judgments, on the other hand, possess only a hypothetical\ncertainty. We recognise that they may be overturned through some\naddition to our present experience, and that they may not hold for\nbeings on other planets or for beings with senses differently\nconstituted. Whereas the opposite of a rational judgment is not even\nconceivable, the opposite of an empirical judgment is always possible.\nThe one depends upon the inherent and inalienable nature of our\nthinking; the other is bound up with the contingent material of sense.\nThe one claims absolute or metaphysical truth: the other is a merely\ntentative _r\u00e9sum\u00e9_ of a limited experience.\n\nThe possibility of such _a priori_ judgments had hitherto been\nquestioned only by those who sought to deny to them all possible\nobjective validity. Kant, as a rationalist, has no doubt as to their\nactual existence. In the _Introduction_ to the second edition he bluntly\nasserts their _de facto_ existence, citing as instances the propositions\nof mathematics and the fundamental principles of physical science. Their\npossibility can be accounted for through the assumption of _a priori_\nforms and principles.[148] But with equal emphasis he questions the\nvalidity of their _metaphysical_ employment. For that is an entirely\ndifferent matter. We then completely transcend the world of the senses\nand pass into a sphere where experience can neither guide nor correct\nus. In this sphere the _a priori_ is illegitimately taken as being at\nonce the source of our professed knowledge and also the sole criterion\nof its own claims.\n\nThis is the problem, semi-Critical, semi-dogmatic, which is formulated\nin the letter of 1772 to Herz.[149] What right have we to regard ideas,\nwhich as _a priori_ originate from within, as being valid of things in\nthemselves? In so doing we are assuming a pre-established harmony\nbetween our human faculties and the ultimately real; and that is an\nassumption which by its very nature is incapable of demonstration. The\nproofs offered by Malebranche and by Leibniz are themselves speculative,\nand consequently presuppose the conclusion which they profess to\nestablish.[150] As above stated, Kant obtained his answer to this\nproblem by way of the logical enquiry into the nature and conditions of\n_a priori_ judgment.\n\nOne of the chief causes, Kant declares, why hitherto metaphysical\nspeculation has passed unchallenged among those who practise it, is the\nconfusion of two very different kinds of judgment, the analytic and the\nsynthetic. Much the greater portion of what reason finds to do consists\nin the analysis of our concepts of objects.\n\n     \u201cAs this procedure yields real knowledge _a priori_, which\n     progresses in secure and useful fashion, reason is so far misled as\n     surreptitiously to introduce, without itself being aware of so\n     doing, assertions of an entirely different order, in which reason\n     attaches to given concepts others completely foreign to them--and\n     moreover attaches them _a priori_. And yet one does not know how\n     reason comes to do this. This is a question which is never as much\n     as thought of.\u201d[151]\n\nThe concepts which are analytically treated may be either empirical or\n_a priori_. When they are empirical, the judgments which they involve\ncan have no wider application than the experience to which they give\nexpression; and in any case can only reveal what has all along been\nthought, though confusedly, in the term which serves as subject of the\nproposition. They can never reveal anything different in kind from the\ncontents actually experienced. This limitation, to which the analysis of\nempirical concepts is subject, was admitted by both empiricists and\nrationalists. The latter sought, however, to escape its consequences by\nbasing their metaphysics upon concepts which are purely _a priori_, and\nwhich by their _a priori_ content may carry us beyond the experienced.\nBut here also Kant asserts a _non possibile_. _A priori_ concepts, he\nseeks to show, are in all cases purely logical functions without\ncontent, and accordingly are as little capable as are empirical concepts\nof carrying us over to the supersensible. This is an objection which\nholds quite independently of that already noted, namely, that their\nobjective validity would involve a pre-established harmony.\n\nWhat, then, is the nature and what are the generating conditions of\nsynthetic judgments that are also _a priori_? In all judgments there is\na relation between subject and predicate, and that can be of two kinds.\nEither the predicate B belongs to the subject A, or B lies outside the\nsphere of the concept A though somehow connected with it. In the former\ncase the judgment is analytic; in the latter it is synthetic. The one\nsimply unfolds what has all along been conceived in the subject concept;\nthe other ascribes to the concept of the subject a predicate which\ncannot be found in it by any process of analysis. Thus the judgment \u2018all\nbodies are extended\u2019 is analytic. The concept of body already contains\nthat of extension, and is impossible save through it. On the other hand,\nthe judgment \u2018all bodies are heavy\u2019 is synthetic. For not body as such,\nbut only bodies which are in interaction with other bodies, are found to\ndevelop this property. Bodies can very well be conceived as not\ninfluencing one another in any such manner.\n\nThere is no difficulty in accounting for analytic judgments. They can\nall be justified by the principle of contradiction. Being analytic, they\ncan be established _a priori_. Nor, Kant here claims, is there any\ndifficulty in regard to synthetic judgments that are empirical. Though\nthe predicate is not contained in the subject concept, they belong to\neach other (though accidentally) as parts of a given empirical whole.\nExperience is the _x_ which lies beyond the concept A, and on which\nrests the possibility of the synthesis of B with A. In regard, however,\nto synthetic judgments which are likewise _a priori_, the matter is very\ndifferent. Hitherto, both by the sensationalists and by the\nrationalists, all synthetic judgments have been regarded as empirical,\nand all _a priori_ judgments as analytic. The only difference between\nthe opposed schools lies in the relative value which they ascribe to the\ntwo types of judgment. For Hume the only really fruitful judgments are\nthe synthetic judgments _a posteriori_; analytic judgments are of quite\nsecondary value; they can never extend our knowledge, but only clarify\nits existing content. For Leibniz, on the other hand, true knowledge\nconsists only in the analysis of our _a priori_ concepts, which he\nregards as possessing an intrinsic and fruitful content; synthetic\njudgments are always empirical, and as such are purely contingent.[152]\n\nThus for pre-Kantian philosophy analytic is interchangeable with _a\npriori_, and synthetic with _a posteriori_. Kant\u2019s Critical problem\narose from the startling discovery that the _a priori_ and the synthetic\ndo not exclude one another. A judgment may be synthetic and yet also _a\npriori_. He appears to have made this discovery under the influence of\nHume, through study of the general principle of causality--every event\nmust have a cause.[153] In that judgment there seems to be no connection\nof any kind discoverable between the subject (the conception of an event\nas something happening in time) and the predicate (the conception of\nanother event preceding it as an originating cause); and yet we not\nmerely ascribe the one to the other but assert that they are necessarily\nconnected. We can conceive an event as sequent upon a preceding empty\ntime; none the less, in physical enquiry, the causal principle is\naccepted as an established truth. Here, then, is a new and altogether\nunique type of judgment, of thoroughly paradoxical nature. So entirely\nis it without apparent basis, that Hume, who first deciphered its\nstrange character, felt constrained to ascribe our belief in it to an\nunreasoning and merely instinctive, \u2018natural\u2019 habit or custom.\n\nKant found, however, that the paradoxical characteristics of the causal\nprinciple also belong to mathematical and physical judgments. This fact\nmakes it impossible to accept Hume\u2019s sceptical conclusion. If even the\nassertion 7 + 5 = 12 is both synthetic and _a priori_, it is obviously\nimpossible to question the validity of judgments that possess these\ncharacteristics. But they do not for that reason any the less urgently\npress for explanation. Such an enquiry might not, indeed, be necessary\nwere we concerned only with scientific knowledge. For the natural\nsciences justify themselves by their practical successes and by their\nsteady unbroken development. But metaphysical judgments are also of this\ntype; and until the conditions which make _a priori_ synthetic judgment\npossible have been discovered, the question as to the legitimacy of\nmetaphysical speculation cannot be decided. Such judgments are plainly\nmysterious, and urgently call for further enquiry.\n\nThe problem to be solved concerns the ground of our ascription to the\nsubject concept, as necessarily belonging to it, a predicate which seems\nto have no discoverable relation to it. What is the unknown _x_ on which\nthe understanding rests in asserting the connection? It cannot be\nrepeated experience; for the judgments in question claim necessity. Nor\ncan such judgments be proved by means of a logical test, such as the\ninconceivability of the opposite. The absence of all apparent connection\nbetween subject and predicate removes that possibility. These, however,\nare the only two methods of proof hitherto recognised in science and\nphilosophy. The problem demands for its solution nothing less than the\ndiscovery and formulation of an entirely novel method of proof.\n\nThe three main classes of _a priori_ synthetic judgments are, Kant\nproceeds, the mathematical, the physical, and the metaphysical. The\nsynthetic character of mathematical judgments has hitherto escaped\nobservation owing to their being proved (as is required of all apodictic\ncertainty) according to the principle of contradiction. It is therefrom\ninferred that they rest on the authority of that principle, and are\ntherefore analytic. That, however, is an illegitimate inference; for\nthough the truth of a synthetic proposition can be thus demonstrated,\nthat can only be if another synthetic principle is first presupposed. It\ncan never be proved that its truth, as a separate judgment, is demanded\nby the principle of contradiction. That 7 + 5 must equal 12 does not\nfollow analytically from the conception of the sum of seven and five.\nThis conception contains nothing beyond the union of both numbers into\none; it does not tell us what is the single number that combines both.\nThat five should be added to seven is no doubt implied in the\nconception, but not that the sum should be twelve. To discover that, we\nmust, Kant maintains, go beyond the concepts and appeal to intuition.\nThis is more easily recognised when we take large numbers. We then\nclearly perceive that, turn and twist our concepts as we may, we can\nnever, by means of mere analysis of them, and without the help of\nintuition, arrive at the sum that is wanted. The fundamental\npropositions of geometry, the so-called axioms, are similarly synthetic,\n_e.g._ that the straight line between two points is the shortest. The\nconcept \u2018straight\u2019 only defines direction; it says nothing as to\nquantity.\n\nAs an instance of a synthetic _a priori_ judgment in physical science\nKant cites the principle: the quantity of matter remains constant\nthroughout all changes. In the conception of matter we do not conceive\nits permanency, but only its presence in the space which it fills. The\nopposite of the principle is thoroughly conceivable.\n\nMetaphysics is _meant_ to contain _a priori_ knowledge. For it seeks to\ndetermine that of which we can have no experience, as _e.g._ that the\nworld must have a first beginning. And if, as will be proved, our _a\npriori_ concepts have no content, which through analysis might yield\nsuch judgments, these judgments also must be synthetic.\n\nHere, then, we find the essential problem of pure reason. Expressed in a\nsingle formula, it runs: How are synthetic _a priori_ judgments\npossible? To ask this question is to enquire, first, how pure\nmathematics is possible; secondly, how pure natural science is possible;\nand thirdly, how metaphysics is possible. That philosophy has hitherto\nremained in so vacillating a state of ignorance and contradiction is\nentirely due to the neglect of this problem of _a priori_ synthesis.\n\u201cIts solution is the question of life and death to metaphysics.\u201d Hume\ncame nearest to realising the problem, but he discovered it in too\nnarrow a form to appreciate its full significance and its revolutionary\nconsequences.\n\n     \u201cGreater firmness will be required if we are not to be deterred by\n     inward difficulties and outward opposition from endeavouring,\n     through application of a method entirely different from any\n     hitherto employed, to further the growth and fruitfulness of a\n     science indispensable to human reason--a science whose every branch\n     may be cut away but whose root cannot be destroyed.\u201d[154]\n\nThese statements are decidedly ambiguous, owing to Kant\u2019s failure to\ndistinguish in any uniform and definite manner between immanent and\ntranscendent metaphysics.[155] The term metaphysics is used to cover\nboth. Sometimes it signifies the one, sometimes the other; while in\nstill other passages its meaning is neutral. But if we draw the\ndistinction, Kant\u2019s answer is that a genuine and valid immanent\nmetaphysics is for the first time rendered possible by his _Critique_;\nits positive content is expounded in the _Analytic_. Transcendent\nmetaphysics, on the other hand, is criticised in the _Dialectic_; it is\nnever possible. The existing speculative sciences transgress the limits\nof experience and yield only a pretence of knowledge. This determination\nof the limits of our possible _a priori_ knowledge is the second great\nachievement of the _Critique_. Thus the _Critique_ serves a twofold\npurpose. It establishes a new _a priori_ system of metaphysics, and also\ndetermines on principles equally _a priori_ the ultimate limits beyond\nwhich metaphysics can never advance. The two results, positive and\nnegative, are inseparable and complementary. Neither should be\nemphasised to the neglect of the other.\n\n\nCOMMENT ON THE ARGUMENT OF KANT\u2019S _INTRODUCTION_\n\nThis _Introduction_, though a document of great historical importance as\nbeing the first definite formulation of the generating problem of Kant\u2019s\nnew philosophy, is extremely unsatisfactory as a statement of Critical\nteaching. The argument is developed in terms of distinctions which are\nborrowed from the traditional logic, and which are not in accordance\nwith the transcendental principles that Kant is professing to establish.\nThis is, indeed, a criticism which may be passed upon the _Critique_ as\na whole. Though Kant was conscious of opening a new era in the history\nof philosophy, and compares his task with that of Thales, Copernicus,\nBacon and Galileo, it may still be said that he never fully appreciated\nthe greatness of his own achievement. He invariably assumes that the\nrevolutionary consequences of his teaching will not extend to the sphere\nof pure logic. They concern, as he believed, only our metaphysical\ntheories regarding the nature of reality and the determining conditions\nof our human experience. As formal logic prescribes the axiomatic\nprinciples according to which all thinking must proceed, its validity is\nnot affected by the other philosophical disciplines, and is superior to\nthe considerations that determine their truth or falsity. Its\ndistinctions may be securely relied upon in the pioneer labours of\nCritical investigation. This was, of course, a very natural assumption\nfor Kant to make; and many present-day thinkers will maintain that it is\nentirely justified. Should that be our attitude, we may approve of\nKant\u2019s general method of procedure, but shall be compelled to dissent\nfrom much in his argument and from many of his chief conclusions. If, on\nthe other hand, we regard formal logic as in any degree adequate only as\na theory of the thought processes involved in the formation and\napplication of the generic or class concept,[156] we shall be prepared\nto find that the equating of this highly specialised logic with logic in\ngeneral has resulted in the adoption of distinctions which may be fairly\nadequate for the purposes in view of which they have been formulated,\nbut which must break down when tested over a wider field. So far from\ncondemning Kant for departing in his later teaching from these hard and\nfast distinctions, we shall welcome every sign of his increasing\nindependence.\n\nKant was not, of course, so blind to the real bearing of his principles\nas to fail to recognise that they have logical implications.[157] He\nspeaks of the new metaphysics which he has created as being a\ntranscendental logic. It is very clear, however, that even while so\ndoing he does not regard it as in any way alternative to the older\nlogic, but as moving upon a different plane, and as yielding results\nwhich in no way conflict with anything that formal logic may teach.\nIndeed Kant ascribes to the traditional logic an almost sacrosanct\nvalidity. Both the general framework of the _Critique_ and the\narrangement of the minor subdivisions are derived from it. It is\nsupposed to afford an adequate account of discursive thinking, and such\nsupplement as it may receive is regarded as simply an extension of its\ncarefully delimited field. There are two logics, that of discursive or\nanalytic reasoning, and that of synthetic interpretation. The one is\nformal; the other is transcendental. The one was created by Aristotle,\ncomplete at a stroke; Kant professes to have formulated the other in an\nequally complete and final manner.\n\nThis latter claim, which is expressed in the most unqualified terms in\nthe _Prefaces_ to the first and second editions, is somewhat startling\nto a modern reader, and would seem to imply the adoption of an\nultra-rationalistic attitude, closely akin to that of Wolff.\n\n     \u201cIn this work I have made completeness my chief aim, and I venture\n     to assert that there is not a single metaphysical problem which has\n     not been solved, or for the solution of which the key at least has\n     not been supplied. Reason is, indeed, so perfect a unity that if\n     its principle were insufficient for the solution of even a single\n     one of all the questions to which it itself gives birth, we should\n     be justified in forthwith rejecting it as incompetent to answer,\n     with perfect certainty, any one of the other questions.\u201d[158]\n     \u201cMetaphysics has this singular advantage, such as falls to the lot\n     of no other science which deals with objects (for _logic_ is\n     concerned only with the form of thought in general), that should\n     it, through this _Critique_, be set upon the secure path of\n     science, it is capable of acquiring exhaustive knowledge of its\n     entire field. It can finish its work and bequeath it to posterity\n     as a capital that can never be added to. For metaphysics has to\n     deal only with principles, and with the limits of their employment\n     as determined by these principles themselves. Since it is a\n     fundamental science, it is under obligation to achieve this\n     completeness. We must be able to say of it: _nil actum reputans, si\n     quid superesset agendum_.\u201d[159]\n\nThese sanguine expectations--by no means supported by the after-history\nof Kant\u2019s system--are not really due to Kant\u2019s immodest over-estimate of\nthe importance of his work. They would rather seem to be traceable, on\nthe one hand to his continuing acceptance of rationalistic assumptions\nproper only to the philosophy which he is displacing, and on the other\nto his failure to appreciate the full extent of the revolutionary\nconsequences which his teaching was destined to produce in the then\nexisting philosophical disciplines. Kant, like all the greatest\nreformers, left his work in the making. Both his results and his methods\ncall for modification and extension in the light of the insight which\nthey have themselves rendered possible. Indeed, Kant was himself\nconstantly occupied in criticising and correcting his own acquired\nviews; and this is nowhere more evident than in the contrast between the\nteaching of this _Introduction_ and that of the central portions of the\n_Analytic_. But even the later expressions of his maturer views reveal\nthe persisting conflict. They betray the need for further\nreconstruction, even in the very act of disavowing it. Not an additional\nlogic, but the demonstration of the imperative need for a complete\nrevisal of the whole body of logical science, is the first, and in many\nrespects the chief, outcome of his Critical enquiries.\n\nThe broader bearings of the situation may perhaps be indicated as\nfollows. If our account of Kant\u2019s awakening from his dogmatic\nslumber[160] be correct, it consisted in his recognition that\nself-evidence will not suffice to guarantee any general principle. The\nfundamental principles of our experience are synthetic. That is to say,\ntheir opposite is in all cases conceivable. Combining this conclusion\nwith his previous conviction that they can never be proved by induction\nfrom observed facts, he was faced with the task of establishing\nrationalism upon a new and altogether novel basis. If neither empirical\nfacts nor intuitive self-evidence may be appealed to, in what manner can\nproof proceed? And how can we make even a beginning of demonstration, if\nour very principles have themselves to be established? Principles are\nnever self-evident, and yet principles are indispensable. Such was\nKant\u2019s unwavering conviction as regards the fundamental postulates alike\nof knowledge and of conduct.\n\nThis is only another way of stating that Kant is the real founder of the\n_Coherence_ theory of truth.[161] He never himself employs the term\nCoherence, and he constantly adopts positions which are more in harmony\nwith a _Correspondence_ view of the nature and conditions of knowledge.\nBut all that is most vital in his teaching, and has proved really\nfruitful in its after-history, would seem to be in line with the\npositions which have since been more explicitly developed by such\nwriters as Lotze, Sigwart, Green, Bradley, Bosanquet, Jones and Dewey,\nand which in their tenets all derive from Hegel\u2019s restatement of Kant\u2019s\nlogical doctrines. From this point of view principles and facts mutually\nestablish one another, the former proving themselves by their capacity\nto account for the relevant phenomena, and the latter distinguishing\nthemselves from irrelevant accompaniments by their conformity to the\nprinciples which make insight possible. In other words, all proof\nconforms in general type to the hypothetical method of the natural\nsciences. Kant\u2019s so-called transcendental method, the method by which he\nestablishes the validity of the categories, is itself, as we have\nalready observed,[162] of this character. Secondly, the distinction\nbetween the empirical and the _a priori_ must not be taken (as Kant\nhimself takes it in his earlier, and occasionally even in his later\nutterances) as marking a distinction between two kinds of knowledge.\nThey are elements inseparably involved in all knowledge. And lastly, the\ncontrast between analysis and synthesis becomes a difference not of kind\nbut of degree. Nothing can exist or be conceived save as fitted into a\nsystem which gives it meaning and decides as to its truth. In the degree\nto which it can be studied in relative independence of the supporting\nsystem analysis will suffice; in the degree to which it refers us to\nthis system it calls for synthetic interpretation. But ultimately the\nneeds of adequate understanding must constrain us to the employment of\nboth methods of enquiry. Nothing can be known save in terms of the wider\nwhole to which it belongs.\n\nThere is, however, one important respect in which Kant diverges in very\nradical fashion from the position of Hegel. The final whole to which all\nthings must be referred is represented to us only through an \u201cIdea,\u201d for\nwhich no corresponding reality can ever be found. The system which\ndecides what is to be regarded as _empirically_ real is the mechanical\nsystem of natural science. We have no sufficient theoretical criterion\nof absolute reality.\n\nThese somewhat general considerations may be made more definite if we\nnow endeavour to determine in what specific respects the distinctions\nemployed in the _Introduction_ fail to harmonise with the central\ndoctrines of the _Analytic_.\n\nIn the first place, Kant states his problem in reference only to the\nattributive judgment. The other types of relational judgment are\nentirely ignored. For even when he cites judgments of other relational\ntypes, such as the propositions of arithmetic and geometry, or that\nwhich gives expression to the causal axiom, he interprets them on the\nlines of the traditional theory of the categorical proposition. As we\nshall find,[163] it is with the relational categories, and consequently\nwith the various types of relational judgment to which they give rise,\nthat the _Critique_ is alone directly concerned. Even the attributive\njudgment is found on examination to be of this nature. What it expresses\nis not the inclusion of an attribute within a given group of attributes,\nbut the organisation of a complex manifold in terms of the dual category\nof substance and attribute.\n\nSecondly, this exclusively attributive interpretation of the judgment\nleads Kant to draw, in his _Introduction_, a hard and fast distinction\nbetween the analytic and the synthetic proposition--a distinction which,\nwhen stated in such extreme fashion, obscures the real implications of\nthe argument of the _Analytic_. For Kant here propounds[164] as an\nexhaustive division the two alternatives: (_a_) inclusion of the\npredicate concept within the subject concept, and (_b_) the falling of\nthe predicate concept entirely outside it. He adds, indeed, that in the\nlatter case the two concepts may still be in some way connected with one\nanother; but this is a concession of which he takes no account in his\nsubsequent argument. He leaves unconsidered the third possibility, that\nevery judgment is both analytic and synthetic. If concepts are not\nindependent entities,[165] as Kant, in agreement with Leibniz, still\ncontinues to maintain, but can function only as members of an\narticulated system, concepts will be distinguishable from one another,\nand yet will none the less involve one another. In so far as the\ndistinguishable elements in a judgment are directly related, the\njudgment may _seem_ purely analytic; in so far as they are related only\nin an indirect manner through a number of intermediaries, they may\n_seem_ to be purely synthetic. But in every case there is an internal\narticulation which is describable as synthesis, and an underlying unity\nthat in subordinating all differences realises more adequately than any\nmere identity the demand for connection between subject and predicate.\nIn other words, all judgments will, on this view, be of the relational\ntype. Even the attributive judgment, as above noted, is no mere\nassertion of identity. It is always expressed in terms of the dual\ncategory of substance and attribute, connecting by a _relation_ contents\nthat as contents may be extremely diverse.\n\nThis would seem to be the view to which Kant\u2019s Critical teaching, when\nconsistently developed, is bound to lead. For in insisting that the\nsynthetic character of a judgment need not render it invalid, and that\nall the fundamental principles and most of the derivative judgments of\nthe positive sciences are of this nature, Kant is really maintaining\nthat the justification of a judgment is always to be looked for beyond\nits own boundaries in some implied context of coherent experience. But\nthough the value of his argument lies in clear-sighted recognition of\nthe synthetic factor in all genuine knowledge, its cogency is greatly\nobscured by his continued acceptance of the possibility of judgments\nthat are purely analytic. Thus there is little difficulty in detecting\nthe synthetic character of the proposition: all bodies are heavy. Yet\nthe reader has first been required to admit the analytic character of\nthe proposition: all bodies are extended. The two propositions are\nreally identical in logical character. Neither can be recognised as true\nsave in terms of a comprehensive theory of physical existence. If matter\nmust exist in a state of distribution in order that its parts may\nacquire through mutual attraction the property of weight, the size of a\nbody, or even its possessing any extension whatsoever, may similarly\ndepend upon specific conditions such as may conceivably not be\nuniversally realised. We find the same difficulty when we are called\nupon to decide whether the judgment 7 + 5 = 12 is analytic or purely\nsynthetic. Kant speaks as if the concepts of 7, 5, and 12 were\nindependent entities, each with its own quite separate connotation. But\nobviously they can only be formed in the light of the various connected\nconcepts which go to constitute our system of numeration. The\nproposition has meaning only when interpreted in the light of this\nconceptual system. It is not, indeed, a self-evident identical\nproposition; but neither is the connection asserted so entirely\nsynthetic that intuition will alone account for its possibility. That,\nhowever, brings us to the third main defect in Kant\u2019s argument.\n\nWhen Kant states[166] that in synthetic judgments we require, besides\nthe concept of the subject, something else on which the understanding\ncan rely in knowing that a predicate, not contained in the concept,\nnevertheless belongs to it, he entitles this something _x_. In the case\nof empirical judgments, this _x_ is brute experience. Such judgments,\nKant implies, are _merely_ empirical. No element of necessity is\ninvolved, not even in an indirect manner; in reference to empirical\njudgments there is no problem of _a priori synthesis_. Now in\nformulating the issue in this way, Kant is obscuring the essential\npurpose of his whole enquiry. He may, without essential detriment to his\ncentral position, still continue to preserve a hard-and-fast distinction\nbetween analytic and synthetic judgments. In so doing he is only failing\nto perceive the ultimate consequences of his final results. But in\nviewing empirical judgments as lacking in every element of necessity, he\nis destroying the very ground upon which he professes to base the _a\npriori_ validity of general principles. All judgments involve relational\nfactors of an _a priori_ character. The appeal to experience is the\nappeal to an implied system of nature. Only when fitted into the context\nyielded by such a system can an empirical proposition have meaning, and\nonly in the light of such a presupposed system can its truth be\ndetermined. It can be true at all, only if it can be regarded as\nnecessarily holding, under the same conditions, for all minds\nconstituted like our own. Assertion of a contingent relation--as in the\nproposition: this horse is white--is not equivalent to contingency of\nassertion. Colour is a variable quality of the genus horse, but in the\nindividual horse is necessarily determined in some particular mode. If a\nhorse is naturally white, it is necessarily white. Though, therefore, in\nthe above proposition, necessity receives no explicit verbal expression,\nit is none the less implied.\n\nIn other words, the distinction between the empirical and the _a priori_\nis not, as Kant inconsistently assumes in this _Introduction_, a\ndistinction between two kinds of synthesis or judgment, but between two\nelements inseparably involved in every judgment. Experience is\ntranscendentally conditioned. Judgment is in all cases the expression of\na relation which implies an organised system of supporting propositions;\nand for the articulation of this system _a priori_ factors are\nindispensably necessary.\n\nBut the most flagrant example of Kant\u2019s failure to live up to his own\nCritical principles is to be found in his doctrine of pure intuition. It\nrepresents a position which he adopted in the pre-Critical period. It is\nprefigured in _Ueber die Deutlichkeit der Grunds\u00e4tze_ (1764),[167] and\nin _Von dem ersten Grunde des Unterschiedes der Gegenden im Raume_\n(1768),[168] and is definitely expounded in the _Dissertation_\n(1770).[169] That Kant continued to hold this doctrine, and that he\nhimself regarded it as an integral part of his system, does not, of\ncourse, suffice to render it genuinely Critical. As a matter of fact, it\nis really as completely inconsistent with his Critical standpoint as is\nthe view of the empirical proposition which we have just been\nconsidering. An appeal to our fingers or to points[170] is as little\ncapable, in and by itself, of justifying any _a priori_ judgment as are\nthe sense-contents of grounding an empirical judgment. Even when Kant is\nallowed the benefit of his own more careful statements,[171] and is\ntaken as asserting that arithmetical propositions are based on a pure _a\npriori_ intuition which can find only approximate expression in sensuous\nterms, his statements run counter to the main tendencies of his Critical\nteaching, as well as to the recognised methods of the mathematical\nsciences. Intuition may, as Poincar\u00e9 and others have maintained, be an\nindispensable element in all mathematical concepts; it cannot afford\n_proof_ of any general theorem. The conceptual system which directs our\nmethods of decimal counting is what gives meaning to the judgment 7 + 5\n= 12; it is also what determines that judgment as true. The appeal to\nintuition in numerical judgments must be regarded only as a means of\nimaginatively realising in a concrete form the abstract relations of\nsome such governing system, or else as a means of detecting relations\nnot previously known. The last thing in the world which such a method\ncan yield is universal demonstration. This is equally evident in regard\nto geometrical propositions. That a straight line is the shortest\ndistance between two points, cannot be proved by any mere appeal to\nintuition. The judgment will hold if it can be assumed that space is\nEuclidean in character; and to justify that assumption it must be shown\nthat Euclidean concepts are adequate to the interpretation of our\nintuitional data. Should space possess a curvature, the above\nproposition might cease to be universally valid. Space is not a simple,\nunanalysable datum. Though intuitionally apprehended, it demands for its\nprecise determination the whole body of geometrical science.[172]\n\nThe comparative simplicity of Kant\u2019s intuitional theory of mathematical\nscience, supported as it is by the seemingly fundamental distinction\nbetween abstract concepts of reflective thinking and the construction of\nconcepts[173] in geometry and arithmetic, has made it intelligible even\nto those to whom the very complicated argument of the _Analytic_ makes\nno appeal. It would also seem to be inseparably bound up with what from\nthe popular point of view is the most striking of all Kant\u2019s theoretical\ndoctrines, namely, his view that space and time are given subjective\nforms, and that the assertion of their independent reality must result\nin those contradictions to which Kant has given the title antinomy. For\nthese reasons his intuitional theory of mathematical science has\nreceived attention out of all proportion to its importance. Its\npre-Critical character has been more or less overlooked, and instead of\nbeing interpreted in the light of Critical principles, it has been\nallowed to obscure the sounder teaching of the _Analytic_. In this\nmatter Schopenhauer is a chief culprit. He not only takes the views of\nmathematical science expounded in the _Introduction_ and _Aesthetic_ as\nbeing in line with Kant\u2019s main teaching, but expounds them in an even\nmore unqualified fashion than does Kant himself.\n\nThere are thus four main defects in the argument of this _Introduction_,\nregarded as representative of Critical teaching. (1) Its problems are\nformulated exclusively in terms of the attributive judgment; the other\nforms of relational judgment are ignored. (2) It maintains that\njudgments are either merely analytic or completely synthetic. (3) It\nproceeds in terms of a further division of judgments into those that are\npurely empirical and those that are _a priori_. (4) It seems to assert\nthat the justification for mathematical judgments is intuitional. All\nthese four positions are in some degree retained throughout the\n_Critique_, but not in the unqualified manner of this _Introduction_. In\nthe _Analytic_, judgment in all its possible forms is shown to be a\nsynthetic combination of a given manifold in terms of relational\ncategories. This leads to a fourfold conclusion. In the first place,\njudgment must be regarded as essentially relational. Secondly, the _a\npriori_ and the empirical must not be taken as two separate kinds of\nknowledge, but as two elements involved in all knowledge. Thirdly,\nanalysis and synthesis must not be viewed as co-ordinate processes;\nsynthesis is the more fundamental; it conditions all analysis. And\nlastly, it must be recognised that nothing is merely given; intuitional\nexperience, whether sensuous or _a priori_, is conditioned by processes\nof conceptual interpretation. Though the consequences which follow from\nthese conclusions, if fully developed, would carry us far beyond any\npoint which Kant himself reached in the progressive maturing of his\nviews, the next immediate steps would still be on the strict lines of\nthe Critical principles, and would involve the sacrifice only of such\npre-Critical doctrines as that of the intuitive character of\nmathematical proof. Such correction of Kant\u2019s earlier positions is the\nnecessary complement of his own final discovery that sense-intuition is\nincapable of grounding even the so-called empirical judgment.\n\n       *       *       *       *       *\n\nThe _Introduction_ to the first edition bears all the signs of having\nbeen written previous to the central portions of the _Analytic_.[174]\nThat it was not, however, written prior to the _Aesthetic_ seems\nprobable. The opening sections of the _Aesthetic_ represent what is\nvirtually an independent introduction which takes no account of the\npreceding argument, and which redefines terms and distinctions that have\nalready been dwelt upon. The extensive additions which Kant made in\nrecasting the _Introduction_ for the second edition are in many respects\na great improvement. In the first edition Kant had not, except when\nspeaking of the possibility of constructing the concepts of mathematical\nscience, referred to the synthetic character of mathematical judgments.\nThis is now dwelt upon in adequate detail. Kant\u2019s reason for not making\nthe revision more radical was doubtless his unwillingness to undertake\nthe still more extensive alterations which this would have involved. Had\nhe expanded the opening statement of the second edition _Introduction_,\nthat even our empirical knowledge is a compound of the sensuous and the\n_a priori_, an entirely new _Introduction_ would have become necessary.\nThe additions made are therefore only such as will not markedly conflict\nwith the main tenor of the argument of the first edition.\n\n\nHOW ARE SYNTHETIC _A PRIORI_ JUDGMENTS POSSIBLE?\n\nTreatment of detailed points will be simplified if we now consider in\nsystematic fashion the many difficulties that present themselves in\nconnection with Kant\u2019s mode of formulating his central problem: _How are\nsynthetic_ a priori _judgments possible?_ This formula is less definite\nand precise than would at first sight appear. The central phrase\n\u2018synthetic _a priori_\u2019 is sufficiently exact (the meaning to be attached\nto the _a priori_ has already been considered[175]), but ambiguities of\nthe most various kinds lurk in the seemingly innocent and simple terms\nwith which the formula begins and ends:\n\n     A. \u2018How\u2019 has two very different meanings:\n\n     (_a_) _How_ possible = _in what manner_ possible = wie.\n\n     (_b_) _How_ possible = _in how far_ possible, _i.e. whether_\n     possible = _ob_.\n\nIn connection with these two meanings of the term \u2018how,\u2019 we shall have\nto consider the distinction between the synthetic method employed in the\n_Critique_ and the analytic method employed in the _Prolegomena_.\n\n     B. \u2018Possible\u2019 has a still wider range of application.\n     Vaihinger[176] distinguishes within it no less than three pairs of\n     alternative meanings:\n\n     (_a_) Psychological and logical possibility.\n\n     (_b_) Possibility of explanation and possibility of existence.\n\n     (_c_) Real and ideal possibility.\n\nA. Kant personally believed that the possibility of valid _a priori_\nsynthetic judgment is proved by the existing sciences of mathematics and\nphysics. And that being so, there were for Kant two very different\nmethods which could be employed in accounting for their possibility, the\nsynthetic or progressive, and the analytic or regressive. The synthetic\nmethod would start from given, ordinary experience (in its simplest\nform, as consciousness of time), to discover its conditions, and from\nthem to prove the validity of knowledge that is _a priori_. The analytic\nmethod would start \u201cfrom the sought as if it were given,\u201d that is, from\nthe existence of _a priori_ synthetic judgments, and, assuming them as\nvalid, would determine the conditions under which alone such validity\ncan be possible. The precise formulation of these two methods, the\ndetermination of their interrelations, of their value and comparative\nscope, is a matter of great importance, and must therefore be considered\nat some length.\n\nThe synthetic method may easily be confounded with the analytic method.\nFor in the process of its argument it makes use of analysis. By\nanalysing ordinary experience in the form in which it is given, it\ndetermines (in the _Aesthetic_ and in the _Analytic of Concepts_) the\nfundamental elements of which knowledge is composed, and the generating\nconditions from which it results. From these the validity of the _a\npriori_ principles that underlie mathematics and physics can (in the\n_Analytic of Principles_) be directly deduced. The fundamental\ndifferentiating feature, therefore, of the so-called synthetic method is\nnot its synthetic procedure, since in great part, in the solution of the\nmost difficult portion of its task, it employs an analytic method, but\nonly its attitude towards the one question of the validity of _a priori_\nsynthetic knowledge. It does not postulate this validity as a premiss,\nbut proves it as a consequence of conditions which are independently\nestablished. By a preliminary regress upon the conditions of our _de\nfacto_ consciousness it acquires data from which it is enabled to\nadvance by a synthetic, progressive or deductive procedure to the\nestablishment of the validity of synthetic _a priori_ judgments. The\nanalytic method, on the other hand, makes no attempt to prove the\nvalidity of _a priori_ knowledge. It seeks only to discover the\nconditions under which such knowledge, if granted to exist, can possess\nvalidity, and in the light of which its paradoxical and apparently\ncontradictory features can be viewed as complementary to one another.\nThe conditions, thus revealed, will render the validity of knowledge\nconceivable, will account for it once it has been assumed; but they do\nnot prove it. The validity is a premiss; the whole argument rests upon\nthe assumption of its truth. The conditions are only postulated _as\nconditions_; and their reality becomes uncertain, if the validity, which\npresupposes them, is itself called in question. Immediately we attempt\nto reverse the procedure, and to prove validity from these conditions,\nour argument must necessarily adopt the synthetic form; and that, as has\nbeen indicated, involves the prior application of a very different and\nmuch more thorough process of analysis. The distinction between the two\nmethods may therefore be stated as follows. In the synthetic method the\ngrounds which are employed to explain _a priori_ knowledge are such as\nalso at the same time suffice to prove its validity. In the analytic\nmethod they are grounds of explanation, but not of proof. They are\nthemselves proved only in so far as the assumption of validity is\npreviously granted.\n\nThe analytic procedure which is involved in the complete synthetic\nmethod ought, however, for the sake of clearness, to be classed as a\nseparate, third, method. And as such I shall henceforth regard it. It\nestablishes by an independent line of argument the existence of _a\npriori_ factors, and also their objective validity as conditions\nnecessary to the very possibility of experience. So viewed, it is the\nmost important and the most fundamental of the three methods. The\nargument which it embodies constitutes the very heart of the _Critique_.\nIt is, indeed, Kant\u2019s new transcendental method; and in the future, in\norder to avoid confusion with the analytic method of the _Prolegomena_,\nI shall refer to it always by this title. It is because the\ntranscendental method is an integral part of the complete, synthetic\nmethod, but cannot be consistently made a part of the analytic method,\nthat the synthetic method alone serves as an adequate expression of the\nKantian standpoint. This new transcendental method is proof by reference\nto the possibility of experience. Experience is given as psychological\nfact. The conditions which can alone account for it, as psychological\nfact, also suffice to prove its objective validity; but at the same time\nthey limit that validity to the phenomenal realm.\n\nWe have next to enquire to what extent these methods are consistently\nemployed in the _Critique_. This is a problem over which there has been\nmuch controversy, but which seems to have been answered in a quite final\nmanner by Vaihinger. It is universally recognised that the _Critique_\nprofesses to follow the synthetic method, and that the _Prolegomena_,\nfor the sake of a simpler and more popular form of exposition, adopts\nthe analytic method. How far these two works live up to their\nprofessions, especially the _Critique_ in its two editions, is the only\npoint really in question. Vaihinger found two diametrically opposed\nviews dividing the field. Paulsen, Riehl, and Windelband maintain the\nview that Kant starts from the fact that mathematics, pure natural\nscience, and metaphysics contain synthetic _a priori_ judgments claiming\nto be valid. Kant\u2019s problem is to test these claims; and his answer is\nthat they are valid in mathematics and pure natural science, but not in\nmetaphysics. Paulsen, and those who follow him, further contend that in\nthe first edition this method is in the main consistently held to, but\nthat in the second edition, owing to the occasional employment\n(especially in the _Introduction_) of the analytic method of the\n_Prolegomena_, the argument is perverted and confused: Kant assumes what\nhe ought first to have proved. Fischer, on the other hand, and in a\nkindred manner also B. Erdmann, maintain that Kant never actually\ndoubted the validity of synthetic _a priori_ judgments; starting from\ntheir validity, in order to explain it, Kant discovers the conditions\nupon which it rests, and in so doing is able to show that these\nconditions are not of such a character as to justify the professed\njudgments of metaphysics.\n\nVaihinger[177] combines portions of both views, while completely\naccepting neither. Hume\u2019s profound influence upon the development and\nformulation of Kant\u2019s Critical problem can hardly be exaggerated, but it\nought not to prevent us from realising that this problem, _in its first\nform_, was quite independently discovered. As the letter of 1772 to Herz\nclearly shows,[178] Kant was brought to the problem, how an idea in us\ncan relate to an object, by the inner development of his own views,\nthrough reflection upon the view of thought which he had developed in\nthe _Dissertation_ of 1770. The conformity between thought and things is\nin that letter presented, not as a sceptical objection, but as an actual\nfact calling for explanation. He does not ask whether there is such\nconformity, but only how it should be possible. Even after the further\ncomplication, that thought is synthetic as well as _a priori_, came into\nview through the influence of Hume, the problem still continued to\npresent itself to Kant in this non-sceptical light. And this largely\ndetermines the wording of his exposition, even in passages in which the\ndemands of the synthetic method are being quite amply fulfilled. Kant,\nas it would seem, never himself doubted the validity of the mathematical\nsciences. But since their validity is not beyond possible impeachment,\nand since metaphysical knowledge, which is decidedly questionable,\nwould appear to be of somewhat similar type, Kant was constrained to\nrecognise that, from the point of view of strict proof, such assumption\nof validity is not really legitimate. Though, therefore, the analytic\nmethod would have resolved Kant\u2019s own original difficulty, only the\nsynthetic method is fully adequate to the situation.\n\nKant accordingly sets himself to prove that whether or not we are ready\n(as he himself is) to recognise the validity of scientific judgments,\nthe correctness of this assumption can be firmly established. And being\nthus able to prove its correctness, he for that very reason does not\nhesitate to employ it in his introductory statement. The problem, he\nsays, is that of \u2018understanding\u2019 how synthetic _a priori_ judgments can\nbe valid. A \u2018difficulty,\u2019 a \u2018mystery,\u2019 a \u2018secret,\u2019 lies concealed in\nthem. How can a predicate be ascribed to a subject term which does not\ncontain it? And even more strangely (if that be possible), how can _a\npriori_ judgments legislate for objects which are independent\nexistences? Such judgments, even if valid beyond all disputing, would\nstill call for explanation. This is, indeed, Kant\u2019s original and ground\nproblem. As already indicated, no one, save only Hume, had hitherto\nperceived its significance. Plato, Malebranche, and Crusius may have\ndwelt upon it, but only to suggest explanations still stranger and more\nmystical than the mysterious fact itself.[179]\n\nPaulsen is justified in maintaining that Kant, in both editions of the\n_Critique_, recognises the validity of mathematics and pure natural\nscience. The fact of their validity is less explicitly dwelt upon in the\nfirst edition, but is none the less taken for granted. The sections\ntransferred from the _Prolegomena_ to the _Introduction_ of the second\nedition make no essential change, except merely in the emphasis with\nwhich Kant\u2019s belief in the existence of valid _a priori_ synthetic\njudgments is insisted upon. As has already been stated, only by virtue\nof this initial assumption is Kant in position to maintain that there is\nan alternative to the strict synthetic method. The _problem_ from which\nhe starts is common to both methods, and for that reason the formulation\nused in the _Prolegomena_ can also be employed in the _Introduction_ to\nthe _Critique_. Only in their manner of solving the problem need they\ndiffer.[180] Kant\u2019s Critical problem first begins with this\npresupposition of validity, and does not exist save through it.[181] He\ndoes not first seek to discover whether such judgments are valid, and\nthen to explain them. He accepts them as valid, but develops a method of\nargument which suffices for proof as well as for explanation. The\nargument being directed to both points simultaneously, and establishing\nboth with equal cogency, it may legitimately be interpreted in either\nway, merely as explanation, or also as proof. Kant does not profess or\nattempt to keep exclusively to any one line of statement. Against the\ndogmatists he insists upon the necessity of _explaining_ the validity of\n_a priori_ synthetic judgments, against the sceptics upon the\npossibility of _proving_ their validity. And constantly he uses\nambiguous terms, such as \u2018justification\u2019 (_Rechtfertigung_),\n\u2018possibility,\u2019 that may indifferently be read in either sense. But\nthough the fundamental demand which characterises the synthetic method\nin its distinction from the analytic thus falls into the background, and\nis only occasionally insisted upon, it is none the less fulfilled. So\nfar as regards the main argument of the _Critique_ in either edition,\nthe validity of synthetic _a priori_ judgments is not required as a\npremiss. It is itself independently proved.\n\nThe manner in which Kant thus departs from the strict application of the\nsynthetic method may be illustrated by an analysis of his argument in\nthe _Aesthetic_.[182] Only in the arguments of the first edition in\nregard to space and time is the synthetic method employed in its ideal\nand rigorous form. For the most part, even in the first edition, instead\nof showing how the _a priori_ character of pure and applied mathematics\nfollows from conclusions independently established, he assumes both pure\nand applied mathematics to be given as valid, and seeks only to show how\nthe independently established results of the _Aesthetic_ enable him to\nexplain and render comprehensible their recognised characteristics. This\nis not, indeed, any very essential modification of the synthetic method;\nfor his independently established results suffice for deducing all that\nthey are used to explain. The validity of mathematics is not employed as\na premiss. Kant\u2019s argument is, however, made less clear by the above\nprocedure.\n\nFurther difficulty is caused by Kant\u2019s occasional employment, even in\nthe first edition, of the analytic method. He several times cites as an\nargument in support of his view of space the fact that it alone will\naccount for the existing science of geometry. That is to say, he employs\ngeometry, viewed as valid, to _prove_ the correctness of his view of\nspace.[183] Starting from that science as given, he enquires what are\nthe conditions which can alone render it possible. These conditions are\nfound to coincide with those independently established. Now this is a\nvalid argument when employed in due subordination to the main synthetic\nmethod. It offers welcome _confirmation_ of the results of that method.\nIt amounts in fact to this, that having proved (by application of the\ntranscendental method) the mathematical sciences to be valid, everything\nwhich their validity necessarily implies must be granted. Kant\u2019s\nreasoning here becomes circular, but it is none the less valid on that\naccount. This further complication of the argument is, however,\ndangerously apt to mislead the reader. It is in great part the cause of\nthe above division among Kant\u2019s commentators. The method employed in the\n_Prolegomena_ is simply this form of argument systematised and cut free\nfrom all dependence upon the transcendental method of proof.[184]\n\nThe whole matter is, however, still further complicated by the\ndistinction, which we have already noted, between real and ideal\npossibility. Are the given synthetic _a priori_ judgments valid? That is\none question. Can the Critical philosophy discover, completely\nenumerate, and prove in a manner never before done, all the possible\nsynthetic _a priori_ principles? That is a very different problem, and\nwhen raised brings us to the further discussion of Kant\u2019s transcendental\nmethod. The question at issue is no longer merely whether or not certain\ngiven judgments are valid, and how, if valid, they are to be accounted\nfor. The question is now that of discovering and of proving principles\nwhich have not been established by any of the special sciences. This\nshifting of the problem is concealed from Kant himself by his omission\nto distinguish between the undemonstrated axioms of the mathematical\nsciences and their derivative theorems, between the principles employed\nby the physicist without enquiry into their validity and the special\nlaws based upon empirical evidence.\n\nAs regards the mathematical axioms, the problem is fairly simple. As we\nshall see later, in the _Aesthetic_, they do not require a deduction in\nthe strict transcendental sense. They really fall outside the\napplication of the transcendental method. They require only an\n\u201cexposition.\u201d But in regard to the fundamental principles of natural\nscience we are presented with the problem of discovery as well as of\nproof. Unlike the axioms of the mathematician, they are frequently left\nunformulated. And many postulates, such as that there is a _lex continui\nin natura_, are current in general thought, and claim equal validity\nwith the causal principle. Kant has thus to face the question whether in\naddition to those principles employed more or less explicitly by the\nscientist, others, such as might go to form an immanent metaphysics of\nnature, may not also be possible.\n\nB. (_a_)[185] =Psychological and logical possibility.=--Both have to be\nrecognised and accounted for. Let us consider each in order.\n\n(1) =Psychological possibility.=--What are the _subjective_ conditions of\n_a priori_ synthetic judgments? _Through what mental faculties_ are they\nrendered possible? Kant replies by developing what may be called a\ntranscendental psychology. They depend upon space and time as forms of\nsensibility, upon the _a priori_ concepts of understanding, and upon the\nsynthetic activities by which the imagination schematises these concepts\nand reduces the given manifold to the unity of apperception. This\ntranscendental psychology is the necessary complement of the more purely\nepistemological analysis.[186] But on this point Kant\u2019s utterances are\nextremely misleading. His Critical enquiry has, he declares, nothing in\ncommon with psychology. In the _Preface_ to the first edition we find\nthe following passage: \u201cThis enquiry ... [into] the pure understanding\nitself, its possibility and the cognitive faculties upon which it rests\n..., although of great importance for my chief purpose, does not form an\nessential part of it.\u201d[187] The question, he adds, \u201chow is the faculty\nof thought itself possible?... is as it were a search for the cause of a\ngiven effect, and therefore is of the nature of an hypothesis [or \u2018mere\nopinion\u2019], though, as I shall show elsewhere, this is not really so.\u201d\nThe concluding words of this passage very fairly express Kant\u2019s\nhesitating and inconsistent procedure. Though he has so explicitly\neliminated from the central enquiry of the _Critique_ all psychological\ndetermination of the mental powers, statements as to their constitution\nare none the less implied, and are involved in his epistemological\njustification alike of _a priori_ knowledge and of ordinary experience.\nIf we bear in mind that Kant is here attempting to outline the possible\ncauses of given effects, and that his conclusions are therefore\nnecessarily of a more hypothetical character than those obtained by\nlogical analysis, we shall be prepared to allow him considerable liberty\nin their formulation. But in certain respects his statements are precise\nand definite--the view, for instance, of sensations as non-spatial, of\ntime as a form of inner sense, of the productive imagination as\npre-conditioning our consciousness, of spontaneity as radically distinct\nfrom receptivity, of the pure forms of thought as not acquired through\nsense, etc. No interpretation which ignores or under-estimates this\npsychological or subjective aspect of his teaching can be admitted as\nadequate.[188]\n\n(2) =Logical or epistemological possibility.=--How can synthetic _a\npriori_ judgments be _valid_? This question itself involves a twofold\nproblem. How, despite their synthetic character, can they possess truth,\n_i.e._ how can we pass from their subject terms to their predicates? And\nsecondly, how, in view of their origin in our human reason, can they be\nobjectively valid, _i.e._ legislate for the independently real? How can\nwe pass beyond the subject-predicate relation to real things? This\nlatter is the Critical problem in the form in which it appears in Kant\u2019s\nletter of 1772 to Herz.[189] The former is the problem of synthesis\nwhich was later discovered.\n\n(_b_) (1) =Possibility of explanation and= (2) =possibility of\nexistence.=--(1) How can synthetic _a priori_ judgments be _accounted\nfor_? How, despite their seemingly inconsistent and apparently\nparadoxical aspects, can their validity (their validity as well as their\nactuality being taken for granted) be rendered _comprehensible_? (2) The\nvalidity of such judgments has been called in question by the\nempiricists, and is likewise inexplicable even from the dogmatic\nstandpoint of the rationalists. How, then, can these judgments _be\npossible at all_? These two meanings of the term \u2018possible\u2019 connect with\nthe ambiguity, above noted, in the term \u2018how.\u2019 The former problem can be\nsolved by an analytic method; the latter demands the application of the\nmore radical method of synthetic reconstruction.\n\n(_c_) =Real and ideal possibility.=[190]--We have to distinguish between\nthe possible validity of those propositions which the mathematical and\nphysical sciences profess to have established and the possible validity\nof those principles such as that of causality, which are postulated by\nthe sciences, but which the sciences do not attempt to prove, and which\nin certain cases they do not even formulate. The former constitute an\nactually existent body of scientific knowledge, demonstrated in\naccordance with the demands of scientific method. The latter are\nemployed by the scientist, but are not investigated by him. The science\ninto which they can be fitted has still to be created; and though some\nof the principles composing it may be known, others remain to be\ndiscovered. All of them demand such proof and demonstration as they have\nnever yet received.[191] This new and ideal science is the scientific\nmetaphysics which Kant professes to inaugurate by means of the\n_Critique_. In reference to the special sciences, possibility means the\nconditions of the actually given. In reference to the new and ideal\nmetaphysics, possibility signifies the conditions of the realisation of\nthat which is sought. In view of this distinction, the formula--How are\nsynthetic _a priori_ judgments possible?--will thus acquire two very\ndifferent meanings. (1) How are the existing _a priori_ synthetic\njudgments to be accounted for? (2) How may all the really fundamental\njudgments of that type be exhaustively discovered and proved? Even in\nregard to immanent metaphysics Kant interprets the formula in both ways.\nThis is due to his frequent confusion of immanent metaphysics with the\nprinciples of natural science. Its propositions are then regarded as\ngiven, and only their general validity calls for proof. It is, however,\nin the problem of ideal possibility that the essential problem of the\n_Critique_ lies; and that is a further reason why it cannot be\nadequately dealt with, save by means of the synthetic method.\n\n       *       *       *       *       *\n\n=Experience.=--Throughout the _Introduction_ the term _experience_[192]\nhas (even at times in one and the same sentence) two quite distinct\nmeanings, (1) as product of sense and understanding acting\nco-operatively, and (2) as the raw material (the impressions) of sense.\nConsiderable confusion is thereby caused.\n\n=Understanding and reason=[193] are here, as often elsewhere in the\n_Critique_, used as equivalent terms. Throughout the entire two first\nsections of the _Introduction_ to the second edition the term reason\ndoes not occur even once. As first mentioned,[194] it is taken as the\nsource of metaphysical judgments.\n\n=General (a priori) truths have an inner necessity and must be clear and\ncertain by themselves.=[195]--These statements are not in accordance with\nKant\u2019s new Critical teaching.[196] They have remained uncorrected from a\nprevious way of thinking. This must be one reason for the recasting of\nthis paragraph in the second edition.\n\n=Even with (unter) our experiences there is mingled knowledge which must\nbe of a priori origin.=[197]--Kant is here distinguishing the _immanent a\npriori_, such as that involved in any causal judgment, from the\n_transcendent a priori_ dwelt upon in the next paragraph. The latter is\nexpressed through metaphysical judgments, such as \u2018God exists,\u2019 \u2018the\nsoul is immortal.\u2019\n\n=Original concepts and judgments derived from them.=[198]--Cf. B 5-6.\n\n=Pure.=--In the title of the section the term _pure_[199] (_rein_) is, as\nthe subsequent argument shows, taken as exactly equivalent to _a\npriori_. As Vaihinger notes, the adjective _apriorisch_ had not yet been\ninvented. The opposite of pure is here empirical (_empirisch_).[200]\n\n=All our knowledge begins with experience.=[201]--This is a stronger\nstatement than any in the corresponding paragraphs of the first edition.\nHad Kant proceeded to develop its consequences, he would have had to\nrecast the entire _Introduction_, setting the problem of empirical\nknowledge alongside that of the _a priori_.[202] As it is, he is\nforced[203] to subdivide the absolutely _a priori_ into the pure and the\nmixed.[204]\n\n=By objects which affect (r\u00fchren) our senses. The raw material of\nsensuous impressions.=[205]--These incidental statements call for\ndiscussion. Cf. below, pp. 80-8, 120-1, 274 ff.\n\n=A knowledge of objects which we call experience.=[206]--Kant does not\nkeep to this definition. The term experience is still used in its other\nand narrower sense, as in the very next paragraph, when Kant states that\nknowledge does not, perhaps, arise solely from experience (= sense\nimpressions).\n\n=In respect of time.=[207]--This statement, taken as an account of Kant\u2019s\nteaching in the _Critique_, is subject to two reservations. In the\n_Aesthetic_[208] Kant sometimes claims a temporal antecedence for the _a\npriori_. And secondly, the _a priori_ is not for Kant merely logical. It\nalso possesses a dynamical priority.[209]\n\n=Even experience itself is a compound.=[210]--The \u201ceven\u201d seems to refer to\nthe distinction drawn in A 2 between the immanent and the transcendent\n_a priori_.[211]\n\n=It is therefore a question whether there exists such knowledge\nindependent of experience.=[212]--This question was not raised in the\nfirst edition.[213] The alternative methods, analytic and synthetic, are\ndiscussed above, p. 44 ff.\n\n=Such knowledge is called a priori and is distinguished from empirical\nknowledge.=[214]--Throughout the _Introduction_, in both editions\nequally, Kant fails to state the problems of the _Critique_ in a\nsufficiently comprehensive manner. He speaks as if the _Critique_ dealt\nonly with the absolutely _a priori_, in its two forms, as immanent\nscientific knowledge and as transcendent speculation. It also deals with\nthe equally important and still more fundamental problem of accounting\nfor the possibility of _experience_.[215] Our empirical knowledge\ninvolves an _a priori_ element, and may not therefore be opposed to _a\npriori_ knowledge in the manner of the passage before us.\n\n=This term a priori is not yet definite enough.=[216]--It is frequently\nemployed in a merely relative sense. Thus we can say of a person who\nundermines the foundations of his house that he might have known _a\npriori_ that it would collapse, that is, that he need not wait for the\nexperience of its actual fall. But still he could not know this entirely\n_a priori_; he had first to learn from experience that bodies are heavy,\nand will fall when their supports are taken away. But as dealt with in\nthe _Critique_ the term _a priori_ is used in an absolute sense, to\nsignify that knowledge which is independent, not of this or that\nexperience only, but of all impressions of the senses. Thus far Kant\u2019s\nposition is comparatively clear; but he proceeds to distinguish two\nforms within the absolutely _a priori_, namely, mixed and pure. The\nabsolutely _a priori_ is mixed when it contains an empirical element,\npure when it does not. (\u201cPure\u201d is no longer taken in the meaning which\nit has in the title of the section.[217] It signifies not the _a priori_\nas such, but only one subdivision of it.) Thus after defining absolutely\n_a priori_ knowledge as independent of all experience, Kant takes it in\none of its forms as involving empirical elements. The example which he\ngives of an absolutely _a priori_ judgment, which yet is not pure, is\nthe principle: every change has its cause. \u201cChange\u201d is an empirical\nconcept, but the synthetic relation asserted is absolutely _a priori_.\nIn the next section[218] this same proposition is cited as a _pure_\njudgment _a priori_--\u201cpure\u201d being again used in its more general meaning\nas synonymous with _a priori_. This confusion results from Kant\u2019s\nexclusive preoccupation with the _a priori_, and consequent failure to\ngive due recognition to the correlative problem of the empirical\njudgment. The omitted factor retaliates by thus forcing its way into\nKant\u2019s otherwise clean-cut divisions. Also, it is not true that the\nrelative _a priori_ falls outside the sphere of the Critical enquiry.\nSuch judgment expresses necessity or objectivity, and for that reason\ndemands a transcendental justification no less urgently than the\nabsolutely _a priori_. The finding of such justification is, indeed, the\ncentral problem of the _Analytic_.[219]\n\nThe subdivisions of the _a priori_ may be tabulated thus:\n\n                      { Relative, _e.g._ every unsupported house must\n                      { fall.\n  _A priori_ knowledge{          { Mixed, _e.g._ every change has its cause.\n                      { Absolute { Pure, _e.g._ a straight line is the shortest\n                      {          { distance between two points.\n\nThe term _pure_ (_rein_) thus acquires a second meaning distinct from\nthat defined above.[220] It is no longer employed as identical with _a\npriori_, but as a subdivision of it, meaning _unmixed_. Its opposite is\nno longer the empirical, but the impure or mixed. Owing, however, to the\nfact that \u201cpure\u201d (in its first meaning) is identical with the _a\npriori_, it shares in all the different connotations of the latter, and\naccordingly is also employed to denote that which is _not relative_. But\n\u201cpure\u201d has yet another meaning peculiar to itself. The phrase\n\u201cindependent of experience\u201d has in reference to \u201cpure\u201d an ambiguity from\nwhich it does not suffer in its connection with \u201c_a priori_\u201d (since\nmathematical knowledge, whether pure or applied, is always regarded by\nKant as _a priori_). It may signify either independence as regards\n_content and validity_, or independence as regards _scope_. The latter\nmeaning is narrower than the former. By the former meaning it denotes\nthat which originates, and can possess truth, independently of\nexperience. By the latter it signifies that which is not only\nindependent of sense but also applies to the non-sensuous. In this\nlatter meaning pure knowledge therefore signifies transcendent\nknowledge. Its opposite is the immanent. The various meanings of \u201cpure\u201d\n(four in number) may be tabulated as follows:\n\n  (_a_) (1) _A priori_: independent of experience as regards origin\n                and validity. (Its opposite = empirical.)\n\n               { (2) Absolutely independent of experience. (Its\n               { opposite = relative.)\n\n               { (3) Unmixed with experience. (Its opposite =\n               { impure or mixed.)\n\n  (_b_) (4) Independent of experience as regards scope = transcendent.\n                     (Its opposite = immanent.)\n\nAll these varied meanings contribute to the ambiguity of the title of\nthe _Critique_. Kant himself employs the title in all of the following\nsenses:\n\n1. Critique of absolutely pure _a priori_ knowledge, determination of\nits sources, conditions, scope and limits.\n\n2. Critique of all _a priori_ knowledge, relative as well as absolute,\nin so far as it depends upon _a priori_ principles, determination, etc.\n\n3. Critique of all knowledge, whether _a priori_ or empirical,\ndetermination, etc.\n\n4. Critique of transcendent knowledge, its sources and limits.\n\nFurther meanings could also be enumerated but can be formulated by the\nreader for himself in the light of the ambiguities just noted.[221] The\nspecial context in each case can alone decide how the title is to be\nunderstood. If a really adequate definition of the purpose and scope of\nthe _Critique_ is sought by the reader, he must construct it for\nhimself. The following may perhaps serve. _The_ Critique _is an enquiry\ninto the sources, conditions, scope and limits of our knowledge, both a\npriori and empirical, resulting in the construction of a new system of\nimmanent metaphysics; in the light of the conclusions thus reached, it\nalso yields an analysis and explanation of the transcendental illusion\nto which transcendent metaphysics, both as a natural disposition and as\na professed science, is due._\n\nKant further complicates matters by offering a second division of the\nabsolutely _a priori_,[222] viz. into the original and the derivative.\nAlso, by implication, he classes relative _a priori_ judgments among the\npropositions to be reckoned with by the _Critique_; and yet in B 4 he\nspeaks of the proposition, all bodies are heavy, as merely\nempirical.[223]\n\n=A criterion.=[224]--Necessity and universality are valid criteria of the\n_a priori_ (= the non-empirical). This follows from Kant\u2019s view[225] of\nthe empirical as synonymous with the contingent (_zuf\u00e4llig_). Experience\ngives only the actual; the _a priori_ alone yields that which cannot be\notherwise.\n\n     \u201cNecessity and strict universality are thus safe criteria of _a\n     priori_ knowledge, and are inseparable from one another. But since\n     in the employment of these criteria the empirical limitation of\n     judgments is sometimes more easily shown than their contingency,\n     or since, as frequently happens, their unlimited universality can\n     be more convincingly proved than their necessity, it is advisable\n     to use the two criteria separately, each being by itself\n     infallible.\u201d[226]\n\nNow Kant is here, of course, assuming the main point to be established,\nnamely, that experience is incapable of accounting for such universality\nand necessity as are required for our knowledge, both ordinary and\nscientific. We have already considered this assumption,[227] and have\nalso anticipated misunderstanding by noting the important qualifications\nto which, from Kant\u2019s new Critical standpoint, the terms \u2018necessity\u2019 and\n\u2018universality\u2019 become subject.[228] The very specific meaning in which\nKant employs the term _a priori_ must likewise be borne in mind. Though\nnegatively the _a priori_ is independent of experience, positively it\noriginates in our human reason. The necessity and universality which\ndifferentiate the _a priori_ distinguish it only from the humanly\naccidental. The _a priori_ has no absolute validity. From a metaphysical\nstandpoint, it is itself contingent. As already stated,[229] all truth\nis for Kant merely _de facto_. The necessary is not that which cannot be\nconceived to be otherwise, nor is it the unconditioned. Our reason\nlegislates only for the world of appearance. But as yet Kant gives no\nhint of this revolutionary reinterpretation of the rationalist criteria.\nOne of the chief unfortunate consequences of the employment in this\n_Introduction_ of the analytic method of the _Prolegomena_ is that it\ntends to mislead the reader by seeming to commit Kant to a logical _a\npriori_ of the Leibnizian type.\n\n=To show that, if experience is to be possible, [pure a priori\npropositions] are indispensable, and so to prove their existence a\npriori.=[230]--At first sight Kant would seem to be here referring to the\nalternative synthetic method of procedure, _i.e._ to the\n_transcendental_ proof of the _a priori_. The next sentence shows,\nhowever, that neither in intention nor in fact is that really so. He\nargues only that _a priori_ principles, such as the principle of\ncausality, are necessary in order to give \u201ccertainty\u201d to our experience;\nsuch a principle must be postulated if inductive inference is to be\nvalid. Experience could have no [scientific] certainty, \u201cif all rules\naccording to which it proceeds were themselves in turn empirical, and\ntherefore contingent. They could hardly be regarded as first\nprinciples.\u201d There is no attempt here to prove that empirical knowledge\n_as such_ necessarily involves the _a priori_. Also the method of\nargument, though it seeks to establish the _necessity_ of the _a\npriori_, is not transcendental or Critical in character. It is merely a\nrepetition of the kind of argument which both Hume and Leibniz had\nalready directed against the sensationalist position.[231] Very\nstrangely, considering that these sentences have been added in the\nsecond edition, and therefore subsequent to the writing of the objective\ndeduction, Kant gives no indication of the deeper problem to which he\nfinally penetrated. The explanation is, probably, that to do so would\nhave involved the recasting of the entire _Introduction_. Even on the\nbriefest reference, the hard-and-fast distinction between the _a priori_\nand the empirical, as two distinct and separate classes of judgment,\nwould have been undermined, and the reader would have been made to feel\nthe insufficiency of the analysis upon which it is based.[232] The\nexistence of the deeper view is betrayed only through careless\nemployment of the familiar phrase \u201cpossibility of experience.\u201d For, as\nhere used, it is not really meant. \u201cCertainty of experience\u201d--a very\ndifferent matter--is the meaning that alone will properly fit the\ncontext.\n\n=Reason and understanding.=[233]--They are here distinguished, having been\nhitherto, in A 1-2, employed as synonymous. The former carries us beyond\nthe field of all possible experience; the latter is limited to the world\nof sense. Thus both _Reason_ and _understanding_ are here used in their\nnarrowest meaning.\n\n=These inevitable problems of pure Reason itself are God, freedom, and\nimmortality. The science which, with all its methods, is in its final\nintention directed solely to the solution of these problems, is called\nmetaphysics.=[234]--These sentences are characteristic of the second\nedition with its increased emphasis upon the positive results of the\n_Critique_ on the one hand, and with its attitude of increased favour\ntowards transcendent metaphysics on the other. The one change would seem\nto be occasioned by the nature of the criticisms passed upon the first\nedition, as, for instance, by Moses Mendelssohn who describes Kant as\n\u201cthe all-destroyer\u201d (_der alles zermalmende_). The other is due to\nKant\u2019s preoccupation with the problems of ethics and of teleology. The\nabove statements are repeated with even greater emphasis in B 395\n_n_.[235] The definition here given of metaphysics is not strictly kept\nto by Kant. As above noted,[236] Kant really distinguishes within it two\nforms, immanent and transcendent. In so doing, however, he still[237]\nregards transcendent metaphysics as the more important. Immanent\nmetaphysics is chiefly of value as contributing to the solution of the\n\u201cinevitable problems of pure Reason.\u201d\n\n=A 3-4 = B 7-8.=--The reasons, here cited by Kant, for the failure of\nphilosophical thinking to recognise the difference between immanent and\ntranscendent judgments are: (1) the misunderstood character, and\nconsequent misleading influence, of _a priori_ mathematical judgments;\n(2) the fact that once we are beyond the sensible sphere, experience can\nnever contradict us; (3) natural delight in the apparent enlargement of\nour knowledge; (4) the ease with which logical contradictions can be\navoided; (5) neglect of the distinction between analytic and synthetic\n_a priori_ judgments. Vaihinger points out[238] that in the\n_Fortschritte_[239] Kant adds a sixth reason--confusion of the concepts\nof understanding with the Ideas of Reason. Upon the first of the above\nreasons the best comment is that of the _Methodology_.[240] But the\nreader must likewise bear in mind that in B xvi Kant develops his new\nphilosophical method on the analogy of the mathematical method. The\nlatter is, he claims, _mutatis mutandis_, the true method of\n_legitimate_ speculation, _i.e._ of immanent metaphysics. The one\nessential difference (as noted by Kant[241]), which has been overlooked\nby the dogmatists, is that philosophy gains its knowledge from concepts,\nmathematics from the construction of concepts.\n\n=Remain investigations only.=[242]--Cf. _Prolegomena_, \u00a7 35.\n\n=The analysis of our concepts of objects.=[243]--Vaihinger\u2019s\ninterpretation, that the concepts here referred to are those which we\n\u201cform _a priori_ of things,\u201d[244] seems correct.[245] The rationalists\nsought to deduce the whole body of rational psychology from the _a\npriori_ conception of the soul as a simple substance, and of rational\ntheology from the _a priori_ conception of God as the all-perfect Being.\n\n     =Analytic and synthetic judgments.=[246]--\u201cAll analytic judgments\n     depend wholly on the law of contradiction, and are in their nature\n     _a priori_ cognitions, whether the concepts that supply them with\n     matter be empirical or not. For the predicate of an affirmative\n     analytic judgment is already contained in the concept of the\n     subject, of which it cannot be denied without contradiction. In the\n     same way its opposite is necessarily denied of the subject in an\n     analytic, but negative, judgment by the same law of\n     contradiction.... For this very reason all analytic judgments are\n     _a priori_ even when the concepts are empirical, as, for example,\n     gold is a yellow metal; for to know this I require no experience\n     beyond my concept of gold as a yellow metal: it is, in fact, the\n     very concept, and I need only analyse it, without looking beyond it\n     elsewhere.... [Synthetic judgments, _a posteriori_ and _a priori_]\n     agree in this, that they cannot possibly spring solely from the\n     principle of analysis, the law of contradiction. They require a\n     quite different principle. From whatever they may be deduced, the\n     deduction must, it is true, always be in accordance with the\n     principle of contradiction. For that principle must never be\n     violated. But at the same time everything cannot be deduced from\n     it.\u201d[247]\n\nIn A 594 = B 622 analytic judgments are also spoken of as identical; but\nin the _Fortschritte_[248] this use of terms is criticised:\n\n     \u201cJudgments are analytic if their predicate only represents clearly\n     (_explicite_) what was thought obscurely (_implicite_) in the\n     concept of the subject, _e.g._ all bodies are extended. Were we to\n     call such judgments identical only confusion would result. For\n     identical judgments contribute nothing to the clearness of the\n     concept, and that must be the purpose of all judging. Identical\n     judgments are therefore empty, _e.g._ all bodies are bodily (or to\n     use another term material) beings. Analytic judgments do, indeed,\n     ground themselves upon identity and can be resolved into it; but\n     they are not identical. For they demand analysis and serve for the\n     explanation of the concept. In identical judgments, on the other\n     hand, _idem_ is defined _per idem_, and nothing at all is\n     explained.\u201d\n\nVaihinger[249] cites the following contrasted examples of analytic and\nsynthetic judgments:\n\n_Analytic._--(_a_) Substance is that which exists only as subject in\nwhich qualities inhere.[250] (_b_) Every effect has a cause.[5] (_c_)\nEverything conditioned presupposes a condition.\n\n_Synthetic._--(_a_) Substance is permanent. (_b_) Every event has a\ncause.[251] (=c=) Everything conditioned presupposes an unconditioned.\n\n=B 11-12.=--The first half of this paragraph is transcribed practically\nword for word from the _Prolegomena_.[252] The second half is a close\nrestatement of an omitted paragraph of the first edition. The chief\naddition lies in the concluding statement, that \u201cexperience is itself a\nsynthetic connection of intuitions.\u201d This is in keeping with statements\nmade in the deduction of the categories in the second edition,[253] and\nin the paragraph inserted in the proof of the second analogy in the\nsecond edition.[254] The _x_ has strangely been omitted in the second\nedition in reference to empirical judgments, though retained in\nreference to synthetic _a priori_ judgments.\n\n=The proposition: everything which happens has its cause.=[255]--As we\nhave already observed,[256] Hume influenced Kant at two distinct periods\nin his philosophical development--in 1756-1763, and again at some time\n(not quite definitely datable) after February 1772. The first influence\nconcerned the character of concrete causal judgments; the second related\nto the causal axiom. Though there are few distinctions which are more\nimportant for understanding the _Critique_ than that of the difference\nbetween these two questions, it has nowhere been properly emphasised by\nKant, and in several of the references to Hume, which occur in the\n_Critique_ and in the _Prolegomena_, the two problems are confounded in\na most unfortunate manner. The passages in the _Introduction_[257] are\nclear and unambiguous; the influence exercised by Hume subsequent to\nFebruary 1772 is quite adequately stated. The causal axiom claims to be\n_a priori_, and is, as Hume asserts, likewise synthetic. Consequently\nthere are only two alternatives, each decisive and far-reaching. Either\nvalid _a priori_ synthesis must, contrary to all previous philosophical\nbelief, be possible, or \u201ceverything which we call metaphysics must turn\nout to be a mere delusion of reason.\u201d The solution of this problem is \u201ca\nquestion of life and death to metaphysics.\u201d To this appreciation of\nHume, Kant adds criticism. Hume did not sufficiently universalise his\nproblem. Had he done so, he would have recognised that pure mathematics\ninvolves _a priori_ synthesis no less necessarily than do the\nmetaphysical disciplines. From denying the possibility of mathematical\nscience \u201chis good sense would probably have saved him.\u201d Hume\u2019s problem,\nthus viewed, finds its final and complete expression in the formula: How\nare synthetic _a priori_ judgments possible?\n\nIn A 760 = B 788 the account differs in two respects: first, it\ndiscusses the metaphysical validity of the causal axiom as well as its\nintrinsic possibility as a judgment; and secondly, reference is made to\nthe conception of causality as well as to the axiom. The implied\ncriticism of Hume is correspondingly modified. Otherwise, it entirely\nharmonises with the passages in the _Introduction_.\n\n     \u201cHume dwelt especially upon the principle of causality, and quite\n     rightly observed that its truth, and even the objective validity of\n     the concept of efficient cause in general, is based on no insight,\n     _i.e._ on no _a priori_ knowledge, and that its authority cannot\n     therefore be ascribed to its necessity, but merely to its general\n     utility in the course of experience and to a certain subjective\n     necessity which it thereby acquires, and which he entitles custom.\n     From the incapacity of our reason to make use of this principle in\n     any manner that transcends experience he inferred the nullity of\n     all pretensions of reason to advance beyond the empirical.\u201d\n\nNow so far, in these references to Hume, Kant has had in view only the\nproblems of mathematical and physical science and of metaphysics. The\nproblems involved in the possibility of empirical knowledge are left\nentirely aside. His account of Hume\u2019s position and of his relation to\nHume suffers change immediately these latter problems are raised. And\nunfortunately it is a change for the worse. The various problems treated\nby Hume are then confounded together, and the issues are somewhat\nblurred. Let us take the chief passages in which this occurs. In A 764 =\nB 792 ff. Kant gives the following account of Hume\u2019s argument. Hume,\nrecognising the impossibility of predicting an effect by analysis of the\nconcept of the cause, or of discovering a cause from the concept of the\neffect, viewed all concrete causal judgments as merely contingent, and\ntherefrom inferred the contingency of the causal axiom. In so doing\nHume, Kant argues, confuses the legitimate and purely _a priori_\ninference from a given event to _some_ antecedent with the very\ndifferent inference, possible only through special experience, to a\n_specific_ cause. Now this is an entire misrepresentation of Hume\u2019s real\nachievement, and may perhaps be explained, at least in part, as being\ndue to the fact that Kant was acquainted with Hume\u2019s _Treatise_ only\nthrough the indirect medium of Beattie\u2019s quotations. Hume committed no\nsuch blunder. He clearly recognised the distinction between the problem\nof the validity of the causal axiom and the problem of the validity of\nconcrete causal judgments. He does not argue from the contingency of\nconcrete causal laws to the contingency of the universal principle, but\nshows, as Kant himself recognises,[258] that the principle is neither\nself-evident nor demonstrable _a priori_. And as necessity cannot be\nrevealed by experience, neither is the principle derivable from that\nsource. Consequently, Hume concludes, it cannot be regarded as\nobjectively valid. It must be due to a subjective instinct or natural\nbelief. (The two problems are similarly confounded by Kant in A 217 = B\n264.)\n\nIn the _Introduction_ to the _Prolegomena_ there is no such confusion\nof the two problems, but matters are made even worse by the omission of\nall reference to Hume\u2019s analysis of the causal axiom. Only Hume\u2019s\ntreatment of the concept of causality is dwelt upon. This is the more\nunfortunate, and has proved the more misleading, in that it is here that\nKant makes his most explicit acknowledgment of his indebtedness to Hume.\nIn \u00a7\u00a7 27 ff. of the _Prolegomena_ both problems reappear, but are again\nconfounded. The section is preceded by sentences in which the problem of\nexperience is emphasised; and in keeping with these prefatory remarks,\nKant represents \u201cHume\u2019s _crux metaphysicorum_\u201d as concerning only the\nconcept of causality (viewed as a synthetic, and professedly _a priori_,\nconnection between concrete existences). Yet in \u00a7 30 the causal axiom is\nalso referred to, and together they are taken as constituting \u201cHume\u2019s\nproblem.\u201d\n\nNow if we bear in mind that Hume awakened Kant to both problems--how _a\npriori_ knowledge is possible, and how experience is possible--this\nconfusion can easily be understood. Kant had already in the early\n\u2018sixties studied Hume with profound admiration and respect.[259] In the\nperiod subsequent to 1772 this admiration had only deepened; and\nconstantly, as we may believe, Kant had returned with fresh relish to\nHume\u2019s masterly analyses of causality and of inductive inference. It is\nnot, therefore, surprising that as the years passed, and as the other\nelements in Hume\u2019s teaching revealed to him, through the inner growth of\nhis own views, their full worth and significance, he should allow the\ncontribution that had more specifically awakened him to fall into the\nbackground, and should, in vague fashion, ascribe to Hume\u2019s teaching as\na whole the specific influence which was really due to one particular\npart. By 1783, the date of the _Prolegomena_, Kant\u2019s first enthusiasm\nover the discovery of the fundamental problem of _a priori_ synthesis\nhad somewhat abated, and the problem of experience had more or less\ntaken its place. This would seem to be the reason why in the\n_Prolegomena_ he thus deals with both aspects of Hume\u2019s problem, and why\nin so doing he gives a subordinate place to Hume\u2019s treatment of the\ncausal axiom. But though the misunderstanding may be thus accounted for,\nit must none the less be deplored. For the reader is seriously misled,\nand much that is central to the Critical philosophy is rendered obscure.\nThe influence which Kant in the _Prolegomena_ thus ascribes to Hume was\nnot that which really awakened him from his dogmatic slumber, but is in\npart that which he had assimilated at least as early as 1763, and in\npart that which acted upon him with renewed force when he was struggling\n(probably between 1778 and 1780) with the problems involved in the\ndeduction of the categories. It was Hume\u2019s treatment of the causal\naxiom, and that alone, which, at some time subsequent to February 1772,\nwas the really effective influence in producing the Copernican\nchange.[260]\n\n=Purely a priori and out of mere concepts.=[261]--Vaihinger\u2019s comment\nseems correct: Kant means only that neither actual experience nor pure\nintuition can be resorted to. This does not contradict the complementary\nassertion,[262] that the principle, everything which happens has its\ncause, can be known _a priori_, not immediately from the concepts\ninvolved in it, but only indirectly[263] through the relation of these\nconcepts to possible experience. \u201cPossible experience,\u201d even though it\nstands for \u201csomething purely contingent,\u201d is itself a concept.\nVaihinger[264] quotes Apelt upon this \u201cmysterious\u201d type of judgment.\n\n     \u201cMetaphysics is synthetic knowledge from mere concepts, not like\n     mathematics from their construction in intuition, and yet these\n     synthetic propositions cannot be known from bare concepts, _i.e._\n     not analytically. The necessity of the connection in those\n     propositions is to be apprehended through thought alone, and yet is\n     not to rest upon the form of thought, the principle of\n     contradiction. The conception of a kind of knowledge which arises\n     from bare concepts, and yet is synthetic, eludes our grasp. The\n     problem is: How can one concept be necessarily connected with\n     another, without also at the same time being contained in it?\u201d\n\nThe paragraphs in B 14 to B 17 are almost verbal transcripts from\n_Prolegomena_, \u00a7 2 _c_, 2 ff.\n\n=Mathematical judgments are one and all (insgesammt)\nsynthetic.=[265]--This assertion is carelessly made, and does not\nrepresent Kant\u2019s real view. In B 16 he himself recognises the existence\nof analytic mathematical judgments, but unduly minimises their number\nand importance.\n\n=All mathematical conclusions proceed according to the principle of\ncontradiction.=[266]--To the objection made by Paulsen that Kant, in\nadmitting that mathematical judgments can be deduced from others by\nmeans of the principle of contradiction, ought consistently to have\nrecognised as synthetic only axioms and principles, Vaihinger replies as\nfollows:[267]\n\n     \u201cThe proposition--the angles of a triangle are together equal to\n     two right angles--Kant regards as synthetic. It is indeed deduced\n     from the axiom of parallels (with the aid of auxiliary lines), and\n     to that extent is understood in accordance with the principle of\n     contradiction.... The angles in the triangle constitute a special\n     case of the angles in the parallel lines which are intersected by\n     other lines. The principle of contradiction thus serves as vehicle\n     in the deduction, because once the identity of A and A\u00b4 is\n     recognised, the predicate _b_, which belongs to A, must also be\n     ascribed to A\u00b4. But the proposition is not for that reason itself\n     analytic in the Kantian sense. In the analytic proposition the\n     predicate is derived from the analysis of the subject concept. But\n     that does not happen in this case. The synthetic proposition can\n     never be derived _in and by itself_ from the principle of\n     contradiction; ... but only with the aid of that principle _from\n     other propositions_. Besides, in this deduction intuition must\n     always be resorted to; and that makes an essential difference.\n     Without it the identity of A and A\u00b4 cannot become known.\u201d\n\n=Pure mathematics.=[268]--\u201cPure,\u201d as thus currently used, is opposed only\nto applied, not to empirical. Kant here arbitrarily reads the latter\nopposition into it. Under this guise he begs the point in dispute.\n\n7 + 5 = 12.[269]--Though 7 + 5 = 12 expresses an identity or equality,\nit is an equality of the _objects_ or _magnitudes_, 7 + 5 and 12, not of\nthe concepts through which we think them.[270] Analysis of the concepts\ncan never reveal this equality. Only by constructing the concepts in\nintuition can it be recognised by the mind. This example has been\nalready cited in the first edition.[271] It is further elaborated in the\n_Prolegomena_, \u00a7 2 _c_, and is here transcribed. Kant\u2019s mode of stating\nhis position is somewhat uncertain. He alternates between \u201cthe\nrepresentation of 7 and 5,\u201d \u201cthe representation of the combination of 7\nand 5,\u201d[272] and \u201cthe concepts 7 and 5.\u201d[273] His view would seem to be\nthat there are _three_ concepts involved. For the concept of 7 we must\nsubstitute the intuition of 7 points, for the concept of 5 the intuition\nof 5 points, and for the concept of their sum the intuitive operation of\naddition.\n\n=Call in the assistance of intuition, for instance our five\nfingers.=[274]--This statement, repeated from the _Prolegomena_,[275]\ndoes not represent Kant\u2019s real position. The views which he has\nexpressed upon the nature of arithmetical science are of the most\ncontradictory character,[276] but to one point he definitely commits\nhimself, namely, that, like geometrical science, it rests, not (as here\nasserted) upon empirical, but upon pure intuition.[277] Except\nindirectly, by the reference to larger numbers, Kant here ignores his\nown important distinction between image and schema.[278] The above\nstatement would also make arithmetic dependent upon space.\n\n=Segner: Anfangsgr\u00fcnde der Arithmetik=,[279] translated from the Latin,\nsecond edition, Halle, 1773.\n\n=Natural science (physica) contains synthetic a priori\njudgments.=[280]--There is here a complication to which Vaihinger[281]\nhas been the first to draw attention. In the _Prolegomena_[282] Kant\nemphasises the distinction between physics and pure or universal science\nof nature.[283] The latter treats only the _a priori_ form of nature\n(_i.e._ its necessary conformity to law), and is therefore a\npropaedeutic to physics which involves further empirical factors. For\ntwo reasons, however, this universal natural science falls short of its\nideal. First, it contains empirical elements, such as the concepts of\nmotion, impenetrability, inertia, etc. Secondly, it refers only to the\nobjects of external sense, and not, as we should expect in a universal\nscience, to natural existences without exception, _i.e._ to the objects\nof psychology as well as of physics.[284] But among its principles there\nare, Kant adds, a few which are purely _a priori_ and possess the\nuniversality required: _e.g._ such propositions as that _substance is\npermanent_, and that _every event has a cause_. Now these are the\nexamples which ought to have been cited in the passage before us. Those\nactually given fall entirely outside the scope of the _Critique_. They\nare treated only in the _Metaphysische Anfangsgr\u00fcnde_. They belong to\nthe relatively, not to the absolutely, pure science of nature. The\nsource of the confusion Vaihinger again traces to Kant\u2019s failure to hold\nfast to the important distinction between immanent and transcendent\nmetaphysics.[285] His so-called pure or universal natural science\n(nature, as above noted, signifying for Kant \u201call that is\u201d) is really\n_immanent metaphysics_, and the propositions in regard to substance and\ncausality ought therefore to be classed as metaphysical. This, indeed,\nis how they are viewed in the earlier sections of the _Prolegomena_. The\ndistinction later drawn in \u00a7 15 is ignored. Pure natural science is\nidentified with mathematical physics, and the propositions which in \u00a7\n15 are spoken of as belonging to pure universal natural science are now\nregarded as metaphysical. \u201cGenuinely metaphysical judgments are one and\nall synthetic.... For instance, the proposition--everything which in\nthings is substance is permanent--is a synthetic, and properly\nmetaphysical judgment.\u201d[286] In \u00a7 5 the principle of causality is also\ncited as an example of a synthetic _a priori_ judgment in metaphysics.\nBut Kant still omits to draw a distinction between immanent and\ntranscendent metaphysics; and as a consequence his classification of\nsynthetic _a priori_ judgments remains thoroughly confused. They are\ntaken as belonging to three spheres, mathematics, physics (in the\nrelative sense), and metaphysics. The implication is that this threefold\ndistinction corresponds to the threefold division of the _Doctrine of\nElements_ into _Aesthetic_, _Analytic_, and _Dialectic_. Yet, as a\nmatter of fact, the propositions of mathematical physics, in so far as\nthey are examples of applied mathematics, are dealt with in the\n_Aesthetic_, and in so far as they involve concepts of motion and the\nlike fall entirely outside the scope of the _Critique_, while the\n_Analytic_ deals with those _metaphysical_ judgments (such as the\nprinciple of causality) which are of immanent employment.[287]\n\nAs the new paragraphs in the _Introduction_ to the second edition are\ntransferred without essential modification from the _Prolegomena_, they\nare open to the same criticism. To harmonise B 17 with the real teaching\nof the _Critique_, it must be entirely recast. Instead of \u201cnatural\nscience\u201d (_physica_) we must read \u201cpure universal natural science [=\nimmanent metaphysics],\u201d and for the examples given we must substitute\nthose principles of substance and causality which are dealt with in the\n_Analytic_. The next paragraph deals with metaphysics in its\ntranscendent form, and accordingly states the problem peculiar to the\n_Dialectic_.\n\n=Metaphysics.=[288]--This paragraph deals _explicitly_ only with\ntranscendent judgments, but as the terms used are ambiguous, it is\npossible that those of immanent metaphysics are also referred to. The\nparagraph is not taken from the _Prolegomena_. The corresponding\npassage[289] in the _Prolegomena_ deals only with the judgments of\nimmanent metaphysics.\n\n=The real problem of pure reason is contained in the question: How are\nsynthetic a priori judgments possible?=[290]--Cf. above, pp. 26 ff., 33\nff., 43 ff.\n\n=David Hume.=[291]--Cf. above, pp. 61 ff.\n\n=A theoretical knowledge.=[292]--_i.e._ Kant explicitly leaves aside the\nfurther problem, whether such judgments may not also be possible in the\npractical (moral) and other spheres.\n\n=How is pure natural science possible?=[293]--The note which Kant appends\nshows that he is here taking natural science in the relative sense.[294]\nThe same irrelevant instances are again cited.\n\n=As these sciences really exist.=[295]--Cf. below, p. 44 ff.\n\n=The poor progress which metaphysics has hitherto made.=[296]--Cf.\n_Preface_ to the second edition; _Prolegomena_, \u00a7 4, and A 175 ff.\n\n=How is metaphysics as a science possible?=[297]--We may now consider how\nthis and the three preceding questions are related to one another and to\nthe various divisions of the _Critique_.[298] The four subordinate\nquestions within the main problem--How are synthetic _a priori_\njudgments possible?--are here stated by Kant as:\n\n1. =How is pure mathematics possible?=\n\n2. =How is pure natural science possible?=\n\n3. =How is metaphysics as natural disposition possible?=\n\n4. =How is metaphysics as science possible?=\n\nThere is little difficulty as regards 1 and 2. The first is dealt with\nin the _Aesthetic_, and the second[299] in the _Analytic_, though, owing\nto the complexity of the problems, the _Aesthetic_ and _Analytic_ are\nwider than either query, and cannot be completely separated. Applied\nmathematics is dealt with in the _Analytic_ as well as in the\n_Aesthetic_, and in both the determination of the limits of scientific\nknowledge is equally important with that of accounting for its positive\nacquisitions. The third and fourth questions raise all manner of\ndifficulties. Notwithstanding the identical mode of formulation, they do\nnot run on all fours with the two preceding. The first two are taken as\nreferring to actually existing and valid sciences. It is the ground of\ntheir _objective validity_ that is sought. But what is investigated in\nthe third question falsely lays claim to the title of science; we can\nenquire only as to the ground of its _subjective_ possibility. In the\nfourth question, the problem takes still another form. Kant now seeks to\ndetermine _whether_ a new, not yet existing, science of metaphysics is\npossible, and _in what manner_ it can be validly constructed. The\nmanifoldness of the problems is thus concealed by the fixity of the\ncommon formula.[300] Now with what divisions of the _Critique_ are the\ntwo last questions connected? It has been suggested[301] that the third\nquestion is dealt with in the _Dialectic_ and the fourth in the\n_Methodology_, the four questions thus corresponding to the four main\ndivisions of the _Critique_. But this view is untenable, especially in\nits view of the fourth question. The division of the _Critique_ is by\ndichotomy into _doctrine of elements_ and _doctrine of methods_, the\nformer including the _Aesthetic_ and _Logic_, and the _Logic_ being\nagain divided into _Analytic_ and _Dialectic_. Its problems stand in an\nequally complex subordination; they cannot be isolated from one another,\nand set merely side by side. Secondly, it has been maintained[302] that\nthe third question is dealt with in the introduction to the _Dialectic_\n(in its doctrine of Ideas), and the fourth in the _Dialectic_ proper.\nThis view is fairly satisfactory as regards the third question, but\nwould involve the conclusion that the fourth question refers only to\ntranscendent metaphysics, and that it therefore receives a negative\nanswer. But that is not Kant\u2019s view of metaphysics _as a science_. The\n_Critique_ is intended to issue in a new and genuine body of\nmetaphysical teaching.\n\nThe key to the whole problem of the four questions is not to be found in\nthe _Critique_. This section is transcribed from \u00a7\u00a7 4-5 of the\n_Prolegomena_, and is consequently influenced by the general arrangement\nof the latter work. This fourfold division was indeed devised for the\npurposes of the argument of the _Prolegomena_, which is developed on the\nanalytic method, and for that reason it cannot be reconciled with the\nvery different structure of the _Critique_. Yet even the _Prolegomena_\nsuffers from confusion, due[303] to Kant\u2019s failure to distinguish\nbetween universal and relative natural science on the one hand, and\nbetween immanent and transcendent metaphysics on the other. The four\nquestions do not coincide with those of the _Critique_. Instead of the\nthird--how is metaphysics as natural disposition possible?--we find:\n_how is metaphysics in general possible?_ In \u00a7\u00a7 4, 5, Kant\u2019s argument is\nclear and straightforward. Pure mathematical science and mathematical\nphysics are actually existing sciences. The synthetic _a priori_\njudgments which they contain must be recognised as valid. Metaphysics\nmakes similar claims. But, as is sufficiently proved by the absence of\nagreement among philosophers, its professions are without ground. It\ntransgresses the limits of possible experience, and contains only\npretended knowledge. This false transcendent metaphysics is refuted in\nthe _Dialectic_. Kant was, however, equally convinced that an _immanent_\nmetaphysics is possible, and that its grounds and justification had been\nsuccessfully given in the _Analytic_. His problem as formulated in the\n_Prolegomena_ is accordingly threefold: (1) how are the existing\nrational sciences, mathematical and physical, possible? (2) _in the\nlight of the insight acquired by this investigation_, what is the origin\nand explanation of the existing pretended sciences of transcendent\nmetaphysics? and (3) in what manner can we establish a positive\nmetaphysics that will harmonise with reason\u2019s true vocation? So far all\nis clear and definite. But the unresolved difficulty, as to the relation\nin which natural science and immanent metaphysics stand to one another,\nbrings confusion in its train. As already noted,[304] in \u00a7 15 natural\nscience is displaced by immanent metaphysics (though not under that\nname); and as a result the fourth question reduces to the second, and\nthe above threefold problem has to be completely restated. The\n_Prolegomena_ has, however, already been divided into four parts; and in\nthe last division Kant still continues to treat the fourth question as\ndistinct from that which has been dealt with in the second division,\nthough, as his answer shows, they are essentially the same. The answer\ngiven is that metaphysics as a science is possible only in and through\nthe _Critique_, and that though the whole _Critique_ is required for\nthis purpose, the _content_ of the new science is embodied in the\n_Analytic_.\n\nIn the second edition of the _Critique_ the confusion between natural\nscience and immanent metaphysics still persists, and a new source of\nambiguity is added through the reformulation of the third question. It\nis now limited to the problem of the _subjective_ origin of metaphysics\nas a natural disposition. The fourth question has therefore to be\nwidened, so as to include transcendent as well as immanent, the old as\nwell as the new, metaphysics. But save for this one alteration the\nentire section is inspired by considerations foreign to the _Critique_;\nthis section, like B 17, must be recast before it will harmonise with\nthe subsequent argument.\n\n=Every kind of knowledge is called pure, etc.=[305]--These sentences are\nomitted in the second edition. They have been rendered unnecessary by\nthe further and more adequate definition of \u201cpure\u201d given in B 3 ff.\n\n=Reason is the faculty which supplies the principles of knowledge a\npriori.=[306]--This statement should, as Vaihinger points out, be\ninterpreted in the light of A 299 = B 355.\n\n     \u201cReason, like understanding, can be employed in a merely formal,\n     _i.e._ logical manner, wherein it abstracts from all content of\n     knowledge. But it is also capable of a real use,[307] since it\n     contains within itself the source of certain concepts and\n     principles, which it does not borrow either from the senses or from\n     the understanding.\u201d\n\nReason is taken in the first of the above meanings. Reason in its real\nuse, when extended so as to include pure sensibility and\nunderstanding,[308] is the pure reason referred to in the next sentence\nof the _Critique. A priori_ is here used to signify the relatively _a\npriori_; in the next sentence it denotes the absolutely _a priori_.\n\n=An Organon of pure reason.=[309]--What follows, from this point to the\nmiddle of the next section, is a good example of Kant\u2019s patchwork method\nof piecing together old manuscript in the composition of the _Critique_.\nThere seems to be no way of explaining its bewildering contradictions\nsave by accepting Vaihinger\u2019s[310] conclusion that it consists of three\nseparate accounts, written at different times, and representing\ndifferent phases in the development of Kant\u2019s views.\n\nI. The first account, beginning with the above words and ending with\n\u201calready a considerable gain\u201d (_schon sehr viel gewonnen ist_), is\nevidently the oldest. It reveals the influence of the _Dissertation_. It\ndistinguishes:\n\n  1. =Critique= of pure reason ( = _Propaedeutic_).\n  2. =Organon= of pure reason.\n  3. =System= of pure reason.\n\n1. =Critique= is a critical examination (_Beurtheilung_) of pure reason,\nits sources and limits. The implication (obscured by the direct relating\nof _Critique_ to _System_) is that it prepares the way for the\n_Organon_.\n\n2. =Organon= comprehends all the principles by which pure knowledge can be\nacquired and actually established.\n\n3. =System= is the complete application of such an _Organon_.\n\nThis classification is, as Paulsen[311] was the first to remark, an\nadaptation of the _Dissertation_ standpoint.\n\nII. The second account begins: \u201cI entitle all knowledge transcendental,\u201d\nbut is broken by the third account--from \u201cSuch a _Critique_\u201d to the end\nof the paragraph--which has been inserted into the middle of it. It is\nthen continued in the next section. It distinguishes:\n\n1. =Critique= of pure reason.\n\n2. =Transcendental philosophy.=\n\n1. =Critique= contains the principles of all _a priori synthetical_\nknowledge, tracing an architectonic plan which guarantees the\ncompleteness and certainty of all the parts.\n\n2. =Transcendental philosophy= contains their complete analytic\ndevelopment, and is therefore the system of such knowledge.\n\nIII. The third account (\u201cSuch a _Critique_\u201d to end of paragraph) in its\nmain divisions follows the first account: 1. _Critique_, 2. _Organon_ or\n_Canon_, 3. _System_. But they are now defined in a different manner.\n_Critique_ is a propaedeutic for the _Organon_. But _Organon_, which\nsignifies the totality of the principles through which pure knowledge is\nattained and extended,[312] may not be possible. In that case the\n_Critique_ is a preparation only for a _Canon_, _i.e._ the totality of\nthe principles of the _proper_ employment of reason.[313] The _Organon_\nor _Canon_, in turn, will render possible a _System_ of the philosophy\nof pure reason, the former yielding a system in extension of _a priori_\nknowledge, the latter a system which defines the limits of _a priori_\nknowledge.\n\nIt is impossible to reduce these divergencies to a single consistent\nview. They illustrate the varying sense in which Kant uses the term\n\u201cmetaphysics.\u201d In the first account, even though that account is based\non a distinction drawn in the _Dissertation_, the _system_ of\nmetaphysics is immanent; in the second it is also transcendent; in the\nthird it is neutral.[314]\n\n=Propaedeutic.=[315]--That the _Critique_ is only propaedeutic to a\n_System_ of pure reason was later denied by Kant in the following\nemphatic terms:\n\n     \u201cI must here observe that I cannot understand the attempt to\n     ascribe to me the view that I have sought to supply only a\n     Propaedeutic to transcendental philosophy, not the System of this\n     philosophy. Such a view could never have entered my thoughts, for I\n     have myself praised the systematic completeness (_das vollendete\n     Ganze_) of the pure philosophy in the _Critique of Pure Reason_ as\n     the best mark of its truth.\u201d[316]\n\nKant thus finally, after much vacillation in his use of the terms, came\nto the conclusion that _Critique_, _Transcendental Philosophy_, and\n_System_ all coincide. Meantime he has forgotten his own previous and\nconflicting utterances on this point.\n\n=As regards speculation negative only.=[317]--\u201cSpeculation\u201d here signifies\nthe theoretical, as opposed to the practical.[318] The qualifying phrase\nis in line with other passages of the second edition, in which it is\nemphasised that the conclusions of the _Critique_ are positive in their\npractical (moral) bearing.[319]\n\n=Transcendental=--=transcendent.=[320]--Kant was the first to distinguish\nbetween these two terms. In the scholastic period, in which they first\nappear, they were exactly synonymous, the term transcendent being the\nmore usual. The verb, to transcend, appears in Augustine in its widest\nmetaphysical sense. \u201cTranscende et te ipsum.\u201d \u201cCuncta corpora\ntranscenderunt [Platonici] quaerentes Deum; omnem animam mutabilesque\nomnes spiritus transcenderunt quaerentes summum Deum.\u201d[321] The first\nemployment of the term in a more specific or technical sense occurs in a\ntreatise, _De natura generis_, falsely ascribed to Thomas Aquinas. In\nthis treatise _ens_, _res_, _aliquid_, _unum_, _bonum_, _verum_ are\nentitled _transcendentia_. To understand the meaning in which the word\nis here used, we have, it would seem,[322] to take account of the\ninfluence exercised upon Aquinas by a mystical work of Arabian origin,\nentitled _De causis_. It contained reference to the Neo-Platonic\ndistinction between the Aristotelian categories, which the\nNeo-Platonists regarded as being derivative, and the more universal\nconcepts, _ens_, _unum_, _verum_, _bonum_. To these latter concepts\nAquinas gave a theological application. _Ens_ pertains to essence,\n_unum_ to the person of the Father, _verum_ to the person of the Son,\n_bonum_ to the person of the Holy Ghost. In the _De natura generis_ the\nnumber of these supreme concepts is increased to six by the addition of\n_res_ and _aliquid_, and as just stated the title _transcendentia_ is\nalso now applied for the first time. In this meaning the term\ntranscendent and its synonym transcendental are of frequent occurrence\nin Scholastic writings. The _transcendentia_ or _transcendentalia_ are\nthose concepts which so transcend the categories as to be themselves\npredicable of the categories. They are the \u201c_termini vel proprietates\nrebus omnibus cuiusque generis convenientes_.\u201d Thus Duns Scotus speaks\nof _ens_ as the highest of the \u201c_transcendental_\u201d concepts. The term\nalso occurs in a more or less similar sense in the writings of\nCampanella, Giordano Bruno, Francis Bacon, and Spinoza. The last named\ngives a psychological explanation of the \u201ctermini _Transcendentales_ ...\nut Ens, Res, Aliquid\u201d as standing for ideas that are in the highest\ndegree confused owing to the multiplicity of the images which have\nneutralised one another in the process of their generation.[323]\nBerkeley also speaks of the \u201ctranscendental maxims\u201d which lie outside\nthe field of mathematical enquiry, but which influence all the\nparticular sciences.[324] Evidently the term has become generalised\nbeyond its stricter scholastic meaning. Lambert employs transcendent in\nan even looser sense to signify concepts which represent what is common\nto both the corporeal and the intellectual world.[325] We may, indeed,\nassert that in Kant\u2019s time the terms transcendent and transcendental,\nwhile still remaining synonymous, and though used on the lines of their\noriginal Scholastic connotation, had lost all definiteness of meaning\nand all usefulness of application. Kant took advantage of this situation\nto distinguish sharply between them, and to impose upon each a meaning\nsuitable to his new Critical teaching.\n\n\u201cTranscendental\u201d is primarily employed by Kant as a name for a certain\nkind of knowledge. Transcendental knowledge is knowledge not of objects,\nbut of the nature and conditions of our _a priori_ cognition of them. In\nother words, _a priori_ knowledge must not be asserted, simply because\nit is _a priori_, to be transcendental; this title applies only to such\nknowledge as constitutes a _theory_ or _science_ of the _a priori_.[326]\nTranscendental knowledge and transcendental philosophy must therefore be\ntaken as coinciding; and as thus coincident, they signify the science of\nthe possibility, nature, and limits of _a priori_ knowledge. The term\nsimilarly applies to the subdivisions of the _Critique_. The _Aesthetic_\nis transcendental in that it establishes the _a priori_ character of the\nforms of sensibility; the _Analytic_ in that it determines the _a\npriori_ principles of understanding, and the part which they play in the\nconstitution of knowledge; the _Dialectic_ in that it defines and limits\nthe _a priori_ Ideas of Reason, to the perverting power of which all\nfalse metaphysics is due. That this is the primary and fundamental\nmeaning common to the various uses of the term is constantly overlooked\nby Max M\u00fcller. Thus in A 15 = B 30 he translates _transcendentale\nSinnenlehre_ \u201cdoctrine of transcendental sense\u201d instead of as\n\u201ctranscendental doctrine of sense.\u201d In transforming _transcendentale\nElementarlehre_ into \u201celements of transcendentalism\u201d he avoids the above\nerror, but only by inventing a word which has no place in Kant\u2019s own\nterminology.\n\nBut later in the _Critique_ Kant employs the term transcendental in a\nsecond sense, namely, to denote the _a priori_ factors in knowledge. All\nrepresentations which are _a priori_ and yet are applicable to objects\nare transcendental. The term is then defined through its distinction\nfrom the empirical on the one hand, and from the transcendent on the\nother. An intuition or conception is transcendental when it originates\nin pure reason, and yet at the same time goes to constitute an _a\npriori_ knowledge of objects. The contrast between the transcendental\nand the transcendent, as similarly determined upon by Kant, is equally\nfundamental, but is of quite different character. That is transcendent\nwhich lies entirely beyond experience; whereas the transcendental\nsignifies those _a priori_ elements which underlie experience as its\nnecessary conditions. The transcendent is always unknowable. The\ntranscendental is that which by conditioning experience renders all\nknowledge, whether _a priori_ or empirical, possible. The direct\nopposite of the transcendent is the immanent, which as such includes\nboth the transcendental and the empirical. Thus while Kant employs the\nterm transcendental in a very special sense which he has himself\narbitrarily determined, he returns to the original etymological meaning\nof the term transcendent. It gains a specifically Critical meaning only\nthrough being used to expound the doctrine that all knowledge is limited\nto sense-experience. The attempt to find some similar etymological\njustification for Kant\u2019s use of the term transcendental has led\nSchopenhauer and Kuno Fischer to assert that Kant entitles his\nphilosophy transcendental because it transcends both the dogmatism and\nthe scepticism of all previous systems![327] Another attempt has been\nmade by Stirling[328] and Watson,[329] who assert, at least by\nimplication, that the transcendental is a species of the transcendent,\nin that while the latter transcends the scope of experience, the former\ntranscends its sense-content. Kant himself, however, nowhere attempts to\njustify his use of the term by any such argument.\n\nA third meaning of the term transcendental arises through its extension\nfrom the _a priori_ intuitions and concepts to the processes and\nfaculties to which they are supposed to be due. Thus Kant speaks of the\ntranscendental syntheses of apprehension, reproduction, and recognition,\nand of the transcendental faculties of imagination and understanding. In\nthis sense the transcendental becomes a title for the conditions which\nrender experience possible. And inasmuch as processes and faculties can\nhardly be entitled _a priori_, Kant has in this third application of the\nterm departed still further from his first definition of it.[330]\n\nThe distinction between the transcendental and the transcendent may be\nillustrated by reference to the Ideas of reason. Regarded as regulative\nonly, _i.e._ merely as ideals which inspire the understanding in the\npursuit of knowledge, they are transcendental. Interpreted as\nconstitutive, _i.e._ as representing absolute realities, they are\ntranscendent. Yet, despite the fundamental character of this\ndistinction, so careless is Kant in the use of his technical terms that\nhe also employs transcendental as exactly equivalent in meaning to\ntranscendent. This is of constant occurrence, but only two instances\nneed here be cited. In the important phrase \u201ctranscendental ideality of\nspace and time\u201d the term transcendental is used in place of the term\ntranscendent. For what Kant is asserting is that judged from a\n_transcendent_ point of view, _i.e._ from the point of view of the thing\nin itself, space is only subjectively real.[331] The phrase is indeed\neasily capable of the orthodox interpretation, but, as the context\nclearly shows, that is not the way in which it is actually being used by\nKant. Another equally surprising example is to be found in the title\n\u201ctranscendental dialectic.\u201d Though it is defined in A 63-4 = B 88 in\ncorrect fashion, in A 297 = B 354 and A 308-9 = B 365-6 it is\ninterpreted as treating of the illusion involved in transcendent\njudgments, and so virtually as meaning _transcendent_ dialectic.[332]\n\n=Not a Critique of books and systems.=[333]--Kant here inserts a statement\nfrom the omitted _Preface_ to the first edition.[334] He now adds that\nthe _Critique_ will supply a criterion for the valuation of all other\nsystems.\n\n=A 13 = B 27.=--Kant\u2019s reason for omitting the title of Section II in the\nsecond edition was no doubt its inconsistency with the assertion of its\nopening sentence, viz. that the _Critique_ is _not_ transcendental\nphilosophy, but only a preparation for it. Instead of it, Kant has\nintroduced the more appropriate heading placed over the preceding\nparagraph.\n\n=The highest principles of morals do not belong to transcendental\nphilosophy.=[335]--Cf. A 801 = B 829. The alteration made in this passage\nin the second edition[336] indicates a transition towards the opposite\nview which Kant developed in the _Critique of Practical Reason_.[337]\n\n=The division of this science.=[338]--Kant in this paragraph alternates in\nthe most bewildering fashion between the _Critique_ and _Transcendental\nPhilosophy_. In this first sentence the _Critique_ seems to be referred\nto. Later it is _Transcendental Philosophy_ that is spoken of.\n\n=Doctrine of Elements and Doctrine of Methods.=[339]--Cf. A 707 ff. = B\n735 ff., and below, pp. 438, 563.\n\n=Two stems, sensibility and understanding, which may perhaps spring from\na common root.=[340]--Kant sometimes seems to suggest[341] that\nimagination is this common root. It belongs both to sensibility and to\nunderstanding, and is passive as well as spontaneous. But when so\nviewed, imagination is virtually regarded as an unknown supersensuous\npower, \u201cconcealed in the depths of the soul.\u201d[342] The supersensuous is\nthe point of union of our disparate human faculties, as well as of\nnature and freedom, mechanism and teleology.\n\n=The transcendental doctrine of sense would necessarily constitute the\nfirst part of the Science of Elements.=[343]--\u201cNecessarily constitute the\nfirst part\u201d translates _zum ersten Theile geh\u00f6ren m\u00fcssen_. This\nVaihinger explains as an archaic mode of expression, equivalent to\n_ausmachen_. The point is important because, if translated quite\nliterally, it might seem to conflict with the division actually\nfollowed, and to support the alternative division given in the _Critique\nof Practical Reason_. The first _Critique_ is divided thus:\n\n  I. Doctrine of Elements.\n    1. Aesthetic.\n    2. Logic.\n      (_a_) Analytic.\n      (_b_) Dialectic.\n\n  II. Doctrine of Methods.\n\nIn the _Critique of Practical Reason_[344] a much more satisfactory\ndivision is suggested:\n\n  I. Doctrine of Elements.\n    1. Analytic.\n      (_a_) Aesthetic (Sense).\n      (_b_) Logic (Understanding).\n    2. Dialectic.\n\n  II. Doctrine of Methods.\n\nThe first division rests on somewhat irrelevant distinctions derived\nfrom the traditional logic; the other is more directly inspired by the\ndistinctions which naturally belong to Kant\u2019s own philosophical system.\n\n\n\n\nTHE TRANSCENDENTAL DOCTRINE OF ELEMENTS\n\n\n\n\nPART I\n\nTHE TRANSCENDENTAL AESTHETIC\n\n\nThe _Aesthetic_ opens with a series of definitions. =Intuition=\n(_Anschauung_) is knowledge (_Erkenntnis_) which is in immediate\nrelation to objects (_sich auf Gegenst\u00e4nde unmittelbar bezieht_). Each\nterm in this definition calls for comment. _Anschauung_ etymologically\napplies only to visual sensation. Kant extends it to cover sensations of\nall the senses. The current term was _Empfindung_. Kant\u2019s reason for\nintroducing the term intuition in place of sensation was evidently the\nfact that the latter could not be made to cover space and time. We can\nspeak of pure intuitions, but not of pure sensations. _Knowledge_ is\nused in a very wide sense, not strictly consistent with A 50-1 = B\n74-5.[345] The phrase _sich bezieht_ is quite indefinite and ambiguous.\nIts meaning will depend upon the interpretation of its context. _Object_\nis used in its widest and most indefinite meaning. It may be taken as\nsignifying content (_Inhalt_, a term which does not occur in this\npassage, but which Kant elsewhere employs[346]). That, at least, is the\nmeaning which best fits the context. For when Kant adds that intuition\nrelates itself to objects _immediately_, it becomes clear that he has in\nmind its distinction from conception (_Begriff_) which as expressing the\nuniversal is related to objects only indirectly, representing some one\nor more attributes of the _given_ objects. Ultimately the whole content\nof conception must be given.[347] The phrase \u201crelates itself to objects\u201d\nmay, therefore, be paraphrased \u201chas some content, such as red or cold,\nas its immediate object.\u201d Through the content of intuition the whole\nmaterial of thought is supplied. Intuition in itself is blind, but not\nempty. \u201cThoughts without content are empty; intuitions without concepts\nare blind.\u201d[348]\n\nBut the phrase \u201cis in relation to objects\u201d has also for Kant a second\nmeaning, implied in the above, but supplementary to it. As he states in\nthe very next sentence, intuition can have an object, meaning thereby a\ncontent, only in so far as that content is _given_. The material of\nthought must be supplied; it cannot be invented.[349] The only mode,\nhowever, in which it can be supplied, at least to the human mind, is\nthrough the affecting of the mind by \u201cthe object.\u201d This is an excellent\ninstance of Kant\u2019s careless mode of expressing himself. In the first\npart of the sentence object means _object of intuition_. In the latter\npart it signifies the _cause of intuition_. And on Kant\u2019s view the two\ncannot coincide. The object which affects the mind is independently\nreal; the immediate object of the intuition is a sense-content, which\nKant, following the universally accepted view of his time, regards as\npurely subjective. The term object is thus used in two quite distinct\nmeanings within one and the same sentence.\n\nKant\u2019s definition of intuition, when stated quite explicitly, and\ncleared of all ambiguity, is therefore as follows. _Intuition is the\nimmediate apprehension of a content which as given is due to the action\nof an independently real object upon the mind._ This definition is\nobviously not meant to be a description of intuition as it presents\nitself to introspection, but to be a reflective statement of its\nindispensable conditions. Also it has in view only empirical intuitions.\nIt does not cover the pure intuitions space and time.[350] Though space\nand time are given, and though each possesses an intrinsic content,\nthese contents are not due to the action of objects upon the\nsensibility.\n\n     \u201cAn intuition is such a representation as immediately depends upon\n     the presence of the object. Hence it seems impossible _originally_\n     to intuit _a priori_ because intuition would in that event take\n     place without either a former or a present object to refer to, and\n     by consequence could not be intuition.\u201d[351]\n\nThis interpretation is borne out by Kant\u2019s answer to Beck when the\nlatter objected that only through subsumption under the categories can a\nrepresentation become objective. Kant replies in a marginal note, the\nmeaning of which, though difficult to decipher, admits of a fairly\ndefinite interpretation.\n\n     \u201cThe determining of a concept through intuition so as to yield\n     knowledge of the object falls within the province of the faculty of\n     judgment, but not the relation of the intuition to an object in\n     general [_i.e._ the view of it as having a content which is given\n     and which is therefore due to some object], for that is merely the\n     logical use of the representation, whereby it is thought as falling\n     within the province of knowledge. On the other hand, if this single\n     representation is related only to the subject, the use is aesthetic\n     (feeling), and the representation cannot be an act of\n     knowledge.\u201d[352]\n\n=Mind= (_Gem\u00fct_) is a neutral term without metaphysical implications.[353]\nIt is practically equivalent to the term which is substituted for it in\nthe next paragraph, power of representation (_Vorstellungsf\u00e4higkeit_).\n=Representation= (_Vorstellung_) Kant employs in the widest possible\nmeaning. It covers any and every cognitive state. The definition here\ngiven of =sensibility=--\u201cthe capacity (receptivity) to obtain\nrepresentations through the mode in which we are affected by\nobjects\u201d--is taken directly over from the _Dissertation_.[354] In this\ndefinition, as in that of intuition, Kant, without argument or question,\npostulates the existence of independently existing objects. The\nexistence of given sensations presupposes the existence of things in\nthemselves. Sensibility is spoken of as the source both of objects and\nof intuitions. This is legitimate since object and intuition mutually\nimply one another; the latter is the apprehension of the former. By\n\u201cobjects\u201d is obviously meant what in the third paragraph is called the\nmatter of appearances, _i.e._ sensations in their objective aspect, as\nqualities or contents. The term \u201cobject\u201d is similarly employed in the\nlast line of this first paragraph.\n\n=Understanding= (_Verstand_) is defined only in its logical or discursive\nemployment. Kant wisely defers all reference to its more fundamental\nsynthetic activities. _In us_ (_bei uns_) is an indirect reference to\nthe possibility of intellectual (non-sensuous) intuition which is\nfurther developed in other parts of the _Aesthetic_.[355] Sensuous\nintuition is due to affection by an object. In intellectual intuition\nthe mind must produce the object in the act of apprehending it.[356]\n\nKant\u2019s definition of intuition applies, as already noted, only to\nempirical intuition. He proceeds[357] to define the relation in which\n=sensation= (_Empfindung_) stands to empirical intuition. What he here\nsays amounts to the assertion that through sensation intuition acquires\nits object, _i.e._ that _sensation is the content of intuition_. And\nthat being so, it is also through sensation that empirical intuition\nacquires its relation to the object (= thing in itself) which causes it.\n(That would seem to be the meaning of the ambiguous second sentence; but\nit still remains uncertain whether the opposition intended is to pure or\nto intellectual intuition.) If this interpretation of the paragraph be\ncorrect, sensation is counted as belonging exclusively to the content\nside of subjective apprehension. But Kant views sensation in an even\nmore definite manner than he here indicates. Though sensation is given,\nit likewise involves a reaction of the mind.\n\n     \u201cWhatever is sensuous in knowledge depends upon the subject\u2019s\n     peculiar nature, in so far as it is capable of this or that\n     modification upon the presence of the object.\u201d[358]\n\nThus for Kant sensation is a modification or state of the subject,\nproduced by affection through an object. The affection produces a\nmodification or state of the subject, and this subjective modification\nis the sensation.\n\n     \u201cSensation is a perception [_Perception_] which relates itself\n     solely to the subject as the modification of its state.\u201d[359]\n\nThis view of sensation, as subjective, was universally held in Kant\u2019s\nday. He accepts it without argument or question. That it could possibly\nbe challenged never seems to have occurred to him. He is equally\nconvinced that it establishes the existence of an actually present\nobject.\n\n     \u201cSensation argues the presence of something, but depends as to its\n     quality upon the nature of the subject.\u201d[360] \u201cSensation\n     presupposes the actual presence of the object.\u201d[361]\n\nKant\u2019s view of sensation, as developed in the _Aesthetic_,[362] thus\ninvolves three points: (1) It must be counted as belonging to the\ncontent side of mental apprehension. (2) Though a quality or content, it\nis purely subjective, depending upon the nature of our sensibility. (3)\nIt is due to the action of some object upon the sensibility.\n\nKant distinguishes between =sensation= (_Empfindung_) and =feeling=\n(_Gef\u00fchl_).[363] It had been usual to employ them as synonyms.\n\n     \u201cWe understand by the word sensation an objective representation of\n     the senses; and in order to preclude the danger of being\n     misunderstood, we shall denote that which must always remain merely\n     subjective and can constitute absolutely no representation of an\n     object by the ordinary (_sonst \u00fcblichen_) term feeling.\u201d[364]\n\n=Appearance= (_Erscheinung_) is here defined as the undetermined object of\nan intuition. By undetermined object is meant, as we have seen, the\nobject in so far as it consists of the given sense contents. When these\ncontents are interpreted through the categories they become _phenomena_.\n\n     \u201cAppearances so far as they are thought as objects according to the\n     unity of the categories are called phenomena.\u201d[365]\n\nBut this distinction between appearance and phenomenon is not held to by\nKant. He more usually speaks of the categorised objects as appearances.\nThe term phenomenon is of comparatively rare occurrence in the\n_Critique_. This has been concealed from English readers, as both\nMeiklejohn and Max M\u00fcller almost invariably translate _Erscheinung_\nphenomenon. The statement that appearance is the _object_ of an\nempirical intuition raises a very fundamental and difficult question,\nnamely, as to the relation in which representation stands to the\nrepresented.[366] Frequently Kant\u2019s argument implies this distinction,\nyet constantly he speaks and argues as if it were non-existent. We have\nto recognise two tendencies in Kant, subjectivist and\nphenomenalist.[367] When the former tendency is in the ascendent, he\nregards all appearances, all phenomena, all empirical objects, as\nrepresentations, modifications of the sensibility, merely subjective.\nWhen, on the other hand, his thinking is dominated by the latter\ntendency, appearances gain an existence independent of the individual\nmind. They are known through subjective representations, but must not be\ndirectly equated with them. They have a genuine objectivity. To this\ndistinction, and its consequences, we shall have frequent occasion to\nreturn.\n\nThe phenomenalist standpoint is dominant in these first two paragraphs\nof the _Aesthetic_, and it finds still more pronounced expression in\nthe opening of the third paragraph. \u201cThat in the appearances which\n_corresponds_ (_correspondirt_) to sensation, I call its matter.\u201d This\nsentence, through the use of the term corresponds, clearly implies a\ndistinction between sensation and the real object apprehended in and\nthrough it. That, in turn, involves a threefold distinction, between\nsensation as subjective content (= appearance in the strict sense), the\nreal enduring object in space (= phenomenon, the categorised object,\nappearance in its wider and more usual sense), and the thing in\nitself.[368] Yet in the immediately following sentence Kant says that\n\u201cthe matter of all appearance is given _a posteriori_.\u201d By \u201cmatter of\nappearance\u201d Kant must there mean sensations, for they alone are given _a\nposteriori_.[369] On this view the phenomena or empirical objects reduce\nto, and consist of, sensations. The intermediate term of the above\nthreefold distinction is eliminated. The matter of appearance does not\ncorrespond to, but itself _is_, sensation. Thus in these successive\nsentences the two conflicting tendencies of Kant\u2019s teaching find verbal\nexpression. They intervene even in the preliminary definition of his\nterms. This fundamental conflict cannot, however, be profitably\ndiscussed at this stage.\n\nThe =manifold of appearance= (_das Mannichfaltige der Erscheinung_). The\nmeaning to be assigned to this phrase must depend upon the settlement of\nthe above question.[370] But in this passage it allows only of a\nsubjectivist interpretation, whereby sensations _are_ appearance. The\ngiven sensations as such constitute a manifold; as objects in space they\nare already ordered. Kant\u2019s more usual phrase is \u201cthe manifold of\nintuition.\u201d His adoption of the term \u201cmanifold\u201d (the _varia_ of the\n_Dissertation_) expresses his conviction that synthesis is indispensable\nfor all knowledge, and also his correlative view that nothing absolutely\nsimple can be apprehended in sense-experience. By the manifold Kant does\nnot mean, however, as some of his commentators would seem to imply, the\nchaotic or disordered. The emphasis is on manifoldness or plurality, as\ncalling for reduction to unity and system. The unity has to be _found_\nin it, not introduced into it forcibly from the outside. The manifold\nhas to be _interpreted_, even though the principles of interpretation\nmay originate independently of it. Though, for instance, the manifold\nas given is not in space and time, the specific space and time relations\nassigned by us are determined for us by the inherent nature of the\nmanifold itself.[371]\n\nThe form of appearance is defined--if the definition given in the first\nedition be translated literally--as \u201cthat which causes (_dasjenige,\nwelches macht dass_) the manifold of appearance to be intuited as\nordered in certain relations.\u201d This phrase is employed by Kant in other\nconnections, and, as Vaihinger points out,[372] need not necessarily\nindicate activity. \u201cSensation is that in our knowledge which causes it\nto be called _a posteriori_ knowledge.\u201d[373] In the second edition Kant\naltered the text from \u201c_geordnet angeschaut wird_\u201d to \u201c_geordnet werden\nkann_.\u201d The reason probably was that the first edition\u2019s wording might\nseem to imply that the form is (as the _Dissertation_ taught) capable in\nand by itself of ordering the manifold. Throughout the second edition\nKant makes more prominent the part which understanding plays in the\napprehension of space.[374]\n\nThis distinction between matter and form is central in Kant\u2019s\nsystem.[375] As he himself says:\n\n     \u201cThese are two conceptions which underlie all other reflection, so\n     inseparably are they bound up with all employment of the\n     understanding. The one [matter] signifies the determinable in\n     general, the other [form] its determination.\u201d[376]\n\nOn the side of matter falls the manifold, given, empirical, contingent\nmaterial of sense; on the side of form fall the unifying, _a priori_,\nsynthetic, relational instruments of sensibility and thought. For Kant\nthese latter are no mere abstractions, capable of being _distinguished_\nby the mind; they differ from the matter of experience in nature, in\nfunction, and in origin. Upon this dualistic mode of conceiving the two\nfactors depends the strength as well as the weakness of his position. To\nits perverting influence most of the unsatisfactory features of his\ndoctrine of space and time can be directly traced. But to it is also due\nhis appreciation of the new Critical problems, with their revolutionary\nconsequences, as developed in the _Analytic_.\n\nKant proceeds to argue: (_a_) that the distinction is between two\nelements of fundamentally different nature and origin. The matter is\ngiven _a posteriori_ in sensation; the form, as distinct from all\nsensation, must lie ready _a priori_ in the mind. (_b_) Kant also argues\nthat form, because of its separate origin, is capable of being\ncontemplated apart from all sensation. The above statements rest upon\nthe unexpressed assumption that sensations have no spatial attributes of\nany kind.[377] In themselves they have only intensive, not extensive,\nmagnitude.[378] Kant assumes this without question, and without the\nleast attempt at proof.[379] The assumption appears in Kant\u2019s writings\nas early as 1768 as a self-evident principle;[380] and throughout the\n_Critique_ is treated as a premiss for argument, never as a statement\ncalling for proof. The only kind of supporting argument which is even\nindirectly suggested by Kant is that space cannot by itself act upon the\nsenses.[381] This would seem to be his meaning when he declares[382]\nthat it is no object, but only an _ens imaginarium_. \u201cSpace is no object\nof the senses.\u201d[383] Such argument, however, presupposes that space can\nbe conceived apart from objects. It is no proof that an extended object\nmay not yield extended sensations. Kant completely ignores the\npossibility that formal relations may be given in and with the\nsensations. If our sensibility, in consequence of the action of objects\nupon it, is able to generate qualitative sensations, why, as Vaihinger\nvery pertinently enquires,[384] should it be denied the power of also\nproducing, in consequence of these same causes, impressions of\nquantitative formal nature? Sensations, on Kant\u2019s view, are the product\nof mind much more than of objects. Why, then, may not space itself be\nsensational?[385] From the point of view of empirical science there is\nno such radical difference between cause and effect in the latter case\nas exists in the former. As Herbert Spencer has remarked,[386] Kant\nmakes the enormous assumption\n\n     \u201c...that no differences among our sensations are determined by any\n     differences in the _non-ego_ (for to say that they are so\n     determined is to say that the form under which the _non-ego_ exists\n     produces an effect upon the _ego_); and as it similarly follows\n     that the order of coexistence and sequence among these sensations\n     is not determined by any order in the _non-ego_; we are compelled\n     to conclude that all these differences and changes in the _ego_ are\n     self-determined.\u201d\n\nKant\u2019s argument in the _Dissertation_ is exactly of this nature.\n\n     \u201cObjects do not strike the senses by their form. In order,\n     therefore, that the various impressions from the object acting on\n     the sense may coalesce into some whole of representation, there is\n     required an inner principle of the mind through which in accordance\n     with stable and innate laws that manifold may take on some\n     form.\u201d[387]\n\nIn the paragraph before us Kant may, at first sight, seem to offer an\nargument. He is really only restating his premiss. \u201cThat wherein alone\nsensations can be arranged (_sich ordnen_[388]) and placed in a certain\nform cannot itself again be sensation.\u201d Now, of course, if the term\nsensation is to be limited to the sense qualities, _i.e._ to content or\nmatter, conceived as existing apart from all formal relations, the\nformal elements cannot possibly be sensational. The legitimacy of that\nlimitation is, however, the question at issue. It cannot be thus decided\nby an arbitrary verbal distinction.\n\n     \u201cWere the contention that the relations of sensations are not\n     themselves sensed correct, the inference to the pure apriority of\n     the form of our perception would be inevitable. For sensation is\n     the sole form of interaction between consciousness and reality....\n     But that contention is false. The relations of sensations, their\n     determined coexistence and sequence, impress consciousness, just as\n     do the sensations. We feel this impression in the compulsion which\n     the determinateness of the empirical manifolds lays upon the\n     perceiving consciousness. The mere affection of consciousness by\n     these relations does not, indeed, by itself suffice for their\n     apprehension; but neither does it suffice for the apprehension of\n     the sensation itself. Thus there is in these respects no difference\n     between the matter and the form of appearance.\u201d[389]\n\nIn this way, then, by means of his definition of sensation, Kant\nsurreptitiously introduces his fundamental assumption. That assumption\nreappears as the conclusion that since the form of appearance cannot be\nsensation, it does not arise through the action of the object, and\nconsequently must be _a priori_. Though the paragraph seems to offer an\nargument in support of the apriority of space and time, it is found on\nexamination merely to unfold a position adopted without the slightest\nattempt at proof.[390]\n\n=The form of appearance must lie ready in the mind.=[391]--Comment upon\nthis, in order to be adequate, had best take the form of a systematic\ndiscussion of Kant\u2019s views, here and elsewhere, of space as an _a\npriori_ form of intuition. As already stated, the definition which Kant\ngives of intuition--as knowledge which stands in immediate relation to\nobjects--applies only to empirical intuition. Though by the term object\nKant, in so far as he is definite, means content, that content is such\nas can arise only through the action of some independent object upon the\nsensibility. In other words, the content apprehended must be sensuous.\nNow such a view of intuition obviously does not apply to pure intuition.\nAs the concluding line of the paragraph before us states, pure intuition\n\u201ccan be contemplated in separation from all sensation;\u201d and as the next\nparagraph adds, it exists in the mind \u201cwithout any actual object of the\nsenses.\u201d Yet Kant does not mean to imply that it is without content of\nany kind. \u201cThis pure form of sensibility may also itself be called pure\nintuition.\u201d[392] \u201cIt can be known before all actual perception, and for\nthat reason is called pure intuition.\u201d[393] Though, therefore, pure\nintuition has an intrinsic content, and is the immediate apprehension of\nthat content, it stands in no relation to any actual independent object.\nThe content as well as the form is _a priori_. That, however, raises\nwider questions, and these we must now discuss.\n\nHere, as in most of his fundamental positions, Kant entertains divergent\nand mutually contradictory doctrines. Only in his later utterances does\nhe in any degree commit himself to one consistent view. The position to\nwhich he finally inclines must not, however, be allowed to dominate the\ninterpretation of his earlier statements. The _Aesthetic_ calls for its\nown separate exegesis, quite as if it formed by itself an independent\nwork. Its problems are discussed from a standpoint more or less peculiar\nto itself. The commentator has the twofold task of stating its\nargumentation both in its conflict with, and in its relation to, the\nother parts of the _Critique_.\n\nOne essential difference between Kant\u2019s earlier and later treatments of\nspace is that in his earlier utterances it is viewed almost exclusively\nas a psychological _a priori_. The logical aspect of the problem first\nreceives anything like adequate recognition in the _Analytic_. If we\nkeep this important fact in mind, two distinct and contradictory views\nof the psychological nature of space intuition can be traced throughout\nthe _Aesthetic_. On one view, it antedates experience as an actual,\ncompleted, conscious intuition. On the other view, it precedes\nexperience only as a potential disposition. We rule ourselves out from\nunderstanding Kant\u2019s most explicit utterances if we refuse to recognise\nthe existence of both views. Kant\u2019s commentators have too frequently\nshut their eyes to the first view, and have then blamed Kant for using\nmisleading expressions. It is always safer to take Kant quite literally.\nHe nearly always means exactly what he says at the time when he says it.\nFrequently he holds views which run completely counter to present-day\npsychology, and on several occasions he flatly contradicts what he has\nwith equal emphasis maintained in other contexts. The aspects of Kant\u2019s\nproblems are so complex and various, and he is so preoccupied in doing\ncomplete justice to each in turn, that the question of the mutual\nconsistency of his results is much less considered than is ideally\ndesirable.\n\nThe two views can be more explicitly formulated. The first view alone is\nstraightforward and unambiguous. Space lies ready (_liegt bereit_) in\nthe mind, _i.e._ it does not arise. Prior even to sense-experience it\nexists as a _conscious_ intuition. For this reason it can be\ncontemplated apart from all sensation. It still remains when all sense\ncontent is thought away, and yet is not a mere form. In independence of\nthe sensuous manifold it possesses a pure manifold of its own. The\nground thesis of the second view--that space, prior to sense-experience,\nexists only as a permanent endowment of the mind--is likewise\nunambiguous. But in its development Kant throws consistency to the\nwinds. The possible ways in which, on the second view, consciousness of\nspace may be gained, can be tabulated as follows:\n\n     (_a_) By reflection upon the activity of the mind in the\n     construction of experience, yielding the intuition of a pure\n     manifold; or (_b_) by reflection upon the space-endowed products of\n     experience.[394] The latter mode of reflection may reveal:\n\n        (\u03b1) A pure manifold distinct from the manifold of\n                     sense; or\n\n        (\u03b2) Space as a form of the sensuous manifold.\n\n\n\n\nThere are thus three different ways (_a_, \u03b1, \u03b2) in\nwhich the second view can be developed: (_a_) represents the view of the\n_Dissertation_ (1770), of the reply to Eberhard (1790), and of those\nparts of the first edition\u2019s deduction of the categories which are of\nvery early origin; (\u03b1) represents the final standpoint of the\n_Analytic_; (\u03b2), the prevailing view of the present day, is\nnowhere accepted by Kant.[395]\n\nKant\u2019s utterances in the _Aesthetic_ are all of them coloured by the\nfirst main view. We can best approach them by way of the contrasted\nteaching of the _Dissertation_ of 1770. The teaching there formulated\npractically coincides, as above stated, with (_a_) of the second main\nview. Space, he maintains, is neither innate nor acquired from\nsense-experience.\n\n     \u201cCertainly both conceptions [of time and of space] are undoubtedly\n     acquired, not indeed by abstraction from our sensations of objects\n     (for sensation gives the matter, not the form of human cognition),\n     but from the mind\u2019s own action in co-ordinating its sensations in\n     accordance with unchanging laws. Each represents, as it were, an\n     immutable type, and so can be known intuitively. Sensations excite\n     this act of mind but do not contribute to the intuition. There is\n     here nothing innate except this law of the mind according to which\n     it conjoins in a certain manner the sensations derived from the\n     presence of some object.\u201d[396]\n\nHow this view is to be reconciled with the contention, no less\nexplicitly maintained,[397] that space is not only a form of intuition\nbut itself a pure intuition, Kant does not make clear. Reflection upon\nan activity of the mind may yield the representation of space as a form;\nit is difficult to comprehend how it should also yield an _a priori_\ncontent.\n\nKant nowhere in the _Critique_ directly discusses the question whether\nthe representation of space is innate or acquired. Such suggestions as\noccur refer (with the solitary exceptions of A 196 = B 241 and B 166\nff.)[398] only to the categories,[399] or as in the _Prolegomena_[400]\nto the Ideas of reason. But in 1790 Kant in his reply to Eberhard[401]\nagain formulates the view of the _Dissertation_. The _Critique_ allows,\nhe there says, of no innate representations. All, without exception, are\nacquired. But of certain representations there is an original\nacquisition (_urspr\u00fcngliche Erwerbung_). Their ground (_Grund_) is\ninborn. In the case of space this ground is the mind\u2019s peculiar capacity\nfor acquiring sensations in accordance with its subjective\nconstitution.[402]\n\n     \u201cThis first formal ground is alone inborn, not the space\n     representation itself. For it always requires impressions to\n     determine the faculty of knowledge to the representation of an\n     object (which in every case is its own action). Thus arises the\n     formal intuition, which we name space, as an originally acquired\n     representation (the form of outer objects in general), the ground\n     of which (as mere receptivity) is likewise inborn, and the\n     acquisition of which long antedates the determinate _conception_ of\n     things which are in accordance with this form.\u201d[403]\n\nThat last remark is confusing. Kant cannot mean that the representation\nof space is acquired prior to sense-experience, but only that since the\nmind gains it by reflection upon its own activity, it is among the first\nthings to be apprehended--an extremely questionable assertion, could\nthe premisses be granted. If \u201cthe determinate conception of things\u201d\ncomes late, still later must come the determinate conception of anything\nso abstract as pure space. The above passage thus repeats without\nessential modification the teaching of the _Dissertation_, and is open\nto the same objections. This teaching coincides with that of Leibniz in\nhis _Nouveaux Essais_; and in formulating it in the _Dissertation_ Kant\nwas very probably influenced by Leibniz. Though it is an improvement\nupon the more extreme forms of the Cartesian doctrine of innate ideas,\nit does not go sufficiently far.\n\nNow while Kant thus in 1770 and in 1790 so emphatically teaches that the\nrepresentation of space is not innate, he none the less, in the\nintermediate period represented by the _Aesthetic_, would seem to\nmaintain the reactionary view. Space is no mere potential disposition.\nAs a conscious representation it lies ready in the mind. What, then,\nwere the causes which constrained Kant to go back upon his own better\nviews and to adopt so retrograde a position? The answer must be\nconjectural, but may perhaps be found in the other main point in which\nthe teaching of the _Aesthetic_ is distinguished from that of the\n_Dissertation_. Throughout the _Critique_ Kant insists that space is a\nform of _receptivity_. It is _given_ to the mind. It has nothing to do\nwith spontaneity or understanding, and therefore cannot be acquired by\nreflection upon any activity of the mind. But neither can it, as _a\npriori_, be acquired from without. Consequently it cannot be acquired at\nall. But if given, and yet not acquired, it must as a representation lie\nready in the mind from the very birth of consciousness. Constrained by\nsuch reasoning, Kant views it as given in all its completeness just as\ntruly as is a sensation of colour or sound. This conclusion may not be\nsatisfactory. Kant\u2019s candid recognition of it is, however, greatly\npreferable to the blurring of the issue by most of his commentators.\n\nKant came, no doubt, to the more consistent position of the _Aesthetic_\nchiefly through further reflection upon the arguments of the\n_Dissertation_,[404] and especially by recognition of the fact that\nthough reflection upon an activity of the mind may be regarded as\nyielding a form of intuition, it can hardly be capable of yielding a\npure manifold which can be substituted for, and take the place of, the\nmanifold of sense. There are for Kant only two ways of escape from this\nunhappy quandary: (_a_) Either he must return to the _Dissertation_\nposition, and admit that the mind is active in the construction of\nspace. This he does in the 1790 reply to Eberhard, but only by\nmisrepresenting his own teaching in the _Critique_. In order\nconsistently to maintain that space is acquired by reflection upon an\nactivity of the mind, he would have to recast the entire _Aesthetic_, as\nwell as much of the _Analytic_, and to do so in ways which cannot\ngenuinely harmonise with the main tendencies of his teaching.[405] (_b_)\nNo such obstacle lay in the way of an alternative modification of his\nposition. Kant might very easily have given up the contention that space\nis a pure intuition. If he had been willing to recognise that the sole\npossible manifold of intuition is sensuous, he could then have\nmaintained that though space is innate as a potential form of\nreceptivity, it is acquired only through reflection upon the\nspace-endowed products of sensibility. So obvious are the advantages of\nthis position, so completely does it harmonise with the facts of\nexperience and with the teaching of modern psychology, and so obscure\nare the various passages in which Kant touches on this central issue,\nthat many of his most competent commentators are prepared to regard it\nas being the actual teaching of the _Critique_. The evidence[406] seems\nto me, however, to refute this interpretation of Kant\u2019s position. The\ntraditional, Cartesian, semi-mystical worship of mathematical truth, as\naltogether independent of the contingencies of sense-experience, and as\na body of knowledge absolutely distinct in origin from the merely\nempirical sciences, influences Kant\u2019s thinking even at the very moment\nwhen he is maintaining, in opposition to the Cartesians, that its\nsubject matter is a merely subjective intuition. Kant, as it would seem,\nstill maintains that there is a pure manifold of intuition distinct from\nthe manifold of sense; and so by the inevitable logic of his thought is\nconstrained to view space as innate in conscious form. This is not, of\ncourse, a conclusion which he could permanently stand by, but its\nelimination would have involved a more radical revision of his whole\nview of pure intuition and of mathematical science than he was willing\nto undertake. Though in the _Analytic_ he has come to recognise[407]\nthat it is acquired by reflection upon _objects_, to the end he would\nseem to persist in the difficult contention that such reflection yields\na pure manifold distinct from the manifold of sense.[408] His belief\nthat mathematical science is based upon pure intuition prevented him\nfrom recognising that though space may be a pure form of intuition, it\ncan never by itself constitute a complete intuition. Its sole possible\n_content_ is the manifold of sense. But even apart from the fact that\nour apprehension of space is always empirically conditioned, Kant\u2019s view\nof mathematical propositions as grounded in intuition is, as already\nobserved, not itself tenable. For though intuitions may perhaps be the\nultimate subject matter of geometry, concepts are its sole possible\ninstruments. Intuitions yield scientific insight in exact proportion to\nour powers of restating their complex content in the terms of abstract\nthought. Until the evidence which they supply has been thus\nintellectually tested and defined, they cannot be accepted as justifying\neven the simplest proposition.[409]\n\nThe complicated ambiguities of Kant\u2019s treatment of space may be\nillustrated and further clarified by discussion of another difficulty.\nIs space a _totum analyticum_ or a _totum syntheticum_? Does the whole\nprecondition the parts, or does it arise through combination of the\nparts? Or to ask another but connected question, do we intuit\ninfinitude, or is it conceptually apprehended only as the presupposition\nof our limited intuitions? To these questions diametrically opposite\nanswers can be cited from the _Critique_. As we have above noted, Kant\nteaches in the _Aesthetic_ that space is given as a whole, and that the\nparts arise only by limitation of it. But in A 162 = B 203 we find him\nalso teaching that a magnitude is to be entitled extensive\n\n     \u201c...when the representation of the parts makes possible, and\n     therefore necessarily precedes, the representation of the whole. I\n     cannot represent to myself a line, however small, without drawing\n     it in thought, _i.e._ generating from a point all its parts one\n     after another, and thus for the first time recording this\n     intuition.\u201d[410]\n\nHe adds in the second edition[411] that extensive magnitude cannot be\napprehended save through a \u201csynthesis of the manifold,\u201d a \u201ccombination\nof the homogeneous.\u201d\n\nThe note which Kant appends to B 136 is a very strange combination of\nboth views. It first of all reaffirms the doctrine of the _Aesthetic_\nthat space and time are not concepts, but intuitions within which as in\na unity a multitude of representations are contained; and then proceeds\nto argue that space and time, as thus _composite_, must presuppose an\nantecedent synthesis. In A 505 = B 533 we find a similar attempt to\ncombine both assertions.\n\n     \u201cThe parts of a given appearance are first given through and in the\n     regress of _decomposing synthesis_ (_decomponirenden Synthesis_).\u201d\n\nThe clash of conflicting tenets which Kant is striving to reconcile\ncould hardly find more fitting expression than in this assertion of an\n_analytic synthesis_. The same conflict appears, though in a less\nviolent form, in A 438 = B 466.\n\n     \u201cSpace should properly be called not _compositum_ but _totum_,\n     since its parts are possible only in the whole, not the whole\n     through the parts. It might, indeed, be said to be a _compositum_\n     that is _ideale_, but not _reale_. That, however, is a mere\n     subtlety.\u201d[412]\n\nThe arguments by which Kant proves space to be an _a priori_ intuition\nrest upon the view that _space is given as infinite_, and that _its\nparts arise through limitation of this prior-existent whole_. But a\nprinciple absolutely fundamental to the entire _Critique_ is the counter\nprinciple, that all analysis rests upon and presupposes a previously\nexercised synthesis. _Synthesis or totality as such can never be given._\nOnly in so far as a whole is synthetically constructed can it be\napprehended by the mind. _Representation of the parts precedes and\nrenders possible representation of the whole._\n\nThe solution of the dilemma arising out of these diverse views demands\nthe drawing of two distinctions. First, between a synthesised totality\nand a principle of synthesis; the former may involve a prior synthesis;\nthe latter does not depend upon synthesis, but expresses the\npredetermined nature of some special form of synthesis. Secondly, it\ndemands a distinction between the _a priori_ manifolds of space and time\nand the empirical manifold which is apprehended in and through them.\nThis, as we have already noted, is a distinction difficult to take quite\nseriously, and is entirely unsupported by psychological evidence. But it\nwould seem to be insisted upon by Kant, and to have been a determining\nfactor in the formulation of several of his main doctrines.\n\nIn terms of the first distinction we are compelled to recognise that the\nview of space which underlies the _Aesthetic_ is out of harmony with the\nteaching of the _Analytic_. In the _Aesthetic_ Kant interprets space not\nmerely as a form of intuition but also as a formal intuition, which is\ngiven complete in its totality, and which is capable of being\napprehended independently of its empirical contents, and even prior to\nthem. That would seem to be the view of space which is presupposed in\nKant\u2019s explanation of pure mathematical science. The passages from the\n_Analytic_, quoted above, are, however, its express recantation. Space,\nas the intuition of a manifold, is a _totum syntheticum_, not a _totum\nanalyticum_. It is constructed, not given. The divergence of views\nbetween the _Aesthetic_ and the _Analytic_ springs out of the difficulty\nof meeting at once the logical demands of a world which Kant conceives\nobjectively, and the psychological demands which arise when this same\nworld is conceived as subjectively conditioned. In principle, the whole\nprecedes the parts; in the process of being brought into existence as an\nintuition, the parts precede the whole. The principle which determines\nour apprehension of any space, however small or however large, is that\nit exists in and through universal space. This is the principle which\nunderlies both the synthetic construction of space and also its\napprehension once it is constructed. In principle, therefore, _i.e._ in\nthe order of logical thought, the whole precedes the parts.[413] The\nprocess, however, which this principle governs and directs, cannot start\nwith space as a whole, but must advance to it through synthesis of\nsmaller parts.\n\nBut Kant does not himself recognise any conflict between this teaching\nand the doctrine of the _Aesthetic_. He seems to himself merely to be\nmaking more definite a position which he has consistently held all\nalong; and this was possible owing to his retention and more efficient\nformulation of the second of the two distinctions mentioned above, viz.\nthat between the manifold of sense and the manifold of intuition. This\ndistinction enables him to graft the new view upon the old, and so in\nthe very act of insisting upon the indispensableness of the conceptual\nsyntheses of understanding, none the less to maintain his view of\ngeometry as an intuitive science.[414]\n\n     \u201cSpace and time contain a manifold of pure _a priori_ intuition,\n     but at the same time are conditions of the receptivity of our\n     mind--conditions under which alone it can receive representations\n     of objects, and which therefore must also affect the concept of\n     them. But if this manifold is to be known, the spontaneity of our\n     thinking requires that it be gone through in a certain way, taken\n     up, and connected. This action I name synthesis.... Such a\n     synthesis is pure, if the manifold is not empirical, but is given\n     _a priori_, as is that of space and of time.\u201d[415]\n\nThus Kant recognises that space, as apprehended by us, is constructed,\nnot given, and so by implication that the infinitude of space is a\nprinciple of apprehension, not a given intuition. But he also holds to\nthe view that it contains a pure, and presumably infinite, manifold,\ngiven as such.[416] In what this pure manifold consists, and how the\ndescription of it as a manifold, demanding synthesis for its\napprehension, is to be reconciled with its continuity, Kant nowhere even\nattempts to explain. Nor does he show what the simple elements are from\nwhich the synthesis of apprehension and reproduction in pure intuition\nmight start. The unity and multiplicity of space are, indeed, as he\nhimself recognises,[417] inseparably involved in one another; and\nrecognition of this fact must render it extremely difficult to assign\nthem to separate faculties. For the same reason it is impossible to\ndistinguish temporally, as Kant so frequently does, the processes of\nsynthesis and of analysis, making the former in all cases precede the\nlatter in time. The very nature of space and time, and, as he came to\nrecognise, the very nature of all Ideas of reason, in so far as they\ninvolve the notion of the unconditioned, conflict with such a view.\n\nEven when Kant is dealing with space as a principle of synthesis, he\nspeaks with no very certain voice. In the _Analytic_ it is ascribed to\nthe co-operation of sensibility and understanding. In the _Dialectic_ it\nis, by implication, ascribed to Reason; and in the _Metaphysical First\nPrinciples_ it is explicitly so ascribed.\n\n     \u201cAbsolute space cannot be object of experience; for space without\n     matter is no object of perception, and yet it is a necessary\n     conception of Reason, and therefore nothing but a mere Idea.\u201d[418]\n     \u201cAbsolute space is not necessary as a conception of an actual\n     object, but as an Idea which can serve as rule....\u201d[419]\n\nKant\u2019s teaching in the _Critique of Judgment_ is a further development\nof this position.\n\n     \u201cThe mind listens to the voice of Reason which, for every given\n     magnitude--even for those that can never be entirely apprehended,\n     although (in sensible representation) they are judged as entirely\n     given--requires totality.... It does not even except the infinite\n     (space and past time) from this requirement; on the contrary, it\n     renders it unavoidable to think the infinite (in the judgment of\n     common reason) as _entirely given_ (in its totality). But the\n     infinite is absolutely (not merely comparatively) great. Compared\n     with it everything else (of the same kind of magnitudes) is small.\n     But what is most important is that the mere ability to think it as\n     _a whole_ indicates a faculty of mind which surpasses every\n     standard of sense.... The _bare capability of thinking_ the given\n     infinite without contradiction requires in the human mind a faculty\n     itself supersensible. For it is only by means of this faculty and\n     its Idea of a noumenon ... that the infinite of the world of sense,\n     in the pure intellectual estimation of magnitude, can be\n     _completely_ comprehended _under_ one concept.... Nature is,\n     therefore, sublime in those of its phenomena, whose intuition\n     brings with it the Idea of its infinity.... For just as imagination\n     and _understanding_, in judging of the beautiful, generate a\n     subjective purposiveness of the mental powers by means of their\n     harmony, so imagination and _Reason_ do so by means of their\n     conflict.\u201d[420]\n\nKant has here departed very far indeed from the position of the\n_Aesthetic_.[421]\n\n\n\n\nTHE TRANSCENDENTAL AESTHETIC\n\n\nSECTION I\n\nSPACE\n\nMETAPHYSICAL EXPOSITION OF THE CONCEPTION OF SPACE[422]\n\n     =Space: First Argument.=--\u201cSpace is not an empirical concept\n     (_Begriff_) which has been abstracted from outer experiences. For\n     in order that certain sensations be related to something outside me\n     (_i.e._ to something in another region of space from that in which\n     I find myself), and similarly in order that I may be able to\n     represent them as outside [and _alongside_][423] one another, and\n     accordingly as not only [qualitatively] different but as in\n     different places, the representation of space must be presupposed\n     (_muss schon zum Grunde liegen_). The representation of space\n     cannot, therefore, be empirically obtained at second-hand from the\n     relations of outer appearance. This outer experience is itself\n     possible at all only through that representation.\u201d[424]\n\nThe first sentence states the thesis of the argument: _space is not an\nempirical concept abstracted from outer experiences_. The use of the\nterm _Begriff_ in the title of the section, and also in this sentence,\nis an instance of the looseness with which Kant employs his terms. It is\nhere synonymous with the term representation (_Vorstellung_), which\ncovers intuitions as well as general or discursive concepts.\nConsequently, the contradiction is only verbal, not real, when Kant\nproceeds to prove that the concept of space is an intuition, not a\nconcept. But this double employment of the term is none the less\nmisleading. When Kant employs it in a strict sense, it signifies solely\nthe general class concept.[425] All true concepts are for Kant of that\nsingle type. He has not re-defined the term concept in any manner which\nwould render it applicable to the relational categories. For\nunfortunately, and very strangely, he never seems to have raised the\nquestion whether categories are not also concepts. The application to\nthe forms of understanding of the separate title categories seems to\nhave contented him. Much that is obscure and even contradictory in his\nteaching might have been prevented had he recognised that the term\nconcept is a generic title which includes, as its sub-species, both\ngeneral notions and relational categories.\n\nKant\u2019s limitation of the term concept to the merely generic,[426] and\nhis consequent equating of the categorical proposition with the\nassertion of the substance-attribute relation,[427] would seem in large\npart to be traceable to his desire to preserve for himself, in the\npioneer labours of his Critical enquiries, the guiding clues of the\ndistinctions drawn in the traditional logic. Kant insists on holding to\nthem, at least in outward appearance, at whatever sacrifice of strict\nconsistency. Critical doctrine is made to conform to the exigencies of\nan artificial framework, with which its own tenets are only in very\nimperfect harmony. Appreciation of the ramifying influence, and, as\nregards the detail of exposition, of the far-reaching consequences, of\nthis desire to conform to the time-honoured rubrics, is indeed an\nindispensable preliminary to any adequate estimate whether of the\nstrength or of the defects of the Critical doctrines. As a separate and\never-present influence in the determining of Kant\u2019s teaching, this\nfactor may conveniently and compendiously be entitled Kant\u2019s logical\n_architectonic_.[428] We shall have frequent occasion to observe its\neffects.[429]\n\nThe second sentence gives expression to the fact through which Kant\nproves his thesis. Certain sensations, those of the special senses as\ndistinguished from the organic sensations,[430] are related to something\nwhich stands in a different region of space from the embodied self, and\nconsequently are apprehended as differing from one another not only in\nquality but also in spatial position. As is proved later in the\n_Analytic_, thought plays an indispensable part in constituting this\nreference of sensations to objects. Kant here, however, makes no mention\nof this further complication. He postulates, as he may legitimately do\nat this stage, the fact that our sensations are thus objectively\ninterpreted, and limits his enquiry to the spatial factor. Now the\nargument, as Vaihinger justly points out,[431] hinges upon the\nassumption which Kant has already embodied[432] in his definition of the\n\u201cform\u201d of sense, viz. that sensations are non-spatial, purely\nqualitative. Though this is an assumption of which Kant nowhere attempts\nto give proof, it serves none the less as an unquestioned premiss from\nwhich he draws all-important conclusions. This first argument on space\nderives its force entirely from it.\n\nThe proof that the representation of space is non-empirical may\ntherefore be explicitly stated as follows. As sensations are non-spatial\nand differ only qualitatively, the representation of space must have\nbeen added to them. And not being supplied by the given sensations, it\nmust, as the only alternative, have been contributed by the mind. The\nrepresentation of space, so far from being derived from external\nexperience, is what first renders it possible. As a subjective form that\nlies ready in the mind, it precedes experience and co-operates in\ngenerating it. This proof of the apriority of space is thus proof of the\npriority of the _representation_ of space to every empirical perception.\n\nIn thus interpreting Kant\u2019s argument as proving more than the thesis of\nthe first sentence claims, we are certainly reading into the proof more\nthan Kant has himself given full expression to. But, as is clearly shown\nby the argument of the next section, we are only stating what Kant\nactually takes the argument as having proved, namely, that the\nrepresentation of space is not only non-empirical but is likewise of\nsubjective origin and precedes experience in temporal fashion.\n\nThe point of view which underlies and inspires the argument can be\ndefined even more precisely. Kant\u2019s conclusion may be interpreted in\neither of two ways. The form of space may precede experience only as a\npotentiality. Existing as a power of co-ordination,[433] it will come to\nconsciousness only indirectly through the addition which it makes to the\ngiven sensations. Though subjective in origin, it will be revealed to\nthe mind only in and through experience. This view may indeed be\nreconciled with the terms of the proof. But a strictly literal\ninterpretation of its actual wording is more in keeping with what, as we\nshall find, is the general trend of the _Aesthetic_ as a whole. We are\nthen confronted by a very different and extremely paradoxical view,\nwhich may well seem too naive to be accepted by the modern reader, but\nwhich we seem forced,[434] none the less, to regard as the view actually\npresented in the text before us. Kant here asserts, in the most explicit\nmanner, that the mind, in order to construe sensations in spatial terms,\nmust already be in possession of a _representation_ of space, and that\nit is in the light of this representation that it apprehends sensations.\nThe conscious representation of space precedes in time external\nexperience. Such, then, would seem to be Kant\u2019s first argument on space.\nIt seeks to establish a negative conclusion, viz. that space is not\nderived from experience. But, in so doing, it also yields a positive\npsychological explanation of its origin.\n\nThose commentators[435] who refuse to recognise that Kant\u2019s problem is\nin any degree psychological, or that Kant himself so regards it, and who\nconsequently seek to interpret the _Aesthetic_ from the point of view of\ncertain portions of the _Analytic_, give a very different statement of\nthis first argument. They state it in purely logical terms.[436] Its\nproblem, they claim, is not that of determining the origin of our\nrepresentation of space, but only its logical relation to our specific\nsense-experiences. The notion of space in general precedes, as an\nindispensable logical presupposition, all particular specification of\nthe space relation. Consciousness of space as a whole is not constructed\nfrom consciousness of partial spaces; on the contrary, the latter is\nonly possible in and through the former.\n\nSuch an argument does of course represent a valuable truth; and it alone\nharmonises with much in Kant\u2019s maturer teaching;[437] but we must not\ntherefore conclude that it is also the teaching of the _Aesthetic_. The\n_Critique_ contains too great a variety of tendencies, too rich a\ncomplexity of issues, to allow of such simplification. It loses more\nthan it gains by such rigorous pruning of the luxuriant secondary\ntendencies of its exposition and thought. And above all, this procedure\ninvolves the adoption by the commentator of impossible responsibilities,\nthose of deciding what is essential and valuable in Kant\u2019s thought and\nwhat is irrelevant. The value and suggestiveness of Kant\u2019s philosophy\nlargely consist in his sincere appreciation of conflicting tendencies,\nand in his persistent attempt to reduce them to unity with the least\npossible sacrifice. But in any case the logical interpretation\nmisrepresents this particular argument. Kant is not here distinguishing\nbetween space in general and its specific modifications. He is\nmaintaining that no space relation can be revealed in sensation. It is\nnot only that the apprehension of any limited space presupposes the\nrepresentation of space as a whole. Both partial and infinite space are\nof mental origin; sensation, as such, is non-spatial, purely subjective.\nAnd lastly, the fact that Kant means to assert that space is not only\nlogically presupposed but is subjectively generated, is sufficiently\nborne out by his frequent employment elsewhere in the _Aesthetic_ of\nsuch phrases as \u201cthe subjective condition of sensibility,\u201d \u201clying ready\nin our minds,\u201d and \u201cnecessarily preceding [as the form of the subject\u2019s\nreceptivity] all intuitions of objects.\u201d\n\n=Second Argument.=--Having proved by the first argument that the\nrepresentation of space is not of empirical origin, Kant in the second\nargument proceeds to establish the positive conclusion that it is _a\npriori_.[438] The proof, when all its assumptions are rendered explicit,\nruns as follows. _Thesis_: Space is a necessary representation, and\nconsequently is _a priori_. _Proof_: It is impossible to imagine the\nabsence of space, though it is possible to imagine it as existing\nwithout objects to fill it. A representation which it is impossible for\nthe mind to be without is a necessary representation. But necessity is\none of the two criteria of the _a priori_. The proof of the necessary\ncharacter of space is therefore also a proof of its being _a priori_.\n\nThe argument, more freely stated, is that what is empirically given from\nwithout can be thought away, and that since space cannot be thus\neliminated, it must be grounded in our subjective organisation, _i.e._\nmust be psychologically _a priori_. The argument, as stated by Kant,\nemphasises the apriority, not the subjectivity, of space, but none the\nless the asserted apriority is psychological, not logical in character.\nFor the criterion employed is not the impossibility of thinking\notherwise, but our incapacity to represent this specific element as\nabsent. The ground upon which the whole argument is made to rest is the\nmerely brute fact (asserted by Kant) of our incapacity to think except\nin terms of space.\n\nThe argument is, however, complicated by the drawing of a further\nconsequence, which follows as a corollary from the main conclusion.\nFrom the subjective necessity of space follows its objective necessity.\nSpace being necessary _a priori_, objects can only be apprehended in and\nthrough it. Consequently it is not dependent upon the objects\napprehended, but itself underlies outer appearances as the condition of\ntheir possibility. This corollary is closely akin to the first argument\non space, and differs from it only in orientation. The first argument\nhas a psychological purpose. It maintains that the representation of\nspace precedes external experience, causally conditioning it. The\ncorollary has a more objective aim. It concludes that space is a\nnecessary constituent of the external experience thus generated. The one\nproves that space is a necessary _subjective antecedent_; the other that\nit is a necessary _objective ingredient_.[439]\n\nTo consider the proof in detail. The exact words which Kant employs in\nstating the _nervus probandi_ of the argument are that we can never\n_represent_ (_eine Vorstellung davon machen_) space as non-existent,\nthough we can very well _think_ (_denken_) it as being empty of objects.\nThe terms _Vorstellung_ and _denken_ are vague and misleading. Kant\nhimself recognises that it is possible to conceive that there are beings\nwho intuit objects in some other manner than in space. He cannot\ntherefore mean that we are unable to _think_ or _conceive_ space as\nnon-existent. He must mean that we cannot in imagination intuit it as\nabsent. It is the necessary form of all our intuitions, and therefore\nalso of imagination, which is intuitive in character. Our consciousness\nis dependent upon given intuitions for its whole content, and to that\nextent space is a form with which the mind can never by any possibility\ndispense. Pure thought enables it to realise this _de facto_ limitation,\nbut not to break free from it. Even in admitting the possibility of\nother beings who are not thus constituted, the mind still recognises its\nown ineluctable limitations.\n\nKant offers no proof of his assertion that space can be intuited in\nimage as empty of all sensible content; and as a matter of fact the\nassertion is false. Doubtless the use of the vague term _Vorstellung_ is\nin great part responsible for Kant\u2019s mistaken position. So long as\nimagination and thought are not clearly distinguished, the assertion is\ncorrespondingly indefinite. Pure space may possibly be _conceived_, but\nit can also be conceived as altogether non-existent. If, on the other\nhand, our imaginative power is alone in question, the asserted fact\nmust be categorically denied. With the elimination of all sensible\ncontent space itself ceases to be a possible image. Kant\u2019s proof thus\nrests upon a misstatement of fact.\n\nIn a second respect Kant\u2019s proof is open to criticism. He takes the\nimpossibility of imagining space as absent as proof that it originates\nfrom within. The argument is valid only if no other psychological\nexplanation can be given of this necessity, as for instance through\nindissoluble association or through its being an invariable element in\nthe given sensations. Kant\u2019s ignoring of these possibilities is due to\nhis unquestioning belief that sensations are non-spatial, purely\nqualitative. That is a presupposition whose truth is necessary to the\ncogency of the argument.\n\n=Third Argument.=--This argument, which was omitted in the second edition,\nwill be considered in its connection with the transcendental exposition\ninto which it was then merged.\n\n=Fourth (in second edition, Third) Argument.=--The next two arguments seek\nto show that space is not a discursive or general concept but an\nintuition. The first proof falls into two parts, (_a_) We can represent\nonly a single space. For though we speak of many spaces, we mean only\nparts of one and the same single space. Space must therefore be an\nintuition. For only intuition is thus directly related to a single\nindividual. A concept always refers indirectly, _per notas communes_, to\na plurality of individuals. (_b_) The parts of space cannot precede the\none all-comprehensive space. They can be thought only in and through it.\nThey arise through limitation of it. Now the parts (_i.e._ the\nattributes) which compose a concept precede it in thought. Through\ncombination of them the concept is formed. Space cannot, therefore, be a\nconcept. Consequently it must, as the only remaining alternative, be an\nintuition. Only in an intuition does the whole precede the parts. In a\nconcept the parts always precede the whole. Intuition stands for\nmultiplicity in unity, conception for unity in multiplicity.\n\nThe first part of the argument refers to the extension, the second part\nto the intension of the space representation. In both aspects it appears\nas intuitional.[440]\n\nKant, in repeating his thesis as a conclusion from the above grounds,\nconfuses the reader by an addition which is not strictly relevant to the\nargument, viz. by the statement that this intuition must be\nnon-empirical and _a priori_. This is simply a recapitulation of what\nhas been established in the preceding proofs. It is not, as might at\nfirst sight appear, part of the conclusion established by the argument\nunder consideration. The reader is the more apt to be misled owing to\nthe fact that very obviously arguments for the non-empirical and for the\n_a priori_ character of space _can_ be derived from proof (_b_). That\nspace is non-empirical would follow from the fact that representation of\nspace as a whole is necessary for the apprehension of any part of it.\nEmpirical intuition can only yield the apprehension of a limited space.\nThe apprehension of the comprehensive space within which it falls must\ntherefore be non-empirical.\n\n     \u201cAs we intuitively apprehend (_anschauend erkennen_) not only the\n     space of the object which affects our senses, but the whole space,\n     space cannot arise out of the actual affection of the senses, but\n     must precede it in time (_vor ihr vorhergehen_).\u201d[441]\n\nBut in spite of its forcibleness this argument is nowhere presented in\nthe _Critique_.\n\nSimilarly, in so far as particular spaces can be conceived only in and\nthrough space as a whole, and in so far as the former are limitations of\nthe one antecedent space, the intuition which underlies all external\nperception must be _a priori_. This is in essentials a stronger and more\ncogent mode of formulating the second argument on space. But again, and\nvery strangely, it is nowhere employed by Kant in this form.\n\nThe concluding sentence, ambiguously introduced by the words _so werden\nauch_, is tacked on to the preceding argument. Interpreted in the light\nof \u00a7 15 C of the _Dissertation_,[442] and of the corresponding\nfourth[443] argument[444] on time, it may be taken as offering further\nproof that space is an intuition. The concepts of line and triangle,\nhowever attentively contemplated, will never reveal the proposition that\nin every triangle two sides taken together are greater than the third.\nAn _a priori_ intuition will alone account for such apodictic knowledge.\nThis concluding sentence thus really belongs to the transcendental\nexposition; and as such ought, like the third argument, to have been\nomitted in the second edition.\n\nKant\u2019s proof rests on the assumption that there are only two kinds of\nrepresentation, intuitions and concepts, and also in equal degree upon\nthe further assumption that all concepts are of one and the same\ntype.[445] Intuition is, for Kant, the apprehension of an individual.\nConception is always the representation of a class or genus. Intuition\nis immediately related to the individual. Conception is reflective or\ndiscursive; it apprehends a plurality of objects indirectly through the\nrepresentation of those marks which are common to them all.[446]\nIntuition and conception having been defined in this manner, the proof\nthat space is single or individual, and that in it the whole precedes\nthe parts, is proof conclusive that it is an intuition, not a\nconception. Owing, however, to the narrowness of the field assigned to\nconception, the realm occupied by intuition is proportionately wide, and\nthe conclusion is not as definite and as important as might at first\nsight appear. By itself, it amounts merely to the statement, which no\none need challenge, that space is not a generic class concept.\nIncidentally certain unique characteristics of space are, indeed,\nforcibly illustrated; but the implied conclusion that space on account\nof these characteristics must belong to receptivity, not to\nunderstanding, does not by any means follow. It has not, for instance,\nbeen proved that space and time are radically distinct from the\ncategories, _i.e._ from the relational forms of understanding.\n\nIn 1770, while Kant still held to the metaphysical validity of the pure\nforms of thought, the many difficulties which result from the ascription\nof independent reality to space and time were, doubtless, a sufficient\nreason for regarding the latter as subjective and sensuous. But upon\nadoption of the Critical standpoint such argument is no longer valid. If\nall our forms of thought may be subjective, the existence of antinomies\nhas no real bearing upon the question whether space and time do or do\nnot have a different constitution and a different mental origin from the\ncategories. The antinomies, that is to say, may perhaps suffice to prove\nthat space and time are subjective; they certainly do not establish\ntheir sensuous character.\n\nBut though persistence of the older, un-Critical opposition between the\nintellectual and the sensuous was partly responsible for Kant\u2019s\nreadiness to regard as radical the very obvious differences between a\ncategory such as that of substance and attribute and the visual or\ntactual extendedness with which objects are endowed, it can hardly be\nviewed as the really decisive influence. That would rather seem to be\ntraceable to Kant\u2019s conviction that mathematical knowledge is unique\nboth in fruitfulness and in certainty, and to his further belief that it\nowes this distinction to the _content_ character of the _a priori_\nforms upon which it rests. For though the categories of the physical\nsciences are likewise _a priori_, they are exclusively\n_relational_,[447] and serve only to organise a material that is\nempirically given. To account for the superiority of mathematical\nknowledge Kant accordingly felt constrained to regard space and time as\nnot merely _forms_ in terms of which we interpret the matter of sense,\nbut as also themselves intuited _objects_, and as therefore possessing a\ncharacter altogether different from anything which can be ascribed to\nthe pure understanding. The opposition between forms of sense and\ncategories of the understanding, in the strict Kantian mode of\nenvisaging that opposition, is thus inseparably bound up with Kant\u2019s\ndoctrine of space and time as being not only forms of intuition, but as\nalso in their purity and independence themselves intuitions. _Even the\nsensuous subject matter of pure mathematics_--so Kant would seem to\ncontend--_is_ a priori _in nature_. If this latter view be\nquestioned--and to the modern reader it is indeed a stone of\nstumbling--much of the teaching of the _Aesthetic_ will have to be\nmodified or at least restated.\n\n=Fifth (in second edition, Fourth) Argument.=--This argument is quite\ndifferently stated in the two editions of the _Critique_, though the\npurpose of the argument is again in both cases to prove that space is an\nintuition, not a general concept. In the first edition this is proved by\nreference to the fact that space is given as an infinite magnitude. This\ncharacteristic of our space representation cannot be accounted for so\nlong as it is regarded as a concept. A general conception of space which\nwould abstract out those properties and relations which are common to\nall spaces, to a foot as well as to an ell, could not possibly determine\nanything in regard to magnitude. For since spaces differ in magnitude,\nany one magnitude cannot be a common quality. Space is, however, given\nus as determined in magnitude, namely, as being of infinite magnitude;\nand if a general conception of space relations cannot determine\nmagnitude, still less can it determine infinite magnitude. Such infinity\nmust be derived from limitlessness in the progression of intuition. Our\nconceptual representations of infinite magnitude must be derivative\nproducts, acquired from this intuitive source.\n\nIn the argument of the second edition the thesis is again established by\nreference to the infinity of space. But in all other respects the\nargument differs from that of the first edition. A general conception,\nwhich abstracts out common qualities from a plurality of particulars,\ncontains an infinite number of possible different representations\n_under_ it; but it cannot be thought as containing an infinite number of\nrepresentations _in_ it. Space must, however, be thought in this latter\nmanner, for it contains an infinite number of coexisting parts.[448]\nSince, then, space cannot be a concept, it must be an intuition.\n\nThe definiteness of this conclusion is somewhat obscured by the further\ncharacterisation of the intuition of space as _a priori_, and by the\nstatement that it is the _original_ (_urspr\u00fcngliche_) representation\nwhich is of this intuitive nature. The first addition must here, again,\njust as in the fourth argument, be regarded as merely a recapitulation\nof what has already been established, not a conclusion from the present\nargument. The introduction of the word \u2018original\u2019 seems to be part of\nKant\u2019s reply to the objections which had already been made to his\nadmission in the first edition that there is a conception as well as an\nintuition of space. It is the _original given intuition_ of space which\nrenders such reflective conception possible.\n\nThe chief difficulty of these proofs arises out of the assertion which\nthey seem to involve that space is given as actually infinite. There are\napparently, on this point, two views in Kant, which were retained up to\nthe very last, and which are closely connected with his two\nrepresentations of space, on the one hand as a _formal intuition_ given\nin its purity and in its completeness, and on the other hand as the\n_form of intuition_, which exists only so far as it is constructed, and\nwhich is dependent for its content upon given matter.\n\n=Third Argument, and Transcendental Exposition of Space.=--The distinction\nbetween the metaphysical and the transcendental expositions, introduced\nin the second edition of the _Critique_,[449] is one which Kant seems to\nhave first made clear to himself in the process of writing the\n_Prolegomena_.[450] It is a genuine improvement, marking an important\ndistinction. It separates out two comparatively independent lines of\nargument. The terms in which the distinction is stated are not, however,\nfelicitous. Kant\u2019s reason for adopting the title metaphysical is\nindicated in the _Prolegomena_:[451]\n\n     \u201cAs concerns the sources of metaphysical cognition, its very\n     concept implies that they cannot be empirical.... For it must not\n     be physical but metaphysical knowledge, _i.e._ knowledge lying\n     beyond experience.... It is therefore _a priori_ knowledge, coming\n     from pure understanding and pure Reason.\u201d\n\nThe metaphysical exposition, it would therefore seem, is so entitled\nbecause it professes to prove that space is _a priori_, not empirical,\nand to do so by analysis of its concept.[452] Now by Kant\u2019s own\ndefinition of the term transcendental, as the theory of the _a priori_,\nthis exposition might equally well have been named the transcendental\nexposition. In any case it is an essential and chief part of the\n_Transcendental Aesthetic_. Such division of the _Transcendental\nAesthetic_ into a metaphysical and a transcendental part involves a\ntwofold use, wider and narrower, of one and the same term. Only as\ndescriptive of the whole _Aesthetic_ is transcendental employed in the\nsense defined.\n\nExposition (_Er\u00f6rterung_, Lat. _expositio_) is Kant\u2019s substitute for the\nmore ordinary term definition. Definition is the term which we should\nnaturally have expected; but as Kant holds that no =given= concept,\nwhether _a priori_ or empirical, can be defined in the strict\nsense,[453] the substitutes the term exposition, using it to signify\nsuch definition of the nature of space as is possible to us. To complete\nthe parallelism Kant speaks of the transcendental enquiry as also an\nexposition. It is, however, in no sense a definition. Kant\u2019s terms here,\nas so often elsewhere, are employed in a more or less arbitrary and\nextremely inexact manner.\n\nThe distinction between the two expositions is taken by Kant as follows.\nThe metaphysical exposition determines the nature of the concept of\nspace, and shows it to be a given _a priori_ intuition. The\ntranscendental exposition shows how space, when viewed in this manner,\nrenders comprehensible the possibility of synthetic _a priori_\nknowledge.\n\nThe omission of the third argument on space from the second edition, and\nits incorporation into the new transcendental exposition, is certainly\nan improvement. In its location in the first edition, it breaks in upon\nthe continuity of Kant\u2019s argument without in any way contributing to the\nfurther definition of the concept of space. Also, in emphasising that\nmathematical knowledge depends upon the _construction_ of concepts,[454]\nKant presupposes that space is intuitional; and that has not yet been\nestablished.\n\nThe argument follows the strict, rigorous, synthetic method. From the\nalready demonstrated _a priori_ character of space, Kant deduces the\napodictic certainty of all geometrical principles. But though the\nparagraph thus expounds a consequence that follows from the _a priori_\ncharacter of space, not an argument in support of it, something in the\nnature of an argument is none the less implied. The fact that this view\nof the representation of space alone renders mathematical science\npossible can be taken as confirming this interpretation of its nature.\nSuch an argument, though circular, is none the less cogent.\nConsideration of Kant\u2019s further statements, that were space known in a\nmerely empirical manner we could not be sure that in all cases only one\nstraight line is possible between two points, or that space will always\nbe found to have three dimensions, must meantime be deferred.[455]\n\nIn the new transcendental exposition Kant adopts the analytic method of\nthe _Prolegomena_, and accordingly presents his argument in independence\nof the results already established. He starts from the assumption of the\nadmitted validity of geometry, as being a body of synthetic _a priori_\nknowledge. Yet this, as we have already noted, does not invalidate the\nargument; in both the first and the last paragraphs it is implied that\nthe _a priori_ and intuitive characteristics of space have already been\nproved. From the synthetic character of geometrical propositions Kant\nargues[456] that space must be an intuition. Through pure concepts no\nsynthetic knowledge is possible. Then from the apodictic character of\ngeometry he infers that space exists in us as pure and _a priori_;[457]\nno experience can ever reveal necessity. But geometry also exists as an\napplied science; and to account for our power of anticipating\nexperience, we must view space as existing only in the perceiving\nsubject as the form of its sensibility. If it precedes objects as the\nnecessary subjective condition of their apprehension, we can to that\nextent predetermine the conditions of their existence.\n\nIn the concluding paragraph Kant says that this is the only explanation\nwhich can be given of the possibility of geometry. He does not\ndistinguish between pure and applied geometry, though the proof which\nhe has given of each differs in a fundamental respect. Pure geometry\npresupposes only that space is an _a priori_ intuition; applied geometry\ndemands that space be conceived as the _a priori_ form of external\nsense. Only in reference to applied geometry does the Critical problem\narise:--viz. how we can form synthetic judgments _a priori_ which yet\nare valid of objects; or, in other words, how judgments based upon a\nsubjective form can be objectively valid. But any attempt, at this\npoint, to define the nature and possibility of applied geometry must\nanticipate a result which is first established in _Conclusion b_.[458]\nThough, therefore, the substitution of this transcendental exposition\nfor the third space argument is a decided improvement, Kant, in\nextending it so as to cover applied as well as pure mathematics,\noverlooks the real sequence of his argument in the first edition. The\nemployment of the analytic method, breaking in, as it does, upon the\nsynthetic development of Kant\u2019s original argument, is a further\nirregularity.[459]\n\nIt may be noted that in the third paragraph Kant takes the fact that\ngeometry can be applied to objects as proof of the subjectivity of\nspace.[460] He refuses to recognise the possibility that space may be\nsubjective as a form of receptivity, and yet also be a mode in which\nthings in themselves exist. This, as regards its conclusion, though not\nas regards its argument, is therefore an anticipation of _Conclusion a_.\nIn the last paragraph Kant is probably referring to the views both of\nLeibniz and of Berkeley.\n\n\nCONCLUSIONS FROM THE ABOVE CONCEPTS[461]\n\n=Conclusion a.=--_Thesis_: Space is not a property of things in\nthemselves,[462] nor a relation of them to one another. Proof: The\nproperties of things in themselves can never be intuited prior to their\nexistence, _i.e. a priori_. Space, as already proved, is intuited in\nthis manner. In other words, the apriority of space is by itself\nsufficient proof of its subjectivity.\n\nThis argument has been the subject of a prolonged controversy between\nTrendelenburg and Kuno Fischer.[463] Trendelenburg was able to prove his\nmain point, namely, that the above argument is quite inconclusive. Kant\nrecognises only two alternatives, either space as objective is known _a\nposteriori_, or being an _a priori_ representation it is subjective in\norigin. There exists a third alternative, namely, that though our\nintuition of space is subjective in origin, space is itself an inherent\nproperty of things in themselves. The central thesis of the rationalist\nphilosophy of the Enlightenment was, indeed, that the independently real\ncan be known by _a priori_ thinking. Even granting the validity of\nKant\u2019s later conclusion, first drawn in the next paragraph, that space\nis the subjective form of all external intuition, that would only prove\nthat it does not belong to _appearances_, prior to our apprehension of\nthem; nothing is thereby proved in regard to the character of things in\nthemselves. We anticipate by _a priori_ reasoning only the nature of\nappearances, never the constitution of things in themselves. Therefore\nspace, even though _a priori_, may belong to the independently real. The\nabove argument cannot prove the given thesis.\n\nVaihinger contends[464] that the reason why Kant does not even attempt\nto argue in support of the principle, that the _a priori_ must be purely\nsubjective, is that he accepts it as self-evident. This explanation does\nnot, however, seem satisfactory. But Vaihinger supplies the data for\nmodification of his own assertion. It was, it would seem, the existence\nof the antinomies which first and chiefly led Kant to assert the\nsubjectivity of space and time.[465] For as he then believed that a\nsatisfactory solution of the antinomies is possible only on the\nassumption of the subjectivity of space and time, he regarded their\nsubjectivity as being conclusively established, and accordingly failed\nto examine with sufficient care the validity of his additional proof\nfrom their apriority. This would seem to be confirmed by the fact that\nwhen later,[466] in reply to criticisms of the arguments of the first\nedition, he so far modified his position as to offer reasons in support\nof the above general principle, even then he nowhere discussed the\nprinciple in reference to the forms of sense. All his discussions\nconcern only the possible independent reality of the forms of\nthought.[467] To the very last Kant would seem to have regarded the\nabove argument as an independent, and by itself a sufficient, proof of\nthe subjectivity of space.\n\nThe refutation of Trendelenburg\u2019s argument which is offered by\nCaird[468] is inconclusive. Caird assumes the chief point at issue,\nfirst by ignoring the possibility that space may be known _a priori_ in\nreference to appearances and yet at the same time be transcendently\nreal; and secondly by ignoring the fact that to deny spatial properties\nto things in themselves is as great a violation of Critical principles\nas to assert them. One point, however, in Caird\u2019s reply to Trendelenburg\ncalls for special consideration, viz. Caird\u2019s contention that Kant did\nactually take account of the third alternative, rejecting it as\ninvolving the \u201cabsurd\u201d hypothesis of a pre-established harmony.[469]\nUndoubtedly Kant did so. But the contention has no relevancy to the\npoint before us. The doctrine of pre-established harmony is a\nmetaphysical theory which presupposes the possibility of gaining\nknowledge of things in themselves. For that reason alone Kant was bound\nto reject it. A metaphysical proof of the validity of metaphysical\njudgments is, from the Critical point of view, a contradiction in terms.\nAs the validity of _all_ speculations is in doubt, a proof which is\nspeculative cannot meet our difficulties. And also, as Kant himself\nfurther points out, the pre-established harmony, even if granted, can\nafford no solution of the Critical problem how _a priori_ judgments can\nbe passed upon the independently real. The judgments, thus guaranteed,\ncould only possess _de facto_ validity; we could never be assured of\ntheir necessity.[470] It is chiefly in these two inabilities that Kant\nlocates the \u201cabsurdity\u201d of a theory of pre-established harmony. The\nrefutation of that theory does not, therefore, amount to a disproof of\nthe possibility which we are here considering.\n\n=Conclusion b.=--The next paragraph maintains two theses: (_a_) that space\nis the form of all outer intuition; (_b_) that this fact explains what\nis otherwise entirely inexplicable and paradoxical, namely, that we can\nmake _a priori_ judgments which yet apply to the objects experienced.\nThe first thesis, that the pure intuition of space is only conceivable\nas the form of appearances of outer sense, is propounded in the opening\nsentence without argument and even without citation of grounds. The\nstatement thus suddenly made is not anticipated save by the opening\nsentences of the section on space.[471] It is an essentially new\ndoctrine. Hitherto Kant has spoken of space only as an _a priori_\nintuition. The further assertion that as such it must necessarily be\nconceived as the form of outer sense (_i.e._ not only as a formal\nintuition but also as a form of intuition), calls for the most definite\nand explicit proof. None, however, is given. It is really a conclusion\nfrom points all too briefly cited by Kant in the general _Introduction_,\nnamely, from his distinction between the matter and the form of sense.\nThe assertions there made, in a somewhat casual manner, are here,\nwithout notification to the reader, employed as premisses to ground the\nabove assertion. His thesis is not, therefore, as by its face value it\nwould seem to profess to be, an inference from the points established in\nthe preceding expositions. It interprets these conclusions in the light\nof points considered in the _Introduction_; and thereby arrives at a new\nand all-important interpretation of the nature of the _a priori_\nintuition of space.\n\nThe second thesis employs the first to explain how prior to all\nexperience we can determine the relations of objects. Since (_a_) space\nis merely the form of outer sense, and (_b_) accordingly exists in the\nmind prior to all empirical intuition, all appearances must exist in\nspace, and we can predetermine them from the pure intuition of space\nthat is given to us _a priori_. Space, when thus viewed as the _a\npriori_ form of outer sense, renders comprehensible the validity of\napplied mathematics.\n\nAs we have already noted,[472] Kant in the second edition obscures the\nsequence of his argument by offering in the new transcendental\nexposition a justification of applied as well as of pure geometry. In so\ndoing he anticipates the conclusion which is first drawn in this later\nparagraph. This would have been avoided had Kant given two separate\ntranscendental expositions. First, an exposition of pure mathematics,\nplaced immediately after the metaphysical exposition; for pure\nmathematics is exclusively based upon the results of the metaphysical\nexposition. And secondly, an exposition of applied mathematics,\nintroduced after _Conclusion b_. The explanation of applied geometry is\nreally the more essential and central of the two, as it alone involves\nthe truly Critical problem, how judgments formed _a priori_ can yet\napply to objects. _Conclusion b_ constitutes, as Vaihinger rightly\ninsists,[473] the very heart of the _Aesthetic_. The arrangement of\nKant\u2019s argument diverts the reader\u2019s attention from where it ought\nproperly to centre.\n\nThe use which Kant makes of the _Prolegomena_ in his statement of the\nnew transcendental exposition is one cause of the confusion. The\nexposition is a brief summary of the corresponding _Prolegomena_[474]\nsections. In introducing this summary into the _Critique_ Kant\noverlooked the fact that in referring to applied mathematics he is\nanticipating a point first established in _Conclusion b_. The real\ncause, however, of the trouble is common to both editions, namely Kant\u2019s\nfailure clearly to appreciate the fundamental distinction between the\nview that space is an _a priori_ intuition and the view that it is the\n_a priori_ form of all external intuition, _i.e._ of outer sense. He\ndoes not seem to have fully realised how very different are those two\nviews. In consequence of this he fails to distinguish between the\ntranscendental expositions of pure and applied geometry.[475]\n\n=Third paragraph.=--Kant proceeds to develop the subjectivist conclusions\nwhich follow from _a_ and _b_.\n\n     \u201cWe may say that space contains all things which can appear to us\n     externally, but not all things in themselves, whether intuited or\n     not, nor again all things intuited by any and every subject.\u201d[476]\n\nThis sentence makes two assertions: (_a_) space does not belong to\nthings in and by themselves; (_b_) space is not a necessary form of\nintuition for all subjects whatsoever.\n\nThe grounds for the former assertion are not here considered, and that\nis doubtless the reason why the _oder nicht_ is excised in Kant\u2019s\nprivate copy of the _Critique_. As we have seen, Kant does not anywhere\nin the _Aesthetic_ even attempt to offer argument in support of this\nassertion. In defence of (_a_) Kant propounds for the first time the\nview of sensibility as a limitation. Space is a limiting condition to\nwhich human intuition is subject. Whether the intuitions of other\nthinking beings are subject to the same limitation, we have no means of\ndeciding. But for all human beings, Kant implies, the same conditions\nmust hold universally.[477]\n\nIn the phrase \u201ctranscendental ideality of space\u201d[478] Kant, it may be\nnoted, takes the term ideality as signifying subjectivity, and the term\ntranscendental as equivalent to transcendent. He is stating that judged\nfrom a _transcendent_ point of view, _i.e._ from the point of view of\nthe thing in itself, space has a merely subjective or \u201cempirical\u201d\nreality. This is an instance of Kant\u2019s careless use of the term\ntranscendental. Space is empirically real, but taken _transcendently_,\nis merely ideal.[479]\n\n\nKANT\u2019S ATTITUDE TO THE PROBLEMS OF MODERN GEOMETRY\n\nThis is an appropriate point at which to consider the consistency of\nKant\u2019s teaching with modern developments in geometry. Kant\u2019s attitude\nhas very frequently been misrepresented. As he here states, he is\nwilling to recognise that the forms of intuition possessed by other\nraces of finite beings may not coincide with those of the human species.\nBut in so doing he does not mean to assert the possibility of other\n_spatial_ forms, _i.e._ of spaces that are non-Euclidean. In his\npre-Critical period Kant had indeed attempted to deduce the\nthree-dimensional character of space as a consequence of the law of\ngravitation; and recognising that that law is in itself arbitrary, he\nconcluded that God might, by establishing different relations of\ngravitation, have given rise to spaces of different properties and\ndimensions.\n\n     \u201cA science of all these possible kinds of space would undoubtedly\n     be the highest enterprise which a finite understanding could\n     undertake in the field of geometry.\u201d[480]\n\nBut from the time of Kant\u2019s adoption, in 1770, of the Critical view of\nspace as being the universal form of our outer sense, he seems to have\ndefinitely rejected all such possibilities. Space, to be space at all,\nmust be Euclidean; the uniformity of space is a presupposition of the _a\npriori_ certainty of geometrical science.[481] One of the criticisms\nwhich in the _Dissertation_[482] he passes upon the empirical view of\nmathematical science is that it would leave open the possibility that\n\u201ca space may some time be discovered endowed with other fundamental\nproperties, or even perhaps that we may happen upon a two-sided\nrectilinear figure.\u201d This is the argument which reappears in the third\nargument on space in the first edition of the _Critique_.[483] The same\nexamples are employed with a somewhat different wording.\n\n     \u201cIt would not even be necessary that there should be only one\n     straight line between two points, though experience invariably\n     shows this to be so. What is derived from experience has only\n     comparative universality, namely, that which is obtained through\n     induction. We should therefore only be able to say that, so far as\n     hitherto observed, no space has been found which has more than\n     three dimensions.\u201d\n\nBut that Kant should have failed to recognise the possibility of other\nspaces does not by itself point to any serious defect in his position.\nThere is no essential difficulty in reconciling the recognition of such\nspaces with his fundamental teaching. He admits that other races of\nfinite beings may perhaps intuit through _non-spatial_ forms of\nsensibility; he might quite well have recognised that those other forms\nof intuition, though not Euclidean, are still spatial. It is in another\nand more vital respect that Kant\u2019s teaching lies open to criticism. Kant\nis convinced that space is given to us in intuition as being definitely\nand irrevocably Euclidean in character. Both our intuition and our\nthinking, when we reflect upon space, are, he implies, bound down to,\nand limited by, the conditions of Euclidean space. And it is in this\npositive assumption, and not merely in his ignoring of the possibility\nof other spaces, that he comes into conflict with the teaching of modern\ngeometry. For in making the above assumption Kant is asserting that we\ndefinitely know physical space to be three-dimensional, and that by no\nelaboration of concepts can we so remodel it in thought that the axiom\nof parallels will cease to hold. Euclidean space, Kant implies, is\n_given_ to us as an unyielding form that rigidly resists all attempts at\nconceptual reconstruction. Being quite independent of thought and being\ngiven as complete, it has no inchoate plasticity of which thought might\ntake advantage. The modern geometer is not, however, prepared to admit\nthat _intuitional_ space has any definiteness or preciseness of nature\napart from the concepts through which it is apprehended; and he\ntherefore allows, as at least possible, that upon clarification of our\nconcepts space may be discovered to be radically different from what it\nat first sight appears to be. In any case, the perfecting of the\nconcepts must have some effect upon their object. But even--as the\nmodern geometer further maintains--should our space be definitely\nproved, upon analytic and empirical investigation, to be Euclidean in\ncharacter, other possibilities will still remain open for speculative\nthought. For though the nature of our intuitional data may constrain us\nto interpret them through one set of concepts rather than through\nanother, the competing sets of alternative concepts will represent\ngenuine possibilities beyond what the actual is found to embody.\n\nThus the defect of Kant\u2019s teaching, in regard to space, as judged in the\nlight of the later teaching of geometrical science, is closely bound up\nwith his untenable isolation of the _a priori_ of sensibility from the\n_a priori_ of understanding.[484] Space, being thus viewed as\nindependent of thought, has to be regarded as limiting and restricting\nthought by the unalterable nature of its initial presentation. And\nunfortunately this is a position which Kant continued to hold, despite\nhis increasing recognition of the part which concepts must play in the\nvarious mathematical sciences. In the deduction of the first edition we\nfind him stating that synthesis of apprehension is necessary to all\nrepresentation of space and time.[485] He further recognises that all\narithmetical processes are syntheses _according to concepts_.[486] And\nin the _Prolegomena_[487] there occurs the following significant\npassage.\n\n     \u201cDo these laws of nature lie in space, and does the understanding\n     learn them by merely endeavouring to find out the fruitful meaning\n     that lies in space; or do they inhere in the understanding and in\n     the way in which it determines space according to the conditions of\n     the synthetical unity towards which its concepts are all directed?\n     Space is something so uniform and as to all particular properties\n     so indeterminate, that we should certainly not seek a store of laws\n     of nature in it. That which determines space to the form of a\n     circle or to the figures of a cone or a sphere, is, on the\n     contrary, the understanding, so far as it contains the ground of\n     the unity of these constructions. The mere universal form of\n     intuition, called space, must therefore be the substratum of all\n     intuitions determinable to particular objects, and in it, of\n     course, the condition of the possibility and of the variety of\n     these intuitions lies. But the unity of the objects is solely\n     determined by the understanding, and indeed in accordance with\n     conditions which are proper to the nature of the understanding....\u201d\n\nObviously Kant is being driven by the spontaneous development of his own\nthinking towards a position much more consistent with present-day\nteaching, and completely at variance with the hard and fast severance\nbetween sensibility and understanding which he had formulated in the\n_Dissertation_ and has retained in the _Aesthetic_. In the above\n_Prolegomena_ passage a plasticity is being allowed to space, sufficient\nto permit of essential modification in the conceptual processes through\nwhich it is articulated. But, as I have just stated, that did not lead\nKant to disavow the conclusions which he had drawn from his previous\nteaching.\n\nThis defect in Kant\u2019s doctrine of space, as expounded in the\n_Aesthetic_, indicates a further imperfection in his argument. He\nasserts that the form of space cannot vary from one human being to\nanother, and that for this reason the judgments which express it are\nuniversally valid. Now, in so far as Kant\u2019s initial datum is\nconsciousness of time,[488] he is entirely justified in assuming that\neverything which can be shown to be a necessary condition of such\nconsciousness must be uniform for all human minds. But as his argument\nis not that consciousness of _Euclidean_ space is necessary to\nconsciousness of time, but only that consciousness of the _permanent_ in\nspace is a required condition, he has not succeeded in showing the\nnecessary uniformity of the human mind as regards the specific mode in\nwhich it intuits space. The permanent might still be apprehended as\npermanent, and therefore as yielding a possible basis for consciousness\nof sequence, even if it were apprehended in some four-dimensional form.\n\n       *       *       *       *       *\n\n=Fourth Paragraph.=--The next paragraph raises one of the central problems\nof the _Critique_, namely, the question as to the kind of reality\npossessed by appearances. Are they subjective, like taste or colour? Or\nhave they a reality at least relatively independent of the individual\npercipient? In other words, is Kant\u2019s position subjectivism or\nphenomenalism? Kant here alternates between these positions. This fourth\nparagraph is coloured by his phenomenalism, whereas in the immediately\nfollowing fifth paragraph his subjectivism gains the upper hand. The\ntaste of wine, he there states, is purely subjective, because dependent\nupon the particular constitution of the gustatory organ on which the\nwine acts. Similarly, colours are not properties of the objects which\ncause them.\n\n     \u201cThey are only modifications of the sense of sight which is\n     affected in a certain manner by the light.... They are connected\n     with the appearances only as effects accidentally added by the\n     particular constitution of the sense organs.\u201d[489]\n\nSpace, on the other hand, is a necessary constituent of the outer\nobjects. In contrast to the subjective sensations of taste and colour,\nit possesses objectivity. This mode of distinguishing between space and\nthe matter of sense implies that extended objects are not mere ideas,\nbut are sufficiently independent to be capable of acting upon the sense\norgans, and of thereby generating the sensations of the secondary\nqualities.\n\nKant, it must be observed, refers only to taste and colour. He says\nnothing in regard to weight, impenetrability, and the like. These are\nrevealed through sensation, and therefore on his view ought to be in\nexactly the same position as taste or colour. But if so, the relative\nindependence of the extended object can hardly be maintained. Kant\u2019s\ndistinction between space and the sense qualities cannot, indeed, be\nmade to coincide with the Cartesian distinction between primary and\nsecondary qualities.\n\nA second difference, from Kant\u2019s point of view, between space and the\nsense qualities is that the former can be represented _a priori_, in\ncomplete separation from everything empirical, whereas the latter can\nonly be known _a posteriori_. This, as we have seen, is a very\nquestionable assertion. The further statement that all determinations of\nspace can be represented in the same _a priori_ fashion is even more\nquestionable. At most the difference is only between a homogeneous\nsubjective form yielded by outer sense and the endlessly varied and\nconsequently unpredictable contents revealed by the special senses. The\ncontention that the former can be known apart from the latter implies\nthe existence of a pure manifold additional to the manifold of sense.\n\n=Fifth Paragraph.=--In the next paragraph Kant emphasises the distinction\nbetween the empirical and the transcendental meanings of the term\nappearance. A rose, viewed _empirically_, as a thing with an intrinsic\nindependent nature, may appear of different colour to different\nobservers.\n\n     \u201cThe _transcendental_ conception of appearances in space, on the\n     other hand, is a Critical reminder that nothing intuited in space\n     is a thing in itself, that space is not a form inhering in things\n     in themselves ... and that what we call outer objects are nothing\n     but mere representations of our sensibility, the form of which is\n     space.\u201d\n\nIn other words, the distinction drawn in the preceding paragraph between\ncolour as a subjective effect and space as an objective existence is no\nlonger maintained. Kant, when thus developing his position on\nsubjectivist lines, allows no kind of independent existence to anything\nin the known world. Objects as known are mere Ideas (_blosse\nVorstellungen unserer Sinnlichkeit_), the sole correlate of which is the\nunknowable thing in itself. But even in this paragraph both tendencies\nfind expression. \u201cColour, taste, etc., must not rightly be regarded as\nproperties of things, but only as changes in the subject.\u201d This implies\na threefold distinction between subjective sensations, empirical objects\nin space, and the thing in itself. The material world, investigated by\nscience, is recognised as possessing a relatively independent mode of\nexistence.\n\n=Substituted Fourth Paragraph of second edition.=--In preparing the second\nedition Kant himself evidently felt the awkwardness of this abrupt\njuxtaposition of the two very different points of view; and he\naccordingly adopts a non-committal attitude, substituting a logical\ndistinction for the ontological. Space yields synthetic judgments _a\npriori_; the sense qualities do not. Only in the concluding sentence\ndoes there emerge any definite phenomenalist implication. The sense\nqualities, \u201cas they are mere sensations and not intuitions, in\nthemselves reveal no object, least of all [an object] _a priori_.\u201d[490]\nThe assertion that the secondary qualities have no _ideality_ implies a\nnew and stricter use of the term ideal than we find anywhere in the\nfirst edition--a use which runs counter to Kant\u2019s own constant\nemployment of the term. On this interpretation it is made to signify\nwhat though subjective is also _a priori_. Here, as in many of the\nalterations of the second edition, Kant is influenced by the desire to\nemphasise the points which distinguish his idealism from that of\nBerkeley.\n\n\n\n\nTHE TRANSCENDENTAL AESTHETIC\n\n\nSECTION II\n\nTIME\n\nMETAPHYSICAL EXPOSITION OF THE CONCEPTION OF TIME\n\n=Time: First Argument.=--This argument is in all respects the same as the\nfirst argument on space. The thesis is that the representation[491] of\ntime is not of empirical origin. The proof is based on the fact that\nthis representation must be previously given in order that the\nperception of coexistence or succession be possible. It also runs on all\nfours with the first argument in the _Dissertation_.\n\n     \u201c_The idea of time does not originate in, but is presupposed by the\n     senses._ When a number of things act upon the senses, it is only by\n     means of the idea of time that they can be represented whether as\n     simultaneous or as successive. Nor does succession generate the\n     conception of time; but stimulates us to form it. Thus the notion\n     of time, even if acquired through experience, is very badly defined\n     as being a series of actual things existing one _after_ another.\n     For I can understand what the word _after_ signifies only if I\n     already know what time means. For those things are _after_ one\n     another which exist at _different_ times, as those are\n     _simultaneous_ which exist at one and the same time.\u201d[492]\n\n=Second Argument.=--Kant again applies to time the argument already\nemployed by him in dealing with space. The thesis is that time is given\n_a priori_. Proof is found in the fact that it cannot be thought away,\n_i.e._ in the fact of its subjective necessity. From this subjective\nnecessity follows its objective necessity, so far as all appearances are\nconcerned. In the second edition Kant added a phrase--\u201cas the general\ncondition of their possibility\u201d--which is seriously misleading. The\nconcluding sentence is thereby made to read as if Kant were arguing from\nthe objective necessity of time, _i.e._ from its necessity as a\nconstituent in the appearances apprehended, to its apriority. It is\nindeed possible that Kant himself regarded this objective necessity of\ntime as contributing to the proof of its apriority. But no such argument\ncan be accepted. Time may be necessary to appearances, _once appearances\nare granted_. This does not, however, prove that it must therefore\nprecede them _a priori_. This alteration in the second edition is an\nexcellent, though unfortunate, example of Kant\u2019s invincible carelessness\nin the exposition of his thought. It has contributed to a misreading by\nHerbart and others of this and of the corresponding argument on space.\n\n     \u201cLet us not talk of an absolute space as the presupposition of all\n     our constructed figures. Possibility is nothing but thought, and it\n     arises only when it is thought. Space is nothing but possibility,\n     for it contains nothing save images of the existent; and absolute\n     space is nothing save the abstracted general possibility of such\n     constructions, abstracted from it after completion of the\n     construction. The necessity of the representation of space ought\n     never to have played any r\u00f4le in philosophy. To think away space is\n     to think away the _possibility_ of that which has been previously\n     posited as _actual_. Obviously that is impossible, and the opposite\n     is necessary.\u201d[493]\n\nWere Kant really arguing here and in the second argument on space solely\nfrom the _objective_ necessity of time and space, this criticism would\nbe unanswerable. But even taking the argument in its first edition form,\nas an argument from the _psychological_ necessity of time, it lies open\nto the same objection as the argument on space. It rests upon a false\nstatement of fact. We cannot retain time in the absence of all\nappearances of outer and inner sense. With the removal of the given\nmanifold, time itself must vanish.\n\n=Fourth Argument.=[494]--This argument differs only slightly, and mainly\nthrough omissions,[495] from the fourth[496] of the arguments in regard\nto space; but a few minor points call for notice. (_a_) In the first\nsentence, instead of intuition, which alone is under consideration in\nits contrast to conception, Kant employs the phrase \u201cpure form of\nintuition.\u201d (_b_) In the third sentence Kant uses the quite untenable\nphrase \u201cgiven through a single object (_Gegenstand_).\u201d Time is not\ngiven from without, nor is it due to an object. (_c_) The concluding\nsentences properly belong to the transcendental exposition. They are\nhere introduced, not in the ambiguous manner of the fourth[1] argument\non space, but explicitly as a further argument in proof of the intuitive\ncharacter of time. The synthetic proposition which Kant cites is taken\nneither from the science of motion nor from arithmetic. It expresses the\nnature of time itself, and for that reason is immediately contained in\nthe intuition of time.\n\n=Fifth Argument.=--This argument differs fundamentally from the\ncorresponding argument on space, whether of the first or of the second\nedition, and must therefore be independently analysed. The thesis is\nagain that time is an intuition. Proof is derived from the fact that\ntime is a representation in which the parts arise only through\nlimitation, and in which, therefore, the whole must precede the parts.\nThe original (_urspr\u00fcngliche_) time-representation, _i.e._ the\nfundamental representation through limitation of which the parts arise\nas secondary products, must be an intuition.\n\nTo this argument Kant makes two explanatory additions. (_a_) As\nparticular times arise through limitation of one single time, time must\nin its original intuition be given as infinite, _i.e._ as unlimited. The\ninfinitude of time is not, therefore, as might seem to be implied by the\nprominence given to it, and by analogy with the final arguments of both\nthe first and the second edition, a part of the proof that it is an\nintuition, but only a consequence of the feature by which its intuitive\ncharacter is independently established. The unwary reader, having in\nmind the corresponding argument on space, is almost inevitably misled.\nAll reference to infinitude could, so far as this argument is concerned,\nhave been omitted. The mode in which the argument opens seems indeed to\nindicate that Kant was not himself altogether clear as to the\ncross-relations between the arguments on space and time respectively.\nThe real parallel to this argument is to be found in the second part of\nthe fourth[1] argument on space. That part was omitted by Kant in his\nfourth argument on time, and is here developed into a separate argument.\nThis is, of course, a further cause of confusion to the reader, who is\nnot prepared for such arbitrary rearrangement. Indeed it is not\nsurprising to find that when Kant became the reader of his own work, in\npreparing it for the second edition, he was himself misled by the\nintricate perversity of his exposition. In re-reading the argument he\nseems to have forgotten that it represents the second part of the\nfourth[497] argument on space. Interpreting it in the light of the\nfifth[498] argument on space which he had been recasting for the second\nedition, it seemed to him possible, by a slight alteration, to bring\nthis argument on time into line with that new proof.[499] This\nunfortunately results in the perverting of the entire paragraph. The\nargument demands an opposition between intuition in which the whole\nprecedes the parts, and conception in which the parts precede the whole.\nIn order to bring the opposition into line with the new argument on\nspace, according to which a conception contains an infinite number of\nparts, not in it, but only under it, Kant substitutes for the previous\nparenthesis the statement that \u201cconcepts contain only partial\nrepresentations,\u201d meaning, apparently, that their constituent elements\nare merely abstracted attributes, not real concrete parts, or in other\nwords, not strictly parts at all, but only partial representations. But\nthis does not at all agree with the context. The point at issue is\nthereby obscured.\n\n(_b_) The main argument rests upon and presupposes a very definite view\nas to the manner in which alone, according to Kant, concepts are formed.\nOnly if this view be granted as true of all concepts without exception\nis the argument cogent. This doctrine[500] of the concept is accordingly\nstated by Kant in the words of the parenthesis. The partial\nrepresentations, _i.e._ the different properties which go to constitute\nthe object or content conceived, precede the representation of the\nwhole. \u201cThe aggregation of co-ordinate attributes (_Merkmale_)\nconstitutes the totality of the concept.\u201d[501] Upon the use which Kant\nthus makes of the traditional doctrine of the concept, and upon its lack\nof consistency with his recognition of relational categories, we have\nalready dwelt.[502]\n\n=Third Argument and the Transcendental Exposition.=--The third argument\nought to have been omitted in the second edition, and its substance\nincorporated in the new transcendental exposition, as was done with the\ncorresponding argument concerning space. The excuse which Kant offers\nfor not making the change, namely, his desire for brevity, is not valid.\nBy insertion in the new section the whole matter could have been stated\njust as briefly as before.\n\nThe purpose of the transcendental exposition has been already defined.\nIt is to show how time, when viewed in the manner required by the\nresults of the metaphysical deduction, as an _a priori_ intuition,\nrenders synthetic _a priori_ judgments possible.\n\nThis exposition, as it appears in the third argument of the first\nedition, grounds the apodictic character of two axioms in regard to\ntime[503] on the proved apriority of the representation of time, and\nthen by implication finds in these axioms a fresh proof of the apriority\nof time.\n\nThe new transcendental exposition extends the above by two further\nstatements: (_a_) that only through the intuition of time can any\nconception of change, and therewith of motion (as change of place), be\nformed; and (_b_) that it is because the intuition of time is an _a\npriori_ intuition that the synthetic _a priori_ propositions of the\n\u201cgeneral doctrine of motion\u201d are possible. To take each in turn. (_a_)\nSave by reference to time the conception of motion is\nself-contradictory. It involves the ascription to one and the same thing\nof contradictory predicates, _e.g._ that an object both is and is not in\na certain place. From this fact, that time makes possible what is not\npossible in pure conception, Kant, in his earlier rationalistic period,\nhad derived a proof of the subjectivity of time.[504] (_b_) In 1786 in\nthe _Metaphysical First Principles of Natural Science_ Kant had\ndeveloped the fundamental principles of the general science of motion.\nHe takes the opportunity of the second edition (1787) of the _Critique_\nto assign this place to them in his general system. The implication is\nthat the doctrine of motion stands to time in the relation in which\ngeometry stands to space. Kant is probably here replying, as Vaihinger\nhas suggested,[505] to an objection made by Garve to the first edition,\nthat no science, corresponding to geometry, is based on the intuition of\ntime. For two reasons, however, the analogy between mechanics and\ngeometry breaks down. In the first place, the conception of motion is\nempirical; and in the second place, it presupposes space as well as\ntime.[506]\n\nKant elsewhere explicitly disavows this view that the science of motion\nis based on time. He had already done so in the preceding year (1786) in\nthe _Metaphysical First Principles_. He there points out[507] that as\ntime has only one dimension, mathematics is not applicable to the\nphenomena of inner sense. At most we can determine in regard to them (in\naddition, of course, to the two axioms already cited) only the law that\nall these changes are continuous. Also in Kant\u2019s _Ueber Philosophie\n\u00fcberhaupt_ (written some time between 1780 and 1790, and very probably\nin or about the year 1789) we find the following utterance:\n\n     \u201cThe general doctrine of time, unlike the pure doctrine of space\n     (geometry), does not yield sufficient material for a whole\n     science.\u201d[508]\n\nWhy, then, should Kant in 1787 have so inconsistently departed from his\nown teaching? This is a question to which I can find no answer.\nApparently without reason, and contrary to his more abiding judgment, he\nhere repeats the suggestion which he had casually thrown out in the\n_Dissertation_[509] of 1770:\n\n     \u201cPure mathematics treats of space in geometry and of time in pure\n     mechanics.\u201d\n\nBut in the _Dissertation_ the point is only touched upon in passing. The\ncontext permits of the interpretation that while geometry deals with\nspace, mechanics deals with time in addition to space.\n\n\nKANT\u2019S VIEWS REGARDING THE NATURE OF ARITHMETICAL SCIENCE\n\nIn the _Dissertation_, and again in the chapter on _Schematism_ in the\n_Critique_ itself, still another view is suggested, namely, that the\nscience of arithmetic is also concerned with the intuition of time. The\npassage just quoted from the _Dissertation_ proceeds as follows:\n\n     \u201cPure mathematics treats of space in geometry and of time in pure\n     mechanics. To these has to be added a certain concept which is in\n     itself intellectual, but which demands for its concrete\n     actualisation (_actuatio_) the auxiliary notions of time and space\n     (in the successive addition and in the juxtaposition of a\n     plurality). This is the concept of number which is dealt with in\n     _Arithmetic_.\u201d[510]\n\nThis view of arithmetic is to be found in both editions of the\n_Critique_. Arithmetic depends upon the synthetic activity of the\nunderstanding; the conceptual element is absolutely essential.\n\n     \u201cOur counting (as is easily seen in the case of large numbers) is a\n     synthesis according to concepts, because it is executed according\n     to a common ground of unity, as, for instance, the decade\n     (_Dekadik_).\u201d[511] \u201cThe pure image ... of all objects of the senses\n     in general is time. But the pure _schema_ of quantity, in so far as\n     it is a concept of the understanding, is _number_, a representation\n     which combines the successive addition of one to one (homogeneous).\n     Thus number is nothing but the unity of the synthesis of the\n     manifold of a homogeneous intuition in general, whereby I generate\n     time itself in the apprehension of the intuition.\u201d[512]\n\nThis is also the teaching of the _Methodology_.[513] Now it may be\nobserved that in none of these passages is arithmetic declared to be the\n_science of time_, or even to be based on the intuition of time. In\n1783, however, in the _Prolegomena_, Kant expresses himself in much more\nambiguous terms, for his words imply that there is a parallelism between\ngeometry and arithmetic.\n\n     \u201cGeometry is based upon the pure intuition of space. Arithmetic\n     produces its concepts of number through successive addition of\n     units in time, and pure mechanics especially can produce its\n     concepts of motion only by means of the representation of\n     time.\u201d[514]\n\nThe passage is by no means explicit; the \u201cespecially\u201d (_vornehmlich_)\nseems to indicate a feeling on Kant\u2019s part that the description which he\nis giving of arithmetic is not really satisfactory. Unfortunately this\ncasual statement, though never repeated by Kant in any of his other\nwritings, was developed by Schulze in his _Erl\u00e4uterungen_.\n\n     \u201cSince geometry has space and arithmetic has counting as its object\n     (and counting can only take place by means of time), it is evident\n     in what manner geometry and arithmetic, that is to say pure\n     mathematics, is possible.\u201d[515]\n\nLargely, as it would seem,[516] through Schulze, whose _Erl\u00e4uterungen_\ndid much to spread Kant\u2019s teaching, this view came to be the current\nunderstanding of Kant\u2019s position. The nature of arithmetic, as thus\npopularly interpreted, is expounded by Schopenhauer in the following\nterms:\n\n     \u201cIn time every moment is conditioned by the preceding. The ground\n     of existence, as law of the sequence, is thus simple, because time\n     has only one dimension, and no manifoldness of relations can be\n     possible in it. Every moment is conditioned by the preceding; only\n     through the latter can we attain to the former; only because the\n     latter was, and has elapsed, does the former now exist. All\n     counting rests upon this nexus of the parts of time; its words\n     merely serve to mark the single steps of the succession. This is\n     true of the whole of arithmetic, which throughout teaches nothing\n     but the methodical abbreviations of counting. Every number\n     presupposes the preceding numbers as grounds of its existence; I\n     can only reach them through all the preceding, and only by means of\n     this insight into the ground of its existence do I know that, where\n     ten are, there are also eight, six, four.\u201d[517]\n\nSchulze was at once challenged to show that this was really Kant\u2019s\nteaching, and the passage which he cited was Kant\u2019s definition of the\nschema of number, above quoted.[518] It is therefore advisable that we\nshould briefly discuss the many difficulties which this passage\ninvolves. What does Kant mean by asserting that in the apprehension of\nnumber we generate time? Does he merely mean that time is required for\nthe process of counting? Counting is a process through which numerical\nrelations are discovered; and it undoubtedly occupies time. But so do\nall processes of apprehension, in the study of geometry no less than of\narithmetic. That this is not Kant\u2019s meaning, and that it is not even\nwhat Schulze, notwithstanding his seemingly explicit mode of statement,\nintends to assert, is clearly shown by a letter written by Kant to\nSchulze in November 1788. Schulze, it appears, had spoken of this very\nmatter.\n\n     \u201c_Time, as you justly remark, has no influence upon the properties\n     of numbers_ (as pure determinations of quantity), such as it may\n     have upon the nature of those changes (of quantity) which are\n     possible only in connection with a specific property of inner sense\n     and its form (time). _The science of number, notwithstanding the\n     succession which every construction of quantity demands, is a pure\n     intellectual synthesis which we represent to ourselves in thought._\n     But so far as _quanta_ are to be numerically determined, they must\n     be given to us in such a way that we can apprehend their intuition\n     in successive order, and such that their _apprehension_ can be\n     subject to time....\u201d[519]\n\nNo more definite statement could be desired of the fact that though in\narithmetical science as in other fields of study our processes of\napprehension are subject to time, the quantitative relations determined\nby the science are independent of time and are intellectually\napprehended.\n\nBut if the above psychological interpretation of Kant\u2019s teaching is\nuntenable, how is his position to be defined? We must bear in mind the\ndoctrine which Kant had already developed in his pre-Critical period,\nthat mathematical differs from philosophical knowledge in that its\nconcepts can have concrete individual form.[520] In the _Critique_ this\ndifference is expressed in the statement that the mathematical sciences\nalone are able to _construct_ their concepts. And as they are _pure_\nmathematical sciences, this construction is supposed to take place by\nmeans of the _a priori_ manifold of space and of time. Now though Kant\nhad a fairly definite notion of what he meant by the construction of\ngeometrical figures in space, his various utterances seem to show that\nin regard to the nature of arithmetical and algebraic construction he\nhad never really attempted to arrive at any precision of view. To judge\nby the passage already quoted[521] from the _Dissertation_, Kant\nregarded space as no less necessary than time to the construction or\nintuition of number. \u201d[The intellectual concept of number] demands for\nits concrete actualisation the auxiliary notions of time _and space_ (in\nthe successive addition and in the _juxtaposition of a plurality_)\u201d A\nsimilar view appears in the _Critique_ in A 140 = B 179 and in B 15. In\nconformity, however, with the general requirements of his doctrine of\n_Schematism_, Kant defines the schema of number in exclusive reference\nto time; and, as we have noted, it is to this definition that Schulze\nappeals in support of his view of arithmetic as the science of counting\nand therefore of time. It at least shows that Kant perceived _some_ form\nof connection to exist between arithmetic and time. But in this matter\nKant\u2019s position was probably simply a corollary from his general view of\nthe nature of mathematical science, and in particular of his view of\ngeometry, the \u201cexemplar\u201d[522] of all the others. Mathematical science,\nas such, is based on intuition;[523] therefore arithmetic, which is one\nof its departments, must be so likewise. No attempt, however, is made\nto define the nature of the intuitions in which it has its source.\nSympathetically interpreted, his statements may be taken as suggesting\nthat arithmetic is the study of _series_ which find concrete expression\nin the order of sequent times. The following estimate, given by\nCassirer,[524] does ample justice both to the true and to the false\nelements in Kant\u2019s doctrine.\n\n     \u201d[Even discounting Kant\u2019s insistence upon the conceptual character\n     of arithmetical science, and] allowing that he derives arithmetical\n     concepts and propositions from the _pure intuition of time_, this\n     teaching, to whatever objections it may lie open, has certainly not\n     the merely _psychological_ meaning which the majority of its\n     critics have ascribed to it. If it contained only the trivial\n     thought, that the empirical act of counting requires time, it would\n     be completely refuted by the familiar objection which B. Beneke has\n     formulated: \u2018The fact that time elapses in the process of counting\n     can prove nothing; for what is there over which time does not\n     flow?\u2019 It is easily seen that Kant is only concerned with the\n     \u2018transcendental\u2019 determination of the concept of time, according to\n     which it appears as the type of an ordered sequence. William\n     [Rowan] Hamilton, who adopts Kant\u2019s doctrine, has defined algebra\n     as \u2018science of pure time or _order in progression_.\u2019 That the whole\n     content of arithmetical concepts can really be obtained from the\n     fundamental concept of _order_ in unbroken development, is\n     completely confirmed by Russell\u2019s exposition. As against the\n     Kantian theory it must, of course, be emphasised, that it is not\n     the _concrete_ form of time intuition which constitutes the\n     _ground_ of the concept of number, but that on the contrary the\n     pure logical concepts of sequence and of order are already\n     implicitly contained and embodied in that concrete form.\u201d\n\nMuch of the unsatisfactoriness of Kant\u2019s argument is traceable to his\nmode of conceiving the \u201cconstruction\u201d[525] of mathematical concepts. All\nconcepts, he seems to hold, even those of geometry and arithmetic, are\nabstract class concepts--the concept of triangle representing the\nproperties common to all triangles, and the concept of seven the\nproperties common to all groups that are seven. Mathematical concepts\ndiffer, however, from other concepts in that they are capable of _a\npriori_ construction, that is, of having their objects represented in\npure intuition. Now this is an extremely unfortunate mode of statement.\nIt implies that mathematical concepts have a dual mode of existence,\nfirst as abstracted, and secondly as constructed. Such a position is not\ntenable. The concept of seven, in its primary form, is not abstracted\nfrom a variety of particular groups of seven; it is already involved in\nthe apprehension of each of them as being seven. Nor is it a concept\nthat is itself constructed. It may perhaps be described as being the\nrepresentation of something constructed; but that something is not\nitself. It represents the process or method generative of the complex\nfor which it stands. Thus Kant\u2019s distinction between the intuitive\nnature of mathematical knowledge and the merely discursive character of\nconceptual knowledge is at once inspired by the very important\ndistinction between the product of construction and the product of\nabstraction, and yet at the same time is also obscured by the quite\ninadequate manner in which that latter distinction has been formulated.\nKant has again adhered to the older logic even in the very act of\nrevising its conclusions; and in so doing he has sacrificed the Critical\ndoctrines of the _Analytic_ to the pre-Critical teaching of the\n_Dissertation_ and _Aesthetic_. _Mathematical concepts are of the same\ngeneral type as the categories; their primary function is not to clarify\nintuitions, but to make them possible._ They are derivable from\nintuition only in so far as they have contributed to its constitution.\nIf intuition contains factors additional to the concepts through which\nit is interpreted, these factors must remain outside the realm of\nmathematical science, until such time as conceptual analysis has proved\nitself capable of further extension.\n\nI may now summarise this general discussion. Though Kant in the first\nedition of the _Critique_ had spoken of the mathematical sciences as\nbased upon the intuition of space and time, he had not, despite his\nconstant tendency to conceive space and time as parallel forms of\nexistence, based any separate mathematical discipline upon time. His\ndefinition of number, in the chapter on _Schematism_, had recognised the\nessentially conceptual character of arithmetic, and had connected it\nwith time only in a quite indirect manner. A passage in the\n_Prolegomena_ is the one place in all Kant\u2019s writings in which he would\nseem to assert, though in brief and quite indefinite terms, that\narithmetic is related to time as geometry is related to space. No such\nview of arithmetic is to be found in the second edition of the\n_Critique_. In the transcendental exposition of time, added in the\nsecond edition, only pure mechanics is mentioned. This would seem to\nindicate that Kant had made the above statement carelessly, without due\nthought, and that on further reflection he found himself unable to stand\nby it. The omission is the more significant in that Kant refers to\narithmetic in the passages added in the second edition _Introduction_.\nThe teaching of these passages, apart from the asserted necessity of\nappealing to fingers or points,[526] harmonises with the view so\nbriefly outlined in the _Analytic_. Arithmetic is a conceptual science;\nthough it finds in ordered sequence its intuitional material, it cannot\nbe adequately defined as being the science of time.\n\n\nCONCLUSIONS FROM THE PRECEDING CONCEPTS[527]\n\nThese _Conclusions_ do not run parallel with the corresponding\n_Conclusions_ in regard to space. In the first paragraph there are two\ndifferences. (_a_) Kant takes account of a view not considered under\nspace, viz. that time is a self-existing substance. He rejects it on a\nground which is difficult to reconcile with his recognition of a\nmanifold of intuition as well as a manifold of sense, namely that it\nwould then be something real without being a real object. In A 39 = B 57\nand B 70 Kant describes space and time, so conceived, as _unendliche\nUndinge_. (_b_) Kant introduces into his first _Conclusion_ the\nargument[528] that only by conceiving time as the form of inner\nintuition can we justify _a priori_ synthetic judgments in regard to\nobjects.\n\n=Second Paragraph (Conclusion b).=--This latter statement is repeated at\nthe opening of the second _Conclusion_. The emphasis is no longer,\nhowever, upon the term \u201cform\u201d but upon the term \u201cinner\u201d; and Kant\nproceeds to make assertions which by no means follow from the five\narguments, and which must be counted amongst the most difficult and\ncontroversial tenets of the whole _Critique_. (_a_) Time is not a\ndetermination of outer appearances. For it belongs neither to their\nshape nor to their position--and prudently at this point the property of\nmotion is smuggled out of view under cover of an etc. _Time does not\ndetermine the relation of appearances to one another, but only_ the\nrelation of representations _in our inner state_.[529] It is the form\nonly of the intuition of ourselves and of our inner state.[530]\nObviously these are assertions which Kant cannot possibly hold to in\nthis unqualified form. In the very next paragraph they are modified and\nrestated. (_b_) As this inner intuition supplies no shape (_Gestalt_),\nwe seek to make good this deficiency by means of analogies. We represent\nthe time-sequence through a line progressing to infinity in which the\nmanifold constitutes a series of only one dimension. From the properties\nof this line, with the one exception that its parts are simultaneous\nwhereas those of time are always successive, we conclude to all the\nproperties of time.\n\nThe wording of the passage seems to imply that such symbolisation of\ntime through space is helpful but not indispensably necessary for its\napprehension. That it is indispensably necessary is, however, the view\nto which Kant finally settled down.[531] But he has not yet come to\nclearness on this point. The passage has all the signs of having been\nwritten prior to the _Analytic_. Though Kant seems to have held\nconsistently to the view that time has, in or by itself, only one\ndimension,[532] the difficulties involved drove him to recognise that\nthis is true only of time as the order of our representations. It is not\ntrue of the objective time apprehended in and through our\nrepresentations. When later Kant came to hold that consciousness of time\nis conditioned by consciousness of space, he apparently also adopted the\nview that, by reference to space, time indirectly acquires simultaneity\nas an additional mode. The objective spatial world is in time, but in a\ntime which shows simultaneity as well as succession. In the\n_Dissertation_[533] Kant had criticised Leibniz and his followers for\nneglecting simultaneity, \u201cthe most important consequence of time.\u201d\n\n     \u201cThough time has only one dimension, yet the _ubiquity_ of time (to\n     employ Newton\u2019s term), through which all things sensuously\n     thinkable are _at some time_, adds another dimension to the\n     quantity of actual things, in so far as they hang, as it were, upon\n     the same point of time. For if we represent time by a straight line\n     extended to infinity, and simultaneous things at any point of time\n     by lines successively erected [perpendicular to the first line],\n     the surface thus generated will represent the _phenomenal world_\n     both as to substance and as to accidents.\u201d\n\nSimilarly in A 182 = B 226 of the _Critique_ Kant states that\nsimultaneity is not a mode of time,[534] since none of the parts of time\ncan be simultaneous, and yet also teaches in A 177 = B 219 that, as the\norder of _appearances_, time possesses in addition to succession the two\nmodes, duration and simultaneity. The significance of this distinction\nbetween time as the order of our inner states, and time as the order of\nobjective appearances, we shall consider immediately.\n\nA connected question is as to whether or not Kant teaches the\npossibility of simultaneous apprehension. In the _Aesthetic_ and\n_Dialectic_ he certainly does so. Space is given as containing\ncoexisting parts, and[535] can be intuited as such without successive\nsynthesis of its parts. In the _Analytic_, on the other hand, the\nopposite would seem to be implied.[536] The apprehension of a manifold\ncan only be obtained through the successive addition or generation of\nits parts.\n\n(_c_) Lastly, Kant argues that the fact that all the relations of time\ncan be expressed in an outer intuition is proof that the representation\nof time is itself intuition. But surely if, as Kant later taught, time\ncan be apprehended at all only in and through space, that, taken alone,\nwould rather be a reason for denying it to be itself intuition. In any\ncase it is difficult to follow Kant in his contention that the intuition\nof time is similar in general character to that of space.[537]\n\n=Third Paragraph (Conclusion c).=--Kant now reopens the question as to the\nrelation in which time stands to outer appearances. As already noted, he\nhas argued in the beginning of the previous paragraph that it cannot be\na determination of outer appearances, but only of representations in our\ninner state. External appearances, however, as Kant recognises, can be\nknown only in and through representations. To that extent they belong to\ninner sense, and consequently (such is Kant\u2019s argument) are themselves\nsubject to time. Time, as the immediate condition of our\nrepresentations, is also the mediate condition of appearances.\nTherefore, Kant concludes, \u201call _appearances_, _i.e._ all _objects_ of\nthe senses, are in time, and necessarily stand in time-relations.\u201d\n\nNow quite obviously this argument is invalid if the distinction between\nrepresentations and their objects is a real and genuine one. For if so,\nit does not at all follow that because our _representations_ of objects\nare in time that the objects themselves are in time. In other words, the\nargument is valid only from the standpoint of extreme subjectivism,\naccording to which objects are, in Kant\u2019s own phraseology, _blosse\nVorstellungen_. But the argument is employed to establish a realist\nconclusion, that outer objects, as objects, stand in time-relations to\none another. In contradiction of the previous paragraph he is now\nmaintaining that time is a determination of outer appearances, and that\nit reveals itself in the motion of bodies as well as in the flux of our\ninner states.\n\nThe distinction between representations and their objects also makes it\npossible for Kant both to assert and to deny that simultaneity is a mode\nof time. \u201cNo two years can be coexistent. Time has only one dimension.\nBut existence (_das Dasein_), measured through time, has two dimensions,\nsuccession and simultaneity.\u201d There are, for Kant, two orders of time,\nsubjective and objective. Recognition of the latter (emphasised and\ndeveloped in the _Analytic_)[538] is, however, irreconcilable with his\ncontention that time is merely the form of inner sense.\n\nWe have here one of the many objections to which Kant\u2019s doctrine of time\nlies open. It is the most vulnerable tenet in his whole system. A mere\nlist of the points which Kant leaves unsettled suffices to show how\ngreatly he was troubled in his own mind by the problems to which it\ngives rise. (1) The nature of the _a priori_ knowledge which time\nyields. Kant ascribes to this source sometimes only the two axioms in\nregard to time, sometimes pure mechanics, and sometimes also arithmetic.\n(2) Whether time only allows of, or whether it demands, representation\nthrough space. Sometimes Kant makes the one assertion, sometimes the\nother. (3) Whether it is possible to apprehend the coexistent without\nsuccessive synthesis of its parts. This possibility is asserted in the\n_Aesthetic_ and _Dialectic_, denied in the _Analytic_. (4) Whether\nsimultaneity is a mode of time. (5) Whether, and in what manner,\nappearances of outer sense are in time. Kant\u2019s answer to 4 and to 5\nvaries according as he identifies or distinguishes representations and\nempirical objects.\n\nThe manifold difficulties to which a theory of time thus lies open are\nprobably the reason why Kant, in the _Critique_, reverses the order in\nwhich he had treated time and space in the _Dissertation_.[539] But the\nplacing of space before time is none the less unfortunate. It greatly\ntends to conceal from the reader the central position which Kant has\nassigned to time in the _Analytic_. Consciousness of time is the\nfundamental fact, taken as bare fact, by reference to which Kant gains\nhis transcendental proof of the categories and principles of\nunderstanding.[540] In the _Analytic_ space, by comparison, falls very\nmuch into the background. A further reason for the reversal may have\nbeen Kant\u2019s Newtonian view of geometry as the mathematical science _par\nexcellence_.[541] In view of his formulation of the Critical problem as\nthat of accounting for synthetic _a priori_ judgments, he would then\nnaturally be led to throw more emphasis on space.\n\nTo sum up our main conclusions. Kant\u2019s view of time as a form merely of\ninner sense, and as having only one dimension, connects with his\nsubjectivism. His view of it as inhering in objects, and as having\nduration and simultaneity as two of its modes, is bound up with his\nphenomenalism. Further discussion of these difficulties must therefore\nbe deferred until we are in a position to raise the more fundamental\nproblem as to the nature of the distinction between a representation and\nits object.[542] Motion is not an inner state. Yet it involves time as\ndirectly as does the flow of our feelings and ideas. Kant\u2019s assertion\nthat \u201ctime can no more be intuited externally than space can be intuited\nas something in us,\u201d[543] if taken quite literally, would involve both\nthe subjectivist assertion that motion of bodies is non-existent, and\nalso the phenomenalist contention that an extended object is altogether\ndistinct from a representation.\n\nThe _fourth_ and _fifth_ paragraphs call for no detailed analysis.[544]\nTime is empirically real, transcendentally ideal--these terms having\nexactly the same meaning and scope as in reference to space.[545] The\nfourth sentence in the fifth paragraph is curiously inaccurate. As it\nstands, it would imply that time is given through the senses. In the\nconcluding sentences Kant briefly summarises and applies the points\nraised in these fourth and fifth paragraphs.\n\n\nELUCIDATION\n\n=First and Second Paragraphs.=--Kant here replies to a criticism which, as\nhe tells us in his letter of 1772 to Herz, was first made by Pastor\nSchulze and by Lambert.[546] In that letter the objection and Kant\u2019s\nreply are stated as follows.\n\n     \u201cIn accordance with the testimony of inner sense, changes are\n     something real. But they are only possible on the assumption of\n     time. Time is, therefore, something real which belongs to the\n     determinations of things in themselves. Why, said I to myself, do\n     we not argue in a parallel manner: \u2018Bodies are real, in accordance\n     with the outer senses. But bodies are possible only under the\n     condition of space. Space is, therefore, something objective and\n     real which inheres in the things themselves.\u2019 The cause [of this\n     differential treatment of space and of time] is the observation\n     that in respect to outer things we cannot infer from the reality of\n     representations the reality of their objects, whereas in inner\n     sense the thought or the existing of the thought and of myself are\n     one and the same. Herein lies the key to the difficulty.\n     Undoubtedly I must think my own state under the form of time, and\n     the form of the inner sensibility consequently gives me the\n     appearance of changes. Now I do not deny that changes are something\n     real any more than I deny that bodies are something real, but I\n     thereby mean only that something real corresponds to the\n     appearance. I may not even say the inner appearance undergoes\n     change (_ver\u00e4ndere sich_), for how could I observe this change\n     unless it appeared to my inner sense? _To the objection that this\n     leads to the conclusion that all things in the world objectively\n     and in themselves are unchangeable, I would reply that they are\n     neither changeable nor unchangeable._ As Baumgarten states in \u00a7 18\n     of his _Metaphysica_, the absolutely impossible is hypothetically\n     neither possible nor impossible, since it cannot be mentally\n     entertained under any condition whatsoever; so in similar manner\n     _the things of the world are objectively or in themselves neither\n     in one and the same state nor in different states at different\n     times, for thus understood [viz. as things in themselves] they are\n     not represented in time at all_.\u201d[547]\n\nThus Kant\u2019s contention, both in this letter and in the passage before\nus, is that even our inner states would not reveal change if they could\nbe apprehended by us or by some other being apart from the subjective\nform of our inner sense. We may not say that our inner states undergo\nchange, or that they succeed one another, but only that to us they\nnecessarily appear as so doing.[548] Time is no more than subjectively\nreal.[549] As K\u00f6rner writes to Schiller: \u201cWithout time man would indeed\n_exist_ but not _appear_. Not his reality but only his appearance is\ndependent upon the condition of time.\u201d \u201cMan _is_ not, but only\n_appears_, when he undergoes change.\u201d[550] The objects of inner sense\nstand in exactly the same position as those of outer sense. Both are\nappearances, and neither can be identified with the absolutely real. As\nKant argues later in the _Critique_,[551] inner processes are not known\nwith any greater certainty or immediacy than are outer objects; the\nreality of time as subjective proves its unreality in relation to things\nin themselves. The statement that the constitution of things in\nthemselves is \u201cproblematic\u201d is an exceptional mode of expression for\nKant. Usually--as indeed throughout the whole context of this\npassage[552]--he asserts that though things in themselves are\nunknowable, we can with absolute certainty maintain that they are\nneither in space nor in time. Upon this point we have already dwelt in\ndiscussing Trendelenburg\u2019s controversy with Fischer.[553]\n\n=Third Paragraph.=--The third and fourth paragraphs of this section ought\nto have had a separate heading. They summarise the total argument of the\n_Aesthetic_ in regard to space as well as time, distinguish its tenets\nfrom those of Newton and of Leibniz, and draw a general conclusion. The\nsummary follows the strict synthetic method. The opening sentences\nillustrate Kant\u2019s failure to distinguish between the problems of pure\nand of applied mathematics, and also show how completely he tends to\nconceive mathematics as typified by geometry. The criticism of\nalternative views traverses the ground of the famous controversy between\nLeibniz and Clarke. Their _Streitschriften_ were, as we have good\ncircumstantial grounds for believing,[554] a chief influence in the\ndevelopment of Kant\u2019s own views. Kant, who originally held the\nLeibnizian position, was by 1768[555] more or less converted to the\nNewtonian teaching, and in the _Dissertation_ of 1770 developed his\nsubjectivist standpoint with the conscious intention of retaining the\nadvantages while remedying the defects of both alternatives.[556] For\nconvenience we may limit the discussion to space. (_a_) The view\npropounded by Newton, and defended by Clarke, is that space has an\nexistence in and by itself, independent alike of the mind which\napprehends it and of the objects with which it is filled. (_b_) The view\nheld by Leibniz is that space is an empirical concept abstracted from\nour confused sense-experience of the relations of real things.[557]\n\nThe criticism of (_a_) is twofold. First, it involves belief in an\neternal and infinite _Unding_. Secondly, it leads to metaphysical\ndifficulties, especially in regard to the existence of God. If space is\nabsolutely real, how is it to be reconciled with the omnipresence of\nGod? Newton\u2019s view of space as the =sensorium Dei= can hardly be regarded\nas satisfactory.\n\nThe objection to (_b_) is that it cannot account for the apodictic\ncertainty of geometry, nor guarantee its application to experience. The\nconcept of space, when regarded as of sensuous origin, is something that\nmay distort (and according to the Leibnizian teaching does actually\ndistort) what it professes to represent, and is something from which\nrestrictions that hold in the natural world have been omitted.[558] As\nempirical, it cannot serve as basis for the universal and necessary\njudgments of mathematical science.[559]\n\nThe first view has, however, the advantage of keeping the sphere of\nappearances open for mathematical science. As space is infinite and\nall-comprehensive, its laws hold universally. The second view has the\nadvantage of not subjecting reality to space conditions. These\nadvantages are retained, while the objections are removed, by the\nteaching of the _Aesthetic_.\n\nKant further criticises the former view in A 46 ff. = B 64 ff. There is\nno possibility of accounting for the _a priori_ synthetic judgments of\ngeometry save by assuming that space is the pure form of outer\nintuition. For though the Newtonian view will justify the assertion that\nthe laws of space hold universally, it cannot explain how we come to\nknow them _a priori_. And assuming, as Kant constantly does, that space\ncannot be both an _a priori_ form of intuition and also independently\nreal, he concludes that it is the former only.\n\nIn B 71 Kant also restates the metaphysical difficulties to which the\nNewtonian view lies open. In natural theology we deal with an existence\nwhich can never be the object of sensuous intuition, and which has to be\nfreed from all conditions of space and time. This is impossible if space\nis so absolutely real that it would remain though all created things\nwere annihilated.\n\n=Fourth Paragraph.=--Space and time are the only two forms of sensibility;\nall other concepts belonging to the senses, such as motion and change,\nare empirical.[560] As Kant has himself stated, no reason can be given\nwhy space and time are the sole forms of our possible intuition:\n\n     \u201cOther forms of intuition than space and time, ... even if they\n     were possible, we cannot render in any way conceivable and\n     comprehensible to ourselves, and even assuming that we could do so,\n     they still would not belong to experience, the only kind of\n     knowledge in which objects are given to us.\u201d[561]\n\nThe further statement,[562] frequently repeated in the _Critique_, that\ntime itself does not change, but only what is in time,[563] indicates\nthe extent to which Kant has been influenced by the Newtonian receptacle\nview. As Bergson very justly points out, time, thus viewed as a\nhomogeneous medium, is really being conceived on the analogy of space.\n\u201cIt is merely the phantom of space obsessing the reflective\nconsciousness.\u201d[564]\n\n\nGENERAL OBSERVATIONS ON THE TRANSCENDENTAL AESTHETIC\n\n=I. First Paragraph.=--\u201cTo avoid all misapprehension\u201d Kant proceeds to\nstate \u201cas clearly as possible\u201d his view of sensuous knowledge. With this\nend in view he sets himself to enforce two main points: (_a_) that as\nspace and time are only forms of sensibility, everything apprehended is\nonly appearance; (_b_) that this is not a mere hypothesis but is\ncompletely certain. Kant expounds (_a_) indirectly through criticism of\nthe opposing views of Leibniz and of Locke. But before doing so he makes\nin the next paragraph a twofold statement of his own conclusions.\n\n=Second Paragraph.=--This paragraph states (_a_) that through intuition we\ncan represent only appearances, not things in themselves, and (_b_) that\nthe appearances thus known exist only in us. Both assertions have\nimplications, the discussion of which must be deferred to the\n_Analytic_. The mention of the \u201crelations of things by themselves\u201d may,\nas Vaihinger suggests,[565] be a survival from the time when (as in the\n_Dissertation_[566]) Kant sought to reduce spatial to dynamical\nrelations. The assertion that things in themselves are completely\nunknown to us goes beyond what the _Aesthetic_ can establish and what\nKant here requires to prove. His present thesis is only that no\nknowledge of things in themselves can be acquired either through the\nforms of space and time or through sensation; space and time are\ndetermined solely by our pure sensibility, and sensations by our\nempirical sensibility. Failure to recognise this is, in Kant\u2019s view, one\nof the chief defects of the Leibnizian system.\n\n=Third and Fourth Paragraphs. Criticism of the Leibniz-Wolff\nInterpretation of Sensibility and of Appearance.=--Leibniz vitiates both\nconceptions. Sensibility does not differ from thought in clearness but\nin content. It is a difference of kind.[567] They originate in different\nsources, and neither can by any transformation be reduced to the other.\n\n     \u201cEven if an appearance could become completely transparent to us,\n     such knowledge would remain _toto coelo_ different from knowledge\n     of the object in itself.\u201d[568] \u201cThrough observation and analysis of\n     appearances we penetrate to the secrets of nature, and no one can\n     say how far this may in time extend.... [But however far we\n     advance, we shall never be able by means of] so ill-adapted an\n     instrument of investigation [as our sensibility] to find anything\n     except still other appearances, the non-sensuous cause of which we\n     yet long to discover.\u201d[569]\n\nWe should still know only in terms of the two inalienable forms of our\nsensibility.[570] The dualism of thought and sense can never be\ntranscended by the human mind. By no extension of its sphere or\nperfecting of its insight can sensuous knowledge be transformed into a\nconceptual apprehension of purely intelligible entities.\n\nLeibniz\u2019s conception of appearances as things in themselves confusedly\napprehended is equally false, and for the same reasons.[571] Appearance\nand reality are related as distinct existences, each of which has its\nown intrinsic character and content. Through the former there can be no\nhope of penetrating to the latter. Appearance is subjective in matter as\nwell as in form. For Leibniz our knowledge of appearances is a confused\nknowledge of things in themselves. Properly viewed, it is the\napprehension, whether distinct or confused, of objects which are never\nthings in themselves. Sense-knowledge, such as we obtain in the science\nof geometry, has often the highest degree of clearness. Conceptual\napprehension is all too frequently characterised by obscurity and\nindistinctness.\n\nThis criticism of Leibniz, as expounded in these two paragraphs, is\nthoroughly misleading if taken as an adequate statement of Kant\u2019s view\nof the relations between sense and understanding, appearance and\nreality. These paragraphs are really a restatement of a passage in the\n_Dissertation_.\n\n     \u201cIt will thus be seen that we express the nature of the sensuous\n     very inappropriately when we assert that it is the _more\n     confusedly_ known, and the nature of the intellectual when we\n     describe it as the _distinctly_ known. For these are merely logical\n     distinctions, and obviously have nothing to do with the given facts\n     which underlie all logical comparison. The sensuous may be\n     absolutely distinct, and the intellectual extremely confused. That\n     is shown on the one hand in _geometry_, the prototype of sensuous\n     knowledge, and on the other in _metaphysics_, the instrument of all\n     intellectual enquiry. Every one knows how zealously metaphysics has\n     striven to dispel the mists of confusion which cloud the minds of\n     men at large and yet has not always attained the happy results of\n     the former science. Nevertheless each of these kinds of knowledge\n     preserves the mark of the stock from which it has sprung. The\n     former, however distinct, is on account of its origin entitled\n     sensuous, while the latter, however confused, remains\n     intellectual--as _e.g._ the _moral_ concepts, which are known not\n     by way of experience, but through the pure intellect itself. I\n     fear, however, that Wolff by this distinction between the sensuous\n     and the intellectual, which for him is merely logical, has checked,\n     perhaps wholly (to the great detriment of philosophy), that noblest\n     enterprise of antiquity, the investigation of _the nature of\n     phenomena and noumena_, turning men\u2019s minds from such enquiries to\n     what are very frequently only logical subleties.\u201d[572]\n\nThe paragraphs before us give expression only to what is common to the\n_Dissertation_ and to the _Critique_, and do so entirely from the\nstandpoint of the _Dissertation_. Thus the illustration of the\nconception of \u201cright\u201d implies that things in themselves can be known\nthrough the understanding. The conception, as Kant says, represents \u201ca\nmoral property which belongs to actions in and by themselves.\u201d\nSimilarly, in distinguishing the sensuous from \u201cthe intellectual,\u201d he\nsays that through the former we do not apprehend things in themselves,\nthus implying that things in themselves can be known through the pure\nintellect. The view developed in the _Analytic_, alike of sensibility\nand of appearance, is radically different. Sensibility and understanding\n_may_ have a common source; and both are indispensably necessary for the\napprehension of appearance. Neither can function save in co-operation\nwith the other. Appearance does not differ from reality solely through\nits sensuous content and form, but also in the intellectual order or\ndispensation to which it is subject. But in the very act of thus\ndeepening the gulf between appearance and reality by counting even\nunderstanding as contributing to the knowledge only of the former, he\nwas brought back to a position that has kinship with the Leibnizian view\nof their interrelation. Since understanding is just as essential as\nsensibility to the apprehension of appearances, and since understanding\ndiffers from sensibility in the universality of its range, it enables us\nto view appearances in their relation to ultimate reality, and so to\napprehend them as being, however subjective or phenomenal, ways in which\nthe thing in itself presents itself to us. Such a view is, however, on\nKant\u2019s principles, quite consistent with the further contention, that\nappearance does not differ from reality in a merely logical manner.\nFactors that are peculiar to the realm of appearance have intervened to\ntransform the real; and in consequence even completed knowledge of the\nphenomenal--if such can be conceived as possible--would not be\nequivalent to knowledge of things in themselves.\n\n=Fifth Paragraph. Criticism of Locke\u2019s View of Appearance.=--This\nparagraph discusses Locke\u2019s doctrine[573] that the secondary qualities\nare subjective, and that in the primary qualities we possess true\nknowledge of things in themselves. The distinction is drawn upon\nempirical grounds, namely, that while certain qualities are uniform for\nmore than one sense, and belong to objects under all conditions, others\nare peculiar to the different senses, and arise only through the\naccidental relation of objects to the special senses.[574] This\ndistinction is, Kant says, entirely justified from the physical\nstandpoint.[575] A rainbow is an appearance of which the raindrops\nconstitute the true empirical reality. But Locke and his followers\ninterpret this distinction wrongly. They ignore the more fundamental\ntranscendental (_i.e._ metaphysical) distinction between empirical\nreality and the thing in itself. From the transcendental standpoint the\nraindrops are themselves merely appearance. Even their round shape, and\nthe very space in which they fall; are only modifications of our\nsensuous intuition. The \u2018transcendental object\u2019[576] remains unknown to\nus.\n\nWhen Kant thus declares that the distinction between primary and\nsecondary qualities is justified (_richtig_) from the physical\nstandpoint, he is again[577] speaking from a phenomenalist point of\nview. And it may be noted that in developing his transcendental\ndistinction he does not describe the raindrops as mere representations.\nHis phrase is much more indefinite. They are \u201cmodifications or\nfundamental forms (_Grundlagen_) of our sensuous intuition.\u201d\n\nKant does not here criticise the view of sensibility which underlies\nLocke\u2019s view of appearance. But he does so in A 271 = B 327, completing\nthe parallel and contrast between Leibniz and Locke.\n\n     \u201cLeibniz _intellectualised_ appearances, just as Locke, according\n     to his system of noogony (if I may be allowed these expressions),\n     _sensualised_ all concepts of the understanding, _i.e._ interpreted\n     them as simply empirical or abstracted concepts of reflection.\n     Instead of interpreting understanding and sensibility as two quite\n     different sources of representations, which yet can supply\n     objectively valid judgments of things only in _conjunction_ with\n     each other, each of these great men holds only to one of the two,\n     viewing it as in immediate relation to things in themselves. The\n     other faculty is regarded as serving only to confuse or to order\n     the representations which this selected faculty yields.\u201d[578]\n\n=Proof that the above View of Space and Time is not a mere Hypothesis,\nbut completely certain.=[579]--The proof, which as here recapitulated and\ndeveloped follows the analytic method, has already been considered in\nconnection with A 39 = B 56. It proceeds upon the assumption that space\ncannot be both an _a priori_ form of intuition and also independently\nreal. The argument as a whole lacks clearness owing to Kant\u2019s failure to\ndistinguish between the problems of pure and applied geometry, between\npure intuition and form of intuition. This is especially obvious in the\nvery unfortunate and misleading second application of the triangle\nillustration.[580] Kant\u2019s tendency to conceive mathematical science\nalmost exclusively in terms of geometry is likewise illustrated.\n\n     \u201cThere is in regard to both [space and time] a large number of _a\n     priori_ apodictic and synthetic propositions. This is especially\n     true of space, which for this reason will be our chief illustration\n     in this enquiry.\u201d[581]\n\n=II. Paragraphs added in the Second Edition.=[582]--Kant proceeds to offer\nfurther proof of the ideality of the appearances (_a_) of outer and\n(_b_) of inner sense. Such proof he finds in the fact that these\nappearances consist solely of relations. (_a_) Outer appearances reduce\nwithout remainder to relations of position in intuition (_i.e._ of\nextension), of change of position (motion), and to the laws which\nexpress in merely relational terms the motive forces by which such\nchange is determined. What it is that is thus present in space, or what\nthe dynamic agencies may be to which the motion is due, is never\nrevealed. But a real existent (_Sache an sich_) can never be known\nthrough mere relations. Outer sense consequently reveals through its\nrepresentations only the relation of an object to the subject, not the\nintrinsic inner nature of the object in itself (_Object an sich_).\nKant\u2019s avoidance of the term _Ding an sich_ may be noted.[583]\n\n(_b_) The same holds true of inner sense, not only because the\nrepresentations of outer sense constitute its proper (_eigentlichen_)\nmaterial, but also because time, in which these are set, contains only\nrelations of succession, coexistence, and duration. This time (which as\nconsisting only of relations can be nothing but a form[584]) is itself,\nin turn, a mere relation. It is only the manner in which through its own\nactivity the mind is affected by itself. But in order to be affected by\nitself it must have receptivity, in other words, sensibility. Time,\nconsequently, must be regarded as the form of this inner sense.\n\nThat everything represented in time, like that which is represented in\nspace, consists solely of relations, Kant does not, however, attempt to\nprove. He is satisfied with repeating the conclusion reached in the\nfirst edition of the _Aesthetic_, that, as time is the object of a\nsense, it must of necessity be appearance. This, like everything which\nKant wrote upon inner sense, is profoundly unsatisfactory. The\nobscurities of his argument are not to be excused on the ground that\n\u201cthe difficulty, how a subject can have an internal intuition of itself,\nis common to every theory.\u201d For no great thinker,[585] except Locke, has\nattempted to interpret inner consciousness on the analogy of the senses.\nDiscussion of the doctrine must meantime be deferred.[586]\n\n=III. B 69.=--Kant here formulates the important distinction between\nappearance (_Erscheinung_) and illusion (_Schein_). The main text is\nclear so far as it goes; but the appended note is thoroughly confused.\nTogether they contain no less than three distinct and conflicting views\nof illusion.[587] According to the main text, _Schein_ signifies a\nrepresentation, such as may occur in a dream, to which nothing real\ncorresponds. _Erscheinung_, on the other hand, is always the appearance\nof a _given_ object; but since the qualities of that object depend\nsolely on our mode of intuition, we have to distinguish the object as\nappearance from the object as thing in itself.\n\n     \u201d[Every appearance] has two sides, the one by which the object is\n     viewed in and by itself, ... the other by which the form of the\n     intuition of the object is taken into account....\u201d[588]\n\nObviously, when illusion is defined in the above manner, the assertion\nthat objects in space are mere appearances cannot be taken as meaning\nthat they are illusory.\n\nBut this view of illusion is peculiar to the passage before us and to A\n38 = B 55. It occurs nowhere else, either in the _Critique_ or in the\n_Prolegomena_; and it is not, as Kant has himself admitted,[589] really\nrelevant to the purposes of the _Critique_. The issues are more\nadequately faced in the appended note, which, however, at the same time,\nshows very clearly that Kant has not yet properly disentangled their\nvarious strands. The above definition of appearance is too wide. It\ncovers illusory sense perception as well as appearance proper. The\nfurther qualification must be added, that the predicates of appearance\nare constant and are inseparable from its representation. Thus the space\npredicates can be asserted of any external object. Redness and scent can\nbe ascribed to the rose. All of these are genuine appearances. If, on\nthe other hand, the two handles, as observed by Galileo, are attributed\nto Saturn, roundness to a distant square tower, bentness to a straight\nstick inserted in water, the result is mere illusion. The predicates, in\nsuch cases, do not stand the test of further observation or of the\nemployment of other senses. Only in a certain position of its rings,\nrelatively to the observer, does Saturn _seem_ (_scheint_) to have two\nhandles. The distant tower only _seems_ to be round. The stick only\n_seems_ to be bent. But the rose _is_ extended and _is_ red. Obviously\nKant is no longer viewing _Schein_ as equivalent to a merely mental\nimage. It now receives a second meaning. It is illusion in the modern,\npsychological sense. It signifies an abnormal perception of an actually\npresent object. The distinction between appearance and illusion is now\nreduced to a merely relative difference in constancy and universality of\nappearance. Saturn necessarily appears to Galileo as possessing two\nhandles. A square tower viewed from the distance cannot appear to the\nhuman eye otherwise than round. A stick inserted in water must appear\nbent. If, however, Saturn be viewed under more favourable conditions, if\nthe distance from the tower be diminished, if the stick be removed from\nthe water, the empirical object will appear in a manner more in harmony\nwith the possible or actual experiences of touch. The distinction is\npractical, rather than theoretical, in its justification. It says only\nthat certain sets of conditions may be expected to remain uniform;\nthose, for instance, physical, physiological, and psychical, which cause\na rose to appear red. Other sets of conditions, such as those which\ncause the stick to appear bent, are exceptional, and for that reason the\nbentness may be discounted as illusion. Among the relatively constant\nare the space and time properties of bodies. To employ the terms of the\nmain text, it is not only by illusion that bodies seem to exist outside\nme; they actually are there.\n\nSo long as we keep to the sphere of ordinary experience, and require no\ngreater exactitude than practical life demands, this distinction is, of\ncourse, both important and valid. But Kant, by his references to Saturn,\nraises considerations which, if faced, must complicate the problem and\nplace it upon an entirely different plane. If, in view of scientific\nrequirements, the conditions of observation are more rigorously\nformulated, and if by artificial instruments of scientific precision we\nmodify the perceptions of our human senses, what before was ranked as\nappearance becomes illusion; and no limit can be set to the\ntransformations which even our most normal human experiences may thus be\nmade to undergo. Even the most constant perceptions then yield to\nvariation. The most that can be asserted is that throughout all change\nin the conditions of observation objects still continue to possess, in\nhowever new and revolutionary a fashion, some kind of space and time\npredicates. The application of this more rigorous scientific standard of\nappearance thus leads to a fourfold distinction between ultimate\nreality, scientific appearances, the appearances of ordinary\nconsciousness, and the illusions of ordinary consciousness. The\nappearances of practical life are the illusions of science, and the\nappearances of science would similarly be illusions to any being who\nthrough \u2018intuitive understanding\u2019 could apprehend things in themselves.\n\nBut if the distinction between appearance and illusion is thus merely\nrelative to the varying nature of the conditions under which observation\ntakes place, it can afford no sufficient answer to the criticisms which\nKant is here professing to meet. Kant has in view those critics (such as\nLambert, Mendelssohn, and Garve) who had objected that if bodies in\nspace are representations existing, as he so often asserts, only \u201cwithin\nus,\u201d their appearing to exist \u201coutside us\u201d is a complete illusion. These\ncritics have, indeed, found a vulnerable point in Kant\u2019s teaching. The\nonly way in which he can effectively meet it is by frank recognition and\ndevelopment of the phenomenalism with which his subjectivism comes into\nso frequent conflict.[590] That certain perceptions are more constant\nthan others does not prove that all alike may not be classed as\nillusory. The criticism concerns only the reality of extended objects.\nFrom Kant\u2019s own extreme subjectivist position they _are_ illusions of\nthe most thoroughgoing kind. If, as Kant so frequently maintains,\nobjects are representations and exist only \u201cwithin us,\u201d their existence\n\u201coutside us\u201d must be denied. The criticism can be met only if Kant is\nprepared consistently to formulate and defend his own alternative\nteaching, that sensations arise through the action of external objects\nupon the sense-organs, and that the world of physical science has\nconsequently a reality not reducible to mere representations in the\nindividual mind.\n\nIt may be objected that Kant has in the main text cited one essential\ndifference between his position and that which is being ascribed to him.\nExtended objects, though mere representations, are yet due to, and\nconditioned by, things in themselves. They are illusory only in regard\nto their properties, not in regard to their existence. But this\ndistinction is not really relevant. The criticism, as just stated, is\ndirected only against Kant\u2019s view of space. The fact that the spatial\nworld is a grounded and necessary illusion is not strictly relevant to\nthe matter in dispute. Kant has, indeed, elsewhere, himself admitted the\njustice of the criticism. In A 780 = B 808 he cites as a possible\nhypothesis, entirely in harmony with his main results, though not in any\ndegree established by them, the view\n\n     \u201cthat this life is an appearance only, that is, a sensuous\n     representation of purely spiritual life, and that the whole\n     sensible world is a mere image (_ein blosses Bild_) which hovers\n     before our present mode of knowledge, and like a dream has in\n     itself no objective reality.\u201d\n\nKant\u2019s reply is thus really only verbal. He claims that illusion, if\nconstant, has earned the right to be called appearance. He accepts the\ncriticism, but restates it in his own terms. The underlying\nphenomenalism which colours the position in his own thoughts, and for\nwhich he has not been able to find any quite satisfactory formulation,\nis the sole possible justification, if any such exists, for his\ncontention that the criticism does not apply. Such phenomenalism crops\nout in the sentence, already partially quoted:\n\n     \u201cIf I assert that the quality of space and time, according to\n     which, as a condition of their existence, I posit both external\n     objects and my own soul, lies in my mode of intuition and not in\n     these objects in themselves, I am not saying that only by illusion\n     do bodies seem to exist outside me or my soul to be given in my\n     self-consciousness.\u201d[591]\n\nBut, so far, I have simplified Kant\u2019s argument by leaving out of account\na third and entirely different view of illusion which is likewise\nformulated in the appended note. In the middle of the second sentence,\nand in the last sentence, illusion is defined as the attribution to the\nthing in itself of what belongs to it only in its relation to the\nsenses. Illusion lies not in the object apprehended, but only in the\n_judgment_ which we pass upon it. It is due, not to sense, but to\nunderstanding.[592] Viewing illusion in this way, Kant is enabled to\nmaintain that his critics are guilty of \u201can unpardonable and almost\nintentional misconception,\u201d[593] since this is the very fallacy which he\nhimself has been most concerned to attack. As he has constantly\ninsisted, appearance is appearance just because it can never be a\nrevelation of the thing in itself.\n\nNow the introduction of this third view reduces the argument of the\nappended note to complete confusion. Its first occurrence as a\nparenthesis in a sentence which is stating an opposed view would seem to\nindicate that the note has been carelessly recast. Originally containing\nonly a statement of the second view, Kant has connected therewith the\nview which he had already formulated in the first edition and in the\n_Prolegomena_. But the two views cannot be combined. By the former\ndefinition, illusion is necessitated but abnormal perception; according\nto the latter, it is a preventable error of our conscious judgment. The\nopposite of illusion is in the one case _appearance_, in the other\n_truth_. The retention of the reference to Saturn, in the statement of\nthe third view at the end of the note, is further evidence of hasty\nrecasting. While the rose and the extended objects are there treated as\nalso things in themselves, Saturn is taken only in its phenomenal\nexistence. In view of the general confusion, it is a minor inconsistency\nthat Kant should here maintain, in direct opposition to A 28-9, that\nsecondary qualities can be attributed to the empirical object.\n\nThis passage from the second edition is a development of _Prolegomena_,\n\u00a7 13, iii. Kant there employs the term appearance in a quite indefinite\nmanner. For the most part he seems to mean by it any and every\nsense-experience, whether normal or abnormal, and even to include under\nit dream images. But it is also employed in the second of the above\nmeanings, as signifying those sense-perceptions which harmonise with\ngeneral experience. Illusion is throughout employed in the third of the\nabove meanings. Kant\u2019s illustration, that of the apparently retrograde\nmovements of the planets, necessitates a distinction between apparent\nand real motion in space, and consequently leads to the fruitful\ndistinction noted above. Kant gives, however, no sign that he is\nconscious of the complicated problems involved.\n\nIn the interval between the _Prolegomena_ (1783) and the second edition\nof the _Critique_ (1787) Mendelssohn had published (1785) his\n_Morgenstunden_. In its introduction, entitled _Vorerkenntniss von\nWahrheit, Schein und Irrthum_,[594] he very carefully distinguishes\nbetween illusion (_Sinnenschein_) and error of judgment (_Irrthum_).\nThis introduction Kant had read. In a letter to Sch\u00fctz[595] he cites it\nby title, and praises it as \u201cacute, original, and of exemplary\nclearness.\u201d It is therefore the more inexcusable that he should again in\nthe second edition of the _Critique_ have confused these two so\nradically different meanings of the term _Schein_. Mendelssohn, however,\ndrew no distinction between _Schein_ and _Erscheinung_. They were then\nused as practically synonymous,[596] though of course _Schein_ was the\nstronger term. Kant seems to have been the first to distinguish them\nsharply and to attempt to define the one in opposition to the other. But\nthe very fact that _Erscheinung_ and _Schein_ were currently employed as\nequivalent terms, and that the distinction, though one of his own\ndrawing, had been mentioned only in the most cursory manner in the first\nedition of the _Critique_,[597] removes all justification for his retort\nupon his critics of \u201cunpardonable misconception.\u201d His anger was really\ndue, not to the objection in itself, but to the implied comparison of\nhis position to that of Berkeley. Such comparison never failed to arouse\nKant\u2019s wrath. For however much this accusation might be justified by his\nown frequent lapses into subjectivism of the most extreme type, even its\npartial truth was more than he was willing to admit. Berkeley represents\nin his eyes, not merely a subjectivist interpretation of the outer\nworld, but the almost diametrical opposite of everything for which he\nhimself stood. Discussion of Kant\u2019s relation to Berkeley had best,\nhowever, be introduced through consideration of the passage immediately\nfollowing in which Kant refers to Berkeley by name.\n\n=III. (Second Part) B 70.=--Kant urges that his doctrine of the ideality\nof space and time, so far from reducing objects to mere illusion, is the\nsole means of defending their genuine reality. If space and time had an\nindependent existence, they would have to be regarded as more real than\nthe bodies which occupy them. For on this view space and time would\ncontinue to exist even if all their contents were removed; they would be\nantecedent necessary conditions of all other existences. But space and\ntime thus interpreted are impossible conceptions.[598] The reality of\nbodies is thereby made to depend upon _Undinge_. If this were the sole\nalternative, \u201cthe good Bishop Berkeley [could] not be blamed for\ndegrading bodies to mere illusion.\u201d We should, Kant maintains, have to\nproceed still further, denying even our own existence. For had Berkeley\ntaken account of time as well as of space, a similar argument,\nconsistently developed in regard to time, would have constrained him to\nreduce the self to the level of mere illusion. Belief in the reality of\nthings in themselves, whether spiritual or material, is defensible only\nif space and time be viewed as subjective. In other words, Berkeley\u2019s\nidealism is an inevitable consequence of a realist view of space. But it\nis also its _reductio ad absurdum_.\n\n     [\u201cBerkeley in his dogmatic idealism] maintains that space, with all\n     the things of which it is the inseparable condition, is something\n     impossible in itself, and he therefore regards the things in space\n     as merely imaginary entities (_Einbildungen_). Dogmatic idealism is\n     inevitable if space be interpreted as a property which belongs to\n     things in themselves. For, when so regarded, space, and everything\n     to which it serves as condition, is a non-entity (_Unding_). The\n     ground upon which this idealism rests we have removed in the\n     _Transcendental Aesthetic_.\u201d[599]\n\nThe term _Schein_ is not employed throughout this passage in either of\nthe two meanings of the appended note, but in that of the main text. It\nsignifies a representation, to which no existence corresponds.\n\n\nKANT\u2019S RELATION TO BERKELEY\n\nBy idealism[600] Kant means any and every system which maintains that\nthe sensible world does not exist in the form in which it presents\nitself to us. The position is typified in Kant\u2019s mind by the Eleatics,\nby Plato, and by Descartes, all of whom are rationalists. With the\ndenial of reality to sense-appearances they combine a belief in the\npossibility of rationally comprehending its supersensible basis. Failing\nto appreciate the true nature of the sensible, they misunderstand the\ncharacter of geometrical science, and falsely ascribe to pure\nunderstanding a power of intellectual intuition. Kant\u2019s criticisms of\nBerkeley show very clearly that it is this more general position which\nhe has chiefly in view. To Berkeley Kant objects that only in\nsense-experience is there truth, that it is sensibility, not\nunderstanding, which possesses the power of _a priori_ intuition, and\nthat through pure understanding, acting in independence of sensibility,\nno knowledge of any kind can be acquired. In other words, Kant classes\nBerkeley with the rationalists. And, as we have already seen, he even\ngoes the length of regarding Berkeley\u2019s position as the _reductio ad\nabsurdum_ of the realist view of space. Kant does, indeed,\nrecognise[601] that Berkeley differs from the other idealists, in\nholding an empirical view of space, and consequently of geometry, but\nthis does not prevent Kant from maintaining that Berkeley\u2019s thinking is\ninfluenced by certain fundamental implications of the realist position.\nBerkeley\u2019s insight--such would seem to be Kant\u2019s line of argument--is\nperverted by the very view which he is attacking. Berkeley appreciates\nonly what is false in the Cartesian view of space; he is blind to the\nimportant element of truth which it contains. Empiricist though he be,\nhe has no wider conception of the function and powers of sensibility\nthan have the realists from whom he separates himself off; and in order\nto comprehend those existences to which alone he is willing to allow\ntrue reality, he has therefore, like the rationalists, to fall back upon\npure reason.[602]\n\nThat Kant\u2019s criticism of Berkeley should be extremely external is not,\ntherefore, surprising. He is interested in Berkeley\u2019s positive teaching\nonly in so far as it enables him to illustrate the evil tendencies of a\nmistaken idealism, which starts from a false view of the functions of\nsensibility and of understanding, and of the nature of space and time.\nThe key to the true idealism lies, he claims, in the Critical problem,\nhow _a priori_ synthetic judgments can be possible. This is the\nfundamental problem of metaphysics, and until it has been formulated and\nanswered no advance can be made.\n\n     \u201cMy so-called (Critical) idealism is thus quite peculiar in that it\n     overthrows ordinary idealism, and that through it alone _a priori_\n     cognition, even that of geometry, attains objective reality, a\n     thing which even the keenest realist could not assert till I had\n     proved the ideality of space and time.\u201d[603]\n\nIn order to make Kant\u2019s account of Berkeley\u2019s teaching really\ncomprehensible, we seem compelled to assume that he had never himself\nactually read any of Berkeley\u2019s own writings. Kant\u2019s acquaintance with\nthe English language was most imperfect, and we have no evidence that he\nhad ever read a single English book.[604] When he quotes Pope and\nAddison, he does so from German translations.[605] Subsequent to 1781 he\ncould, indeed, have had access to Berkeley\u2019s _Dialogues between Hylas\nand Philonous_[606] in a German translation; but in view of the account\nwhich he continues to give of Berkeley\u2019s teaching, it does not seem\nlikely[607] that he had availed himself of this opportunity. As to what\nthe indirect sources of Kant\u2019s knowledge of Berkeley may have been, we\ncannot decide with any certainty, but amongst them must undoubtedly be\nreckoned Hume\u2019s statements in regard to Berkeley in the _Enquiry_,[608]\nand very probably also the references to Berkeley in Beattie\u2019s _Nature\nof Truth_.[609] From the former Kant would learn of Berkeley\u2019s\nempirical view of space and also of the sceptical tendencies of his\nidealist teaching. From it he might also very naturally infer that\nBerkeley denies all reality to objects. By Beattie Kant would be\nconfirmed in this latter view, and also in his contention that Berkeley\nis unable to supply a criterion for distinguishing between reality and\ndreams. Kant may also have received some impressions regarding Berkeley\nfrom Hamann.\n\nTo take Kant\u2019s criticisms of Berkeley more in detail. In the first\nedition of the _Critique_[610] Kant passes two criticisms, without,\nhowever, mentioning Berkeley by name: first, that he overlooks the\nproblem of time, and, like Descartes, ascribes complete reality to the\nobjects of inner sense. This is the cause of a second error, namely,\nthat he views the objects of outer sense as mere illusion (_blosser\nSchein_). Proceeding, Kant argues that inner and outer sense are really\nin the same position. Though they yield only appearances, these\nappearances are conditioned by things in themselves. Through this\nrelation to things in themselves they are distinguished from all merely\nsubjective images. Berkeley is again referred to in the fourth\n_Paralogism_.[611] His idealism is distinguished from that of Descartes.\nThe one is dogmatic; the other is sceptical. The one denies the\nexistence of matter; the other only doubts whether it is possible to\nprove it. Berkeley claims, indeed, that there are contradictions in the\nvery conception of matter; and Kant remarks that this is an objection\nwhich he will have to deal with in the section on the _Antinomies_. But\nthis promise Kant does not fulfil; and doubtless for the reason that,\nhowever unwilling he may be to make the admission, on this point his own\nteaching, especially in the _Dialectic_, frequently coincides with that\nof Berkeley. So little, indeed, is Kant concerned in the first edition\nto defend his position against the accusation of subjectivism, that in\nthis same section he praises the sceptical idealist as a \u201cbenefactor of\nhuman reason.\u201d\n\n     \u201cHe compels us, even in the smallest advances of ordinary\n     experience, to keep on the watch, lest we consider as a well-earned\n     possession what we perhaps obtain only in an illegitimate manner.\n     We are now in a position to appreciate the value of the objections\n     of the idealist. They drive us by main force, unless we mean to\n     contradict ourselves in our commonest assertions, to view all our\n     perceptions, whether we call them inner or outer, as a\n     consciousness only of what is dependent on our sensibility. They\n     also compel us to regard the outer objects of these perceptions not\n     as things in themselves, but only as representations, of which, as\n     of every other representation, we can become immediately conscious,\n     and which are entitled outer because they depend on what we call\n     \u2018outer sense\u2019 whose intuition is space. Space itself, however, is\n     nothing but an inner mode of representation in which certain\n     perceptions are connected with one another.\u201d[612]\n\nThese criticisms are restated in A 491-2 = B 519-20, with the further\naddition that in denying the existence of extended beings \u201cthe empirical\nidealist\u201d removes the possibility of distinguishing between reality and\ndreams. This is a new criticism. Kant is no longer referring to the\ndenial of unknowable things in themselves. He is now maintaining that\nonly the Critical standpoint can supply an immanent criterion whereby\nreal experiences may be distinguished from merely subjective happenings.\nThis point is further insisted upon in the _Prolegomena_,[613] but is\nnowhere developed with any direct reference to Berkeley\u2019s own personal\nteaching. Kant assumes as established that any such criterion must rest\nupon the _a priori_; and in this connection Berkeley is conveniently\nmade to figure as a thoroughgoing empiricist.\n\nThe _Critique_, on its publication, was at once attacked, especially in\nthe Garve-Feder review, as presenting an idealism similar to that of\nBerkeley. As Erdmann has shown, the original plan of the _Prolegomena_\nwas largely modified in order to afford opportunity for reply to this\n\u201cunpardonable and almost intentional misconception.\u201d[614] Kant\u2019s\nreferences to Berkeley, direct and indirect, now for the first time\nmanifest a polemical tone, exaggerating in every possible way the\ndifference between their points of view. Only the transcendental\nphilosophy can establish the possibility of _a priori_ knowledge, and so\nit alone can afford a criterion for distinguishing between realities and\ndreams. It alone will account for the possibility of geometrical\nscience; Berkeley\u2019s idealism would render the claims of that science\nwholly illusory. The Critical idealism transcends experience only so far\nas is required to discover the conditions which make empirical cognition\npossible; Berkeley\u2019s idealism is \u2018visionary\u2019 and \u2018mystical.\u2019[615] Even\nsceptical idealism now comes in for severe handling. It may be called\n\u201cdreaming idealism\u201d; it makes things out of mere representations, and\nlike idealism in its dogmatic form it virtually denies the existence of\nthe only true reality, that of things in themselves. Sceptical idealism\nmisinterprets space by making it empirical, dogmatic idealism by\nregarding it as an attribute of the real. Both entirely ignore the\nproblem of time. For these reasons they underestimate the powers of\nsensibility (to which space and time belong as _a priori_ forms), and\nexaggerate those of pure understanding.\n\n     \u201cThe position of all genuine idealists from the Eleatics to\n     Berkeley is contained in this formula: \u2018All cognition through the\n     senses and experience is nothing but mere illusion, and only in the\n     ideas of pure understanding and Reason is there truth.\u2019 The\n     fundamental principle ruling all my idealism, on the contrary, is\n     this: \u2018All cognition of things solely from pure understanding or\n     pure Reason is nothing but mere illusion and only in experience is\n     there truth.\u2019\u201d[616]\n\nThis is an extremely inadequate statement of the Critical standpoint,\nbut it excellently illustrates Kant\u2019s perverse interpretation of\nBerkeley\u2019s teaching.\n\nTo these criticisms Kant gives less heated but none the less explicit\nexpression in the second edition of the _Critique_. He is now much more\ncareful to avoid subjectivist modes of statement. His phenomenalist\ntendencies are reinforced, and come to clearer expression of all that\nthey involve. The fourth _Paralogism_ with its sympathetic treatment of\nempirical idealism is omitted, and in addition to the above passage Kant\ninserts a new section, entitled _Refutation of Idealism_, in which he\nstates his position in a much more adequate manner.\n\n=IV. B 71.=--Kant continues the argument of A 39.[617] If space and time\ncondition all existence, they will condition even divine existence, and\nso must render God\u2019s omniscience, which as such must be intuitive, not\ndiscursive, difficult of conception. Upon this point Kant is more\nexplicit in the _Dissertation_.[618]\n\n     \u201c_Whatever is, is somewhere and sometime_, is a spurious axiom....\n     By this spurious principle all beings, even though they be known\n     intellectually, are restricted in their existence by conditions of\n     space and time. Philosophers therefore discuss every form of idle\n     question regarding the locations in the corporeal universe of\n     substances that are immaterial--and of which for that very reason\n     there can be no sensuous intuition nor any possible spatial\n     representation--or regarding the seat of the soul, and the like.\n     And since the sensuous mixes with the intellectual about as badly\n     as square with round, it frequently happens that the one disputant\n     appears as holding a sieve into which the other milks the he-goat.\n     The presence of immaterial things in the corporeal world is\n     virtual, not local, although it may conveniently be spoken of as\n     local. Space contains the conditions of possible interaction only\n     when it is between material bodies. What, however, in immaterial\n     substances constitutes the external relations of force between them\n     or between them and bodies, obviously eludes the human\n     intellect.... But when men reach the conception of a highest and\n     extra-mundane Being, words cannot describe the extent to which they\n     are deluded by these shades that flit before the mind. They picture\n     God as present in a place: they entangle Him in the world where He\n     is supposed to fill all space at once. They hope to make up for the\n     [spatial] limitation they thus impose by thinking of God\u2019s place\n     _per eminentiam_, _i.e._ as infinite. But to be present in\n     different places at the same time is absolutely impossible, since\n     different places are mutually external to one another, and\n     consequently what is in several places is outside itself, and is\n     therefore present to itself outside itself--which is a\n     contradiction in terms. As to time, men have got into an\n     inextricable maze by releasing it from the laws that govern sense\n     knowledge, and what is more, transporting it beyond the confines of\n     the world to the Being that dwells there, as a condition of His\n     very existence. They thus torment their souls with absurd\n     questions, for instance, why God did not fashion the world many\n     centuries earlier. They persuade themselves that it is easily\n     possible to conceive how God may discern present things, _i.e._\n     what is actual in the time in which He is. But they consider that\n     it is difficult to comprehend how He should foresee the things\n     about to be, _i.e._ the actual in the time in which He is not yet.\n     They proceed as if the existence of the Necessary Being descended\n     successively through all the moments of a supposed time, and having\n     already exhausted part of His duration, foresaw the eternal life\n     that still lies before Him together with the events which [will]\n     occur simultaneously [with that future life of His]. All these\n     speculations vanish like smoke when the notion of time has been\n     rightly discerned.\u201d\n\nThe references in B 71-2 to the intuitive understanding are among the\nmany signs of Kant\u2019s increased preoccupation, during the preparation of\nthe second edition, with the problems which it raises. Such\nunderstanding is not sensuous, but intellectual; it is not derivative,\nbut original; the object itself is created in the act of intuition. Or,\nas Kant\u2019s position may perhaps be more adequately expressed, all of\nGod\u2019s activities are creative, and are inseparable from the non-sensuous\nintuition whereby both they and their products are apprehended by Him.\nKant\u2019s reason for again raising this point may be Mendelssohn\u2019s\ntheological defence of the reality of space in his\n_Morgenstunden_.[619] Mendelssohn has there argued that just as\nknowledge of independent reality is confirmed by the agreement of\ndifferent senses, and is rendered the more certain in proportion to the\nnumber of senses which support the belief, so the validity of our\nspatial perceptions is confirmed in proportion as men are found to agree\nin this type of experience with one another, with the animals, and with\nangelic beings. Such inductive inference will culminate in the proof\nthat even the Supreme Being apprehends things in this same spatial\nmanner.[620] Kant\u2019s reply is that however general the intuition of space\nmay be among finite beings, it is sensuous and derivative, and therefore\nmust not be predicated of a Divine Being. For obvious reasons Kant has\nnot felt called upon to point out the inadequacy of this inductive\nmethod to the solution of Critical problems. In A 42 Kant, arguing that\nour forms of intuition are subjective, claims that they do not\nnecessarily belong to all beings, though they must belong to all\nmen.[621] He is quite consistent in now maintaining[622] that their\ncharacteristics, as sensuous and derivative, do not necessarily preclude\ntheir being the common possession of all finite beings.\n\n\nTHE PARADOX OF INCONGRUOUS COUNTERPARTS\n\nThe purpose, as already noted, of the above sections II. to IV., as\nadded in the second edition, is to afford \u2018confirmation\u2019 of the ideality\nof space and time. That being so, it is noticeable that Kant has omitted\nall reference to an argument embodied, for this same purpose, in \u00a7 13 of\nthe _Prolegomena_. The matter is of sufficient importance to call for\ndetailed consideration.[623]\n\nAs the argument of the _Prolegomena_ is somewhat complicated, it is\nadvisable to approach it in the light of its history in Kant\u2019s earlier\nwritings. It was to his teacher Martin Knutzen that Kant owed his first\nintroduction to Newton\u2019s cosmology; and from Knutzen he inherited the\nproblem of reconciling Newton\u2019s mechanical view of nature and absolute\nview of space with the orthodox Leibnizian tenets. In his first\npublished work[624] Kant seeks to prove that the very existence of\nspace is due to gravitational force, and that its three-dimensional\ncharacter is a consequence of the specific manner in which gravity acts.\nSubstances, he teaches, are unextended. Space results from the\nconnection and order established between them by the balancing of their\nattractive and repulsive forces. And as the law of gravity is merely\ncontingent, other modes of interaction, and therefore other forms of\nspace, with more than three dimensions, must be recognised as possible.\n\n     \u201cA science of all these possible kinds of space would undoubtedly\n     be the highest enterprise which a finite understanding could\n     undertake in the field of geometry.\u201d[625]\n\nIn the long interval between 1747 and 1768 Kant continued to hold to\nsome such compromise, retaining Leibniz\u2019s view that space is derivative\nand relative, and rejecting Newton\u2019s view that it is prior to, and\npre-conditions, all the bodies that exist in it. But in that latter year\nhe published a pamphlet[626] in which, following in the steps of the\nmathematician, Euler,[627] he drew attention to certain facts which\nwould seem quite conclusively to favour the Newtonian as against the\nLeibnizian interpretation of space. The three dimensions of space are\nprimarily distinguishable by us only through the relation in which they\nstand to our body. By relation to the plane that is at right angles to\nour body we distinguish \u2018above\u2019 and \u2018below\u2019; and similarly through the\nother two planes we determine what is \u2018right\u2019 and \u2018left,\u2019 \u2018in front\u2019 and\n\u2018behind.\u2019 Through these distinctions we are enabled to define\ndifferences which cannot be expressed in any other manner. All species\nof hops--so Kant maintains--wind themselves around their supports from\nleft to right, whereas all species of beans take the opposite direction.\nAll snail shells, with some three exceptions, turn, in descending from\ntheir apex downwards, from left to right. This determinate direction of\nmovement, natural to each species, like the difference in spatial\nconfiguration between a right and a left hand, or between a right hand\nand its reflection in a mirror, involves in all cases a reference of the\ngiven object to the wider space within which it falls, and ultimately to\nspace as a whole. Only so can its determinate character be distinguished\nfrom its opposite counterpart. For as Kant points out, though the right\nand the left hand are _counterparts_, that is to say, objects which have\na common definition so long as the arrangement of the parts of each is\ndetermined in respect to its central line of reference, they are none\nthe less inwardly _incongruent_, since the one can never be made to\noccupy the space of the other. As he adds in the _Prolegomena_, the\nglove of one hand cannot be used for the other hand. This inner\nincongruence compels us to distinguish them as different, and this\ndifference is only determinable by location of each in a single absolute\nspace that constrains everything within it to conform to the conditions\nwhich it prescribes. In three-dimensional space everything must have a\nright and a left side, and must therefore exhibit such inner differences\nas those just noted. Spatial determinations are not, as Leibniz teaches,\nsubsequent to, and dependent upon, the relations of bodies to one\nanother; it is the former that determine the latter.\n\n     \u201cThe reason why that which in the shape of a body exclusively\n     concerns its relation to pure space can be apprehended by us only\n     through its relation to other bodies, is that absolute space is not\n     an object of any outer sensation, but a fundamental conception\n     which makes all such differences possible.\u201d[628]\n\nKant enforces his point by arguing that if the first portion of creation\nwere a human hand, it would have to be either a right or a left hand.\nAlso, a different act of creation would be demanded according as it was\nthe one or the other. But if the hand alone existed, and there were no\npre-existing space, there would be no inward difference in the relations\nof its parts, and nothing outside it to differentiate it. It would\ntherefore be entirely indeterminate in nature, _i.e._ would suit either\nside of the body, which is impossible.\n\nThis adoption of the Newtonian view of space in 1768 was an important\nstep forward in the development of Kant\u2019s teaching, but could not, in\nview of the many metaphysical difficulties to which it leads, be\npermanently retained; and in the immediately following year--a year\nwhich, as he tells us,[629] \u201cgave great light\u201d--he achieved the final\nsynthesis which enabled him to combine all that he felt to be essential\nin the opposing views. Though space is an absolute and preconditioning\nsource of differences which are not conceptually resolvable, it is a\nmerely subjective form of our sensibility.\n\nNow it is significant that when Kant expounds this view in the\n_Dissertation_ of 1770, the argument from incongruous counterparts is no\nlonger employed to establish the absolute and pre-conditioning\ncharacter of space, but only to prove that it is a pure non-conceptual\nintuition.\n\n     \u201cWhich things in a given space lie towards one side, and which lie\n     towards the other, cannot by any intellectual penetration be\n     discursively described or reduced to intellectual marks. For in\n     solids that are completely similar and equal, but incongruent, such\n     as the right and the left hand (conceived solely in terms of their\n     extension), or spherical triangles from two opposite hemispheres,\n     there is a diversity which renders impossible the coincidence of\n     their spatial boundaries. This holds true, even though they can be\n     substituted for one another in all those respects which can be\n     expressed in marks that are capable of being made intelligible to\n     the mind through speech. It is therefore evident that the\n     diversity, that is, the incongruity, can only be apprehended by\n     some species of pure intuition.\u201d[630]\n\nThere is no mention of this argument in the first edition of the\n_Critique_, and when it reappears in the _Prolegomena_ it is interpreted\nin the light of an additional premiss, and is made to yield a very\ndifferent conclusion from that drawn in the _Dissertation_, and a\n_directly opposite conclusion_ from that drawn in 1768. Instead of being\nemployed to establish either the intuitive character of space or its\nabsolute existence, it is cited as evidence in proof of its\nsubjectivity. As in 1768, it is spoken of as strange and paradoxical,\nand many of the previous illustrations are used. The paradox consists in\nthe fact that bodies and spherical figures, conceptually considered, can\nbe absolutely identical, and yet for intuition remain diverse. This\nparadox, Kant now maintains[631] in opposition to his 1768 argument,\nproves that such bodies and the space within which they fall are not\nindependent existences. For were they things in themselves, they would\nbe adequately cognisable through the pure understanding, and could not\ntherefore conflict with its demands. Being conceptually identical, they\nwould necessarily be congruent in every respect. But if space is merely\nthe form of sensibility, the fact that in space the part is only\npossible through the whole will apply to everything in it, and so will\ngenerate a fundamental difference between conception and intuition.[632]\nThings in themselves are, as such, unconditioned, and cannot, therefore,\nbe dependent upon anything beyond themselves. The objects of intuition,\nin order to be possible, must be merely ideal.\n\nNow the new premiss which differentiates this argument from that of\n1768, and which brings Kant to so opposite a conclusion, is one which is\nentirely out of harmony with the teaching of the _Critique_. In this\nsection of the _Prolegomena_ Kant has unconsciously reverted to the\ndogmatic standpoint of the _Dissertation_, and is interpreting\nunderstanding in the illegitimate manner which he so explicitly\ndenounces in the section on _Amphiboly_.\n\n     \u201cThe mistake ... lies in employing the understanding contrary to\n     its vocation transcendentally [_i.e._ transcendently] and in making\n     objects, _i.e._ possible intuitions, conform to concepts, not\n     concepts to possible intuitions, on which alone their objective\n     validity rests.\u201d[633]\n\nThe question why no mention of this argument is made in the second\nedition of the _Critique_ is therefore answered. Kant had meantime, in\nthe interval between 1783 and 1787,[634] become aware of the\ninconsistency of the position. So far from being a paradox, this assumed\nconflict rests upon a false view of the function of the\nunderstanding.[635] The relevant facts may serve to confirm the view of\nspace as an intuition in which the whole precedes the parts;[636] but\nthey can afford no evidence either of its absoluteness or of its\nideality. In 1768 they seem to Kant to prove its absoluteness, only\nbecause the other alternative has not yet occurred to him. In 1783 they\nseem to him to prove its ideality, only because he has not yet\ncompletely succeeded in emancipating his thinking from the dogmatic\nrationalism of the _Dissertation_.\n\nAs already noted,[637] Kant\u2019s reason for here asserting that space is\nintuitive in nature, namely, that in it the parts are conditioned by the\nwhole, is also his reason for elsewhere describing it as an Idea of\nReason. The further implication of the argument of the _Prolegomena_,\nthat in the noumenal sphere the whole is made possible only by its\nunconditioned parts, raises questions the discussion of which must be\ndeferred. The problem recurs in the _Dialectic_ in connection with\nKant\u2019s definition of the Idea of the unconditioned. In the Ideas of\nReason Kant comes to recognise the existence of concepts which do not\nconform to the reflective type analysed by the traditional logic, and to\nperceive that these Ideas can yield a deeper insight than any possible\nto the discursive understanding. The above rationalistic assumption must\nnot, therefore, pass unchallenged. It may be that in the noumenal sphere\nall partial realities are conditioned by an unconditioned whole.\n\n       *       *       *       *       *\n\n=Concluding Paragraph.=[638]--The wording of this paragraph is in keeping\nwith the increased emphasis which in the _Introduction_ to the second\nedition is given to the problem, how _a priori_ synthetic judgments are\npossible. Kant characteristically fails to distinguish between the\nproblems of pure and applied mathematics, with resulting\ninconsecutiveness in his argumentation.\n\n\n\n\nTHE TRANSCENDENTAL DOCTRINE OF ELEMENTS\n\n\n\n\nPART II\n\nTHE TRANSCENDENTAL LOGIC\n\n\nINTRODUCTION\n\n=I. Concerning Logic in General.=--This _Introduction_[639] which falls\ninto four divisions, is extremely diffuse, and contributes little that\nis of more than merely architectonic value. It is a repetition of the\nlast section of the general _Introduction_, and of the introductory\nparagraphs of the _Aesthetic_, but takes no account of the definitions\ngiven in either of those two places. It does not, therefore, seem likely\nthat it could have been written in immediate sequence upon the\n_Aesthetic_. It is probably later than the main body of the\n_Analytic_.[640] In any case it is externally tacked on to it; as\nAdickes has noted,[641] it is completely ignored in the opening section\nof the _Analytic_.[642]\n\nIn treating of intuition in the first sentence, Kant seems to have in\nview only empirical intuition.[643] Yet he at once proceeds to state\nthat intuition may be pure as well as empirical.[644] Also, in asserting\nthat \u201cpure intuition contains only the form under which something is\nintuited,\u201d Kant would seem to be adopting the view that it does not\nyield its own manifold, a conclusion which he does not, however, himself\ndraw.\n\nIn defining sensibility,[645] Kant again ignores pure intuition.\nSensuous intuition, it is stated, is the mode in which we are affected\nby objects.[646] Understanding, in turn, is defined only in its\nopposition to sensibility, in the ordinary meaning of that term.\nUnderstanding is the faculty which yields thought of the object to which\nsense-affection is due. It is \u201cthe power of thinking the object of\nsensuous intuition\u201d; and acts, it is implied, in and through pure\nconcepts which it supplies out of itself.\n\n     \u201cWithout sensibility objects would not be given to us [_i.e._ the\n     impressions, in themselves merely subjective contents, through\n     which alone independent objects can be revealed to us, would be\n     wanting]; without understanding they would not be thought by us\n     [_i.e._ they would be apprehended only in the form in which they\n     are given, viz. as subjective modes of our sensibility].\u201d\n\nKant has not yet developed the thesis which the central argument of the\n_Analytic_ is directed to prove, namely, that save through the\ncombination of intuition and conception no consciousness whatsoever is\npossible. In these paragraphs he still implies that though concepts\nwithout intuition are empty they are not meaningless, and that though\nintuitions without concepts are blind they are not empty.[647] Their\nunion is necessary for genuine knowledge, but not for the existence of\nconsciousness as such.\n\n     \u201cIt is just as necessary to make our concepts sensuous, _i.e._ to\n     add to them their object in intuition, as to make our intuitions\n     intelligible, _i.e._ to bring them under concepts.\u201d\n\nKant\u2019s final Critical teaching is very different from this. Concepts are\nnot first given in their purity, nor is \u201ctheir object\u201d added in\nintuition. Only through concepts is apprehension of an object possible,\nand only in and through such apprehension do concepts come to\nconsciousness. Nor are intuitions \u201cmade intelligible\u201d by being \u201cbrought\nunder concepts.\u201d Only as thus conceptually interpreted can they exist\nfor consciousness. The co-operation of concept and intuition is\nnecessary for consciousness in any and every form, even the simplest and\nmost indefinite. Consciousness of the subjective is possible only in and\nthrough consciousness of the objective, and _vice versa_. The dualistic\nseparation of sensibility from understanding persists, however, even in\nKant\u2019s later utterances; and, as above stated,[648] to this sharp\nopposition are due both the strength and the weakness of Kant\u2019s\nteaching. Intuition and conception must, he here insists, be carefully\ndistinguished. _Aesthetic_ is the \u201cscience of the rules of sensibility\nin general.\u201d _Logic_ is the \u201cscience of the rules of understanding in\ngeneral.\u201d\n\nKant\u2019s classification of the various kinds of logic[649] may be\nexhibited as follows:\n\n                  { pure\n        { general { applied\n  Logic { special\n        { transcendental\n\nAdickes[650] criticises Kant\u2019s classification as defective, owing to the\nomission of the intermediate concept \u2018ordinary.\u2019 Adickes therefore gives\nthe following table:\n\n                 Logic\n  ===========================================\n  ||               |\n  ||transcendental |       ordinary\n  ||               |_________________________\n  ||               |         |\n  ||               | special |  general\n  ||               |         |_______________\n  ||               |         |      |\n  ||               |         | pure | applied\n  ||               |         |      |\n\nGeneral logic is a logic of elements, _i.e._ of the absolutely necessary\nlaws of thought, in abstraction from all differences in the objects\ndealt with, _i.e._ from all content, whether empirical or\ntranscendental. It is a canon of the understanding in its general\ndiscursive or analytic employment. When it is pure, it takes no account\nof the empirical psychological conditions under which the understanding\nhas to act. When it is developed as an applied logic, it proceeds to\nformulate rules for the employment of understanding under these\nsubjective conditions. It is then neither canon, nor organon, but simply\na catharticon of the ordinary understanding. Special logic is the\norganon of this or that science, _i.e._ of the rules governing correct\nthinking in regard to a certain class of objects. Only pure general\nlogic is a pure doctrine of reason. It alone is absolutely independent\nof sensibility, of everything empirical, and therefore of psychology.\nSuch pure logic is a body of demonstrative teaching, completely _a\npriori_. It stands to applied logic in the same relation as pure to\napplied ethics.\n\n     \u201cSome logicians, indeed, affirm that logic presupposes\n     _psychological_ principles. But it is just as inappropriate to\n     bring principles of this kind into logic as to derive the science\n     of morals from life. If we were to take the principles from\n     psychology, that is, from observations on our understanding, we\n     should merely see _how_ thought takes place, and _how_ it is\n     affected by the manifold subjective hindrances and conditions; so\n     that this would lead only to the knowledge of _contingent_ laws.\n     But in logic the question is not of _contingent_, but of\n     _necessary_ laws; not how we do think, but how we ought to think.\n     The rules of logic, then, must not be derived from the\n     _contingent_, but from the _necessary_ use of the understanding\n     which without any psychology a man finds in himself. In logic we do\n     not want to know how the understanding is and thinks, and how it\n     has hitherto proceeded in thinking, but how it ought to proceed in\n     thinking. Its business is to teach us the correct use of reason,\n     that is, the use which is consistent with itself.\u201d[651]\n\nBy a canon Kant means a system of _a priori_ principles for the correct\nemployment of a certain faculty of knowledge.[652] By an organon Kant\nmeans instruction as to how knowledge may be extended, how new knowledge\nmay be acquired. A canon formulates positive principles through the\napplication of which a faculty can be directed and disciplined. A canon\nis therefore a discipline based on positive principles of correct use.\nThe term discipline is, however, reserved by Kant[653] to signify a\npurely negative teaching, which seeks only to prevent error and to check\nthe tendency to deviate from rules. When a faculty has no correct use\n(as, for instance, pure speculative reason), it is subject only to a\ndiscipline, not to a canon. A discipline is thus \u201ca separate, negative\ncode,\u201d \u201ca system of caution and self-examination.\u201d It is further\ndistinguished from a canon by its taking account of other than purely _a\npriori_ conditions. It is related to a pure canon much as applied is\nrelated to general logic. As a canon supplies principles for the\n_directing_ of a faculty, its distinction from an organon obviously\ncannot be made hard and fast. But here as elsewhere Kant, though\nrigorous and almost pedantic in the drawing of distinctions, is\ncorrespondingly careless in their application. He describes special\nlogic as the organon of this or that science.[654] We should expect from\nthe definition given in the preceding sentence that it would rather be\nviewed as a canon. In A 46 = B 63 Kant speaks of the _Aesthetic_ as an\norganon.\n\n=II. Concerning Transcendental Logic.=--It is with the distinction between\ngeneral and transcendental logic that Kant is chiefly concerned. It is a\ndistinction which he has himself invented, and which is of fundamental\nimportance for the purposes of the _Critique_. Transcendental logic is\nthe new science which he seeks to expound in this second main division\nof the _Doctrine of Elements_. The distinction, from which all the\ndifferences between the two sciences follow, is that while general logic\nabstracts from all differences in the objects known, transcendental\nlogic abstracts only from empirical content. On the supposition, not yet\nproved by Kant, but asserted in anticipation, that there exist pure _a\npriori_ concepts which are valid of objects, there will exist a science\ndistinct in nature and different in purpose from general logic. The two\nlogics will agree in being _a priori_, but otherwise they will differ in\nall essential respects.\n\nThe reference in A 55 = B 79 to the forms of intuition is somewhat\nambiguous. Kant might be taken as meaning that in transcendental logic\nabstraction is made not only from everything empirical but also from all\nintuition. That is not, however, Kant\u2019s real view, or at least not his\nfinal view. In sections A 76-7 = B 102, A 130-1 = B 170, and A 135-6 = B\n174-5, which are probably all of later origin, he states his position in\nthe clearest terms. Transcendental logic, he there declares, differs\nfrom general logic in that it is not called upon to abstract from the\npure _a priori_ manifolds of intuition.[655] This involves, it may be\nnoted, the recognition, so much more pronounced in the later\ndevelopments of Kant\u2019s Critical teaching, of space and time as not\nmerely forms for the apprehension of sensuous manifolds but as\nthemselves presenting to the mind independent manifolds of _a priori_\nnature.\n\nAs the term _transcendental_ indicates, the new logic will have as its\ncentral problems the origin, scope, conditions and possibility of valid\n_a priori_ knowledge of objects. None of these problems are treated in\ngeneral logic, which deals only with the understanding itself. The\nquestion which it raises is, as Kant says in his _Logic_,[656] _How can\nthe understanding know itself?_ The question dealt with by\ntranscendental logic we may formulate in a corresponding way: _How can\nthe understanding possess pure a priori knowledge of objects?_ It is a\ncanon of pure understanding in so far as that faculty is capable of\nsynthetic, objective knowledge _a priori_.[657] General logic involves,\nit is true, the idea of reference to objects,[658] but the possibility\nof such reference is not itself investigated. In general logic the\nunderstanding deals only with itself. It assumes indeed that all objects\nmust conform to its laws, but this assumption plays no part in the\nscience itself.\n\nA further point, not here dwelt upon by Kant, calls for notice, namely,\nthat the activities of understanding dealt with by general logic are its\nmerely discursive activities,--those of discrimination and comparison;\nwhereas those dealt with by transcendental logic are the originative\nactivities through which it produces _a priori_ concepts from within\nitself, and through which it attains, independently of experience, to an\n_a priori_ determination of objects. Otherwise stated, general logic\ndeals only with analytic thinking, transcendental logic with the\nsynthetic activities that are involved in the generation of the complex\ncontents which form the subject matter of the analytic procedure.\n\n=III. Concerning the Division of General Logic into Analytic and\nDialectic.=[659]--The following passage from Kant\u2019s _Logic_[660] forms an\nexcellent and sufficient comment upon the first four paragraphs of this\nsection:\n\n     \u201cAn important perfection of knowledge, nay, the essential and\n     inseparable condition of all its perfection, is truth. Truth is\n     said to consist in the agreement of knowledge with the object.\n     According to this merely verbal definition, then, my knowledge, in\n     order to be true, must agree with the object. Now I can only\n     compare the object with my knowledge by this means, namely _by\n     having knowledge of it_. My knowledge, then, is to be verified by\n     itself, which is far from being sufficient for truth. For as the\n     object is external to me, I can only judge whether my knowledge of\n     the object agrees with my knowledge of the object. Such a circle in\n     explanation was called by the ancients Diallelos. And, indeed, the\n     logicians were accused of this fallacy by the sceptics, who\n     remarked that this account of truth was as if a man before a\n     judicial tribunal should make a statement, and appeal in support of\n     it to a witness whom no one knows, but who defends his own\n     credibility by saying that the man who had called him as a witness\n     is an honourable man. The charge was certainly well-founded. The\n     solution of the problem referred to is, however, absolutely\n     impossible for any man.\n\n     \u201cThe question is in fact this: whether and how far there is a\n     certain, universal, and practically applicable criterion of truth.\n     For this is the meaning of the question, What is truth?...\n\n     \u201cA _universal material_ criterion of truth is not possible; the\n     phrase is indeed self-contradictory. For being _universal_ it would\n     necessarily abstract from all distinction of objects, and yet being\n     a material criterion, it must be concerned with just this\n     distinction in order to be able to determine whether a cognition\n     agrees with the very object to which it refers, and not merely with\n     some object or other, by which nothing would be said. But\n     _material_ truth must consist in this agreement of a cognition with\n     the definite object to which it refers. For a cognition which is\n     true in reference to one object may be false in reference to other\n     objects. It is therefore absurd to demand a universal material\n     criterion of truth, which is at once to abstract and not to\n     abstract from all distinction of objects.\n\n     \u201cBut if we ask for a universal _formal_ criterion of truth, it is\n     very easy to decide that there may be such a criterion. For formal\n     truth consists simply in the agreement of the cognition with itself\n     when we abstract from all objects whatever, and from every\n     distinction of objects. And hence the universal formal criteria of\n     truth are nothing but universal logical marks of agreement of\n     cognitions with themselves, or, what is the same thing, with the\n     general laws of the understanding and the Reason. These formal\n     universal criteria are certainly not sufficient for objective\n     truth, but yet they are to be viewed as its _conditio sine qua\n     non_. For before the question, whether the cognition agrees with\n     the object, must come the question, whether it agrees with itself\n     (as to form). And this is the business of logic.\u201d[661]\n\nThe remaining paragraphs[662] of Section III. may similarly be compared\nwith the following passage from an earlier section of Kant\u2019s\n_Logic_:[663]\n\n     \u201cAnalytic discovers, by means of analysis, all the activities of\n     reason which we exercise in thought. It is therefore an analytic of\n     the form of understanding and of Reason, and is justly called the\n     logic of truth, since it contains the necessary rules of all\n     (formal) truth, without which truth our knowledge is untrue in\n     itself, even apart from its objects. It is therefore nothing more\n     than a canon for deciding on the formal correctness of our\n     knowledge.\n\n     \u201cShould we desire to use this merely theoretical and general\n     doctrine as a practical art, that is, as an organon, it would\n     become a _dialectic_, _i.e. a logic of semblance_ (_ars sophistica\n     disputatoria_), arising out of an abuse of the analytic, inasmuch\n     as by the mere logical form there is contrived the semblance of\n     true knowledge, the characters of which must, on the contrary, be\n     derived from agreement with objects, and therefore from the\n     _content_.\n\n     \u201cIn former times dialectic was studied with great diligence. This\n     art presented false principles in the semblance of truth, and\n     sought, in accordance with these, to maintain things in semblance.\n     Amongst the Greeks the dialecticians were advocates and\n     rhetoricians who could lead the populace wherever they chose,\n     because the populace lets itself be deluded with semblance.\n     Dialectic was therefore at that time the art of semblance. In\n     logic, also, it was for a long time treated under the name of the\n     _art of disputation_, and during that period all logic and\n     philosophy was the cultivation by certain chatterboxes of the art\n     of semblance. But nothing can be more unworthy of a philosopher\n     than the cultivation of such an art. Dialectic in this form,\n     therefore, must be altogether suppressed, and instead of it there\n     must be introduced into logic a critical examination of this\n     semblance.\n\n     \u201cWe should then have two parts of logic: the _analytic_, which will\n     treat of the formal criteria of truth, and the _dialectic_, which\n     will contain the marks and rules by which we can know that\n     something does not agree with the formal criteria of truth,\n     although it seems to agree with them. Dialectic in this form would\n     have its use as a _cathartic_ of the understanding.\u201d\n\nDialectic is thus interpreted in a merely negative sense. It is, Kant\nsays, a catharticon. So far from being an organon, it is not even a\ncanon. It is merely a discipline.[664] By this manner of defining\ndialectic Kant causes some confusion. It does not do justice to the\nscope and purpose of that section of the _Critique_ to which it gives\nits name.[665]\n\n=IV. Concerning the Division of Transcendental Logic into Transcendental\nAnalytic and Dialectic.=--The term object[666] is used throughout this\nsection in two quite distinct senses. In the second and third sentences\nit is employed in its wider meaning as equivalent to content or matter.\nIn the fourth sentence it is used in the narrower and stricter sense,\nmore proper to the term, namely, as meaning \u2018thing.\u2019 Again, in the fifth\nsentence content (_Inhalt_) would seem to be identified with object in\nthe narrower sense, while in the sixth sentence matter (_Materie_, a\nsynonym for content) appears to be identified with object in the wider\nsense. _Transcendental Dialectic_, in accordance with the above account\nof its logical correlate, is defined in a manner which does justice only\nto the negative side of its teaching. Its function is viewed as merely\nthat of protecting the pure understanding against sophistical\nillusions.[667]\n\n\n\n\nTHE TRANSCENDENTAL LOGIC\n\nDIVISION I\n\nTHE TRANSCENDENTAL ANALYTIC\n\n\nThe chief point of this section[668] lies in its insistence that, as the\n_Analytic_ is concerned only with the pure understanding, the _a priori_\nconcepts with which it deals must form a unity or system. Understanding\nis viewed as a separate faculty, and virtually hypostatised. As a\nseparate faculty, it must, it is implied, be an independent unity,\nself-containing and complete. Its concepts are determined in number,\nconstitution, and interrelation, by its inherent character. They\noriginate independently of all differences in the material which they\nare employed to organise.\n\n\n\n\nBOOK I\n\nTHE ANALYTIC OF CONCEPTS\n\n\n=Introductory Paragraph.=--Kant\u2019s view of the understanding as a separate\nfaculty is in evidence again in this paragraph.[669] The _Analytic_ is a\n\u201cdissection of the faculty of the understanding.\u201d _A priori_ concepts\nare to be sought nowhere but in the understanding itself, as their\nbirthplace. There \u201cthey lie ready till at last, on the occasion of\nexperience, they become developed.\u201d But such statements fail to do\njustice to Kant\u2019s real teaching. They would seem to reveal the\npersisting influence of the pre-Critical standpoint of the\n_Dissertation_.\n\n\n\n\nCHAPTER I\n\nTHE CLUE TO THE DISCOVERY OF ALL PURE CONCEPTS OF THE UNDERSTANDING\n\n\nThat the understanding is \u201can absolute unity\u201d is repeated. From this\nassertion, thus dogmatically made, without even an attempt at argument,\nKant deduces the important conclusion that the pure concepts,\noriginating from such a source, \u201cmust be connected with each other\naccording to _one_ concept or idea (_Begriff oder Idee_).\u201d And he adds\nthe equally unproved assertion:\n\n     \u201cBut such a connection supplies a rule by which we are enabled to\n     assign its proper place to each pure concept of the understanding\n     and by which we can determine in an _a priori_ manner their\n     systematic completeness. Otherwise we should be dependent in these\n     matters on our own discretionary judgment or merely on chance.\u201d\n\nIn the next section he sets himself to discover from an examination of\nanalytic thinking what this rule or principle actually is, and in so\ndoing he for the first time discloses, in any degree at all adequate,\nthe real nature of the position which he is seeking to develop. He\nconnects the required principle with the nature of the act of judging,\nconsidered as a function of unity.\n\n\n=Section I. The Logical Use of the Understanding.=--This section,[670]\nviewed as introductory to the metaphysical deduction of the categories,\nis extremely unsatisfactory. It directs attention to the wrong points,\nand conceals rather than defines Kant\u2019s real position. Its argumentation\nis also contorted and confused, and only by the most patient analysis\ncan it be straightened out. The commentator has presented to him a\ntwofold task from which there is no escape. He must render the argument\nconsistent by such modification as will harmonise it with Kant\u2019s later\nand more deliberate positions, and he must explain why Kant has\npresented it in this misleading manner.\n\nThe title of the section would seem to imply that only the discursive\nactivities of understanding are to be dealt with. That is, indeed, in\nthe main true. Confusion results, however, from the clashing of this\navowed intention with the ultimate purpose in view of which the argument\nis propounded. Kant is seeking to prove that we can derive from the more\naccessible procedure of the discursive understanding a clue sufficient\nfor determining those pre-logical activities which have to be postulated\nin terms of his new Copernican hypothesis. But though that is the real\nintention of this section, it has, unfortunately, not been explicitly\nrecognised, and can be divined by the reader only after he has mastered\nthe later portions of the _Analytic_. Kant\u2019s argument has also the\nfurther defect that no sufficient statement is given either of the\nnature of the discursive concept or of its relation to judgment. These\nlacunae we must fill out as best we can from his utterances elsewhere. I\nshall first state Kant\u2019s view of the distinction between discursive and\nsynthetic thinking, and then examine his treatment of the nature of the\nconcept and of its relation to judgment.\n\nAs already noted,[671] the distinction between transcendental and\ngeneral logic marks for Kant all-important differences in the use of the\nunderstanding. In the one employment the understanding, by creative\nsynthetic activities, generates from the given manifold the complex\nobjects of sense-experience. In so doing it interprets and organises the\nmanifold through concepts which originate _from within itself_. By the\nother it discriminates and compares, and thereby derives _from the\ncontent of sense-experience_ the generic concepts of the traditional\nlogic. Now Kant would seem to argue in this section that if the\ndifference in the origin of the concepts in those two cases be left out\nof account, and if we attend only to the quite general character of\ntheir respective activities, they will be found to agree in one\nfundamental feature, namely, that they express functions of unity. Each\nis based on the spontaneity of thought--on the spontaneity of synthetic\ninterpretation on the one hand, of discrimination and comparison on the\nother. This feature common to the two types of activity can be further\ndefined as being the unity of the act whereby a multiplicity is\ncomprehended under a single representation. In the judgment \u201cevery metal\nis a body\u201d the variety of metals is reduced to unity through the concept\nbody. In an analogous manner the synthetic understanding organises a\nmanifold of intuition through some such form of unity as that of\nsubstance and attribute. That is the category which underlies the above\nproposition, and which renders possible the specific unity of the total\njudgment. To quote the sentence with which in a later section Kant\nintroduces his table of categories:\n\n     \u201cThe same understanding, and by the same operations by which in\n     concepts, by means of analytic unity, it has produced the logical\n     form of a judgment, introduces, by means of the synthetic unity of\n     the manifold in intuition in general, a transcendental element into\n     its representations....\u201d[672]\n\nNow Kant\u2019s exposition is extremely misleading. As his later utterances\nshow, his real argument is by no means that which is here given. We\nshall have occasion to observe that Kant is unable to prove, and does\nnot ultimately profess to prove, that it is \u201cthe same understanding,\u201d\nand still less that it is \u201cthe same operations,\u201d which are exercised in\ndiscursive and in creative thinking. But this is a criticism which it\nwould be premature to introduce at this stage. We must proceed to it by\nway of preliminary analysis of the above exposition. Kant\u2019s argument\ndoes not rest upon any such analogy as that just drawn, between the\nconcepts formed by consciously comparing contents and the concepts which\noriginate from within the understanding itself. Both, it is true, are\nfunctions of unity, but otherwise there is, according to Kant\u2019s own\nteaching, not the least resemblance between them. A generic or abstract\nconcept expresses common qualities found in each of a number of complex\ncontents. It is itself a content. A category, on the other hand, is\nalways a function of unity whereby contents are interpreted. It is not a\ncontent, but a form for the organisation of content.[673] It can gain\nexpression only in the total act of judging, not in any one element such\nas the discursive concept. But though the analogy drawn by Kant thus\nbreaks down, his argument is continued in a new and very different form.\nIt is no longer made to rest on any supposed resemblance between\ndiscursive and creative thinking, regarded as co-ordinate and\nindependent activities. It now consists in the proof that the former\npresupposes and is conditioned by the latter. Through study of the\nunderstanding in its more accessible discursive procedure, we may hope\nto discover the synthetic forms according to which it has proceeded in\nits pre-logical activities. When we determine the various forms of\nanalytic judgment, the categories which are involved in synthetic\nthinking reveal themselves to consciousness.\n\nThus in spite of Kant\u2019s insistence upon the conceptual predicate, and\nupon the unity to which it gives expression, immediately he proceeds to\nthe deduction of the categories, the emphasis is shifted to the unity\nwhich underlies the judgment as a whole. What constitutes such\npropositions as \u201call bodies are divisible,\u201d \u201cevery metal is a body,\u201d a\nunique and separate type of judgment is not the character of the\npredicate, but the category of substance and attribute whereby the\npredicate is related to the subject. To that category they owe their\nspecific form; and it is a function of unity for which the discursive\nunderstanding can never account. As Kant states in the _Prolegomena_, if\ngenuine judgments, that is, judgments that are \u201cobjectively valid,\u201d are\nanalysed,\n\n     \u201c...it will be found that they never consist of mere intuitions\n     connected only (as is commonly believed) by comparison in a\n     judgment. They would be impossible were not a pure concept of the\n     understanding superadded to the concepts abstracted from\n     intuition. The abstract concepts are subsumed under a pure concept,\n     and in this manner only can they be connected in an objectively\n     valid judgment.\u201d[674]\n\nThus the analogy between discursive and _a priori_ concepts is no sooner\ndrawn than it is set aside as irrelevant. Though generic concepts rest\nupon functions of unity, and though (as we shall see immediately) they\nexist only as factors in the total act of judging, there is otherwise\nnot the least resemblance between them and the categories.[675] The clue\nto the categories is not to be found in the inherent characteristics of\nanalytic thinking, or of its specific products (namely, concepts), but\nsolely in what, after all abstraction, it must still retain from the\nproducts which synthetic thinking creates. Each type of analytic\njudgment will be found on examination to involve some specific function\nwhereby the conceptual factors are related to, and unified with, the\nother elements in the judgment. This function of unity is in each case\nan _a priori_ category of the understanding. That is the thesis which\nunderlies the concluding sentence of this section.\n\n     \u201cThe functions of the understanding [_i.e._ the _a priori_ concepts\n     of understanding] can be discovered in their completeness, if it is\n     possible to state exhaustively the functions of unity [_i.e._ the\n     forms of relation] in judgments.\u201d\n\nThe adoption of such a position involves, it may be noted, the giving up\nof the assertion, which is so emphatically made in the passage above\nquoted, that it is _by the same activities_ that the understanding\ndiscursively forms abstract concepts and creatively organises the\nmanifold of sense. That is in no respect true. There is no real\nidentity--there is not even analogy--between the processes of comparison\nand abstraction on the one hand and those of synthetic interpretation on\nthe other. The former are merely reflective: the latter are genuinely\ncreative. Discursive activities are conscious processes, and are under\nour control: the synthetic processes, are non-conscious; only their\nfinished products appear within the conscious field. This, however, is\nto anticipate a conclusion which was among the last to be realised by\nKant himself, namely that there is no proof that these two types of\nactivity are ascribable to one and the same source. The synthetic\nactivities--as he himself finally came to hold--are due to a faculty of\nimagination.\n\n     \u201cSynthesis in general ... is the mere result of the power of\n     imagination, a blind but indispensable function of the soul,\n     without which we should have no knowledge whatsoever, but of which\n     we are scarcely ever conscious.\u201d[676]\n\nThis sentence occurs in a passage which is undoubtedly a later\ninterpolation.[677] The \u201cscarcely ever\u201d (_selten nur einmal_) indicates\nKant\u2019s lingering reluctance to recognise this fundamental fact,\ndestructive of so much in his earlier views, even though it completes\nand reinforces his chief ultimate conclusions. With this admission Kant\nalso gives up his sole remaining ground for the contention that there\nmust be a complete parallelism between discursive and creative thinking.\nIf they arise from such different sources, we have no right to assume,\nwithout specific proof, that they must coincide in the forms of their\nactivity. This is a point to which we shall return in discussing Kant\u2019s\nformulation of the principle which is supposed to guarantee the\ncompleteness of the table of categories.\n\nThis unavowed change in point of view is the main cause of confusion in\nthis section. Its other defects are chiefly those of omission. Kant\nfails to develop in sufficient detail his view of the nature of the\ndiscursive concept, or to make sufficiently clear the grounds for his\nassertion that conception as an activity of the understanding is\nidentical with judgment. To take the former point first. Kant\u2019s mode of\nviewing the discursive concept finds expression in the following passage\nin the _Introduction_ to his _Logic_:[678]\n\n     \u201cHuman knowledge is on the side of the understanding _discursive_;\n     that is, it takes place by means of ideas which make what is common\n     to many things the ground of knowledge: and hence by means of\n     attributes as such. We therefore cognise things only by means of\n     attributes. An attribute is that in a thing which constitutes part\n     of our cognition of it; or, what is the same, a partial conception\n     so far as it is considered as a ground of cognition of the whole\n     conception. _All our concepts, therefore, are attributes, and all\n     thought is nothing but conception by means of attributes._\u201d\n\nThe limitations of Kant\u2019s view of the concept could hardly find more\ndefinite expression. The only type of judgment which receives\nrecognition is the categorical, interpreted in the traditional\nmanner.[679]\n\n     \u201cTo compare something as a mark with a thing, is called \u2018to judge.\u2019\n     The thing itself is the subject, the mark [or attribute] is the\n     predicate. The comparison is expressed by the word \u2018is,\u2019 ... which\n     when used without qualification indicates that the predicate is a\n     mark [or attribute] of the subject, but when combined with the sign\n     of negation states that the predicate is a mark opposed to the\n     subject.\u201d[680]\n\nKant\u2019s view of analytic thinking is entirely dominated by the\nsubstance-attribute teaching of the traditional logic. A concept must,\nin its connotation, be an abstracted attribute, and in its denotation\nrepresent a class. Relational thinking, and the concepts of relation,\nare ignored. Thus, in the _Aesthetic_, as we have already noted,[681]\nKant maintains that since space and time are not generic class concepts\nthey must be intuitions. This argument, honestly employed by Kant, shows\nhow completely unconscious he was of the revolutionary consequences of\nhis new standpoint. Even in the very act of insisting upon the\nrelational character of the categories, he still continues to speak of\nthe concept as if it must necessarily conform to the generic type. In\nthis, as in so many other respects, transcendental logic is not, as he\nwould profess, supplementary to general logic; it is its tacit\nrecantation. Modern logic, as developed by Lotze, Sigwart, Bradley, and\nBosanquet, is, in large part, the recasting of general logic in terms of\nthe results reached by Kant\u2019s transcendental enquiries. Meantime,\nsufficient has been said to indicate the strangely limited character of\nKant\u2019s doctrine of the logical concept.\n\nBut on one fundamental point Kant breaks entirely free from the\ntraditional logic. The following passage occurs in the above-quoted\npamphlet on _The Mistaken Subtlety of the Four Syllogistic Figures_:\n\n     \u201cIt is clear that in the ordinary treatment of logic there is a\n     serious error in that distinct and complete concepts are treated\n     before judgments and ratiocinations, although the former are only\n     possible by means of the latter.\u201d \u201cI say, then, first, that a\n     _distinct_ concept is possible only by means of a _judgment_, a\n     _complete_ concept only by means of a _ratiocination_. In fact, in\n     order that a concept should be distinct, I must clearly recognise\n     something as an attribute of a thing, and this is a judgment. In\n     order to have a distinct concept of body, I clearly represent to\n     myself impenetrability as an attribute of it. Now this\n     representation is nothing but the thought, \u2018a body is\n     impenetrable.\u2019 Here it is to be observed that this judgment is not\n     the distinct concept itself, but is the act by which it is\n     realised; for the idea of the thing which arises after this act is\n     distinct. It is easy to show that a complete concept is only\n     possible by means of a ratiocination: for this it is sufficient to\n     refer to the first section of this essay. We might say, therefore,\n     that a distinct concept is one which is made clear by a judgment,\n     and a complete concept one which is made distinct by a\n     ratiocination. If the completeness is of the first degree, the\n     ratiocination is simple; if of the second or third degree, it is\n     only possible by means of a chain of reasoning which the\n     understanding abridges in the manner of a sorites.... Secondly, as\n     it is quite evident that the completeness of a concept and its\n     distinctness do not require different faculties of the mind (since\n     the same capacity which recognises something immediately as an\n     attribute in a thing is also employed to recognise in this\n     attribute another attribute, and thus to conceive the thing by\n     means of a remote attribute), so also it is evident that\n     understanding and reason, that is, the power of cognising\n     distinctly and the power of forming ratiocinations, are not\n     different faculties. Both consist in the power of judging, but when\n     we judge mediately we reason.\u201d[682]\n\nIn the section before us this same standpoint is maintained, but is\nexpressed in a much less satisfactory manner. Concepts are no longer\nspoken of as complete judgments. In the above passages Kant always\nspeaks of the concept as the subject of the proposition; it is now\ntreated only as a predicate.[683] This difference is significant. The\nconcept as subject can represent the judgment as a whole (or at least it\ndoes so from the traditional standpoint to which Kant holds); the\nconcept as predicate is merely one element, even though it be a unifying\nelement, in the total act of judging. This falling away from his own\nmaturer standpoint would seem to be due to Kant\u2019s lack of clearness as\nto the nature of the analogy which he is here drawing between analytic\nand synthetic thinking. It is connected with his mistaken, and merely\ntemporary, comparison of _a priori_ with discursive concepts. His\nposition in 1762 alone harmonises with his essential teaching. Now, as\nthen, he is prepared to view judgment as the sole ultimate activity of\nthe understanding, and therefore to define understanding as the faculty\nof judging.\n\nBut the new Critical standpoint compels Kant to reinterpret this\ndefinition in a manner which involves a still more radical\ntransformation of the traditional doctrine. The categories constitute a\nunique type of concept, and condition the processes of discursive\nthought. They are embodied in the complex contents from which analytic\nthinking starts; and however far the processes of discursive comparison\nand abstraction be carried, one or other of these categories must still\npersist, determining the form which the analytic judgment is to take.\nThe categorical judgment can formulate itself only by means of the _a\npriori_ concept of subject and attribute, the hypothetical only by means\nof the pure concept of ground and consequence, and so with the others.\nAnd there are in consequence just as many categories as there are forms\nof the analytic judgment. This is how the principle of the metaphysical\ndeduction must be interpreted when the later and deeper results of the\ntranscendental deduction are properly taken into account. In deducing\nthe forms of the understanding from the modes of discursive judgment\nKant is virtually maintaining that analytic judgment involves the same\nproblems as does judgment of the synthetic type. The categories can be\nderived from the forms of discursive judgment only because they are the\nconditions in and through which it becomes possible.\n\nBut though Kant, both here and in the central portions of the\n_Analytic_, seems to be on the very brink of this conclusion, it is\nnever explicitly drawn. As we shall see,[684] it would have involved the\nfurther admission that there is no absolute guarantee of the\ncompleteness of the table of categories, and no satisfactory method of\ndetermining their interrelations. To the very last general logic is\nisolated from transcendental logic. The Critical enquiry is formulated\nas if it concerned only such judgments as are explicitly synthetic. The\nprinciple of the metaphysical deduction is not, therefore, stated by\nKant himself in the above manner; and we have still to decide the\ndifficult question as to what the principle employed by Kant in the\ndeduction actually is.\n\nKant makes a twofold demand upon the principle. It must enable us to\ndiscover the categories, and it must also in so doing enable us to view\nthem as together forming a systematic whole, and so as having their\ncompleteness guaranteed by other than merely empirical considerations.\nThe principle is stated sometimes in a broader and sometimes in a more\nspecific form; for on this point also Kant speaks with no very certain\nvoice.[685] The broader formulation of the principle is that all acts of\nunderstanding are judgments, and that therefore the possible ultimate _a\npriori_ forms of understanding are identical with the possible ultimate\nforms of the judgment.[686] The more specific and correct formulation is\nthat to every form of analytic judgment there corresponds a pure concept\nof understanding. The first statement of the principle is obviously\ninadequate. It merely reformulates the problem as being a problem not of\nconception but of judgment. If a principle is required to guarantee the\ncompleteness of our list of _a priori_ concepts, it will equally be\nrequired to guarantee the completeness of our list of judgments. Even if\nthe above principle be more explicitly formulated, as in the\n_Prolegomena_,[687] where judging is defined as the act of understanding\nwhich comprises all its other acts, it will not enable us to guarantee\nthe completeness of any list of the forms of judgment or to determine\ntheir systematic interrelation. We are therefore thrown back upon the\nsecond view. This, however, only brings us face to face with the further\nquestion, what principle guarantees the completeness of the table of\nanalytic judgments. And to that query Kant has absolutely no answer. The\nreader\u2019s questionings break vainly upon his invincible belief in the\nadequacy and finality of the classification yielded by the traditional\nlogic.\n\nThe _fons et origo_ of all the confusions and obscurities of this\nsection are thus traceable to Kant\u2019s attitude towards formal logic. He\nmight criticise it for ignoring the interdependence of conception,\njudgment, and reasoning; he might reject the second, third, and fourth\nsyllogistic figures; and he might even admit that its classification of\nthe forms of judgment is not as explicit as might be desired; but\nhowever many provisos he made and defects he acknowledged, they were to\nhim merely minor matters, and he accepted its teaching as complete and\nfinal. This unwavering faith in the fundamental distinctions of the\ntraditional logic was indeed, as we shall have constant occasion to\nobserve, an ever present influence in determining alike the general\nframework and much of the detail of Kant\u2019s Critical teaching. The\ndefects of the traditional logic were very clearly indicated in his own\ntranscendental logic. He showed that synthetic thinking is fundamental;\nthat by its distinctions the forms and activities of analytic thought\nare predetermined; that judgment in its various forms can be understood\nonly by a regress upon the synthetic concepts to which these forms are\ndue; that notions are not merely of the generic type, but that there are\nalso categories of relation. None the less, to the very last, Kant\npersisted in regarding general logic as a separate discipline, and as\nquite adequate in its current form. He continued to ignore the fact that\nthe analytic judgment, no less than the synthetic judgment, demands a\ntranscendental justification.\n\nThe resulting situation is strangely perverse. In the very act of\nrevolutionising the traditional logic, Kant relies upon its prestige and\nupon the assumed finality of its results to make good the shortcomings\nof the logic which is to displace it. By Kant\u2019s own admission\ntranscendental logic is incapable of guaranteeing that completeness upon\nwhich, throughout the whole _Critique_, so great an emphasis is laid.\nGeneral logic is allowed an independent status, sufficient to justify\nits authority being appealed to; and the principle which is supposed to\nguarantee the completeness of the table of categories is so formulated\nas to contain no suggestion of the dependence of discursive upon\nsynthetic thinking. Formal logic, Kant would seem to hold, can supply a\ncriterion for the classification of the ultimate forms of judgment just\nbecause its task is relatively simple, and is independent of all\nepistemological views as to the nature, scope, and conditions of the\nthought process. Since formal logic is a completed and perfectly _a\npriori_ science, which has stood the test of 2000 years, and remains\npractically unchanged to the present day, its results can be accepted as\nfinal, and can be employed without question in all further enquiries.\nAnalytic thinking is scientifically treated in general logic; the\n_Critique_ is concerned only with the possibility and conditions of\nsynthetic judgment. The table of analytic judgments therefore supplies a\ncomplete and absolutely guaranteed list of the possible categories of\nthe understanding. But the perverseness of this whole procedure is shown\nby the manner in which, as we shall find, Kant recasts, extends, or\nalters, to suit his own purposes, the actual teaching of the traditional\nlogic.\n\nAs noted above,[688] the asserted parallelism of analytic and synthetic\njudgment rests upon the further assumption that discursive thinking and\nsynthetic interpretation are the outcome of one and the same faculty of\nunderstanding. It is implied, in accordance with the attitude of the\npre-Critical _Dissertation_, that understanding, viewed as the faculty\nto which all thought processes are due, has certain laws in accordance\nwith which it necessarily acts in all its operations, and that these\nmust therefore be discoverable from analytic no less than from synthetic\nthinking. The mingling of truth and falsity in this assumption has\nalready been indicated. Such truth as it contains is due to the fact\nthat analytic thinking is not co-ordinate with, but is dependent upon,\nand determined by, the forms of synthetic thinking. Its falsity consists\nin its ignoring of what thus gives it partial truth. The results of the\ntranscendental deduction call for a complete recasting of the entire\nargument of the metaphysical deduction. And when this is done, there is\nno longer any ground for the contention that the number of the\ncategories is determinable on _a priori_ grounds. On Kant\u2019s own\nfundamental doctrine of the synthetic, and therefore merely _de facto_,\ncharacter of all _a priori_ principles, the necessity of the categories\nis only demonstrable by reference to the contingent fact of actual\nexperience. The possible conceptual forms are relative to actual and\nultimate differences in the contingent sensuous material; and being thus\nrelative, they cannot possibly be systematised on purely _a priori_\ngrounds. This Kant has himself admitted in a passage added in the second\nedition,[689] though apparently without full consciousness of the\nimportant consequences which must follow.\n\n     \u201cThis peculiarity of our understanding that it can produce _a\n     priori_ unity of apperception solely by means of the categories,\n     and only by such and so many, is as little capable of further\n     explanation as why we have just these and no other functions of\n     judgment, or why space and time are the only forms of our possible\n     intuition.\u201d\n\n\nSTAGES IN THE DEVELOPMENT OF KANT\u2019S METAPHYSICAL DEDUCTION\n\nThe character of the metaphysical deduction will be placed in a clearer\nlight if we briefly trace the stages, so far as they can be\nreconstructed, through which it passed in Kant\u2019s mind. We may start from\nthe _Dissertation_ of 1770. Kant there modifies his earlier Wolffian\nstandpoint, developing it, probably under the direct influence of the\nrecently published _Nouveaux Essais_, on more genuinely Leibnizian\nlines.\n\n     \u201cThe use of the intellect ... is twofold. By the one use concepts,\n     both of things and of relations, are themselves _given_. This is\n     the _real use_. By the other use concepts, whencesoever given, are\n     merely _subordinated_ to each other, the lower to the higher (the\n     common attributes), and compared with one another according to the\n     principle of contradiction. This is called the _logical use_....\n     Empirical concepts, therefore, do not become intellectual in the\n     _real sense_ by reduction to greater universality, and do not pass\n     beyond the type of sensuous cognition. However high the abstraction\n     be carried, they must always remain sensuous. But in dealing with\n     things _strictly intellectual_, in regard to which the _use of the\n     intellect_ is _real_, intellectual concepts (of objects as well as\n     of relations), are given by the very nature of the intellect. They\n     are not abstracted from any use of the senses, and do not contain\n     any form of sensuous knowledge as such. We must here note the\n     extreme ambiguity of the word _abstract_.... An intellectual\n     concept _abstracts_ from everything sensuous; it is not\n     _abstracted_ from things sensuous. It would perhaps be more\n     correctly named _abstracting_ than _abstract_. It is therefore\n     preferable to call the intellectual concepts _pure ideas_, and\n     those which are given only empirically _abstract ideas_.\u201d[690] \u201cI\n     fear, however, that Wolff, by this distinction between the sensuous\n     and the intellectual, which for him is merely logical, has checked,\n     perhaps wholly (to the great detriment of philosophy), that noblest\n     enterprise of antiquity, the investigation of _the nature of\n     phenomena and noumena_, turning men\u2019s minds from such enquiries to\n     what are very frequently only logical subtleties. Philosophy, in so\n     far as it contains the _first principles_ of the use of the _pure\n     intellect_, is _metaphysics_.... As empirical principles are not to\n     be found in metaphysics, the concepts to be met with in it are not\n     to be sought in the senses but in the very nature of the pure\n     intellect. They are not _connate_ concepts, but are _abstracted_\n     from laws inherent in the mind (_legibus menti insitis_), and are\n     therefore _acquired_. Such are the concepts of possibility,\n     existence, necessity, substance, cause, etc. with their opposites\n     or correlates. They never enter as parts into any sensuous\n     representation, and therefore cannot in any fashion be abstracted\n     from such representations.\u201d[691]\n\nThe _etcetera_, with which in that last passage Kant concludes his list\nof pure intellectual concepts, indicates a problem that must very soon\nhave made itself felt. That it did so, appears from his letter to Herz\n(February 21, 1772). He there informs his correspondent, that, in\ndeveloping his _Transcendentalphilosophie_ (the first occurrence of that\ntitle in Kant\u2019s writings), he has\n\n     \u201c...sought to reduce all concepts of completely pure reason to a\n     fixed number of categories [this term also appearing for the first\n     time], not in the manner of Aristotle, who in his ten predicaments\n     merely set them side by side in a sort of order, just as he might\n     happen upon them, but as they distribute themselves of themselves\n     according to _some few principles_ of the understanding.\u201d[692]\n\nThough in this same letter Kant professes to have solved his problems,\nand to be in a position to publish his _Critique of Pure Reason_ (this\ntitle is already employed) \u201cwithin some three months,\u201d the phrase \u201csome\nfew principles\u201d clearly shows that he has not yet developed the teaching\nembodied in the metaphysical deduction. For its keynote is insistence\nupon the necessity of a _single_ principle, sufficient to reduce them\nnot merely to classes but to system. The difficulty of discovering such\na principle must have been one of the causes which delayed completion of\nthe _Critique_. The only data at our disposal for reconstructing the\nvarious stages through which Kant\u2019s views may have passed in the period\nbetween February 1772 and 1781 are the _Reflexionen_, but they are\nsufficiently ample to allow of our doing so with considerable\ndefiniteness.[693]\n\nIn the _Dissertation_ Kant had traced the concepts of space and time, no\nless than the concepts of understanding, to mental activities.\n\n     \u201cBoth concepts [space and time] are undoubtedly acquired. They are\n     not, however, abstracted from the sensing of objects (for sensation\n     gives the matter, not the form of human cognition). As immutable\n     types they are intuitively apprehended from the activity whereby\n     the mind co-ordinates its sensuous data in accordance with\n     perpetual laws.\u201d[694]\n\nNow the _Dissertation_ is quite vague as to how the \u201cmind\u201d (_animus_),\nactive in accordance with laws generative of the intuitions space and\ntime, differs from \u201cunderstanding\u201d (_intellectus_), active in accordance\nwith laws generative of pure concepts. Kant\u2019s reasons, apart from the\nintuitive character of space and time, for contrasting the former with\nthe latter, as the sensuous with the intellectual, were the existence of\nthe antinomies and his belief that through pure concepts the absolutely\nreal can be known. When, however, that belief was questioned by him, and\nhe had come to regard the categories as no less subjective than the\nintuitional forms, the antinomies ceased to afford any ground for thus\ndistinguishing between them. The intuitional nature of space and time,\nwhile certainly peculiar to them, is in itself no proof that they belong\nto the sensuous side of the mind.[695]\n\nA difficulty which immediately faced Kant, from the new Critical\nstandpoint, was that of distinguishing between space and time, on the\none hand, and the categories on the other. This is borne out by the\n_Reflexionen_ and by the following passage in the _Prolegomena_.[696]\n\n     \u201cOnly after long reflection, expended in the investigation of the\n     pure non-empirical elements of human knowledge, did I at last\n     succeed in distinguishing and separating with certainty the pure\n     elementary concepts of sensibility (space and time) from those of\n     the understanding.\u201d\n\nThe first stage in the development of the metaphysical deduction would\nseem to have consisted in the attempt to view the categories as acquired\nby reflection upon the activities of the understanding in \u201ccomparing,\ncombining, or separating\u201d;[697] and among the _notiones rationales_,\n_notiones intellectus puri_, thus gained, the idea of space is specially\nnoted. The following list is also given:\n\n     \u201cThe concepts of existence (reality), possibility, necessity,\n     ground, unity and plurality, parts, all, none, composite and\n     simple, space, time, change, motion, substance and accident, power\n     and action, and everything that belongs to ontology proper.\u201d[698]\n\nIn _Reflexionen_, ii. 507 and 509, the fundamental feature of such\nrational concepts is found in their _relational_ character. They all\nagree in being concepts of form.[699]\n\nQuite early, however, Kant seems to have developed the view, which has\ncreated so many more difficulties than it resolves, that space and time\nare given to consciousness through outer and inner sense. Though still\nfrequently spoken of as concepts, they are definitely referred to the\nreceptive, non-spontaneous, side of the mind. This is at once a return\nto the _Dissertation_ standpoint, and a decided modification of its\nteaching. It holds to the point of view of the _Dissertation_ in so far\nas it regards them as sensuous, and departs from it in tracing them to\nreceptivity.[700]\n\nThe passage quoted from the letter of 1772 to Herz may perhaps be\nconnected with the stage revealed in the _Reflexionen_ already cited.\n\u201cComparing, combining, and separating\u201d may be the \u201csome few principles\nof the understanding\u201d there referred to. That, however, is doubtful, for\nthe next stage in the development likewise resulted in a threefold\ndivision. This second stage finds varied expression in _Reflexionen_,\nii. 483, 522, 528, 556-63. These, in so far as they agree, distinguish\nthree classes of categories--of thesis, of analysis, and of synthesis.\nThe first covers the categories of quality and modality, the second\nthose of quantity, the third those of relation.\n\n_Reflexionen_, ii. 528 is as follows:\n\n  [Thesis = ]    \u201cThe metaphysical concepts are, first, absolute:\n                     possibility and existence; secondly, relative:\n                (_a_) Unity and plurality: _omnitudo_ and _particularitas_.\n\n  [Analysis = ] (_b_) Limits: the first, the last: _infinitum_, _finitum_.\n                    [Anticipates the later category of limitation.]\n\n                 (_c_) Connection: co-ordination: whole and part\n  [Synthesis = ]     [anticipates the later category of reciprocity],\n                     simple and compound; subordination:\n                      (1) Subject and predicate.\n                      (2) Ground and consequence.\n\nThis, and the connected _Reflexionen_ enumerated above, are of interest\nas proving that Kant\u2019s table of categories was in all essentials\ncomplete before the idea had occurred to him of further systematising it\nor of guaranteeing its completeness by reference to the logical\nclassification of the forms of judgment. They also justify us in the\nbelief that when Kant set himself to discover such a unifying principle\nthe above list of categories and the existing logical classifications\nmust have mutually influenced one another, each undergoing such\nmodification as seemed necessary to render the parallelism complete.\nThis, as we shall find, is what actually happened. The logical table,\nfor instance, induced Kant to distinguish the categories of quality from\nthose of modality, while numerous changes were made in the logical table\nitself in order that it might yield the categories required.\n\nBut the most important alteration, the introduction of the threefold\ndivision of each sub-heading, is not thus explicable, as exclusively due\nto one or other of the two factors. The adoption of this threefold\narrangement in place of the dichotomous divisions of the logical\nclassification and of the haphazard enumerations of Kant\u2019s own previous\nlists, seems to be due to the twofold circumstance that he had already\ndistinguished three categories of synthesis or relation (always the most\nimportant for Kant), and that this sufficiently harmonised with the\nlogical distinction between categorical, hypothetical, and disjunctive\njudgments. He then sought to modify the logical divisions by addition in\neach case of a third, and finding that this helped him to obtain the\ncategories required, the threefold division became for him (as it\nremained for Hegel) an almost mystical dogma of transcendental\nphilosophy.[701] In so far as it involved recognition that the hard and\nfast opposites of the traditional logic (such as the universal and the\nparticular, the affirmative and the negative) are really aspects\ninseparably involved in every judgment and in all existence, it\nconstituted an advance in the direction both of a deeper rationalism and\nof a more genuine empiricism. But in so far as it was due to the desire\nto guarantee completeness on _a priori_ grounds, and so was inspired by\na persistent overestimate of our _a priori_ powers, it has been\ndecidedly harmful. Much of the useless \u201carchitectonic\u201d of the _Critique_\nis due to this scholastic prejudice.\n\nThis fundamental alteration in the table of logical judgments is\nintroduced with the naive assertion that \u201cvarieties of thought in\njudgments,\u201d unimportant in general logic, \u201cmay be of importance in the\nfield of its pure _a priori_ knowledge.\u201d In the _Critique of\nJudgment_[702] we find the following passage:\n\n     \u201cIt has been made a difficulty that my divisions in pure philosophy\n     have almost always been threefold. But this lies in the nature of\n     the case. If an _a priori_ division is to be made, it must be\n     either analytic, according to the principle of contradiction, and\n     then it is always twofold (_quodlibet ens est aut A aut non A_); or\n     else synthetic. And if in this latter case it is derived from _a\n     priori_ concepts (not as in mathematics from the _a priori_\n     intuition corresponding to the concept) the division must\n     necessarily be a trichotomy. For according to what is requisite for\n     synthetic unity in general, there must be (1) a condition, (2) a\n     conditioned, and (3) the concept which arises from the union of\n     these two.\u201d\n\nThe last stage, as expressed in the _Critique_, was, as we have already\nnoted, merely an application of his earlier position that all thinking\nis judging. This appreciation of the inseparable connection of the\ncategories with the act of judging is sound in principle, and is\npregnant with many of the most valuable results of the Critical\nteaching. But these fruitful consequences follow only upon the lines\ndeveloped in the _transcendental_ deduction. They are bound up with\nKant\u2019s fundamental Copernican discovery that the categories are forms of\n_synthesis_, and accordingly express _functions_ or _relations_. The\ncategories can no longer be viewed, in the manner of the\n_Dissertation_,[703] as yielding concepts of _objects_. The view of the\nconcept which we find in the _Dissertation_ is, indeed, applied in the\n_Critique_ to space and time--they are taken as in themselves\nintuitions, not as merely _forms_ of intuition--but the categories are\nrecognised as being of an altogether relational character. Though _a\npriori_, they are not, in and by themselves, complete objects of\nconsciousness, and accordingly can reveal no object. They are\n_functions_, not _contents_. That, however, is to anticipate. We must\nfirst discharge, as briefly as possible, the ungrateful task of dwelling\nfurther upon the laboured, arbitrary, and self-contradictory character\nof the detailed working out of the metaphysical deduction. The deduction\nis given in Sections II. and III.\n\n\n=Section II. The Logical Function of the Understanding in\nJudgment.=[704]--Kant\u2019s introductory statement may here be noted. If, he\nsays, we leave out of consideration the content of any judgment, and\nattend only to the mere form, we \u201cfind\u201d that the function of thought in\na judgment \u201ccan\u201d be brought under four heads, each with three\nsubdivisions. But Kant himself, in this same section, recognises in the\nfrankest and most explicit manner, that the necessary distinctions are\nonly to be obtained by taking account of the matter as well as of the\nform of judgments. And even after this contradiction is discounted, the\nterm \u201cfind\u201d may be allowed as legitimate only if the word \u201ccan\u201d is\ncorrespondingly emphasised. The distinctions were not derived from any\nexisting logic. They were reached only by the freest possible handling\nof the classifications currently employed. Examination of the table of\njudgments, and comparison of it with the table of categories, supply\nconclusive evidence that the former has been rearranged, in highly\nartificial fashion, so as to yield a more or less predetermined list of\nrequired categories.\n\n=1. Quantity.=--Kant here frankly departs from the classification of\njudgments followed in formal logic; and the reason which he gives for so\ndoing is in direct contradiction to his demand that only the form of\njudgment must be taken into account. The \u201cquantity of knowledge\u201d here\nreferred to is determinable, not from the form, but only from the\ncontent of the judgment. Also, the statement that the singular judgment\nstands to the universal as unity to infinity (_Unendlichkeit_) is\ndecidedly open to question. The universal is itself a form of unity, as\nKant virtually admits in deriving, as he does, the category of unity\nfrom the universal judgment.\n\n=2. Quality.=--Kant makes a similar modification in the logical treatment\nof quality, by distinguishing between affirmative and infinite\njudgments. The proposition, A is not-B, is to be viewed as neither\naffirmative nor negative. As the _content_ of the predicate includes the\ninfinite number of things that are not-B, the judgment is infinite.\nKant, in a very artificial and somewhat arbitrary manner, contrives to\ndefine it as limitative in character, and so as sharing simultaneously\nin the nature both of affirmation and of negation. The way is thus\nprepared for the \u201cdiscovery\u201d of the category of limitation.\n\n=3. Relation.=--Wolff, Baumgarten, Meier, Baumeister, Reimarus, and\nLambert, with very minor differences, agree in the following\ndivision:[705]\n\n           {Simple = Categorical\n           {\n  Judgments{         {Copulative (_i.e._ categorical with more\n           {         {than one subject or more than one\n           {Complex  {predicate).\n                     {Hypothetical.\n                     {Disjunctive.\n\nKant omits the copulative judgment, and by ignoring the distinction\nbetween simple and complex judgments (which in Reimarus, and also less\ndefinitely in Wolff, is connected with the distinction between\nconditional and unconditional judgments) contrives to bring the\nremaining three types of judgment under the new heading of \u201crelation.\u201d\nThey had never before been thus co-ordinated, and had never before been\nsubsumed under this particular title. It is by no means clear why such\ndistinctions as those between simple and complex, conditioned and\nunconditioned, should be ignored, and why the copulative judgment should\nnot be recognised as well as the hypothetical. Kant\u2019s criterion of\nimportance and unimportance in the distinctions employed by the\nlogicians of his day was wholly personal to himself; and, though hard to\ndefine, was certainly not dictated by any logic that is traceable to\nAristotelian sources. His exposition is throughout controlled by\nforeknowledge of the particular categories which he desires to\n\u201cdiscover.\u201d\n\n=4. Modality.=--Neither Wolff nor Reimarus gives any account of\nmodality.[706] Baumgarten classifies judgments as pure or modal\n(existing in _four_ forms, necessity, contingency, possibility,\nimpossibility). Baumeister and Thomasius also recognise four forms of\nmodality. Meier distinguishes between pure judgment (_judicium purum_)\nand impure judgment (_judicium modale_, _modificatum_, _complexum qua\ncopula_), but does not classify the forms of modality. Lambert\nalone[707] classifies judgments as \u201cpossible, actual (_wirklich_),\nnecessary, and their opposite.\u201d But when Kant adopts this threefold\ndivision, the inclusion of actuality renders the general title\n\u201cmodality\u201d inapplicable in its traditional sense. The expression of\nactuality in the assertoric judgment involves no adverbial modification\nof the predicate. Also, in its \u201caffirmative\u201d and \u201ccategorical\u201d forms it\nhas already been made to yield two other categories.\n\nKant speaks of the problematic, the assertoric, and the apodictic forms\nof judgment as representing the stages through which knowledge passes in\nthe process of its development.\n\n     \u201cThese three functions of modality are so many momenta of thought\n     in general.\u201d\n\nThis statement has been eulogised by Caird,[708] as being an\nanticipation of the Hegelian dialectic. As a matter of fact, Kant\u2019s\nremark is irrelevant and misleading. The advance from consciousness of\nthe problematic, through determination of it as actual to its\nexplanation as necessary, represents only a psychological order in the\nmind of the individual. Logically, knowledge of the possible rests on\nand implies prior knowledge of the actual and of the necessities that\nconstitute the actual.[709]\n\n\n=Section III.[710] The Categories or Pure Concepts of the\nUnderstanding.=--The first three pages of this section, beginning\n\u201cGeneral logic abstracts,\u201d and concluding with the word \u201crest on the\nunderstanding,\u201d would seem to be a later interpolation. Embodying, as\nthey do, some of the fundamental ideas of the transcendental deduction,\nthey express Kant\u2019s final method of distinguishing between general and\ntranscendental logic. But they are none the less out of harmony with the\nother sections of the metaphysical deduction. They are of the nature of\nan after-thought, even though that afterthought represents a more mature\nand adequate standpoint. In A 55-7, where Kant defines the distinction\nbetween general and transcendental logic, the latter is formulated in\nentire independence of all reference to pure intuition.[711] Kant,\nindeed, argues[712] that just as there are both pure and empirical\nintuitions, so there are both pure and empirical concepts. But there is\nno indication that he has yet realised the close interdependence of the\ntwo types of _a priori_ elements. Even when he proceeds in A 62 to\nremark that the empirical employment of pure concepts is conditioned by\nthe fact that objects are given in intuition, no special reference is\nmade to \u201cthe manifold of pure _a priori_ intuition.\u201d Now, however, Kant\nemphasises, as the fundamental characteristic of transcendental logic,\nits possession of a pure manifold through reference to which its pure\nconcepts gain meaning. Thus not only does transcendental logic not\nabstract from the pure _a priori_ concepts, it likewise possesses an _a\npriori_ material.[713] It is in this twofold manner that it is now\nregarded as differing from formal logic.\n\nThe accounts given of the metaphysical deduction by Cohen,[714]\nCaird,[715] Riehl,[716] and Watson[717] are vitiated by failure to\nremark that this latter standpoint is a late development, and is out of\nkeeping with the rest of the deduction. Riehl\u2019s exposition has, however,\nthe merit of comparative consistency. He explicitly recognises the\nimportant consequence which at once follows from acceptance of this\nlater view, namely, that it is by their implying space and time that the\ncategories differ from the notions which determine the forms of\njudgment; in other words, that the _categories are actualised only as\nschemata_. The category of substance, for instance, differs from the\nmerely logical notion of a propositional subject, in being the concept\nof that which is _always_ a subject, and _never_ a predicate; and such a\nconception has specific meaning for us only as the _permanent in time_.\nLogical subjects and predicates, quantitative relations apart, are\ninterchangeable. The relation between them is the analytic relation of\nidentity. The concept of subject, on the other hand, transcendentally\nviewed, that is, as a _category_, is the apprehension of what is\n_permanent_, in synthetic distinction from, and relation to, its\nchanging attributes. In other words, the transcendental distinction\nbetween _substance_ and _accidents_ is substituted for that of _subject_\nand _predicate_. Similarly the logical relation of _ground_ and\n_consequence_, conceived as expressive of logical identity, gives way to\nthe synthetic temporal relation of _cause_ and _effect_. And so with all\nthe other pure forms. As categories, they are schemata. Kant has\n_virtually_ recognised this by the names which he gives to the\ncategories of relation. But the proper recognition of the necessary\ninterdependence of the intuitional and conceptual forms came too late to\nprevent him from distinguishing between categories and schemata, and so\nfrom creating for himself the artificial difficulties of the section on\nschematism.\n\nIn A 82 Kant states that he intentionally omits definitions of the\ncategories. He had good reason for so doing. The attempt would have\nlanded him in manifold difficulties, since his views were not yet\nsufficiently ripe to allow of his perceiving the way of escape. In A 241\n(omitted in second edition) Kant makes, however, the directly counter\nstatement that definition of the categories is not possible, giving as\nhis reason that, in isolation from the conditions of sensibility, they\nare merely logical functions, \u201cwithout the slightest indication as to\nhow they can possess meaning and objective validity.\u201d[718]\n\nIt cannot be too often repeated that the _Critique_ is not a unitary\nwork, but the patchwork record of twelve years of continuous\ndevelopment. Certain portions of the transcendental deduction, of which\nA 76-9 is one, represent the latest of all the many stages; and their\nteaching, when accepted, calls for a radical recasting of the\nmetaphysical deduction. The bringing of the entire _Critique_ into line\nwith its maturest parts would have been an Herculean task; and it was\none to which Kant, then fifty-seven years of age, was very rightly\nunwilling to sacrifice the time urgently needed for the writing of his\nother _Critiques_. The passage before us is one of the many\ninterpolations by which Kant endeavoured to give an external unity to\nwhat, on close study, is found to be the plain record of successive and\nconflicting views. Meantime, in dealing with this passage, we are\nconcerned only to note that if this later mode of defining\ntranscendental logic be accepted, far-reaching modifications in Kant\u2019s\nCritical teaching have to be made. The other points developed in A 76-9\nwe discuss below[719] in their proper connection.\n\n=The same Function, etc.=[720]--This passage has already been sufficiently\ncommented upon.[721] Kant here expresses in quite inadequate fashion the\nstandpoint of the transcendental deduction. The implication is that\nanalytic and synthetic thinking are co-ordinate, one and the same\nfaculty exercising, on these two levels, the same operations. The true\nCritical teaching is that synthetic thinking is alone fundamental, and\nthat only by a regress upon it can judgments be adequately accounted\nfor. This passage, like the preceding, may be of later origin than the\nmain sections of the metaphysical deduction.\n\n=Term \u201cCategories\u201d[722] borrowed from Aristotle.=--Cf. below, p. 198.\n\n=Table of Categories. Quantity.=--Kant derives the category of unity from\nthe universal,[723] and that of totality (_Allheit_)[724] from the\nsingular. These derivations are extremely artificial. In _Reflexionen_,\nii. 563, Kant takes the more natural line of identifying totality with\nthe universal, and unity with the singular. Probably[725] the reason of\nKant\u2019s change of view is the necessity of obtaining totality by\ncombining unity with multiplicity. That can only be done if universality\nis thus equated with unity. Watson\u2019s explanation,[726] that Kant has\nreversed the order of the categories, seems to be erroneous.\n\n=Quality.=--Cf. above, p. 192.\n\n=Relation.=--The correlation of the categorical judgment with the\nconception of substance and attribute is only possible[727] owing to\nKant\u2019s neglect of the relational judgment and to the dominance in his\nlogical teaching of the Aristotelian substance-attribute view of\npredication. The correlation is also open to question in that the\nrelation of subject and predicate terms in a logical judgment is a\nreversible one. It is a long step from the merely grammatical subject to\nthe conception of that which is always a subject and never a predicate.\n\nKant\u2019s identification of the category of community or reciprocity with\nthe disjunctive judgment, though at first sight the most arbitrary of\nall, is not more so than many of the others. Its essential correctness\nhas been insisted upon in recent logic by Sigwart, Bradley, and\nBosanquet. In Kant\u2019s own personal view[728] co-ordination in the form of\nco-existence is only possible through reciprocal interaction. The\nrelation of whole and part (the parts in their relations of reciprocal\nexclusion exhausting and constituting a genuine whole) thus becomes, in\nits application to actual existences, that of reciprocal causation. The\nreverse likewise holds; interaction is only possible between existences\nwhich together constitute a unity.[729] Kant returns to this point in\n_Note_ 3, added in the second edition.[730] The objection which Kant\nthere considers has been very pointedly stated by Schopenhauer.\n\n     \u201cWhat real analogy is there between the problematical determination\n     of a concept by disjunctive predicates and the thought of\n     reciprocity? The two are indeed absolutely opposed, for in the\n     disjunctive judgment the actual affirmation of one of the two\n     alternative propositions is also necessarily the negation of the\n     other; if, on the other hand, we think of two things in the\n     relation of reciprocity, the affirmation of one is also necessarily\n     the affirmation of the other, and _vice versa_.\u201d[731]\n\nThe answer to this criticism is on the lines suggested by Kant. The\nvarious judgments which constitute a disjunction do not, _when viewed as\nparts of the disjunction_, merely negate one another; they mutually\npresuppose one another in the total complex. Schopenhauer also fails to\nobserve that in locating the part of a real whole in one part of space,\nwe exclude it from all the others.[732]\n\n=Modality.=--The existence of separate categories of modality seems\nhighly doubtful. The concepts of the possible and of the probable may be\nviewed as derivative; the notion of existence does not seem to differ\nfrom that of reality; and necessity seems in ultimate analysis to reduce\nto the concept of ground and consequence. These are points which will be\ndiscussed later.[733]\n\nAristotle\u2019s ten categories[734] are enumerated by Kant in _Reflexionen_,\nii. 522,[735] as: (1) _substantia_, _accidens_, (2) _qualitas_, (3)\n_quantitas_, (4) _relatio_, (5) _actio_, (6) _passio_, (7) _quando_, (8)\n_ubi_, (9) _situs_, (10) _habitus_; and the five post-predicaments as:\n_oppositum_, _prius_, _simul_, _motus_, _habere_. Eliminating _quando_,\n_ubi_, _situs_, _prius_, and _simul_ as being modes of sensibility;\n_actio_ and _passio_ as being complex and derivative; and also omitting\n_habitus_ (condition) and _habere_, as being too general and indefinite\nin meaning to constitute separate categories; we are then left with\n_substantia_, _qualitas_, _quantitas_, _relatio_, and _oppositum_. The\nmost serious defect in this reduced list, from the Kantian point of\nview, is its omission of causality. It is, however, a curious\ncoincidence that when substance is taken as a form of _relatio_, and\n_oppositum_ as a form of quality, we are left with the three groups,\nquality, quantity, relation. Only modality is lacking to complete Kant\u2019s\nown fourfold grouping. None the less, as the study of Kant\u2019s\n_Reflexionen_ sufficiently proves,[736] it was by an entirely different\nroute that Kant travelled to his metaphysical deduction. Watson does not\nseem to have any ground for his contention[737] that the above modified\nlist of Aristotle\u2019s categories \u201cgave Kant his starting-point.\u201d It was\nthere indeed, as the reference to Aristotle in his letter of 1772 to\nHerz shows, that he first looked for assistance, only, however, to be\ndisappointed in his expectations.\n\n=Derivative concepts.=[738]--Cf. above, pp. 66, 71-2.\n\n=I reserve this task for another occasion.=[739]--Cf. A 204 = B 249; A 13;\nabove, p. 66 ff., and below, pp. 379-80.\n\n=Definitions of categories omitted.=[740]--Cf. above, pp. 195-6, and A 241\nthere cited; also below, pp. 339-42, 404-5.\n\n=Note 1.=[741]--On this distinction between mathematical and dynamical\ncategories cf. below, pp. 345-7, 510-11.\n\n=Note 2.=[742]--This remark is inserted to meet a criticism which had been\nmade by Johann Schulze,[743] and to which Kant in February 1784 had\nreplied in terms almost identical with those of the present passage.\n\n     \u201cThe third category certainly springs from the connection of the\n     first and second, not, indeed, from their mere combination, but\n     from a connection the possibility of which constitutes a concept\n     that is a special category. For this reason the third category may\n     not be applicable in instances in which the other two apply: _e.g._\n     _one_ year, _many_ years of future time, are real concepts, but the\n     _totality_ of future years, that is, the collective unity of a\n     future eternity, conceived as entire (so to say, as completed), is\n     something that cannot be thought. But even in those cases in which\n     the third category is applicable, it always contains something more\n     than the first and the second taken separately and together, namely\n     the _derivation_ of the second from the first, a process which is\n     not always practicable. Necessity, for example, is nothing else\n     than existence, _in so far as_ it can be inferred from possibility.\n     Community is the reciprocal _causality_ of _substances_ in respect\n     of their determinations. But that determinations of one substance\n     can be produced by another substance, is something that we may not\n     simply assume; it is one of those connections without which there\n     could be no reciprocal relation of things in space, and therefore\n     no outer experience. In a word, I find that just as the conclusion\n     of a syllogism indicates, in addition to the operations of\n     understanding and judgment in the premisses, _a special operation\n     peculiar to reason_ ..., so also the third category is a special,\n     and in part original, concept. For instance, the concepts,\n     _quantum_, _compositum_, _totum_, come under the categories unity,\n     plurality, totality, but a _quantum_ thought as _compositum_ would\n     not yield the concept of totality unless the concept of the\n     _quantum_ is thought as determinable through the _composition_, and\n     in certain _quanta_, such as infinite space, that cannot be\n     done.\u201d[744]\n\nKant\u2019s assertion that in certain cases the third category is not\napplicable is misleading. His proof of the validity of the category of\nreciprocity in the third _Analogy_ really consists in showing that it is\nnecessary to the apprehension of spatial co-existence;[745] and if, as\nKant maintains, consciousness of space is necessary to consciousness of\ntime, it is thereby proved to be involved in each and every act of\nconsciousness. It is presupposed in the apprehension even of substantial\nexistence and of causal sequence. His proof that it is a unique\ncategory, distinct from the mere combination of the categories of\nsubstance and causality, does not, therefore, assume what his words in\nthe above letter would seem to imply, that it is only occasionally\nemployed. The same remark holds in regard to totality; it is presupposed\neven in the apprehension of a single year. Kant\u2019s references, both here\nand in other parts of the _Critique_,[746] to totality in its bearing\nupon the conception of infinitude, reveal considerable lack of\nclearness as to the relation in which it stands to the Idea of the\nunconditioned. Sometimes, as in this letter, he would seem to be\nidentifying them; elsewhere this confusion is avoided. In B 111 totality\nis defined as multiplicity regarded as unity, and in A 142-3 = B 182 its\nschema is defined as number. (The identification of totality with number\nhas led Kant to say in B 111 that number is not applicable in the\nrepresentation of the infinite, a much more questionable assertion than\nthat of the letter above quoted.) The statement that necessity is\nexistence in so far as it can be inferred from possibility, or that it\nis existence given through possibility, is similarly misleading. Kant\u2019s\ntrue position is that all three are necessary to the conception of any\none of the three.\n\nThus Kant\u2019s reply to Schulze, alike in his letter and in _Note 2_, fails\nto indicate with any real adequacy the true bearing of Critical teaching\nin this matter; and consequently fails to reveal the full force of his\nposition. Only in terms of totality can unity and plurality be\napprehended; only through the reciprocal relations which determine\nco-existence can we acquire consciousness of either permanence or\nsequence; only in terms of necessity can either existence or possibility\nbe defined. The third category is not derived from a prior knowledge of\nthe subordinate categories. It represents in each case a higher complex\nwithin which alone the simpler relations defined by the simpler concepts\ncan exist or have meaning.\n\n=B 113-16, \u00a7 12.=--This section, of no intrinsic importance, is an example\nof Kant\u2019s loving devotion to this \u201carchitectonic.\u201d His reasoning is\nextremely artificial, especially in its attempt to connect \u201cunity,\ntruth, and perfection\u201d with the three categories of quantity. The\n_Reflexionen_ show how greatly Kant was preoccupied with these three\nconcepts, seeking either to base a table of categories upon them (B.\nErdmann\u2019s interpretation), or to reduce them to categories (Adickes\u2019\ninterpretation). For some time Kant himself ranked with those who[747]\n\u201cincautiously made these criteria of thought to be properties of the\nthings in themselves.\u201d In _Reflexionen_, ii. 903,[748] we find the\nfollowing statement: \u201cUnity (connection, agreement), truth (quality),\ncompleteness (quantity).\u201d In ii. 916[749] Kant makes trial to connect\nthem, as conceptions of possibility, with the categories of relation. In\nii. 911 and 912 the later view, that they are logical in character and\nfunction, appears, but leads to their being set in relation to the three\nfaculties of understanding, judgment, and reason. This is conjectured\nby B. Erdmann to have been Kant\u2019s view at the time of the first edition.\nii. 915, 919, 920 present the view expounded in the section before\nus.[750] Erdmann[751] remarks that in this section Kant \u201cis settling\naccounts with certain thoughts which in the \u2019seventies had yielded\nsuggestions for the transformation of ontology into the transcendental\nanalytic.\u201d\n\n\n\n\nCHAPTER II\n\nDEDUCTION OF THE PURE CONCEPTS OF THE UNDERSTANDING\n\n\n=First edition Subjective and Objective Deductions.=--In dealing with the\ntranscendental deduction, as given in the first edition, we can make use\nof the masterly and convincing analysis which Vaihinger[752] (building\nupon Adickes\u2019 previous results, but developing an independent and quite\noriginal interpretation) has given of its inconsecutive and strangely\nbewildering argumentation. Vaihinger\u2019s analysis is an excellent example\nof detective genius in the field of scholarship. From internal evidence,\ncircumstantially supported by the _Reflexionen_ and _Lose Bl\u00e4tter_, he\nis able to prove that the deduction is composed of manuscripts,\nexternally pieced together, and representing no less than four distinct\nstages in the slow and gradual development of Kant\u2019s views. Like\ngeological deposits, they remain to record the processes by which the\nfinal result has come to be. Though they do not in their present setting\nrepresent the correct chronological order, that may be determined once\nthe proper clues to their disentanglement have been duly discovered.\nThat discovery is itself, however, no easy task; for the unexpected,\nwhile lending colour and incident to the commentator\u2019s enterprise,\nbaffles his natural expectations at every turn. The first stage is one\nin which Kant _dispenses_ with the categories, and in which, when they\nare referred to, they are taken as applying _to things in themselves_.\nThe last stage, worked out, as there is ground for believing, in the\nhaste and excitement of the final revision, is not represented in the\n_Prolegomena_ or in the second edition of the _Critique_, the author\nretracing his steps and resuming the standpoint of the stage which\npreceded it. The fortunate accident of Kant\u2019s having jotted down upon\nthe back of a dated paper the record of his passing thought (one of the\nfew _Lose Bl\u00e4tter_ that are thus datable) is the culminating incident in\nthis philosophical drama. It felicitously serves as a keystone in the\nbody of evidence supported by general reasoning.\n\nBefore becoming acquainted with Vaihinger\u2019s analysis I had observed\nKant\u2019s ascription to empirical concepts of the functions elsewhere\nallotted to the categories, but had been hopelessly puzzled as to how\nsuch teaching could be fitted into his general system. Vaihinger\u2019s view\nof it as a pre-Critical survival would seem to be the only possible\nsatisfactory solution. For the view which I have taken of Kant\u2019s\ndoctrine of the transcendental object as also pre-Critical, and for its\nemployment as a clue to the dating of passages, I am myself alone\nresponsible.\n\nThe order of my exposition will be as follows:[753]\n\nI. Enumeration, in chronological order, of the four stages which compose\nthe deduction of the first edition, and citation of the passages which\nrepresent each separate stage.\n\nII. Detailed analysis, again in chronological order, of each successive\nstage, with exposition of the views which it embodies.\n\nIII. Examination of the evidence yielded by the _Reflexionen_ and _Lose\nBl\u00e4tter_ in support of the above analysis.\n\nIV. Connected statement and discussion of the total argument of the\ndeduction.\n\n\n=I. Enumeration of the Four Stages=\n\n(1) FIRST STAGE: THAT OF THE TRANSCENDENTAL OBJECT, WITHOUT CO-OPERATION\nOF THE CATEGORIES.--This stage is represented by[754]: (_a_) II. 3 (from\nbeginning of the third paragraph to end of 3) = A 104-10; (_b_) I. \u00a7 13\n(the entire section) = A 84-92 (retained in second edition as B 116-24).\n_a_ discusses the problem of the reference of sensations to an object,\n_b_ that of the objective validity of the categories. _b_ is therefore\ntransitional to the second stage.\n\n(2) SECOND STAGE: THAT OF THE CATEGORIES, WITHOUT CO-OPERATION OF THE\nPRODUCTIVE IMAGINATION.--This stage is represented by: (_a_) I. [\u00a7 14]\n(with the exception of its concluding paragraph) = A 92-4 (retained in\nsecond edition as B 124-7); (_b_) II. (the first four paragraphs) = A\n95-7; (_c_) II. 4 (the entire section) = A 110-14.\n\n(3) THIRD STAGE: THAT OF THE PRODUCTIVE IMAGINATION, WITHOUT MENTION OF\nTHE THREEFOLD TRANSCENDENTAL SYNTHESIS.--This stage is represented by\n(_a_) III. \u03b2 (from beginning of seventh paragraph to end of\ntwelfth) = A 119-23; (_b_) III. \u03b1 (from beginning of third\nparagraph to end of sixth) = A 116-19; (_c_) I. \u00a7 14 (Concluding\nparagraph) = A 94-5; (_d_) III. \u03b4 (from beginning of sixteenth\nparagraph to end of section preceding summary) = A 126-8; (_e_)\nS(ummary) (in conclusion to III.) = A 128-30; (_f_) III. \u03b3 (from\nbeginning of thirteenth paragraph to end of fifteenth) = A 123-6; (_g_)\nI(ntroduction) (from beginning of section to end of second paragraph) =\nA 115-16; (_h_) \u00a7 10 T(ransitional to the fourth stage) = A 76-9\n(retained as B 102-4).\n\n(4) FOURTH STAGE: THAT OF THE THREEFOLD TRANSCENDENTAL SYNTHESIS.--This\nstage is represented by: (_a_) II. 1-3 (from opening of 1 to end of\nsecond paragraph in 3) = A 98-104; (_b_) II. (the two paragraphs\nimmediately preceding _a_) = A 97-8.\n\n\n=II. Detailed Analysis of the Four Stages=\n\n=First Stage.=--A 104-10; A 84-92 (B 116-24).\n\n=A 104-10=; II. \u00a7 3.--This is the one passage in the _Critique_ in which\nKant explicitly defines his doctrine of the \u201ctranscendental object\u201d; and\ncareful examination of the text shows that by it he means the _thing in\nitself_, conceived as being the object of our representations. Such\nteaching is, of course, thoroughly un-Critical; and as I shall try to\nshow, this was very early realised by Kant himself. The passages in\nwhich the phrase \u201ctranscendental object\u201d occurs are, like the section\nbefore us, in every instance of early origin. It is significant that the\ntranscendental object is not again referred to in the deduction of the\nfirst edition.[755] Though it reappears in the chapter on phenomena and\nnoumena, it does so in a passage which Kant excised in the second\nedition. The paragraphs which he then substituted make no mention of it.\nThe doctrine is of frequent occurrence in the _Dialectic_, and combines\nwith other independent evidence to show that the larger part of the\n_Dialectic_ is of early origin. That the doctrine of the transcendental\nobject is thus a pre-Critical or semi-Critical survival has, so far as I\nam aware, not hitherto been observed by any writer upon Kant. It has\ninvariably been interpreted in the light of the sections in which it\ndoes not occur, and, as thus toned down and tempered to something\naltogether different from what it really stands for, has been taken as\nan essential and characteristic tenet of the Critical philosophy. It was\nin the course of an attempt to interpret Kant\u2019s entire argument in the\nlight of his doctrine of the transcendental object that I first came to\ndetect its absence from all his later utterances. But it is important to\nrecognise that the difficulties which would result from its retention\nare quite insuperable, and would by themselves, even in the absence of\nall external evidence of Kant\u2019s rejection of it, compel us to regard it\nas a survival of pre-Critical thinking. As Vaihinger does not seem to\nhave detected the un-Critical character of this doctrine, it is the more\nsignificant that he should, on other grounds, have felt constrained to\nregard the passage in which it is expounded as embodying the earliest\nstage in the development of the deduction. He would seem to continue in\nthe orthodox view so far as to hold that though the doctrine of the\ntranscendental object is here stated in pre-Critical terms, it was\npermanently retained by Kant in altered form.\n\nThe doctrine of the transcendental object, as here expounded, is as\nfollows:\n\n     \u201cAppearances are themselves nothing but sensuous representations\n     which must not be taken as capable of existing in themselves (_an\n     sich_) with exactly the same character (_in ebenderselben Art_)\n     outside our power of representation.\u201d[756]\n\nThese sense-representations are our only possible representations, and\nwhen we speak of an object corresponding to them, we must be conceiving\nan object in general, equal to _x_.\n\n     \u201cThey have their object, but an object which can never be intuited\n     by us, and which may therefore be named the non-empirical, _i.e._\n     transcendental object = _x_.\u201d[757]\n\nThis object is conceived as being that which prevents our\nrepresentations from occurring at haphazard, necessitating their order\nin such manner that, manifold and varied as they may be, they can yet be\nself-consistent in their several groupings, and so possess that unity\nwhich is essential to the concept of an object.\n\n     \u201cThe pure concept of this transcendental object, which in fact\n     throughout all our knowledge is always one and the same, is that\n     which can alone confer upon all our empirical concepts relation in\n     general to an object, _i.e._ objective reality.\u201d[758]\n\nWhat renders this doctrine impossible of permanent retention was that it\nallowed of no objective existence mediate between the merely subjective\nand the thing in itself. On such teaching there is no room for the\nempirical object; and immediately upon the recognition of that latter\nphenomenal form of existence in space, Kant was constrained to recognise\nthat it is in the empirical object, not in the thing in itself, that the\ncontents of our representations are grounded and unified. Any other view\nmust involve the application of the categories, especially those of\nsubstance and causality, to the thing in itself. The entire empirical\nworld has still to be conceived as grounded in the non-empirical, but\nthat is a very different contention from the thesis that the thing in\nitself is the object and the sole object of our representations. The\ndoctrine of the transcendental object has thus a twofold defect: it\nadvocates an extreme subjectivism, and yet at the same time applies the\ncategories to the thing in itself.\n\nBut the latter consequence is one which could not, at the stage\nrepresented by this section, be appreciated by Kant. For, as we shall\nfind, he is endeavouring to solve the problem of the reference of\nsense-representation to an object without assumption of _a priori_\ncategories. It is in _empirical_ concepts, conditioned only by a\ntranscendental apperception, that he professes to discover the grounds\nand conditions of this objective reference. Let us follow Kant\u2019s\nargument in detail. The section opens[759] with what may be a reference\nto the _Aesthetic_, and proceeds to deal with the first of the two\nproblems cited in the 1772 letter to Herz[760]--how\n_sense_-representations stand related to their object. The exact terms\nin which this question was there formulated should be noted.\n\n     \u201cI propounded to myself this question: on what ground rests the\n     relation of that in us which we name representation (_Vorstellung_)\n     to the object. If the representation contains only the mode in\n     which the subject is affected by the object, it is easily\n     understood how it should accord (_gem\u00e4ss sei_) with that object as\n     an effect with its cause, and how [therefore] this determination of\n     our mind should be able to _represent_ something, _i.e._ have an\n     object. The passive or sensuous representations have thus a\n     comprehensible (_begreifliche_) relation to objects, and the\n     principles, which are borrowed from the nature of our soul, have a\n     comprehensible validity for all things in so far as they are to be\n     objects of the senses.\u201d[761]\n\nThus in 1772 there was here no real problem for Kant. The assumed fact,\nthat our representations are generated in us by the action of\nindependent existences, is taken as sufficient explanation of their\nbeing referred to objects.\n\nThe section of the _Critique_ under consideration shows that Kant had\ncome to realise the inadequacy of this explanation quite early, indeed\nprior to his solution of the second and further question which in that\nsame letter is spoken of as \u201cthe key to the whole secret\u201d of\nmetaphysics. On what grounds, he now asks, is a subjective idea, _even\nthough it be a sense impression_, capable of yielding consciousness of\nan object? In the letter to Herz the use of the term representation\n(_Vorstellung_) undoubtedly helped to conceal this problem. It is now\nemphasised that appearances are nothing but sense representations, and\nmust never be regarded as objects capable of existing in themselves,\nwith exactly the same character, outside our power of representation.\nNow also Kant employs, in place of the phrase \u201cin accord with,\u201d the much\nmore definite term \u201ccorresponding to.\u201d He points out that when we speak\nof an object _corresponding_ to our knowledge, we imply that it is\ndistinct from that knowledge. Consciousness of such an object must\ntherefore be acquired from some other source than the given impressions.\nIn other words, Kant is now prepared to withdraw his statement that \u201cthe\npassive or sensuous representations have an [easily] comprehensible\nrelation to objects.\u201d In and by themselves they are purely subjective,\nand can involve no such concept. The latter is a thought (_Gedanke_), a\nconcept (_Begriff_), additional to, and distinct from, the given\nimpressions. Its possibility, as regards both origin[762] and validity,\nmust be \u201cdeduced.\u201d\n\nThere then results this first and very peculiar form of the\ntranscendental deduction. That part of it which persists in the\nsuccessive stages rests upon an explicitly developed distinction between\nempirical and transcendental apperception. Kant teaches, in agreement\nwith Hume, though, as we may believe, independently of his direct\ninfluence, that there is no single empirical state of the self which is\nconstant throughout experience.[763]\n\n     \u201cThe consciousness of the self, according to the determinations of\n     our state in inner perception, is merely empirical, and always in\n     process of change.... That which has to be represented as of\n     necessity numerically identical cannot be thought as such through\n     empirical data. There must be a condition which precedes all\n     experience, and renders experience itself possible, if a\n     transcendental pre-supposition of this kind is to be rendered\n     valid.... This pure, original, unchangeable consciousness I shall\n     name transcendental apperception.\u201d[764]\n\nKant would seem to have first developed this view in a quite crude form.\nThe consciousness of the self, he seems to have held, consists in its\nawareness of its own unceasing activities. As consciousness of\n_activity_, it is entirely distinct in nature and in origin from all\napprehension of sense impressions.[765] This teaching is a natural\nextension of the doctrine of the _Dissertation_,[766] that such pure\nnotions as those of possibility, existence, necessity, substance, cause,\nare \u201cacquired by attending to the actions of the mind on the occasion of\nexperience.\u201d Kant would very naturally hold that consciousness of the\nidentity and unity of the self is obtained in a similar manner. Such,\nindeed, is the teaching of the section before us.\n\n     \u201cNo knowledge can take place in us ... without that unity of\n     consciousness which precedes all data of intuitions, and in\n     relation to which all representation of objects is alone\n     possible.\u201d[767] \u201cIt is precisely this transcendental apperception\n     that constructs out of (_macht aus_) all possible appearances,\n     which are capable of coexisting in one experience, a connection of\n     all these representations according to laws. For this unity of\n     consciousness would be impossible if the mind could not become\n     conscious, in the knowledge of the manifold, of the identity of the\n     function whereby it combines it synthetically in one knowledge.\n     Thus the mind\u2019s original and necessary consciousness of the\n     identity of itself is at the same time a consciousness of an\n     equally necessary unity of the synthesis of all appearances\n     according to concepts, _i.e._ according to rules.... For the mind\n     could not possibly think the identity of itself in the manifold of\n     its representations, and indeed _a priori_, if it did not have\n     before its eyes the identity of its action....\u201d[768]\n\nThat is to say, the self is the sole source of all unity. As a pure and\noriginal unity it precedes experience; to its synthetic activities all\nconceptual unity is due; and by reflection upon the constancy of these\nactivities it comes to consciousness of its own identity.\n\n     \u201c...even the purest objective unity, namely that of the _a priori_\n     concepts (space and time), is possible only through relation of\n     the intuitions to [transcendental apperception]. The numerical\n     unity of this apperception is therefore the _a priori_ condition of\n     all concepts, just as the manifoldness of space and of time is of\n     the intuitions of sensibility.\u201d[769]\n\nTo this consciousness of the abiding unity of the self Kant also traces\nthe notion of the transcendental object. The latter, he would seem to\nargue, is formed by analogy from the former.\n\n     \u201cThis object is nothing else than the subjective representation (of\n     the subject) itself, but made general, for I am the original of all\n     objects.\u201d[770] \u201cThe mind, through its original and underived\n     thinking, is itself the pattern (_Urbild_) of such a\n     synthesis.\u201d[771] \u201cI would not represent anything as outside me, and\n     so make [subjective] appearances into objective experience if the\n     representations were not related to something which is parallel to\n     my ego, and so in that way referred by me to another subject.\u201d[772]\n\nThese quotations from the _Lose Bl\u00e4tter_ would seem to contain the key\nto Kant\u2019s extremely enigmatic statement in A 105, that \u201cthe unity which\nthe object makes necessary can be nothing else than the formal unity of\nconsciousness in its synthesis of the manifold of its representations,\u201d\nand again in A 109, that \u201cthis relation [of representations to an\nobject] is nothing else than the necessary unity of consciousness.\u201d[773]\n\nBut this does not complete the sum-total of the functions which Kant is\nat this stage prepared to assign to apperception. It mediates our\nconsciousness of the transcendental object in still another manner,\nnamely, by rendering possible the formation of the _empirical_ concepts\nwhich unify and direct its synthetic activities. This is, indeed, the\nfeature in which this form of the deduction diverges most radically from\nall later positions. Space and time are, it would seem, regarded as\nbeing the sole _a priori_ concepts.[774] The instruments through which\nthe unity of apperception acts, and through which the thought of an\nobject becomes possible, are _empirical_ concepts. Such general concepts\nas \u201cbody\u201d or \u201ctriangle\u201d serve as rules constraining the synthetic\nprocesses of apprehension and reproduction to take place in such\nunitary fashion as is required for unitary consciousness. The notion of\nobjectivity is specified in terms of the necessities which these\nempirical concepts thus impose.\n\n     \u201cWe think a triangle as object in so far as we are conscious of the\n     combination of three straight lines according to a rule by which\n     such an intuition can at all times be generated. This unity of rule\n     determines the whole manifold and limits it to conditions which\n     make the unity of apperception possible; and the concept of this\n     unity [of rule] is the representation of the object.... All\n     knowledge demands a concept, ... and a concept is always, as\n     regards its form, something general, something that serves as a\n     rule. Thus the concept of body serves as a rule to our knowledge of\n     outer appearances, in accordance with the unity of the manifold\n     which is thought through it.... The concept of body necessitates\n... the representation of extension, and therewith of\n     impenetrability, shape, etc.\u201d[775]\n\nSuch is the manner in which Kant accounts for our concept of the\ntranscendental object. It consists of two main elements: first, the\nnotion of an unknown _x_, to which representations may be referred; and\nsecondly, the consciousness of this _x_ as exercising compulsion upon\nthe order of our thinking. The former notion is framed on the pattern of\nthe transcendental subject; it is conceived as another but unknown\nsubject. The consciousness of it as a source of external necessity is\nmediated by the empirical concepts which transcendental apperception\nalso makes possible. And from this explanation of the _origin_ of the\nconcept of the transcendental object Kant derives the proof of its\n_validity_.[776] It is indispensable for the realisation by the unitary\nself of a unitary consciousness.\n\n     \u201cThis relation [of representations to an object] is nothing else\n     than the necessary unity of consciousness, and therefore also of\n     the synthesis of the manifold, by a common (_gemeinschaftlich_)\n     functioning of the mind, which unites it in one\n     representation.\u201d[777]\n\nThrough instruments empirical in origin, and subjectively necessary, the\nnotion of an objective necessity is rendered possible to the mind.\n\nIt is not surprising that Kant did not permanently hold to this view of\nthe empirical concept. The objections are obvious. Such a view of the\nfunction of general concepts renders unintelligible their own first\nformation. For as they are empirical, they can only be acquired by\nconscious processes that do not involve them. That is to say,\nconsciousness of objects follows upon a prior consciousness in and\nthrough which concepts, such as that of body, are discovered and formed.\nYet, as the argument claims, general concepts are the indispensable\nconditions of unitary consciousness. How through a consciousness that is\nnot yet unified can general concepts be formed? Also it is difficult to\nsee how empirical concepts can be viewed as directly conditioned by, and\nas immediately due to, anything so general as pure apperception. These\nobjections Kant must have come very quickly to recognise. This was the\nfirst part of his teaching to be modified. In the immediately succeeding\nstage,[778] so far as the stages can be reconstructed from the survivals\nin the _Critique_, the empirical concepts are displaced once and for all\nby the _a priori_ categories.\n\nThe only sentences which can be regarded as possibly conflicting with\nthe above interpretation are those two (in the second last and in the\nlast paragraphs) in which the phrase \u201crules _a priori_\u201d occurs. Even\ngranting (what is at least questionable as regards the first) that the\nwords are meant to be taken together, it does not follow that Kant is\nhere speaking of categories. For contrary to his usual teaching he\nspeaks of the concept of body as a source of necessity. If so, it may\nwell, with equal looseness, be spoken of as _a priori_. That is indeed\ndone, by implication, in the second and third paragraphs, where he\nspeaks of a rule (referring to \u201cbody and triangle\u201d) as making the\nsynthesis of reproduction \u201c_a priori_ necessary.\u201d Such assertions are\ncompletely inconsistent with Kant\u2019s Critical teaching, but so is the\nentire section.\n\nThe setting in which the passage before us occurs has its own special\ninterest.[779] When Kant, as it would seem, on the very eve of the\npublication of the _Critique_, developed the doctrine of a threefold\nsynthesis culminating in a \u201csynthesis of recognition in the concept,\u201d he\nmust have bethought himself of this earlier position, and have completed\nhis subjective deduction by incorporation, probably with occasional\nalterations of phrasing, of the older manuscript. This procedure has\nbewildered even the most discerning among Kant\u2019s readers; but now,\nthanks to Vaihinger\u2019s convincing analysis, it may be welcomed as of\nilluminating interest in the historical study of Kant\u2019s development.\n\nI may here draw attention to the two important respects in which the\npositions revealed in this section continued to influence Kant\u2019s later\nteaching: namely, in the emphasis laid upon the transcendental unity of\napperception, and in the view of objectivity as involving the thought of\nthe thing in itself.\n\nThe excessive emphasis which in this first stage is laid upon the\ntranscendental unity of apperception persists throughout the later forms\nof the deduction, and, as I shall try to show, does so to the detriment\nof the argument. Though its functions are considerably diminished, they\nare still exaggerated; this is perhaps in part due to its having been in\nthis early stage regarded as in and by itself the sole ultimate ground\nof unitary experience. There were, however, two other influences at\nwork. Kant continued to employ the terminology of his earlier view, and\nin his less watchful moments was betrayed thereby into conflict with his\nconsidered teaching. But even more important was the influence of his\npersonal convictions. He was irrevocably committed in his own private\nthinking to a belief in the spiritual and abiding character of the self;\nand this belief frequently colours, in illegitimate ways, the expression\nof his views. This is especially evident in some of the alterations[780]\nof the second edition, written as they were at a time when he was\nchiefly preoccupied with moral problems.\n\nAs regards the other factor, the view adopted in regard to the nature of\nobjectivity, there is ample evidence that even after the empirical\nconcepts had been displaced by the categories Kant still continued for\nsome time (possibly for several years in the earlier and middle\n\u2019seventies) to hold to his doctrine of the transcendental object.\nPassages which expound it in this later form occur in the _Note on\nAmphiboly_ and throughout the _Dialectic_.[781] That this may not be\ntaken for his final teaching is equally certain. The entire first layer\nof the deduction of the first edition, all the relevant passages in the\nchapter on phenomena and noumena, and some of those in the _Dialectic_,\nwere omitted in the second edition; and nowhere, either in the other\nportions of the deduction of the first edition, or in the deduction of\nthe second edition, or in any passages added elsewhere in the second\nedition, is such teaching to be found.\n\nA brief statement of Kant\u2019s doctrine of the transcendental object _in\nits later form_ seems advisable at this point; it is required in order\nto complete and to confirm the interpretation which I have given of the\nearlier exposition. At the same time I shall endeavour to show that the\nsections in which the doctrine occurs, though later than the first layer\nof the deduction of the first edition, are all of comparatively early\norigin, and that they reveal not the least trace of Kant\u2019s more mature,\nphenomenalist view of the empirical world in space.\n\nWe may begin with the passages in the chapter on phenomena and noumena.\nThe meaning in which the term transcendental is employed is there made\nsufficiently clear.\n\n     \u201cThe transcendental employment of a concept in any principle\n     consists in its being referred to things _in general and in\n     themselves_.\u201d[782]\n\nThat is to say, the term transcendental, as used in the phrase\ntranscendental object, is not employed in any sense which would oppose\nit to the transcendent. In so far as the thought of the thing in itself\nis a necessary ingredient in the concept of objectivity, it is a\ncondition of apperception, and therefore of possible experience; in\nother words, _the thought of a transcendent object is one of the\ntranscendental conditions of our experience_. As Kant is constantly\ninterchanging the terms transcendent and transcendental, such an\nexplanation of the phrase is perhaps superfluous; but if any is called\nfor, the above would seem to suffice. As we shall have occasion to\nobserve,[783] other factors besides the _a priori_ must be reckoned\namong the conditions of experience; and to both types of conditions Kant\napplies the epithet transcendental.\n\nIn the chapter on phenomena and noumena Kant enquires at considerable\nlength whether the categories (meaning, of course, the pure forms of\nunderstanding, not their schematised correlates) allow of transcendental\n(_i.e._ transcendent) employment. The passages in which this discussion\noccurs[784] would seem, however, to be highly composite; many\nparagraphs, or portions of paragraphs, are of much later date than\nothers. We may therefore limit our attention to those in which the\nphrase transcendental object is actually employed, _i.e._ to those which\nappear only in the first edition.\n\n     \u201cAll our representations are referred by the understanding to some\n     object; and since appearances are merely representations, the\n     understanding refers them to a _something_ as the object of\n     sensuous intuition. But this something, thus conceived (_in so\n     fern_), is only the transcendental object; and by that is meant a\n     something = x, of which we know, and with the present constitution\n     of our understanding can know, nothing whatsoever, but which, as a\n     correlate of the unity of apperception, can serve only for the\n     unity of the manifold in sensuous intuition. By means of this\n     unity the understanding combines the manifold into the concept of\n     an object. This transcendental object cannot be separated from the\n     sense data, for nothing then remains over through which it might be\n     thought. Consequently it is not in itself an object of knowledge,\n     but only the representation of appearances under the concept of an\n     object in general which is determinable through the manifold of\n     those appearances. Precisely for this reason also the categories do\n     not represent a special object given to the understanding alone,\n     but only serve to specify the transcendental object (the concept of\n     something in general) through that which is given in sensibility,\n     in order thereby to know appearances empirically under concepts of\n     objects.\u201d[785] \u201cThe object to which I relate appearance in general\n     is the transcendental object, _i.e._ the completely indeterminate\n     thought of _something_ in general. This cannot be entitled the\n     _noumenon_ [_i.e._ the thing in itself more specifically determined\n     as being the object of a purely intelligible intuition];[786] for I\n     know nothing of what it is in itself, and have no concept of it\n     save as the object of a sensuous intuition in general, and so as\n     being one and the same for all appearances.\u201d[787]\n\nOtherwise stated, Kant\u2019s teaching is as follows. The thought of the\nthing in itself remains altogether indeterminate; it does not _specify_\nits object, and therefore yields no knowledge of it; none the less it is\na necessary ingredient in the concept of objectivity as such. The object\nas specified in terms of sense is _mere representation_; the object as\ngenuinely objective can only be thought. The correlate of the unity of\napperception is the thought of the thing in itself. This is what Kant is\nreally asserting, though in a hesitating manner which would seem to\nindicate that he is himself already more or less conscious of its\nunsatisfactory and un-Critical character.\n\nThe phrase transcendental object occurs once in the second\n_Analogy_[788] and twice in the _Note on Amphiboly_.[789] The passage in\nthe second _Analogy_ may very well, in view of the kind of subjectivism\nwhich it expounds, be of early date of writing. By transcendental object\nKant there quite obviously means the thing in itself. From the first\nreference in the _Note on Amphiboly_ no definite conclusions can be\ndrawn. The argument is too closely bound up with his criticism of\nLeibniz to allow of his own independent standpoint being properly\ndeveloped. There is, however, nothing in it which compels us to regard\nit as of late origin; and quite evidently Kant here means by the\ntranscendental object the thing in itself. The phrase _substantia\nphaenomenon_ is not, as might at first sight seem, equivalent to the\nempirical object of Kant\u2019s phenomenalist teaching. It is an adaptation\nof Leibnizian phraseology.[790] The second reference in the _Note on\nAmphiboly_ occurs in a passage which may perhaps be of later\norigin;[791] but the transcendental object is there mentioned only in\norder to afford opportunity for the statements that it cannot be thought\nthrough any of the categories, that we are completely ignorant whether\nit is within or without us, and whether if sensibility were removed it\nwould vanish or remain, and that it can therefore serve only as a\nlimiting concept. We here observe it in the very process of being\neliminated. As we shall find, Kant\u2019s teaching is ill-expressed in the\nsections on _Amphiboly_; so much so that they could not be recast\nwithout seriously disturbing the balance of his architectonic. They were\ntherefore allowed to remain unaltered in the second edition.\n\nWe may now pass to the _Dialectic_. The subjectivist doctrine of the\ntranscendental object is there expressed in a much more uncompromising\nmanner. Let us first consider the references to the transcendental\nobject in the _Paralogisms_ and in the subsequent _Reflection_. The\nphrase transcendental object occurs twice in the second _Paralogism_,\nonce in the third, twice in the fourth, and three times in the\n_Reflection_;[792] and in all these cases there is not the least\nuncertainty as to its denotation. It is taken as equivalent to the thing\nin itself, and is expounded as a necessary ingredient in the\nconsciousness of our subjective representations as noumenally grounded.\n\n     \u201cWhat matter may be as a thing in itself (transcendental object) is\n     completely unknown to us, though, owing to its being represented as\n     something external, its permanence as appearance can indeed be\n     observed.\u201d[793] \u201cWe can indeed admit that something, which may be\n     (in the transcendental[794] sense) \u2018outside us,\u2019 is the cause of\n     our outer intuitions, but this is not the object of which we are\n     thinking in the representations of matter and of corporeal things,\n     for these are merely appearances, _i.e._ mere kinds of\n     representation which are never to be met with save in us, and whose\n     actuality depends on immediate consciousness just as does the\n     consciousness of my own thoughts. The transcendental object is\n     equally unknown in respect to inner and to outer intuition.\u201d[795]\n\nHere Kant at one and the same time distinguishes between, and confounds\ntogether, representation and its empirical object. What is alone clear\nis that by the transcendental object he means simply the thing in itself\nviewed as the cause of our sensations. In A 358 it is used in a wider\nsense as also comprehending the noumenal conditions which underlie the\nconscious subject.\n\n     \u201c...this something which underlies the outer appearances and which\n     so affects our sense that it obtains the representations of space,\n     matter, shape, etc., this something viewed as noumenon (or better\n     as transcendental object) might also at the same time be the\n     subject that does our thinking....\u201d\n\nSimilarly in A 379-80:\n\n     \u201cThough the I, as represented through inner sense in time, and\n     objects in space outside me, are specifically quite distinct\n     appearances, they are not for that reason thought as being\n     different things. Neither the transcendental object which underlies\n     outer appearances, nor that which underlies inner intuition, is in\n     itself either matter or a thinking being, but is a ground (to us\n     unknown) of the appearances which supply to us the empirical\n     concepts of the former as well as of the latter kind.\u201d\n\nThe references in the _Reflection on the Paralogisms_ are of the same\ngeneral character and are equally definite.[796] A 390-1 has special\ninterest in that it explicitly states that to appearances, taken as Kant\ninvariably takes them throughout the _Paralogisms_ in the first edition\nas mere subjective representations, the category of causality, and\ntherefore by implication the category of substance, is inapplicable.\n\n     \u201cNo one could dream of asserting that that which he has once come\n     to recognise as mere representation is an outer cause.\u201d\n\nWe may now turn to the passages in the chapter on the _Antinomies_.\n\n     \u201cThe non-sensuous cause of our representations is completely\n     unknown to us, and therefore we cannot intuit it as object.... We\n     may, however, entitle the purely intelligible cause of appearances\n     in general the transcendental object.... To this transcendental\n     object we can ascribe the whole extent and connection of our\n     possible perceptions....\u201d[797]\n\nAppearances can be regarded as real only to the extent to which they are\nactually experienced. Otherwise they exist only in some unknown noumenal\nform of which we can acquire no definite concept, and which is\ntherefore really nothing to us. This, Kant declares, is true even of\nthat immemorial past of which we are ourselves the product.\n\n     \u201c...all the events which have taken place in the immense periods\n     that have preceded my own existence mean really nothing but the\n     possibility of extending the chain of experience from the present\n     perception back to the conditions which determine it in time.\u201d[798]\n\nIn other words, we may not claim that such events, empirically\nconceived, have ever actually existed in any such empirical form. A\nsimilar interpretation is given to the assertion of the present reality\nof what has never been actually experienced.\n\n     \u201cMoreover, in outcome it is a matter of indifference whether I say\n     that in the empirical progress in space I can meet with stars a\n     hundred times farther removed than the outermost now perceptible to\n     me, or whether I say that they are perhaps to be met with in\n     cosmical space even though no human being has ever perceived or\n     ever will perceive them. For though they might be given as things\n     in themselves, without relation to possible experience, they are\n     still nothing for me, and therefore are not objects, save in so far\n     as they are contained in the series of the empirical regress.\u201d[799]\n     \u201cThe cause of the empirical conditions of this process, that which\n     determines what members I shall meet with and how far by means of\n     such members I can carry out the regress, is transcendental and is\n     therefore necessarily unknown to me.\u201d[800]\n\nSuch is the form in which Kant\u2019s pre-Critical doctrine of the\ntranscendental object survives in the _Critique_.[801] It contains no\ntrace of the teaching of the objective deduction of the first and second\nedition or of the teaching of the refutation of idealism in the second\nedition. It closely resembles Mill\u2019s doctrine of the permanent\npossibilities of sensation, and is almost equally subjectivist in\ncharacter. As already noted,[802] it also lies open to the further\nobjection that it involves an illegitimate application of the categories\nto things in themselves. As Kant started from the na\u00efve and natural\nassumption that reference of representations to objects must be their\nreference to things in themselves, he also took over the current\nCartesian view that it is by an inference in terms of the category of\ncausality that we advance from a representation to its cause. The thing\nin itself is regarded as the sole true substance and as the real cause\nof everything which happens in the natural world. Appearances, being\nrepresentations merely, are wholly transitory and completely\ninefficacious. _Not only, therefore, are the categories regarded as\nvalid of things in themselves, they are also declared to have no\npossible application to phenomena._ Sense appearances do not, on this\nview, constitute the mechanical world of the natural sciences; they have\na purely subjective, more or less epi-phenomenal, existence in the mind\nof each separate observer. It was very gradually, in the process of\ndeveloping his own Critical teaching, that Kant came to realise the very\ndifferent position to which he was thereby committed. The categories,\nincluding that of causality, are pre-empted for the _empirical_ object\nwhich is now regarded as immediately apprehended; and the function of\nmediating the reference of phenomena to things in themselves now falls\nto the Ideas of Reason. The distinction between appearance and reality\nis no longer that between representations and their noumenal causes, but\nbetween the limited and relative character of the entire world in space\nand time and the unconditioned demanded by Reason. But these are\nquestions whose discussion must meantime be deferred.[803]\n\nI may now briefly summarise the evidence in favour of the view that the\ndoctrine of the transcendental object is a pre-Critical or semi-Critical\nsurvival and must not be taken as forming part of Kant\u2019s final and\nconsidered position. (I) Of the six sections in which the phrase\ntranscendental object occurs, three[804] were omitted in the second\nedition, and in the passages which were substituted for them it receives\nno mention. There are various reasons which can be suggested in\nexplanation of the retention of the other three[805] in the second\nedition. The _Note on Amphiboly_ was too unsatisfactory as a whole to\nencourage Kant to improve upon it in detail. The other two are outside\nthe limit at which Kant thought good to terminate all attempts to\nimprove, whether in major or in minor matters, the text of the first\nedition.[806] To have recast the _Antinomies_ as he had recast the\n_Paralogisms_ would have involved alterations much too extensive. Also,\nthere were no outside polemical influences--or at least none acting\nquite directly--such as undoubtedly reinforced his other reasons for\nrevising the _Paralogisms_. (2) Secondly, the transcendental object is\nnot mentioned in the later layers of the deduction of the first edition,\nnor in the deduction of the second edition, nor in any passage or note\nadded in the second edition. That Kant should thus suddenly cease to\nemploy a phrase to which he had accustomed himself is the more\nsignificant in view of his conservative preference for the adapting of\nfamiliar terminology to new uses. It can only be explained as due to his\nrecognition of the completely untenable character of the teaching to\nwhich it had given expression. As the object of knowledge is always\nempirical, it can never legitimately be called transcendental. (3)\nThirdly, the general teaching of the passages in which the phrase\ntranscendental object occurs is by itself sufficient proof of their\nearly origin. They reveal not the least trace of the deepened insight of\nhis final standpoints. As we know, it was certain difficulties involved\nin the working out of the objective deduction that delayed the\npublication of the _Critique_ for so many years; and the sections which\ndeal with these difficulties contain Kant\u2019s maturest teaching. In them\nhe seems to withdraw definitely from the positions to which he had\nunwarily committed himself by his un-Critical doctrine of the\ntranscendental object. I now pass to the second section constitutive of\nthe first stage.\n\n=A 84-92=B 116-24, I. \u00a7 13.=--Just as in II. \u00a7 3 Kant deals solely with\nthe first of the two questions formulated in the letter of 1772 to\nHerz--the reference of _sense_-representations to an object,--so in I. \u00a7\n13 he raises only the second--that of the objective validity of\n_intellectual_ representations (now spoken of as pure concepts of\nunderstanding, or pure _a priori_ concepts, and only in one sentence as\ncategories). And just as in the former section he carries the problem a\nstep further, yet without attaining to the true Critical position, so in\nthis latter he still assumes that it is the application of these pure\nconcepts to real independent objects, _i.e._ to _things in themselves_,\nwhich calls for justification. We must again consider the exact terms in\nwhich this problem is formulated in the letter to Herz.[807]\n\n     \u201cSimilarly, if that in us which is called a representation, were\n     active in relation to the object, that is to say, if the object\n     itself were produced by the representation (as on the view that the\n     ideas in the Divine Mind are the archetypes of things), the\n     conformity of representations with objects might be understood. We\n     can thus render comprehensible at least the possibility of two\n     kinds of intelligence--of an _intellectus archetypus_, on whose\n     intuition the things themselves are grounded, and of an\n     _intellectus ectypus_ which derives the data of its logical\n     procedure from the sensuous intuition of things. But our\n     understanding (leaving moral ends out of account) is not the cause\n     of the object through its representations, nor is the object the\n     cause of its intellectual representations (_in sensu reali_).\n     Hence, the pure concepts of the understanding cannot be abstracted\n     from the data of the senses, nor do they express our capacity for\n     receiving representations through the senses. But, whilst they have\n     their sources in the nature of the soul, they originate there\n     neither as the result of the action of the object upon it, nor as\n     themselves producing the object. In the _Dissertation_ I was\n     content to explain the nature of these intellectual representations\n     in a merely negative manner, viz. as not being modifications of the\n     soul produced by the object. But I silently passed over the further\n     question, how such representations, which refer to an object and\n     yet are not the result of an affection due to that object, can be\n     possible. I had maintained that the sense representations represent\n     things as they appear, the intellectual representations things as\n     they are. But how then are these things given to us, if not by the\n     manner in which they affect us? And if such intellectual\n     representations are due to our own inner activity, whence comes the\n     agreement which they are supposed to have with objects, which yet\n     are not their products? How comes it that the axioms of pure reason\n     about these objects agree with the latter, when this agreement has\n     not been in any way assisted by experience? In mathematics such\n     procedure is legitimate, because its objects only _are_ quantities\n     for us, and can only be represented as quantities, in so far as we\n     can generate their representation by repeating a unit a number of\n     times. Hence the concepts of quantity can be self-producing, and\n     their principles can therefore be determined _a priori_. But when\n     we ask how the understanding can form to itself completely _a\n     priori_ concepts of things in their _qualitative_ determination,\n     with which these things must of necessity agree, or formulate in\n     regard to their possibility principles which are independent of\n     experience, but with which experience must exactly conform,--we\n     raise a question, that of the origin of the agreement of our\n     faculty of understanding with the things in themselves, over which\n     obscurity still hangs.\u201d[808]\n\nThe section before us represents the same general standpoint as that\ngiven in the above letter. Here, too, it is the validity of the _a\npriori_ concepts in reference to _things in themselves_ that is under\nconsideration. The implication of Kant\u2019s argument is that the\ncategories, being neither determinable nor discoverable by means of\nexperience, will only apply to appearances if they determine, or rather\nreveal, the actual non-experienced nature of things in themselves. These\npure concepts, it is implied, owing to their combined _a priori_ and\nintellectual characteristics, make this inherent claim. Either they are\naltogether empty and illusory, or such unlimited validity must be\ngranted to them. Kant, that is to say, still holds, as in the\n_Dissertation_, that sense-representations reveal things as they\n_appear_, intellectual representations things as they _are_.\n\n     \u201cWe have either to surrender completely all claims to judgments of\n     pure reason, in the most esteemed of all fields, that which extends\n     beyond the limits of all possible experience, or we must bring this\n     Critical investigation to perfection.\u201d[809]\n\nThe pure concepts, unlike space, \u201capply to objects generally, apart from\nthe conditions of sensibility.\u201d[810] But here also, as in the letter to\nHerz, the strange and problematic character of such knowledge is clearly\nrecognised.\n\nKant\u2019s discussion of the concept of causality in A 90 may seem to\nconflict with the above contention--that it is its applicability to\nthings in themselves which Kant is considering. But this difficulty\nvanishes if we bear in mind that here, as in the _Dissertation_, there\nis no such distinction as we find in Kant\u2019s later more genuinely\nphenomenalist position, between the objects causing our sensations and\n_things in themselves_.[3] The purely intelligible object, supposed to\nremain after elimination of the empirical and _a priori_ sensory\nfactors, is the thing in itself. The objects apprehended through sense\nare real, only not in their sensuous form.\n\nThere are two connected facts which together may perhaps be taken as\nevidence that I. \u00a7 13 is later than II. 3 _b_. Intellectual concepts are\nreinstated alongside the _a priori_ concepts of space and time. Kant has\nevidently in the meantime given up the attempt to construe the former as\nempirical in origin. That that attempt was earlier in time would seem to\nbe proved by the further fact, that the _a priori_ concepts are here\nviewed as performing the same kind of function as that ascribed in II. 3\n_b_ to concepts that are empirical. They are conditions of the\n\u201csynthetic unity of thought.\u201d[811] This view of the function of concepts\nis certainly fundamental and important, and Kant permanently retained it\nfrom his previous abortive method of \u2018deduction.\u2019 But it was a long step\nfrom the discovery of the distinction between empirical and _a priori_\nconcepts to its fruitful application. That involved appreciation of the\nfurther fact that the two problems, separately stated in the letter to\nHerz and separately dealt with in II. 3 _b_ and in I. \u00a7 13--the problem\nof the relation of _sense_-representations, and the problem of the\nrelation of _intellectual_ representations, to an object,--are indeed\none and the same, soluble from one and the same standpoint, by one and\nthe same method of deduction, namely, by reference to the possibility of\nexperience. Only in and through relation to an object can\nsense-representations be apprehended; and only as conditions of such\nsense-experience are the categories objectively valid. Relation to an\nobject is constituted by the categories, and is necessary in reference\nto sense-representations, because only thereby is consciousness of any\nkind possible at all.\n\nThat this truly Critical position had not been attained when I. \u00a7 13 was\nwritten,[812] is shown not only by its concentration on the single\nproblem of the validity of _a priori_ concepts, but also by its repeated\nassertion that representations can be consciously apprehended\nindependently of all relation to the faculty of understanding. The\ndirectly counter assertion appears, however, in the sections (I. \u00a7 14,\nII.: first four paragraphs) which immediately follow in the text of the\n_Critique_--indicating that in the period represented by these latter\nthe revolutionary discovery, the truly Copernican hypothesis, had at\nlast been achieved. They constitute the second stage, and to it we may\nnow proceed.\n\n\n=Second Stage.=--A 92-4 = B 124-7; A 95-7; A 110-14.\n\n=A 92-4=, I. \u00a7 14 (with the exception of the concluding classification of\nmental powers).--This section makes a fresh start; it stands in no\nnecessary relation to any preceding section. The problem is still\nformulated, in its opening sentences, in terms reminiscent of the letter\nto Herz; but otherwise the standpoint is entirely new, and save for the\nwording of a single sentence (A 93: \u201cif not intuited, yet\u201d), is\ngenuinely Critical. The phrase \u201cpossibility of experience\u201d now appears,\nand is at once assigned the central r\u00f4le. The words \u201cif not intuited,\nyet\u201d in A 93 may possibly have been inserted later in order to tone down\nthe flagrant contradiction with the preceding paragraphs. In any case,\neven this qualification is explicitly retracted in A 94.\n\n=A 95-7.=--The same standpoint appears in the first three paragraphs of\nSection II. The categories are \u201cthe _a priori_ conditions on which the\npossibility of experience depends.\u201d[813] By the categories alone \u201ccan an\nobject be thought.\u201d[814] The further important point that only in their\nempirical employment do the categories have use and meaning is\nexcellently developed.\n\n     \u201cAn _a priori_ concept not referring to experience would be the\n     logical form only of a concept, but not the concept itself by which\n     something is thought.\u201d[815]\n\n=A 110-14, II. 4.=--In this section also the argument starts afresh,\nindicating (if such evidence were required) that, like I. \u00a7 14, it must\nhave been written independently of its present context. But the argument\nis now advanced one step further. The categories are recognised as\nsimultaneously conditioning both unity of consciousness and objectivity.\n\n     \u201cThere is but one experience ... as there is but one space and one\n     time....\u201d \u201cThe _a priori_ conditions of a possible experience are\n     at the same time conditions of the possibility of objects of\n     experience\u201d[816] \u201c...the necessity of these categories rests on the\n     relation which our whole sensibility, and with it also all possible\n     appearances, have to the original unity of apperception....\u201d[817]\n\nNow also it is _emphasised_ that save in and through _a priori_ concepts\nno representations can exist for consciousness.\n\n     \u201cThey would then belong to no experience, would be without an\n     object, a blind play of representations, less even than a\n     dream.\u201d[818] They \u201cwould be to us the same as nothing.\u201d[819]\n\nThe wording is still not altogether unambiguous, but the main point is\nmade sufficiently clear.\n\nThese paragraphs are the earliest in which traces of a genuine\nphenomenalism can be detected. The transcendental object, one and the\nsame for all our knowledge, is not referred to. \u2018Objects\u2019 (in the\nplural) is the term which is used wherever the context permits. The\nempirical object is thus made to intervene between the thing in itself\nand the subjective representations. But the distinction between\nempirical objects and subjective representations on the one hand, and\nbetween empirical objects and things in themselves on the other, is not\nyet drawn in any really clear and definite manner.\n\nA similar phenomenalist tendency crops out in Kant\u2019s distinction[820]\nbetween objective affinity and subjective association.\n\n     \u201cThe ground of the possibility of the association of the manifold,\n     so far as it lies in the object, is named the affinity of the\n     manifold.\u201d\n\nNone the less Kant\u2019s subjectivism finds one of its most decided\nexpressions in A 114.\n\n=Third Stage.=--A 119-23 = III. \u03b2; A 116-19 = III. \u03b1; A\n94-5 = I. \u00a7 14 C(oncluding paragraph); A 126-8 = III. \u03b4; A\n128-30 = S(ummary); A 123-6 = III. \u03b3; A 115-16 = III.\nI(ntroduction); A 76-9 (B 102-4) = \u00a7 10 T(ransition to fourth stage).\n\n=A 119-23, III. \u03b2= (from the beginning of the seventh paragraph\nto the end of the twelfth). The doctrine of objective affinity already\ndeveloped in the above sections is now made to rest upon a new faculty,\nthe productive imagination. As Vaihinger remarks, the wording of this\nsection would seem to indicate that it is Kant\u2019s first attempt at\nformulating that new doctrine. He has not as yet got over his own\nsurprise at the revolutionary nature of the conclusions to which he\nfeels himself driven by the exigencies of Critical teaching. He finds\nthat it is deepening into consequences which may lead very far from the\ncurrent psychology and from his own previous views regarding the nature\nand conditions of the knowing process and of personality. As evidence\nthat this section was not written continuously with II. 4,[1] we have\nthe further fact that though the doctrine of objective affinity is dwelt\nupon, it is described afresh, with no reference to the preceding\naccount. Also, the empirical processes of apprehension and reproduction,\nalready mentioned in A 104-10, are now ascribed to the empirical\nimagination which is carefully distinguished from the productive.\n\nIII. \u03b1 repeats \u201cfrom above\u201d the argument given in III. \u03b2 \u201cfrom below.\u201d\nIt insists upon the close connection between the categories (first\nintroduced in II. 4[821]) with the productive imagination of III. \u03b2.\n\nVaihinger places III. \u03b4 next in order, on account of the\nconnection of its argument with III. \u03b1.[822] But it dwells only\nupon the chief outcome of the total argument, viz. that the orderliness\nof nature is due to understanding. That productive imagination is not\nmentioned, is taken by Vaihinger to signify Kant\u2019s recognition that it\ncan be postulated only hypothetically, and that as doctrine it is not\nabsolutely essential to the strict deduction.\n\nS summarises the entire argument, and in it \u201cpure imagination\u201d receives\nmention.\n\nWithin this third stage III. \u03b3 is subsequent to the above four\nsections. For it carries the doctrine of productive imagination one step\nfurther. In III. \u03b2, III. \u03b1, and S, productive\nimagination has been treated merely as an auxiliary function of pure\nunderstanding.\n\n     \u201cThe unity of apperception in relation to the synthesis of\n     imagination is the _understanding_; and the same unity with\n     reference to the transcendental synthesis of the imagination is the\n     _pure understanding_.\u201d[823]\n\nIt is now treated as a separate and distinct faculty. So far from being\na function of understanding, its synthesis \u201cby itself, though carried\nout _a priori_, is always sensuous.\u201d[824] It is\n\n     \u201cone of the fundamental faculties of the human soul.... The two\n     extreme ends, sensibility and understanding, must be brought into\n     connection with each other by means of this transcendental function\n     of imagination.\u201d[825]\n\nIn this section there also appears a new element which would seem to\nconnect it with the next following stage, namely, the addition to the\nseries, apprehension, association, and reproduction, of the further\nprocess, recognition. As here introduced it is extremely ambiguous in\ncharacter. It is counted as being empirical, and yet as containing _a\npriori_ concepts. This decidedly hybrid process would seem to represent\nKant\u2019s first formulation of the even more ambiguous process, which\ncorresponds to it in the fourth stage.\n\nIn III. I recognition is again mentioned, but this time in a form still\nmore akin to its treatment in the fourth stage. It is not recognition\nthrough categories, but, as a form in apperception, is the\n\n     \u201cempirical consciousness of the identity of the reproductive\n     representations with the appearances by which they were\n     given.\u201d[826]\n\nIn all other respects, however, the above six sections agree (along with\nI. \u00a7 14 C) in holding to a threefold division of mental powers:\nsensibility, imagination, and apperception. This third stage is thereby\nmarked off sufficiently clearly from the second stage in which pure\nimagination is wanting, and from the fourth stage in which it is\ndissolved into a threefold _a priori_ synthesis.\n\nIn both I. \u00a7 14 C and in III. I the classification which underlies the\nthird stage is explicitly formulated. Their statements harmoniously\ncombine to yield the following tabular statement:\n\n1. The _synopsis_ of the manifold--_a priori_ through sense, _i.e._ in\npure intuition.\n\n2. The synthesis of this manifold--through pure transcendental\nimagination.\n\n3. The unity of this synthesis--through pure original transcendental\napperception.\n\nAt this point Vaihinger adds to the above section the earlier passage \u00a7\n10 T.[827] It is even more definitely than III. \u03b3 and III. I\ntransitional to the fourth stage. It must be classed within the third\nstage, as it holds to the above threefold classification. But it\nmodifies that classification in two respects. First, in that it does not\nemploy the term _synopsis_, but only speaks of pure intuition as\nrequired to yield us a manifold. The term synopsis, as used by Kant, is,\nhowever, decidedly misleading.[828] His invariable teaching is that all\nconnection is due to synthesis. By synopsis, therefore, which he\ncertainly does not employ as synonymous with synthesis, can be meant\nonly apprehension of external side-by-sideness. It never signifies\nanything except apprehension of the lowest possible order. Kant\u2019s\nomission of the term, therefore, tends to clearness of statement.\nSecondly, the classification is also modified by the substitution of\nunderstanding for the unity of apperception. Apperception is, however,\nso obscurely treated in all of the above sections, that this cannot be\nregarded as a vital alteration. What is new in this section, and seems\nto connect it in a curious and interesting manner with sections in the\nfourth stage, is its doctrine of\n\n     \u201ca manifold of _a priori_ sensibility.\u201d \u201cSpace and time contain a\n     manifold of pure _a priori_ intuition.\u201d[829]\n\nThat is, in this connection, an entirely new doctrine. In all the\nprevious sections of the deduction (previous in the assumed order of\noriginal writing) the manifold supplied through intuition is taken as\nbeing empirical, and as consisting of sensations. Kant here also adds\nthat the manifold, \u201cwhether given empirically or _a priori_,\u201d[830] must\nbe synthesised before it can be known.\n\n     \u201cThe spontaneity of our thought requires that this manifold [of\n     pure _a priori_ intuition] should be run through in a certain\n     manner, taken up, and connected, in order that a knowledge may be\n     formed out of it. This action I call synthesis.\u201d\n\n=Fourth Stage.--A 98-104; A 97-8.=--As already noted, there are in Kant\ntwo persistent but conflicting interpretations of the nature of the\nsynthetic processes exercised by imagination and understanding, the\nsubjectivist and the phenomenalist.[831] Now, on the former view,\nimagination is simply understanding _at work_. In other words,\nimagination is merely the active synthesising side of a faculty whose\ncomplementary aspect appears in the logical unity of the concept. From\nthis point of view the transcendental and the empirical factors may be\ntaken as forming a single series. The transcendental and the empirical\nprocesses will vary together, some form of transcendental activity\ncorresponding to every fundamental form of empirical activity and _vice\nversa_. Such an inference only follows _if_ the subjectivist standpoint\nbe accepted to the exclusion of the phenomenalist point of view. But\nsince Kant constantly alternates between them, and never quite\ndefinitely formulates them in their distinction and opposition; since,\nin fact, they were rather of the nature of obscurely felt tendencies\nthan of formulated standpoints, it is quite intelligible that an\ninference derived from the one should be drawn even at the very time\nwhen the other is being more explicitly developed. This, it would seem,\nis what actually happened. When we come to consider the evidence\nderivable from the _Reflexionen_ and _Lose Bl\u00e4tter_, we shall find\nsupport for the view that after January 1780, on the very eve of the\npublication of the _Critique_, while the revolutionary, phenomenalist\nconsequences of the Critical hypothesis were becoming clearer to him, he\nunguardedly allowed the above inference to lead him to recast his\nprevious views in a decidedly subjectivist manner. The view that\ntranscendental imagination has a special and unique activity altogether\ndifferent in type from any of its empirical processes, namely, the\n\u201cproductive,\u201d is now allowed to drop; and in place of it Kant develops\nthe view that transcendental functions run exactly parallel with the\nempirical processes of apprehension, reproduction, and recognition.\nAccordingly, in place of the classification presented in the third\nstage, we find a new and radically different one introduced into the\ntext, without the least indication that Kant\u2019s standpoint has meantime\nchanged. It is given in A 97:\n\nA. =Synopsis= of the manifold through sense.\n\nB. =Synthesis.=\n\n    1. Synthesis of apprehension of representations in [inner] intuition.\n    2. Synthesis of reproduction of representations in imagination.\n    3. Synthesis of recognition of representations in the concept.\n\nAnd Kant adds in explanation that \u201cthese point to three subjective\nsources of knowledge which make the understanding itself possible, and\nwhich in so doing make all experience possible, in so far as it is an\nempirical product of the understanding.\u201d What, now, are these three\nsubjective sources of knowledge? They certainly are not those classified\nin the table of the third stage. _A_ roughly coincides with its first\nmember; consequently _B_ 1 is left without proper correlate. _B_ 2 is\naltogether different from the previous synthesis of imagination, for in\nthe earlier table transcendental imagination is regarded as being solely\nproductive, _never_ reproductive.[832] It is now asserted to be\nreproductive--a contradiction of one of his own most emphatic\ncontentions, which can only be accounted for by some such explanation as\nwe are here stating. Nothing is lacking as regards explicitness in the\nstatement of this new position. \u201c...the _reproductive_ synthesis of\nimagination belongs to the transcendental acts of the soul, and, in\nreference to it [viz. to the reproductive synthesis], we will call this\npower too the transcendental power of the imagination.\u201d[833] Lastly,\neven B 3 does not coincide with the pure apperception of the other\ntable. B 3 is more akin to the recognition which in the third stage is\ndeclared to be always empirical. In any case, it is recognition _in the\nconcept_; and though that may ultimately involve and condition\ntranscendental apperception, it remains, in the manner in which it is\nhere developed by Kant, something very different. But this is a point to\nwhich we shall return. There is an added complication, running through\nthis entire stage, which first requires to be disentangled. The\ntranscendental syntheses are declared to condition the pure\nrepresentations of space and time no less than those of\nsense-experience.\n\n     \u201cThis synthesis of apprehension also must be executed _a priori_,\n     _i.e._ in reference to representations which are not empirical. For\n     without it we could not have the _a priori_ representations either\n     of space or of time, since these can be generated only through the\n     synthesis of the manifold which sensibility presents in its\n     original receptivity. Thus we have a pure synthesis of\n     apprehension\u201d[834] \u201c...if I draw a line in thought or desire to\n     think of the time from one noon to another, or merely represent to\n     myself a certain number, I must, firstly, apprehend these manifold\n     representations one after the other. But if the preceding\n     representations (the first parts of the line, the antecedent parts\n     of time or the units serially represented) were always to drop out\n     of my thought, and were not reproduced when I advance to those that\n     follow, no complete representation, and none of all the\n     aforementioned thoughts, not even the purest and first basal\n     representations of space and time, could ever arise.\u201d[835]\n\nThis, as Vaihinger remarks, is a point of sufficient importance to\njustify separate treatment. But it is introduced quite incidentally by\nKant, and obscures quite as much as it clarifies the main argument.\n\nIt is convenient to start with the second synthesis. Kant\u2019s argument is\nmuch clearer in regard to it than in regard to the other two. He\ndistinguishes between empirical and transcendental reproduction.\nReproduction in ordinary experience, in accordance with the laws of\nassociation, is merely empirical. The _de facto_ conformity of\nappearances to rules is what renders such empirical reproduction\npossible;\n\n     \u201c...otherwise our faculty of empirical imagination would never find\n     any opportunity of action suited to its capacities, and would\n     remain hidden within the mind as a dead, and to us unknown\n     power.\u201d[836]\n\nKant proceeds to argue, consistently with his doctrine of objective\naffinity, that empirical reproduction is itself transcendentally\nconditioned. The form, however, in which this argument is developed is\npeculiar to the section before us, and is entirely new.\n\n     \u201cIf we can show that even our purest _a priori_ intuitions yield no\n     knowledge, save in so far as they contain such connection of the\n     manifold as will make possible a thoroughgoing synthesis of\n     reproduction, this synthesis of the imagination must be grounded,\n     prior to all experience, on _a priori_ principles; and since\n     experience necessarily presupposes that appearances can be\n     reproduced, we shall have to assume a pure transcendental synthesis\n     of the imagination as conditioning even the possibility of all\n     experience.\u201d[837]\n\nIn the concluding paragraph Kant makes clear that he regards this\n_transcendental_ activity as being exercised in a twofold manner: in\nrelation to the _empirically_ given manifold as well as in relation to\nthe _a priori_ given manifold. How this transcendental activity is to be\ndistinguished from the empirical is not further explained. I discuss\nthis point below.[838]\n\nThe argument of the section on the synthesis of _apprehension_, to which\nwe may now turn back, suffers from serious ambiguity. It is not clear\nwhether a distinction, analogous to that between empirical and\ntranscendental reproduction, is being made in reference to apprehension.\nThe actual wording of its two last paragraphs would lead to that\nconclusion. That, however, is a view which would seem to be excluded by\nthe wider context. Kant is dealing with the synthesis of apprehension in\n_inner_ intuition, _i.e._ in time. By the fundamental principles of his\nteaching such intuition must always be transcendental. Empirical\napprehension can only concern the data of the special senses. The\nprocess of apprehension referred to in the middle paragraph must\ntherefore itself be transcendental.\n\nBut it is in dealing with the synthesis of _recognition_ that the\nargument is most obscure. It is idle attempting to discover any possible\ndistinction between an empirical and a transcendental process of\nrecognition. For the transcendental process here appears as being the\nconsciousness that what we are thinking now is the same as what we\nthought a moment before; and it is illustrated not by reference to the\npure intuitions of space and time, but only by the process of counting.\nIt may be argued that empirical recognition is mediated by\ntranscendental factors--by pure concepts and by apperception. But unless\nwe are to take transcendental recognition as synonymous with\ntranscendental apperception, which Kant\u2019s actual teaching does not seem\nto justify us in doing, such considerations will not enable us to\ndistinguish two forms of recognition. Apart, however, from this\ndifficulty, there is the further one that the concepts in and through\nwhich the recognition is executed are here described as being empirical.\nThe only key that will solve the mystery of this extraordinary section,\nhopelessly inexplicable when viewed as a single continuous whole, is, it\nwould seem, the theory of Vaihinger, namely,[839] that from the third\nparagraph onwards (already dealt with as forming the first stage of the\ndeduction) Kant is making use of manuscript which represents the\n_earliest_ form in which his explanation of the consciousness of objects\nwas developed, with the strange result that this section is a\ncombination of the latest and of the earliest forms of the deduction.\nWhile seeking to make out a parallelism between the empirical, conscious\nactivities of imagination and understanding on the one hand, and its\ntranscendental functions on the other, he must have bethought himself of\nthe earlier attempt to explain consciousness of objects through\nempirical concepts conditioned by transcendental apperception, and so\nhave attempted to expound the third form of synthesis by means of it.\nAs thus extended it involves a distinction between transcendental and\nempirical apperception, and upon that the discussion, so far as it\nconcerns anything akin to recognition, altogether turns. But there is\nnot the least further mention of recognition itself. As transcendental,\nit cannot be taken as the equivalent of empirical apperception; and as a\nsynthesis through concepts, can hardly coincide with pure apperception.\nThe title of the section, \u201cthe synthesis of recognition in the concept,\u201d\nis thus no real indication of the astonishing fare prepared for the\nreader. The doctrine of a threefold synthesis seems to have occurred to\nKant on the very eve of the publication of the _Critique_. The passage\nexpounding it may well have been hurriedly composed, and when unforeseen\ndifficulties accumulated, especially in regard to recognition as a\ntranscendental process, Kant must have resolved simply to close the\nmatter by inserting the older manuscript.\n\n\n=III. Evidence yielded by the \u201cReflexionen\u201d and \u201cLose Bl\u00e4tter\u201d in support\nof the above analysis.=\n\nThe evidence, derived by Vaihinger from the _Reflexionen_ and _Lose\nBl\u00e4tter_, briefly outlined, is as follows.[840] (1) In the _Reflexionen\nzur Anthropologie_ relevant passages are few in number, and represent a\nstandpoint very close to that of the 1770 _Dissertation_. Imagination is\ntreated only as an empirical faculty.[841] Recognition, which is only\nonce mentioned,[842] is also viewed as merely empirical. The\nunderstanding is spoken of as the faculty through which objects are\nthought.[843] The categories are not mentioned, and it is stated that\nthe understanding yields only ideas of reflection. \u201cAll knowledge of\nthings is derived, as regards its matter, from sensation--the\nunderstanding gives only ideas of reflection.\u201d[844] So far, these\n_Reflexionen_ would seem to coincide, more or less, with the first stage\nof the deduction. They contain, however, no reference to\n_transcendental_ apperception; and are therefore regarded by Vaihinger\nas representing a still earlier standpoint.\n\n(2) In the _Reflexionen zur Kritik der reinen Vernunft_ there is a very\nlarge and valuable body of relevant passages. No. 925 must be of the\nsame date as the letter of 1772 to Herz; it formulates its problem in\npractically identical terms.[845] Nos. 946-52 and 955 may belong to the\nperiod of the first stage. For though the doctrine of the transcendental\nobject as the opposite counterpart of the transcendental subject is not\nmentioned, the spiritualist view of the self is prominent. In No. 946 it\nis asserted that the representation of an object is \u201cmade by us through\nfreedom.\u201d\n\n     \u201cFree actions are already given _a priori_, namely our own.\u201d[846]\n     \u201cTo pass universal objective judgments, and to do so apodictically,\n     reason must be free from subjective grounds of determination. For\n     were it so determined the judgment would be merely accidental,\n     namely in accordance with its subjective cause. Thus reason is\n     conscious _a priori_ of its freedom in objectively necessary\n     judgments in so far as it apprehends them as exclusively grounded\n     through their relation to the object.\u201d[847] \u201cTranscendental freedom\n     is the necessary hypothesis of all rules, and therefore of all\n     employment of the understanding.\u201d[848] \u201cAppearances are\n     representations whereby we are affected. The representation of our\n     free self-activity does not involve affection, and accordingly is\n     not appearance, but apperception.\u201d[849]\n\nIt is significant that the categories receive no mention.\n\nAlmost all the other _Reflexionen_ would seem to have originated in the\nperiod of the second stage of the deduction; but they still betray a\nstrong spiritualist bias.\n\n     \u201cImpressions are not yet representations, for they must be related\n     to something else which is an action. Now the reaction of the mind\n     is an action which relates to the impression, and which if taken\n     alone[850] may in its special forms receive the title\n     categories.\u201d[851] \u201cWe can know the connection of things in the\n     world only if we produce it through a universal action, and so out\n     of a principle of inner power (_aus einem Prinzip der inneren\n     Potestas_): substance, ground, combination.\u201d[852]\n\nThese _Reflexionen_ recognise only the categories of relation,[853] and\nmust therefore be prior to the twelvefold classification. There is not\nthe least trace of the characteristic doctrines of the third and fourth\nstages of the deduction, viz. of the transcendental function of the\nimagination or of a threefold transcendental synthesis. The nature of\napprehension is also most obscure. It is frequently equated with\napperception.\n\n(3) The _Lose Bl\u00e4tter aus Kants Nachlass_ (Heft I.) contains fragments\nwhich also belong to the second stage of the deduction, but which would\nseem to be of somewhat earlier date than the above _Reflexionen_.[854]\nThey have interesting points of contact with the first stage. Thus\nthough the phrase transcendental object does not occur in them, the\nobject of knowledge is equated with _x_, and is regarded in the manner\nof the first stage as the opposite counterpart of the unity of the\nself.[855] These fragments belong, however, to the second stage in\nvirtue of their recognition of the _a priori_ categories of relation.\nThere is also here, as is in the _Reflexionen_, great lack of clearness\nregarding the nature of apprehension; and there is still no mention of\nthe transcendental faculty of imagination. Fragment 8 is definitely\ndatable. It covers the free spaces of a letter of invitation dated May\n20, 1775.[856] Fragment B 12[857] belongs to a different period from the\nabove. This is sufficiently evident from its contents; but fortunately\nthe paper upon which it is written--an official document in the\nhandwriting of the Rector of the Philosophical Faculty of\nK\u00f6nigsberg--enables us to decide the exact year of its origin. It is\ndated January 20, 1780. The fragment must therefore be subsequent to\nthat date. Now in it transcendental imagination appears as a third\nfaculty alongside sensibility and understanding, and a distinction is\ndefinitely drawn between its empirical and its transcendental\nemployment. The former conditions the synthesis of apprehension; the\nlatter conditions the synthetic unity of apperception. It further\ndistinguishes between reproductive and productive imagination, and\nascribes the former exclusively to the empirical imagination. In all\nthese respects it stands in complete agreement with the teaching of the\nthird stage of the deduction. The fact that this fragment is subsequent\nto January 1780 would seem to prove that even at that late date Kant was\nstruggling with his deduction.[858] But the most interesting of all\nVaihinger\u2019s conclusions has still to be mentioned. He points out that at\nthe time when this fragment was composed Kant had not yet developed the\ndoctrine characteristic of the fourth stage, namely, of a threefold\ntranscendental synthesis. Moreover, as he observes, the statement which\nit explicitly contains, that reproductive imagination is always\nempirical, is inconsistent with any such doctrine. The teaching of the\nfourth stage must consequently be ascribed to an even later date.[859]\n\n(4) The _Lose Bl\u00e4tter_ (Heft II.), though almost exclusively devoted to\nmoral and legal questions, contain in E 67[860] a relevant passage which\nReicke regards as belonging to the \u2018eighties, but which Adickes and\nVaihinger agree in dating \u201cshortly before 1781.\u201d On Vaihinger\u2019s view it\nis a preliminary study for the passages of the fourth stage of the\ndeduction. But such exact dating is not essential to Vaihinger\u2019s\nargument. It is undoubtedly quite late, and contains the following\nsentence:\n\n     \u201cAll representations, whatever their origin, are yet ultimately as\n     representations modifications of inner sense, and their unity must\n     be viewed from this point of view. A spontaneity of synthesis\n     corresponds to their receptivity: either of apprehension as\n     sensations or of reproduction as images (_Einbildungen_) or of\n     recognition as concepts.\u201d\n\nThis is the doctrine from which the deduction of the first edition\n_starts_; it was, it would seem, _the last to be developed_.[861] That\nwe find no trace of it in the _Prolegomena_, and that it is not only\neliminated from the second edition, but is expressly disavowed,[862]\nwould seem to indicate that it had been hastily adopted on the very eve\nof publication, and that upon reflection Kant had felt constrained\ndefinitively to discard it. The threefold synthesis can be verified on\nthe empirical level, but there is no evidence that there exist\ncorresponding transcendental activities.\n\n\n=IV. Connected Statement and Discussion of Kant\u2019s Subjective and\nObjective Deductions in the First Edition=\n\nSuch are the varying and conflicting forms in which Kant has presented\nhis deduction of the categories. We may now apply our results to obtain\na connected statement of the essentials of his argument. The following\nexposition, which endeavours to emphasise its main broad features, to\ndistinguish its various steps, and to disentangle its complex and\nconflicting tendencies, will, I trust, yield to the reader such steady\norientation as is necessary in so bewildering a labyrinth. In the\nmeantime I shall take account only of the deductions of the first\nedition,[863] and from them shall strive to construct the ideal\nstatement to which they severally approximate. Any single relatively\nconsistent and complete deduction that is thus to serve as a standard\nexposition must, like the root-languages of philology, be typical or\narchetypal, representing the argument at which Kant aimed; it cannot be\none of the alternative expositions which he himself gives. Such\nreconstruction of an argument which Kant has failed to express in a\nfinal and genuinely adequate form must, of course, lie open to all the\ndangers of arbitrary and personal interpretation. It is an extremely\nadventurous undertaking, and will have to be carefully guarded by\nconstant reference to Kant\u2019s _ipsissima verba_. Proof of its historical\nvalidity will consist in its capacity to render intelligible Kant\u2019s own\ndepartures from it, and in its power of explaining the reasons of his so\ndoing. Its expository value will be in proportion to the assistance\nwhich it may afford to the reader in deciphering the actual texts.\n\nOur first task is to make clear the nature of the distinction which Kant\ndraws between the \u201csubjective\u201d and the \u201cobjective\u201d deductions. This is a\ndistinction of great importance, and raises issues of a fundamental\ncharacter. In regard to it students of Kant take widely different views.\nFor it brings to a definite issue many of the chief controversies\nregarding Critical teaching. Kant has made some very definite statements\nin regard to it; and one of the opposing schools of interpretation finds\nits chief and strongest arguments in the words which he employs. But for\nreasons which will appear in due course, adherence to the letter of the\n_Critique_ would in this case involve the commentator in great\ndifficulties. We have no option except to adopt the invidious position\nof maintaining that we may now, after the interval of a hundred years\nand the labours of so many devoted students, profess to understand Kant\nbetter than he understood himself. For such procedure we may indeed cite\nhis own authority.\n\n     \u201cNot infrequently, upon comparing the thoughts which an author has\n     expressed in regard to his subject, whether in ordinary\n     conversation or in writing, we find that we can understand him\n     better than he understood himself. As he has not sufficiently\n     determined his concept, he has sometimes spoken, or even thought,\n     in opposition to his own intention.\u201d[864]\n\nLet us, then, consider first the distinction between the two types of\ndeduction in the form in which it is drawn by Kant. In the _Preface_ to\nthe first edition,[865] Kant states that his transcendental deduction of\nthe categories has two sides, and assigns to them the titles subjective\nand objective.\n\n     \u201cThis enquiry, which is somewhat deeply grounded, has two sides.\n     The one refers to the objects of pure understanding, and is\n     intended to expound and render intelligible the objective validity\n     of its _a priori_ concepts. It is therefore essential to my\n     purposes. The other seeks to investigate the pure understanding\n     itself, its possibility and the cognitive faculties upon which it\n     rests. Although this latter exposition is of great importance for\n     my chief purpose, it does not form an essential part of it. For the\n     chief question is always simply this,--what and how much can the\n     understanding and Reason know apart from all experience? not--how\n     is the faculty of thought itself possible? The latter is as it were\n     a search for the cause of a given effect; and therefore is of the\n     nature of an hypothesis (though, as I shall show elsewhere, this is\n     not really so); and I would appear to be taking the liberty simply\n     of expressing an opinion, in which case the reader would be free to\n     express a different opinion.[866] For this reason I must forestall\n     the reader\u2019s criticism by pointing out that the objective\n     deduction, with which I am here chiefly concerned, retains its full\n     force even if my subjective deduction should fail to produce that\n     complete conviction for which I hope....\u201d\n\nThe subjective deduction seeks to determine the subjective conditions\nwhich are required to render knowledge possible, or to use less\nambiguous terms the generative processes to whose agency human knowledge\nis due. It is consequently psychological in character. The objective\ndeduction, on the other hand, is so named because it deals not with\npsychological processes but with questions of objective validity. It\nenquires how concepts which are _a priori_, and which as _a priori_ must\nbe taken to originate in pure reason, can yet be valid of objects. In\nother words, the objective deduction is logical, or, to use a\npost-Kantian term, epistemological in character.\n\nIt is indeed true, as Kant here insists, that the subjective deduction\ndoes not concern itself in any quite direct fashion with the Critical\nproblem--how _a priori_ ideas can relate to objects. \u201cAlthough of great\nimportance for my chief purpose, it does not form an essential part of\nit.\u201d This, no doubt, is one reason why Kant omitted it when he revised\nthe _Critique_ for the second edition.[867] None the less it is, as he\nhere says, important; and what exactly that importance amounts to, and\nwhether it is really true that it has such minor importance as to be\nrightly describable as unessential, is what we have to decide.\n\nThough empirical psychology, in so far as it investigates the temporal\ndevelopment of our experience, is, as Kant very justly claims, entirely\ndistinct in aim and method from the Critical enquiry, the same cannot be\nsaid of a psychology which, for convenience, and on the lines of Kant\u2019s\nown employment of terms, may be named transcendental.[868] For it will\ndeal, not with the temporal development of the concrete and varied\naspects of consciousness, but with the more fundamental question of the\ngenerative conditions indispensably necessary to consciousness as such,\n_i.e._ to consciousness in each and every one of its possible\nembodiments. In the definition above given of the objective deduction, I\nhave intentionally indicated Kant\u2019s unquestioning conviction that the _a\npriori_ originates independently of the objects to which it is applied.\nThis independent origin is only describable in mental or psychological\nterms. The _a priori_ originates from within; it is due to the specific\nconditions upon which human thinking rests. Now this interpretation of\nthe _a priori_ renders the teaching contained in the subjective\ndeduction much more essential than Kant is himself willing to recognise.\nThe conclusions arrived at may be highly schematic in conception, and\nextremely conjectural in detail; they are none the less required to\nsupplement the results of the more purely logical analysis. For though\nin the second edition the sections devoted to the subjective deduction\nare suppressed, their teaching, and the distinctions which they draw\nbetween the different mental processes, continue to be employed in the\nexposition of the objective deduction, and indeed are presupposed\nthroughout the _Critique_ as a whole. They are indispensably necessary\nin order to render really definite many of the contentions which the\nobjective deduction itself contains. To eliminate the subjective\ndeduction is not to cut away these presuppositions, but only to leave\nthem in the obscure region of the undefined. They will still continue to\ninfluence our mode of formulating and of solving the Critical problem,\nbut will do so as untested and vaguely outlined assumptions, acting as\nunconscious influences rather than as established principles. For these\nreasons the omission of the subjective deduction is to be deplored. The\nexplicit statement of the implied psychological conditions is preferable\nto their employment without prior definition and analysis. The deduction\nof the second edition rests throughout upon the initial and\nindispensable assumption, that though connection or synthesis can never\nbe given, it is yet the generative source of all consciousness of order\nand relation. Factors which are transcendental in the strict or logical\nmeaning of the term rest upon processes that are transcendental in a\npsychological sense.\n\nThis last phrase, \u2018transcendental in a psychological sense,\u2019 calls for a\nword of justification. The synthetic processes generative of experience\nare not, of course, transcendental in the strict sense. For they are not\n_a priori_ in the manner of the categories. None the less they are\ndiscoverable by the same transcendental method, namely, as being, like\nthe categories, indispensably necessary to the possibility of\nexperience. They differ from the categories in that they are not\nimmanent in experience, constituent of it, and cannot therefore be known\nin their intrinsic nature. As they fall outside the field of\nconsciousness, they can only be hypothetically postulated. None the\nless, formal categories and generative processes, definable elements and\nproblematic postulates, alike agree in being conditions _sine qua non_\nof experience. And further, in terms of Kant\u2019s presupposed psychology,\nthe latter are the source to which the former are due. There would thus\nseem to be sufficient justification for extending the term\ntranscendental to cover both; and in so doing we are following the path\nwhich Kant himself willingly travelled. For such would seem to have been\nhis unexpressed reasons for ascribing, as he does, the synthetic\ngenerative processes to what he himself names transcendental faculties.\n\nThis disposes of Kant\u2019s chief reason for refusing to recognise the\nsubjective deduction as a genuine part of the Critical enquiry, namely,\nthe contention upon which he lays such emphasis in the prefaces both of\nthe first and of the second edition,[869] that in transcendental\nphilosophy nothing hypothetical, nothing in any degree dependent upon\ngeneral reasoning from contingent fact, can have any place. That\ncontention proves untenable even within the domain of his purely logical\nanalyses. The very essence of his transcendental method consists in the\nestablishment of _a priori_ elements through proof of their connection\nwith factual experience. Kant is here revealing how greatly his mind is\nstill biased by the Leibnizian rationalism from which he is breaking\naway. His _a priori_ cannot establish itself save in virtue of\nhypothetical reasoning.[870] His transcendental method, rightly\nunderstood, does not differ in essential nature from the hypothetical\nmethod of the natural sciences; it does so only in the nature of its\nstarting-point, and in the character of the analyses which that\nstarting-point prescribes. And if hypothetical reasoning may be allowed\nin the establishment of the logical _a priori_, there is no sufficient\nreason why it may not also be employed for the determination of\ndynamical factors. The sole question is as to whether the hypotheses\nconform to the logical requirements and so raise themselves to a\ndifferent level from mere opinion and conjecture.[871] As Kant himself\nsays,[872] though his conclusions in the subjective deduction may seem\nto be hypothetical in the illegitimate sense, they are not really so.\nFrom the experience in view of which they are postulated they receive at\nonce the proof of their actuality and the material for their\nspecification.\n\nWe may now return to the question of the nature of the two deductions.\nThe complex character of their interrelations may be outlined as\nfollows:\n\n1. Though the subjective deduction is in its later stages coextensive\nwith its objective counterpart, in its earlier stages it moves wholly on\nwhat may be called the empirical level. The data which it analyses and\nthe conditions which it postulates are both alike empirical. The\nobjective deduction, on the other hand, deals from start to finish with\nthe _a priori_.\n\n2. The later stages of the subjective deduction are based upon the\nresults of the objective deduction. The existence and validity of _a\npriori_ factors having been demonstrated by transcendental, _i.e._\nlogical, analysis, the subjective deduction can be extended from the\nlower to the higher level, and can proceed to establish for the _a\npriori_ elements what in its earlier stages it has determined for\nempirical consciousness, namely, the nature of the generative processes\nwhich require to be postulated as their ground and origin. When the two\ndeductions are properly distinguished the objective deduction has,\ntherefore, to be placed midway between the initial and the final stages\nof the subjective deduction.\n\n3. The two deductions concentrate upon different aspects of experience.\nIn the subjective deduction experience is chiefly viewed as a _temporal\nprocess_ in which the given falls apart into successive events, which,\nin and by themselves, are incapable of constituting a unified\nconsciousness. The fundamental characteristic of human experience, from\nthis point of view, is that it is _serial_ in character. Though it is an\napprehension of time, it is itself also a process in time. In the\nobjective deduction, on the other hand, the time element is much less\nprominent. Awareness of _objects_ is the subject-matter to which\nanalysis is chiefly devoted. This difference very naturally follows from\nthe character of the two deductions. The subjective enquiry is mainly\ninterested in the conditions generative of experience, and finds its\nnatural point of departure in the problem by what processes a unified\nexperience is constructed out of a succession of distinct happenings.\nThe objective deduction presents the logical problem of validity in its\nmost striking form, in our awareness of objects; the objective is\ncontrasted with the subjective as being that which is universally and\nnecessarily the same for all observers. Ultimately each of the two\ndeductions must yield an analysis of both types of\nconsciousness--awareness of time and awareness of objects; _a priori_\nfactors are involved in the former no less than in the latter, and both\nare conditioned by generative processes. Unfortunately the manner in\nwhich this is done in the _Critique_ causes very serious\nmisunderstanding. The problem of the psychological conditions generative\nof consciousness of objects is raised[873] before the logical analysis\nof the objective deduction has established the data necessary for its\nprofitable discussion. The corresponding defect in the objective\ndeduction is of a directly opposite character, but is even more\nunfortunate in its effects. The results obtained from the analysis of\nour awareness of objects are not, within the limits of the objective\ndeduction, applied in further analysis of our consciousness of time.\nThat is first done, and even then by implication rather than by explicit\nargument, in the _Analytic of Principles_. This has the twofold evil\nconsequence, that the relations holding between the two deductions are\nvery greatly obscured, and that the reader is not properly prepared for\nthe important use to which the results of the objective deduction are\nput in the _Analytic of Principles_. For it is there assumed--a quite\nlegitimate inference from the objective deduction, but one whose\nlegitimacy Kant has nowhere dwelt upon and explained--that to be\nconscious of time we must be conscious of it as existing in two distinct\norders, subjective and objective. To be conscious of time we must be\nconscious of objects, and to be conscious of objects we must be able to\ndistinguish between the order of our ideas and the order of the changes\n(if any) in that which is known by their means.\n\nThus the two deductions, properly viewed in their full scope, play into\none another\u2019s hands. The objective deduction is necessary to complete\nthe analysis of time-consciousness given in the subjective deduction,\nand the extension of the analysis of object-consciousness to the\nexplanation of time-consciousness is necessary in order to make quite\ndefinite and clear the full significance of the conclusions to which the\nobjective enquiry has led.[874]\n\nOne last point remains for consideration. Experience is a highly\nambiguous term, and to fulfil the r\u00f4le assigned to it by Kant\u2019s\ntranscendental method--that of establishing the reality of the\nconditions of its own possibility--its actuality must lie beyond the\nsphere of all possible controversy. It must be itself a datum, calling\nindeed for explanation, but not itself making claims that are in any\ndegree subject to possible challenge. Now if we abstract from all those\nparticularising factors which are irrelevant in this connection, we are\nleft with only three forms of experience--experience of self, experience\nof objects, and experience of time. The two former are open to question.\nThey may be illusory, as Hume has argued. And as their validity, or\nrather actuality, calls for establishment, they cannot fulfil the\ndemands which the transcendental method exacts from the experience whose\npossibility is to yield proof of its discoverable conditions.\nConsciousness of time, on the other hand, is a fact whose actuality,\nhowever problematic in its conditions, and however mysterious in its\nintrinsic nature, cannot, even by the most metaphysical of subtleties,\nbe in any manner or degree challenged. It is an unquestioned possession\nof the human mind. Whether time itself is real we are not metaphysically\ncertain, but that, whatever be its reality or unreality, we are\nconscious of it in the form of change, is beyond all manner of doubt.\nConsciousness of time is the _factual_ experience, as conditions of\nwhose possibility the _a priori_ factors are transcendentally proved.\nIn so far as they can be shown to be its indispensable conditions, its\nmere existence proves their reality. And such in effect is the ultimate\ncharacter of Kant\u2019s proof of the objective validity of the categories.\nThey are proved in that it is shown that only in and through them is\nconsciousness of time possible.\n\nThe argument gains immeasurably in clearness when this is\nrecognised;[875] and the deduction of the first edition of the\n_Critique_, in spite of its contorted character, remains in my view\nsuperior to that of the second edition owing to this more explicit\nrecognition of the temporal aspect of consciousness and to employment of\nit as the initial starting-point. Analysis at once reveals that though\nconsciousness of time is undeniably actual, it is conditioned in complex\nways, and that among the conditions indispensably necessary to its\npossibility are both consciousness of self and consciousness of an\nobjective order of existence. Starting from the undeniable we are thus\nbrought to the problematic; but owing to the factual character of the\nstarting-point we can substantiate what would otherwise remain open to\nquestion.\n\nAs this method of formulating Kant\u2019s argument gives greater prominence\nto the temporal factor than Kant himself does in his statement of the\ndeductions, the reader may very rightly demand further evidence that I\nam not, by this procedure, setting the deductions in a false or\narbitrary perspective. Any statement of Kant\u2019s position in other than\nhis own _ipsissima verba_ is necessarily, in large part, a matter of\ninterpretation, and proof of its correctness must ultimately consist in\nthe success with which it can be applied in unravelling the manifold\nstrands that compose his tortuous and many-sided argument; but the\nfollowing special considerations may be cited in advance. Those parts of\nthe _Critique_, such as the chief paragraphs of the subjective deduction\nand the chapter on _Schematism_, which are demonstrably late in date of\nwriting, agree in assigning greater prominence to the temporal aspect of\nexperience. This is also true of those numerous passages added in the\nsecond edition which deal with inner sense. All of these show an\nincreasing appreciation of the central r\u00f4le which time must play in the\nCritical enquiries. Secondly, proof of the validity of _specific_\ncategories is given, as we shall find,[876] not in the objective\ndeduction of the _Analytic of Concepts_, but only in the _Analytic of\nPrinciples_. What Kant gives in the former is only the quite general\ndemonstration that forms of unity, such as are involved in all judgment,\nare demanded for the possibility of experience. Now when proof of the\nspecific categories does come, in the _Analytic of Principles_, it is\nmanifestly based on the analysis of time-experience. In the three\n_Analogies_, for example, Kant\u2019s demonstration of the objective validity\nof the categories of relation consists in the proof that they are\nnecessary conditions of the possibility of our time-consciousness. That\nis to say, the transcendental method of proof, when developed in full\ndetail, in reference to some specific category, agrees with the\nformulation which I have given of the subjective and objective\ndeductions. In the third place, Kant started from a spiritualist\nstandpoint, akin to that of Leibniz,[877] and only very gradually broke\naway from the many illegitimate assumptions which it involves. But this\noriginal starting-point reveals its persisting influence in the\nexcessive emphasis which Kant continued to lay upon the unity of\napperception. He frequently speaks[878] as if it were an ultimate\nself-justifying principle, by reference to which the validity of all\npresupposed conditions can be established. But that, as I have already\nargued, is a legitimate method of procedure only if it has previously\nbeen established that self-consciousness is involved in all\nconsciousness, that is, involved even in consciousness of sequence and\nduration. And as just stated, the deductions of specific categories,\ngiven in the _Analytic of Principles_, fulfil these requirements of\ncomplete proof. They start from the time-consciousness, not from\napperception.\n\nI shall now summarise these introductory discussions in a brief\ntabulated outline of the main steps in the argument of the two\ndeductions, and shall add a concluding note upon their interconnection.\n\n=Subjective Deduction.=--1. Consciousness of time is an experience whose\nactuality cannot be questioned; by its actuality it will therefore\nestablish the reality of everything that can be proved to be its\nindispensable condition.\n\n2. Among the conditions indispensably necessary to all consciousness of\ntime are synthetic processes whereby the contents of consciousness,\noccurring in successive moments, are combined and unified. These\nprocesses are processes of apprehension, reproduction, and recognition.\n\n3. Recognition, in turn, is conditioned by self-consciousness.\n\n4. As no consciousness is possible without self-consciousness, the\nsynthetic processes must have completed themselves before such\nself-consciousness is possible, and consequently are not verifiable by\nintrospection but only by hypothetical construction.\n\n[1, 2, 3, and 4 are steps which can be stated independently of the\nargument of the objective deduction.]\n\n5. Self-consciousness presupposes consciousness of objects, and\nconsciousness of objects presupposes the synthetic activities of\nproductive imagination whereby the matter of sense is organised in\naccordance with the categories. These productive activities also are\nverifiable only by conjectural inference, and only upon their completion\ncan consciousness of any kind make its appearance.\n\n6. Consciousness of self and consciousness of objects thus alike rest\nupon a complexity of non-phenomenal conditions. For anything that\ncritical analysis can prove to the contrary, consciousness and\npersonality may not be ultimates. They may be resultants due to\nrealities fundamentally different from themselves.\n\n[5 is a conclusion obtained only by means of the argument of the\nobjective deduction. 6 is a further conclusion, first explicitly drawn\nby Kant in the _Dialectic_.]\n\n=Objective Deduction.=--1. The starting-point coincides with that of the\nsubjective deduction. Consciousness of time is an experience by whose\nactuality we can establish the reality of its indispensable conditions.\n\n2. Among the conditions necessary to all consciousness of time is\nself-consciousness.\n\n3. Self-consciousness, in turn, is itself conditioned by consciousness\nof objects.\n\n4. Consciousness of objects is possible only if the categories have\nvalidity within the sphere of sense-experience.\n\n5. Conclusion.--The empirical validity of the categories, and\nconsequently the empirical validity of our consciousness alike of the\nself and of objects, must be granted as a _conditio sine qua non_ of our\nconsciousness of time. They are the indispensable conditions of that\nfundamental experience.\n\nAs above stated,[879] the preliminary stages of the subjective deduction\nprepare the way for the argument of the objective deduction, while the\nresults obtained by the latter render possible the concluding steps of\nthe former. That is to say, the objective deduction has to be\nintercalated midway between the opening and the concluding stages of the\nsubjective deduction. It may also be observed that whereas the objective\ndeduction embodies the main positive teaching of the _Analytic_, in that\nit establishes the possibility of natural science and of a metaphysics\nof experience, the subjective deduction is more directly concerned with\nthe subject-matter of the _Dialectic_, reinforcing, as it does, the more\nnegative consequences which follow from the teaching of the objective\ndeduction--the impossibility of transcendent speculation. It stands in\npeculiarly close connection with the teaching of the section on the\n_Paralogisms_. We may now proceed to a detailed statement of the\nargument of the two deductions.\n\n\nTHE SUBJECTIVE DEDUCTION IN ITS INITIAL EMPIRICAL STAGES\n\nIn the opening of the subjective deduction Kant is careful to give due\nprominence to the temporal aspect of our human experience.\n\n     \u201c...all the contents of our knowledge are ultimately subject to the\n     formal condition of inner sense, that is, to time, as that wherein\n     they must all be ordered, connected, and brought into relation to\n     one another. This is a general remark which the reader must bear in\n     mind as being a fundamental presupposition of my entire\n     argument.\u201d[880]\n\nConsciousness of time is thus the starting-point of the deduction.\nAnalysis reveals it as highly complex; and the purpose of the deduction\nis to discover, and, as far as may be possible, to define its various\nconditions. The argument can best be expounded by reference to a single\nconcrete example--say, our experience of a series of contents, _a_, _b_,\n_c_, _d_, _e_, _f_, as in succession to one another and as together\nmaking up the total six. In order that such an experience may be\npossible the successive members of the series must be held together\nsimultaneously before the mind. Obviously, if the earlier members\ndropped out of consciousness before the mind reached _f_, _f_ could not\nbe apprehended as having followed upon them. There must be a synthesis\nof apprehension of the successive items.\n\nSuch a synthesis of apprehension is, however, only possible through\nreproduction of the earlier experiences. If when the mind has passed\nfrom _a_ to _f_, _f_ is apprehended as having followed upon _a_, _b_,\n_c_, _d_, _e_, such consciousness is only possible in so far as these\nearlier contents are reproduced in image. Synthesis of apprehension is\nconditioned by synthesis of reproduction in imagination.\n\n     \u201cBut if the preceding representations (the first parts of [a] line,\n     the earlier moments of time or the units represented in sequent\n     order) were always to drop out of my thought, and were not\n     reproduced when I advance to those that follow, no complete\n     representation, and none of all the aforementioned thoughts, not\n     even the purest and first basal representations of space and time,\n     could ever arise.\u201d[881]\n\nIn order, however, that the reproduced images may fulfil their function,\nthey must be recognised as standing for or representing contents which\nthe self has just experienced.\n\n     \u201cWithout the consciousness that what we are thinking is the same as\n     what we thought a moment before, all reproduction in the series of\n     representations would be in vain.\u201d[882]\n\nEach reproduced image would in its present state be a new experience,\nand would not help in the least towards gaining consciousness of order\nor number in the succession of our experiences. Recognition is,\ntherefore, a third form of synthesis, indispensably necessary to\nconsciousness of time. But further, the recognition is recognition of a\nsuccession as forming a unity or whole, and that unity is always\nconceptual.\n\n     \u201cThe word concept (_Begriff_) might of itself have suggested this\n     remark. For it is this unitary consciousness which unites into a\n     single representation a manifold that has been successively\n     intuited and then subsequently reproduced.\u201d[883] \u201cIf in counting I\n     forgot that the units ... have been added to one another in\n     succession, I should never recognise what the sum-total is that is\n     being produced through the successive addition of unit to unit; and\n     so would remain ignorant of the number. For the concept of this\n     number is nothing but the consciousness of this unity of\n     synthesis.\u201d[884]\n\nThe synthesis of recognition is thus a synthesis which takes place in\nand through empirical concepts. In the instance which we have chosen,\nthe empirical concept is that of the number six.\n\nThe analysis, however, is not yet complete. Just as reproduction\nconditions apprehension and both rest on recognition, so in turn\nrecognition presupposes a still further condition, namely,\nself-consciousness. For it is obvious, once the fact is pointed out,\nthat the recognition of reproduced images as standing for past\nexperiences can only be possible in so far as there is an abiding self\nwhich is conscious of its identity throughout the succession. Such an\nact of recognition is, indeed, merely one particular form or concrete\ninstance of self-consciousness. The unity of the empirical concept in\nand through which recognition takes place finds its indispensable\ncorrelate in the unity of an empirical self. Thus an analysis of our\nconsciousness, even though conducted wholly on the empirical level, that\nis, without the least reference to the _a priori_, leads by simple and\ncogent argument to the conclusion that it is conditioned by complex\nsynthetic processes, and that these syntheses in turn presuppose a unity\nwhich finds twofold expression for itself, objectively through a concept\nand subjectively in self-consciousness.\n\nSo far I have stated the argument solely in reference to serial\nconsciousness. Kant renders his argument needlessly complex and\ndiminishes its force by at once extending it so as to cover the\nconnected problem, how we become aware of objects. This occurs in the\nsection on the synthesis of reproduction. An analysis of our\nconsciousness of objects, as distinct from consciousness of the\nimmediately successive, forces us to postulate further empirical\nconditions. Since the reproductive imagination, to whose agency the\napprehension of complex unitary existences is psychologically due, acts\nthrough the machinery of association, it presupposes constancy in the\napprehended manifold.\n\n     \u201cIf cinnabar were sometimes red, sometimes black, sometimes light,\n     sometimes heavy, if a man changed sometimes into this and sometimes\n     into that animal form, if the country on the longest day were\n     sometimes covered with fruits, sometimes with ice and snow, my\n     empirical imagination would never even have occasion when\n     representing red colour to bring to mind heavy cinnabar....\u201d[885]\n\nThis passage may be compared with the one which occurs in the section on\nthe synthesis of recognition. Our representations, in order to\nconstitute knowledge, must have the unity of some concept; the manifold\ncannot be apprehended save in so far as this is possible.\n\n     \u201cAll knowledge demands a concept, though that concept may be quite\n     imperfect or obscure. But a concept is always, as regards its form,\n     something general which serves as a rule. The concept of body, for\n     instance, as the unity of the manifold which is thought through it,\n     serves as a rule to our knowledge of outer appearances.... It\n     necessitates in the perception of something outside us the\n     representation of extension, and therewith the representations of\n     impenetrability, form, etc.\u201d[886]\n\nSo far the deduction still moves on the empirical level. When Kant,\nhowever, proceeds to insist[887] that this empirical postulate itself\nrests upon a transcendental condition, the argument is thrown into\ncomplete confusion, and the reader is bewildered by the sudden\nanticipation of one of the most difficult and subtle conclusions of the\nobjective deduction. The same confusion is also caused throughout these\nsections as a whole by Kant\u2019s description of the various syntheses as\nbeing transcendental.[888] They cannot properly be so described. The\nconcepts referred to as unifying the syntheses, and the\nself-consciousness which is proved to condition the syntheses, are all\nempirical. They present themselves in concrete form, and presuppose\ncharacteristics due to the special contingent nature of the given\nmanifold; as Kant states in so many words in the second edition.\n\n     \u201cWhether I can become _empirically_ conscious of the manifold as\n     simultaneous or as successive depends on circumstances or empirical\n     conditions. The empirical unity of consciousness, through\n     association of representations, therefore itself relates to an\n     appearance, and is wholly contingent.\u201d[889]\n\nThe argument in these preliminary stages of the subjective deduction, in\nso far as it is employed to yield proof that all consciousness involves\nthe unity of concepts and the unity of self-consciousness, is\nindependent of any reference to the categories, and consequently to\ntranscendental conditions. In accordance with the plan of exposition\nabove stated, we may now pass to the objective deduction.\n\n\nOBJECTIVE DEDUCTION AS GIVEN IN THE FIRST EDITION\n\nThe transition from the preliminary stages of the subjective deduction\nto the objective deduction may be made by further analysis either of the\nobjective unity of empirical concepts or of the subjective unity of\nempirical self-consciousness. It is the former line which the argument\nof the first edition follows. Kant is asking what is meant by an object\ncorresponding to our representations,[890] and answers by his objective\ndeduction. He substitutes the empirical for the transcendental\nobject,[891] and in so doing propounds one of the central and most\nrevolutionary tenets of the Critical philosophy. Existence takes a\nthreefold, not a merely dual form. Besides representations and things in\nthemselves, there exist the objects of our representations--the extended\nworld of ordinary experience and of science. Such a threefold\ndistinction is prefigured in the Leibnizian metaphysics, and is more or\nless native to every philosophy that is genuinely speculative. Kant\nhimself claims Plato as his philosophical progenitor. The originality is\nnot in the bare thesis, but in the fruitful, tenacious, and consistent\nmanner in which it is developed through detailed analysis of our actual\nexperience.\n\nIn its first stages the argument largely coincides with the argument of\nthe paragraphs which deal with the transcendental object. When we\nexamine the objective, we find that the primary characteristic\ndistinguishing it from the subjective is that it lays a compulsion upon\nour minds, constraining us to think about it in a certain way. By an\nobject is meant something which will not allow us to think at haphazard.\nCinnabar is an object which constrains us to think it as heavy and red.\nAn object is thus the external source of a necessity to which our\nthinking has to conform. The two arguments first begin to diverge when\nKant sets himself to demonstrate that our consciousness of this external\nnecessity is made possible by categories which originate from within.\n\nFor this conclusion Kant prepares the way by an analysis of the second\nmain characteristic constitutive of an object, viz. its unity. This\nunity is of a twofold nature, involving either the category of substance\nand attribute or the category of cause and effect. The two categories\nare ultimately inseparable, but lead us to conceive the object in two\ndistinct modes. When we interpret an object through the _a priori_\nconcept of substance and attribute, we assert that all the contents of\nour perceptions of it are capable of being regarded as qualities of one\nand the same identical substance. No one of its qualities can be\nincongruent with any other, and all of them together, in their unity,\nmust be expressive of its substantial nature.\n\nThe causal interpretation of the object is, however, the more important,\nand is that which is chiefly emphasised by Kant. It is, indeed, simply a\nfurther and more adequate mode of expressing the substantial unity of\nthe object. All the qualities must be causally bound up with one another\nin such a way that the nature of each is determined by the nature of all\nthe others, and that if any one quality be changed all the others must\nundergo corresponding alterations. Viewed in this manner, in terms of\nthe category of causality, an object signifies a necessitated\ncombination of interconnected qualities or effects. But since no such\nform of _necessitation_ can be revealed in the manifold of sense, our\n_consciousness_ of compulsion cannot originate from without, and must be\ndue to those _a priori_ forms which, though having their source within,\ncontrol and direct our interpretation of the given. Though the objective\ncompulsion is not itself due to the mind, our _consciousness_ of it has\nthis mental _a priori_ source. The concept of an object consists in the\nthought of a manifold so determined in its specific order and groupings\nas to be interpretable in terms of the categories of substance and\ncausality.\n\nBut the problem of the deduction proper is not yet raised. On the one\nhand, Kant has defined what the concept of the objective must be taken\nas involving, and on the other, has pointed out that since the given as\ngiven is an unconnected manifold, any categories through which it may be\ninterpreted must be of independent origin; but it still remains to be\nproved that the above is a valid as well as a possible mode of\nconstruing the given appearances. The categories, as _a priori_\nconcepts, originate from within. By what right may we assert that they\nnot only relate to an object, but even constitute the very concept of\nit? Are appearances legitimately interpretable in any such manner? It\nwas, we may believe, in the process of answering this question that Kant\ncame to realise that the objects of our representations must no longer\nbe regarded as things in themselves. For, as he finds, a solution is\npossible only on the further assumption that the mind is legislating\nmerely for the world of sense-experience, and is making no assertion in\nregard to the absolutely and independently real. Kant\u2019s method of proof\nis the transcendental, _i.e._ he seeks to demonstrate that this\ninterpretation of the given is indispensably necessary as being a _sine\nqua non_ of its possible apprehension. This is achieved by means of the\nconclusion already established through the preliminary steps of the\nsubjective deduction, namely, that all consciousness involves\nself-consciousness. Kant\u2019s proof of the objective validity of the\ncategories consists in showing that only by means of the interpretation\nof appearances as _empirically_ objective is self-consciousness possible\nat all.\n\nThe self-consciousness of the subjective deduction, in the preliminary\nform above stated, is, however, itself empirical. Kant, developing on\nmore strictly Critical lines the argument which had accompanied his\nearlier doctrine of the transcendental object, now proceeds to maintain\nin what is at once the most fruitful and the most misleading of his\ntenets, that the ultimate ground of the possibility of consciousness and\ntherefore also of empirical self-consciousness is the transcendental\nunity of apperception. Such apperception, to use Kant\u2019s ambiguous\nphraseology, precedes experience as its _a priori_ condition. The\ninterpretation of given appearances through _a priori_ categories is a\nnecessity of consciousness because it is a condition of\nself-consciousness; and it is a condition of self-consciousness because\nit alone will account for the transcendental apperception upon which all\nempirical self-consciousness ultimately depends.\n\nOne chief reason why Kant\u2019s deduction is found so baffling and illusive\nis that it rests upon an interpretation of the unity of apperception\nwhich is very definitely drawn, but to which Kant himself gives only the\nbriefest and most condensed expression. I shall therefore take the\nliberty of restating it in more explicit terms. The true or\ntranscendental self has no content of its own through which it can gain\nknowledge of itself. It is mere identity, I am I. In other words,\nself-consciousness is a mere _form_ through which contents that never\nthemselves constitute the self are yet apprehended as being objects to\nthe self. Thus though the self in being conscious of time or duration\nmust be conscious of itself as _identical_ throughout the succession of\nits experiences, that identity can never be discovered in those\nexperiences; it can only be thought as a condition of them. The\ncontinuity of memory, for instance, is not a possible substitute for\ntranscendental apperception. As the subjective deduction demonstrates,\nself-consciousness conditions memory, and cannot therefore be reduced to\nor be generated by it.[892] When, however, such considerations are\nallowed their due weight, the necessity of postulating a transcendental\nunity becomes only the more evident. Though it can never itself be found\namong appearances, it is an interpretation which we are none the less\ncompelled to give to appearances.\n\nTo summarise before proceeding. We have obtained two important\nconclusions: first, that all consciousness involves self-consciousness;\nand secondly, that self-consciousness is a mere form, in terms of which\ncontents that do not constitute the self are apprehended as existing for\nthe self. The first leads up to the second, and the second is equivalent\nto the assertion that there can be no such thing as a pure\nself-consciousness, _i.e._ a consciousness in which the self is aware of\nitself and of nothing but itself. Self-consciousness, to be possible at\nall, must at the same time be a consciousness of something that is\nnot-self. Only one further step is now required for the completion of\nthe deduction, namely, proof that this not-self, consciousness of which\nis necessary to the possibility of self-consciousness, must consist in\nempirical objects apprehended in terms of the categories. For proof Kant\nagain appeals to the indispensableness of apperception. As no intuitions\ncan enter consciousness which are not capable of being related to the\nself, they must be so related to one another that, notwithstanding their\nvariety and diversity, the self can still be conscious of itself as\nidentical throughout them all. In other words, no intuition can be\nrelated to the self that is incapable of being combined together with\nall the other intuitions to form a unitary consciousness. I may here\nquote from the text of the second edition:[893]\n\n     \u201c...only in so far as I can grasp the manifold of the\n     representations in one consciousness, do I call them one and all\n     mine. For otherwise I should have as many-coloured and diverse a\n     self as I have representations of which I am conscious to myself.\u201d\n\nOr as it is stated in the first edition:[894]\n\n     \u201cWe are _a priori_ aware of the complete identity of the self in\n     respect of all representations which belong to our knowledge ... as\n     a necessary condition of the possibility of all representations.\u201d\n\nThese are the considerations which lead Kant to entitle the unity of\napperception _transcendental_. He so names it for the reason that,\nthough it is not itself _a priori_ in the manner of the categories, we\nare yet enabled by its means to demonstrate that the unity which is\nnecessary for possible experience can be securely counted upon in the\nmanifold of all possible representations, and because (as he believed)\nit also enables us to prove that the forms of such unity are the\ncategories of the understanding.\n\nTo the argument supporting this last conclusion Kant does not give the\nattention which its importance would seem to deserve. He points out that\nas the given is an unconnected manifold, its unity can be obtained only\nby synthesis, and that such synthesis must conform to the conditions\nprescribed by the unity of apperception. That these conditions coincide\nwith the categories he does not, however, attempt to prove. He\napparently believes that this has been already established in the\nmetaphysical deduction.[895] The forms of unity demanded by\napperception, he feels justified in assuming, are the categories. They\nmay be regarded as expressing the minimum of unity necessary to the\npossibility of self-consciousness. If sensations cannot be interpreted\nas the diverse attributes of unitary substances, if events cannot be\nviewed as arising out of one another, if the entire world in space\ncannot be conceived as a system of existences reciprocally\ninterdependent, all unity must vanish from experience, and apperception\nwill be utterly impossible.[896]\n\nThe successive steps of the total argument of the deduction, as given in\nthe first edition, are therefore as follows: Consciousness of time\ninvolves empirical self-consciousness; empirical self-consciousness is\nconditioned by a transcendental self-consciousness; and such\ntranscendental self-consciousness is itself, in turn, conditioned by\nconsciousness of objects. The argument thus completed becomes the proof\nof mutual interdependence. Self-consciousness and consciousness of\nobjects, as polar opposites, mutually condition one another. Only\nthrough consciousness of both simultaneously can consciousness of either\nbe attained. Only in and through reference to an object can an idea be\nrelated to a self, and so be accompanied by that self-consciousness\nwhich conditions recognition, and through recognition all the varying\nforms in which our consciousness can occur. From the point of view,\nhowever, of a Critical enquiry apperception is the more important of the\ntwo forms of consciousness. For though each is the _causa existendi_ of\nthe other, self-consciousness has the unique distinction of being the\n_causa cognoscendi_ of the objective and _a priori_ validity of the\nforms of understanding.\n\n     \u201cThe synthetic proposition, that all the variety of empirical\n     consciousness must be combined in a single self-consciousness, is\n     the absolutely first and synthetic principle of our thought in\n     general.\u201d[897]\n\nWe may at this point consider Kant\u2019s doctrine of \u201cobjective affinity.\u201d\nIt excellently enforces the main thesis which he is professing to\nestablish, namely, that the conditions of _unitary_ consciousness are\nthe conditions of _all_ consciousness. The language, however, in which\nthe doctrine is expounded is extremely obscure and difficult; and before\ncommenting upon Kant\u2019s own methods of statement, it seems advisable to\nparaphrase the argument in a somewhat free manner, and also to defer\nconsideration of the transcendental psychology which Kant has employed\nin its exposition.[898] Association can subsist only between ideas, both\nof which have occurred within the same conscious field. Now the\nfundamental characteristic of consciousness, the very condition of its\nexisting at all, is its unity; and until this has been recognised, there\ncan be no understanding of the associative connection which arises under\nthe conditions which consciousness supplies. To attempt to explain the\nunity of consciousness through the mechanism of association is to\nexplain an agency in terms of certain of its own effects. It is to\nexplain the fundamental in terms of the derivative, the conditions in\nterms of what they have themselves made possible. Kant\u2019s argument is\ntherefore as follows. Ideas do not become associated merely by\nco-existing. They must occur together in a unitary consciousness; and\namong the conditions necessary to the possibility of association are\ntherefore the conditions of the possibility of experience. Association\nis transcendentally grounded. So far from accounting for the unity of\nconsciousness, it presupposes the latter as determining the conditions\nunder which alone it can come into play.\n\n     \u201c...how, I ask, is association itself possible?... On my principles\n     the thorough-going affinity of appearances is easily explicable.\n     All possible appearances belong as representations to the totality\n     of a possible self-consciousness. But as this self-consciousness is\n     a transcendental representation, numerical identity is inseparable\n     from it and is _a priori_ certain. For nothing can come to our\n     knowledge save in terms of this original apperception. Now, since\n     this identity must necessarily enter into the synthesis of all the\n     manifold of appearances, so far as the synthesis is to yield\n     empirical knowledge, the appearances are subject to _a priori_\n     conditions, with which the synthesis of their apprehension must be\n     in complete accordance.... Thus all appearances stand in a\n     thorough-going connection according to necessary laws, and\n     therefore in a transcendental affinity of which the empirical is a\n     mere consequence.\u201d[899]\n\nIn other words, representations must exist in consciousness before they\ncan become associated; and they can exist in consciousness only if they\nare consciously apprehended. But in order to be consciously apprehended,\nthey must conform to the transcendental conditions upon which all\nconsciousness rests; and in being thus apprehended they are set in\nthoroughgoing unity to one another and to the self. They are apprehended\nas belonging to an objective order or unity which is the correlate of\nthe unity of self-consciousness. This is what Kant entitles their\nobjective affinity; it is what conditions and makes possible their\nassociative or empirical connection.\n\nThis main point is very definitely stated in A 101.\n\n     \u201cIf we can show that even our purest _a priori_ intuitions yield no\n     knowledge, save in so far as they contain such a connection of the\n     manifold as will make possible a thoroughgoing synthesis of\n     reproduction, this synthesis of the imagination\u201d [which acts\n     through the machinery of association] \u201cmust be grounded, prior to\n     all experience, on _a priori_ principles, and since experience\n     necessarily presupposes that appearances can be reproduced, we\n     shall have to assume a pure transcendental synthesis of the\n     imagination\u201d [_i.e._ such synthesis as is involved in the unity of\n     consciousness] \u201cas conditioning even the possibility of all\n     experience.\u201d[900]\n\nIn A 121-2 Kant expresses his position in a more ambiguous manner. He\nmay seem to the reader merely to be arguing that a certain minimum of\nregularity is necessary in order that representations may be associated,\nand experience may be possible.[901] But the general tenor of the\npassage as a whole, and especially its concluding sentences, enforce the\nstronger, more consistent, thesis.\n\n     \u201d[The] subjective and empirical ground of reproduction according to\n     rules is named the association of representations. If this unity of\n     association did not also have an objective ground, which makes it\n     impossible that appearances should be apprehended by the\n     imagination except under the condition of a possible synthetic\n     unity of this apprehension, it would be entirely accidental that\n     appearances should fit into a connected whole of human knowledge.\n     For even though we had the power of associating perceptions, it\n     would remain entirely undetermined and accidental whether they\n     would themselves be associable; and should they not be associable,\n     there might exist a multitude of perceptions, and indeed an entire\n     sensibility, in which much empirical consciousness would arise in\n     my mind, but in a state of separation, and without belonging to one\n     consciousness of myself. That, however, is impossible. For only in\n     so far as I ascribe all perceptions to one consciousness (original\n     apperception), can I say in all perceptions that I am conscious of\n     them. There must therefore be an objective ground (that is, one\n     that can be recognised _a priori_, antecedently to all empirical\n     laws of the imagination) upon which may rest the possibility, nay\n     the necessity, of a law that extends to all appearances....\u201d\n\nKant is not merely asserting that the associableness of ideas, and the\nregularity of connection which that implies, must be postulated as a\ncondition of experience. That would be a mere begging of the issue; the\ncorrectness of the postulate would not be independently proved. Kant is\nreally maintaining the much more important thesis, that the unity of\nexperience, _i.e._ of consciousness, is what makes association possible\nat all. And since consciousness must be unitary in order to exist, there\ncannot be any empirical consciousness in which the conditions of\nassociation, and therefore of reproduction, are not to be found.\n\nA further misunderstanding is apt to be caused by Kant\u2019s statement that\nassociative affinity rests upon objective affinity. This seems to imply,\nin the same manner as the passage which we have just considered, that\ninstead of proving that appearances are subject to law and order, he is\nmerely postulating that an abiding ground of such regularity must exist\nin the noumenal conditions of the sense manifold. But he himself again\nsupplies the needful correction.\n\n     \u201cThis [objective ground of all association of appearances] can\n     nowhere be found, except in the principle of the unity of\n     apperception in respect of all forms of knowledge which can belong\n     to me. In accordance with this principle all appearances must so\n     enter the mind, or be so apprehended, that they fit together to\n     constitute the unity of apperception. This would be impossible\n     without synthetic unity in their connection, and that unity is\n     therefore also objectively necessary. The objective unity of all\n     empirical consciousness in one consciousness, that of original\n     apperception, is therefore the necessary condition of all (even of\n     all _possible_) perception; and the affinity of all appearances,\n     near or remote, is a necessary consequence of a synthesis in\n     imagination which is grounded _a priori_ on rules.\u201d[902]\n\nThe fundamental characteristic of consciousness is the unified form in\nwhich alone it can exist; only when this unity is recognised as\nindispensably necessary, and therefore as invariably present whenever\nconsciousness exists at all, can the inter-relations of the contents of\nconsciousness be properly defined.\n\nIf this main contention of the Critical teaching be accepted, Hume\u2019s\nassociationist standpoint is no longer tenable. Association cannot be\ntaken to be an ultimate and inexplicable property of our mental states.\nNor is it a property which can be regarded as belonging to presentations\nviewed as so many independent existences. It is conditioned by the unity\nof consciousness, and therefore rests upon the \u201ctranscendental\u201d\nconditions which Critical analysis reveals. Since the unity of\nconsciousness conditions association, it cannot be explained as the\noutcome and product of the mechanism of association.\n\nIn restating the objective deduction in the second edition, Kant has\nomitted all reference to this doctrine of objective affinity. His\nreasons for this omission were probably twofold. In the first place, it\nhas been expounded in terms of a transcendental psychology, which, as we\nshall find, is conjectural in character. And secondly, the phrase\n\u201cobjective affinity\u201d is, as I have already pointed out, decidedly\nmisleading. It seems to imply that Kant is postulating, without\nindependent proof, that noumenal conditions must be such as to supply an\norderly manifold of sense data. But though the doctrine of objective\naffinity is eliminated, its place is to some extent taken[903] by the\nproof that all apprehension is an act of judgment and therefore involves\nfactors which cannot be reduced to, or explained in terms of,\nassociation.\n\nThere are a number of points in the deduction of the first edition which\ncall for further explanatory and critical comment. The first of these\nconcerns the somewhat misleading character of the term _a priori_ as\napplied to the categories. It carries with it rationalistic associations\nto which the Critical standpoint, properly understood, yields no\nsupport. The categories are for Kant of merely _de facto_ nature. They\nhave no intrinsic validity. They are proved only as being the\nindispensable conditions of what is before the mind as brute fact,\nnamely, conscious experience. By the _a priori_ is meant merely those\n_relational_ factors which are required to supplement the given manifold\nin order to constitute our actual consciousness. And, as Kant is careful\nto point out, the experience, as conditions of which their validity is\nthus established, is of a highly specific character, resting upon\nsynthesis of a manifold given in space and time. That is to say, their\nindispensableness is proved only for a consciousness which in these\nfundamental respects is constituted like our own.[904] And secondly, the\nvalidity of the _a priori_ categories, even in our human thinking, is\nestablished only in reference to that empirical world which is\nconstructed out of the given manifold in terms of the intuitive forms,\nspace and time. Their validity is a merely phenomenal validity. They are\nvalid of appearances, but not of things in themselves. The _a priori_ is\nthus doubly _de facto_: first as a condition of brute fact, namely, the\nactuality of our human consciousness; and secondly, as conditioning a\nconsciousness whose knowledge is limited to appearances. It is a\nrelative, not an absolute _a priori_. Acceptance of it does not,\ntherefore, commit us to rationalism in the ordinary meaning of that\nterm. Its credentials are conferred upon it by what is mere fact; it\ndoes not represent an order superior to the actual and legislative for\nit. In other words, it is Critical, not Leibnizian in character. No\ntranscendent metaphysics can be based upon it. In formulating this\ndoctrine of the _a priori_ as yielding objective insight and yet as\nlimited in the sphere of its application, the _Critique of Pure Reason_\nmarks an epoch in the history of scepticism, no less than in the\ndevelopment of Idealist teaching.\n\nThere is one important link in the deduction, as above given, which is\nhardly calculated to support the conclusions that depend upon it. Kant,\nas we have already noted,[905] asserts that the categories express the\nminimum of unity necessary for the possibility of apperception. A\ncontention so essential to the argument calls for the most careful\nscrutiny and a meticulous exactitude of proof. As a matter of fact, such\nproof is not to be found in any part of the deductions, whether of the\nfirst or of the second editions. It is attempted only in the later\nsections on the _Principles of Understanding_, and even there it is\ndeveloped, in any really satisfactory fashion, only in regard to the\ncategories of causality and reciprocity.[906] This proof, however, as\nthere given, is an argument which in originality, subtlety and force\ngoes far to atone for all shortcomings. It completes the objective\ndeduction by developing in masterly fashion (in spite of the diffuse and\nill-arranged character of the text) the central contention for which the\ndeduction stands. But in the transcendental deduction itself, we find\nonly such an argument--if it may be called an argument--as follows from\nthe identification of apperception with understanding.\n\n     \u201cThe unity of apperception, in relation to the synthesis of\n     imagination, is the understanding.... In understanding there are\n     pure _a priori_ forms of knowledge which contain the necessary\n     unity of pure synthesis of imagination in respect of all possible\n     appearances. But these are the categories, _i.e._ pure concepts of\n     understanding.\u201d[907]\n\nThe point is again merely assumed in A 125-6. So also in A 126:\n\n     \u201cAlthough through experience we learn many laws, these are only\n     special determinations of still higher laws, of which the highest,\n     under which all others stand, originate _a priori_ in the\n     understanding itself....\u201d[908]\n\nAgain in A 129 it is argued that as we prescribe _a priori_ rules to\nwhich all experience must conform, those rules cannot be derived from\nexperience, but must precede and condition it, and can do so only as\noriginating from ourselves (_aus uns selbst_).\n\n     \u201d[They] precede all knowledge of the object as [their] intellectual\n     form, and constitute a formal _a priori_ knowledge of all objects\n     in so far as they are thought (categories).\u201d\n\nBut this is only to repeat that such forms of unity as are necessary to\nself-consciousness must be realised in all synthesis. It is no\nsufficient proof that those forms of relation coincide with the\ncategories. As we shall find in considering the deduction of the second\nedition, Kant to some extent came to recognise the existence of this gap\nin his argument and sought to supply the missing steps. But his method\nof so doing still ultimately consists in an appeal to the results of the\nmetaphysical deduction, and therefore rests upon his untenable belief in\nthe adequacy of formal logic. It fails to obviate the objection in any\nsatisfactory manner.\n\nAs regards the negative aspect of the conclusion reached--that the\nvalidity of the categories is established only for appearances--Kant\nmaintains that this is a necessary corollary of their validity being _a\npriori_. That things in themselves must conform to the conditions\ndemanded by the nature of our self-consciousness is altogether\nimpossible of proof. Even granting, what is indeed quite possible, that\nthings in themselves embody the pure forms of understanding, we still\ncannot have any ground for maintaining that they must do so of necessity\nand will be found to do so universally. For even if we could directly\nexperience things in themselves, and apprehend them as conforming to the\ncategories, such conformity would still be known only as contingent. But\nwhen it is recognised that nature consists for us of nothing but\nappearances, existing only in the mode in which they are experienced,\nand therefore as necessarily conforming to the conditions under which\nexperience is alone possible, the paradoxical aspect of the apriority\nascribed to the categories at once vanishes. Proof of their _a priori_\nvalidity presupposes the phenomenal character of the objects to which\nthey apply. They can be proved to be universal and necessarily valid of\nobjects only in so far as it can be shown that they have antecedently\nconditioned and constituted them. The sole sufficient reason for\nasserting them to be universally valid throughout experience is that\nthey are indispensably necessary for rendering it possible.[909] The\ntranscendental method of proof, _i.e._ proof by reference to the very\npossibility of experience, is for this reason, as Kant so justly\nemphasises, the sole type of argument capable of fulfilling the demands\nwhich have to be met. It presupposes, and itself enforces, the truth of\nthe fundamental Critical distinction between appearances and things in\nthemselves.\n\nKant entitles the unity of apperception _original_\n(_urspr\u00fcnglich_);[910] and we may now consider how far and in what sense\nthis title is applicable.[911] _From the point of view of method_ there\nis the same justification for employing the term \u2018original\u2019 as for\nentitling the unity of apperception transcendental.[912]\nSelf-consciousness is more fundamental or original than consciousness of\nobjects, in so far as[913] it is only from the subjective standpoint\nwhich it represents that the objective deduction can demonstrate the\nnecessity of synthesis, and the empirical validity of the pure forms of\nunderstanding. It is as a condition of the possibility of\nself-consciousness that the objective employment of the categories is\nproved to be legitimate. _In the development of the deduction_\nself-consciousness is, therefore, more original than consciousness of\nobjects. Kant\u2019s employment of the term is, however, extremely\nmisleading. For it would seem to imply that the self has been proved to\nbe original or ultimate in an ontological sense, as if it preceded\nexperience, and through its antecedent reality rendered objective\nexperience possible of achievement. Such a view is undoubtedly\nreinforced by Kant\u2019s transformation of apperception into a faculty--_das\nRadicalverm\u00f6gen aller unsrer Erkenntniss_[914]--and his consequent\nidentification of it with the understanding.[915] It then seems as if he\nwere maintaining that the transcendental ego is ultimate and is\nindependent of all conditions, and that to its synthetic activities the\nvarious forms of objective consciousness are due.[916]\n\nThis unfortunate phraseology is directly traceable to the spiritualistic\nor Leibnizian character of Kant\u2019s earlier standpoint. In the\n_Dissertation_ the self is viewed as an ultimate and unconditioned\nexistence, antecedent to experience and creatively generative of it. We\nhave already noted that a somewhat similar view is presented in the\n_Critique_ in those paragraphs which Vaihinger identifies as embodying\nthe earliest stage in the development of the argument of the deduction.\nThe self is there described as coming to consciousness of its permanence\nthrough reflection upon the constancy of its own synthetic activities.\nOur consciousness of a transcendental object, and even the possibility\nof the empirical concepts through which such consciousness is, in these\nparagraphs, supposed to be mediated, are traced to this same source. To\nthe last this initial excess of emphasis upon the unity of apperception\nremained characteristic of Kant\u2019s Critical teaching; and though in the\nlater statements of his theory, its powers and prerogatives were very\ngreatly diminished, it still continued to play a somewhat exaggerated\nr\u00f4le. The early spiritualistic views were embodied in a terminology\nwhich he continued to employ; and unless the altered meaning of his\nterms is recognised and allowed for, misunderstanding is bound to\nresult. The terms, having been forged under the influence of the older\nviews, are but ill adapted to the newer teaching which they are employed\nto formulate.\n\nThere was also a second influence at work. When Kant was constrained in\nthe light of his new and unexpected results to recognise his older views\nas lacking in theoretical justification, he still held to them in his\nown personal thinking. For there is ample evidence that they continued\nto represent his _Privatmeinungen_.[917]\n\nOnly, therefore, when these misleading influences, verbal, expository,\nand personal, are discounted, do the results of the deduction appear in\ntheir true proportions. Kant\u2019s Critical philosophy does not profess to\nprove that it is self-consciousness, or apperception, or a\ntranscendental ego, or anything describable in kindred terms, which\nultimately renders experience possible. The most that we can\nlegitimately postulate, as noumenally conditioning experience, are\n\u201csyntheses\u201d (themselves, in their generative character, not\ndefinable)[918] in accordance with the categories. For only upon the\ncompletion of such syntheses do consciousness of self and consciousness\nof objects come to exist. Consciousness of objects does, indeed,\naccording to the argument of the deduction, involve consciousness of\nself; self-consciousness is the form of all consciousness. But, by the\nsame argument, it is equally true that only in and through\nconsciousness of objects is any self-consciousness possible at all.\nConsciousness of self and consciousness of objects _mutually_ condition\none another. Only through consciousness of both simultaneously can\nconsciousness of either be attained. Self-consciousness is not\ndemonstrably in itself any more ultimate or original than is\nconsciousness of objects. Both alike are forms of experience which are\nconditioned in complex ways. Upon the question as to whether or not\nthere is any such thing as abiding personality, the transcendental\ndeduction casts no direct light. Indeed consciousness of self, as the\nmore inclusive and complex form of awareness, may perhaps be regarded as\npointing to a greater variety of contributory and generative conditions.\n\nUnfortunately Kant, for the reasons just stated, has not sufficiently\nemphasised this more negative, or rather noncommittal, aspect of the\nresults of the deduction. But when later in the chapter on the\n_Paralogisms_ he is brought face to face with the issue, and has\noccasion to pronounce upon the question, he speaks with no uncertain\nvoice. In the theoretical sphere there is, he declares, no sufficient\nproof of the spirituality, or unitary and ultimate character, of the\nself. Like everything else the unity of apperception must be noumenally\nconditioned, but it cannot be shown that in itself, _as\nself-consciousness or apperception_, it represents any noumenal reality.\nIt may be a resultant, resting upon, and due to, a complexity of\ngenerative conditions; and these conditions may be fundamentally\ndifferent in character from itself. They may, for all that we can prove\nto the contrary, be of a non-conscious and non-personal nature. There is\nnothing in our cognitive experience, and no result of the Critical\nanalysis of it, which is inconsistent with such a possibility.[919]\nThose commentators, such as Cohen, Caird, and Watson, who more or less\nfollow Hegel in his criticism of Kant\u2019s procedure, give an\ninterpretation of the transcendental deduction which makes it\ninconsistent with the sceptical conclusions which the _Critique_ as a\nwhole is made by its author to support. Unbiassed study of the\n_Analytic_, even if taken by itself in independence of the _Dialectic_,\ndoes not favour such a view. The argument of the transcendental\ndeduction itself justifies no more than Kant is willing to allow in his\ndiscussion of the nature of the self in the section on the\n_Paralogisms_. It may, indeed, as Caird has so forcibly shown in his\nmassive work upon the Critical philosophy, be developed upon Hegelian\nlines, but only through a process of essential reconstruction which\ndeparts very far from many of Kant\u2019s most cherished tenets, and which\ndoes so in a spirit that radically conflicts with that which dominates\nthe _Critique_ as a whole.\n\n\nTHE LATER STAGES OF THE SUBJECTIVE DEDUCTION\n\nThe reader will have noted that several of the factors in Kant\u2019s\nexposition have so far been entirely ignored. The time has now come for\nreckoning with them. They constitute, in my view, the later stages of\nthe subjective deduction. That is to say, they refer to the\ntranscendental generative powers which Kant, _on the strength of the\nresults obtained in the more objective enquiry_, feels justified in\npostulating. Separate consideration of them tends to clearness of\nstatement. Kant\u2019s constant alternation between the logical and the\ndynamical standpoints is one of the many causes of the obscurity in his\nargument. In this connection we shall also find opportunity to discuss\nthe fundamental conflict, to which I have already had occasion to refer,\nbetween the subjectivist and the phenomenalist modes of developing the\nCritical standpoint.\n\nThe conclusions arrived at in the objective deduction compelled Kant to\nrevise his previous psychological views. Hitherto he had held to the\nLeibnizian theory that _a priori_ concepts are obtained by reflection\nupon the mind\u2019s native and fundamental modes of action. In the\n_Dissertation_ he carefully distinguishes between the _logical_ and the\n_real_ employment of the understanding. Through the former empirical\nconcepts are derived from concrete experience. Through the latter pure\nconcepts are creatively generated. Logical and real thinking agree,\nhowever, Kant there argues, in being activities of the _conscious_ mind.\nBoth can be apprehended and adequately determined through the revealing\npower of reflective consciousness. Such a standpoint is no longer\ntenable for Kant. Now that he has shown that the consciousness of self\nand the consciousness of objects mutually condition one another, and\nthat until both are attained neither is possible, he can no longer\nregard the mind as even possibly conscious of the activities whereby\nexperience is brought about. The activities generative of consciousness\nhave to be recognised as themselves falling outside it. Not even in its\npenumbra, through some vague form of apprehension, can they be detected.\nOnly the finished products of such activities, not the activities\nthemselves, can be presented to consciousness; and only by general\nreasoning, inferential of agencies that lie outside the conscious\nfield, can we hope to determine them.\n\nNow Kant appears to have been unwilling to regard the \u2018understanding\u2019 as\never unconscious of its activities. Why he was unwilling, it does not\nseem possible to explain; at most his rationalist leanings and Wolffian\ntraining may be cited as contributing causes. To the end he continued to\nspeak of the understanding as the faculty whereby the _a priori_ is\nbrought to consciousness. In order to develop the distinctions demanded\nby the new Critical attitude, he had therefore to introduce a new\nfaculty, capable of taking over the activities which have to be\nrecognised as non-conscious. For this purpose he selected the\nimagination, giving to it the special title, _productive_ imagination.\nThe empirical reproductive processes hitherto alone recognised by\npsychologists are not, he declares, exhaustive of the nature of the\nimagination. It is also capable of _transcendental_ activity, and upon\nthis the \u201cobjective affinity\u201d of appearances and the resulting\npossibility of their empirical apprehension is made to rest. The\nproductive imagination is also viewed as rendering possible the\nunderstanding, that is, the conscious apprehension of the _a priori_ as\nan element embedded in objective experience. Such apprehension is\npossible because in the pre-conscious elaboration of the given manifold\nthe productive imagination has conformed to those _a priori_ principles\nwhich the understanding demands for the possibility of its own exercise\nin conscious apprehension. Productive imagination acts in the manner\nrequired to yield experiences which are capable of relation to the unity\nof self-consciousness, _i.e._ of being found to conform to the unity of\nthe categories. Why it should act in this manner cannot be explained;\nbut it is none the less, on Critical principles, a legitimate\nassumption, since only in so far as it does so can experience, which _de\nfacto_ exists, be possible in any form. As a condition _sine qua non_ of\nactual and possible experience, the existence of such a faculty is, Kant\nargues, a legitimate inference from the results of the transcendental\ndeduction.\n\nThough Kant\u2019s insistence upon the conscious character of understanding\ncompels him to distinguish between it and the imagination, he has also\nto recognise their kinship. If imagination can never act save in\nconformity with the _a priori_ forms of understanding, some reason must\nexist for their harmony. This twofold necessity of at once\ndistinguishing and connecting them is the cause of the hesitating and\nextremely variable account which in both editions of the _Critique_ is\ngiven of their relation. In several passages the understanding is\nspoken of as simply imagination which has attained to consciousness of\nits activities.[920] Elsewhere he explicitly states that they are\ndistinct and separate. From this second point of view Kant regards\nimagination as mediating between sense and understanding, and, though\nreducible to neither, akin to both.\n\nOnly on one point is Kant clear and definite, namely, that it is to\nproductive imagination that the _generation_ of unified experience is\nprimarily due. In it something of the fruitful and inexhaustible\ncharacter of noumenal reality is traceable. Doubtless one chief reason\nfor his choice of the title imagination is the creative character which\nin popular thought has always been regarded as its essential feature. As\nKant, speaking of schematism, which is a process executed by the\nimagination, states in A 141: \u201cThis schematism ... is an _art_ (_Kunst_)\nconcealed in the depths of the human soul.\u201d[921] This description may\nperhaps be interpreted in the light of Kant\u2019s account of the creative\ncharacter of artistic genius in the _Critique of Judgment_, for there\nalso imagination figures as the truly originative or creative faculty of\nthe human spirit. To its noumenal character we may also trace its\ncapacity of combining those factors of sense and understanding which in\nthe realm of appearance remain persistently opposed.[922] Imagination\ndiffers from the understanding chiefly in that it is at once more\ncomprehensive and also more truly creative. It supplements the\nfunctional forms with a sensuous content, and applies them dynamically\nin the generation of experience.\n\nThe schemata, which the productive imagination is supposed to construct,\nare those generalised forms of temporal and spatial existence in which\nalone the unity of experience necessary to apperception can be realised.\nThey are\n\n     \u201cpure (without admixture of anything empirical), and yet are in one\n     aspect intellectual and in another sensuous.\u201d[923]\n\nOr as Kant describes the process in the chapter before us:[924]\n\n     \u201cWe name the synthesis of the manifold in imagination\n     transcendental, if without distinction of intuitions it is directed\n     exclusively to the _a priori_ combination of the manifold; and the\n     unity of this synthesis is entitled transcendental, if it is\n     represented as _a priori_ necessary in relation to the original\n     unity of apperception. As this unity of apperception conditions\n     the possibility of all knowledge, the transcendental unity of the\n     synthesis of imagination is the pure form of all possible\n     knowledge. Hence, through it all objects of possible experience\n     must be represented _a priori_.\u201d\n\nThe schemata, thus transcendentally generated, are represented by Kant\nas limiting and controlling the empirical processes of apprehension,\nreproduction, and recognition. As no experience is attainable save in\nterms of the schemata, they enable us to determine, on _a priori_\ngrounds, the degree of constancy and regularity that can be securely\ncounted upon in all experience. This is Kant\u2019s psychological explanation\nof what he has entitled \u201cobjective affinity.\u201d[925] The empirical ground\nof reproduction is the association of ideas; its transcendental ground\nis an objective affinity which is \u201ca necessary consequence of a\nsynthesis in imagination, grounded _a priori_ on rules.\u201d[926]\n\n     \u201d[The] subjective and empirical ground of reproduction according to\n     rules is named the association of representations. If this unity of\n     association did not also have an objective ground, which makes it\n     impossible that appearances should be apprehended by the\n     imagination except under the condition of a possible synthetic\n     unity of this apprehension, it would be entirely accidental that\n     appearances should fit into a connected whole of human\n     knowledge.... There might exist a multitude of perceptions, and\n     indeed an entire sensibility, in which much empirical consciousness\n     would arise in my mind, but in a state of separation, and without\n     belonging to one consciousness of myself. That, however, is\n     impossible.\u201d [As the subjective and objective deductions have\n     demonstrated, where there is no self-consciousness there is no\n     consciousness of any kind.] \u201cThere must therefore be an objective\n     ground (that is, one that can be determined _a priori_,\n     antecedently to all empirical laws of the imagination) upon which\n     may rest the possibility, nay, the necessity of a law that extends\n     to all appearances--the law, namely, that all appearances must be\n     regarded as data of the senses which are associable in themselves\n     and subject to general rules of universal connection in their\n     reproduction. This objective ground of all association of\n     appearances I entitle their _affinity_.... The objective unity of\n     all empirical consciousness in one consciousness, that of original\n     apperception, is the necessary condition of all possible\n     perception; and the affinity of all appearances, near or remote, is\n     a necessary consequence of a synthesis in imagination which is\n     grounded _a priori_ on rules.\u201d[927]\n\nThis part of Kant\u2019s teaching is apt to seem more obscure than it is. For\nthe reader is not unnaturally disinclined to accept it in the very\nliteral sense in which it is stated. That Kant means, however, exactly\nwhat he says, appears from the further consequence which he himself not\nonly recognises as necessary, but insists upon as valid. The doctrine of\nobjective affinity culminates in the conclusion[928] that it is \u201cwe\nourselves who introduce into the appearances that order and regularity\nwhich we name nature.\u201d The \u201cwe ourselves\u201d refers to the mind in the\ntranscendental activities of the productive imagination. The conscious\nprocesses of apprehension, reproduction, and recognition necessarily\nconform to schemata, non-consciously generated, which express the\ncombined _a priori_ conditions of intuition and understanding required\nfor unitary consciousness.\n\nMany points in this strange doctrine call for consideration. It rests,\nin the first place, upon the assumption of a hard and fast distinction,\nvery difficult of acceptance, between transcendental and empirical\nactivities of the mind. Secondly, Kant\u2019s assertion, that the empirical\nmanifolds can be relied upon to supply a satisfactory content for the\nschemata, calls for more adequate justification than he himself adduces.\nIt is upon independent reality that the fixity of empirical\nco-existences and sequences depends. Is not Kant practically assuming a\npre-established harmony in asserting that as the mind creates the _form_\nof nature it can legislate _a priori_ for all possible experience?\n\nAs regards the first assumption Kant would seem to have been influenced\nby the ambiguities of the term transcendental. It means, as we have\nalready noted,[929] either the science of the _a priori_, or the _a\npriori_ itself, or the conditions which render experience possible. Even\nthe two latter meanings by no means coincide. The conditions of the\npossibility of experience are not in all cases _a priori_. The manifold\nof outer sense is as indispensable a precondition of experience as are\nthe forms of understanding, and yet is not _a priori_ in any valid sense\nof that term. It does not, therefore, follow that because the activities\nof productive imagination \u201ctranscendentally\u201d condition experience, they\nmust themselves be _a priori_, and must, as Kant also maintains,[930]\ndeal with a pure _a priori_ manifold. Further, the separation between\ntranscendental and empirical activities of the mind must defeat the very\npurpose for which the productive imagination is postulated, namely, in\norder to account for the generation of a complex consciousness in which\nno one element can temporally precede any of the others. If the\nproductive imagination generates only schemata, it will not account for\nthat complex experience in which consciousness of self and consciousness\nof objects are indissolubly united. The introduction of the productive\nimagination seems at first sight to promise recognition of the dynamical\naspect of our temporally sequent experience, and of that aspect in which\nas appearance it refers us beyond itself to non-experienced conditions.\nAs employed, however, in the doctrines of schematism and of objective\naffinity, the imagination exhibits a formalism hardly less extreme than\nthat of the understanding whose shortcomings it is supposed to make\ngood.\n\nIn his second assumption Kant, as so often in the _Critique_, is\nallowing his old-time rationalistic leanings to influence him in\nunderestimating the large part which the purely empirical must always\noccupy in human experience, and in exaggerating the scope of the\ninferences which can be drawn from the presence of the formal,\nrelational factors. But this is a point which we are not yet in a\nposition to discuss.[931]\n\nFortunately, if Vaihinger\u2019s theory be accepted,[932] section A 98-104\nenables us to follow the movement of Kant\u2019s mind in the interval between\nthe formulating of the doctrine of productive imagination and the\npublication of the _Critique_. He himself would seem to have recognised\nthe unsatisfactoriness of dividing up the total conditions of experience\ninto transcendental activities that issue in schemata, and supplementary\nempirical processes which transform them into concrete, specific\nconsciousness. The alternative theory which he proceeds to propound is\nat first sight much more satisfactory. It consists in duplicating each\nof the various empirical processes with a transcendental faculty. There\nare, he now declares, three transcendental powers--a transcendental\nfaculty of apprehension, a transcendental faculty of reproduction\n(=imagination), and a transcendental faculty of recognition. Thus Kant\u2019s\nprevious view that transcendental imagination has a special and unique\nactivity, namely, the productive, altogether different in type from any\nof its empirical processes, is now allowed to drop; in place of it Kant\ndevelops the view that the transcendental functions run exactly parallel\nwith the empirical processes.[933] But though such a position may at\nfirst seem more promising than that which it displaces, it soon reveals\nits unsatisfactoriness. The two types of mental activity, transcendental\nand empirical, no longer, indeed, fall apart; but the difficulty now\narises of distinguishing in apprehension, reproduction, and recognition\nany genuinely transcendental aspect.[934] Apprehension, reproduction,\nand recognition are so essentially conscious processes that to view them\nas also transcendental does not seem helpful. They contain elements that\nare transcendental in the logical sense, but cannot be shown to\npresuppose in any analogous fashion mental powers that are\ntranscendental in the dynamical sense. This is especially evident in\nregard to recognition, which is described as being \u201cthe _consciousness_\nthat what we are thinking is the same as what we thought a moment\nbefore.\u201d In dealing with apprehension and reproduction the only real\ndifference which Kant is able to suggest, as existing between their\ntranscendental and their empirical activities, is that the former\nsynthesise the pure _a priori_ manifolds of space and time, and the\nlatter the contingent manifold of sense. But even this unsatisfactory\ndistinction he does not attempt to apply in the case of recognition. Nor\ncan we hold that by the transcendental synthesis of recognition Kant\nmeans transcendental apperception. That is, of course, the suggestion\nwhich at once occurs to the reader. But however possible it might be to\ninject such a meaning into kindred passages elsewhere, it cannot be made\nto fit the context of this particular section.\n\nVaihinger\u2019s theory seems to be the only thread which will guide us\nthrough this labyrinth. Kant, on the eve of the publication of the\n_Critique_, recognising the unsatisfactoriness of his hard and fast\nseparation of transcendental from empirical processes, adopted the view\nthat some form of transcendental activity corresponds to every\nfundamental form of empirical activity and _vice versa_. Hastily\ndeveloping this theory, he incorporated it into the _Critique_ alongside\nhis older doctrine. It does not, however, reappear in the _Prolegomena_,\nand its teaching is explicitly withdrawn in the second edition of the\n_Critique_. Its plausibility had entrapped him into its temporary\nadoption, but the defects which it very soon revealed speedily led him\nto reject it.\n\nOne feature of great significance calls for special notice. The\nbreakdown of this doctrine of a threefold transcendental synthesis did\nnot, as might naturally have been expected from what is stated in the\nprefaces to the _Critique_ regarding the unessential and seemingly\nconjectural character of the subjective deduction, lead Kant to despair\nof developing a transcendental psychology. Though in the second edition\nhe cuts away the sections containing the earlier stages of the\nsubjective deduction,[935] and in recasting the other sections gives\ngreater prominence to the more purely logical analyses, the older\ndoctrine of productive imagination is reinstated in full force,[936] and\nis again developed in[937] connection with the doctrine of pure _a\npriori_ manifolds. Evidently, therefore, Kant was not disheartened by\nthe various difficulties which lie in the path of a transcendental\npsychology, and it seems reasonable to conclude that there were powerful\nreasons inclining him to its retention. I shall now attempt, to the best\nof my powers, to explain--the task is a delicate and difficult one--what\nwe may believe these reasons to have been.[938]\n\n\nTHE DISTINCTION BETWEEN PHENOMENALISM AND SUBJECTIVISM\n\nA wider set of considerations than we have yet taken into account must\nbe borne in mind if certain broader and really vital implications of\nKant\u2019s enquiry are to be properly viewed. The self has a twofold aspect.\nIt is at once animal in its conditions and potentially universal in its\npowers of apprehension. Though man\u2019s natural existence is that of an\nanimal organism, he can have consciousness of the spatial world out of\nwhich his organism has arisen, and of the wider periods within which his\ntransitory existence falls. Ultimately such consciousness would seem to\nconnect man cognitively with reality as a whole. Now it is to this\nuniversal or absolutist aspect of our consciousness, to its\ntranscendence of the embodied and separate self, that Kant is seeking to\ndo justice in his transcendental deductions, especially in his doctrine\nof the transcendental unity of apperception. For he views that\napperception as conditioned by, and the correlate of, the consciousness\nof objectivity. It involves the consciousness of a single cosmical time\nand of a single cosmical space within which all events fall and within\nwhich they form a whole of causally interdependent existences. That is\nwhy he names it the _objective_ unity of apperception. It is that aspect\nin which the self correlates with a wider reality, and through which it\nstands in fundamental contrast to the merely subjective states and to\nthe individual conditions of its animal existence. The transcendental\nself, so far from being identical with the empirical self, would seem to\nbe of directly opposite nature. The one would seem to point beyond the\nrealm of appearance, the other to be in its existence merely natural.\nThe fact that they are inextricably bound up with one another, and\nco-operate in rendering experience possible, only makes the more\nindispensable the duty of recognising their differing characters. Even\nshould they prove to be inseparable aspects of sense-experience, without\nmetaphysical implications, that would not obviate the necessity of\nclearly distinguishing them. The distinction remains, whatever\nexplanation may be adopted of its speculative or other significance.\n\nNow obviously in so fundamental an enquiry, dealing as it does with the\nmost complicated and difficult problem in the entire field of\nmetaphysics, no brief and compendious answer can cover all the various\nconsiderations which are relevant and determining. The problem of the\ndeduction being what it is, the section dealing with it can hardly fail\nto be the most difficult portion of the whole _Critique_. The\nconclusions at which it arrives rest not merely upon the argument which\nit contains but also upon the results more or less independently reached\nin the other sections. The doctrine of the empirical object as\nappearance requires for its development the various discussions\ncontained in the _Aesthetic_, in the sections on _Inner Sense_ and on\nthe _Refutation of Idealism_, in the chapters on _Phenomena and Noumena_\nand on the _Antinomies_. The metaphysical consequences and implications\nof Kant\u2019s teaching in regard to the transcendental unity of apperception\nare first revealed in the chapter on the _Paralogisms_. The view taken\nof productive imagination is expanded in the section on _Schematism_. In\na word, the whole antecedent teaching of the _Critique_ is focussed, and\nthe entire subsequent development of the Critical doctrine is\nanticipated, in this brief chapter.\n\nBut there are, of course, additional causes of the difficulty and\nobscurity of the argument. One such cause has already been noted,\nnamely, that the _Critique_ is not a unitary work, developed from a\npreviously thought-out standpoint, but in large part consists of\nmanuscripts of very various dates, artificially pieced together by the\naddition of connecting links. In no part of the _Critique_ is this so\nobvious as in the _Analytic of Concepts_. Until this is recognised all\nattempts to interpret the text in any impersonal fashion are doomed to\nfailure. For this reason I have prefaced our discussion by a statement\nof Vaihinger\u2019s analysis. No one who can accept it is any longer in\ndanger of underestimating this particular cause of the obscurity of\nKant\u2019s deduction.\n\nBut the chief reason is one to which I have thus far made only passing\nreference, and to which we may now give the attention which its\nimportance demands, namely, the tentative and experimental character of\nKant\u2019s own final solutions. The arguments of the deduction are only\nintelligible if viewed as an expression of the conflicting tendencies to\nwhich Kant\u2019s thought remained subject. He sought to allow due weight to\neach of the divergent aspects of the experience which he was analysing,\nand in so doing proceeded, as it would seem, simultaneously along the\nparallel lines of what appeared to be the possible, alternative methods\nof explanation. And to the end these opposing tendencies continued side\nby side, to the confusion of those readers who seek for a single unified\nteaching, but to the great illumination of those who are looking to\nKant, not for clear-cut or final solutions, but for helpful analysis and\nfor partial disentanglement of the complicated issues which go to\nconstitute these baffling problems.\n\nThe two chief tendencies which thus conflicted in Kant\u2019s mind may be\nnamed the subjectivist and the phenomenalist respectively. This conflict\nremained, so to speak, underground, influencing the argument at every\npoint, but seldom itself becoming the subject of direct discussion. As\nwe shall find, it caused Kant to develop a twofold view of inner sense,\nof causality, of the object of knowledge, and of the unity of\napperception. One of the few sections in the _Critique_ where it seems\non the point of emerging into clear consciousness is the section, added\nin the second edition, on the _Refutation of Idealism_. But this section\nowes its origin to polemical causes. It represents a position peculiar\nto the maturer portions of the _Analytic_; the rest of the _Critique_ is\nnot rewritten so as to harmonise with it, or to develop the consequences\nwhich consistent holding to it must involve.\n\nI shall use the term _subjectivism_ (and its equivalent _subjective\nidealism_) in the wide sense[939] which makes it applicable to the\nteaching of Descartes and Locke, of Leibniz and Wolff, no less than to\nthat of Berkeley and Hume. A common element in all these philosophies is\nthe belief that subjective or mental states, \u201cideas\u201d in the Lockean\nsense, are the objects of consciousness, and further are the sole\npossible objects of which it can have any direct or immediate awareness.\nKnowledge is viewed as a process entirely internal to the individual\nmind, and as carrying us further only in virtue of some additional\nsupervening process, inferential, conjectural, or instinctive. This\nsubjectivism also tends to combine with a view of consciousness as an\nultimate self-revealing property of a merely individual existence.[940]\nFor Descartes consciousness is the very essence, both of the mind and of\nthe self. It is indeed asserted to be exhaustive of the nature of both.\nThough the self is described as possessing a faculty of will as well as\na power of thinking, all its activities are taken as being disclosed to\nthe mind through the revealing power of its fundamental attribute. The\nindividual mind is thus viewed as an existence in which everything takes\nplace in the open light of an all-pervasive consciousness. Leibniz, it\nis true, taught the existence of subconscious perceptions, and so far\nmay seem to have anticipated Kant\u2019s recognition of non-conscious\nprocesses; but as formulated by Leibniz that doctrine has the defect\nwhich frequently vitiates its modern counterpart, namely that it\nrepresents the subconscious as analogous in nature to the conscious, and\nas differing from it only in the accidental features of intensity and\nclearness, or through temporary lack of control over the machinery of\nreproductive association. The subconscious, as thus represented, merely\nenlarges the private content of the individual mind; it in no respect\ntranscends it.\n\nThe genuinely Critical view of the generative conditions of experience\nis radically different from this Leibnizian doctrine of _petites\nperceptions_. It connects rather with Leibniz\u2019s mode of conceiving the\norigin of _a priori_ concepts. But even that teaching it restates in\nsuch fashion as to free it from subjectivist implications. Leibniz\u2019s\ncontention that the mind is conscious of its fundamental activities, and\nthat it is by reflection upon them that it gains all ultimate _a priori_\nconcepts, is no longer tenable in view of the conclusions established in\nthe objective deduction. Mental processes, in so far as they are\ngenerative of experience, must fall outside the field of consciousness,\nand as activities dynamically creative cannot be of the nature of ideas\nor contents. They are not subconscious ideas but non-conscious\nprocesses. They are not the submerged content of experience, but its\nconditioning grounds. Their most significant characteristic has still,\nhowever, to be mentioned. They must no longer be interpreted in\nsubjectivist terms, as originating in the separate existence of an\nindividual self. In conditioning experience they generate the only self\nfor which experience can vouch, and consequently, in the absence of full\nand independent proof, must not be conceived as individually\ncircumscribed. The problem of knowledge, properly conceived, is no\nlonger how consciousness, individually conditioned, can lead us beyond\nits own bounds, but what a consciousness, which is at once consciousness\nof objects and also consciousness of a self, must imply for its\npossibility. Kant thus obtains what is an almost invariable concomitant\nof scientific and philosophical advance, namely a more correct and\nscientific formulation of the problem to be solved. The older\nformulation assumes the truth of the subjectivist standpoint; the\nCritical problem, when thus stated, is at least free from preconceptions\nof that particular brand. Assumptions which hitherto had been quite\nunconsciously held, or else, if reflected upon, had been regarded as\naxiomatic and self-evident, are now brought within the field of\ninvestigation. Kant thereby achieves a veritable revolution; and with it\nmany of the most far-reaching consequences of the Critical teaching are\nclosely bound up.\n\nThis new standpoint, in contrast to _subjective_ idealism, may be named\n_Critical_, or to employ the term which Kant himself applies both to his\ntranscendental deduction and to the unity of apperception, _objective_\nidealism. But as the distinction between appearance and reality is no\nless fundamental to the Critical attitude, we shall perhaps be less\nlikely to be misunderstood, or to seem to be identifying Kant\u2019s\nstandpoint with the very different teaching of Hegel, if by preference\nwe employ the title _phenomenalism_.\n\nIn the transcendental deduction Kant, as above noted, is seeking to do\njustice to the universal or absolutist aspect of our consciousness, to\nits transcendence of the embodied and separate self. The unity of\napperception is entitled _objective_, because it is regarded as the\ncounterpart of a single cosmical time and of a single cosmical space\nwithin which all events fall. Its objects are not mental states peculiar\nto itself, nor even ideal contents numerically distinct from those in\nother minds. It looks out upon a common world of genuinely independent\nexistence. In developing this position Kant is constrained to revise and\nindeed completely to recast his previous views both as to the nature of\nthe synthetic processes, through which experience is constructed, and of\nthe given manifold, upon which they are supposed to act. From the\nsubjectivist point of view the synthetic activities consist of the\nvarious cognitive processes of the individual mind, and the given\nmanifold consists of the sensations aroused by material bodies acting\nupon the special senses. From the objective or phenomenalist standpoint\nthe synthetic processes are of a noumenal character, and the given\nmanifold is similarly viewed as being due to noumenal agencies acting,\nnot upon the sense-organs, which as appearances are themselves\nnoumenally conditioned, but upon what may be called \u201couter sense.\u201d These\ndistinctions may first be made clear.\n\nSensations, Kant holds, have a twofold origin, noumenal and mechanical.\nThey are due in the first place to the action of things in themselves\nupon the noumenal conditions of the self, and also in the second place\nto the action of material bodies upon the sense-organs and the brain. To\ntake the latter first. Light reflected from objects, and acting on the\nretina, gives rise to sensations of colour. For such causal\ninterrelations there exists, Kant teaches, the same kind of empirical\nevidence as for the causal interaction of material bodies.[941] Our\nsensational experiences are as truly events in time as are mechanical\nhappenings in space. In this way, however, we can account only for the\nexistence of our sensations and for the order in which they make their\nappearance in or to consciousness, not for our awareness of them. To\nstate the point by means of an illustration. The impinging of one\nbilliard ball upon another accounts causally for the motion which then\nappears in the second ball. But no one would dream of asserting that by\nitself it accounts for our consciousness of that second motion. We may\ncontend that in an exactly similar manner, to the same extent, no more\nand no less, the action of an object upon the brain accounts only for\nthe occurrence of a visual sensation as an event in the empirical time\nsequence. A sensation just as little as a motion can carry its own\nconsciousness with it. To regard that as ever possible is ultimately to\nendow events in time with the capacity of apprehending objects in space.\nIn dealing with causal connections in space and time we do not require\nto discuss the problem of knowledge proper, namely, how it is possible\nto have or acquire knowledge, whether of a motion in space or of a\nsensation in time. When we raise that further question we have to adopt\na very different standpoint, and to take into account a much greater\ncomplexity of conditions.\n\nKant applies this point of view no less rigorously to feelings,\nemotions, and desires than to the sensations of the special senses. All\nof them, he teaches, are \u2018animal\u2019[942] in character. They are one and\nall conditioned by, and explicable only in terms of, the particular\nconstitution of the animal organism. They one and all belong to the\nrealm of appearance.[943]\n\nThe term \u2018sensation\u2019 may also, however, be applied in a wider sense to\nsignify the material of knowledge in so far as it is noumenally\nconditioned. Thus viewed, sensations are due, not to the action of\nphysical stimuli upon the bodily organs, but to the affection by things\nin themselves of those factors in the noumenal conditions of the self\nwhich correspond to \u201csensibility.\u201d Kant is culpably careless in failing\nto distinguish those two very different meanings of the phrase \u2018given\nmanifold.\u2019 The language which he employs is thoroughly ambiguous. Just\nas he frequently speaks as if the synthetic processes were conscious\nactivities exerted by the self, so also he frequently uses language\nwhich implies that the manifold upon which these processes act is\nidentical with the sensations of the special senses. But the sensations\nof the bodily senses, even if reducible to it, can at most form only\npart of it. The synthetic processes, interpreting the manifold in\naccordance with the fixed forms, space, time, and the categories,\ngenerate the spatial world within which objects are apprehended as\ncausally interacting and as giving rise through their action upon the\nsense-organs to the various special sensations as events in time.\nSensations, as mechanically caused, are thus on the same plane as other\nappearances. They depend upon the same generating conditions as the\nmotions which produce them. As minor incidents within a more\ncomprehensive totality they cannot possibly represent the material out\nof which the whole has been constructed. To explain the phenomenal world\nas constructed out of the sensations of the special senses is virtually\nto equate it with a small selection of its constituent parts. Such\nprofessed explanation also commits the further absurdity of attempting\nto account for the origin of the phenomenal world by means of events\nwhich can exist only under the conditions which it itself supplies. The\nmanifold of the special senses and the primary manifold are radically\ndistinct. The former is due to material bodies acting upon the material\nsense-organs. The latter is the product of noumenal agencies acting upon\n\u201couter sense,\u201d _i.e._ upon those noumenal conditions of the self which\nconstitute our \u201csensibility\u201d; it is much more comprehensive than the\nformer; it must contain the material for all modes of objective\nexistence, including many that are usually regarded as purely\nmental.[944]\n\nTo turn, now, to the other aspect of experience. What are the factors\nwhich condition its form? What must we postulate in order to account for\nthe existence of consciousness and for the unitary form in which alone\nit can appear? Kant\u2019s answer is again ambiguous. He fails sufficiently\nto insist upon distinctions which yet are absolutely vital to any\ngenuine understanding of the new and revolutionary positions towards\nwhich he is feeling his way. The synthetic processes which in the\nsubjective and objective deductions are proved to condition all\nexperience may be interpreted either as conscious or as non-conscious\nactivities, and may be ascribed either to the agency of the individual\nself or to noumenal conditions which fall outside the realm of possible\ndefinition. Now, though Kant\u2019s own expositions remain thoroughly\nambiguous, the results of the Critical enquiry would seem--at least so\nlong as the fundamental distinction between matter and form is held to\nand the temporally sequent aspect of experience is kept in view--to be\ndecisive in favour of the latter alternative in each case. The synthetic\nprocesses must take place and complete themselves before any\nconsciousness can exist at all. And as they thus precondition\nconsciousness, they cannot themselves be known to be conscious; and not\nbeing known to be conscious, it is not even certain that they may\nlegitimately be described as mental. We have, indeed, to conceive them\non the analogy of our mental processes, but that may only be because of\nthe limitation of our knowledge to the data of experience. Further, we\nhave no right to conceive them as the activities of a noumenal self. We\nknow the self only as conscious, and the synthetic processes, being the\ngenerating conditions of consciousness, are also the generating\nconditions of the only self for which our experience can vouch. Kant,\nviewing as he does the temporal aspect of human experience as\nfundamental, would seem to be justified in naming these processes\n\u201csynthetic.\u201d For consciousness in its very nature would seem to involve\nthe carrying over of content from one time to other times, and the\nconstruction of a more comprehensive total consciousness from the\nelements thus combined. Kant is here analysing in its simplest and most\nfundamental form that aspect of consciousness which William James has\ndescribed in the _Principles of Psychology_,[945] and which we may\nentitle the telescoping of earlier mental states into the successive\nexperiences that include them. They telescope in a manner which can\nnever befall the successive events in a causal series, and which is not\nexplicable by any scheme of relations derivable from the physical\nsphere.\n\nObviously, what Kant does is to apply to the interpretation of the\nnoumenal conditions of our conscious experience a distinction derived by\nanalogy from conscious experience itself--the distinction, namely,\nbetween our mental processes and the sensuous material with which they\ndeal. The application of such a distinction may be inevitable in any\nattempt to explain human experience; but it can very easily, unless\ncarefully guarded, prove a source of serious misunderstanding. Just as\nthe synthetic processes which generate consciousness are not known to be\nthemselves conscious, so also the manifold cannot be identified with the\nsensations of the bodily senses. These last are events in time, and are\neffects not of noumenal but of mechanical causes.\n\nKant\u2019s conclusion when developed on consistent Critical lines, and\ntherefore in phenomenalist terms, is twofold: positive, to the effect\nthat consciousness, for all that our analysis can prove to the contrary,\nmay be merely a resultant, derivative from and dependent upon a\ncomplexity of conditions; and negative, to the effect that though these\nconditions may by analogy be described as consisting of synthetic\nprocesses acting upon a given material, they are in their real nature\nunknowable by us. Even their bare possibility we cannot profess to\ncomprehend. We postulate them only because given experience is\ndemonstrably not self-explanatory and would seem to refer us for\nexplanation to some such antecedent generative grounds.\n\nKant, as we have already emphasised, obscures his position by the way in\nwhich he frequently speaks of the transcendental unity of apperception\nas the supreme condition of our experience. At times he even speaks as\nif it were the source of the synthetic processes. That cannot, however,\nbe regarded as his real teaching. Self-consciousness (and the unity of\napperception, in so far as it finds expression through\nself-consciousness) rests upon the same complexity of conditions as does\nouter experience, and therefore may be merely a product or resultant. It\nis, as he insists in the _Paralogisms_, the emptiest of all our\nconcepts, and can afford no sufficient ground for asserting the self to\nbe an abiding personality. We cannot by theoretical analysis of the\nfacts of experience or of the nature of self-consciousness prove\nanything whatsoever in regard to the ultimate nature of the self.\n\nNow Kant is here giving a new, and quite revolutionary, interpretation\nof the distinction between the subjective and the objective. The\nobjective is for the Cartesians the independently real;[946] the\nsubjective is that which has an altogether different kind of existence\nin what is entitled the field of consciousness. Kant, on the other hand,\nfrom his phenomenalist standpoint, views existences as objective when\nthey are determined by purely physical causes, and as subjective when\nthey also depend upon physiological and psychological conditions. On\nthis latter view the difference between the two is no longer a\ndifference of kind; it becomes a difference merely of degree. Objective\nexistences, owing to the simplicity and recurrent character of their\nconditions, are uniform. Subjective existences resting upon conditions\nwhich are too complex to be frequently recurrent, are by contrast\nextremely variable. But both types of existence are objective in the\nsense that they are objects, and immediate objects, for consciousness.\nSubjective states do not run parallel with the objective system of\nnatural existences, nor are they additional to it. For they do not\nconstitute our consciousness of nature; they are themselves part of the\nnatural order which consciousness reveals. That they contrast with\nphysical existences in being unextended and incapable of location in\nspace is what Kant would seem by implication to assert, but he\nchallenges Descartes\u2019 right to infer from this particular difference a\ncomplete diversity in their whole nature. Sensations, feelings,\nemotions, and desires, so far as they are experienced by us, constitute\nthe empirical self which is an objective existence, integrally connected\nwith the material environment, in terms of which alone it can be\nunderstood. In other words, the distinction between the subjective and\nthe objective is now made to fall within the system of natural law. The\nsubjective is not opposite in nature to the objective, but is a\nsubspecies within it.\n\nThe revolutionary character of this reformulation of Cartesian\ndistinctions may perhaps be expressed by saying that what Kant is really\ndoing is to substitute the distinction between appearance and reality\nfor the Cartesian dualism of the mental and the material. The psychical\nis a title for a certain class of known existences, _i.e._ of\nappearances; and they form together with the physical a single system.\nBut underlying this entire system, conditioning both physical and\npsychical phenomena, is the realm of noumenal existence; and when the\nquestion of the possibility of knowledge, that is, of the experiencing\nof such a comprehensive natural system, is raised, it is to this\nnoumenal sphere that we are referred. Everything experienced, even a\nsensation or desire, is an event; but the experiencing of it is an act\nof awareness, and calls for an explanation of an altogether different\nkind.\n\nThus Kant completely restates the problem of knowledge. The problem is\nnot how, starting from the subjective, the individual can come to\nknowledge of the independently real; but how, if a common world is alone\nimmediately apprehended, the inner private life of the self-conscious\nbeing can be possible, and how such inner experience is to be\ninterpreted. How does it come about that though sensations, feelings,\netc., are events no less mechanically conditioned than motions in space,\nand constitute with the latter a single system conformed to natural law,\nthey yet differ from all other classes of natural events in that they\ncan be experienced only by a single consciousness. To this question Kant\nreplies in terms of his fundamental distinction between appearance and\nreality. Though everything of which we are conscious may legitimately be\nstudied in terms of the natural system to which it belongs,\nconsciousness itself cannot be so regarded. In attempting to define it\nwe are carried beyond the phenomenal to its noumenal conditions. In\nother words, it constitutes a problem, the complete data of which are\nnot at our disposal. This is by itself a sufficient reason for our\nincapacity to explain why the states of each empirical self can never be\napprehended save by a single consciousness, or otherwise stated, why\neach consciousness is limited, as regards sensations and feelings,\nexclusively to those which arise in connection with some one animal\norganism. It at least precludes us from dogmatically asserting that this\nis due to their being subjective in the dualistic and Cartesian sense of\nthat term--namely, as constituting, or being states of, the knowing\nself.\n\nA diagram may serve, though very crudely, to illustrate Kant\u2019s\nphenomenalist interpretation of the cognitive situation.\n\n                                    NC^EW\n                                      |\n                                      \\\n                                   /    \\\n               /------------------/------\\---------------------\\\n             / The empirical world of our human consciousness   \\\n           /  +------+------+       +---+      +--------------+  \\\n          |   |             |       | l |      |              |  |\n          |   |    ES^A     |       +---+      |     ES^B     |  |\n          |   |             |       +---+      |              |  |\n          |   |x^1 y^1 z^1  |       | m |      |  x^2 y^2 z^2 |  |\n          |   |             |       +---+      |              |  |\n          |   +-------------+       +---+      +--------------+  |\n          |                         | n |                        |\n           \\                        +---+                       /\n            \\--------------------------------------------------/\n             \\           /                       \\           /\n               \\       /                           \\       /\n                 \\   /                               \\   /\n                   V                                   V\n                  NC^A                                NC^B\n\n            ES^A = Empirical self of the conscious Being A.\n\n            ES^B = Empirical self of the conscious Being B.\n\n            NC^A = Noumenal conditions of the conscious Being A.\n\n            NC^B = Noumenal conditions of the conscious Being B.\n\n            l, m, n = Objects in space.\n\n            x^1, y^1, z^1 = Sensations caused by objects l, m, n\n            acting on the sense-organs of the empirical self A.\n\n            x^2, y^2, z^2= Sensations caused by 1, m, n acting\n            on the sense-organs of the empirical self B.\n\n            NC^EW = Noumenal conditions of the empirical world.\n\nEverything in this empirical world is equally open to the consciousness\nof both A and B, save only certain psychical events that are conditioned\nby physiological and psychological factors. x^1, y^1, z^1 can be\napprehended only by A; x^2, y^2, z^2 can be apprehended only by B.\nOtherwise A and B experience one and the same world; the body of B is\nperceived by A in the same manner in which he perceives his own body.\nThis is true _a fortiori_ of all other material existences. Further,\nthese material existences are known with the same immediacy as the\nsubjective states. As regards the relation in which NC^A, NC^B, and\nNC^EW stand to one another, no assertions can be made, save, as above\nindicated,[947] such conjectural statements as may precariously be\nderived through argument by analogy from distinctions that fall within\nour human experience.[948]\n\nKant\u2019s phenomenalism thus involves an objectivist view of individual\nselves and of their interrelations. They fall within the single common\nworld of space. Within this phenomenal world they stand in external,\nmechanical relations to one another. They are apprehended as embodied,\nwith known contents, sensations, feelings, and desires, composing their\ninner experience. There is, from this point of view, no problem of\nknowledge. On this plane we have to deal only with events known, not\nwith any process of apprehension. Even the components of the empirical\nself, the subject-matter of empirical psychology, are not processes of\napprehension, but apprehended existences. It is only when we make a\nregress beyond the phenomenal as such to the conditions which render it\npossible, that the problem of knowledge arises at all. And with this\nregress we are brought to the real crux of the whole question--the\nreconciliation of this phenomenalism with the conditions of our\nself-consciousness. For we have then to take into account the\nfundamental fact that each self is not only an animal existence within\nthe phenomenal world, but also in its powers of apprehension coequal\nwith it. The self known is external to the objects known; the self that\nknows is conscious of itself as comprehending within the field of its\nconsciousness the wider universe in infinite space.\n\nSuch considerations would, at first sight, seem to force us to modify\nour phenomenalist standpoint in the direction of subjectivism. For in\nwhat other manner can we hope to unite the two aspects of the self, the\nknown conditions of its finite existence and the consciousness through\nwhich it correlates with the universe as a whole? In the one aspect it\nis a part of appearance; in the other it connects with that which makes\nappearance possible at all.\n\nQuite frequently it is the subjectivist solution which Kant seems to\nadopt. Objects known are \u201cmere representations,\u201d \u201cstates of the\nidentical self.\u201d Everything outside the individual mind is real;\nappearances are purely individual in origin. But such a position is\ninconsistent with the deeper implications of Kant\u2019s Critical teaching,\nand would involve the entire ignoring of the many suggestions which\npoint to a fundamentally different and much more adequate standpoint.\nThe individual is himself known only as appearance, and cannot,\ntherefore, be the medium in and through which appearances exist. Though\nappearances exist only in and through consciousness, they are not due to\nany causes which can legitimately be described as individual. From this\nstandpoint Kant would seem to distinguish between the grounds and\nconditions of phenomenal existence and the special determining causes\nof individual consciousness. Transcendental conditions generate\nconsciousness of the relatively permanent and objective world in space\nand time; empirical conditions within this space and time world\ndetermine the sensuous modes through which special portions of this\ninfinite and uniform world appear diversely to different minds.\n\nThis, however, is a point of view which is only suggested, and, as we\nhave already observed,[949] the form in which it is outlined suggests\nmany objections and difficulties. Consciousness of the objective world\nin space and time does not exist complete with one portion of it more\nspecifically determined in terms of actual sense-perceptions. Rather the\nconsciousness of the single world in space and time is gradually\ndeveloped through and out of sense experience of limited portions of it.\nWe have still to consider the various sections in the _Analytic of\nPrinciples_ (especially the section added in the second edition on the\n_Refutation of Idealism_) and in the _Dialectic_, in which Kant further\ndevelops this standpoint. But even after doing so, we shall be forced to\nrecognise that Kant leaves undiscussed many of the most obvious\nobjections to which his phenomenalism lies open. To the very last he\nfails to state in any really adequate manner how from the phenomenalist\nstandpoint he would regard the world described in mechanical terms by\nscience as being related to the world of ordinary sense-experience,[950]\nor how different individual consciousnesses are related to one another.\nThe new form, however, in which these old-time problems here emerge is\nthe best possible proof of the revolutionary character of Kant\u2019s\nCritical enquiries. For these problems are no longer formulated in\nterms of the individualistic presuppositions which govern the thinking\nof all Kant\u2019s predecessors, even that of Hume. The concealed\npresuppositions are now called in question, and are made the subject of\nexplicit discussion. But further comment must meantime be deferred.[951]\n\n\nTRANSCENDENTAL DEDUCTION OF THE CATEGORIES, IN THE SECOND EDITION\n\nThe argument of the second edition transcendental deduction can be\nreduced to the following eight points:\n\n(1)[952] It opens with the statement of a fundamental assumption which\nKant does not dream of questioning and of which he nowhere attempts to\noffer proof. The representation of combination is the one kind of\nrepresentation which can never be given through sense. It is not so\ngiven even in the pure forms of space and time yielded by outer and\ninner sense.[953] It is due to an act of spontaneity, which as such must\nbe performed by the understanding. As it is one and the same for every\nkind of combination, it may be called by the general name of synthesis.\nAnd as all combination, without exception, is due to this source, its\ndissolution, that is, analysis, which seems to be its opposite, always\npresupposes it.\n\n(2)[954] Besides the manifold and its synthesis a further factor is\ninvolved in the conception of combination, namely, _the representation\nof the unity_ of the manifold. The combination which is necessary to and\nconstitutes knowledge is _representation_ of the synthetical unity of\nthe manifold. This is a factor additional to synthesis and to the\nmanifold synthesised. For such representation cannot arise out of any\nantecedent consciousness of synthesis. On the contrary, it is only\nthrough supervention upon the unitary synthesis that the conception of\nthe combination becomes possible. In other words, the representation of\nunity conditions _consciousness_ of synthesis, and therefore cannot be\nthe outcome or product of it. This is an application, or rather\ngeneralisation, of a position which in the first edition is developed\nonly in reference to the empirical process of recognition. Recognition\npreconditions consciousness, and therefore cannot be subsequent upon it.\n\n(3)[955] The unity thus represented is not, however, that which is\nexpressed through the category of unity. The consciousness of unity\nwhich is involved in the conception of synthesis is that of apperception\nor transcendental self-consciousness. This is the highest and most\nuniversal form of unity, for it is a presupposition of the unity of all\npossible concepts, whether analytic or synthetic, in the various forms\nof judgment.\n\n(4)[956] A manifold though given is not for that reason also\nrepresented. It must be possible for the \u2018I think\u2019 to accompany it and\nall my other representations:\n\n     \u201c...for otherwise something would be represented in me which could\n     not be thought at all; and that is equivalent to saying that the\n     representation would be impossible or at least would be nothing to\n     me.\u201d[957]\n\nBut to ascribe a manifold as my representations to the identical self is\nto comprehend them, as synthetically connected, in one\napperception.[958] Only what can be combined in one consciousness can be\nrelated to the \u2018I think.\u2019 The analytic unity of self-consciousness\npresupposes the synthetic unity of the manifold.\n\n(5)[959] The unity of apperception is analytic or self-identical. It\nexpresses itself through the proposition, _I am I._ But being thus pure\nidentity without content of its own, it cannot be conscious of itself in\nand by itself. Its unity and constancy can have meaning only through\ncontrast to the variety and changeableness of its specific experiences;\nand yet, at the same time, it is also true that such manifoldness will\ndestroy all possibility of unity unless it be reconcileable with it. The\nvariety can contribute to the conditioning of apperception only in so\nfar as it is capable of being combined into a single consciousness.\nThrough synthetic unifying of the manifold the self comes to\nconsciousness both of itself and of the manifold.\n\n(6)[960] The transcendental original unity of apperception is an\nobjective, not a merely subjective, unity. Its conditions are also the\nconditions in and through which we acquire consciousness of objects. An\nobject is that in the conception of which the manifold of given\nintuitions is combined. (This point, though central to the argument, is\nmore adequately developed in the first than in the second edition.) Such\ncombination requires unity of consciousness. Thus the same unity which\nconditions apperception likewise conditions the relation of\nrepresentations to an object. The unity of pure apperception may\ntherefore be described as an _objective_ unity for two reasons: first,\nbecause it can apprehend its own analytical unity only through discovery\nof unity in the given, and secondly, for the reason that such\nsynthetical unifying of the manifold is also the process whereby\nrepresentations acquire reference to objects.\n\n(7)[961] Kant reinforces this conclusion, and shows its further\nsignificance, by analysis of the act of judgment. The logical definition\nof judgment, as the representation of a relation between two concepts,\nhas many defects. These, however, are all traceable to its initial\nfailure to explain, or even to recognise, the nature of the assertion\nwhich judgment as such claims to make. Judgment asserts relations of a\nquite unique kind, altogether different from those which exist between\nideas connected through association. If, for instance, on seeing a body\nthe sensations of weight due to the attempt to raise it are suggested by\nassociation, there is nothing but subjective sequence; but if we form\nthe judgment that the body is heavy, the two representations are then\nconnected together _in the object_. This is what is intended by the\ncopula \u2018is.\u2019 It is a relational term through which the objective unity\nof given representations is distinguished from the subjective. It\nindicates that the representations stand in objective relation under the\npure unity of apperception, and not merely in subjective relation owing\nto the play of association in the individual mind. \u201cJudgment is nothing\nbut the mode of bringing cognitions to the objective unity of\napperception,\u201d _i.e._ of giving to them a validity which holds\nindependently of the subjective processes through which it is\napprehended. Objective relations are not, of course, all necessary or\nuniversal; and a judgment may, therefore, assert a relation which is\nempirical and contingent. None the less the fundamental distinction\nbetween it and any mere relation of association still persists. The\nempirical relation is still in the judgment asserted to be objective.\nThe subject and the predicate are asserted, in the particular case or\ncases to which the judgment refers, to be connected in the object and\nnot merely in the mind of the subject. Or otherwise stated, though\nsubject and predicate are not themselves declared to be necessarily and\nuniversally related to one another, their contingent relation has to be\nviewed as objectively, and therefore necessarily, grounded. Judgment\nalways presupposes the existence of necessary relations even when it is\nnot concerned to assert them. Judgment is the organ of objective\nknowledge, and is therefore bound up, indirectly when not directly,\nwith the universality and necessity which are the sole criteria of\nknowledge. The judgment expressive of contingency is still judgment, and\nis therefore no less necessary in its conditions, and no less objective\nin its validity, than is a universal judgment of the scientific type. To\nuse Kant\u2019s own terminology, judgment acquires objective validity through\nparticipation in the necessary unity of apperception. In so doing it is\nmade to embody those principles of the objective determination of all\nrepresentations through which alone cognition is possible.\n\n(8)[962] As judgment is nothing but the mode of bringing cognitions to\nthe objective unity of apperception, it follows that the categories,\nwhich in the metaphysical deduction have been proved to be the possible\nfunctions in judging, are the conditions in and through which such pure\napperception becomes possible. Apperception conditions experience, and\nthe unity which both demand for their possibility is that of the\ncategories.\n\n       *       *       *       *       *\n\nBefore passing to the remaining sections of the deduction,[963] which\nare supplementary rather than essential, I may add comment upon the\nabove points. Only (7) and (8) call for special consideration. They\nrepresent a form of argument which has no counterpart in the first\nedition. As we noted,[964] the first edition argument is defective owing\nto its failure to demonstrate that the categories constitute the unity\nwhich is necessary to knowledge. By introducing in the second edition\nthis analysis of judgment, and by showing the inseparable connection\nbetween pure apperception, objective consciousness and judgment, this\ndefect is in some degree removed. As the categories correspond to the\npossible functions of judgment, their objective validity is thereby\nestablished. By this means also the connection which in Kant\u2019s view\nexists between the metaphysical and the transcendental deductions\nreceives for the first time proper recognition. The categories which in\nthe former deduction are discovered and systematised through _logical_\nanalysis of the _form_ of judgment, are in the latter deduction, through\n_transcendental_ analysis of the _function_ of judgment, shown to be\njust those forms of relation which are necessary to the possibility of\nknowledge. It must, however, be noted that the transcendental argument\nis brought to completion only through assumption of the adequacy of the\nmetaphysical deduction. No independent attempt is made to show that the\nparticular categories obtained in the metaphysical deduction are those\nwhich are required, that there are no others, or that all the twelve are\nindispensable.\n\n(7) is a development of an argument which first appears in the\n_Prolegomena_. The statement of it there given is, however, extremely\nconfused, owing to the distinction which Kant most unfortunately\nintroduces[965] between judgments of experience and judgments of\nperception. That distinction is entirely worthless and can only serve to\nmislead the reader. It cuts at the very root of Kant\u2019s Critical\nteaching. Judgments of perception involve, Kant says, no category of the\nunderstanding, but only what he is pleased to call the \u201clogical\nconnection of perceptions in a thinking subject.\u201d What that may be he\nnowhere explains, save by adding[966] that in it perceptions are\n\u201ccompared and conjoined in a consciousness of my state\u201d (also spoken of\nby Kant as \u201cempirical consciousness\u201d), and not \u201cin consciousness in\ngeneral.\u201d\n\n     \u201cAll our judgments are at first mere judgments of perception; they\n     hold good merely for us (that is, for the individual subject), and\n     we do not till afterwards give them a new reference, namely, to an\n     object.... To illustrate the matter: that the room is warm, sugar\n     sweet, and wormwood bitter--these are merely subjectively valid\n     judgments. I do not at all demand that I myself should at all\n     times, or that every other person should, find the facts to be what\n     I now assert; they only express a reference of two sensations to\n     the same subject, to myself, and that only in my present state of\n     perception. Consequently they are not intended to be valid of the\n     object. Such judgments I have named those of perception. Judgments\n     of experience are of quite a different nature. What experience\n     teaches me under certain circumstances, it must teach me always and\n     teach everybody, and its validity is not limited to the subject or\n     to its state at a particular time.\u201d[967]\n\nThe illegitimacy and the thoroughly misleading character of this\ndistinction hardly require to be pointed out. Obviously Kant is here\nconfusing assertion of contingency and contingency of assertion.[968] A\njudgment of contingency, in order to be valid, must itself be necessary.\nEven a momentary state of the self is referable to an object in judgment\nonly if that object is causally, and therefore necessarily, concerned in\nits production.[969]\n\nThe distinction is repeated in \u00a7 22 as follows:\n\n     \u201cThinking is the combining of representations in one consciousness.\n     This combination is either merely relative to the subject, and is\n     contingent and subjective, or is absolute, and is necessary or\n     objective. The combination of representations in one consciousness\n     is judgment. Thinking, therefore, is the same as judging, or the\n     relating of representations to judgments in general. Judgments,\n     therefore, are either merely subjective, or they are objective.\n     They are subjective when representations are related to a\n     consciousness in one subject only, and are combined in it alone.\n     They are objective when they are united in a consciousness in\n     general, that is, necessarily.\u201d[970]\n\nTo accept this distinction is to throw the entire argument into\nconfusion. This Kant seems to have himself recognised in the interval\nbetween the _Prolegomena_ and the second edition of the _Critique_. For\nin the section before us there is no trace of it. The opposition is no\nlonger between subjective and objective judgment, but only between\nassociation of ideas and judgment which as such is always objective. The\ndistinction drawn in the _Prolegomena_ is only, indeed, a more definite\nformulation of the distinction which runs through the first edition of\nthe _Critique_ between the indeterminate and the determinate object of\nconsciousness. The more definite formulation of it seems, however, to\nhave had the happy effect of enabling Kant to realise the illegitimacy\nof any such distinction.\n\n       *       *       *       *       *\n\nWe may now proceed to consider the remaining sections.[971] In section\n21[972] Kant makes a very surprising statement. The above argument,\nwhich he summarises in a sentence, yields, he declares, \u201cthe _beginning_\nof a deduction of the pure concepts of understanding.\u201d This can hardly\nbe taken as representing Kant\u2019s real estimate of the significance of the\npreceding argument, and would seem to be due to a temporary\npreoccupation with the problems that centre in the doctrine of\nschematism. So far, Kant adds in explanation, no account has been taken\nof the particular manner in which the manifold of empirical intuition is\nsupplied to us.[973] The necessary supplement, consisting of a very\nbrief outline statement of the doctrine of schematism, is given in\nsection 26.[974] It differs from the teaching of the special chapter\ndevoted to schematism in emphasising space equally with time. The\ndoctrine of pure _a priori_ manifolds is incidentally asserted.[975]\nSection 26 concludes by consideration of the question why appearances\nmust conform to the _a priori_ categories. It is no more surprising,\nKant claims, than that they should agree with the _a priori_ forms of\nintuition. The categories and the intuitional forms are relative to the\nsame subject to which the appearances are relative; and the appearances\n\u201cas mere representations are subject to no law of connection save that\nwhich the combining faculty prescribes.\u201d\n\nThe summary of the deduction given in section 27 discusses the three\npossible theories regarding the origin of pure concepts, viz. those of\n_generatio aequivoca_ (out of experience), _epigenesis_, and\n_preformation_. The first is disproved by the deduction. The second is\nthe doctrine of the deduction and fulfils all the requirements of\ndemonstration. The proof that the categories are at once independent of\nexperience and yet also universally valid for all experience is of the\nstrongest possible kind, namely, that they make experience itself\npossible. The third theory, that the categories, while subjective and\nself-discovered, originate in faculties which are implanted in us by our\nCreator and which are so formed as to yield concepts in harmony with the\nlaws of nature, lies open to two main objections. In the first place,\nthis is an hypothesis capable of accounting equally well for any kind of\n_a priori_ whatsoever; the predetermined powers of judgment can be\nmultiplied without limit. But a second objection is decisive, namely,\nthat on such a theory the categories would lack the particular kind of\nnecessity which is required. They would express only the necessities\nimposed upon our thinking by the constitution of our minds, and would\nnot justify any assertion of necessary connection _in the object_. Kant\nmight also have added,[976] that this hypothesis is metaphysical, and\ntherefore offers in explanation of the _empirical_ validity of _a\npriori_ concepts a theory which rests upon and involves their\n_unconditioned_ employment. That is a criticism which is reinforced by\nthe teaching of the _Dialectic_.\n\nTo return now to the omitted sections 22 to 25. Section 22 makes no\nfresh contribution to the argument of the first edition. Its teaching in\nregard to pure intuition and mathematical knowledge has already been\ncommented upon. In section 23 Kant dwells upon an interesting\nconsequence of the argument of the deduction. The categories have a\nwider scope than the pure forms of sense. Since the argument of the\ndeduction has shown that judgment is the indispensable instrument both\nfor reducing a manifold to the unity of apperception and also for\nconferring upon representations a relation to an object, it follows that\nthe categories which are simply the possible functions of unity in\njudgment are valid for any and every consciousness that is sensuously\nconditioned and whose knowledge is therefore acquired through synthesis\nof a given manifold. Though such consciousness may not intuit in terms\nof space and time, it must none the less apprehend objects in terms of\nthe categories. The categories thus extend to _objects of sensuous\nintuition in general_. They are not, however, valid of objects as such,\nthat is, of things in themselves. As empty relational forms they have\n_meaning_ only in reference to a given matter; and as instruments for\nthe reduction of variety to the unity of apperception their _validity_\nhas been proved only for conscious and sensuous experience. Even if the\npossibility of a non-sensuous intuitive understanding, capable of\napprehending things in themselves, be granted, we have no sufficient\nground for asserting that the forms which such understanding will employ\nmust coincide with the categories.[977] These are points which will come\nup for discussion in connection with Kant\u2019s more detailed argument in\nthe chapter on the distinction between phenomena and noumena.[978]\n\nThe heading to section 24 is decidedly misleading. The phrase \u201cobjects\nof the senses in general\u201d might be synonymous with \u201cobjects of intuition\nin general\u201d of the preceding section. To interpret it, however, by the\ncontents of the section, it means \u201cobjects of _our_ senses.\u201d This\nsection ought, therefore, to form part of section 26, which in its\nopening sentences supplies its proper introduction. (It may also be\nnoted that the opening sentences of section 24 are a needless repetition\nof section 23. This would seem to show that it was not written in\nimmediate continuation of it.) The first three paragraphs of section 24\nexpound the same doctrine of schematism as that outlined in section 26,\nsave that time alone is referred to. The remaining paragraphs of section\n24 deal with the connected doctrine of inner sense. Section 25 deals\nwith certain consequences which follow from that doctrine of inner\nsense.[979]\n\n\nTHE DOCTRINE OF INNER SENSE\n\nWe have still to consider a doctrine of great importance in Kant\u2019s\nthinking, that of inner sense. The significance of this doctrine is\nalmost inversely proportionate to the scantiness and obscurity of the\npassages in which it is expounded and developed. Much of the\nindefiniteness and illusiveness of the current interpretations of Kant\nwould seem to be directly traceable to the commentator\u2019s failure to\nappreciate the position which it occupies in Kant\u2019s system. Several of\nKant\u2019s chief results are given as deductions from it, while it itself,\nin turn, is largely inspired by the need for a secure basis upon which\nthese positions may be made to rest. The relation of the doctrine to its\nconsequences is thus twofold. Kant formulates it in order to safeguard\nor rather to justify certain conclusions; and yet these conclusions have\nthemselves in part been arrived at owing to his readiness to accept such\na doctrine, and to what would seem to have been his almost instinctive\nfeeling of its kinship (notwithstanding the very crude form in which\nalone he was able to formulate it) with Critical teaching. It was\nprobably one of the earliest of the many new tenets which Kant adopted\nin the years immediately subsequent to the publication of the inaugural\n_Dissertation_, but it first received adequate statement in the second\nedition of the _Critique_. Kant took advantage of the second edition to\nreply to certain criticisms to which his view of time had given rise,\nand in so doing was compelled to formulate the doctrine of inner sense\nin a much more explicit manner. Hitherto he had assumed its truth, but\nhad not, as it would seem, sufficiently reflected upon the various\nconnected conclusions to which he was thereby committed. This is one of\nthe many instances which show how what is most fundamental in Kant\u2019s\nthinking is frequently that of which he was himself least definitely\naware. Like other thinkers, he was most apt to discuss what he himself\nwas inclined to question and feel doubt over. The sources of his insight\nas well as the causes of his failure often lay beyond the purview of his\nexplicitly developed tenets; and only under the stimulus of criticism\nwas he constrained and enabled to bring them within the circle of\nreasoned conviction. We may venture the prophecy that if Kant had been\nable to devote several years more to the maturing of the problems which\nin the face of so many difficulties he had brought thus far, the\ndoctrine of inner sense, or rather the doctrines to which it gives\nexpression, would have been placed in the forefront of his teaching, and\ntheir systematic interconnection, both in the way of ground and of\nconsequence, with all his chief tenets would have been traced and\nsecurely established.\n\nThis would have involved, however, two very important changes. In the\nfirst place, Kant would have had to recognise the unsatisfactory\ncharacter of the supposed analogy between inner and outer sense. As\nalready remarked,[980] no great thinker, except Locke, has attempted to\ninterpret inner consciousness on the analogy of the senses; and the\nobscurities of Kant\u2019s argument are not, therefore, to be excused on the\nground that \u201cthe difficulty, how a subject can have an internal\nintuition of itself, is common to every theory.\u201d Secondly, Kant would\nhave had to define the relation in which he conceived this part of his\nteaching to stand to his theory of consciousness. But both these changes\ncould have been made without requiring that he should give up the\ndoctrines which are mainly responsible for his theory of inner sense,\nnamely, that there can be no awareness of awareness, but only of\nexistences which are objective, and that there is consequently no\nconsciousness of the generative, synthetic processes[981] which\nconstitute consciousness on its subjective side. It is largely in virtue\nof these conclusions that Kant\u2019s phenomenalism differs from the\nsubjective idealism of his predecessors. If we ignore or reject them,\nmerely because of the obviously unsatisfactory manner in which alone\nKant has been able to formulate them, we rule ourselves out from\nunderstanding the intention and purpose of much that is most\ncharacteristic of Critical teaching.\n\nThe doctrine of inner sense, as expounded by Locke, suffers from an\nambiguity which seems almost inseparable from it, namely, the confusion\nbetween inner sense, on the one hand as a _sense_ in some degree\nanalogous in nature to what may be called outer sense, and on the other\nas consisting in self-conscious reflection. This same confusion is\ntraceable throughout the _Critique_, and is, as we shall find, in large\npart responsible for Kant\u2019s failure to recognise, independently of\noutside criticism, the central and indispensable part which this\ndoctrine is called upon to play in his system.\n\nThe doctrine is stated by Kant as follows. Just as outer sense is\naffected by noumenal agencies, and so yields a manifold arranged in\nterms of a form peculiar to it, namely, space, so inner sense is\naffected by the mind itself and its inner state.[982] The manifold\nthereby caused is arranged in terms of a form peculiar to inner sense,\nnamely, time. The content thus arranged falls into two main divisions.\nOn the one hand we have feelings, desires, volitions, that is, states of\nthe mind in the strict sense, subjective non-spatial existences. On the\nother hand we have sensations, perceptions, images, concepts, in a word,\nrepresentations (_Vorstellungen_) of every possible type. These latter\nall refer to the external world in space, and yet, according to Kant,\nspeaking from the limited point of view of a critique of _knowledge_,\nform the proper content of inner sense. \u201c...the representations of the\nouter senses constitute the actual material with which we occupy our\nminds,\u201d[983] \u201cthe whole material of knowledge even for our inner\nsense.\u201d[984] (These statements, it may be observed, are first made in\nthe second edition.) As Kant explains himself in B 67-8, he would seem\nto mean that the mind in the process of \u201csetting\u201d representations of\nouter sense in space affects itself, and is therefore constrained to\narrange the given representations likewise in time. No new content,\nadditional to that of outer sense, is thereby generated, but what\npreviously as object of outer sense existed merely in space is now also\nsubjected to conditions of time. The representations of outer sense are\nall by their very nature likewise representations of inner sense. To\nouter sense is due both their content and their spatial form; to inner\nsense they owe only the additional form of time; their content remains\nunaffected in the process of being taken over by a second sense. This\nyields such explanation as is possible of Kant\u2019s assertion in A 33 that\n\u201ctime can never be a determination of outer appearances.\u201d He may be\ntaken as meaning that time is never a determination of outer sense _as\nsuch_, but only of its contents as always likewise subject to the form\nof inner sense.[985]\n\nThis is how Kant formulates his position from the extreme subjectivist\npoint of view which omits to draw any distinction between representation\nand its object, between inner states of the self and appearances in\nspace. All representations, he says,[986] all appearances without\nexception, are states of inner sense, modifications of the mind. Some\nexist only in time, some exist both in space and in time; but all alike\nare modes of the identical self, mere representations (_blosse\nVorstellungen_). Though appearances may exist outside one another in\nspace, space itself exists only as representation, merely \u201cin us.\u201d\n\nNow without seeking to deny that this is a view which we find in the\nsecond edition of the _Critique_ as well as in the first,[987] and that\neven in passages which are obviously quite late in date of writing Kant\nfrequently speaks in terms which conform to it, we must be no less\ninsistent in maintaining that an alternative view more and more comes\nto the front in proportion as Kant gains mastery over the conflicting\ntendencies that go to constitute his new Critical teaching. From the\nvery first he uses language which implies that _some_ kind of\ndistinction must be drawn between representations and objects\nrepresented, between subjective cognitive states in the proper sense of\nthe term and existences in space.\n\n     \u201cTime can never be a determination of outer appearances. It belongs\n     neither to form nor position, etc. On the other hand it determines\n     the relation of representations in our inner state.\u201d[988]\n\nSimilarly in those very sentences in which he asserts all appearances to\nbe _blosse Vorstellungen_, a distinction is none the less implied.\n\n     \u201cTime is the formal _a priori_ condition of all appearances in\n     general. Space, as the pure form of all outer intuition, is as _a\n     priori_ condition limited exclusively (_bloss_) to outer\n     appearances. On the other hand as all representations, _whether\n     they have outer things as their object or not_, still in themselves\n     belong, as determinations of the mind, to the inner state, and this\n     inner state is subject to the formal condition of inner intuition,\n     that is of time, time is an _a priori_ condition of all appearance\n     whatever. It is, indeed, the immediate condition of the inner\n     appearance (of our souls), and thereby mediately likewise of outer\n     appearances.\u201d[989]\n\nAs the words which I have italicised show, Kant, even in the very\nsentence in which he asserts outer representations to be inner states,\nnone the less recognises that appearances in space are not\nrepresentations in the same meaning of that term as are subjective\nstates. They are the _objects_ of representation, not representation\nitself. The latter alone is correctly describable as a state of the\nmind. The former may be conditioned by representation, and may therefore\nbe describable as appearances, but are not for that reason to be equated\nwith representation. But before the grounds and nature of this\ndistinction can be formulated in the proper Critical terms, we must\nconsider the reasons which induced Kant to commit himself to this\nobscure and difficult doctrine of inner sense. As I shall try to show,\nit is no mere excrescence upon his system; on the contrary, it is\ninseparably bound up with all his main tenets.\n\nOne of the chief influences which constrained Kant to develop this\ndoctrine is the conclusion, so essential to his position, that knowledge\nmust always involve an intuitional manifold in addition to _a priori_\nforms and concepts. That being so, he was bound to deny to the mind all\npower of gaining knowledge by mere reflection. If our mental activities\nand states lay open to direct inspection, we should have to recognise in\nthe mind a non-sensuous intuitional power. Through self-consciousness or\nreflection we should acquire knowledge independently of sense. Such\napprehension, though limited to the mind\u2019s own operations and states,\nwould none the less be _knowledge_, and yet would not conform to the\nconditions which, as the transcendental deduction has shown, are\ninvolved in all knowledge. In Kant\u2019s view the belief that we possess\nself-consciousness of this type, a power of reflection thus conceived,\nis wholly illusory. To assume any such faculty would be to endow the\nmind with occult or mystical powers, and would throw us back upon the\nLeibnizian rationalism, which traces to such reflection our\nconsciousness of the categories, and which rears upon this foundation\nthe entire body of metaphysical science.[990]\n\nThe complementary _negative_ conclusion of the transcendental deduction\nis a no less fundamental and constraining influence in compelling Kant\nto develop a doctrine of inner sense. If all knowledge is knowledge of\nappearances, or if, as he states his position in the _Analytic of\nPrinciples_,[991] our knowledge can extend no further than sense\nexperience and inference from such experience, either knowledge of our\ninner states must be mediated, like our knowledge of outer objects, by\nsensation, or we can have no knowledge of them whatsoever. On Critical\nprinciples, consistently applied, there can be no middle course between\nacceptance of an indirect empirical knowledge of the mind and assertion\nof its unknowableness. Mental activities may perhaps be thought in terms\nof the pure forms of understanding, but in that case their conception\nwill remain as purely problematic and as indeterminate as the conception\nof the thing in itself. It is impossible for Kant to admit _immediate_\nconsciousness of the mind\u2019s real activities and states, and at the same\ntime to deny that we can have knowledge of things in themselves. The\n_Aesthetic_, in proving that everything in space and time is appearance,\nimplicitly assumes the impossibility of direct self-conscious\nreflection; and the transcendental deduction in showing that all\nknowledge involves as correlative factors both sense and thought, has\nreinforced this conclusion, and calls for its more explicit\nrecognition, in reference to the more inward aspect of experience.\n\nAs we have already noted,[992] Kant\u2019s doctrine of inner sense was\nprobably adopted in the early \u2019seventies, and though it is not itself\ndefinitely formulated in the first edition, the chief consequence that\nfollows from it is clearly recognised. Thus in the _Aesthetic_ Kant\ndraws the conclusion that, as time is the form of inner sense,\neverything apprehended in time, and consequently all inner states and\nactivities, can be known only as appearances. The mind (meaning thereby\nthe ultimate conditioning grounds of consciousness) is as indirectly\nknown as is any other mode of noumenal existence. In the _Analytic_,\nwhenever he is called upon to express himself upon this and kindred\npoints, he continues to hold to this position; and in the section on the\n_Paralogisms_ all the main consequences that follow from its acceptance\nare drawn in the most explicit and unambiguous manner. It is argued that\nas the inner world, the feelings, volitions and representations of which\nwe are conscious, is a world constructed out of a given manifold yielded\nby inner sense, and is therefore known only as the appearance of a\ndeeper reality which we have no power of apprehending, it possesses no\nsuperiority either of certainty or of immediacy over the outer world of\nobjects in space. We have immediate consciousness of both alike, but in\nboth cases this immediate consciousness rests upon the transcendental\nsynthetic processes whereby such consciousness is conditioned and\ngenerated. The transcendental activities fall outside the field of\nempirical consciousness and therefore of knowledge.\n\nThus Kant would seem to be maintaining that the radical error committed\nby the subjective idealists, and with which all the main defects of\ntheir teaching are inseparably bound up, lies in their ascription to the\nmind of a power of direct self-conscious reflection, and consequently in\ntheir confusion of the transcendental activities which condition\nconsciousness with the inner states and processes which such\nconsciousness reveals. This has led them to ascribe priority and\nindependence to our inner states, and to regard outer objects as known\nonly by an inference from them. The Critical teaching insists on the\ndistinction between appearance and reality, applies it to the inner\nlife, and so restores to our consciousness of the outer world the\ncertainty and immediacy of which subjective idealism would profess to\ndeprive it. Such are the important conclusions at which Kant arrives in\nhis various \u201crefutations of idealism\u201d; and it will be advisable to\nconsider these refutations in full detail before attempting to complete\nour statement of his doctrine of inner sense.\n\n\nKANT\u2019S REFUTATIONS OF IDEALISM\n\nKant has in a number of different passages attempted to define his\nCritical standpoint in its distinction from the positions of Descartes\nand Berkeley. Consideration of these will enable us to follow Kant in\nhis gradual recognition of the manifold consequences to which he is\ncommitted by his substitution of inner sense for direct self-conscious\nintuition or reflection, or rather of the various congenial tenets which\nit gives him the right consistently to defend and maintain. In Kant\u2019s\nCritical writings we find no less than seven different statements of his\nrefutation of idealism: (I.) in the fourth _Paralogism_ of the first\nedition of the _Critique_; (II.) in section 13 (_Anm._ ii. and iii.) of\nthe _Prolegomena_; (III.) in section 49 of the _Prolegomena_; (IV.) in\nthe second appendix to the _Prolegomena_; (V.) in sections added in the\nsecond edition at the conclusion of the _Aesthetic_ (B 69 ff.); (VI.) in\nthe \u201crefutation of idealism\u201d (B 274-8), in the supplementary section at\nthe end of the section on the _Postulates_ (B 291-4), and in the note to\nthe new preface (B xxxix-xl); (VII.) in the \u201crefutation of problematic\nidealism\u201d given in the _Seven Small Papers_ which originated in Kant\u2019s\nconversations with Kiesewetter. Consideration of these in the above\norder will reveal Kant\u2019s gradual and somewhat vacillating recognition of\nthe new and revolutionary position which alone genuinely harmonises with\nCritical principles. But first we must briefly consider the various\nmeanings which Kant at different periods assigned to the term idealism.\nEven in the _Critique_ itself it is employed in a great variety of\ndiverse connotations.\n\nIn the pre-Critical writings[993] the term idealism is usually employed\nin what was its currently accepted meaning, namely, as signifying any\nphilosophy which denied the existence of an independent world\ncorresponding to our subjective representations. But even as thus used\nthe term is ambiguous.[994] It may signify either denial of a\n_corporeal_ world independent of our representations or denial of an\nimmaterial world \u201ccorresponding to\u201d the represented material world,\n_i.e._ the denial of _Dinge an sich_. For there are traceable in\nLeibniz\u2019s writings two very different views as to the reality of the\nmaterial world. Sometimes the monads are viewed as purely intelligible\nsubstances without materiality of any kind. The kingdom of the extended\nis set into the representing subjects; only the immaterial world of\nunextended purely spiritual monads remains as independently real. At\nother times the monads, though in themselves immaterial, are viewed as\nconstituting through their coexistence an independent material world and\na materially occupied space. Every monad has a spatial sphere of\nactivity. The material world is an objective existence due to external\nrelations between the monads, not a merely subjective existence internal\nto each of them. This alternation of standpoints enabled Leibniz\u2019s\nsuccessors to deny that they were idealists; and as the more daring and\nspeculative aspects of Leibniz\u2019s teaching were slurred over in the\nprocess of its popularisation, it was the second, less consistent view,\nwhich gained the upper hand. Wolff, especially in his later writings,\ndenounces idealism; and in the current manuals, sections in refutation\nof idealism became part of the recognised philosophical teaching.\nIdealism still, however, continued to be used ambiguously, as signifying\nindifferently either denial of material bodies or denial of things in\nthemselves. This is the dual meaning which the term presents in Kant\u2019s\npre-Critical writings. In his _Dilucidatio_ (1755)[995] he refutes\nidealism by means of the principle that a substance cannot undergo\nchanges unless it is a substance independent of other substances.\nObviously this argument can at most prove the existence of an\nindependent world, not that it is spatial or material. And as Vaihinger\nadds, it does not even rule out the possibility that changes find their\nsource in a Divine Being. In the _Dreams of a Visionseer_ (1766)[996]\nSwedenborg is described as an idealist, but without further\nspecification of the exact sense in which the term is employed. In the\ninaugural _Dissertation_ (1770)[997] idealism is again rejected, on the\nground that sense-affection points to the presence of an intelligible\nobject or _noumenon_.\n\nIn Kant\u2019s class lectures on metaphysics,[998] which fall, in part at\nleast, between 1770 and 1781, the term idealism is employed in a very\ndifferent sense, which anticipates its use in the _Appendix_ to the\n_Prolegomena_.[999] The teaching of the _Dissertation_, that things in\nthemselves are knowable, is now described as dogmatic, Platonic,\nmystical (_schw\u00e4rmerischer_) idealism. He still rejects the idealism of\nBerkeley, and still entitles it simply idealism, without limiting or\ndescriptive predicates. But now also he employs the phrase \u201cproblematic\nidealism\u201d as descriptive of his own new position. This is, of course,\ncontrary to his invariable usage elsewhere, but is interesting as\nshowing that about this time his repugnance to the term idealism begins\nto give way, and that he is willing to recognise that the relation of\nthe Critical teaching to idealism is not one of simple opposition. He\nnow begins to regard idealism as a factor, though a radically\ntransformed factor, in his own philosophy.\n\nStudy of the _Critique_ reinforces this conclusion. In the _Aesthetic_\nKant teaches the \u201ctranscendental ideality\u201d of space and time; and in the\n_Dialectic_ (in the fourth _Paralogism_) describes his position as\nidealism, though with the qualifying predicate transcendental.[1000] But\nthough this involves an extension of the previous connotation of the\nterm idealism, and might therefore have been expected to increase the\nexisting confusion, it has the fortunate effect of constraining Kant to\nrecognise and discriminate the various meanings in which it may be\nemployed. This is done somewhat clumsily, as if it were a kind of\nafterthought. In the introductory syllogism of the fourth _Paralogism_\nDescartes\u2019 position and his own are referred to simply as idealism and\ndualism respectively. The various possible sub-species of idealism as\npresented in the two editions of the _Critique_ and in the _Prolegomena_\nmay be tabulated as follows:\n\n            {Material {Sceptical {Problematic (the position of Descartes).\n            {         {          {Sceptical in the stricter and more usual\n  Idealism  {         {          { sense (the position of Hume).\n            {         {Dogmatic (the position of Berkeley).\n            {Formal or Critical or Transcendental (Kant\u2019s own position).\n\nThe distinction between problematic idealism and idealism of the more\nstrictly sceptical type is not clearly drawn by Kant.[1001] Very\nstrangely Kant in this connection never mentions Hume: the reference in\nB xxxix _n._ is probably not to Hume but to Jacobi. Transcendental\nidealism is taken as involving an empirical realism and dualism, and is\nset in opposition to transcendental realism which is represented as\ninvolving empirical idealism. In B xxxix _n._ Kant speaks of\n\u201cpsychological idealism,\u201d meaning, as it would seem, material or\nnon-Critical idealism.\n\nIn the second appendix to the _Prolegomena_ Kant draws a further\ndistinction, in line with that already noted in his lectures on\nmetaphysics. Tabulated it is as follows:\n\n          {Mystical, in the sense of belief in and reliance on a supposed\n          { human power of intellectual intuition. It is described as\n  Idealism{idealism in the strict (_eigentlich_) sense--the position of the\n          {Eleatics, of Plato and Berkeley.\n          {Formal or Critical--Kant\u2019s own position.\n\nThis latter classification can cause nothing but confusion. The\nobjections that have to be made against it from Kant\u2019s own critical\nstandpoint are stated below.[1002]\n\nLet us now consider, in the order of their presentation, the various\nrefutations of idealism which Kant has given in his Critical writings.\n\n=I. Refutation of Idealism as given in First Edition of \u201cCritique\u201d= (A\n366-80).--This refutation is mainly directed against Descartes, who is\nmentioned by name in A 367. Kant, as Vaihinger suggests, was very\nprobably led to recognise Descartes\u2019 position as a species of idealism\nin the course of a re-study of Descartes before writing the section on\nthe _Paralogisms_. As already pointed out, this involves the use of the\nterm idealism in a much wider sense than that which was usually given to\nit in Kant\u2019s own day. In the development of his argument Kant also\nwavers between two very different definitions of this idealism, as being\ndenial of _immediate_ certainty and as denial of all certainty.[1003]\nThe second interpretation, which would make it apply to Hume rather than\nto Descartes, is strengthened in the minds of his readers by his further\ndistinction[1004] between dogmatic and sceptical idealism, and the\nidentification of the idealism under consideration with the latter. The\ntitle problematic which Kant in the second edition[1005] applies to\nDescartes\u2019 position suffers from this same ambiguity. As a matter of\nfact, Kant\u2019s refutation applies equally well to either position. The\nteaching of Berkeley, which coincides with dogmatic idealism as here\ndefined by Kant, namely, as consisting in the contention that the\nconception of matter is inherently contradictory, is not dwelt upon, and\nthe appended promise of refutation is not fulfilled.\n\nDescartes\u2019 position is stated as follows: only our own existence and\ninner states are immediately apprehended by us; all perceptions are\nmodifications of inner sense; and the existence of external objects can\ntherefore be asserted only by an inference from the inner perceptions\nviewed as effects. In criticism, Kant points out that since an effect\nmay result from more than one cause, this inference to a quite\ndeterminate cause, viz. objects as bodies in space, is doubtfully\nlegitimate. The cause of our inner states may lie within and not without\nus, and even if external, need not consist in spatial objects. Further,\nleaving aside the question of a possible alternative to the assumption\nof independent material bodies, the assertion of the existence of such\nobjects would, on Descartes\u2019 view, be merely conjectural. It could never\nhave certainty in any degree equivalent to that possessed by the\nexperiences of inner sense.\n\n     \u201cBy an idealist, therefore, we must not understand one who denies\n     the existence of outer objects of the senses, but only one who does\n     not admit that their existence is known through immediate\n     perception, and who therefore concludes that we can never, by way\n     of any possible experience, be completely certain of their\n     reality.\u201d[1006]\n\nNo sooner is the term idealist thus clearly defined than Kant, in\nkeeping with the confused character of the entire section, proceeds to\nthe assertion (_a_) that there are idealists of another type, namely,\ntranscendental idealists,[1007] and (_b_) that the non-transcendental\nidealists sometimes also adopt a dogmatic position, not merely\nquestioning the immediacy of our knowledge of matter, but asserting it\nto be inherently contradictory. All this points to the composite origin\nof the contents of this section.\n\nTranscendental idealism is opposed to empirical idealism. It maintains\nthat phenomena are representations merely, not things in themselves.\nSpace and time are the sensuous forms of our intuitions. Empirical\nidealism, on the other hand, goes together with transcendental realism.\nIt maintains that space and time are given as real in themselves, in\nindependence of our sensibility. (Transcendental here, as in the phrase\n\u201ctranscendental ideality,\u201d[1008] is exactly equivalent to transcendent.)\nBut such a contention is inconsistent with the other main tenet of\nempirical idealism. For if our inner representations have to be taken as\nentirely distinct from their objects, they cannot yield assurance _even\nof the existence_ of these objects. To the transcendental idealist no\nsuch difficulty is presented. His position naturally combines with\nempirical realism, or, as it may also be entitled, empirical dualism.\nMaterial bodies in space, being merely subjective representations, are\nimmediately apprehended. The existence of matter can be established\n\u201cwithout our requiring to issue out beyond our bare self-consciousness\nor to assume anything more than the certainty of the representations in\nus, _i.e._ of the _cogito ergo sum_.\u201d[1009] Though the objects thus\napprehended are outside one another in space, space itself exists only\nin us.\n\n     \u201cOuter objects (bodies) are mere appearances, and are therefore\n     nothing but a species of my representations, the objects of which\n     are something only through these representations. Apart from them\n     they are nothing. Thus outer things exist as well as I myself, and\n     both, indeed, upon the immediate witness of my\n     self-consciousness....\u201d[1010]\n\nThe only difference is that the representation of the self belongs only\nto inner, while extended bodies also belong to outer sense. There is\nthus a dualism, but one that falls entirely within the field of\nconsciousness, and which is therefore empirical, not transcendental.\nThere is indeed a transcendental object which \u201cin the transcendental\nsense may be outside us,\u201d[1011] but it is unknown and is not in\nquestion. It ought not to be confused with our representations of matter\nand corporeal things.\n\nFrom this point[1012] the argument becomes disjointed and repeats\nitself, and there is much to be said in support of the contention of\nAdickes that the remainder of the section is made up of a number of\nseparate interpolations.[1013] First, Kant applies the conclusion\nestablished in the _Postulates of Empirical Thought_, viz. that reality\nis revealed only in sensation. As sensation is an element in all outer\nperception, perception affords immediate certainty of real existence,\nKant next enters[1014] upon a eulogy of sceptical idealism as \u201ca\nbenefactor of human reason.\u201d It brings home to us the utter\nimpossibility of proving the existence of matter on the assumption that\nspatial objects are things in themselves, and so constrains us to\njustify the assertions which we are at every moment making. And such\njustification is, Kant here claims, only possible if we recognise that\nouter objects as mere representations are immediately known. In the next\nparagraph we find a sentence which, together with the above eulogistic\nestimate of the merits of idealism, shows how very far Kant, at the\ntime of writing, was from feeling the need of differentiating his\nposition from that of subjectivism. The sentence is this:\n\n     \u201cWe cannot be sentient of what is outside ourselves, but only of\n     what is in ourselves, and the whole of our self-consciousness\n     therefore yields nothing save merely our own determinations.\u201d\n\nIt is probable, indeed, that the paragraph in which this occurs is of\nvery early origin, prior to the development of the main body of the\n_Analytic_; for in the same paragraph we also find the assertion,\nutterly at variance with the teaching of the _Analytic_ and with that of\nthe first and third _Paralogisms_, that \u201cthe thinking ego\u201d is known\nphenomenally as _substance_.[1015] We seem justified in concluding that\nthe various manuscripts which have gone to form this section on the\nfourth _Paralogism_ were written at an early date within the Critical\nperiod.\n\nWe may note, in passing, two sentences in which, as in that quoted\nabove, a distinction between representations and their objects is\nrecognised in wording if not in fact.\n\n     \u201cAll outer perception furnishes immediate proof of something actual\n     in space, or rather is the actual itself. To this extent empirical\n     realism is beyond question, _i.e._ there corresponds to our outer\n     perceptions something actual in space.\u201d[1016]\n\nAgain in A 377 the assertion occurs that \u201cour outer senses, as regards\nthe data from which experience can arise, have their actual\ncorresponding objects in space.\u201d Certainly these statements, when taken\ntogether with the other passages in this section, form a sufficiently\nstrange combination of assertion and denial. Either there is a\ndistinction between representation and its object or there is not; if\nthe former, then objects in space are not merely representations; if the\nlatter, then the \u201ccorrespondence\u201d is merely that of a thing with\nitself.[1017]\n\n       *       *       *       *       *\n\nThis refutation of idealism will not itself stand criticism. For two\nseparate reasons it entirely fails to attain its professed end. In the\nfirst place, it refutes the position of Descartes only by virtually\naccepting the still more extreme position of Berkeley. Outer objects,\nKant argues, are immediately known because they are ideas merely. There\nis no need for inference, because there is no transcendence of the\ndomain of our inner consciousness. In other words, Kant refutes the\nproblematic idealism of Descartes by means of the more subjective\nidealism of Berkeley. The \u201cdogmatic\u201d idealism of Berkeley in the form in\nwhich Kant here defines it,[1018] namely, as consisting in the assertion\nthat the notion of an independent spatial object involves inherent\ncontradictions, is part of his own position. For that reason he was\nbound to fail in his promise[1019] to refute such dogmatic idealism.\nFortunately he never even attempts to do so. In the second place, Kant\nignores the fact that he has himself adopted an \u201cidealist\u201d view of inner\nexperience. Inner experience is not for him, as it was for Descartes,\nthe immediate apprehension of genuine reality. As it is only appearance,\nthe incorporation of outer experience within it, so far from\nestablishing the reality of the objects of outer sense, must rather\nprove the direct contrary. No more is really established than Descartes\nhimself invariably assumes, namely, the actual existence of mental\nrepresentations of a corporeal world in space. Descartes\u2019 further\nassertion that the world of things in themselves can be inferred to be\nmaterial and spatial, Kant, of course, refuses to accept. On this latter\npoint Kant is in essential agreement with Berkeley.\n\nIt is by no means surprising that Kant\u2019s first critics,[1020] puzzled\nand bewildered by the obscurer and more difficult portions of the\n_Critique_, should have based their interpretation of Kant\u2019s general\nposition largely upon the above passages; and that in combining the\nextreme subjective idealism which Kant there advocates with his doctrine\nthat the inner life of ever-changing experiences is itself merely ideal,\nshould have come to the conclusion that Kant\u2019s position is an extension\nof that of Berkeley. Pistorius objected that in making outer appearances\nrelative to an inner consciousness which is itself appearance, Kant is\nreducing everything to mere illusion. Hamann came to the somewhat\nsimilar conclusion, that Kant, notwithstanding his very different\nmethods of argument, is \u201ca Prussian Hume,\u201d in substantial agreement with\nhis Scotch predecessor.\n\n\nII. =\u201cProlegomena,\u201d Section 13, Notes II and III.=--In the _Prolegomena_\nKant replies to the criticism which the first edition of the _Critique_\nhad called forth, that his position is an extension of the idealism of\nDescartes, and even more thoroughgoing than that of Berkeley. Idealism\nhe redefines in a much narrower sense, which makes it applicable only to\nBerkeley\n\n     \u201c...as consisting in the assertion that there are none but thinking\n     beings, and that all other things which we suppose ourselves to\n     perceive in intuition are nothing but representations in the\n     thinking beings, to which no object external to them corresponds in\n     fact.\u201d[1021]\n\nIn reply Kant affirms his unwavering belief in the reality of _Dinge an\nsich_\n\n     \u201c...which though quite unknown to us as to what they are in\n     themselves, we yet know by the representations which their\n     influence on our sensibility procures us.... Can this be termed\n     idealism? It is the very contrary.\u201d[1022]\n\nKant adds that his position is akin to that of Locke, differing only in\nhis assertion of the subjectivity of the primary as well as of the\nsecondary qualities.\n\n     \u201cI should be glad to know what my assertions ought to have been in\n     order to avoid all idealism. I suppose I ought to have said, not\n     only that the representation of space is perfectly conformable to\n     the relation which our sensibility has to objects (for that I have\n     said), but also that it is completely similar to them--an assertion\n     in which I can find as little meaning as if I said that the\n     sensation of red has a similarity to the property of cinnabar which\n     excites this sensation in me.\u201d[1023]\n\nKant is here very evidently using the term idealism in the narrowest\npossible meaning, as representing only the position of Berkeley, and as\nexcluding that of Descartes and Leibniz. Such employment of the term is\nat variance with his own previous usage. Though idealism here\ncorresponds to the \u201cdogmatic idealism\u201d of A 377, it is now made to\nconcern the assertion or denial of things in themselves, not as\npreviously the problem of the reality of material objects and of space.\nKant is also ignoring the fact, which he more than once points out in\nthe _Critique_, that his philosophy cannot prove that the cause of our\nsensations is without and not within us. His use of \u201cbody\u201d[1024] as a\nname for the thing in itself is likewise without justification. This\npassage is mainly polemical; it is hardly more helpful than the\ncriticism to which it was designed to reply.\n\nIn Section 13, Note iii., Kant meets the still more extreme criticism\n(made by Pistorius), that his system turns all the things of the world\ninto mere illusion (_Schein_). He distinguishes transcendental idealism\nfrom \u201cthe mystical and visionary idealism of Berkeley\u201d on the one hand,\nand on the other from the Cartesian idealism which would convert mere\nrepresentations into things in themselves. To obviate the ambiguities of\nthe term transcendental, he declares that his own idealism may perhaps\nmore fitly be entitled Critical. This distinction between mystical and\nCritical idealism connects with the contents of the second part of the\nAppendix, treated below.\n\nIII. =\u201cProlegomena,\u201d Section 49.=--This is simply a repetition of the\nargument of the fourth _Paralogism_. The Cartesian idealism, now (as in\nB 274) named material idealism, is alone referred to. The Cartesian\nidealism does nothing, Kant says, but distinguish external experience\nfrom dreaming. There is here again the same confusing use of the term\n\u201ccorresponds.\u201d\n\n     \u201cThat something actual without us not only corresponds but must\n     correspond to our external perceptions can likewise be\n     proved....\u201d[1025]\n\nIV. =\u201cProlegomena,\u201d Second Part of the Appendix.=--Kant here returns to\nthe distinction, drawn in Section 13, Note iii., between what he now\ncalls \u201cidealism proper (_eigentlicher_),\u201d[1026] _i.e._ visionary or\nmystical idealism, and his own.\n\n     \u201cThe position of all genuine idealists from the Eleatics to Bishop\n     Berkeley is contained in this formula: \u2018All cognition through the\n     senses and experience is nothing but mere illusion, and only in the\n     ideas of pure understanding and Reason is there truth.\u2019 The\n     fundamental principle ruling all my idealism, on the contrary, is\n     this: \u2018All cognition of things solely from pure understanding or\n     pure Reason is nothing but mere illusion and only in experience is\n     there truth.\u2019\u201d[1027]\n\nThis mode of defining idealism can, in this connection, cause nothing\nbut confusion. Its inapplicability to Berkeley would seem to prove that\nKant had no first-hand knowledge of Berkeley\u2019s writings.[1028] As Kant\u2019s\nNote to the Appendix to the _Prolegomena_[1029] shows, he also had Plato\nin mind. But the definition given of \u201cthe fundamental principle\u201d of his\nown idealism is almost equally misleading. It omits the all-essential\npoint, that for Kant experience itself yields truth only by conforming\nto _a priori_ concepts. As it is, he proceeds to criticise Berkeley for\nfailure to supply a sufficient criterion of distinction between truth\nand illusion. Such criterion, he insists, is necessarily _a priori_. The\nCritical idealism differs from that of Berkeley in maintaining that\nspace and time, though sensuous, are _a priori_, and that in combination\nwith the pure concepts of understanding they\n\n     \u201c...prescribe _a priori_ its law to all possible experience: the\n     law which at the same time yields the sure criterion for\n     distinguishing within experience truth from illusion. My so-called\n     idealism--which properly speaking is Critical idealism--is thus\n     quite peculiar in that it overthrows ordinary idealism, and that\n     through it all _a priori_ cognition, even that of geometry, now\n     attains objective reality, a thing which even the keenest realist\n     could not assert till I had proved the ideality of space and\n     time.\u201d[1030]\n\nV. =Sections added in Second Edition at the Conclusion of the Aesthetic.=\n(B 69 ff.)--Kant here again replies to the criticism of Pistorius that\nall existence has been reduced to the level of illusion (_Schein_). His\ndefence is twofold: first, that in naming objects appearances he means\nto indicate that they are independently grounded, or, as he states it,\nare \u201csomething actually given.\u201d If we _mis_interpret them, the result is\nindeed illusion, but the fault then lies with ourselves and not with the\nappearances as presented. Secondly, he argues that the doctrine of the\nideality of space and time is the only secure safeguard against\nscepticism. For otherwise the contradictions which result from regarding\nspace and time as independently real will likewise hold of their\ncontents, and everything, including even our own existence, will be\nrendered illusory. \u201cThe good Berkeley [observing these contradictions]\ncannot, indeed, be blamed for reducing bodies to mere illusion.\u201d This\nlast sentence may perhaps be taken as supporting the view that\nnotwithstanding the increased popularity of Berkeley in Germany and the\nappearance of new translations in these very years, Kant has not been\nsufficiently interested to acquire first-hand knowledge of Berkeley\u2019s\nwritings.[1031] The epithet employed is characteristic of the rather\ndepreciatory attitude which Kant invariably adopts in speaking of\nBerkeley.\n\nVI. =\u201cRefutation of Idealism\u201d in Second Edition of the \u201cCritique.\u201d= (B\n274-9, supplemented by note to B xxxix.).--The refutation opens by\nequating idealism with material idealism (so named in contradistinction\nto his own \u201cformal or rather Critical\u201d teaching). Within material\nidealism Kant distinguishes between the problematic idealism of\nDescartes, and the dogmatic idealism of Berkeley. The latter has, he\nsays, been overthrown in the _Aesthetic_. The former alone is dealt with\nin this refutation. This is the first occurrence in the _Critique_ of\nthe expression \u201cproblematic idealism\u201d: it is nowhere employed in the\nfirst edition.[1032] Problematic idealism consists in the assertion that\nwe are incapable of having experience of any existence save our own;\nonly our inner states are immediately apprehended; all other existences\nare determined by inference from them. The refutation consists in the\nproof that we have experience, and not mere imagination of outer\nobjects. This is proved by showing that inner experience, unquestioned\nby Descartes, is possible only on the assumption of outer experience,\nand that this latter is as immediate and direct as is the former.\n\n=Thesis.=--The empirically determined consciousness of my own existence\nproves the existence of objects in space outside me.[1033]\n\n=Proof.=--I am conscious of my own existence as determined in time. Time\ndetermination presupposes the perception of something permanent. But\nnothing permanent is intuitable in the empirical self. On the cognitive\nside (_i.e._ omitting feelings, etc., which in this connection are\nirrelevant), it consists solely of representations; and these demand a\npermanent, distinct from ourselves, in relation to which their changes,\nand so my own existence in the time wherein they change, may be\ndetermined.[1034] Thus perception of this permanent is only possible\nthrough a thing outside, and not through the mere representation of a\nthing outside. And the same must hold true of the determination of my\nexistence in time, since this also depends upon the apprehension of the\npermanent. That is to say, the consciousness of my existence is at the\nsame time an immediate awareness of the existence of other things\noutside me.\n\nIn the note to the _Preface_ to the second edition[1035] occurs the\nfollowing emphatic statement.\n\n     \u201cRepresentation of something permanent in existence is not the same\n     as permanent representation. For though the representation [of the\n     permanent] may be very changing and variable like all our other\n     representations, not excepting those of matter, it yet refers to\n     something permanent. This latter must therefore be an external\n     thing distinct from all my representations, and its existence must\n     be included in the determination of my own existence, constituting\n     with it but a single experience such as would not take place even\n     internally if it were not also at the same time, in part, external.\n     How this should be possible we are as little capable of explaining\n     further as we are of accounting for our being able to think the\n     abiding in time, the coexistence of which with the variable\n     generates the conception of change.\u201d\n\nThe argument of this note varies from that of B 274 ff. only in its use\nof an ambiguous expression which is perhaps capable of being taken as\nreferring to things in themselves, but which does not seem to have that\nmeaning. \u201cI am just as certainly conscious that there are things outside\nme which relate to my sense....\u201d\n\nIn B 277-8 Kant refers to the empirical fact that determination of time\ncan be made only by relation to outer happenings in space, such as the\nmotion of the sun. This is a point which is further developed in another\npassage which Kant added in the second edition.\n\n     \u201c...in order to understand the possibility of things in conformity\n     with the categories, and so to demonstrate the objective reality of\n     the latter, we need not merely intuitions, but intuitions that are\n     in all cases outer intuitions. When, for instance, we take the pure\n     concepts of relation, we find firstly that in order to obtain\n     something _permanent_ in intuition corresponding to the concept of\n     substance, and so to demonstrate the objective reality of this\n     concept, we require an intuition in space (of matter). For space\n     alone is determined as permanent, while time, and therefore\n     everything that is in inner sense, is in constant flux. Secondly,\n     in order to exhibit _change_ as the intuition corresponding to the\n     concept of _causality_, we must take as our example motion, _i.e._\n     change in space. Only in this way can we obtain the intuition of\n     changes, the possibility of which can never be comprehended through\n     any pure understanding. For change is combination of\n     contradictorily opposed determinations in the existence of one and\n     the same thing. Now how it is possible that from a given state of a\n     thing an opposite state should follow, not only cannot be conceived\n     by any reason without an example, but is actually incomprehensible\n     to reason without intuition. The intuition required is the\n     intuition of the movement of a point in space. The presence of the\n     point in different spaces (as a sequence of opposite\n     determinations) is what first yields to us an intuition of change.\n     For in order that we may afterwards make inner changes likewise\n     thinkable, we must represent time (the form of inner sense)\n     figuratively as a line, and the inner change through the drawing of\n     this line (motion), and so in this manner by means of outer\n     intuition make comprehensible the successive existence of\n     ourselves in different states. The reason of this is that all\n     change, if it is indeed to be perceived as change, presupposes\n     something permanent in intuition, and that in inner sense no\n     permanent intuition is to be met with. Lastly, the possibility of\n     the category of _community_ cannot be comprehended through mere\n     reason alone. Its objective reality is not to be understood without\n     intuition and indeed outer intuition in space.\u201d[1036]\n\nIn this passage Kant is modifying the teaching of the first edition in\ntwo very essential respects. In the first place, he is now asserting\nthat consciousness of both space and motion is necessary to\nconsciousness of time;[1037] and in the second place, he is maintaining\nthat the _categories_ can acquire meaning only by reference to outer\nappearances. Had Kant made all the necessary alterations which these new\npositions involve, he would, as we shall find,[1038] have had entirely\nto recast the chapters on _Schematism_ and on the _Principles of\nUnderstanding_. Kant was not, however, prepared to make such extensive\nalterations, and these chapters are therefore left practically\nunmodified. This is one of the many important points in which the reader\nis compelled to reinterpret passages of earlier date in the light of\nKant\u2019s later utterances. There is also a further difficulty. Does Kant,\nin maintaining that the categories can _acquire_ significance only in\nreference to outer perception, also mean to assert that their subsequent\nemployment is limited to the mechanical world of the material sciences?\nThis is a point in regard to which Kant makes no quite direct statement;\nbut indirectly he would seem to indicate that that was not his\nintention.[1039] He frequently speaks of the states of inner sense as\nmechanically conditioned. Sensations,[1040] feelings, and desires,[1041]\nare, he would seem to assert, integral parts of the unitary system of\nphenomenal existence. Such a view is not, indeed, easily reconcilable\nwith his equating of the principle of substance with the principle of\nthe conservation of matter.[1042] There are here two conflicting\npositions which Kant has failed to reconcile: the traditional dualistic\nattitude of Cartesian physics and the quite opposite implications of his\nCritical phenomenalism. When the former is being held to, Kant has to\nmaintain that psychology can never become a science;[1043] but his\nCritical teaching consistently developed seems rather to support the\nview that psychology, despite special difficulties peculiar to its\nsubject matter, can be developed on lines strictly analogous to those of\nthe material sciences.\n\nWe may now return to Kant\u2019s main argument. This new refutation of\nidealism in the second edition differs from that given in the fourth\n_Paralogism_ of the first edition, not only in method of argument but\nalso in the nature of the conclusion which it seeks to establish. Indeed\nit proves the _direct opposite_ of what is asserted in the first\nedition. The earlier proof sought to show that, as regards immediacy of\napprehension and subjectivity of existence, outer appearances stand on\nthe same level as do our inner experiences. The proof of the second\nedition, on the other hand, argues that though outer appearances are\nimmediately apprehended they must be existences distinct from the\nsubjective states through which the mind represents them. The two\narguments agree, indeed, in establishing immediacy, but as that which is\ntaken as immediately known is in the one case a subjective state and in\nthe other is an independent existence, the immediacy calls in the two\ncases for entirely different methods of proof. The first method\nconsisted in viewing outer experiences as a subdivision within our inner\nexperiences. The new method views their relation as not that of\nincluding and included, but of conditioning and conditioned; and it is\nnow to outer experience that the primary position is assigned. So far is\nouter experience from being possible only as part of inner experience,\nthat on the contrary inner experience, consciousness of the flux of\ninner states, is only possible in and through experience of independent\nmaterial bodies in space. A sentence from each proof will show how\ncompletely their conclusions are opposed.\n\n     \u201cOuter objects (bodies) are mere appearances, and are therefore\n     nothing but a species of my representations, the objects of which\n     are something only through these representations. Apart from them\n     they are nothing.\u201d[1044] \u201cPerception of this permanent is possible\n     only through a _thing_ outside me, and not through the mere\n     _representation_ of a thing outside me.\u201d[1045]\n\nThe one sentence asserts that outer objects are representations; the\nother argues that they must be existences distinct from their\nrepresentations. The one inculcates a subjectivism of a very extreme\ntype; the other results in a realism, which though ultimately\nphenomenalist, is none the less genuinely objective in character. This\ndifference is paralleled by the nature of the idealisms to which the two\nproofs are opposed and which they profess to refute. The argument of the\n_Paralogism_ of the first edition is itself Berkeleian, and refutes only\nthe problematic idealism of Descartes. The argument of the second\nedition, though formally directed only against Descartes, constitutes a\nno less complete refutation of the position of Berkeley. In its realism\nit has kinship with the positions of Arnauld and of Reid, while, in\nattempting to combine this realism with due recognition of the force and\nvalidity of Hume\u2019s sceptical philosophy, it breaks through all previous\nclassifications, formulates a profoundly original substitute for the\npreviously existing theories, and inaugurates a new era in the theory of\nknowledge.\n\nAs already pointed out,[1046] Kant restates the distinction between the\nsubjective and the objective in a manner which places the problem of\nknowledge in an entirely new light. The subjective is not to be regarded\nas opposite in nature to the objective, but as a subspecies within it.\nIt does not proceed parallel with the sequence of natural existences,\nbut is itself part of the natural system which consciousness reveals.\nSensations, in the form in which they are consciously apprehended by us,\ndo not constitute our consciousness of nature, but are themselves\nevents which are possible only under the conditions which the natural\nworld itself supplies.[1047] The Cartesian dualism of the subjective and\nthe objective is thus subordinated to the Critical distinction between\nappearance and reality. Kant\u2019s phenomenalism is a genuine alternative to\nthe Berkeleian teaching, and not, as Schopenhauer and so many others\nhave sought to maintain, merely a variant upon it.\n\nThe striking contradiction between Kant\u2019s various refutations of\nidealism has led some of Kant\u2019s most competent critics to give a\ndifferent interpretation of the argument of the second edition from that\ngiven above. These critics take the independent and permanent objects\nwhich are distinguished from our subjective representations to be things\nin themselves. That is to say, they interpret this refutation as based\nupon Kant\u2019s semi-Critical doctrine of the transcendental object (in the\nform in which it is employed for the solution of the _Antinomies_), and\nso as agreeing with the refutation given in the _Prolegomena_.[1048]\nKant is taken as rejecting idealism because of his belief in things in\nthemselves. This is the view adopted by Benno Erdmann,[1049]\nSidgwick,[1050] A. J. Balfour.[1051]\n\nAs Vaihinger,[1052] Caird,[1053] and Adamson[1054] have shown, such an\ninterpretation is at complete variance with the actual text. This is,\nindeed, so obvious upon unbiassed examination that the only point which\nrepays discussion is the question, why Benno Erdmann and those who\nfollow him should have felt constrained to place so unnatural an\ninterpretation upon Kant\u2019s words. The explanation seems to lie in\nErdmann\u2019s convinced belief, plainly shown in all his writings upon Kant,\nthat the _Critique_ expounds a single consistent and uniform\nstandpoint.[1055] If such belief be justified, there is no alternative\nsave to interpret Kant\u2019s refutation of idealism in the manner which\nErdmann adopts. For as the subjectivism of much of Kant\u2019s teaching is\nbeyond question, consistency can be obtained only by sacrifice of all\nthat conflicts with it. Thus, and thus alone, can Erdmann\u2019s rendering of\nthe refutation of the second edition be sustained; the actual wording,\ntaken in and by itself, does not support it. Kant here departs from his\nown repeated assertion, in the second hardly less than in the first\nedition of the _Critique_, of the subjectivity of outer appearances.\nBut, as Vaihinger justly contends, Kant was never greater than in this\nviolation of self-consistency, \u201cnever more consistent than in this\ninconsistency.\u201d Tendencies, previously active but hitherto inarticulate,\nare at last liberated. If the chrysalis stage of the intense brooding of\nthe twelve years of Critical thinking was completed in the writing of\nthe first edition of the _Critique_, the philosophy which then emerged\nonly attains to mature stature in those extensions of the _Critique_,\nscattered through it from _Preface_ to _Paralogisms_, which embody this\nrealistic theory of the independent existence of material nature. For\nthis theory is no mere external accretion, and no mere reversal of\nsubordinate tenets, but a ripening of germinal ideas to which, even in\ntheir more embryonic form, the earlier Critical teaching owed much of\nits inspiration, and which, when consciously adopted and maturely\nformulated, constitute such a deepening of its teaching as almost\namounts to transformation. The individual self is no longer viewed as\nbeing the bearer of nature, but as its offspring and expression, and as\nbeing, like nature, interpretable in its twofold aspect, as appearance\nand as noumenally grounded. The bearer of appearance is not the\nindividual subject, but those transcendental creative agencies upon\nwhich man and nature alike depend. Both man and nature transcend the\nforms in which they are apprehended; and nothing in experience justifies\nthe giving of such priority to the individual mind as must be involved\nin any acceptance of subjectivist theory. Though man is cognisant of\nspace and time, comprehending them within the limits of his\nconsciousness, and though in all experience unities are involved which\ncannot originate within or be explained by experience, it is no less\ntrue that man is himself subject to the conditions of space and time,\nand that the synthetic unities which point beyond experience do not\ncarry us to a merely individual subject. If man is not a part or product\nof nature, neither is nature the product of man. Kant\u2019s\ntranscendentalism, in its maturest form, is genuinely phenomenalist in\ncharacter. That is the view which has already been developed above, in\nthe discussion of Kant\u2019s transcendental deduction. I shall strive to\nconfirm it by comparison of the teaching of the two editions of the\n_Critique_ in regard to the reality of outer appearances.\n\nSchopenhauer, to whom this new development of the Critical teaching was\naltogether anathema, the cloven hoof of the Hegelian heresies,\ndenounced it as a temporary and ill-judged distortion of the true\nCritical position, maintaining that it is incapable of combination with\nKant\u2019s central teaching, and that it finds no support in the tenets,\npure and unperverted, of the first edition. Kant, he holds, is here\nuntrue to himself, and temporarily, under the stress of polemical\ndiscussion, lapses from the heights to which he had successfully made\nhis way, and upon which he had securely established, in agreement with\nPlato and in extension of Berkeley, the doctrine of all genuine\nphilosophical thinking, the doctrine of the _Welt als Vorstellung_.\n\nWe may agree with Schopenhauer in regarding those sections of the first\nedition of the _Critique_ which were omitted in the second edition as\nbeing a permanently valuable expression of Kantian thought, and as\ncontaining much that finds no equally adequate expression in the\npassages which were substituted for them; and yet may challenge his\ninterpretation of both editions alike. If, as we have already been\narguing, we must regard Kant\u2019s thinking as in large degree tentative,\nthat is, as progressing by the experimental following out of divergent\ntendencies, we may justly maintain that among the most characteristic\nfeatures of his teaching are the readiness with which he makes changes\nto meet deeper insight, and the persistency with which he strives to\nattain a position in which there will be least sacrifice or blurring of\nany helpful distinction, and fullest acknowledgment of the manifold and\ndiverse considerations that are really essential. Recognising these\nfeatures, we shall be prepared to question the legitimacy of\nSchopenhauer\u2019s opposition between the teaching of the two editions. We\nshall rather expect to find that the two editions agree in the\nalternating statement and retraction of conflicting positions, and that\nthe later edition, however defective in this or that aspect as compared\nwith the first edition, none the less expresses the maturer insight, and\nrepresents a further stage in the development of ideas that have been\npresent from the start. It may perhaps for this very reason be more\ncontradictory in its teaching; it will at least yield clearer and more\nadequate formulation of the diverse consequences and conflicting\nimplications of the earlier tenets. It will be richer in content, more\nopen-eyed in its adoption of mutually contradictory positions, freer\ntherefore from unconscious assumptions, and better fitted to supply the\ndata necessary for judgment upon its own defects. Only those critics who\nare blind to the stupendous difficulties of the tasks which Kant here\nsets himself, and credulous of their speedy and final completion, can\ncomplain of the result. Philosophical thinkers of the most diverse\nschools in Germany, France, and England, have throughout the nineteenth\ncentury received from the _Critique_ much of their inspiration. The\nprofound influence which Kant has thus exercised upon succeeding thought\nmust surely be reckoned a greater achievement than any that could have\nresulted from the constructing of a system so consistent and unified,\nthat the alternative would lie only between its acceptance and its\nrejection. Ultimately the value of a philosophy consists more in the\nrichness of its content and the comprehensiveness of its dialectic, than\nin the logical perfection of its formal structure. The latter quality is\nespecially unfitted to a philosophy which inaugurated a new era, and\nformulated the older problems in an altogether novel manner. Under such\nconditions fertility of suggestion and readiness to modify or even\nrecast adopted positions, openness to fuller insight acquired through\nthe very solutions that may at first have seemed to satisfy and close\nthe issues, are more to be valued than the power to remove\ncontradictions and attain consistency. This is the point of view which I\nshall endeavour to justify in reference to the matters now before us. In\nparticular there are two points to be settled: first, whether and how\nfar the argument of the second edition is prefigured in the first\nedition; and secondly, whether and to what extent it harmonises with,\nand gives expression to, all that is most central and genuinely Critical\nin both editions.\n\nIn the first place we must observe that the fourth _Paralogism_ occurs\nin a section which bears all the signs of having been independently\nwritten and incorporated later into the main text. It is certainly of\nearlier origin than those sections which represent the third and fourth\nlayers of the deduction of the first edition, and very possibly was\ncomposed in the middle \u2019seventies. Indeed, apart from single paragraphs\nwhich may have been added in the process of adapting it to the main\ntext, it could quite well, so far as its refutation of idealism is\nconcerned, be of even earlier date. The question as to the consistency\nof the refutation of the second edition with the teaching of the first\nedition must therefore chiefly concern those parts of the _Analytic_\nwhich connect with the later forms of the transcendental deduction, that\nis to say, with the transcendental deduction itself, with the\n_Analogies_ and _Postulates_, and with particular paragraphs that have\nbeen added in other sections. We have already noted how Kant from the\nvery first uses terms which involve the drawing of a distinction between\nrepresentations and their objects. Passages in which this distinction\noccurs can be cited from both the _Aesthetic_ and the _Analytic_, and\ntwo such occur in the fourth _Paralogism_ itself.[1056] Objects, he\nsays, \u201ccorrespond\u201d to their representations. A variation in expression\nis found in such passages as the following:\n\n     \u201c...the objects of outer perception also actually exist (_auch\n     wirklich sind_) in that very form in which they are intuited in\n     space....\u201d[1057]\n\nSuch language is meaningless, and could never have been chosen, if Kant\nhad not, even in the earlier stages of his thinking, postulated a\ndifference between the existence of an object and the existence of its\nrepresentation. He must at least have distinguished between the\nrepresentations and their content. That, however, he could have done\nwithout advancing to the further assertion of their independent\nexistence. Probably he was not at all clear in his own mind, and was too\npreoccupied with the other complexities of his problem, to have thought\nout his position to a definite decision. When, however, as in the fourth\n_Paralogism_, he made any attempt so to do, he would seem to have felt\nconstrained to adopt the extreme subjectivist position. Expressions to\nthat effect are certainly very much more common than those above\nmentioned. This is what affords Schopenhauer such justification,\ncertainly very strong, as he can cite for regarding subjectivism as the\nundoubted teaching of the first edition.\n\nWhen, however, we also take account of the very different teaching which\nis contained in the important section on the _Postulates of Empirical\nThought_, the balance of evidence is decisively altered. The\ncounter-teaching, which is suggested by certain of the conflicting\nfactors of the transcendental deduction and of the _Analogies_, here\nagain receives clear and detailed expression. This is the more\nsignificant, as it is in this section that Kant sets himself formally to\ndefine what is to be understood by empirical reality. It thus contains\nhis, so to speak, official declaration as to the mode of existence\npossessed by outer appearances. The passage chiefly relevant is as\nfollows:\n\n     \u201cIf the existence of the thing is bound up with some perceptions\n     according to the principles of their empirical connection (the\n     Analogies), we can determine its existence antecedently to the\n     perception of it, and consequently, to that extent, in an _a\n     priori_ manner. For as the existence of the thing is bound up with\n     our perceptions in a possible experience, we are able in the series\n     of possible perceptions, and under the guidance of the Analogies,\n     to make the transition from our actual perception to the thing in\n     question. Thus we discover the existence of a magnetic matter\n     pervading all bodies from the perception of the attracted iron\n     filings, although the constitution of our organs cuts us off from\n     all immediate perception of that matter. For in accordance with the\n     laws of sensibility and the connection of our perceptions in a\n     single experience, we should, were our senses more refined,\n     actually experience it in an immediate empirical intuition. The\n     grossness of our senses does not in any way decide the form of\n     possible experience in general.\u201d[1058]\n\nNow it cannot, of course, be argued that the above passage is altogether\nunambiguous. We can, if we feel sufficiently constrained thereto, place\nupon it an interpretation which would harmonise it with Kant\u2019s more\nusual subjectivist teaching, namely as meaning that in the progressive\nconstruction of experience, or in the ideal completion which follows\nupon assumption of more refined sense-organs, possible empirical\nrealities are made to become, or are assumed to become, real, but that\nuntil the possible experiences are thus realised in fact or in ideal\nhypothesis, they exist outwardly only in the form of their noumenal\nconditions. And as a matter of fact, this is how Kant himself interprets\nthe teaching of this section in the process of applying it in solution\nof the antinomies.\n\n     \u201cAccordingly, if I represent to myself the aggregate of all objects\n     of the senses existing in all time and all places, I do not set\n     them, antecedently to experience, in space and time. The\n     representation is nothing but the thought of a possible experience\n     in its absolute completeness. Since the objects are mere\n     representations, only in such a possible experience are they given.\n     To say that they exist prior to all my experience, can only be\n     taken as meaning that they will be met with, if, starting from\n     actual perception, I advance to that part of experience to which\n     they belong. The cause of the empirical conditions of this advance\n     (that which determines what members I shall meet with, or how far I\n     can meet with any such in my regress) is transcendental, and is\n     therefore necessarily unknown to me. We are not, however, concerned\n     with this transcendental cause, but only with the rule of\n     progression in that experience in which objects, that is to say,\n     appearances, are given. Moreover, in outcome it is a matter of\n     indifference whether I say that in the empirical progress in space\n     I can meet with stars a hundred times farther removed than the\n     outermost now perceptible to me, or whether I say that they are\n     perhaps to be met with in cosmical space even though no human being\n     has ever perceived or ever will perceive them. For though they\n     might be given as things in themselves, without relation to\n     possible experience, they are still nothing to me, and therefore\n     are not objects, save in so far as they are contained in the series\n     of the empirical regress.\u201d[1059]\n\nBut though this is a possible interpretation of the teaching of the\n_Postulates_, and though further it is Kant\u2019s own interpretation in\nanother portion of the _Critique_, it is not by any means thereby\ndecided that this is what the section itself actually teaches. Unbiassed\nstudy of the section, in independence of the use to which it is\nelsewhere put, can find within it no such limitation to its assertion of\nthe actual independent existence of non-perceived bodies. We have to\nremember that the doctrine and solution of the _Antinomies_ was\ncompleted prior to the writing of the central portions of the\n_Critique_. The section treating of their _solution_ seems, indeed, in\ncertain parts to be later[1060] than the other main portions of the\nchapter on the _Antinomies_, and must have been at least recast after\ncompletion of the _Postulates_. But the subjectivist solution is so much\nsimpler in statement, so much more fully worked out, and indeed so much\nmore capable of definite formulation, and also so much more at one with\nthe teaching developed in the preceding chapter on the _Paralogisms_,\nthat even granting the doctrine expounded in the section on the\n_Postulates_ to be genuinely phenomenalist, it is not surprising that\nKant should have been unwilling to recast his older and simpler solution\nof the _Antinomies_. In any case we are not concerned to argue that\nKant, even after formulating the phenomenalist view, yields to it an\nunwavering adherence. As I have already insisted, his attitude continues\nto the very last to be one of alternation between two opposed\nstandpoints.\n\nBut the most significant feature of Kant\u2019s treatment of the argument of\nthe _Postulates_ still remains for consideration. It was in immediate\nsuccession to the paragraph above quoted[1061] that Kant, in the second\nedition, placed his \u201c_Refutation of Idealism_\u201d with the emphatic\nstatement that this (not as in the first edition in connection with the\n_Paralogisms_) was its \u201ccorrect location.\u201d It is required, he says, as a\nreply to an objection which the teaching of the _Postulates_ must at\nonce suggest. The argument of the second edition in proof of the\nindependent reality of material bodies, and in disproof of subjectivism,\nis thus given by Kant as a necessary extension and natural supplement of\nthe teaching of the first edition.\n\nThere is therefore reason for concluding that the same preconception\nwhich has led to such radical misinterpretation of Kant\u2019s _Refutation of\nIdealism_ has been at work in inducing a false reading of Kant\u2019s\nargument in the _Postulates_, namely the belief that Kant\u2019s teaching\nproceeds on consistent lines, and that it must at all costs be\nharmonised with itself. Finding subjectivism to be emphatically and\nunambiguously inculcated in all the main sections of the _Critique_, and\nthe phenomenalist views, on the other hand, to be stated in a much less\ndefinite and somewhat elusive manner, commentators have impoverished the\nCritical teaching by suppression of many of its most subtile and\nprogressive doctrines. Kant\u2019s experimental, tentative development of\ndivergent tendencies is surely preferable to this artificial product of\nhigh-handed and unsympathetic emendation.\n\n\nINNER SENSE AND APPERCEPTION\n\nWe are now in position to complete our treatment of inner sense. When\nthe inner world of feelings, volitions, and representations is placed on\nthe same empirical level as the outer world of objects in space, when\nthe two are correlated and yet also at the same time sharply\ndistinguished, when, further, it is maintained that objects in space\nexist independently of their representations, and that in this\nindependence they are necessary for the possibility of the latter, the\nwhole aspect of the Critical teaching undergoes a genial and welcome\ntransformation. Instead of the forbidding doctrine that the world in\nspace is merely my representation, we have the very different teaching\nthat only through consciousness of an independent world in space is\nconsciousness of the inner subjective life possible at all, and that as\neach is \u201cexternal\u201d to the other, neither can be reduced to, or be\nabsorbed within, the other. The inner representations do not produce or\ngenerate the spatial objects, do not even condition their existence, but\nare required only for the individual\u2019s empirical consciousness of them.\nIndeed the relations previously holding between them are now reversed.\nIt is the outer world which renders the subjective representations\npossible. The former is prior to the latter; the latter exist in order\nto reveal the former. The outer world in space must, indeed, be regarded\nas conditioned by, and relative to, the noumenal conditions of its\npossibility; but these, on Kant\u2019s doctrine of outer and inner sense, are\ndistinct from all experienced contents and from all experienced mental\nprocesses. This will at once be recognised as holding of the noumenal\nconditions of the given manifold. But it is equally true, Kant\nmaintains, in regard to the noumenal conditions of our mental life. We\nhave no immediate knowledge of the transcendental syntheses that\ncondition all consciousness, and in our complete ignorance of their\nspecific nature they cannot legitimately be equated with any individual\nor personal agent. As the empirical self is only what it is known as,\nnamely, appearance, it cannot be the bearer of appearance. This function\nfalls to that which underlies both inner and outer appearances equally,\nand which within experience gains twofold expression for itself, in the\nconception of the thing in itself = _x_ on the one hand, and in the\ncorrelative conception of a transcendental subject, likewise = _x_, on\nthe other.\n\nBut with mention of the transcendental subject we are brought to a\nproblem which in the second edition invariably accompanies Kant\u2019s\ndiscussion of inner sense. The \u2018I think\u2019 of apperception can find\nexpression only in an empirical judgment, and yet, so far from being the\noutcome of inner sense, preconditions its possibility. What then is its\nrelation to inner sense? Does not its recognition conflict with Kant\u2019s\ndenial of the possibility of self-conscious reflection, of direct\nintuitive apprehension by the self of itself? The pure apperception, \u2018I\nthink,\u2019 is equivalent, Kant declares, to the judgment \u2018I am,\u2019 and\ntherefore involves the assertion of the subject\u2019s existence.[1062] Does\nnot this conflict on the one hand with the Critical doctrine that\nknowledge of existence is only possible in terms of sense, and on the\nother with the Critical limitation of the categories to the realm of\nappearance? How are such assertions as that the \u2018I think\u2019 of pure\napperception refers to a non-empirical reality, and that it predicates\nits existence, to be reconciled with the doctrine of inner sense as\nabove stated?\n\nAs we have already observed,[1063] Kant\u2019s early doctrine of the\ntranscendental object was developed in a more or less close parallelism\nwith that of the transcendental unity of apperception. They were\nregarded as correlative opposites, the dual centres of noumenal\nreference for our merely subjective representations. Kant\u2019s further\nexamination of the nature of apperception, as embodied in alterations in\nthe second edition, was certainly, as we shall find, inspired by the\ncriticisms which the first edition had called forth. His replies,\nhowever, are merely more explicit statements of the distinction which he\nhad already developed in the first edition between the transcendental\nand the empirical self, and that distinction in turn was doubtless\nitself largely determined by his own independent recognition of the\nuntenability of his early view of the transcendental object. Though it\nis much more difficult to differentiate between the empirical and the\ntranscendental self than to distinguish between the empirical object and\nthe thing in itself, both distinctions are from a genuinely Critical\nstandpoint equally imperative, and rest upon considerations that are\nsomewhat similar in the two cases.\n\nOne of the chief and most telling criticisms directed against the\nteaching of the first edition was that Kant\u2019s doctrine of a\ntranscendental consciousness of the self\u2019s existence, _i.e._ of the\nexistence of a noumenal being, \u201cthis I or he or it (the thing) which\nthinks,\u201d[1064] is inconsistent with the teaching of the _Postulates of\nEmpirical Thought_. In that section, as also later in the section on the\ntheological _Ideal_, Kant had declared most emphatically that existence\nis never discoverable in the content of any mere concept. It is revealed\nin perception, and in perception alone, in virtue of the element of\nsensation contained in the latter.\n\n     \u201c...to know the _actuality_ of things demands _perception_, and\n     therefore sensation.... For that the concept precedes perception,\n     signifies the concept\u2019s mere possibility; the perception which\n     supplies the content [_Stoff_] to the concept, is the sole\n     criterion [_Charakter_] of actuality.\u201d[1065]\n\nYet Kant had also maintained that the \u2018I think\u2019 is equivalent to \u2018I\nam,\u2019[1066] and that in this form, as an intellectual consciousness of\nthe self\u2019s existence, it precedes all experience. The teaching of the\n_Postulates_ is, however, the teaching of the _Critique_ as a whole, and\nsuch critics as Pistorius seemed therefore to be justified in\nmaintaining that Kant, in reducing the experiences of inner sense to\nmere appearance, destroys the possibility of establishing reality in any\nform. Appearance, in order to be appearance, presupposes the reality not\nonly of that which appears, but also of the mental process whereby it is\napprehended. But if reality is given only in sensation, and yet all\nexperience that involves sensation is merely appearance, there is no\nself by which appearance can be conditioned; and only illusion\n(_Schein_), not appearance (_Erscheinung_), is left. To quote Pistorius\u2019\nexact words:\n\n     \u201d[If our inner representations are not things in themselves but\n     only appearances] there will be nothing but illusion (_Schein_),\n     for nothing remains to which anything can appear.\u201d[1067]\n\nKant evidently felt the force of this criticism, for in the second\nedition he replies to it on no less than seven different\noccasions.[1068] In three of these passages[1069] the term _Schein_ is\nemployed, and in the note to B xxxix the term _Erdichtung_ appears. This\nshows very conclusively that it is such criticism as the above that Kant\nhas in mind. The most explicit passage is B 428:\n\n     \u201cThe proposition, \u2018I think,\u2019 or \u2018I exist thinking,\u2019 is an empirical\n     proposition. Such a judgment, however, is conditioned by empirical\n     intuition, and the object that is thought therefore underlies it as\n     appearance. It would consequently seem that on our theory the soul\n     is completely transformed, even in thinking [_selbst im Denken_],\n     into appearance, and that in this way our consciousness itself, as\n     being a mere illusion [_Schein_], must refer in fact to nothing.\u201d\n\nKant, in his reply, is unyielding in the contention that the \u2018I think,\u2019\neven though it involves an empirical judgment, is itself intellectual.\n\u201cThis representation is a thinking, not an intuiting,\u201d[1070] or as he\nadds, \u201cThe \u2018I think\u2019 expresses the _actus_ whereby I determine my\nexistence.\u201d Existence is therefore already given thereby.[1071] Kant\nalso still maintains that the self thus revealed is not \u201cappearance and\nstill less illusion.\u201d\n\n     \u201cI am conscious of myself ..., not as I appear to myself, nor as I\n     am in myself, but only that I am.\u201d[1072] \u201cI thereby represent\n     myself to myself neither as I am nor as I appear to myself. I think\n     myself only as I do any object in general from whose mode of\n     intuition I abstract.\u201d[1073]\n\nKant\u2019s method of meeting the criticism, while still holding to these\npositions, is twofold. It consists in the first place in maintaining\nthat the \u2018I think,\u2019 though intellectual, can find expression only in\nempirical judgments--in other words, that it is in and by itself formal\nonly, and presupposes as the occasion of its employment a given manifold\nof inner sense; and secondly, by the statement that the \u2018existence\u2019\nwhich is involved in the \u2018I think\u2019 is not the category of existence. Let\nus take in order each of these two points.\n\nKant\u2019s first method of reply itself appears in two forms, a stronger and\na milder. The milder mode of statement[1074] is to the effect that\nthough the representation \u2018I am\u2019 already immediately involves the\n_thought_ of the existence of the subject, it yields no knowledge of it.\nKnowledge would involve intuition, namely, consciousness of inner\ndeterminations in time, which in turn would itself presuppose\nconsciousness of outer objects. As a merely intellectual\nrepresentation,\n\n     \u201c...this \u2018I\u2019 has not the least predicate of intuition which, in its\n     character of permanence, could, somewhat after the manner of\n     impenetrability in the empirical intuition of matter, serve as\n     correlate of time determination in inner sense.\u201d[1075]\n\nThe stronger and more definite mode of statement is that the \u2018I think\u2019\nis an empirical proposition.[1076] Though it involves as one factor the\nintellectual representation, \u2018I think,\u2019 it is none the less empirical.\n\n     \u201cWithout some empirical representation supplying the material for\n     thought, the _actus_, \u2018I think,\u2019 would not take place....\u201d[1077]\n\nThe empirical is indeed \u201conly the condition of the application or\nemployment of the pure intellectual faculty,\u201d but as such is\nindispensable. This is repeated in even clearer terms in B 429.\n\n     \u201cThe proposition, \u2018I think,\u2019 in so far as it amounts to the\n     assertion, \u2018I exist thinking,\u2019 is no mere logical function but\n     determines the subject (which is then at the same time object) in\n     respect of existence, and cannot take place without inner\n     sense....\u201d\n\nThis admission is the more significant in that it follows immediately\nupon a passage in which Kant has been arguing that thinking, taken in\nand by itself, is a merely logical function.\n\nThe real crux lies in the question as to the legitimacy of Kant\u2019s\napplication of the predicate existence to the transcendental subject.\nIts employment in reference to the empirical self in time is part of the\nproblem of the _Refutation of Idealism_ in the second edition; and the\nanswer there given is clear and definite. Consciousness of the empirical\nself as existing in time involves consciousness of outer objects in\nspace. But as Kant recognises that a transcendental ego, not in time, is\npresupposed in all consciousness of the empirical self, the question\nwhether the predicate of existence is also applicable to the\ntranscendental self cannot be altogether avoided, and is indeed referred\nto in B 277. The attitude to be taken to this latter question is not,\nhowever, defined in that section.\n\nIn the first edition Kant has insisted that the categories as pure forms\nof the understanding, in isolation from space and time, are merely\nlogical functions \u201cwithout content.\u201d Interpreted literally, this would\nsignify that they are devoid of meaning, and therefore are incapable of\nyielding the thought of any independent object or existence. As merely\nlogical forms of relation, they presuppose a material, and that is\nsupplied only through outer and inner sense. Such is not, however, the\nway in which Kant interprets his own statement. It is qualified so as to\nsignify only that they are without _specific_ or _determinate_ content.\nThey are taken as yielding the conception of object in general. Passages\nin plenty can be cited from the first edition[1078]--passages allowed to\nremain in the second edition--in which Kant teaches that the pure forms\nof understanding, as distinct from the schematised categories, yield the\nconception of things in themselves. This view is, indeed, a survival\nfrom his earlier doctrine of the transcendental object.[1079] In all\npassages added in the second edition the consequences of his argument\nare more rigorously drawn, and the doctrine of the transcendental object\nis entirely eliminated. It is now unambiguously asserted that the pure\nforms of understanding, the \u201c_modes_ of self-consciousness in\nthinking,\u201d[1080] are not intellectual concepts of _objects_. They \u201cyield\nno object whatsoever.\u201d The only object is that given through sense. And\nsince in thinking the transcendental subject we do, by Kant\u2019s own\naccount, think an \u201cobject,\u201d he is led to the conclusion, also explicitly\navowed, that the notion of existence involved in the \u2018I think\u2019 is not\nthe category of the same name.[1081] So also of the categories of\nsubstance and causality.\n\n     \u201cIf I represent myself as _subject_ of thoughts or as _ground_ of\n     thinking, these modes of representation do not signify the\n     categories of substance or of cause....\u201d[1082]\n\nThe notion of the self, like the notion of things in themselves, is a\nconcept distinct from all the categories.[1083]\n\nThis conclusion is reinforced by means of an argument which is employed\nin the section of the first edition on _Paralogisms_. Apperception is\nthe ground of the possibility of the categories, and these latter on\ntheir side represent only the synthetic unity which that apperception\ndemands. Self-consciousness is therefore the representation of that\nwhich is the condition of all unity, and which yet is itself\nunconditioned.\n\n     \u201c...it does not represent itself through the categories, but knows\n     the categories and through them all objects in the absolute unity\n     of apperception, and so through itself. Now it is, indeed, very\n     evident that I cannot know as an object that which I must\n     presuppose in order to know any object....\u201d[1084]\n\nThis argument recurs in B 422.\n\n     \u201cThe subject of the categories cannot by thinking the categories\n     acquire a conception of itself as an object of the categories. For,\n     in order to think them, its pure self-consciousness, which is what\n     was to be accounted for, must itself be presupposed.\u201d\n\nIt is extremely difficult to estimate the value and cogency of this\nargument.[1085] Many objections or rather qualifications must be made\nbefore it can be either accepted or rejected. If it be taken only as\nasserting that the unity of self-consciousness is not _adequately_\nexpressible through any of the categories, it is undoubtedly valid. If,\nfurther, the categories be identified with the schemata, it is also true\nthat they are not applicable in any degree or manner. The schemata are\napplicable only to natural existences in space and time.\nSelf-consciousness can never be reduced to a natural existence of that\ntype. On the other hand, if it is not self-consciousness as such, but\nthe self-conscious _subject_, which on Kant\u2019s view is always\nnoumenal--\u201cthis I or he or it (the thing) which thinks\u201d[1086]--that is\nreferred to, and if we distinguish between the categories strictly so\ncalled, that is, the pure forms of understanding, and the schemata, it\nis not at all evident that the self-conscious subject may not be\ndescribed as being an existence that is always a subject and never a\npredicate, and as being related to experience as a ground or condition.\nThese indefinite assertions leave open alternative possibilities. They\ndo not even decide whether the self is \u201cI or he or it.\u201d[1087] In so far\nas they advance beyond the mere assertion that the self rests upon\nnoumenal conditions they are, indeed, incapable of proof, but by no\nCritical principle can they be shown to be inapplicable. When,\ntherefore, Kant may seem to extract a more definite conclusion from the\nabove argument,[1088] he advances beyond what it can be made to\nsupport.\n\nKant is here influenced by the results of the ethical enquiries with\nwhich in the period subsequent to 1781 he was chiefly preoccupied. He\nbelieved himself to have proved that the self, as a self-conscious\nbeing, is a genuinely noumenal existence. That being so, he was bound to\nhold that the categories, even as pure logical forms, are inadequate to\nexpress its real determinate nature. But he confounds this position with\nthe assertion that they are not only inadequate, but in and by\nthemselves are likewise inapplicable. That is not a legitimate\nconclusion, for even if the self is more than mere subject or mere\nground, it will at least be so much. When ethical considerations are\nleft out of account, the only proper conclusion is that the\napplicability of the categories to the self-conscious subject is capable\nneither of proof nor of disproof, but that when the distinction between\nappearance and reality (which as we shall find is ultimately based upon\nthe Ideas of Reason) has been drawn, the categories can be employed to\ndefine the possible difference between self-conscious experience and its\nunknown noumenal conditions. Any other conclusion conflicts with the\nteaching of the section on the _Paralogisms_.\n\nIt is important to observe--a point ignored by such critics as Caird and\nWatson--that in the sections under consideration[1089] Kant most\nexplicitly declares self-consciousness to be merely \u201cthe\n_representation_ of that which is the condition of all unity.\u201d He\nmaintains that this representation, as standing for \u201cthe determining\nself (the thinking), is to be distinguished from the self which we are\nseeking to determine (the subject which thinks) as knowledge from its\nobject,\u201d[1090] or in other words, that, without special proof,\nunattainable on theoretical grounds, \u201cthe unity of thought\u201d may not be\ntaken as equivalent to the unity of the thinking subject.[1091] They may\nbe as diverse as unity of representation and unity of object represented\nare frequently found to be. We may never argue from simplicity in a\nrepresentation to simplicity in its object.\n\nBut to return to the main thesis, it may be observed that these\narguments, with the exception of that which we have just been\nconsidering from the nature of self-consciousness, lead to the\nconclusion that the categories are as little applicable to the thing in\nitself as to the transcendental subject. Even the argument from the\nnecessary and invariable presence of self-consciousness in each and\nevery act of judgment is itself valid only from a point of view which\nregards self-consciousness in the manner of Kant\u2019s early semi-Critical\nview of the transcendental subject[1092] as an ultimate. But if, as is\nmaintained in the section in which this argument occurs, viz. that on\nthe _Paralogisms_, self-consciousness may be complexly conditioned, and\nmay indeed have conditions similar in nature to those which underlie\nouter experience, the categories may be just as applicable, or as\ninapplicable, to its noumenal nature as to the nature of the thing in\nitself. It is noticeable that in the second edition, doubtless under the\ninfluence of preoccupation with ethical problems, some of Kant\u2019s\nutterances betray a tendency to relax the rigour of his thinking, and to\nbring his theoretical teaching into closer agreement with his ethical\nresults than the theoretical analysis in and by itself at all justifies.\nThis tendency was, of course, reinforced by the persisting influence of\nthat view of the transcendental subject which he had held in the middle\n\u2019seventies, and from which he never completely emancipated either his\nlanguage or his thinking.[1093] Indeed in several of the passages added\nin the second edition[1094] Kant even goes so far as to adopt language\nwhich if taken quite literally would mean that the \u2018I think\u2019 is an\nimmediate consciousness of the mind\u2019s purely intellectual activity--a\nview which, as we have seen,[1095] is altogether alien to the Critical\nposition. It would, as he argues so forcibly elsewhere, involve a kind\nof experience which does not conform to Critical requirements, and which\nwould lie open to the attacks of sceptics such as Hume.\n\nIn B 157-8 the difficulties of Kant\u2019s position are again manifest.\nSpeaking of the representation of the self, he declares that \u201cI am\nconscious of myself ..., not as I appear to myself, nor as I am in\nmyself, but only that I am.\u201d This may seem to imply that existence is\npredicable of the transcendental self. He adds that though the\ndetermination, _i.e._ specification in empirical form, of my existence\n(_mein eigenes Dasein_) is possible only in inner sensuous intuition, it\nis \u201cnot appearance and still less mere illusion.\u201d But in the appended\nnote it is urged that my existence (_Dasein_) as self-active being is\nrepresented in purely indeterminate fashion. Only my existence _as\nsensuous_, and therefore as appearance, can be known, _i.e._ can be made\ndeterminate.\n\nThe problem is more directly and candidly faced in the note to B 422.\nThat note is interesting for quite a number of reasons. It reveals Kant\nin the very act of recasting his position, and in the process of\nsearching around for a mode of formulation which will enable him to hold\nto a transcendental consciousness of the self\u2019s existence and at the\nsame time not to violate the definition of existence given in the\n_Postulates_, _i.e._ both to posit the transcendental self as actual and\nyet to deny the applicability to it of any of the categories. After\nstating that the \u2018I think\u2019 is an empirical proposition in which my\nexistence is immediately involved, he proceeds further to describe it as\nexpressing \u201can undetermined empirical intuition, _i.e._ perception,\u201d and\nso as showing that sensation underlies its assertion of existence. Kant\ndoes not, however, mean by these words that the existence asserted is\nmerely that of the empirical self; for he proceeds:\n\n     \u201c...existence is here not a category, which as such does not apply\n     to an indeterminately-given object.... An indeterminate perception\n     here signifies only something real that is given, given indeed to\n     thought in general, and so not as appearance, nor as thing in\n     itself (_Noumenon_), but as something which actually [_in der\n     That_] exists, and which in the proposition, I think, is denoted\n     [_bezeichnet_] as such.\u201d\n\nThe phrases here employed are open to criticism on every side. Kant\ncompletely departs from his usual terminology when he asserts that\nthrough an \u201cindeterminate perception\u201d the self is given, and \u201cgiven to\nthought in general\u201d as \u201csomething real.\u201d The contention, that the\nexistence asserted is not a category, is also difficult to accept.[1096]\nIt is equally surprising to read that its reality is given \u201cneither as\nappearance nor as thing [_Sache_] in itself (_Noumenon_)\u201d; for hitherto\nno such alternative form of real existence has been recognised.\n\nBut to press such criticisms is to ignore the spirit for the sake of the\nletter. Kant here breaks free from all his habitual modes of expression\nfor the very good and sufficient reason that he is striving to develop a\nposition more catholic and comprehensive than any previously adopted. He\nis seeking to formulate a position which, without in any way justifying\nor encouraging the transcendent employment of the categories, will yet\nretain for thought the capacity of self-limitation, that is, of forming\nconcepts which will reveal the existence of things in themselves and so\nwill enable the mind to apprehend the radical distinction between things\nin themselves and things experienced. But he has not yet discovered that\nin so doing he is committing himself to the thesis that the distinction\nis mediated, not by the understanding, but by Reason, not by categories,\nbut by Ideas.[1097] As I have already indicated, this tendency is\ncrossed by another derived from his preoccupation with moral problems,\nnamely, the desire to defend, in a manner which his Critical teaching\ndoes not justify, the noumenal existence of the self as a _thinking_\nbeing.\n\n\n\n\nTHE TRANSCENDENTAL ANALYTIC\n\n\n\n\nBook II\n\nTHE ANALYTIC OF PRINCIPLES\n\n\nThe distinction which Kant here introduces for the first time between\nunderstanding (now viewed as the faculty only of concepts) and the\nfaculty of judgment (_Urtheilskraft_) is artificial and extremely\narbitrary.[1098] As we have seen,[1099] his own real position involves a\ncomplete departure from the traditional distinction between conceiving,\njudging, and reasoning, as separate processes. All thinking without\nexception finds expression in judgment. Judgment is the fundamental\nactivity of the understanding. It is \u201can act which contains all its\nother acts.\u201d Kant is bent, however, upon forcing the contents of the\n_Critique_ into the external framework supplied by the traditional\nlogic, viewed as an architectonic; and we have therefore no option save\nto take account of his exposition in the actual form which he has chosen\nto give to it. Since general logic develops its teaching under three\nseparate headings, as the logic of conception, the logic of judgment,\nand the logic of reasoning, the _Critique_ has to be made to conform to\nthis tripartite division. The preceding book is accordingly described as\ndealing with concepts, and this second book as dealing with judgments or\nprinciples; while understanding and the faculty of judgment, now viewed\nas independent, are redefined to meet the exigencies of this new\narrangement, the former as being \u201cthe faculty of rules,\u201d and the latter\nas being \u201cthe faculty of subsuming under rules, _i.e._ of distinguishing\nwhether something does or does not stand under a given rule (_casus\ndatae legis_).\u201d\n\nThe reader need not strive to discover any deep-lying ground or\njustification for these definitions.[1100] Architectonic, that \u2018open\nsesame\u2019 for so many of the secrets of the _Critique_, is the\nall-sufficient spell to resolve the mystery. As a matter of fact, Kant\nis here taking advantage of the popular meaning of the term judgment in\nthe sense in which we speak of a man of good judgment; and in order that\njudgment and understanding may be distinguished he then imposes an\nartificial limitation upon the meaning in which the latter term is to be\nemployed.\n\nAs formal logic abstracts from all content, it cannot, Kant maintains,\nsupply rules for the exercise of \u201cjudgment.\u201d It is otherwise with\ntranscendental logic, which in the pure forms of sensibility possesses a\ncontent enabling it to define in an _a priori_ manner the specific cases\nto which concepts must be applicable. The _Analytic of Principles_ is\nthus able to supply \u201ca canon for the faculty of judgment, instructing it\nhow to apply to appearances the concepts of understanding which contain\nthe condition of _a priori_ rules.\u201d[1101] This will involve (1) the\ndefining of the sensuous conditions under which the _a priori_ rules may\nbe applied--the problem of the chapter on schematism; and (2) the\nformulating of the rules in their sensuous, though _a priori_,\nconcreteness--the problem of the chapter on \u201cthe system of all\nprinciples of pure understanding.\u201d\n\nSuch is Kant\u2019s own very misleading account of the purposes of these two\nchapters. There are other and sounder reasons why they should be\nintroduced. In the _Analytic of Concepts_, as we have seen,[1102] the\ntranscendental deduction only succeeds in proving that _a priori_ forms\nof unity are required for the possibility of experience. No proof is\ngiven that the various categories are just the particular forms\nrequired, and that they are one and all indispensable. This omission can\nbe made good only by a series of proofs, directed to showing, in\nreference to each separate category, its validity within experience and\nits indispensableness for the possibility of experience. These proofs\nare given in the second of the two chapters. The chapter on schematism\nis preparatory in character; it draws attention to the importance of the\ntemporal aspect of human experience, and defines the categories in the\nform in which they present themselves in an experience thus conditioned\nby _a priori_ intuition.\n\n\n\n\nCHAPTER I\n\nTHE SCHEMATISM OF PURE CONCEPTS OF UNDERSTANDING[1103]\n\n\nThe more artificial aspect of Kant\u2019s argument again appears in the\nreason which he assigns for the existence of a problem of schematism,\nnamely, that pure concepts, and the sensuous intuitions which have to be\nsubsumed under them, are completely opposite in nature. No such\nexplanation can be accepted. For if category and sensuous intuition are\nreally heterogeneous, no subsumption is possible; and if they are not\nreally heterogeneous, no such problem as Kant here refers to will exist.\nThe heterogeneity which Kant here asserts is merely that difference of\nnature which follows from the diversity of their functions. The category\nis formal and determines structure; intuition yields the content which\nis thereby organised. Accordingly, the \u201cthird thing,\u201d which Kant\npostulates as required to bring category and intuition together, is not\nproperly so describable; it is simply the two co-operating in the manner\nrequired for the possibility of experience. Kant\u2019s method of stating the\nproblem of schematism is, however, so completely misleading, that before\nwe can profitably proceed, the various strands in his highly artificial\nargument must be further disentangled. This is an ungrateful task, but\nhas at least the compensating interest of admirably illustrating the\nkind of influence which Kant\u2019s logical architectonic is constantly\nexercising upon his statement of Critical principles.\n\nThe architectonic has in this connection two very unfortunate\nconsequences. It leads Kant to describe schematism as a process of\n_subsumption_, and to speak of the transcendental schema as \u201ca _third\nthing_.\u201d Neither assertion is legitimate. Schematism, properly\nunderstood, is not a process of subsumption, but, as Kant has already\nrecognised in A 124, of synthetic interpretation. Creative synthesis,\nwhereby contents are apprehended in terms of functional relations, not\nsubsumption of particulars under universals that are homogeneous with\nthem, is what Kant must ultimately mean by the schematism of the pure\nforms of understanding. A category, that is to say, may not be viewed as\n_a predicate_ of a possible judgment, and as being applied to a subject\nindependently apprehended; its function is to articulate the judgment as\na whole. The category of substance and attribute, for instance, is the\n_form_ of the categorical judgment, and may not be equated with any one\nof its single parts.\n\nThus the criticisms which we have already passed upon Kant\u2019s mode of\nformulating the distinction between formal and transcendental\nlogic,[1104] are no less applicable to the sections now before us. The\nterminology which Kant is here employing is borrowed from the\ntraditional logic, and is out of harmony with his Critical principles.\n\nKant\u2019s description of the schema as a third thing, additional to\ncategory and intuition, and intermediate between them, is also a result\nof his misleading mode of formulating his problem. What Kant professes\nto do is to interpret the relation of the categories to the intuitional\nmaterial as analogous to that holding between a class concept and the\nparticulars which can be subsumed under it. This is implied in his use\nof the plate and circle illustration.[1105] But as the relation holding\nbetween categories and the material of sense is that of form and matter,\nstructure and content, the analogy is thoroughly misleading. As _all_\ncontent, strictly so called, falls on the side of the intuitional\nmaterial, there is no content, _i.e._ no quality or attribute, which is\ncommon to both. And thus it happens that the _inappropriateness_ of the\nanalogy which Kant is seeking to enforce is ultimately the sole ground\nwhich he is able to offer in support of his description of the schema as\n\u201ca third thing.\u201d\n\n     \u201cNow it is clear [!] that there must be a third thing, which is\n     homogeneous on the one hand with the category and on the other with\n     the appearance, and which thus makes the application of the one to\n     the other possible.\u201d[1106]\n\nOn the contrary, the true Critical teaching is that category and\nintuition, that is to say, form and content, mutually condition one\nanother, and that the so-called schema is simply a name for the latter\nas apprehended in terms of the former.\n\nBut there is a further complication. Kant, as we have already\nobserved,[1107] defines judgment as being\n\n     \u201c...the faculty of subsuming under _rules_, _i.e._ of\n     distinguishing whether something does or does not stand under a\n     given rule (_casus datae legis_).\u201d\n\nNow this view of judgment really connects with the syllogism, not with\nthe proposition.[1108] As Kant states in his _Logic_, there are\n\n     \u201c...three essential elements in all inference: (1) a universal rule\n     which is entitled the major premiss; (2) the proposition which\n     subsumes a cognition under the condition of the universal rule, and\n     which is entitled the minor premiss; and lastly, (3) the\n     conclusion, the proposition which asserts or denies of the subsumed\n     cognition the predicate of the rule.\u201d[1109]\n\nRegarded in this way, as the application of a _rule_, subsumption is\nmore broadly viewed and becomes a more appropriate analogy for the\nrelation of category to content. And obviously it is this comparison\nthat Kant has chiefly in mind in these introductory sections. For only\nwhen the subsumption is that of a particular instance under a universal\n_rule_, can the necessity of a _mediating_ condition be allowed.\n\nSuch, then, are the straits to which Kant is reduced in the endeavour to\nhold loyally to his architectonic. He has to identify the two very\ndifferent kinds of subsumption which find expression in the proposition\nand in the syllogism respectively; and when his analogy between logical\nsubsumption, thus loosely interpreted, and synthetic interpretation,\nproves inapplicable, he uses the failure of the analogy as an argument\nto prove the necessity of \u201ca third thing.\u201d On his own Critical teaching,\nas elsewhere expounded, no such third thing need be postulated. Even the\ndefinitions which he proceeds to give of the various schemata do not\nreally support this description of them.\n\nBut though Kant\u2019s method of introducing and expounding the argument of\nthis chapter is thus misleading, the contents themselves are of\nintrinsic value, and have a threefold bearing: (_a_) on the doctrine of\nproductive imagination; (_b_) on the relation holding between image and\nconcept; and (_c_) on the nature of the categories in their distinction\nfrom the pure forms of understanding.\n\n(_a_) Kant gives definite and precise expression to the two chief\ncharacteristics of the productive imagination, namely, that it deals\nwith an _a priori_ manifold of pure intuition[1110] and that it\nexercises a \u201chidden art in the depths of the human soul.\u201d[1111] Kant\u2019s\ndescription of the schema as \u201ca third thing,\u201d _at once intellectual and\nsensuous_, seems to be in large part due to the transference to it of\npredicates already applied to the faculty which is supposed to be its\nsource. The distinction between the transcendental schema and the\nparticularised image is also given as analogous to that between the pure\nand the empirical faculties of imagination. In A 141-2 = B 180-1, Kant\nspeaks of the _empirical_ faculty of _productive_ imagination, and so is\nled, to the great confusion of his exposition, though also to the\nenrichment of his teaching, to allow of empirical as well as of\ntranscendental schemata, and thus contrary to his own real position to\nrecognise schemata of such empirical objects as dog or horse--a view\nwhich empirical psychology has since adopted in its doctrine of the\nschematic image. This passage was doubtless written at the time when he\nwas inclining to the view that the empirical processes run parallel with\nthe transcendental.[1112] Kant\u2019s final view is that empirical\nimagination is always reproductive. This brings us, however, to our\nsecond main point.\n\n(_b_) Kant makes a statement which serves as a valuable corrective of\nhis looser assertions in other parts of the _Critique_.[1113] Five\npoints set after one another, thus,....., form an image of the number\nfive. The schema of the number five is, however, of very different\nnature, and must not be identified with any such image. It is\n\n     \u201c...rather _the representation of a method_ whereby a multiplicity\n     [in this case five] may be represented in an image in accordance\n     with a certain concept, than this image itself....\u201d[1114]\n\nThis becomes more evident in the case of large numbers, such as a\nthousand. The thought or schema of the number remains just as clear and\ndefinite as in the case of smaller numbers, but cannot be so adequately\nembodied and surveyed in a concrete image.\n\n     \u201cThis representation of a general procedure of imagination in\n     providing its image for a concept, I name the schema to this\n     concept.\u201d[1115]\n\nBut even in the simplest cases an image can never be completely adequate\nto the concept. The image of a triangle, for instance, is always some\nparticular triangle, and therefore represents only a part of the total\nconnotation. As the schema represents a universal rule of production in\naccordance with a concept, it resembles the concept in its incapacity to\nsubsist in an objective form. Images become possible only through and in\naccordance with schemata, but can never themselves be identified with\nthem. Schemata, therefore, and not images--such is the implied\nconclusion--form the true subject-matter of the mathematical sciences.\nImages are always particular; schemata are always universal. Images\nrepresent existences; schemata represent methods of construction.\n\nThere are three criticisms which must be passed upon this position. In\nthe first place, the selection of the triangle as an illustration tends\nto obscure the main point of Kant\u2019s argument. As there are three very\ndifferent species of triangle, the concept triangle is a class concept\nin a degree and manner which is not to be found in the concepts, say, of\nthe circle or of the number five. So that while Kant may seem to be\nchiefly insisting upon the _inadequacy_[1116] of the image to represent\nmore than a part of the connotation of the corresponding concept, his\nreal intention is to emphasise that the schema expresses the conceptual\nrule whereby, even in images that cover the whole connotation, the true\nmeaning of the image can alone be determined.\n\nSecondly, the above definition of the schema as being \u201cthe\nrepresentation of a general procedure of imagination in providing an\nimage for a concept\u201d is obviously bound up with Kant\u2019s view of it as \u201ca\nthird thing,\u201d additional to the concept, and as intermediate between it\nand the image.[1117] But as we have already found occasion to note, in\ndiscussing Kant\u2019s doctrine of the \u201cconstruction\u201d of mathematical\nconcepts,[1118] this threefold distinction is out of harmony with his\nCritical principles. It results from his retention of the traditional\nview of the concept as in all cases a mere concept, _i.e._ an abstracted\nor class concept. In defining the schema Kant is defining the true\nnature of the concept as against the false interpretation of it in the\ntraditional class-theory; he misrepresents the logic of his own\nstandpoint when he interpolates a third kind of representation\nintermediate between the concept and the image. The concept \u2018triangle,\u2019\n_as a concept_, is (to employ Kant\u2019s own not very satisfactory terms)\nthe representation of the method of constructing a certain type of\nobject; and the only other mode of representing this kind of object is\nthe image. There may, indeed, as Kant has himself suggested, be a\nspecies of image that may be entitled schematic; but if that be\nidentified with a blurred or indeterminate or merely symbolic form of\nrepresentation, it can have nothing in common with the _transcendental_\nor _conceptual_ schema, save the name.\n\nThirdly, the entire discussion of the nature of the schemata of\n\u201csensuous concepts\u201d and of their relation to the sense image, is out of\norder in this chapter; and however valuable in itself, bewilders the\nreader who very properly assumes for it a relevancy which it does not\npossess. The pure concepts of the understanding, whose schemata Kant is\nendeavouring to define, are altogether different in nature from sensuous\nrepresentations, and can never be reduced in any form or degree to an\nimage. They are wholly transcendental, representing pure syntheses\nunified through categories in accordance with the form of inner sense.\nThis, however, brings us to our last main point.\n\n(_c_) Kant\u2019s manner of employing the term category is a typical example\nof his characteristic carelessness in the use of his technical terms.\nSometimes it signifies the pure forms of understanding. But more\nfrequently it stands for what he now, for the first time, entitles\nschemata, namely, the pure conceptual forms as modified through relation\nto time. To take as examples the two chief categories of relation. The\nfirst category of relation, viewed as a form of the pure understanding,\nis the merely logical conception of that which is always a subject and\nnever a predicate. The corresponding schema is the conception of that\nwhich has permanent existence in time; it is not the logical notion of\n_subject_, but the transcendental conception of _substance_. The pure\nlogical conception of ground and consequence is similarly distinguished\nfrom the transcendental schema of cause and effect.\n\nThis contrast is of supreme importance in the Critical philosophy, and\nought therefore to have been marked by a careful distinction of terms.\nHad Kant restricted the term category to denote the pure forms, and\ninvariably employed the term schemata to signify their more concrete\ncounterparts, many ambiguities and confusions would have been prevented.\nThe table of categories, in its distinction from the table of logical\nforms, would then have been named the table of schemata, and the\ndefinitions given in this chapter would have been appended to it, as the\nproper supplement to the metaphysical deduction, completing it by a\ncareful definition of each separate schema. For what Kant usually means\nwhen he speaks of the categories _are_ the schemata; and the chapter\nbefore us therefore contains their delayed definitions.[1119] As Kant\nhas constantly been insisting, and as he again so emphatically teaches\nin this chapter, the pure forms of understanding, taken in and by\nthemselves, apart from the forms of intuition, have no relation to any\nobject, and are mere logical functions without content or determinate\nmeaning.\n\nFrom this point of view the misleading influence of Kant\u2019s architectonic\nmay again be noted. It forces him to preface his argument by\nintroductory remarks which run entirely counter to the very point which\nhe is chiefly concerned to illustrate and enforce, namely, the\ninseparability of conception and intuition in all experience and\nknowledge. He does, indeed, draw attention to the fact that the\nconditions which serve to realise the pure concepts of understanding\nalso at the same time restrict them, but it is with their empirical\nemployment that he is here chiefly concerned.\n\nCaird\u2019s[1120] mode of expounding Kant\u2019s doctrine of schematism may serve\nas an example of the misleading influence of Kant\u2019s artificial method of\nintroducing his argument. As Caird accepts Kant\u2019s initial statements at\ntheir face value, he is led to read the entire chapter in accordance\nwith them, and so to interpret it as being a virtual recantation of the\nassumptions which underlie the statement of its problem. The truer view\nwould rather seem to be that the introduction is demanded by the\nexigencies of Kant\u2019s architectonic, and therefore yields no true account\neither of the essential purpose of the chapter or of its actual\ncontents. Cohen not unjustly remarks that\n\n     \u201c...recent writers are guilty of a very strange misreading of Kant\n     when they maintain, as if in opposition to him, a thought to which\n     his doctrine of schematism gives profound expression, namely, that\n     intuition and conception do not function independently, and that\n     thought, and still more knowledge, is and must always be\n     intuitive.\u201d[1121]\n\nCohen fails, however, to draw attention to the cause of the\nmisunderstanding for which Kant must certainly share the blame.\nRiehl,[1122] while adopting a somewhat similar view to that here given,\ntraces Kant\u2019s misleading mode of stating the problem to his holding a\nfalse view of the universality of the concept. Such criticism of Kant,\nlike that passed by Caird, is in many respects justified, but the\noccasion upon which the admonition is made to follow would none the less\nseem to be ill-chosen.\n\nIt may be asked why Kant in this chapter so completely ignores space. No\nreally satisfactory answer seems to present itself. It is true that time\nis the one universal form of all intuition, of outer as well as of inner\nexperience. It is also true that, as Kant elsewhere shows, consciousness\nof time presupposes consciousness of space for its own possibility, and\nso to that extent may be regarded as including the latter form of\nconsciousness within itself. Nevertheless Kant\u2019s concentration on the\ntemporal aspect of experience is exceedingly arbitrary, and results in\ncertain unfortunate consequences. Owing to the manner in which Kant\nenvisages his problem[1123] he is bound, indeed, to lay the greater\nemphasis upon time, but that need not have involved so exclusive a\nrecognition of its field and function. Possibly Kant\u2019s very natural\npreoccupation with his new and revolutionary doctrines of inner sense\nand productive imagination has something to do with the matter.\n\nThough the definitions given of the various schemata, especially of\nthose of reality and existence, raise many difficulties, consideration\nof them must be deferred.[1124] They can be properly discussed only in\nconnection with the principles which Kant bases upon them. Only one\nfurther point calls for present remark. Kant does not give a schema for\neach of the categories. In the first two groups of pure conceptual\nforms, those of quantity and of quality, he gives a schema only for the\nthird category in each case. Number is strictly not the schema of\nquantity as such, but of _totality_. The schema of quality is a\ndefinition only of _limitation_.[1125] This departure from the demands\nof strict architectonic is made without comment or explanation of any\nkind. Kant delights to insist upon the confirmation given to his\nteaching by the fulfilment of architectonic requirements; he is for the\nmost part silent when they fail to correspond. This architectonic was a\nhobby sufficiently serious to yield him keen pleasure in its\nelaboration, but was not so vital to his main purposes as to call for\nstronger measures when shortcomings occurred.\n\nIn concluding this chapter Kant draws attention to the fact that the\nsensuous conditions which serve to realise the pure concepts also at the\nsame time restrict their meaning. Their wider meaning is, however, of\nmerely logical character.[1126] Their function, as pure concepts, lies\nsolely in establishing unity of representation; they do not therefore\nsuffice to yield knowledge of any object. Objective application \u201ccomes\nto them solely from sensibility.\u201d In these statements Kant expounds one\nof his fundamental doctrines, but in a manner which does less than\njustice to the independent value of pure thought. As he elsewhere\nteaches,[1127] it is not sense that sets limits to understanding; it is\nthe pure forms of thought that enable the mind to appreciate the limited\nand merely phenomenal character of the world experienced.\n\n\n\n\nCHAPTER II\n\nSYSTEM OF ALL PRINCIPLES OF PURE UNDERSTANDING\n\n\nThe introductory remarks to this important chapter are again dictated by\nKant\u2019s architectonic, and set its actual contents in an extremely false\nlight. Kant would seem to imply that as the _Analytic of Concepts_ has\ndetermined all the various conceptual elements constitutive of\nexperience, and has proved that they serve as predicates of possible\njudgments, it now remains to show in an _Analytic of Principles_ what _a\npriori_ synthetic judgments, or in other words what principles, can\nactually be based upon them. Though this is a quite misleading account\nof the relation holding between the two books of the _Analytic_, it has\nbeen accepted by many commentators.[1128] For several reasons it must be\nrejected. The pure forms of understanding are not predicates for\npossible judgments. They underlie judgment as a whole, expressing the\nrelation through which its total contents are organised. Thus in the\nproposition \u201ccinnabar is heavy\u201d the category of substance and attribute\nis not in any sense the predicate; it articulates the entire judgment,\ninterpreting the experienced contents in terms of the dual relation of\nsubstance and attribute. Judgment, its nature and conditions, is the\nreal problem of the misnamed _Analytic of Concepts_. As already\nindicated,[1129] the two main divisions of the _Analytic_ deal with one\nand the same problem. But while doing so, they differ in two respects.\nIn the first place, as above noted, the _Analytic of Concepts_ supplies\nno proof of the validity of particular categories, but only a quite\ngeneral demonstration that forms of unity, such as are involved in all\njudgment, are demanded for the possibility of apperception. The proofs\nof the indispensableness of _specific_ categories are first given in the\n_Analytic of Principles_. Secondly, in the _Analytic of Concepts_ the\ntemporal aspect of experience falls somewhat into the background,\nwhereas in the _Analytic of Principles_ it is emphasised.\n\nFrom these two fundamental points of difference there arises a third\ndistinguishing feature. When the categories, or rather schemata, are\nexplicitly defined, and receive individual proof, they are found to be\njust those principles that are demanded for the possibility of the\npositive sciences. This is, from Kant\u2019s point of view, no mere\ncoincidence. Scientific knowledge is possible only in so far as\nexperience is grounded on _a priori_ conditions; and the conditions of\n_sense_-experience are also the conditions of its conceptual\ninterpretation. But while the _Analytic of Concepts_ deals almost\nexclusively with ordinary experience, in the _Analytic of Principles_\nthe physical sciences receive their due share of consideration.\n\n=First and Second Sections. The Highest Principles of Analytic and\nSynthetic Judgments.=--These two sections contain nothing not already\ndeveloped earlier in the _Critique_. Though the principle of\nnon-contradiction is a merely negative test of truth, it can serve as a\nuniversal and completely adequate criterion in the case of all judgments\nthat are analytic of given concepts. The principle of synthetic\njudgments, on the other hand, is the principle whereby we are enabled to\nadvance beyond a given concept so as to attach a predicate which does\nnot stand to it in the relation either of identity or of contradiction.\nThis principle is the principle of the possibility of experience. Though\n_a priori_ synthetic judgments cannot be _logically_ demonstrated as\nfollowing from higher and more universal propositions,[1130] they are\ncapable of a _transcendental_ proof, that is, as being conditions of\nsense-experience.\n\n     \u201cThe possibility of experience is what gives objective reality to\n     all our _a priori_ knowledge.\u201d[1131] \u201cAlthough we know _a priori_\n     in synthetic judgments a great deal regarding space in general and\n     the figures which productive imagination describes in it, and can\n     obtain such judgments without actually requiring any experience;\n     yet even this knowledge would be nothing but a playing with a mere\n     figment of the brain, were it not that space has to be regarded as\n     a condition of the appearances which constitute the material for\n     outer experience....\u201d[1132]\n\nIn the first part of the last sentence, as in the page which precedes\nit, Kant would seem to be inculcating his doctrine of a pure _a priori_\nmanifold, but the latter part of the statement would not be affected by\nthe admission that space is not an independent intuition but only the\nform of outer sense.\n\n=Third Section. Systematic Representation of all the Synthetic Principles\nof Understanding.=--Kant is not concerned in this section with the\nfundamental propositions of mathematical science, since, on his view,\nthey rest upon the evidence of intuition. He claims, however, that their\nobjective validity depends upon two principles, which, though not\nthemselves mathematical in the strict sense, may conveniently be so\ndescribed from the transcendental standpoint--the principle of the\n\u201caxioms of intuition,\u201d and the principle of the \u201canticipations of\nexperience.\u201d The physicist, who takes the legitimacy of applied\nmathematics for granted, has no occasion to formulate these principles.\nThat he none the less presupposes them is shown, however, by his\nunquestioning assumption that nature conforms to the strict requirements\nof pure mathematics. And since the principles involve pure concepts, the\none embodying the schema of number, and the other the schema of quality,\nthey fall outside the scope of the _Transcendental Aesthetic_, and call\nfor a deduction similar to that of the other categories.\n\nAs already indicated, Kant\u2019s procedure is extremely arbitrary, and is\ndue to the perverting influence of his architectonic. Proof of the\nvalidity of applied mathematics has already been given in the\n_Aesthetic_[1133] of the first edition--a proof which is further\ndeveloped in the _Prolegomena_,[1134] and recast in the second edition\nso as to constitute a separate \u201ctranscendental exposition.\u201d[1135] As\nKant teaches in these passages, the objective validity of applied\nmathematics rests upon proof that space and time are the _a priori_\nforms of outer and inner sense. The new deductions of the schemata of\nnumber and quality, which he now proceeds to formulate, are quite\nunnecessary, and also are by no means conclusive in the manner of their\nproof. This, however, is more than compensated by the extremely valuable\nproofs of the schematised categories of relation which he gives in the\nsection on the _Analogies of Experience_. The section on the _Postulates\nof Empirical Experience_, which deals with the principles of modality,\nalso contains matter of very real importance.\n\nThe principles with which this chapter has to deal can thus be arranged\naccording to the fourfold division of the table of categories: (1)\n_Axioms of Intuition_, (2) _Anticipations of Perception_, (3) _Analogies\nof Experience_, (4) _Postulates of Empirical Thought_. And following the\ndistinction already drawn in the _Analytic of Concepts_,[1136] Kant\ndistinguishes between the Axioms and Anticipations on the one hand, and\nthe Analogies and Postulates on the other. The former determine the\nconditions of intuition in space and time, and may therefore be called\nmathematical and constitutive. They express what is necessarily involved\nin every intuition as such. The latter are dynamical. They are\nprinciples according to which we must think the existence of an object\nas determined in its relation to others. While, therefore, the first set\nof principles can be intuitively verified, the second set have only an\nindirect relation to the objects experienced. Whereas a relation of\ncausality can never be intuited as holding between two events, but only\nthought into them, spatial and temporal relations are direct objects of\nthe mind. Similarly, the relation of substance and attribute cannot be\nintuited; it can only be thought into what is intuited. The mathematical\nprinciples thus acquire an immediate (though, be it remembered, merely\n_de facto_) evidence; the _a priori_ certainty, equally complete, of the\ndynamical principles can be verified only through the circuitous channel\nof transcendental proof.\n\nThe composite constitution of these sections finds striking illustration\nin the duplicated account of this distinction which precedes and follows\nthe table of principles. The two accounts can hardly have been written\nin immediate succession to one another. The earlier in location[1137] is\nprobably the later in date. It would seem to rest upon some such\nuncritical distinction as that drawn in the _Prolegomena_ between\njudgments of perception and judgments of experience.[1138] The second\nand briefer account[1139] is not open to this objection.\n\nIn A 178-80 = B 220-3 Kant develops a further point of difference\nbetween the mathematical and the dynamical principles, or rather\nexplains what he means by his all too brief and consequently ambiguous\nreference in the first of the above accounts to \u201cexistence\u201d (_Dasein_).\nThe mathematical principles are _constitutive_; the dynamical are\n_regulative_. That is to say, the mathematical principles lay down the\nconditions for the generation or construction of appearances. The\ndynamical only specify rules whereby we can define the relation in which\nexistences contingently given are connected. As existence can never be\nconstructed _a priori_, we are limited to the determination of the\ninterrelations between existences all of which must be given. Thus the\nprinciple of causality enables us to predict _a priori_ that for every\nevent there must exist some antecedent cause; but only through empirical\ninvestigation can we determine which of the particular given antecedents\nmay be so described. That is to say, the principle defines conditions to\nwhich experience must conform, but does not enable us to construct it\nin advance. This distinction is inspired by the contrast between\nmathematical and physical science, and is valuable as defining the\nempirically regulative function of the _a priori_ dynamical principles;\nbut its somewhat forced character[1140] becomes apparent when we bear in\nmind Kant\u2019s previous distinction between the principles of pure\nmathematical science and the transcendental principles which justify\ntheir application to experience. Those latter principles concern\nexistence as apprehended through schematised categories, and are\nconsequently, as regards certainty and method of proof, in exactly the\nsame position as the dynamical principles. This is sufficiently evident\nfrom his own illustration of sunlight.[1141] There is as little\npossibility of \u201cconstructing\u201d its intensity as of determining _a priori_\nthe cause of an effect.\n\n\nI. THE AXIOMS OF INTUITION\n\n_All appearances are in their intuition extensive magnitudes._ Or as in\nthe second edition: _All intuitions are extensive magnitudes._\n\n\u2018Extensive\u2019 is here used in a very wide sense to include temporal as\nwell as spatial magnitude. Kant bases this principle upon the schema of\nnumber, and the proof which he propounds in its support is therefore\ndesigned to show that apprehension of an object of perception, whether\nspatial or temporal, is only possible in so far as we bring that schema\ninto play. But though this is the professed purpose of the argument,\nnumber is itself never even mentioned; and the reason for the omission\nis doubtless Kant\u2019s consciousness of the obvious objections to any such\nposition. That aspect of the argument is therefore, no doubt without\nexplicit intention, kept in the background. But even as thus given, the\nargument must have left Kant with some feeling of dissatisfaction.\nLoyalty to his architectonic scheme prevents such doubt and disquietude\nfrom finding further expression.\n\nThe argument, in its first-edition statement, starts from the\nformulation of a view of space and time directly opposed to that of the\n_Aesthetic_:[1142]\n\n     \u201cI entitle a magnitude extensive when the representation of the\n     parts makes possible, and therefore necessarily precedes, the\n     representation of the whole. I cannot represent to myself a line,\n     however small, without drawing it in thought, _i.e._ generating\n     from a point all its parts one after another, and thus for the\n     first time recording this intuition.\u201d\n\nSimilarly with even the smallest time. And as all appearances are\nintuited in space or time, every appearance, so far as intuited, is an\nextensive magnitude, that is to say, can be apprehended only through\nsuccessive generation of its parts. All appearances are \u201caggregates,\n_i.e._ manifolds of antecedently given parts.\u201d\n\nThis definition of extensive magnitude involves an assumption which Kant\nalso employs elsewhere in the _Critique_,[1143] but which he nowhere\nattempts to establish by argument; namely, that it is impossible to\napprehend a manifold save in succession. This assumption is, of course,\nentirely false (at least as applied to our empirical consciousness), as\nhas since been amply demonstrated by experimental investigation. Kant\nadopted it in the earlier subjectivist stage of his teaching, before he\nhad come to recognise that consciousness of space is involved in\nconsciousness of time. But even after he had done so, the earlier view\nstill tended to gain the upper hand whenever the doctrines of inner\nsense and of productive imagination were under consideration. For in\nregard to the transcendental activities of productive imagination, which\nare essentially synthetic, Kant continued to treat time as more\nfundamental than space. But, as already noted,[1144] a directly opposite\nview of the interrelations of space and time is expounded in passages\nadded in the second edition.\n\nThe two central paragraphs are very externally connected with the main\nargument, and are probably later interpolations.[1145] In the first of\nthese two paragraphs Kant ascribes the synthetic activity involved in\nthe \u201cgeneration of figures\u201d to the productive imagination, and maintains\nthat geometry is rendered possible by this faculty. In the other\nparagraph Kant deals with arithmetic, but makes no reference to the\nproductive imagination. Its argument is limited to the contention that\npropositions expressive of numerical relation, though synthetic, are not\nuniversal. They are not axioms, but numerical formulae. This distinction\nhas no very obvious bearing on the present argument, and serves only to\nindicate Kant\u2019s recognition that no rigid parallelism can be established\nbetween geometry and arithmetic. There are, it would seem, no\narithmetical axioms corresponding to the axioms of Euclid.[1146]\n\nThe concluding paragraph is a restatement of the argument of the\n_Aesthetic_ and of \u00a7 13, _Note_ i. of the _Prolegomena_. Appearances are\nnot things in themselves. They are conditioned by the pure intuitional\nforms, and are therefore subject to pure mathematics \u201cin all its\nprecision.\u201d Were we compelled to regard the objects of the senses as\nthings in themselves, an applied science of geometry (again taken, in\nKant\u2019s habitual manner, as typically representing the mathematical\ndisciplines) would not be possible. The only new element in the argument\nis the reference to synthesis as presupposed in all apprehension.\n\nThe additional proof with which in the second edition Kant prefaces the\nentire argument calls for no special comment. It may, however, be noted\nthat though in the argument of the first edition the need of synthesis\nin all apprehension is clearly taught, the term synthesis is not itself\nemployed except in the central and final paragraphs. In the proof given\nin the second edition both the term and what it stands for are allowed\ndue prominence.\n\n\n2. THE ANTICIPATIONS OF PERCEPTION\n\n_In all appearances sensation and the real which corresponds to it in\nthe object (realitas phaenomenon) has an intensive magnitude or degree._\nOr as in the second edition: _In all appearances the real, which is an\nobject of sensation, has intensive magnitude or degree._\n\nWe may first analyse the total section. The first paragraph[1147]\nexplains the term anticipation. The second and third paragraphs give a\nfirst proof of the principle. Paragraphs four to ten treat of continuity\nin space, time and change, and of the impossibility of empty space, and\nalso afford Kant the opportunity to develop his dynamical theory of\nmatter, and so to indicate the contribution which transcendental\nphilosophy is able to make towards a more adequate understanding of the\nprinciples of physical science. The eleventh and twelfth paragraphs,\nevidently later interpolations, give a second proof of the principle\nwhich in one important respect varies from the first proof. In the\nsecond edition a third proof akin to this second proof, but carrying it\na stage further, is added in the form of a new first paragraph.\n\nKant\u2019s reason for changing the formulation of the principle in the\nsecond edition is evidently the unsatisfactoriness of the phrase\n\u201csensation _and_ the real.\u201d[1148] The principle, properly interpreted,\napplies not, as the first edition title and also the second proof would\nlead us to expect, to sensation itself, but to its object, _realitas\nphaenomenon_. It is phenomenalist in its teaching. The emphatic term\n\u201canticipation\u201d is adopted by Kant to mark that in this principle we are\nable in _a priori_ fashion to determine something in regard to what in\nitself is purely empirical. Sensation as such, being the matter of\nexperience, can never be known _a priori_. Its quality, as being a\ncolour or a taste, depends upon factors which are for us, owing to the\nlimitations of our knowledge, wholly contingent. None the less in one\nparticular respect we can predetermine the object of all sensation, and\nso can _anticipate_ experience, even in its material aspect.\n\nThe first proof is as follows. Apprehension, so far as it takes place\nthrough a sensation, occupies only a single moment; it does not involve\nany successive synthesis proceeding from parts to the complete\nrepresentation. That which is apprehended cannot, therefore, possess\nextensive magnitude. But, as already stated in the chapter on\n_Schematism_, reality is that in appearance which corresponds to a\nsensation. It is _realitas phaenomenon_. The absence of it is negation =\n0. Now every sensation is capable of diminution; between reality in the\nappearance and negation there is a continuous series of many possible\nintermediate sensations, the difference between any two of which is\nalways smaller than the difference between the given sensation and zero.\nThat is to say, the real in appearance has intensive magnitude or\ndegree. The argument is from capability of variation in the intensity of\nsensation to existence of degree in its object or cause. For the most\npart this reality is spoken of as that which is apprehended in\nsensation, but Kant adds that if it be\n\n     \u201c...viewed as cause either of sensation or of other reality in\n     appearance, such as change, the degree of its reality ... is then\n     entitled a moment, as for instance the moment of gravity.\u201d\n\nThe obscurity of what in itself is a very simple and direct argument\nwould seem to be traceable to the lack of clearness in Kant\u2019s own mind\nas to what is to be signified by reality. The implied distinction\nbetween sensation and its object has not been clearly formulated.\nDefinitions have, indeed, been given of reality in the chapter on\n_Schematism_;[1149] but they are extremely difficult to decipher. Kant\nnever varies from the assertion that reality is \u201cthat which corresponds\nto sensation in general.\u201d Our difficulty is with the additional\nqualifications. This reality, he further declares, is\n\n     \u201c...that, the concept of which in itself points to an existence\n     [_Sein_] in time.\u201d[1150]\n\nThe words \u2018in time\u2019 would seem to show that what is referred to is\nreality _in the realm of appearance_, the _realitas phaenomenon_ of the\n_Anticipations_. But immediately below we find the following sentence:\n\n     \u201cAs time is only the form of intuition, and consequently of objects\n     as appearances, what corresponds in them to sensation is the\n     transcendental matter of all objects _as things in themselves_,\n     thinghood [_Sachheit_], reality.\u201d[1151]\n\nThe teaching of the first sentence is phenomenalist; that of the other\nis subjectivist.\n\nNow in the section on _Anticipations of Perception_ the phenomenalist\ntendencies of Kant\u2019s thought are decidedly the more prominent. The\nimplied distinction is threefold, between sensation as subjective state\npossessing intensive magnitude, spatial realities that possess both\nintensive and extensive magnitude, and the thing in itself. Objects as\nappearances are regarded as causes of sensation and as producing changes\nin one another.\n\nThe explanation of the phenomenalist character of this section is not\nfar to seek. Kant\u2019s chief purpose in it, as we shall find, is to develop\nthe dynamical theory of matter to which he had long held, and which, as\nhe was convinced, would ultimately be substituted for the mechanistic\nview to which almost all physicists then adhered. We can easily\nunderstand how in this endeavour the realist tendencies of his thinking\nshould at once come to the surface, and why he should have been\nconstrained to develop a position more precise and less ambiguous than\nthat expressed in the definitions of reality and degree given in the\nchapter on _Schematism_. With these preliminary explanations we may pass\nto Kant\u2019s second proof of his principle.\n\nA link of connection between the two proofs may be found in the reason\nwhich Kant in the first proof gives for his assertion that sensation\ncannot possess extensive magnitude--the reason, namely, that as its\napprehension takes place in a single moment, it involves no element of\nsynthesis. In his second proof Kant modifies this contention, and\nmaintains that we can abstract from the extensive magnitude of the\nappearance, and yet can recognise a synthesis as being involved.\n\n     \u201cThe real which corresponds to sensations in general, as opposed to\n     negation = 0, represents only something the very conception of\n     which contains an existence [_ein Sein_], and signifies nothing but\n     the synthesis in an empirical consciousness in general.\u201d[1152]\n\nKant adds that in a single moment we can represent to ourselves as\ninvolved in the bare sensation\n\n     \u201c...a synthesis of the uniform progression from zero to the given\n     empirical consciousness.\u201d\n\nThese statements are far from clear; but it is hardly necessary to\ncriticise them in detail. Since Kant is endeavouring to prove that a\nschema, that of reality or limitation, is involved in the apprehension\nof sensation, he is bound in consistency to maintain, in accordance with\nthe teaching of his deduction of the categories, that the application of\nthe schema demands some species of synthesis.\n\nThe third proof, added in the second edition,[1153] is somewhat more\nexplicit, and represents a further and last stage in Kant\u2019s vain\nendeavour to harmonise the teaching of this section with his general\nprinciples. In the empirical consciousness of sensation there is\n\n     \u201c...a synthesis of the different quantities involved in the\n     generation of a sensation from its beginning in pure intuition = 0\n     to its particular required magnitude.\u201d\n\nOr again, apprehension of magnitude is apprehension\n\n     \u201c...in which the empirical consciousness can in a certain time\n     increase from zero up to its given measure.\u201d\n\nHere, again, what Kant asserts as occurring in our awareness of\nsensation calls for much more rigorous demonstration. Like the argument\nof the second proof, it is not independently established; it is a mere\ncorollary to the general principles of his deduction of the categories.\n\nThus Kant\u2019s thesis, that the apprehension of sense qualities as\nintensive magnitudes presupposes a synthesis according to an _a priori_\nschema, is both obscure in statement, and unconvincing in argument; and\nsome of the assertions made, especially in reference to the occurrence\nof synthesis, would seem to be hardly less arbitrary than the connection\nwhich Kant professes to trace between logical \u201cquality,\u201d as affirmation\nor negation, and the dynamical intensity of sensuous qualities. For, as\nalready indicated,[1154] logical \u201cquality\u201d and intensive magnitude have\nnothing in common save the name.\n\nKant next proceeds to a discussion of the general problem of continuity.\nThe connection is somewhat forced. But if we overlook the artificial\nordering of the argument and are content to regard what is given as in\nthe nature of parenthetical comment, we find in the middle paragraph of\nthis section an excellent statement of his view of the nature of\ncontinuity and a very clear statement of his dynamical theory of matter.\n\nKant develops the conception of continuity (_a_) in reference to space\nand time, and (_b_) in its application to the intensity of sensations\nand of their causes.\n\n(_a_) Kant\u2019s own words require no comment:\n\n     \u201cSpace and time are _quanta continua_ because no part of them can\n     be given, save as enclosed between limits (points or moments), and\n     therefore as being itself a space or a time. Space therefore\n     consists only of spaces, time only of times. Points and moments are\n     only limits, i.e. mere positions that limit space and time. But\n     positions always presuppose the intuitions which they limit or are\n     intended to limit; and out of mere positions, viewed as\n     constituents capable of being given prior to space and time,\n     neither space nor time can be constructed. Such magnitudes may also\n     be called _flowing_, since the synthesis of productive imagination\n     involved in their production is a progression in time, and the\n     continuity of time is ordinarily denoted by the expression\n     _flowing_.\u201d[1155]\n\n(_b_) When Kant proceeds to apply the principle of continuity to\nintensive magnitude, his conclusion rests upon a somewhat different\nbasis. He argues that appearances must be continuous owing to the fact\nthat they are apprehended in space and time.[1156] So far as they are\nextended in space and enduring in time that may perhaps be true; but\nKant\u2019s assertion has a wider sweep. It implies that sensations and the\nphysical conditions of sensation, as for instance the sensation of red\nor the force of gravity, are capable of existing in every possible\ndegree between zero and any given intensity. This affords the key to his\nmethod of formulating his second and third proofs of the principle of\n_Anticipations of Perception_, which, in the form in which he interprets\nit, contains this further implication of continuity. These proofs are\ninspired by the desire to make all apprehension, even that of simple\nsensation, a temporal process, and by that indirect means to establish\nfor sensuous intensity and its objective conditions a continuity similar\nto that of space and time. The proof is, however, as we have seen,\ninconclusive. This application of continuity must be regarded as more in\nthe nature of a mere hypothesis than Kant is willing to recognise. As\nregards sensations, it would seem to have been positively disproved by\nthe results of experimental psychology.\n\nFrom his supposed proof of the continuity of all intensive magnitudes\nKant draws two further conclusions: first, that experience can never be\nmade to yield proof of the void in either space or time. For if all\nreality can exist in innumerable degrees, and if each sense has a\ndeterminate degree of receptivity, the complete absence of reality can\nnever be itself experienced. Inference to such absence is also\nimpossible for a second reason, namely, that one and the same extensive\nmagnitude may be completely occupied by an infinite number of different\nintensive degrees, indefinitely approximating to, and yet also\nindefinitely differing from, zero. Kant is here referring to the\ndynamical theory of matter which he had long held,[1157] and which he\nexpounds in opposition to the current mechanistic view.[1158] The\nmechanistic theory rests, he contends, upon an assumption purely\nmetaphysical and therefore wholly dogmatic, that the real in space has\nno internal differences, but is uniform like the empty space in which it\nexists.[1159] In accordance with this assumption physicists infer that\nall qualitative differences in our sensations must be due to merely\nquantitative differences in their material causes, and ultimately to\ndifferences in the number and distribution of the constituent parts of\nmaterial bodies. If two bodies of the same volume differ in weight or\nin inertia, the variation must be traced to differences in the amount of\nmatter, or, otherwise stated, to differences in the amount of unoccupied\nspace, in the two bodies. To this view Kant opposes his own\nhypothesis--for it is in this more modest form that it is presented in\nthese paragraphs--namely, that matter occupies space by intensity and\nnot by mere bulk, and that it may therefore be diminished indefinitely\nin degree without for that reason ceasing completely to fill the same\nextensive area. Thus an expanded force such as heat, filling space\nwithout leaving the smallest part of it empty, may be indefinitely\ndiminished in degree, and yet may still with these lesser degrees\ncontinue to occupy that space as completely as before. This may not,\nKant admits, be the true explanation of physical differences, but it at\nleast has the merit of freeing the understanding from metaphysical\npreconceptions, and of demonstrating the possibility of an alternative\nto the current view. If matter has intensity as well as extensity, and\nso can vary in quality as well as in quantity, physical science may\nperhaps be fruitfully developed on dynamical lines.\n\n\n3. THE ANALOGIES OF EXPERIENCE\n\nThe principle of the _Analogies_ is: _Experience is possible only\nthrough the representation of a necessary connection of\nperceptions._[1160]\n\nKant introduces the three analogies with the statement of an underlying\nprinciple, which corresponds to the central thesis of the transcendental\ndeduction. In the second edition this general principle is reformulated,\nand a new proof is added. These alterations do not seem, however, to be\nof any special significance. The two proofs repeat the main argument of\nthe transcendental deduction, but with special emphasis upon the\ntemporal aspect of experience. The categories of relation, as\nschematised, yield the _Analogies_, which acquire objective validity in\nso far as they render experience possible. The first proof (given in the\nsecond paragraph of the first edition) maintains that they are\nindispensable for apperception, and the second proof (that of the second\nedition) that they are indispensable for knowledge of objects. The\nreferences to time in the second proof are too condensed to be\nintelligible save in the light of the more explicit arguments given in\nsupport of the three _Analogies_.\n\nThe first paragraph in the first edition must be a later interpolation,\nas its assertion that simultaneity is a mode of time conflicts with the\nproof given of the first _Analogy_, but agrees with what must be\nregarded as a later interpolated passage introductory to that\nproof.[1161] This paragraph is also peculiar in another respect.\nHitherto Kant has traced the existence of the three analogies to the\nthree categories of relation, each of which conditions a separate\nschema. But in this paragraph he bases their threefold form on the fact\nthat time has three modes, duration, sequence,[1162] and coexistence,\nand that there is therefore a threefold problem: first, what is involved\nin consciousness of duration; secondly, what is involved in\nconsciousness of succession; and thirdly, what is involved in\nconsciousness of coexistence. This is not, however, a satisfactory mode\nof stating the matter, for it might seem to imply that the three aspects\nof time can be separately apprehended, and that each has its own\nindependent conditions. What Kant really proves is that all three\ninvolve one another. We can only be conscious of duration in contrast to\nsuccession, and of succession in contrast to the permanent, while both\ninvolve consciousness of coexistence. The three analogies thus treat of\nthree aspects of the same problem, the first connecting with the\ncategory of substance, the second with that of causality, and the third\nwith that of reciprocity.\n\nThe only point that calls for further comment[1163] concerns Kant\u2019s\nadoption of the term _Analogy_ as a title for the three principles of\n\u201crelation.\u201d The term is employed in contra-distinction to constitutive\nprinciple or axiom; and Kant points out that this usage of the term must\nbe carefully distinguished from the other or mathematical. \u201cIn\nphilosophy analogy is not the likeness of two quantitative but of two\nqualitative relations.\u201d In mathematical analogy a fourth term can be\ndiscovered from three given terms; but in an \u2018analogy of experience\u2019 we\npossess a rule that suffices only for the determination of the\n_relation_ to a term not given, never for knowledge of this term itself.\nThus if we are informed that 15 is to _x_ as 5 is to 10, the value of\n_x_ can be determined as 30. But if it be stated that a given event\nstands to an antecedent event as effect to cause, only the relation\nholding between the events can be specified, not the actual cause\nitself. The principle of causality thus serves only as a regulative\nprinciple, directing us to search for the cause of an event among its\nantecedents.\n\nRiehl has suggested a very different explanation of the term, namely, as\nsignifying that the categories of relation are employed only on the\nanalogy of the corresponding, pure logical forms.\n\n     \u201cIn so far as I know matter in terms of its empirical properties as\n     the substance of outer experiences, I do not gain knowledge of the\n     nature of matter but only of its relation to my thinking. In all\n     judgments upon outer things I employ matter as the _subject_. That\n     knowledge is therefore nothing but an _analogy_ to the conceptual\n     relation of a subject to its predicates. Matter is related to its\n     properties and effects in the realm of appearance as the subject of\n     a categorical judgment is related to its predicates. In so far as\n     an antecedent is entitled the cause of an event, we do not gain\n     knowledge of its nature but only of the analogy of the relation of\n     cause and effect with that of antecedent and consequent in a\n     hypothetical proposition; the connection of the changes is\n     analogous to the conceptual relation of ground and consequence; the\n     principle of the sufficient ground of changes is an _analogy of\n     experience_.\u201d[1164]\n\nThis explanation may at first sight seem to be supported by Kant\u2019s own\nstatement in the concluding paragraph of the section before us.\n\n     \u201cThrough these principles we are justified in combining appearances\n     only according to an analogy with the logical and general unity of\n     concepts ...\u201d[1165]\n\nThis assertion is, however, incidental to Kant\u2019s explanation that the\nanalogies are not principles of \u201ctranscendental\u201d (_i.e._ transcendent),\nbut only of empirical application--an explanation itself in turn\noccasioned by his desire to connect his present argument with the\nchapter on _Schematism_. This interpretation of the term analogy is\nprobably, therefore, of the nature of an afterthought. Having adopted\nthe term on the grounds above stated in A 179-80 = B 222, he finds in it\nan opportunity to reinforce his previous assertion of the restricting\ncharacter of the time condition through which categories are transformed\ninto schemata. The entire paragraph is probably, as Adickes remarks, a\nlater interpolation. But there are further reasons why we cannot accept\nthis passage as representing the real origin of the term analogy. It\nwould involve adoption of the subjectivist standpoint from which Riehl,\ndespite his otherwise realistic reading of Kant, interprets Kant\u2019s\nphenomenalist doctrines. For it implies that it is only in the noumenal,\nand not also in the phenomenal sphere, that substantial existences and\ngenuinely dynamical activities are to be found.[1166] It would also seem\nto imply, what is by no means Kant\u2019s invariable position, the absolute\nvalidity of the logical forms. And lastly, it would involve the priority\nof the logical to the real use of the categories, a violation of\nCritical principles of which Kant is himself occasionally guilty, but\nnever, as it would seem, in this exaggerated form.\n\nA. =First Analogy.=--_All appearances contain the permanent (substance) as\nthe object itself, and the changeable as its mere determination, i.e. as\na mode in which the object exists._ Or as in the second edition: _In all\nchange of appearances substance is permanent; its quantum in Nature\nneither increases nor diminishes._\n\nThe second paragraph[1167] is of composite character. Its first part\n(consisting of the first three sentences) and its second part give\nseparate proofs, involving assertions directly contradictory of one\nanother. The one asserts change and simultaneity to be modes of time;\nthe other denies this. They cannot, therefore, be of the same date. The\nfirst would seem to be the later; it connects with the first paragraph\nof the preceding section.\n\nIn the first edition the principle is defined as expressing the schema\nof the dual category of substance and attribute. In the second edition\nit is reformulated in much less satisfactory form, as being the\nscientific principle of the conservation (_i.e._ indestructibility) of\nmatter. This second formulation emphasises the weaker side of the\nargument of the first edition, and is largely due to the perverting\ninfluence of Kant\u2019s method of distinguishing between the _Analytic of\nConcepts_ and the _Analytic of Judgments_. It reveals Kant\u2019s growing\ntendency to contrast the two divisions of the _Analytic_, as dealing,\nthe one with ordinary experience, and the other with its scientific\nreorganisation.[1168]\n\nThe first proof in the first edition gives explicit expression to a\npresupposition underlying this entire section, namely, that all\napprehension is necessarily successive, or in other words that it is\nimpossible to apprehend a manifold save in succession.[1169] From this\nassumption it follows that if such succession is not only to occur but\nis to be apprehended as occurring, and if we are to be able to\ndistinguish between the successive order of all our apprehensions and\nthe order of coexisting independent existences, a permanent must be\nthought into the succession, that is to say, the successive experiences\nmust be interpreted into an objective order in terms of the category of\nabiding substance and changing attributes. Kant neither here nor\nelsewhere makes any attempt to explain how this position is to be\nreconciled with his doctrine that space can be intuited as well as time;\nand there is equal difficulty in reconciling it with the doctrine\ndeveloped in his second proof (in the second division of this same\nparagraph) that time itself does not change but only the appearances in\nit.\n\nAs above shown,[1170] there are two tendencies in Kant\u2019s treatment of\ntime, each of which carries with it its own set of connected\nconsequences. There is the view that consciousness of time _as a whole_\npreconditions consciousness of any part of it. This tends to recognition\nof simultaneity as a mode of time and of the simultaneous as apprehended\nin a single non-successive act of apprehension. On the other hand, there\nis the counter-view that consciousness of time is only possible through\nthe successive combination of its parts. This leads to the assertion\nthat simultaneity is not a mode of time, and that time itself cannot be\napprehended save as the result of synthesis in accordance with unifying\ncategories. Through the categories there arises consciousness of\nobjectivity, and so for the first time consciousness of a distinction\nbetween the subjective which exists invariably and exclusively in\nsuccession, and the objective which may exist either as successive or as\npermanent, and in whose existence both elements are, indeed, inseparably\ninvolved.\n\nTo turn now to Kant\u2019s second[1171] proof of the principle;[1172] it is\nas follows. All our perceptions are in time, and in time are represented\nas either coexistent or successive. Time itself cannot change,[1173] for\nonly as in it can change be represented. Time, however, cannot by\nitself be apprehended. As such, it is the mere empty form of our\nperceptions. There must be found in the objects of perception some\nabiding substrate or substance which will represent the permanence of\ntime in consciousness, and through relation to which coexistence and\nsuccession of events may be perceived. And since only in relation to\nthis substrate can time relations be apprehended, it must be altogether\nunchangeable, and may therefore[1174] be called substance. And being\nunchangeable it can neither increase nor diminish in quantity. Kant,\nwithout further argument, at once identifies this substance with matter.\n\nThis proof may be restated in briefer fashion.[1175] The consciousness\nof events in time involves the dating of them in time. But that is only\npossible in so far as we have a representation of the time in which they\nare to be dated. Time, however, not being by itself experienced, must be\nrepresented in consciousness by an abiding substrate in which all change\ntakes place, and since, as the substrate of _all_ change, it will\nnecessarily be unchangeable, it may be called substance.\n\nThe argument, in both proofs, is needlessly abstract, and as already\nremarked,[1176] the reason of this abstractness is that Kant here, as in\nthe chapter on _Schematism_, unduly ignores space, limiting his analysis\nto inner sense. He defines the schema of substance as the permanence of\nthe real in time, _i.e._ as the representation of the real which\npersists while all else changes. As the second edition of the _Critique_\nshows,[1177] Kant himself came to recognise the inadequacy of this\ndefinition, and therefore of the proof of the first _Analogy_.\nConsciousness is only possible through the representation of objects in\n_space_. Only in outer sense is a permanent given in contrast to which\nchange may be perceived. The proof ought therefore to have proceeded in\nthe following manner. Time can be conceived only as motion, and motion\nis perceivable only against a permanent background in space.\nConsciousness of time therefore involves consciousness of a permanent in\nspace. He might have added that consciousness of relative time involves\nconsciousness of change in relation to something relatively permanent,\nand that the scientific conception of all changes as taking place in a\nsingle absolute time involves the determining of change through relation\nto something absolutely permanent, this ultimate standard being found in\nthe heavenly bodies. By the permanent is not meant the immovable, but\nonly that which is uniform and unchanging in its motions. The uniform\nmotions of the heavenly bodies constitute our ultimate standard of time.\nThe degree of their uniformity is the measure of our approximation to an\nabsolute standard. A marginal note upon this _Analogy_ in Kant\u2019s private\ncopy of the _Critique_ reveals Kant\u2019s late awakened recognition of the\nnecessity of this mode of restating the argument.\n\n     \u201cHere the proof must be so developed as to apply only to substances\n     as phenomena of outer sense, and must therefore be drawn from\n     space, which with its determinations exists at all times. In space\n     all change is motion....\u201d[1178]\n\nThat the new argument of the second edition still proceeds on the same\nlines as the second argument of the first edition is probably due, as\nErdmann remarks,[1179] to Kant\u2019s unwillingness to make the extensive\nalterations which would have been called for in the chapter on\n_Schematism_ as well as in the statement of this _Analogy_.\n\nA second serious objection to Kant\u2019s treatment of the first _Analogy_\nfollows at once from the above. Kant identifies the permanent which\nrepresents time in consciousness with matter, and seeks to prove by\nmeans of this identification the principle of the conservation of\nmatter.[1180] That principle is not really capable of transcendental\nproof. It is not a presupposition of possible experience, but merely a\ngeneralisation empirically grounded. Kant is here confounding a\nparticular theory as to the manner in which the element of permanence,\nnecessary to possible experience, is realised, with the much more\ngeneral conclusion which alone can be established by transcendental\nmethods. His argument also conflicts with his own repeated assertion\nthat the notion of change, in so far as it is distinct from that of\ntemporal succession or of motion in space, is empirical, and\nconsequently falls outside the scope of transcendental enquiry. By the\nconservation of matter we mean the constancy of the weight of matter\nthroughout all changes. But the only permanent which can be postulated\nas necessary to render our actual consciousness of time possible,\nconsists of spatial objects sufficiently constant to act as a standard\nby comparison with which motions may be measured against one another.\nAnd as this first _Analogy_, properly understood, thus deals solely with\nspatial changes of bodies, the principle of the conservation of matter\nhas no real connection with it.\n\nThen thirdly, and lastly, Kant takes this first _Analogy_ as showing the\nindispensable function performed in experience by the category of\nsubstance and attribute. Substance, he argues, corresponds to the time\nin which events happen, and its attributes correspond to the changing\nevents. Just as all events are only to be conceived as happening in\ntime, so too all changes are only to be conceived as changes in an\nabiding substance. These, he would seem to hold, are simply two ways of\nmaking one and the same assertion. Now Kant may perhaps be right in\ninsisting that all change is change in, and not of, time. Unity of\nconsciousness would seem to demand consciousness of a single time in\nwhich all events happen. But this relation of time to its events does\nnot justify the same assertion being made of substance. Substance may be\nwhat corresponds to time in general, and may represent it in\nconsciousness, but we cannot for that reason say that changes are also\nonly in and not of it. To regard the changes in this way as attributes\ninhering in substance directly contradicts the view developed in the\nsecond _Analogy_. For the notion of substance is there treated as an\nimplication of the principle of causality. Substance, Kant there\ninsists, is not a bare static existence in which changes take place, but\na dynamic energy which from its very nature is in perpetual necessitated\nchange. Change is not change in, but change of, substance.\n\nEven in the passage in which Kant identifies the notion of the permanent\nin change with that of substance and attribute, he shows consciousness\nof this difficulty. We must not, he says, separate the substance from\nits accidents, treating it as a separate existence. The accidents are\nmerely the special forms of its existence. But all the same, he adds,\nwithdrawing the words which he has just uttered, such a separation of\nthe changing accidents from the abiding substance is \u201c_unavoidable,\nowing to the conditions of the logical employment of our\nunderstanding_.\u201d[1181] Kant is here so hard pressed to account for the\nuse of the category of substance and attribute in experience, and to\nexplain the contradictions to which it gives rise, that the only way he\nsees out of the difficulty is to refer the contradictions involved in\nthe category to the constitution of our understanding in its _logical_\nemployment. Yet as such employment of understanding is, according to his\nown showing, secondary to, and dependent upon, its \u201creal\u201d employment,\nthe category of substance and attribute can hardly have originated in\nthis way.\n\nWe must, then, conclude that Kant offers no sufficient deduction or\nexplanation of the category of substance and attribute, and as he does\nso nowhere else, we are driven to the further conclusion that he is\nunable to account for its use in experience, or at least to reconcile it\nin any adequate fashion with the principle of causality.\n\nB. =Second Analogy.=--_Everything that happens, i.e. begins to be,\npresupposes something on which it follows according to a rule._ Or as in\nthe second edition: _All changes take place in conformity with the law\nof the connection of cause and effect._\n\nThis section, as Kant very rightly felt, contains one of the most\nimportant and fundamental arguments of the entire _Critique_; and this\nwould seem to be the reason why he has so multiplied the proofs which he\ngives of the _Analogy_. Within the limits of the section no less than\nfive distinct proofs are to be found, and still another was added in the\nsecond edition. As Adickes[1182] argues, it is extremely unlikely that\nKant should have written five very similar proofs in immediate\nsuccession. The probability is that they are of independent origin and\nwere later combined to constitute this section; or, if we hold with\nAdickes that Kant first composed a \u201cbrief outline,\u201d we may conclude that\nhe combined the one or more proofs, which that outline contained, with\nothers of earlier or of later origin. The first to the fourth paragraphs\nof the first edition contain a first proof; the fifth to the seventh a\nsecond proof (a repetition of the first proof but in indirect form); the\neighth to the tenth a third proof (almost identical with the first); the\neleventh to the thirteenth a fourth proof (different in character from\nall the others); the fourteenth a fifth proof (probably the latest in\ntime of writing; an anticipation of the argument in the second edition).\nThe paragraph added in the second edition (the second paragraph in the\ntext of the second edition) gives a sixth and last proof.\n\nWe may first state the central argument, deferring treatment of such\nadditional points as arise in connection with Kant\u2019s varying\nformulations of it in his successive proofs. The second _Analogy_,\nthough crabbedly, diffusely, and even confusedly stated, is one of the\nfinest and most far-reaching pieces of argument in the whole _Critique_.\nIt is of special historical importance as being Kant\u2019s answer to Hume\u2019s\ndenial of the validity of the causal principle. Hume had maintained that\nwe can never be conscious of anything but mere succession. Kant in reply\nseeks to prove that consciousness of succession is only possible through\nconsciousness of a necessity that determines the order of the successive\nevents.\n\nKant, we must bear in mind, accepts much of Hume\u2019s criticism of the\ncategory of causality. The general principle that every event must have\nan antecedent cause is, Kant recognises, neither intuitively certain nor\ndemonstrable by general reasoning from more ultimate truths. It is not\nto be accounted for by analytic thought, but like all synthetic\njudgments _a priori_ can only be proved by reference to the contingent\nfact of actual experience. Secondly, Kant makes no attempt, either in\nthis _Analogy_ or elsewhere in the _Critique_, to explain the nature and\npossibility of causal connection, that is, to show how one event, the\ncause, is able to give rise to another and different event, the effect.\nWe can never by analysis of an effect discover any reason why it must\nnecessarily be preceded by a cause.[1183] Thirdly, the principle of\ncausality, as deduced by Kant and shown to be necessarily involved in\nall consciousness of time, is the quite general principle that every\nevent must have _some_ cause in what immediately precedes it. What in\neach special case the cause may be, can only be empirically discovered;\nand that any selected event is really the cause can never be absolutely\ncertain. The particular causal laws are discovered from experience, not\nby means of the general principle but only in accordance with it, and\nare therefore neither purely empirical nor wholly _a priori_. As even J.\nS. Mill teaches, the general principle is assumed in every inference to\na causal law, and save by thus assuming the general principle the\nparticular inference to causal connection cannot be proved. But at the\nsame time, since the proof of causal connection depends upon\nsatisfaction of those empirical tests which Mill formulates in his\ninductive methods, such special causal laws can be gathered only from\nexperience.\n\nThe starting-point of Kant\u2019s analysis is our consciousness of an\n_objective_ order in time. This is for Kant a legitimate starting-point\nsince he has proved in the _Transcendental Deduction_ that only through\nconsciousness of the objective is consciousness of the subjective in any\nform possible. The independent argument by which it is here supported is\nmerely a particular application of the general principle of that\ndeduction. When we apprehend any very large object, such as a house,\nthough we do so by successively perceiving the different parts of it, we\nnever think of regarding these successive perceptions as representing\nanything successive in the house. On the other hand, when we apprehend\nsuccessive events in time, such as the successive positions of a ship\nsailing down stream, we do regard the succession of our experiences as\nrepresenting objective succession in what is apprehended. Kant therefore\nfeels justified in taking as fact, that we have the power of\ndistinguishing between subjective and objective succession, _i.e._\nbetween sequences which are determined by the order of our attentive\nexperience and sequences which are given as such. It is this fact which\naffords Kant a precise method of formulating the problem of the second\n_Analogy_, viz. _how consciousness of objective change, as distinguished\nfrom subjective succession, is possible_?\n\nSchopenhauer, owing to the prominence in his system of the principle of\nsufficient reason, has commented upon this second _Analogy_ in\nconsiderable detail;[1184] and we may here employ one of his chief\ncriticisms to define more precisely the general intention of Kant\u2019s\nargument. The succession in our experiences of the parts of a house and\nof the positions of a ship is, Schopenhauer maintains, in both cases of\ngenuinely objective character. In both instances the changes are due to\nthe position of two bodies relatively to one another. In the first\nexample one of these bodies is the body of the observer, or rather one\nof his bodily organs, namely the eye, and the other is the house, in\nrelation to the parts of which the position of the eye is successively\naltered. In the second example the ship changes its position relatively\nto the stream. The motion of the eye from roof to cellar is one event;\nits motion from cellar to roof is a second event; and both are events of\nthe same nature as the sailing of the ship. Had we the same power of\ndragging the ship upstream that we have of moving the eye in a direction\nopposite to that of its first movement, the positions of the ship could\nbe reversed in a manner exactly analogous to our reversal of the\nperceptions of the house.\n\nThis criticism is a typical illustration of Schopenhauer\u2019s entire\nfailure to comprehend the central thesis of Kant\u2019s Critical\nidealism.[1185] The _Analytic_, so far as the main argument of its\nobjective deduction is concerned, was to him a closed book; and as this\nsecond analogy is little else than a special application of the results\nof the deduction, he was equally at a loss in its interpretation. Kant\nwas himself, of course, in large part responsible for the\nmisunderstanding. The distinction which would seem to be implied by\nKant\u2019s language between sequence that is objective and sequence that is\n_merely_ subjective is completely inconsistent with Critical\nprinciples,[1186] and is as thoroughly misleading as that other\ndistinction which he so frequently employs between the _a priori_ and\nthe merely empirical. Schopenhauer, however, regarded these distinctions\nas valid, and accordingly applies them in the interpretation of Kant\u2019s\nmethod of argument. If inner and outer experience are to be contrasted\nas two kinds of experience, there is, as Schopenhauer rightly insists,\nno sufficient ground for regarding changes due to movements of the eye\nas being subjective and those that are due to movements of a ship as\nbeing objective. That is not, however, Kant\u2019s intention in the\nemployment of these illustrations. He uses them only to make clear the\nfairly obvious fact that while in certain cases the order of our\nperceptions is subjectively initiated, in other cases we apprehend the\nsubjective order of our experiences as corresponding to, and explicable\nonly through, the objective sequence of events. In holding to this\ndistinction Kant is not concerned to deny that even in the order which\nis determined by the subject\u2019s purposes or caprice objective factors are\nlikewise involved. The fact that the foundations of a house support its\nroof, and will therefore determine what it is that we shall apprehend\nwhen we turn the eye upwards, does not render the order of our\napprehensions any the less subjective in character. But that this order\nis _purely_ subjective, Kant could never have asserted. His Critical\nprinciples definitely commit him to the view that even sensations and\ndesires are integral parts of the unitary system of natural law. Kant,\nas we shall find, is maintaining that some such distinction between\nsubjective and objective sequence as is illustrated in the above\ncontrasted instances must be present from the very start of our\nexperience--must, indeed, be constitutive of experience as such. Out of\na consciousness of the purely subjective the notion of the objective can\nnever arise.[1187] Or otherwise stated, consciousness of a time order,\neven though subjective, must ultimately involve the application of some\nnon-subjective standard.\n\n     \u201cI shall be obliged ... to derive the subjective sequence of\n     apprehension from the objective sequence of appearances, because\n     otherwise the former is entirely undetermined, and does not\n     distinguish any one appearance from any other.\u201d[1188]\n\nWe interpret the subjective order in terms of an objective system;\nconsciousness of the latter is the necessary presupposition of all\nawareness. It is as necessary to the interpretation of what is\napprehended through the rotating eyeballs as to the apprehension of a\nmoving ship. So far from refusing to recognise that the subjective order\nof our experiences is objectively conditioned, Kant is prepared to\nadvance to the further assertion that it is only apprehensible when so\nconceived.\n\nIn the third _Analogy_ Kant proceeds to the connected problem, how we\ncan apprehend the parts of a house as simultaneous notwithstanding the\nsequent relation of our perceptions of them, and what justification we\nhave for thus interpreting the subjectively sequent experiences as\nrepresenting objective coexistence. Just as Kant in this second\n_Analogy_ does not argue that irreversibility is by itself proof of\ncausal relation, but only that the consciousness of such irreversibility\ndemands the employment of the conception of causality, so in the third\n_Analogy_ he does not attempt to reduce the consciousness of coexistence\nto the consciousness of reversibility, but to prove that only through\nthe application of the conception of reciprocity can the reversibility\nbe properly interpreted. In each case the category conditions the\nempirical consciousness; the latter is an apprehension of determinate\norder only in so far as it presupposes the category. Though Kant\u2019s\ntreatment of the third _Analogy_ has less historical importance, and\nperhaps less intrinsic interest, than the proof of the second _Analogy_,\nit is even more significant of the kind of position which he is\nendeavouring to establish, and I may therefore forewarn the reader that\nhe must not spare himself the labour of mastering its difficult, and\nsomewhat illusive, argument. The doctrines which it expounds at once\nreinforce and extend the results of the second _Analogy_, while the\nfurther difficulties which it brings to view, but which it is not itself\ncapable of meeting, indicate that the problems of the _Analytic_ call\nfor reconsideration in the light of certain wider issues first broached\nin the _Dialectic_.\n\nWe may now return to Kant\u2019s main argument. His problem, as we have\nfound, is how consciousness of objective change, as distinguished from\nsubjective succession, is possible. The problem, being formulated in\nthis particular way, demands, Kant felt, careful definition of what is\nmeant by the term \u2018objective,\u2019 upon which so much depends. To apply the\nillustration above used, the house as apprehended is not a thing in\nitself but only an appearance to the mind. What, then, do we mean by the\nhouse, as distinguished from our subjective representations of it, when\nthat house is nothing but a complex (_Inbegriff_) of\nrepresentations?[1189] The question and Kant\u2019s answer to it are stated\nin subjectivist fashion, in terms of his earlier doctrine of the\ntranscendental object. To contrast an object with the representations\nthrough which we apprehend it, is only possible if these representations\nstand under a rule which renders necessary their combination in some one\nparticular way, and so distinguishes this one particular mode of\nrepresentation as the only true mode from all others. The origin,\ntherefore, of our distinction between the subjectively successive and\nthe succession which is also objective must be due in the one case to\nthe presence of a rule compelling us to combine the events in some\nparticular successive order, and in the other to the absence of such a\nrule. Our apprehension of the house, for instance, may proceed in any\norder, from the roof downwards or _vice versa_, and as the order may\nalways be reversed there is no compulsion upon the mind to regard the\norder of its apprehension as representing objective sequence. But since\nin our apprehension of an event B in time, the apprehension of B follows\nupon the apprehension of a previous event A, and we cannot reverse the\norder, the mind is compelled to view the order of succession, in terms\nof the category of causality, as necessitated, and therefore as\nobjective. The order is a necessary order not in the sense that A must\nalways precede B, that A is the cause of B, but that the order, if we\nare to apprehend it correctly, must in this particular case be conceived\nas necessary. The succession, that is, need not be conceived as a causal\none, but in order to be conceived as objective succession it must be\nconceived as rendered necessary by connections that are causal.\n\nHaving, in this general fashion, shown the bearing of his previous\nanalysis of objective experience upon the problem in hand, Kant proceeds\nto develop from it his proof of the special principle of causality. The\nschema of causality is necessary succession in time, and it is through\nthis, its time aspect, that Kant approaches the principle. It has to do\nwith the special case of _change_. To be conscious of change we must be\nconscious of an _event_, that is, of something as happening at a\nparticular point in time. The change, in other words, requires to be\ndated, and as we are not conscious of time in general, it must be dated\nby reference to other events, and obviously in this case in relation to\nthe preceding events, in contrast to which it is apprehended _as\nchange_. But according to the results of our analysis of what\nconstitutes objective experience, it can be fixed in its position in\nobjective time only if it be conceived as related to the preceding\nevents according to a necessary law; and the law of necessary connection\nin time is the law of causality. In order, then, that something which\nhas taken place may be apprehended as having occurred, that is, as being\nan objective change, it must be apprehended as necessarily following\nupon that which immediately precedes it in time, _i.e._ as causally\nnecessary.\n\nThe principle of causality thus conditions consciousness of objective\nsuccession, and Hume, in asserting that we are conscious of the\nsuccession of _events_, therefore admits all that need be assumed in\norder to prove the principle. The reason why Hume failed to recognise\nthis, is that he ignored the distinction between consciousness of the\nsubjective order of our apprehensions and consciousness of the objective\nsequence of events. Yet that is a distinction upon which his own\nposition rested. For he teaches that determination of causal laws,\nsufficiently certain to serve the purposes alike of practical life and\nof natural science, can be obtained through observation of those\nsequences which remain constant. Such is also the position of all\nempiricists. They hold that causal relation is discovered by comparison\nof _given_ sequences. Kant\u2019s contention is that the apprehension of\nchange as change, and therefore ultimately the apprehension even of an\narbitrarily determined order of subjective succession,[1190]\npresupposes, and is only possible through, an application of the\ncategory of causality. The primary function of the understanding does\nnot consist in the clarification of our representation of an event, but\nin making such representation possible at all.[1191] The primary field\nof exercise for the understanding lies not in the realm of reflective\ncomparison, but in the more fundamental sphere of creative\nsynthesis.[1192] In determining the nature of the given it predetermines\nthe principles to which all reflection upon the given must conform. The\ndiscursive activities of scientific reflection are secondary to, and\nconditioned by, the transcendental processes which generate the\nexperience of ordinary consciousness. Only an experience which conforms\nto the causal principle can serve as foundation either for the empirical\njudgments of sense experience, or for that ever-increasing body of\nscientific knowledge into which their content is progressively\ntranslated. The principle of causality is applicable to everything\nexperienced, for the sufficient reason that experience is itself\npossible only in terms of it. This conclusion finds its most emphatic\nand adequate statement in the _Methodology_.\n\n     \u201c...through concepts of understanding pure reason establishes\n     secure principles, not however directly from concepts, but always\n     only indirectly through relation of these concepts to something\n     altogether contingent, namely, _possible experience_. For when such\n     experience (_i.e._ something as object of possible experience) is\n     presupposed, the principles are apodictically certain, though by\n     themselves (directly) _a priori_ they cannot even be recognised at\n     all. Thus no one can acquire insight into the proposition that\n     everything which happens has its cause, merely from the concepts\n     involved. It is not, therefore, a dogma, although from another\n     point of view, namely, from that of the sole field of its possible\n     employment, _i.e._ experience, it can be proved with complete\n     apodictic certainty. But though it needs proof, it should be\n     entitled a _principle_, not a _theorem_, because _it has the\n     peculiar character that it makes possible the very experience which\n     is its own ground of proof, and in this experience must always\n     itself be presupposed_.\u201d[1193]\n\nBefore making further comment upon Kant\u2019s central argument, it is\nadvisable to consider the varying statements which Kant has given of it.\nWe may take his successive proofs in the order in which they occur in\nthe first edition.\n\n=First Proof.=[1194]--The argument is developed in terms of Kant\u2019s early\ndoctrine of the transcendental object. The only points specially\ncharacteristic of the statement here given of that doctrine consist\n(_a_) in the emphasis with which it is asserted that representations can\nbe experienced only in succession to one another, and that they can\nnever stand in the relation of coexistence,[1195] and (_b_) in the\nalmost complete ignoring of the transcendental object as source or\nground of the rule in terms of which the successive representations are\norganised. (_a_) This is a point common to the arguments of all three\n_Analogies_. In the first and third the problem is how, from\nrepresentations merely successive, permanence and coexistence can be\ndetermined. In the second _Analogy_ the problem is how from\nrepresentations invariably successive a distinction can be drawn between\nthe subjectively determined order of our apprehensions and the objective\nsequence of events. Or in other words: how under such conditions we can\nrecognise an order as given, and so as prescribing the order in which it\nmust be apprehended. Or to state the same point in still another manner:\nhow we can distinguish between an arbitrary or reversible order and an\nimposed or fixed order, and so come to apprehend the subjective order of\nour apprehensions as in certain cases controlled by, and explicable only\nthrough, the objective sequence of events.[1196]\n\n(_b_) The reason why the transcendental object, as source of the\ndeterminate and prescribed order of the given events, falls into the\nbackground in this passage is that Kant is concerned only with the\ngeneral principle or category by means of which the order is apprehended\nas necessary. That principle has a subjective origin even though the\nparticular sequences of concrete events have by means of that concept to\nbe conceived as inexorably determined by their noumenal\nconditions.[1197] The principle accounts for the _comprehension_ of the\norder as objective, and that is the only point with which Kant is here\nimmediately concerned. That the assertion of the subjective origin of\nthe category is not inconsistent with recognition of the imposed order\nof the given has already been shown above.[1198] Kant\u2019s own\nillustration, in this section, of the ship sailing down stream shows\nthat he was prepared to assume without question that they are\ncompatible. His argument is, however, obscure, owing to his failure to\ndistinguish between the two senses in which the term \u2018rule\u2019 may be\nemployed. The term may signify either the universal and merely formal\nprinciple that every event must have a cause, or it may be used to\ndenote the fixed order in which concrete events are presented to\nsense-perception. The latter order need not represent a series the\nmembers of which are causally connected _with one another_, but only one\nthat is due to causal necessities. Thus the successive positions of a\nship sailing down stream are not interrelated as cause and effect, and\nyet in order to be apprehended as objectively successive must be\nconceived as causally conditioned. The term \u2018rule\u2019 has very different\nmeanings in the two cases. \u2018Rule\u2019 in the first sense is of subjective\norigin. It is _formal_, and can never be given. It is read into the\ngiven. \u2018Rule\u2019 in the second sense is given merely, and being due to\nnoumenal conditions constitutes the _material_ element in natural\nscience, the empirical content of some particular causal law. Owing to\nKant\u2019s failure explicitly to distinguish between these two very\ndifferent connotations of the term, such a sentence as the following is\nambiguous: \u201cThat in appearance which contains the condition of this\nnecessary rule of apprehension is the object.\u201d Kant may mean that the\nprescribed order of the concrete events is due to the transcendental\nobject; but in that case it is not given as _necessary_. Necessity, as\nhe constantly insists, is the one thing that can never be given. The\nsentence is also misleading through its use of the term \u2018appearance.\u2019\nThat term has no legitimate place in a passage inspired by the doctrine\nof the transcendental object; there can be no such middle term between\nsubjective representations and the thing in itself. As Kant himself\nstates,[1199] appearance defined in terms of that doctrine is \u201cnothing\nsave a complex of representations.\u201d\n\nThere is a very essential difference in the view which Kant takes of the\ncausal relation according as he is proceeding upon subjectivist or upon\nphenomenalist lines. From the one point of view appearances are\nrepresentations merely, and accordingly are entirely devoid of causal\nefficacy. They are not causes and effects of one another. They have not\nthe independence or self-persistence necessary for the exercise of\ndynamical energy or even for the reception of modifications. Being\n\u201cstates of the identical self,\u201d all causal relation, dynamically\nconceived, must lie solely in their noumenal conditions. Causality\nreduces to the thought of necessitated (not necessitating) sequence. It\nis, as Kant has suggested in A 181 = B 224, a mere \u2018analogy\u2019 in terms of\nwhich we apply the logical relation of ground and consequence[1200] to\nthe interpretation of our subjective representations, and so view them\nas grounded not in one another but exclusively in the thing in itself.\nCausality in the strict sense, _i.e._ dynamical agency, can be looked\nfor only in the noumenal sphere.\n\nCaird, while adopting this explanation of the term \u2018analogy,\u2019[1201] is,\nas might be expected from his Hegelian standpoint, extremely indefinite\nand non-committal as to whether or not empirical objects can be genuine\ncauses. Riehl, notwithstanding his professedly realistic interpretation\nof Kant, adopts the above subjectivist view of natural causation. So\nalso do Benno Erdmann and Paulsen. The latter[1202] speaks with no\nuncertain voice.\n\n     \u201cCausality in the phenomenal world signifies for Kant, as for Hume,\n     nothing but regularity in the sequence of phenomena. Real causal\n     efficiency cannot of course occur here, for phenomena are\n     ideational products. As such they can no more produce an effect\n     than concepts can.\u201d\n\nThe corresponding phenomenalist view of the causal relation receives no\nquite definite formulation either in this section or elsewhere in the\n_Critique_, but may be gathered from the general trend of Kant\u2019s\nphenomenalist teaching.[1203] It is somewhat as follows. The term\n\u2018analogy\u2019 is viewed as having a meaning very different from that above\nsuggested. The causal relation is not a mere analogy from the logical\nrelation of ground and consequence; it is the representation of\ngenuinely dynamical activities in the objects apprehended. Those objects\nare not mere states of the self, subjective representations. They are\npart of an independent order which in the form known to us is a\nphenomenalist transcript of a deeper reality. If the causal relation is\nthe analogy of anything distinguishable from itself, it is an analogon\nor interpretation of dynamical powers exercised by things in\nthemselves,[1204] not of the merely logical relation between premisses\nand conclusion. The objects of representation may exercise powers which\nrepresentations as such can never be conceived as possessing. Between\nthe individual\u2019s subjective states and things in themselves stands the\nphenomenal world of the natural sciences. Its function, whether as\ndirectly experienced through sense-perception or as conceptually\nreconstructed through scientific hypothesis, is to stand as the\nrepresentative in human consciousness of that noumenal realm in which\nall existence is ultimately rooted. The causal interactions of material\nbodies in space are as essentially constitutive of those bodies as are\nany of their quantitative properties. Causal relation, even in the\nphenomenal sphere, must not be identified with mere conformity to law.\nThe true and complete purpose of the natural sciences is not to be found\nin the Berkeleian or sceptical ideal of simplification, but in the older\nand sounder conception of causal explanation. That, at least, is the\nview which Kant invariably defends whenever he has occasion to discuss\nthe principles of physical science.\n\n=Second Proof.=[1205]--The argument of the first proof is here developed\nin indirect fashion. In the _absence_ of any rule prescribing necessary\nsequence, no distinction can be made between subjective and objective\nsuccession. The justification for such a rule lies therefore, not in an\ninductive inference from repeated experience, but in its necessity for\nthe possibility of experience. It is an expression of the synthetic\nunity in which experience consists.\n\n=Third Proof.=[1206]--This is for the most part merely a restatement of\nthe first proof. It differs from it in making rather more explicit that\nthe objective reference involved in the notion of the transcendental\nobject is one that carries the mind beyond all representations to the\nthought of something which determines their order according to a rule.\nOtherwise the ambiguities of the terms employed are identical with those\nof the first proof. Its concluding paragraph, however, is a much clearer\nstatement of the difficult argument of A 192-3 = B 238-9.\n\n=Fourth Proof.=[1207]--This proof differs from all the others. It argues\nfrom the characteristics of pure time to the properties necessary to the\nempirical representation of the time-series. As time cannot be\nexperienced in and by itself, all its essential characteristics must be\ncapable of being represented in terms of appearance. \u201cOnly in\nappearances can we empirically recognise continuity in the connection of\ntimes.\u201d The primary function of the understanding is to make such\nrecognition possible, and it does so by \u201ctransferring the time order to\nthe appearances and their existence.\u201d It is a necessary law of time that\nwe can only advance to the succeeding through the preceding. Each moment\nof time is the indispensable condition of the existence of that which\nfollows it. We can pass to the year 1915 only by way of the preceding\nyear 1914. And since, as just noted, time is not cognisable by itself\nbut only as the form of our perceptions, this law must be applicable to\nthem. We can only be conscious of all times as successively conditioning\none another in one single time, and that means in one single _objective_\ntime, if we are conscious of all the phenomena perceived as conditioning\none another in their order in time.\n\nIt is somewhat difficult to understand how Kant came to formulate the\nargument in this form. The explanation may perhaps be found in his\npreoccupation[1208] with the doctrine of a transcendental activity of\nthe productive imagination and with the connected doctrine of a pure _a\npriori_ manifold. For this proof would seem to rest upon the assumption\nthat the characteristics of time are known purely _a priori_ and\ntherefore with complete certainty, independently of sense experience.\nThe unusual and somewhat scholastic character of the proof also appears\nin Kant\u2019s substitution of the principle of sufficient reason for the\nprinciple of causality. But despite the artificial character of the\nstandpoint, the argument serves to bring prominently forward Kant\u2019s\ncentral thesis, viz. that the principle of causality is presupposed in\nall consciousness of time, even of the subjectively successive. Also,\nby emphasising that time in and by itself can never be \u201can object of\nperception,\u201d and that the relating of appearances to \u201cabsolute time\u201d is\npossible only through the determining of them in their relations to one\nanother, it supplies the data for correction of its own starting-point.\n\n=Fifth Proof.=[1209]--This proof is probably later than the preceding\nproofs. Though its essential content coincides with that of the opening\nproof, its formulation would seem to be a first attempt at statement of\nthe sixth proof, _i.e._ of the argument which Kant added in the second\nedition. Adickes considers this proof to be earlier in date than the\nfirst four proofs, but the reason which he assigns for so regarding it,\nviz. that Kant here postulates a synthesis of the imagination\nindependent of the categories as preceding a synthesis of apprehension\nin terms of the categories, seems to be based upon a much too literal\nreading of Kant\u2019s loose mode of statement. The argument rather appears\nto be, as in the sixth proof, that synthesis of the imagination may be\neither subjective or objective; and the term \u201capprehension\u201d would seem\nto be used as signifying that the manifold synthesised is _given_ to the\nimagination through actual sense experience, and that as thus given it\nhas a determinate order of its own. The argument concludes with the\nstatement (more definite than any to be found in the preceding\narguments), that the proof of the principle of causality consists in its\nindispensableness as a condition of all empirical judgments, and so of\nexperience as such. As a ground of the possibility of experience it must\nbe valid of all the objects of experience.\n\n=Sixth Proof.=[1210]--The argument of the fifth proof is here more clearly\nstated. All synthesis is due to \u201cthe faculty of imagination which\ndetermines inner sense in respect of the time relation.\u201d Such synthesis\nmay, however, yield the consciousness either of subjective succession or\nof succession \u201cin the object.\u201d In the latter form it presupposes the\nemployment of a pure concept of the understanding, that of the relation\nof cause and effect. And the conclusion reached is again that only so is\nempirical knowledge possible. This mode of stating the argument is far\nfrom satisfactory. It tends to obscure Kant\u2019s central thesis, that only\nthrough consciousness of an objective order is consciousness of\nsubjective sequence possible, and that the principle of causality is\ntherefore a conditioning factor of all consciousness. The misleading\ndistinction drawn in the _Prolegomena_ between judgments of perception\nand judgments of experience also crops out in Kant\u2019s use of the phrase\n\u201cmere perception.\u201d[1211]\n\n       *       *       *       *       *\n\nWe may again return to Kant\u2019s central argument. For we have still to\nconsider certain objections to which it may seem to lie open, and also\nto comment upon Kant\u2019s further explanations in the remaining paragraphs\nof the section.[1212] Kant\u2019s imperfect statement of his position has\nsuggested to Hutchison Stirling and others a problem which is largely\nartificial, namely, how the mind is enabled to recognise the proper\noccasions upon which to apply the category of causality. On the one hand\nsequence as such cannot be the criterion, since many sequences are not\ncausal, and on the other hand the absence of sequence does not appear to\ndebar its application, since cause and effect would frequently seem to\nbe co-existent. This difficulty arises from failure to appreciate the\ncentral thesis upon which Kant\u2019s proof of the principle of causality\nultimately rests. Kant\u2019s diffuse and varying mode of statement may\nconceal but never conflicts with that thesis, which consists in the\ncontention that the category of causality is a necessary and invariable\nfactor in all consciousness. Nothing can be apprehended save in terms of\nit.[1213] It prescribes an interpretation which the mind has no option\nsave to apply in the consciousness of each and every event, of the\ncoexistent no less than of the sequent. Whether two changes are\ncoexistent or are successive, each must be conceived as possessing an\nantecedent cause. The only difference is that in the case of sequent\nevents one of them (_i.e._ the antecedent change) may, upon empirical\ninvestigation, be found to be itself the cause of the second and\nsubsequent event, whereas with coexistent events this can never be\npossible. As the principle of causality is that every event must have an\nantecedent cause, it follows that where there is no sequence there can\nbe no causation. But when Kant states that sequence is \u201cthe sole\nempirical criterion\u201d[1214] of the causal relation, he does less than\njustice to the position he is defending. The empirical criteria are\nmanifold in number, and are such as John Stuart Mill has attempted to\nformulate in his inductive methods.\n\nSchopenhauer has objected[1215] that Kant\u2019s argument proves too much,\nsince it would involve that all objective sequences, such as that of\nnight and day or of the notes in a piece of music, are themselves causal\nsequences. This criticism has been replied to by Stadler[1216] in the\nfollowing terms:\n\n     \u201cWhen Schopenhauer adduces the sequence of musical notes or of day\n     and night, as objective sequences which can be known without the\n     causal law, we need only meet him with the question, Where in these\n     cases is the substance that changes? So soon as he is forced to put\n     his objection into the form required to bring it into relation to\n     the question of the possibility of knowledge, his error becomes\n     obvious. His instances must then be expressed thus:--The instrument\n     passes from one state of sound into another; the earth changes from\n     the measure of enlightenment which makes day, to that which makes\n     night. Of such changes no one will say that they are not referred\n     to a cause. And we may quote in this reference the appropriate\n     saying of Kant himself, \u2018Days are, as it were, the children of\n     Time, since the following day with that which it contains is the\n     product of the previous day.\u2019\u201d\n\nNight and day, in so far as they are sequent events, must be conceived\nin terms of causality, not in the sense that night causes day, but as\nbeing determined by causes that account not only for each separately,\nbut also for the alternating sequence of the one upon the other. Such\ncauses are found by the astronomer to lie in the changing positions of\nthe earth relatively to the sun.\n\nSchopenhauer adds a further objection of a more subtle nature, which has\nagain been excellently stated and answered by Stadler:\n\n     \u201cSchopenhauer points out that what we call chance is just a\n     sequence of events which do not stand in causal connexion. \u2018I come\n     out of the house and a tile falls from the roof which strikes me;\n     in such a case there is no causal connexion between the falling of\n     the tile and my coming out of the house, yet the succession of\n     these two events is objectively determined in my apprehension of\n     them.\u2019 How have we to criticise this case from the transcendental\n     point of view? We know that successions become necessary, _i.e._\n     objective, for our consciousness, when we regard them as changes of\n     a substance which are determined by a cause. But it is shown here\n     that there are successions in which the single members are changes\n     of different substances. If substance S changes its state A into B\n     on account of the cause X, and substance S\u00b4 changes its state A\u00b4\n     into B\u00b4 on account of the cause X\u00b4, and if I call the first change\n     V and the second V\u00b4, the question arises how the objectivity of the\n     succession V V\u00b4 is related to the law of causality. Sequences such\n     as V V\u00b4 are very frequent, and our consciousness of the objectivity\n     is certain. Do we owe this consciousness to the same rule as holds\n     good in other cases? Certainly. The distinction is not qualitative,\n     but rests only on the greater complication of the change in\n     question. The sequence V V\u00b4 can become objective only if I think it\n     as a necessary connexion. It must be so determined that V can only\n     follow V\u00b4 in \u2018consciousness in general\u2019; there must be a U, the\n     introduction of which is the cause that V\u00b4 follows V. To be\n     convinced of this, I do not need actually to know U. I know that on\n     every occasion U causes the succession V V\u00b4. Of course, this\n     presupposes that all data of the states considered, A and A\u00b4,\n     remain identical. But whether these data are very simple or\n     endlessly complex, whether they are likely to combine to the given\n     result frequently or seldom, is indifferent for the objectifying of\n     the event; it is not the perception of U, but the presupposition of\n     it, which makes the change necessary and so objective for\n     us.\u201d[1217]\n\nTo turn now to the other difficulty which Kant himself raises in A 202-3\n= B 247-8, viz. that cause and effect would frequently seem to be\ncoexistent, and the \u201csole empirical criterion\u201d to be therefore absent.\nIt may from this point of view be maintained that the great majority of\ncauses occur simultaneously with their effects, and that such time\nsequence as occurs is due solely to the fact that the cause cannot\nexecute itself in one single instant. Kant has little difficulty in\ndisposing of this objection. Causality concerns only the _order_, not\nthe _lapse_, of time; and the sequence relation must remain even though\nthere is no interval between the two events. If a leaden ball lies upon\na cushion it makes a depression in it. The ball and the depression are\ncoexistent. None the less, when viewed in their dynamical relation, the\nlatter must be regarded as sequent upon the former. If the leaden ball\nis placed upon a smooth cushion a hollow is at once made, but if a\nhollow exists in a cushion a ball need not appear. In other words, the\ncriteria for the determination of specific causal relations are neither\nthe presence nor the absence of sequence, but are empirical\nconsiderations verifiable only upon special investigation.[1218] The\nobserver is called upon to disentangle the complicated web of given\nappearances under the guidance of the quite general and formal principle\nthat every event is due to some antecedent cause. He must do so as best\nhe can through the application of his acquired insight, and, when\nnecessary, by means of the requisite experimental variation of\nconditions.\n\nIn the two following paragraphs (A 204-5 = B 249-51) Kant raises points\nwhich he later discussed more fully in the _Metaphysical First\nPrinciples of Natural Science_.[1219] As adequate explanation of the\nargument would be a very lengthy matter, and not of any very real\nimportance for the understanding of the general Critical position, we\nmay omit all treatment of it. In the sections of the _Metaphysical First\nPrinciples_ just cited, the reader will find the necessary comment and\nexplanation. Such bearing as these two paragraphs have upon Kant\u2019s view\nof the nature of the causal relation has been noted above.[1220]\n\nIn the section on _Anticipations of Perception_[1221] Kant has stated\nthat the principle of the continuity of change involves empirical\nfactors, and therefore falls outside the limits of transcendental\nphilosophy. To this more correct attitude Kant, unfortunately, did not\nhold. In A 207-11 = B 252-6 he professes to establish the principle in\n_a priori_ transcendental fashion as a necessary consequence of the\nnature of time. This proof is indeed thrice repeated with unessential\nvariations, thereby clearly showing that these paragraphs also are of\ncomposite origin. The argument in all three cases consists in inferring\nfrom the continuity of time the continuity of all changes in time. As\nthe parts of time are themselves times, of which no one is the smallest,\nso in all generation in time, the cause must in its action pass through\nall the degrees of quantity from zero to that of the final effect.\n\n     \u201cEvery change has a cause which evinces its causality in the whole\n     time in which the change takes place. This cause, therefore, does\n     not engender the change suddenly (at once or in one moment), but in\n     a time, so that, as the time increases from its initial moment _a_\n     to its completion in _b_, the quantity of the reality (_b-a_) is in\n     like manner generated through all lesser degrees which are\n     contained between the first and the last.\u201d[1222]\n\nThis argument is inconclusive. As Kant himself recognises in regard to\nspace,[1223] we may not without special proof assume that what is true\nof time must be true of the contents of time. If time, change, and\ncausation can be equated, what is true of one will be true of all three.\nBut the assumption upon which the argument thus rests has not itself\nbeen substantiated.\n\nIn the third proof[1224] the argument is stated in extreme subjectivist\nterms which involve the further assumption that what is true of\napprehension is _ipso facto_ true of everything apprehended. The\npossibility of establishing the law of dynamical continuity follows,\nKant declares, as a consequence of its being a law of our subjective\napprehension.\n\n     \u201cWe anticipate only our own apprehension, the formal condition of\n     which, inasmuch as it inheres in the mind prior to all given\n     appearances, must certainly be capable of being known _a\n     priori_.\u201d[1225]\n\nKant\u2019s attitude towards the physical principle of continuity underwent\nconsiderable change. In his _New Doctrine of Motion and Rest_\n(1758)[1226] he maintains that it cannot be proved, and that physicists\nmay rightly refuse to recognise it even as an hypothesis. It is in the\nEssay on _Negative Quantity_ (1763)[1227] that Kant first adopts the\nattitude of the _Critique_, and rejects the \u201cspeculative\u201d objections\nraised against the mathematical conception of the infinitely small. In\nthe _Metaphysical First Principles of Natural Science_[1228] the\nprinciple of continuity is defended and developed, but only in its\napplication to material existence, not in its relation to the causal\nprocess.\n\nC. =Third Analogy.=--_All substances, in so far as they are coexistent,\nstand in thoroughgoing communion,[1229] i.e. in reciprocity with one\nanother._ Or, as in the second edition: _All substances, so far as they\ncan be perceived to coexist in space, are in thoroughgoing reciprocity._\n\nThis section contains four separate proofs. The first three paragraphs\nin the text of the first edition contain the first proof. The fourth\nparagraph supplies a second proof, and the fifth paragraph a third. In\nthe second edition Kant adds a fourth proof (the first paragraph of the\ntext of the second edition).\n\nWe may lead up to these proofs by first formulating (_a_) the\nfundamental assumption upon which they proceed, and (_b_) the thesis\nwhich they profess to establish. (_a_) The argument involves the same\ninitial assumption as the preceding _Analogies_, viz. that\nrepresentations exist exclusively in succession, or stated in\nphenomenalist terms, that the objectively coexistent can be apprehended\nonly in and through representations that are sequent to one another in\ntime.[1230] Upon this assumption the problem of the third _Analogy_ is\nto explain how from representations all of which are in succession we\ncan determine the objectively coexistent. (_b_) In the\n_Dissertation_[1231] Kant had maintained that though the possibility of\ndynamical communion of substances is not necessarily involved in their\nmere existence, such interaction may be assumed as a consequence of\ntheir common origin in, and dependence upon, a Divine Being. In the\n_Critique_ no such metaphysical speculations are any longer in order,\nand Kant recognises that as regards things in themselves it is not\npossible to decide whether dynamical interaction is, or is not,\nnecessarily involved in coexistence. The problem of this third _Analogy_\nconcerns only appearances, which as such must be subject to the\nconditions of unitary experience; and one such condition is that they be\napprehended as belonging to a single objective order of nature, and\ntherefore as standing in reciprocal relations of interaction. _The\napprehension of substances as reciprocally determining one another is_,\nKant contends, _an indispensable condition of their being known even as\ncoexistent_. Such is Kant\u2019s thesis. The proof may first be stated in\nwhat may be called its typical or generic form. Kant\u2019s four successive\nproofs can then be related to it as to a common standard.\n\nTwo things, A and B, can be apprehended as coexistent only in so far as\nwe can experience them in either order, _i.e._ when the order of our\nperceptions of them is reversible. If they existed in succession, this\ncould never be possible. The earlier member of a time series is past\nwhen the succeeding member is present, and what belongs to the past can\nnever be an object of perception. The fact that the order in which\nthings can be perceived is reversible would thus seem to prove that they\ndo not exist successively to one another in time.[1232] That, however,\nis not the case. By itself such experience does not really suffice to\nyield consciousness of coexistence. It can yield only consciousness of\nan alternating succession.[1233] A further factor, namely,\ninterpretation of the reversibility of our perceptions as due to their\nbeing conditioned by objects which stand in the relation of reciprocal\ndetermination, must first be postulated. If these objects mutually\ndetermine one another to be what they are, no one of them can be\nantecedent to or subsequent upon the others; and by their mutual\nreference each will date the others as simultaneous with itself. In\nother words, the perception of the coexistence of objects involves the\nconception of them as mutually determining one another. The principle of\ncommunion or reciprocity conditions the experience of coexistence, and\nis therefore valid for objects apprehended in that manner.\n\nKant also maintains, more by implication than by explicit statement,\nthat as A and B need not stand in any direct relation, the apprehension\nof them as coexistent involves the conception of an all-embracing order\nof nature within which they fall and which determines them to be what\nthey are. If any one of them, even the most minute and insignificant,\nwere conceived as altered, corresponding simultaneous variations would\nhave to be postulated for all the others. The unity of the phenomenal\nworld is the counterpart of the unity of apperception. Unity of\nexperience involves principles which prescribe a corresponding unity in\nthe natural realm. Dynamical _communion_ is the sufficient and necessary\nfulfilment of this demand. It carries to completion the unity demanded\nby the preceding _Analogies_ of _substance_ and _causality_. Kant sums\nup his position in a note to A 218 = B 265.\n\n     \u201cThe unity of the world-whole, in which all appearances have to be\n     connected, is evidently a mere consequence of the tacitly assumed\n     principle of the communion of all substances which are coexistent.\n     For if they were isolated, they would not as parts constitute a\n     whole. And if their coexistence alone did not necessitate their\n     connection (the reciprocal action of the manifold) we could not\n     argue from the former, which is a merely ideal relation, to the\n     latter, which is a real relation. We have, however, in the proper\n     context, shown that communion is really the ground of the\n     possibility of an empirical knowledge of coexistence, and that\n     therefore the actual inference is merely from this empirical\n     knowledge to communion as its condition.\u201d\n\nTo turn now to Kant\u2019s successive proofs. The first[1234] calls for no\nspecial comment. It coincides with the above. The second[1235] proof is\nan incompletely stated argument, which differs from the first only in\nits more concrete statement of the main thesis and in its limitation of\nthe argument to spatial existences. Dynamical community is the\nindispensable condition of our apprehension of any merely spatial\nside-by-sideness. Kant now adds that it is the dynamical continuity of\nthe _spatial_ world which enables us to apprehend the coexistence of its\nconstituents. The important bearing of this argument we shall consider\nin its connection with the proof which Kant added in the second\nedition.\n\nThe third[1236] proof is probably the earliest in date of writing. It\ndraws a misleading distinction between subjective and objective\ncoexistence, and seems to argue that only the latter form of coexistence\nneed presuppose the employment of the category of reciprocity. That runs\ndirectly counter to the central thesis of the other proofs, that only in\nterms of dynamical relation is coexistence at all apprehensible. That\nthe above distinction indicates an early date of writing would seem to\nbe confirmed by the obscure phrase \u201ccommunity of apperception\u201d which is\nreminiscent of the prominence given to apperception in Kant\u2019s earlier\nviews, and by the concluding sentence in which Kant employs\nterms--inherence, consequence, and composition--that are also\ncharacteristic of the earlier stages of his Critical enquiries.[1237]\n\nIt is significant that in the new argument[1238] of the second edition\nthe space factor, emphasised in the second proof of the first edition,\nis again made prominent.[1239] The principle is, indeed, reformulated in\nsuch manner as to suggest its limitation to spatial existences. \u201cAll\nsubstances, so far as they can be perceived to coexist _in space_, are\nin thoroughgoing reciprocity.\u201d Now it is decidedly doubtful whether Kant\nmeans to limit the category of reciprocity to spatial existences. As we\nhave already noted,[1240] he would seem to hold that though the category\nof causality can _acquire_ meaning only in its application to events in\nspace, it may in its subsequent employment be extended to the states of\ninner sense. The latter are effects dynamically caused, and among their\ncausal conditions are mechanical processes in space. The extension of\nthe category of reciprocity to include sensations and desires\nundoubtedly gives rise to much greater difficulties than those involved\nin the universal application of the causal principle. On the other hand,\nits limitation to material bodies must render the co-ordination of\nmental states and mechanical processes highly doubtful, and would carry\nwith it all the difficulties of an epiphenomenal view of psychical\nexistences. The truth probably is that in this matter Kant had not\nthought out his position in any quite definite manner; and that owing to\nthe influence, on the one hand of the dualistic teaching of the\ntraditional Cartesian physics, and on the other of his increasing\nappreciation of the part which space must play in the definition and\nproof of the principles of understanding, he limited the category of\nreciprocity to spatial existences, without considering how far such\nprocedure is capable of being reconciled with his determinist view of\nthe empirical self. His procedure is also open to a second objection,\nnamely, that while thus reformulating the principle, he fails to remodel\nhis proof in a sufficiently thoroughgoing fashion. The chief stress is\nstill laid upon the temporal element; and in order to obtain a proof of\nthe principle that will harmonise with the prominence given to the\nspace-factor, we are thrown back upon such supplementary suggestions as\nwe can extract from the second argument of the first edition. It is\nthere stated that \u201cwithout dynamical communion even spatial community\n(_communio spatii_) could never be known empirically.\u201d[1241] That is an\nassertion which, if true, will yield a proof of the principle of\nreciprocity analogous to that which has been given of the principle of\ncausality; for it will show that just as the conception of causality is\ninvolved in, and makes possible, the awareness of time, so the\nconception of reciprocity is involved in, and makes possible, the\nawareness of space.\n\nThe proof will be as follows. The parts of space have to be conceived as\nspatially interrelated. Space is not a collection of independent spaces;\nparticular spaces exist only in and through the spaces which enclose\nthem. In other words, the parts of space mutually condition one another.\nEach part exists only in and through its relations, direct or indirect,\nto all the others; the awareness of their coexistence involves the\nawareness of this reciprocal determination. But space cannot, any more\nthan time, be known in and by itself;[1242] and what is true of space\nmust therefore hold of the contents, in terms of the interrelations of\nwhich space can alone be experienced. How, then, can the reciprocal\ndetermination of substances in space be apprehended by a consciousness\nwhich is subject in all its experiences to the conditions of time? As\nKant has pointed out in A 211 = B 258,[1243] objective coexistence is\ndistinguished from objective sequence by reversibility of the\nperceptions through which it is apprehended. When A and B coexist, our\nperceptions can begin with A and pass to B, or start from B and proceed\nto A. There is also, as Kant observes in the second proof, a further\ncondition, namely, that the transition is in each case made through a\n_continuous_ series of changing perceptions.\n\n     \u201cOnly the continuous influences in all parts of space can lead our\n     senses from one object to another. The light, which plays between\n     our eye and the celestial bodies, produces a mediate communion\n     between us and them, and thereby establishes the coexistence of the\n     latter. We cannot empirically change our position (perceive such a\n     change), unless matter in all parts of space makes the perception\n     of our position possible to us. Only by means of its reciprocal\n     influence can matter establish the simultaneous existence of its\n     parts, and thereby, though only mediately, their coexistence with\n     even the most remote objects. Without communion, every perception\n     of an appearance in space is broken off from every other, and the\n     chain of empirical representations, _i.e._ experience, would have\n     to begin entirely anew with every new object, without the least\n     connection with preceding representations, and without standing to\n     them in any relation of time.\u201d[1244]\n\nBut even such reversibility of _continuous_ series does not by itself\nestablish coexistence. For in the imagination[1245] we can represent\nsuch series, without thereby acquiring the right to assert that they\nexist not as series but as simultaneous wholes. And as Kant might also\nhave pointed out, even in sense-perception we can experience reversible\ncontinuous series that do not in any way justify the inference to\ncoexistence. We may, for instance, produce on a musical instrument a\nseries of continuously changing sounds, and then in immediate succession\nproduce the same series in reverse order. An additional factor is\ntherefore required, namely, the interpretation of the reversibility of\nour perceptions as being grounded in objects which, because spatially\nextended, and spatially continuous with one another, can yield\ncontinuous series of perceptions, and which, because of their\nthoroughgoing reciprocity, make possible the reversing of these series.\nTo summarise the argument in a sentence: as the objectively coexistent,\nif it is to be known at all, can only be known through sequent\nrepresentations, the condition of its apprehension is the possibility of\ninterpreting reversible continuous series as due to the reciprocal\ninteraction of spatially ordered substances.\n\nThis argument has a twofold bearing. Its most obvious consequence is\nthat all things apprehended as coexistent must be conceived as standing\nin relations of reciprocal interaction; but by implication this involves\nthe further consequence that the conceptual principle of reciprocity is\nan integral factor in all apprehension of space. Space, though intuitive\nin character, has a meaning that demands this concept for its\narticulation. Just as consciousness of temporal sequence is only\npossible in terms of causation, so consciousness of spatial coexistence\nis only possible through application of the category of reciprocity. And\nsince, on Kant\u2019s view, awareness of space conditions awareness of time,\nthese conclusions carry the Critical analysis of our consciousness of\ntime a stage further. In confirmation of the more general argument of\nthe objective deduction, reciprocity is added to the already large\nsum-total of the indispensable conditions of our time-consciousness;\nwhile in regard to time itself it is shown that, owing to its\nspace-reference, coexistence may be counted among its possible modes.\n\nI have made occasional reference to the positions adopted by Stout in\nhis _Manual of Psychology_, and may here indicate their relation to the\npresent argument. Stout cites four \u201ccategories\u201d or ultimate principles\nof unity which \u201cbelong even to rudimentary perceptual consciousness as a\ncondition of its further development,\u201d[1246] namely, spatial unity,\ntemporal unity, causal unity, and the unity of different attributes as\nbelonging to the same thing. The criticism which, from the standpoint of\nthe _Analogies_, has to be passed upon this list,[1247] is that it\nignores the category of reciprocity, _i.e._ of systematic\ninterconnection, and that it fails to recognise the close relation in\nwhich the various principles stand to one another. The temporal unity\nmust not be isolated from causal unity, nor either of them from the\nspatial unity, with which the category of reciprocity is inseparably\nbound up. Further, Kant maintains that these principles are demanded,\nnot merely for the _development_ of perceptual consciousness, but for\nits very existence.\n\nBut Kant\u2019s argument suggests many difficulties which we have not yet\nconsidered, and we may again employ Schopenhauer\u2019s criticisms to define\nthe issues involved.\n\n     \u201cThe conception of reciprocity ought to be banished from\n     metaphysics. For I now intend, quite seriously, to prove that there\n     is no reciprocity in the strict sense, and this conception, which\n     people are so fond of using, just on account of the indefiniteness\n     of the thought, is seen, if more closely considered, to be empty,\n     false, and invalid.... It implies that both the states A and B are\n     cause and that both are effect of each other; but this really\n     amounts to saying that each of the two is the earlier and also the\n     later; thus it is an absurdity.\u201d[1248]\n\nThis criticism proceeds on the assumption that the category of\nreciprocity reduces to a dual application of the category of causality.\nIf that were the case, there would, of course, be no separate category\nof reciprocity,[1249] and further it would, as Schopenhauer maintains,\nbe impossible to regard A and B as being at one and the same time both\ncause and effect of one another. Causality determines the order of the\nstates of substances in the time series; reciprocity must be distinct\nfrom causality if it is to be capable of defining the order of their\ncoexistent states in space. A deduction from the dual application of the\nconception of causality has, therefore, no bearing upon the question of\nthe possibility of this further category. Kant has laid himself open to\nthis criticism by a passage which occurs in the first proof, and which\nshows that he was not quite clear in his own mind as to how reciprocity\nought to be conceived.\n\n     \u201cThat alone can determine the position of anything else in time,\n     which is its cause or the cause of its determinations. Every\n     substance (inasmuch as only in its determinations can it be an\n     effect) must therefore contain in itself the causality of certain\n     determinations in the other substance, and at the same time the\n     effects of the causality of that other, _i.e._ they must stand in\n     dynamical communion (immediately or mediately), if their\n     coexistence is to be known in any possible experience.\u201d[1250]\n\nIt should be noted that in the new proof[1251] in the second edition\nKant is careful to employ the terms ground and influence in place of the\nterms cause and causality.\n\nSecondly, Schopenhauer argues that if the two states necessarily belong\nto each other and exist at one and the same time, they will not be\nsimultaneous, but will constitute only _one_ state.[1252] Schopenhauer\nis again refusing to recognise the conditions under which alone a\nspecial category of reciprocity is called for. We can speak of\nsimultaneity only if a _multiplicity_ be given; and if it be given, its\nnature as _simultaneous plurality_ cannot be comprehended through a\ncausal law, which, as such, applies only to sequent order.\n\nLastly, Schopenhauer endeavours to confirm his position by examination\nof the supposed instances of reciprocity.\n\n     \u201d[In the continuous burning of a fire] the combination of oxygen\n     with the combustible body is the cause of heat, and heat, again, is\n     the cause of the renewed occurrence of the chemical combination.\n     But this is nothing more than a chain of causes and effects, the\n     links of which have alternately _the same name_.... We see before\n     us only an application of the single and simple law of causality\n     which gives the rule to the sequence of states, but never anything\n     which must be comprehended by means of a new and special function\n     of the understanding.\u201d[1253]\n\nSchopenhauer is again misled by his equating of reciprocity with causal\naction. Combustion is quite obviously a case of sequent processes.\nInstead of proving that coexistence does not involve reciprocity,\nSchopenhauer is only showing that cause and effect may sometimes, as\nKant himself observes,[1254] seem to be simultaneous.[1255] Action\n_followed_ by reaction is not equivalent to what Kant means by\nreciprocal determination. Schopenhauer also cites the instance of a pair\nof scales brought to rest by equal weights.\n\n     \u201cHere there is no effect produced, for there is no change; it is a\n     state of rest; gravity acts, equally divided, as in every body\n     which is supported at its centre of gravity, but it cannot show its\n     force by any effect.\u201d[1256]\n\nThis example is more in line with what Kant would seem to have in view,\nbut is still defined in reference to the problem of causation, and not\nin reference to that of coexistence. Kant is not enquiring whether\ncoexistent bodies are related as causes and effects, though, as we have\nalready observed, his language betrays considerable lack of clearness on\nthis very point. He is endeavouring to define the conditions under which\nwe are enabled to recognise that bodies, external to one another in\nspace and apprehensible only through sequent perceptions, are none the\nless coexistent. And the answer which he gives is that coexistence can\nonly be determined by reference of each existence to the totality of\nsystematic relations within which it is found, its particular spatial\nlocation being one of the factors which condition this reference. Causal\nexplanation in the most usual meaning of that highly ambiguous phrase,\nnamely, as explanation of an artificially isolated event by reference to\nantecedents similarly isolated from their context, may partially account\nfor this event being of one kind rather than another, but will not\nexplain why it is to be found at this particular time in this particular\nplace. That is to say, it will not answer the question which is asked\nwhen we are enquiring as to what events are coexistent with it.\n\nBut the considerations which thus enable us to dispose of Schopenhauer\u2019s\ncriticisms have the effect of involving us in new, and much more\nformidable, difficulties. Indeed they disclose the incomplete, and quite\ninadequate, character of Kant\u2019s proof of the third _Analogy_. For must\nnot spatial co-existence be independently known if it is to serve as one\nof the factors determinant of reciprocity? Can the apprehension of\nextended bodies wait upon a prior knowledge of the system of nature to\nwhich they belong?\n\nThe mere propounding of these questions does not, however, suffice to\noverthrow Kant\u2019s contention. For he is prepared--that is indeed the\nreason why the _Critique_ came to be written--to answer them in a manner\nthat had never before been suggested, save perhaps in the philosophies\nof Plato and Aristotle. This answer first emerges in the _Dialectic_, in\nthe course of its treatment of the wider problem, of which the above\ndifficulties are only special instances, how if conditioned parts can\nonly be known in terms of an unconditioned whole, any knowledge\nwhatsoever can be acquired by us. But though Kant in the _Dialectic_\ngives due prominence to this fundamental problem, the hard and fast\ndivisions of his architectonic--and doubtless other influences which\nwould be difficult to define--intervene to prevent him from recognising\nits full implications. For the problem is viewed in the _Dialectic_ as\ninvolving considerations altogether different from those dwelt upon in\nthe _Analogies_, and as being without application to the matters of\nwhich they treat.\n\nThe situation thus created is very similar to that which is occasioned\nby Kant\u2019s unfortunate separation of the problems of space and time in\nthe _Aesthetic_ from the treatment of the categories in the _Analytic_.\nIn the _Aesthetic_ space and time are asserted to be intuitive, not\nconceptual, in nature; and yet in the _Analytic_ we find Kant\ndemonstrating that the principles of causality and reciprocity are\nindispensably involved in their apprehension. But even more misleading\nis the separation of the problems of the _Aesthetic_ and _Analytic_ from\nthose of the _Dialectic_. Kant\u2019s primary and prevailing interest is in\nthe metaphysics, not in the mere methodology, of experience; and it is\nin the _Dialectic_ that the metaphysical principles which underlie and\ninspire all his other tenets first find adequate statement. Since the\nthird _Analogy_ defines the criterion of coexistence in entire\nindependence of all reference to the Ideas of Reason, Kant is thereby\nprecluded from even so much as indicating the true grounds upon which\nhis position, if it is to be really tenable, must be made to rest. For\nas he ultimately came to recognise, the intuition of space not only\ninvolves the conceptual category of reciprocal determination, but\nlikewise demands for its possibility an Idea of Reason. In space the\nwider whole is always prior in thought to the parts which go to\nconstitute it. But though Kant states[1257] that this characteristic of\nspace justifies its being entitled an Idea of Reason, he nowhere takes\nnotice of the obvious and very important bearing which this must have\nupon the problem, how we are to formulate the criterion of coexistence.\n\nThe general character of time is analogous to that of space, and our\nformulation of the criterion of causal sequence is therefore similarly\naffected. The system of nature is not the outcome of natural laws which\nare independently valid; natural laws are the expression of what this\nsystem prescribes; they are the modes in which it defines and embodies\nits inherent necessities.\n\nThe situation which these considerations would seem to disclose may,\ntherefore, be stated as follows. If the empirical criteria of truth are\nindependent of the Ideas of Reason, the _Analytic_ may be adequate to\ntheir discussion, but will be unable to justify the assertion that there\nis a category of reciprocal or systematic connection distinct from that\nof causality. If, however, it should be found that these criteria are\nmerely special applications of standards _metaphysical in\ncharacter_--and that would seem to be Kant\u2019s final conclusion,--only in\nthe light of the wider considerations first broached in the _Dialectic_,\ncan we hope to define their nature and implications with any approach to\ncompleteness.\n\n\n4. THE POSTULATES OF EMPIRICAL THOUGHT IN GENERAL\n\n=First Postulate.=--_That which agrees, in intuition and in concepts, with\nthe formal conditions of experience is possible._\n\n=Second Postulate.=--_That which is connected with the material conditions\nof experience (that is, with sensation) is actual._\n\n=Third Postulate.=--_That which is determined, in its connection with the\nactual, according to universal conditions of experience is (that is,\nexists as) necessary._\n\nIn this section Kant maintains that when the Critical standpoint is\naccepted, possibility, actuality and necessity can only be defined in\nterms of the conditions which render sense-experience possible. In other\nwords, the Critical position, that all truth, even that of _a priori_\nprinciples, is merely _de facto_, involves acceptance of the view that\nthe actual reduces to the experienced, and that only by reference to the\nactual as thus given can possibility and necessity be defined. The\nLeibnizian view that possibility is capable of being defined\nindependently of the actual, and antecedently to all knowledge of it,\nmust be rejected.\n\nAn analysis of the text can be profitably made only after a detailed\nexamination of Kant\u2019s general argument; and to that task we may at once\napply ourselves. The section affords further illustration of the\nperverting influence of Kant\u2019s architectonic, as well as of the\ninsidious manner in which the older rationalism continued to pervert his\nthinking in his less watchful moments.\n\n=First Postulate.=--In the opening paragraphs Kant uses (as it would seem\nwithout consciousness of so doing) the term possibility in two very\ndifferent senses.[1258] When the possible is distinguished from the\nactual and the necessary, it acquires the meaning defined in this first\n_Postulate_; it is \u201cthat which agrees with the _formal_ conditions of\nexperience.\u201d But it is also employed in a much narrower sense to signify\nthat which can have \u201cobjective reality, _i.e._ transcendental\ntruth.\u201d[1259] The possibility of the objectively real rests upon\nfulfilment of a threefold condition: (1) that it agree with the formal\nconditions of experience; (2) that it stand in connection with the\nmaterial of the sensuous conditions of experience; and (3) that it\nfollow with necessity upon some preceding state in accordance with the\nprinciple of causality, and so form part of a necessitated order of\nnature. In other words, it must be causally necessitated in order to be\nempirically actual; and only the empirically actual is genuinely\npossible. Such is also the meaning that usually attaches to the term\npossible in the other sections of the _Critique_. A \u2018possible\nexperience\u2019 is one that can become actual when the specific conditions,\nall of which must themselves be possible, are fulfilled. An experience\nwhich is not capable of being actual has no right to be described even\nas possible. As a term applicable to the objectively real, the possible\nis not wider than the actual, but coextensive with it. As Kant himself\nremarks, those terms refer exclusively to differences in the subjective\nattitude of the apprehending mind.\n\nThis ambiguity in the term \u2018possibility\u2019 has caused a corresponding\nambiguity in Kant\u2019s employment of the term \u2018actuality.\u2019 It leads him to\nendeavour to define the actual, not in its connection with the\nconditions of possibility, but in distinction from them. The possible\nhaving been defined (in the first _Postulate_) solely in terms of the\n_formal_ factors of experience, he proceeds to characterise the actual\nin a similarly one-sided fashion, exclusively in terms of the\n_material_ element of given sensation. Doubtless the element of\nsensation must play a prominent part in enabling us to decide what is or\nis not actually existent, but no definition which omits to take account\nof relational factors can be an adequate expression of Critical\nteaching. Indeed, we only require to substitute the words \u2018sensuously\ngiven\u2019 for \u2018actual\u2019 in Kant\u2019s definition of the third _Postulate_\n(_i.e._ of the necessary) in order to obtain a correct statement of the\ntrue Critical view of actual existence: it is \u201cthat which is determined\nin its connection with the sensuously given according to universal\nconditions of experience.\u201d For Kant the actual and the necessary,\nobjectively viewed, coincide. Necessity is for the human mind always\nmerely _de facto_; and nothing can be objectively actual that is not\ncausally determined. As the empirically possible cannot, in its\nobjective reference, be wider than the empirically necessary, one and\nthe same definition adequately covers all three terms alike. While the\ndistinctions between them will, of course, remain, they will be\napplicable, not to objects, but only to the subjective conditions of\nexperience in so far as these may vary from one individual to another.\nExperiences capable of being actual for one individual may be merely\npossible for another. And what is merely actual to one observer may by\nothers be comprehended in its necessitating connections. The terms will\nnot denote differences in the real, but only variations in the cognitive\nattitude of the individual.\n\nThus in professing to show that the three _Postulates_ are\ntranscendental _principles_, Kant does less than justice to his own\nteaching. For though both here and in the opening sections of the\nchapter[1260] he speaks of them in this manner, _i.e._ as being\nconditions alike of ordinary and of scientific experience, he has\nhimself admitted in so many words the inappropriateness of such a\ndescription.\n\n     \u201cThe principles of modality are nothing more than explanations\n     [not, it may be noted, proofs] of the concepts of possibility,\n     actuality and necessity, in their empirical use, and are therefore\n     at the same time restrictions of all the categories to this merely\n     empirical use, ruling out and forbidding their transcendental [=\n     transcendent] employment.\u201d[1261]\n\nThat is to say, these so-called principles are not really principles;\nthey merely embody explanatory statements designed to render the\npreceding results more definite, and especially to guard against the\nillegitimate meanings which the Leibnizian metaphysics had attached to\ncertain of the terms involved.\n\nThese considerations bring us to the real source of Kant\u2019s perverse\nargumentation, namely, the artificial (but none the less imperious)\ndemands of his architectonic. He is constrained to provide a set of\n_principles_ corresponding to the categories of modality. The\ndefinitions of the modal categories have therefore to be called by that\ninappropriate name. But that is not the end of the matter. In order to\nmeet the needs of his logical framework, Kant proceeds even further than\nhe had ventured to do in the sections on the _Axioms of Intuition_ and\n_Anticipations of Perception_. There he fell so far short as to provide\nonly a single principle in each case. In dealing, however, with the\ncategories of relation he has been able to define each of the three\ncategories separately, and to derive from each a separate principle.\nMany of the defects in his argument are, indeed, traceable to this\nsource. The close interrelations of the three principles are, as we have\nhad occasion to note, seriously obscured. But still, in the main,\nseparate treatment of each has proved feasible. Kant, encouraged, as we\nmay believe, by this successful fulfilment of architectonic\nrequirements, now sets himself to develop, in similar fashion, a\nseparate principle for each modal category. But for any such enterprise\nthe conditions are less favourable than in the case of the categories of\nrelation. For, as just indicated, no one of the three can, on Critical\nprinciples, possess any genuine meaning save in its relation to the\nothers. Before following out this line of criticism, we must however\nnote some further points in Kant\u2019s argument.\n\nIn A 219 = B 266, and again in A 225 = B 272, Kant makes the statement\nthat a concept can be complete prior to any decision as to its\npossibility, actuality, or necessity. This contention is capable of\nbeing interpreted in two quite independent ways, and in only one of\nthose ways is it tenable. He may mean that the distinction between the\npossible, the actual, and the necessary, does not concern the\nobjectively real, which as such is always both actual and necessary, but\nonly the subjective attitude of the individual towards the objects of\nhis thought and experience. From the Critical standpoint, as we have\nbeen arguing, such a contention is entirely just. But Kant would seem in\nthe above statement to be chiefly concerned to maintain that a\nconception may be complete and determinate, even while we remain in\ndoubt whether the existence for which it stands is even possible.[1262]\nSuch a view is merely a relic of the Leibnizian rationalism from which\nhe is striving to break away. All existences have their place in a\nsystematic order of experience, and no conception of them can be either\ncomplete or determinate which fails to specify the causal context to\nwhich they belong. The process of specifying the detail of a concept is\nthe only process whereby we can define its possibility, actuality, or\nnecessity.[1263] Were it capable of complete statement without\ndetermination of its modal character, it could never form part of a\nunified experience. The examples of \u201cfictitious\u201d concepts, which Kant\ncites, are either so determinate as to be demonstrably inconsistent with\nexperience, and therefore empirically impossible, or so indeterminate as\nto afford no sufficient means of deciding even as to their possibility.\n\nThere is a further objection to the definition given of possibility in\nthe first _Postulate_. After stating that the possible is what _agrees_\nwith the formal conditions of experience, Kant proceeds, on the one\nhand, to argue that the forms of intuition and the categories of\nunderstanding may, in accordance with this criterion, be viewed as\npossible, and, on the other hand, to maintain that no other concepts can\nbe so regarded.[1264] That is to say, the possible, as thus interpreted,\ndoes not consist in something additional to, and in harmony with, the\nconditions of experience, but reduces without remainder to those very\nforms. Now Kant is not betrayed merely by inadvertence into thus\nnarrowing the sphere of the possible; such limitation is an almost\ninevitable consequence of the one-sided manner in which he has treated\nthe concept of the possible in this first _Postulate_. He professes to\nbe proceeding in the light of the results obtained in the transcendental\ndeduction, and to be defining the possible in terms of the conditions\nwhich make sense-experience possible. But the deduction has shown that\nexperience is possible only in so far as the material factors co-operate\nwith the formal. And when this is recognised, it becomes obvious that a\ndefinition of the possible in terms of sensation,--namely, as that which\nis capable of being presented in sense-perception,--is equally\nlegitimate, and is indeed required in order to correct the deficiencies\nof the definition which Kant has himself given. As both factors are\nindispensable in all possible experience, both must be reckoned with in\ndefining the possible.\n\nKant\u2019s argument in the fifth paragraph is somewhat obscured by its\ncontext. He is contending that fictitious (_gedichtete_) concepts,\nelaborated from the contents presented in perception, cannot be\ndetermined as possible. As they involve sensuous contents, the formal\nelements of experience do not suffice for proof of their possibility;\nand since the contents are supposed to have been recombined in ways not\nsupported by experience, an empirical criterion is equally inapplicable.\nObviously Kant is here using the term \u2018possible\u2019 not in the meaning of\nthe first _Postulate_, but in its narrower connotation as signifying\nthat which is capable of objective reality. Such fictitious concepts may\ncompletely fulfil all the demands prescribed by space, time, and the\ncategories, and yet, as he here insists, be none the less incapable of\nobjective existence.\n\nThe argument is still further obscured by the character of the concrete\nexamples which Kant cites. They involve modes of action or of intuition\nwhich contradict the very conditions of human experience, and so for\nthat reason alone fall outside the realm of the empirically possible.\nThat would not, however, seem to be Kant\u2019s meaning in employing them.\nAssumed powers of anticipating the future or of telepathic communication\nwith other minds are, he says, concepts\n\n     \u201c...the possibility of which is altogether groundless, as they\n     cannot be based on experience and its known laws, and without such\n     confirmation are arbitrary combinations of thoughts, which,\n     although indeed free from contradiction, can make no claim to\n     objective reality and so to the possibility of an object such as we\n     here profess to think.\u201d[1265]\n\nThe mathematical examples which Kant gives in A 223 = B 271[1266] are no\nless misleading. The concept of a triangle can, it is implied, be\ndetermined as possible in terms of the first _Postulate_, since it\nharmonises with a formal condition of experience, namely, space. This is\ntrue only if it be granted that construction in space can be executed\nabsolutely _a priori_, in independence of all sense-experience. Such is,\nof course, Kant\u2019s most usual view; and to that extent the argument is\nconsistent. Mathematical concepts will from this point of view represent\nthe only possible exception to the general statement that the formal\nconditions of experience constitute a criterion of possibility for no\nconcepts save themselves. Kant\u2019s final conclusion is clearly and\nexplicitly stated in the following terms:\n\n     \u201cI leave aside everything the possibility of which can be derived\n     only from its reality in experience, and have here in view only the\n     possibility of things through _a priori_ concepts; and I maintain\n     the thesis that the possibility of such things can never be\n     established from such concepts taken in and by themselves, but\n     only when they are viewed as formal and objective conditions of\n     experience in general.\u201d[1267]\n\nWe are now in a position to appreciate the reasons which have induced\nAdickes to regard the text as of composite origin.[1268] Adickes argues\nthat Kant\u2019s original intention was to treat the three concepts together,\nshowing that they can be defined only in empirical terms, and that their\nsignificance is consequently limited to the world of appearance. Such is\nthe content of the first, second, fourth (excepting the first sentence),\nand fifth paragraphs. No attempt is made to separate the three\n_Postulates_, and the term possibility is throughout employed\nexclusively as referring to objective reality. (In the third paragraph\nit is used in both senses.) The other paragraphs were, according to\nAdickes\u2019 theory, added later, when Kant unfortunately resolved to fulfil\nmore exactly the requirements of his architectonic. That involved the\nformulation of three separate _Postulates_, with all the many evil\nconsequences which that attempt carried in its train. He must then have\ninterpolated the third paragraph, added the first sentence to the fourth\nparagraph, corrected the too extensive sweep of the older paragraphs\nthrough the introduction of the sixth paragraph, further supplemented\nthe exposition of the first _Postulate_ by the seventh paragraph, and\nadded independent treatments of the postulates of actuality and\nnecessity. This may seem a very complicated and hazardous hypothesis;\nbut careful examination of the text, with due recognition of the\nconfused character of the argument as it stands, will probably convince\nthe reader that Adickes is in the right.\n\n=Second Postulate.=[1269]--Perception is necessary to all determination of\nactuality. The actual is either itself given in perception or can be\nshown, in accordance with the _Analogies_, to stand within the unity of\nobjective experience, in connection with what is thus given. So long as\nKant expresses himself in these terms his statements are entirely valid.\nNothing which cannot be shown to be bound up with the contingent\nmaterial of sense-experience can be admitted as actual. He proceeds,\nhowever, to give a definition of actuality which entirely omits all\nreference to the _Analogies_, and which is open to the same fundamental\ncriticism as his characterisation of possibility in the first\n_Postulate_. Though the earlier statements give due recognition both to\nthe material content and to the relational forms constitutive of\ncomplete experience, Kant now contrasts the mere or bare (_blosser_)\nconcept and the given perception in a manner which suggests the\nunfortunate distinction drawn in the _Prolegomena_, and repeated in the\nsecond edition of the _Critique_, between judgments of perception and\njudgments of experience.[1270] Kant\u2019s reference to \u201cthe mere concept of\na thing\u201d[1271] is on the same lines as the opening paragraph of the\nsection. However complete the concept may be, it yields not the least\nground for deciding as to the existence of its object.\n\nKant\u2019s thinking, as I have already pointed out, is here perverted by the\ncontinuing influence of the Leibnizian rationalism. He is forgetting\nthat, on Critical principles, even the categories are meaningless except\nin their reference to the contingently given. If that be true of the\nstrictly _a priori_, it must hold with even greater force of empirical\nconcepts with sensuous content. As the sole legitimate function of\nconcepts, whether _a priori_ or empirical, is to organise and unify the\nmaterial of sense, there can be no such thing as the mere or bare\nconcept. Such a combination of words is without Critical significance. A\nconcept as such must refer to, and embody insight into, the real. Only\nin proportion to its incompleteness, that is, to its indefiniteness, can\nit remain without specific and quite determinate location within the\ncontext of unified experience. It may, indeed, be found convenient to\nretain the phrase \u201cmere concept\u201d notwithstanding its misleading\ncharacter and rationalistic origin. It must, however, be used only to\nmark the indefiniteness, indeterminateness, or incompleteness which\nprevents it from adequately revealing the denotation to which through\nthe nature of its content it necessarily refers. Meaning and existence,\nconnotation and denotation, are complementary the one to the other, and\nthough not, perhaps, coextensive (if that term has itself meaning in\nthis connection), are none the less inseparably conjoined. When Kant\u2019s\nutterances, as frequently happens, imply the contrary, they may be taken\nas revealing the strength and insidious tenacity of the influences from\nwhich he was sufficiently courageous, but not always sufficiently\nwatchful, to break away.\n\nThe doctrine of the \u201cmere concept\u201d finds its natural supplement in the\nequally un-Critical assertion that\n\n     \u201c...perception [evidently employed in the less pregnant sense, as\n     signifying \u2018sensation accompanied by consciousness\u2019], which\n     supplies the material to the concept, is _the sole character of\n     actuality_.\u201d[1272]\n\nThis same position is expressed equally strongly by Kant in his\n_Reflexionen_ (ii. 1095).\n\n     \u201cPossibility is thought without being given; actuality is given\n     without being thought; necessity is given through being thought.\u201d\n\nSuch statements are entirely out of harmony with Kant\u2019s central\nteaching. There is no lack of passages in the _Critique_ which inculcate\nthe direct contrary. Though the element of sensation is a _sine qua non_\nof all experience of the actual, the formal elements are no less\nindispensable. In their absence the merely given would reduce to less\nthan a dream; for even in dreams images are interpreted and are referred\nto some connected context. The given, merely as such, cannot enter the\nfield of consciousness, and is therefore \u201cfor us as good as nothing.\u201d As\nCaird has pointed out, we find in Kant\n\n     \u201c...two apparently contradictory forms of expression--(1) that the\n     understanding by means of its conceptions refers our preceptions to\n     objects, and (2) that conceptions are referred to objects only\n     indirectly through perceptions. The former mode of expression is\n     preferred whenever Kant has to show that \u2018perceptions without\n     conceptions are blind\u2019; the latter when he has to show that\n     \u2018conceptions without perceptions are empty.\u2019\u201d[1273] \u201cWe can\n     understand the possibility of Kant\u2019s looking at the subject in\n     these two opposite ways, only if we remember the reciprocal\n     presupposition of perception and conception in the judgment of\n     knowledge, and the way in which Kant tries to explain it, now from\n     the point of view of perception, and now from the point of view of\n     conception. The effect of this is, no doubt, a formal contradiction\n     which Kant himself never disentangles, but which _we_ must\n     endeavour to disentangle, if we would do justice to him.\u201d[1274]\n\nThe one-sidedness of Kant\u2019s definition of actuality is certainly due to\nthe cause suggested by Caird. The definition, notwithstanding its\nmisleading character, serves to enforce against the older rationalism,\nwith which Kant throughout this section is almost exclusively concerned,\nthe central tenet through which the Critical teaching is distinguished\nfrom that of Leibniz, namely, that neither existence, possibility, nor\nnecessity, can be established save by reference to the contingent nature\nof the sensuously given. Proof by reference to the possibility of\nexperience can establish only those conditions which can be shown to be\n_de facto_ necessary in order that consciousness of time may be\naccounted for. The formal conditions of experience, which in and by\nthemselves are determinable neither as actual nor as possible, are\nestablished as actual, and so as necessary, by reference to the merely\ngiven; they are necessary only in this merely relative fashion, as being\nindispensable to what can never itself be viewed as other than\ncontingent.\n\n     \u201cOur knowledge of the existence of things reaches, then, only so\n     far as perception and its continuation according to laws of nature\n     can extend. If we do not start from experience, or do not proceed\n     according to laws of the empirical connection of appearances, our\n     guessing or enquiring into the existence of anything will only be\n     an idle pretence.\u201d[1275]\n\nPolemically, therefore, Kant\u2019s formulation of the second _Postulate_ is\nnot without its advantages, though from the inner standpoint of Critical\nteaching it is altogether inadequate.\n\nFor comment upon A 226 = B 273, and upon the general teaching of this\n_Postulate_ in its important bearing upon Kant\u2019s phenomenalism, cf.\nabove, pp. 318-19.\n\n=B 274-9.=--_Refutation of Idealism_, cf. above, p. 308 ff.\n\n       *       *       *       *       *\n\n_Third Postulate._[1276]--In the opening sentence Kant draws the\ndistinction which was lacking in his treatment of the first _Postulate_\nbetween \u2018material\u2019 and \u2018formal\u2019 modality. (No distinction, however, is\ndrawn between the \u2018formal\u2019 possibility of the first _Postulate_ and\nlogical possibility, which consists in absence of contradiction.) It is\nwith the former alone that we have to deal. As existence cannot be\ndetermined completely _a priori_, necessity can never be known from\nconcepts, but only by reference to the actually given, in accordance\nwith the universal principles that condition experience. Further, since\nsuch empirical necessity does not concern the existence of substances,\nbut only the existence of their states, viewed as dynamically caused,\nthe criterion of empirical necessity reduces to the second _Analogy_,\nviz. that everything which happens is determined by an antecedent\nempirical cause. This criterion does not extend beyond the field of\npossible experience, and even within that field applies only to those\nexistences which can be viewed as effects, _i.e._ as _events_ which come\ninto existence in time, and of which therefore the causes are of the\nsame temporal and conditioned character. The necessity is a hypothetical\nnecessity; given an empirical event, it can always be legitimately\nviewed as necessitated by an antecedent empirical cause.\n\nKant introduces, reinterprets, and in this altered form professes to\njustify, four of the central principles of the Leibnizian metaphysics.\n_In mundo non datur casus_ gives expression to the above empirical\nprinciple. _Non datur fatum_ may be taken as meaning that natural\n(_i.e._ empirical) necessity is a conditioned and therefore\ncomprehensible necessity, and is consequently not rightly described as\nblind. The other two principles, _non datur saltus_, and _non datur\nhiatus_ connect with the principle of continuity already established in\nthe _Anticipations of Perception_ and in the second _Analogy_.\n\nKant\u2019s further remarks reveal an uneasy feeling that he is neglecting to\nassign these principles to the pigeon-holes provided in his\narchitectonic. The reader, he states, may easily do so for himself. That\nmay be so, but only if the reader be permitted the same high-handed\nmethods of adjustment that are here illustrated in Kant\u2019s location of\n_non datur fatum_ with the principles of modality.[1277]\n\nIn the next paragraph (A 230 = B 282) Kant suddenly, without warning or\nexplanation, attaches to the term possibility a meaning altogether\ndifferent from any yet assigned to it. He now takes it as equivalent to\nthe absolutely or metaphysically possible. Combining this with the\nmeanings previously given to it by Kant we obtain the following table:--\n\n             {Logical: equivalent to absence of contradiction.\n             {Empirical: in the wider sense, equivalent to agreement\n             {with the formal conditions of experience; in the narrower\n  Possibility{or stricter sense, involving in addition the capacity\n             {of being presented in sense-experience.\n             {Metaphysical: equivalent to absolute possibility, a conception\n             {not of understanding but of Reason.\n\nWhen this last meaning is given to the term, an entirely new set of\nproblems arises, to the confusion of the reader who very properly\ncontinues to employ the term possibility in the empirical sense which,\nas Kant has been insisting, is alone legitimate. Kant has temporarily\nchanged over to the standpoint of the metaphysical view which he has\nbeen criticising, and accordingly uses the term \u2018possibility\u2019 in the\nLeibnizian sense. Is Leibniz, he asks, justified in maintaining that the\nfield of the possible is wider than the realm of the actual, and the\nlatter in turn wider in extent than the necessary? In reply Kant accepts\nthe metaphysical meaning assigned to the term \u2018possibility,\u2019 but\nrestates the problem in Critical fashion. Do all things belong as\nappearances to the context of a single experience, or are other types of\nexperience possible? Do other forms of intuition besides space and\ntime, other forms of understanding besides the discursive through\nconcepts, come within the range of the possible? These are questions\nwhich fall to be answered, not by the mere understanding, the sole\nfunction of which is empirical, but by Reason, which transcends the\nworld of appearance.\n\nKant introduces these questions, as he is careful to state,[1278] only\nbecause they are currently believed to be within the competence of the\nunderstanding; and he now for the first time points out that\npossibility, in this sense, means absolute possibility, that which is\nindependent of all limiting conditions, a meaning ruled out by the\npreceding treatment of the modal categories. Like all other absolute\nconceptions, it belongs to Reason, and must therefore await treatment in\nthe _Dialectic_. These admissions come, however, only after the\ndiscussion has been completed. Had Kant reversed the order of the two\nparagraphs which constitute this digression, and marked them off as\nbeing a digression, he would have greatly assisted the reader in\nfollowing the argument.\n\nKant adds a refutation of the merely logical arguments by which Leibniz\nhad professed to establish the priority and greater scope of the\npossible. From the proposition, everything actual is possible, we can\ninfer by immediate inference that some possible things are actual. That,\nhowever, would seem to imply that part of the possible is not actual,\nand that something must be added to the possible in order to constitute\nthe actual. But this, Kant replies, is obviously an untenable view. The\nsomething additional to the possible, not being itself possible, we\nshould be constrained to regard as impossible. _For our\nunderstanding_,[1279] the possible is that which connects with some\nperception in agreement with the formal conditions of experience. (Kant\nhere gives the correct Critical definition of the possible, by combining\nthe two first postulates.) Whether, and how far, other existences beyond\nthe field of sense experience are possible, we have no means of\ndeciding.\n\n=B 288-294.=--This second edition section emphasises the fact that\npossibility cannot be determined through the categories alone, but only\nthrough the categories in their relation to intuition, and indeed to\nouter intuition. Possibility is throughout taken as referring to\nobjective reality. The section is chiefly important in connection with\nthe problems bearing on the relation of inner and outer sense and on the\nnature of our consciousness of time.[1280]\n\nIn B 289-91 Kant criticises those rationalistic arguments which rest\nupon the equating of necessity of thought with necessity of existence.\nWhen it is sought by mere analysis of concepts to prove that all\naccidental existence has a cause, the most that can be shown is that the\nexistence of the accidental cannot be _comprehended_ by us, unless the\nexistence of a cause be assumed. But we may not argue that a condition\nof possible understanding is likewise a condition of possible\nexistence.[1281] What is or is not possible for thought is, without\nspecial proof, no sufficient criterion of what is or is not possible in\nthe real. If, again, the term accidental be taken as meaning that which\ncan exist only as a consequence of some other existence, the general\nprinciple becomes merely analytic, and must not be taken as establishing\nthe synthetic principle of causality. The latter demands transcendental\nproof by reference to the possibility of contingent experience.\n\n\n\n\nCHAPTER III\n\nON THE GROUND OF THE DISTINCTION OF ALL OBJECTS WHATEVER INTO PHENOMENA\nAND NOUMENA\n\n\nThis chapter, as Kant himself states,[1282] can yield no new results. It\nwill serve merely to summarise those already established in the\n_Analytic_, showing how they one and all converge upon a conclusion of\nsupreme importance for understanding the nature and scope of human\nexperience--the conclusion, that though the objective employment of the\ncategories can be justified only within the realm of sense-experiences,\nthey have a wider significance whereby they define a distinction between\nappearances and things in themselves. This is the conclusion which Kant\nnow sets himself to illustrate and enforce in somewhat greater detail.\nIt may be observed that the title of the chapter makes mention only of\ngrounds for _distinguishing_ between phenomena and noumena. That things\nin themselves really exist, Kant, as we shall find, never seriously\nthought of questioning.\n\nKant begins by recalling a main point in the preceding argument. The\ncategories apart from the manifold of sensibility are merely logical\nfunctions without content.[1283] Though _a priori_, they require to be\nsupplemented through empirical intuition.\n\n     \u201cApart from this relation to possible experience they have no\n     objective validity of any sort, but are a mere play of the\n     imagination or the understanding with their respective\n     representations.\u201d[1284]\n\nAs evidence of the truth of this conclusion Kant now adds a further\nargument, namely, the impossibility of defining the categories except in\nterms that involve reference to the conditions of sensibility.[1285]\nWhen these conditions are omitted, the categories are without relation\nto any object and consequently without meaning. They are no longer\nconcepts of possible empirical employment, but only of \u201cthings in\ngeneral.\u201d When, for instance, the permanence of existence in time, which\nis the condition of the empirical application of the concept of\nsubstance, is omitted, the category reduces merely to the notion of\nsomething that is always a subject and never a predicate.\n\n     \u201cBut not only am I ignorant of all conditions under which this\n     logical pre-eminence may belong to anything, I can neither put such\n     a concept to any use nor draw the least inference from it. For\n     under these conditions no object is determined for its employment,\n     and consequently we do not at all know whether it signifies\n     anything whatsoever.\u201d[1286]\n\nIn abstraction from sense-data, the categories still remain as concepts\nor _thoughts_, logically possible; but that is not to be taken as\nsignifying that they still continue to possess meaning, _i.e._ reference\nto an object.[1287] And in the absence of ascertainable meaning they\ncannot, of course, be defined.\n\nIn A 244[1288] Kant states his position in somewhat different fashion.\nIn abstraction from sense the categories have meaning, but not\ndeterminate meaning; they relate not to any specific object, but only to\nthings in general. In this latter reference, however, they possess no\nobjective validity, since in the absence of intuition there is no means\nof deciding whether or not any real existence actually corresponds to\nthem.\n\nBut whichever mode of statement be adopted, the same conclusion follows.\n\n     \u201cAccordingly, the transcendental _Analytic_ has this important\n     result, that the most the understanding can achieve _a priori_ is\n     to anticipate the form of a possible experience in general. And\n     since that which is not appearance cannot be an object of\n     experience, the understanding can never transcend those limits of\n     sensibility within which alone objects are given to us. Its\n     principles are merely rules for the exposition of appearances; and\n     the proud title of an Ontology, which presumptuously claims to\n     supply, in systematic doctrinal form, synthetic _a priori_\n     knowledge of things in general (_e.g._ the principle of causality),\n     must therefore give place to the modest claims of a mere Analytic\n     of pure understanding.\u201d[1289]\n\nA 248-9[1290] opens a new line of argument which starts from the results\nobtained in the _Aesthetic_. The proof that space and time are\nsubjective forms establishes the merely phenomenal character of\neverything which can be apprehended in and through them, and is\nmeaningless except on the assumption that things in themselves exist.\nThis assumption, Kant argues, is already involved in the very word\n\u2018appearance,\u2019 and unless it be granted, our thinking will revolve in a\nperpetual circle.[1291] But, he proceeds, this conclusion may easily be\nmisinterpreted. It might be taken as proving the objective reality of\nnoumena, and as justifying us in maintaining a distinction between the\nsensible and the intelligible worlds, and therefore in asserting that\nwhereas the former is the object of intuition, the latter is apprehended\nby the understanding in pure thought. We should then be arguing that\nthough in experience things are known only as they appear, through pure\nunderstanding a nobler world than that of sense, \u201c_eine Welt im Geiste\ngedacht_,\u201d is opened to our view.\n\nBut any such interpretation, Kant insists, runs directly counter to the\nteaching of the _Analytic_, and is ruled out by the conclusions to which\nit has led. Categories yield only \u201crules for the exposition of\nappearances,\u201d and cannot be extended beyond the field of possible\nexperience. It is true that all our sense-representations are related by\nthe understanding to an object that is \u201ctranscendental.\u201d But that\nobject, in its transcendental aspect, signifies only a something = _x_.\nIt cannot be thought apart from the sense-data which are referred to it.\nWhen we attempt to isolate it, and so to conceive it in its independent\nnature, nothing remains through which it can be thought.\n\n     \u201cIt is not in itself an object of knowledge, but only the\n     representation of appearances under the concept of an object in\n     general, viewed as determinable through the manifold of those\n     appearances.\u201d\n\nKant is here again expounding his early doctrine of the transcendental\nobject.[1292] Evidently, at the time at which this passage was written,\nhe had not yet come to realise that such teaching is not in harmony with\nhis Critical principles. It is, as we have seen above, a combination of\nsubjectivism and of dogmatic rationalism.[1293] The very point which he\nhere chiefly stresses was bound, however, when consistently followed\nout, to reveal the untenableness of the doctrine of the transcendental\nobject; and in the second edition Kant so recast this chapter on\nphenomena and noumena as to eliminate all passages in which the\ntranscendental object is referred to.[1294]\n\nBut to return to Kant\u2019s own argument: the reason why the mind is \u201cnot\nsatisfied with this substrate of sensibility,\u201d[1295] and therefore\nproceeds to duplicate the phenomenal world by a second world of noumena,\nlies in the character of the agency whereby sensibility is limited.\nSensibility is limited by the understanding; and the understanding,\noverestimating its powers and prerogatives, proceeds to transform the\nnotion of the transcendental object = _x_ into the concept of a\nnoumenon, viewed in a manner conformable to its etymological\nsignificance, as something apprehended by reason or pure intuition,\n_i.e._ as intuited in some non-sensuous fashion. For only by postulating\nthe possibility of a non-sensuous species of intuition, can the notion\nof a noumenon, thus positively conceived, be saved from\nself-contradiction. Otherwise we should be asserting the apprehension of\nan object independently of appearances, and yet at the same time denying\nthe only means through which such apprehension is possible. Statement of\nthe postulate suffices, however, to reveal its unsupported character. We\nhave no such power of non-sensuous, intuitive apprehension;[1296] nor\ncan we in any way prove that such a power is possible even in a Divine\nBeing. Though, therefore, the concept of noumena is not\nself-contradictory, it involves more than we have the right to assert;\nthe process whereby the empty notion of a transcendental object = _x_ is\ntransformed into the positive concept of a noumenon is easily\ncomprehensible,[1297] but it is none the less illegitimate. We must,\nKant insists, keep strict hold of the central doctrine of Critical\nteaching, namely, that the categories are applicable only to the data of\nsense. We can still employ them as pure logical functions, yielding the\nnotion of objects in general (of the transcendental object = _x_). But\nthis does not widen the sphere of _known_ existences. It only enables us\nto comprehend the limited and merely phenomenal character of the world\nexperienced.\n\nAt this point[1298] Kant\u2019s argument takes a strange and misleading turn.\nThe concept of object in general (the transcendental object = _x_) has\nbeen proved to be involved in the apprehension of appearances as\nappearances, and in this capacity to be a limiting concept\n(_Grenzbegriff_), which, though negative in function, is indispensably\ninvolved in the constitution of human experience. Now, however, Kant\nproceeds to ascribe this function to _the concept of the noumenon_. That\nconcept is, he repeats, purely problematic. Even the mere possibility of\nits object, presupposing as it does the possibility of an understanding\ncapable through non-sensuous intuition of apprehending it, we have no\nright to assert. That the concept is not self-contradictory is the most\nthat we can say of it. None the less, it is to this concept that Kant\nhere ascribes the indispensable limiting function.\n\n     \u201cThe concept of a noumenon is a merely limiting concept, the\n     function of which is to curb the pretensions of sensibility; and it\n     is therefore only of negative employment. At the same time it is\n     no arbitrary invention, and it is bound up with the limitation of\n     sensibility, though it cannot affirm anything positive beyond the\n     field of sensibility.\u201d[1299]\n\nThis confusion, between the concept of a noumenon and the less definite\nconcept of object in general, which is probably due to the combining of\nmanuscripts of different dates, is corrected in the second edition by\nmeans of a new distinction which Kant introduces, evidently for this\nvery purpose. The term noumenon may, he there says,[1300] be used either\npositively or negatively. Taken positively, it signifies \u201can object of a\nnon-sensuous intuition\u201d; regarded negatively, it means only \u201ca thing so\nfar as it is not an object of our sensuous intuition.\u201d Only in its\nnegative employment, he states, is it required as a limiting concept;\nand it is then, as he recognises, indistinguishable from the notion of\nthe unknown thing in itself.\n\nBut despite this variation in mode of expression, in the main Kant holds\nconsistently to his fundamental teaching.\n\n     \u201c...understanding is not limited through sensibility; on the\n     contrary, it itself limits sensibility by applying the term noumena\n     to things in themselves (things not regarded as appearances). But\n     in so doing it at the same time sets limits to itself, recognising\n     that it cannot know these noumena through any of the categories,\n     and that it must therefore think them only under the title of an\n     unknown something.\u201d[1301]\n\nOr as Kant adds in the concluding sentence of this chapter:\n\n     \u201c...the problematic thought which leaves open a place for\n     [intelligible objects], serves only, like an empty space, for the\n     limitation of empirical principles, without itself containing or\n     revealing any other object of knowledge beyond their sphere.\u201d\n\nA sentence in A 258 = B 314 deserves special notice.\n\n     \u201c...we can never know whether such a transcendental or exceptional\n     knowledge is possible under any conditions, least of all if it is\n     to be regarded as of the sort that stands under our ordinary\n     categories.\u201d\n\nThis sentence clearly shows that Kant was willing to recognise that the\ncategories may be inapplicable, not merely owing to lack of data for\ntheir specification, but because of their inherent character. They may\nbe intrinsically inapplicable, expressing only the _modi_ of our\nself-consciousness. They may be merely the instruments of our human\nthinking, not forms necessary to knowledge as such.\n\n\nRELEVANT PASSAGES IN THE SECTION ON AMPHIBOLY\n\nBefore passing to consideration of the extensive alterations made in\nthis chapter in the second edition, it is advisable to take account of\nthe two passages dealing with this problem in the first edition section\non _Amphiboly_: namely, A 277-280 = B 333-6, and A 285-9 = B 342-6. The\nfirst of these passages is of great interest in other connections;[1302]\nits chief importance in reference to the present problem lies in its\nconcluding paragraph. Kant there declares that the representation of an\nobject \u201cas thing in general\u201d is not only, in the absence of specific\ndata, insufficient for the determination of an object, but is\nself-contradictory. For we must either abstract from all reference to an\nobject, and so be left with a merely logical representation; or, in\nassuming an object, we must postulate a special form of intuition which\nwe do not ourselves possess, and which therefore we cannot employ in\nforming our concept of the object. _Here again Kant is substituting the\nconcept of a noumenon for the less definite concept of the thing in\nitself._ This is still more explicitly done in the second passage. The\npure categories are, Kant there declares, incapable of yielding the\nconcept of an object. Apart from the data of sense they have no relation\nto any object. As purely logical functions, they are altogether lacking\nin content or meaning. By objects as things in themselves we must\ntherefore mean objects of a non-sensuous intuition.[1303] Kant still,\nindeed, continues to maintain that to them the categories do not apply,\nand that we cannot, therefore, have any knowledge of them, either\nintuitional or conceptual.\n\n     \u201cEven if we assume a non-sensuous form of intuition, our functions\n     of thought would still have no meaning in reference to it.\u201d[1304]\n\nBut Kant now insists that the notion of noumena, viewed in the above\nmanner, differs from the notion of \u201cobjects in general\u201d (transcendental\n= _x_) in being a legitimate non-contradictory conception; and he also\ninsists that though more positive in content, it is for that very reason\nless open to misunderstanding. Its function is not to extend our\nknowledge, but merely to limit it.\n\n     \u201cFor it merely says that our species of intuition does not extend\n     to all things, but only to objects of our senses; that its\n     objective validity is consequently limited; and that a place\n     therefore remains open for some other species of intuition, and so\n     for things as its objects.\u201d[1305]\n\nThe concept of a noumenon, as thus employed to signify the objects of a\nnon-sensuous intuition, is, Kant proceeds, merely problematic. As we\nhave neither intuition nor (it may be) categories fitted for its\napprehension, it represents something upon the possibility or\nimpossibility of which we are quite unable to pronounce.\n\n     \u201c...as the problematic concept of an object for a quite different\n     intuition and a quite different understanding than ours, it is\n     itself a problem.\u201d\n\nWe may not therefore assert the existence of noumena, but we must none\nthe less form to ourselves the concept of them. This concept is\nindispensably involved in the constitution of our empirical knowledge,\nand is demanded for its proper interpretation. Only when viewed as a\nself-sufficient representation of an absolute existence does it become\ndogmatic and therefore illegitimate. In its Critical aspects it stands\nfor a problem which human reason is constrained by its very nature to\npropound.\n\n     \u201cThe concept of the noumenon is, therefore, not the concept of an\n     object, but is a problem unavoidably bound up with the limitation\n     of our sensibility--the problem, namely, as to whether there may\n     not be objects entirely disengaged from our sensuous species of\n     intuition. This is a question which can only be answered in an\n     indeterminate manner, by saying that, as sense intuition does not\n     extend to all things without distinction, a place remains open for\n     other different objects; and consequently that these latter must\n     not be absolutely denied, though--since we are without a\n     determinate concept of them (inasmuch as no category can serve that\n     purpose)--neither can they be asserted as objects for our\n     understanding.\u201d[1306]\n\nThe fact that these fundamental concepts have not yet been quite\ndefinitely and precisely formulated in Kant\u2019s own mind, appears very\nclearly from the immediately following paragraph. For he there again\nintroduces the concept of the transcendental object, and adds that if\n\u201cwe are pleased to name it noumenon for the reason that its\nrepresentation is not sensuous, we are free so to do.\u201d[1307] The\ncharacterisation given in this paragraph of the transcendental object\ndeserves special notice, for in it Kant goes further in the sceptical\nexpression of his position, though not indeed in the modification of it,\nthan in any other passage.\n\n     \u201d[The understanding in limiting sensibility] thinks for itself an\n     object in itself, but only as transcendental object which is the\n     cause of appearance and therefore not itself appearance, and which\n     can be thought neither as quantity nor as reality nor as substance,\n     etc.... We are completely ignorant whether it is to be met with in\n     us or outside us, whether it would be at once removed with the\n     cessation of sensibility, or whether in the absence of sensibility\n     it would still remain.\u201d[1308]\n\nThis sentence reveals Kant as at once holding unquestioningly to the\nexistence of things in themselves, and yet at the same time as teaching\nthat they must not be conceived in terms of the categories, not even of\nthe categories of reality and existence.\n\n\nALTERATIONS IN SECOND EDITION\n\nIn the second edition certain paragraphs of the chapter on _Phenomena\nand Noumena_ are omitted, and new paragraphs are inserted to take their\nplace. Though these alterations do not give adequate expression to the\nCritical teaching in its maturest form, there are three important\nrespects in which they indicate departures from the teaching of the\nfirst edition. In the first place, those paragraphs in which the\ndoctrine of the transcendental object finds expression are entirely\neliminated, and the phrase \u2018transcendental object\u2019 is no longer\nemployed. This, as we have already noted, is in harmony with the changes\nsimilarly made in the second edition _Transcendental Deduction and\nParalogisms_.[1309]\n\nSecondly, Kant is even more emphatic than in the first edition, that the\ncategories must not be employed save in reference to sense intuitions.\nIn the first edition he still allows that their application to things in\nthemselves is logically possible, though without objective validity. In\nthe second edition he goes much further. Save in their empirical\nemployment the categories \u201cmean nothing whatsoever.\u201d[1310]\n\n     \u201d[In the absence of sensibility] their whole employment, and indeed\n     all their meaning entirely ceases; for we have then no means of\n     determining whether things in harmony with the categories are even\n     possible....\u201d[1311]\n\nIn the third place, Kant, as already noted, distinguishes between a\nnegative and a positive meaning of the term noumenon. Noumenon in its\nnegative sense is defined as being merely that which is not an object of\nsensuous intuition. By noumenon in the positive sense, on the other\nhand, is meant an object of non-sensuous intuition. Kant now claims that\nit is the concept of noumenon in the negative sense, as equivalent\ntherefore simply to the thing in itself, that alone is involved, as a\n_Grenzbegriff_, in the \u201cdoctrine of sensibility.\u201d For its determination\nthe categories cannot be employed; that would demand a faculty of\nnon-sensuous intuition, which we do not possess, and would amount to the\nillegitimate assertion of noumena in the positive sense. The limiting\nconcept, indispensably presupposed in human experience, is therefore the\nbare notion of things in themselves. And accordingly, in modification of\nthe conclusion arrived at in the first edition, viz. that \u201cthe division\nof objects into phenomena and noumena ... is not in any way\nadmissible,\u201d[1312] Kant now adds to the term noumena the qualifying\nphrase \u201cin the positive sense.\u201d In this way the assumption that things\nin themselves actually exist becomes quite explicit, despite Kant\u2019s\ngreater insistence upon the impossibility of applying any of the\ncategories to them.\n\nBut beyond thus placing in still bolder contrast the two counter\nassertions, on the one hand that the categories must not be taken by us\nas other than merely subjective thought-functions, and on the other that\na limiting concept is indispensably necessary, Kant makes no attempt in\nthese new passages to meet the difficulties involved. With the assertion\nthat the categories as such, and therefore by implication those of\nreality and existence, are inapplicable to things in themselves,[1313]\nhe combines, without any apparent consciousness of conflict, the\ncontention that things in themselves must none the less be postulated as\nactually existing.\n\n       *       *       *       *       *\n\nThe teaching of this chapter must be regarded as only semi-Critical. The\nfact that it is formulated in terms of the doctrine of the\ntranscendental object, itself suffices to determine the date at which it\nmust have been composed as comparatively early; and such changes as Kant\ncould make in the second edition were necessarily of a minor character.\nMore extensive alterations would have involved complete reconstruction\nof the entire chapter, and indeed anticipation of the central teaching\nof the _Dialectic_.\n\nKant is also hampered by the unfortunate location to which he has\nassigned this chapter. At this point in the development of his argument,\nnamely, within the limits of the _Analytic_, Kant could really do no\nmore than recapitulate the _negative_ consequences which follow from the\nteaching of the transcendental deduction. For though these might justify\nhim in asserting that it is understanding that limits sensibility, he\nwas not in a position to explain that the term understanding, as thus\nemployed, has a very wide meaning, and that within this faculty he is\nprepared to distinguish between understanding in the strict sense as the\nsource of the categories, and a higher power to which he gives the title\nReason, and which he regards as originating a unique concept, that of\nthe unconditioned. Yet only when these distinctions, and the\nconsiderations in view of which they are drawn, have been duly reckoned\nwith, can the problem before us be discussed in its full significance.\n\nThis placing of the chapter within the _Analytic_, and therefore prior\nto the discussions first broached in the _Dialectic_, has indeed the\nunfortunate consequence of concealing not only from the reader, but\nalso, as it would seem, to some extent from Kant himself, the ultimate\ngrounds upon which, from the genuinely Critical standpoint, the\ndistinction between phenomena and noumena must be based. For neither in\nthis chapter, nor in any other passage in the _Critique_, has Kant\nsought to indicate, in any quite explicit manner, the bearing which the\nimportant conclusions arrived at in the _Dialectic_ may have in regard\nto it. Like so many of the most important and fruitful of his tenets,\nthese consequences are suggested merely by implication; or rather remain\nto be discovered by the reader\u2019s own independent efforts, in proportion\nas he thinks himself into the distinctions upon which, in other\nconnections, Kant has himself insisted. They are never actually\nformulated in and by themselves.\n\nIn seeking, therefore, to decide upon what basis the distinction between\nappearance and reality ought to be regarded as resting, we are\nattempting to determine how the argument of this chapter would have\nproceeded had it been located at the close of the _Dialectic_. The task\nis by no means easy, but the difficulties are hardly as formidable as\nmay at first sight appear. The general outlines of the argument are\nfairly definitely prescribed by Kant\u2019s treatment of kindred questions,\nand may perhaps, with reasonable correctness, be hypothetically\nconstructed in view of the following considerations.\n\nJust as Kant started from the natural assumption that reference of\nrepresentations to objects must be their reference to things in\nthemselves, so he similarly adopted the current Cartesian view that it\nis by an inference, in terms of the category of causality, that we\nadvance from a representation to its external ground. It was very\ngradually, in the process of developing his own Critical teaching, and\nespecially his phenomenalist view of the empirical world in space, that\nhe came to realise the very different position to which he stood\ncommitted.[1314] When the doctrine of the transcendental object is\neliminated from his teaching, and when the categories, including that of\ncausality, are pre-empted for the _empirical_ object, and that object is\nregarded as directly apprehended, the function of mediating the\nreference of phenomenal nature to a noumenal basis falls to the Ideas of\nReason. For the distinction is no longer between representations and\ntheir noumenal causes, but between the limited and relative character of\nthe entire world in space and time, and the unconditioned reality which\nReason demands for its own satisfaction. To regard the world in space as\nmerely phenomenal, because failing to satisfy our standards of genuine\nreality, is to adopt an entirely different attitude from any to be found\nin Descartes or Locke. The position may be outlined in the following\nmanner, in anticipation of its more adequate statement in connection\nwith the problems of the _Dialectic_.\n\nThe concept, whereby Reason limits sensibility, is not properly\ndescribable as being that of the thing in itself; it is the unique\nconcept of the unconditioned. Our awareness of the conditioned as being\nconditioned presupposes, over and above the categories, an antecedent\nawareness of Ideal[1315] standards; and to that latter more fundamental\nform of consciousness all our criteria of truth and reality are\nultimately due. The criteria by means of which we empirically\ndistinguish sense-appearance from sense-illusion, when rigorously\napplied, lead us to detect deficiencies in the empirical as such. We\nhave then no alternative save to conceive absolute reality in terms of\nthe rational Ideals, of which the empirical criteria are merely\nspecialised forms.\n\nThere are thus two distinct, but none the less interdependent, elements\ninvolved in Kant\u2019s more mature teaching, phenomenalism, and what may be\ncalled the Idealist, or absolutist, interpretation of the function of\nReason. Each demands the other for its own establishment. There must be\na genuinely objective world, by reflection upon which we may come to\nconsciousness of the standards which are involved in our judgments upon\nit; and we must possess a faculty through which our consciousness of\nthese standards may be accounted for. The standards of judgment cannot\nbe acquired by means of judgments which do not already presuppose them;\nthe processes by which they are brought to clear consciousness cannot be\nthe processes in which they originate. They must be part of the _a\npriori_ conditions of experience and combine with space, time and the\ncategories to render experience of the kind which we\npossess--self-transcending and self-limiting--actually possible.\n\nFrom this point of view the distinction between appearance and reality\nis not a contrast between experience and the non-experienced, but a\ndistinguishing of factors, which are essential to all experience, and\nthrough which we come to consciousness of an irresolvable conflict\nbetween the Ideals which inspire us in the acquisition of experience,\nand the limiting conditions under which alone experience is attainable\nby us. In the higher field of Reason, as in the lower field of\nunderstanding, it is not through the given, but only through the given\nas interpreted by conditioning forms of an Ideal nature, that a\nmeaningful reality can disclose itself to the mind. The ultimate meaning\nof experience lies in its significance when tested by the standards\nwhich are indispensably involved in its own possibility. That meaning is\nessentially metaphysical; more is implied in experience than the\nexperienced can ever itself be found to be.[1316]\n\nSuch is the central thesis of the Critical philosophy, when the teaching\nof the _Analytic_ is supplemented by that of the _Dialectic_. Though the\n_Critique_ is, indeed, the record of the manifold ways in which Kant\ndiverged from this position, not a systematic exposition of its\nimplications and consequences, the above thesis represents the goal upon\nwhich his various lines of thought tend to converge. It is the guiding\nmotive of his devious and complex argument in the three main divisions\nof the _Dialectic_. On no other interpretation can the detail of his\nexposition be satisfactorily explained.\n\nThere are two chief reasons why Kant failed to draw the above\nconclusions in any quite explicit manner. One reason has already been\nsufficiently emphasised, namely, that the thesis, which I have just\nformulated, rests upon a phenomenalist view of the natural world,\nwhereas the _Dialectic_ is inspired by the earlier, subjectivist\ndoctrine of the transcendental object. Upon the other main reason I\nshall have frequent occasion to insist. As we shall find, Kant was\nunable to arrive at any quite definitive decision as to the nature of\nthe Ideals of Reason. He alternates between the sceptical and the\nabsolutist view of their origin and function, and in the process of\nseeking a comprehensive mid-way position which would do justice to all\nthat is valid in the opposing arguments, the further question as to the\nbearing of his conclusions upon the problem of the distinction between\nappearance and reality was driven into the background. But we are\nanticipating matters the discussion of which must meantime be deferred.\n\n\n\n\nAPPENDIX\n\nTHE AMPHIBOLY OF THE CONCEPTS OF REFLECTION[1317]\n\n\nIn this appendix Kant gives a criticism of the Leibnizian rationalism--a\ncriticism already partially stated in the section on the\n_Postulates_--and he does this in a manner which very clearly reveals\nthe influence which that rationalism continued to exercise upon his own\nthinking. Thus Kant speaks of the \u201cmere concept,\u201d[1318] and in doing so\nevidently means to imply that it exists in its own right, with a nature\ndetermined solely by intrinsic factors of a strictly _a priori_\ncharacter, in complete independence of the specific material of\nsense-experience. He denies, it is true, the objective validity of such\nconcepts, and maintains that in their empirical employment they are\ncompletely transformed through the addition of new factors. None the\nless he allows to the concepts an intrinsic nature, and practically\nmaintains that from the point of view of the pure concept, and therefore\nfrom the point of view of a logic based upon it, the Leibnizian\nrationalism is the one true system of metaphysics. For pure thought,\nLeibniz\u2019s system is the ultimate and only possible philosophy; and were\nthought capable of determining the nature of things in themselves, we\nshould be constrained to adopt it as metaphysically valid. This is the\nstandpoint which underlies much of Kant\u2019s argument in the _Dialectic_.\nIt leads him to maintain that the self must necessarily, in virtue of an\nunavoidable transcendental illusion, believe in its own independent\nsubstantial reality, that the mind is constrained to conceive reality as\nan unconditioned unity, and that the notions of God, freedom, and\nimmortality are Ideas necessarily involved in the very constitution of\nhuman thought.\n\nBut we must not regard Kant\u2019s doctrine of the pure concept merely as a\nsurvival from a standpoint which the Critical teaching is destined to\ndisplace and supersede. For Kant is not led through inconsistency, or\nthrough any mere lack of thoroughness in the development of his Critical\nprinciples, to retain this rationalistic doctrine. To understand the\nreally operative grounds of Kant\u2019s argumentation, and so to place the\ncontents of this section in proper focus, we must recall the fundamental\nantithesis, developed in my introduction,[1319] between the alternative\npositions, which are represented for Kant by the philosophies of Hume\nand Leibniz. Kant, as already observed, is profoundly convinced of the\nessential truth of the Leibnizian position. He holds to the Leibnizian\nview of reason. Human reason is essentially metaphysical; its ultimate\nfunction is to emancipate us from the limiting conditions of animal\nexistence; it reveals its nature in those Ideas of the unconditioned,\nthe discussion of which Kant reserves for the _Dialectic_.\n\nThe chief defect in Kant\u2019s criticism of Leibniz, as developed in this\nsection, is that the deeper issues, which determine the extent of his\nagreement with Leibniz, are not raised or even indicated. Consequently,\nhis references to pure thought, and his assertion[1320] that from the\npoint of view of pure thought Leibniz is entirely justified in his\nteaching, bewilder the reader, who has been made to adopt a Critical\nstandpoint, and therefore to believe that thought can function only in\nconnection with the data of sense-experience. Kant would seem, indeed,\nto have lapsed into the dogmatic standpoint of the _Dissertation_,\ndistinguishing between a sensible and an intelligible world, and\nmaintaining that pure thought is capable of determining the nature of\nthe latter. The only difference between his teaching here and in the\n_Dissertation_ consists in the admission that all knowledge is limited\nto sense-experience, and that we are therefore unable to determine\nwhether this intelligible world which we must _think_, and _think in the\nprecise manner defined by Leibniz_, does or does not exist.\n\nThis section is, indeed, like the chapter on _Phenomena and Noumena_,\nwrongly located. Giving, as it does, Kant\u2019s criticism of the Leibnizian\nontology, it discusses problems of metaphysics; and ought therefore to\nhave found its place in the _Dialectic_, in natural connection with the\ncorresponding examination of the metaphysical sciences of rational\npsychology, cosmology, and theology. Architectonic, that ever-present\nsource of so many of Kant\u2019s idiosyncrasies, has again interposed its\ndespotic mandate. As there are only three forms of syllogism, only three\nmain divisions can be recognised in the _Dialectic_; and the criticism\nof ontology, to its great detriment, must therefore be located, where\nit does not in the least belong, in the concluding section of the\n_Analytic_.[1321]\n\nBut we must follow Kant\u2019s argument as here given. Leibniz views thought\nas capable of prescribing, antecedently to all experience, the\nfundamental conditions to which reality must conform. The possible is\nprior to, and independent of, the actual; and can be adequately\ndetermined by pure reason from its own inherent resources. Kant does not\nhere question this assertion of the independence and priority of pure\nthought. He is content to maintain that what is valid for thought need\nnot hold of those appearances which are the only possible objects of\nhuman knowledge, since in sense-experience conditions, unforeseen by\npure thought, partly limitative and partly extensive of its concepts,\nintervene to modify the conclusions which from its own point of view are\nlogically valid. Leibniz, through failure to realise the dual character\nof thought and sense, overlooked this all-important fact; and, in\nasserting that what is true for pure thought is valid of the sensuously\nreal, fell victim to the fallacy which Kant entitles transcendental\namphiboly.\n\nKant\u2019s clearest statement of the fallacy is in A 280 = B 336. It\nreduces, formally stated, to the fallacy of denying the antecedent. In\naccordance with the _dictum de omni et nullo_, we can validly assert\nthat what belongs to or contradicts a universal concept, belongs to or\ncontradicts the particulars which fall under that concept. Leibniz\nemploys the principle in a negative and invalid form. He argues that\nwhat is _not_ contained in a universal concept is also not contained in\nthe particulars to which it applies. \u201cThe entire intellectualist system\nof Leibniz is reared upon this latter principle.\u201d And as Kant points\nout,[1322] the reason why so acute and powerful a thinker succumbed to\nthis obvious fallacy is to be found in his view of sense as merely\nconfused thought; or, to state the same point in another way, in his\ninterpretation of appearances as being the confused representations of\nthings in themselves.[1323] All differences between appearance and\nreality are, on this view, due merely to lack of clearness in our\napprehension of the given. Sense, when completely clarified, reduces\nwithout remainder to pure thought; and in the concepts, which thought\ndevelops from within itself, lie the whole content alike of knowledge\nand of real existence. Owing to a metaphysical theory of the nature of\nthe real, itself due to a false interpretation of the nature and\nfunction of pure thought, and ultimately traceable to an excessive\npreoccupation with knowledge of the strictly mathematical type,[1324]\nLeibniz failed to do justice to the fundamental characteristics of our\nhuman experience, and in especial to the actual given nature of space,\ntime, and dynamical causality. His rationalistic metaphysics has its\nroots in the Cartesian philosophy,[1325] and is, in Kant\u2019s view, the\nperfected product of philosophical thinking, when developed on dogmatic,\n_i.e._ non-Critical, lines. It is the opposite counterpart of the\nempirical or sceptical type of philosophy which in modern times found\nits first great supporter in Locke, and which, as Kant held, obtained\nits perfected expression in the philosophy of Hume. While Descartes and\nLeibniz intellectualise appearances, Locke and Hume regard the _a\npriori_ concepts of understanding as merely empirical products of\ndiscursive reflection. Both commit the same fundamental error of failing\nto recognise that understanding and sensibility are two distinct sources\nof representations.[1326] Both consequently strive, in equally one-sided\nfashion, to reduce the complexity of experience to one alone of its\nconstituent elements. This section of the _Critique_ ought to have\ndeveloped the Critical teaching in its opposition to both these\nalternative attitudes; Kant arbitrarily limits it to criticism of the\nLeibnizian rationalism.\n\nKant\u2019s method of introducing and arranging his criticism is artificial,\nand need be no more than mentioned. Critical reflection upon the sources\nof our knowledge, which Kant, in order to distinguish it from reflection\nof the ordinary type, entitles _transcendental_ reflection, is, he\nstates, a duty imposed upon all who would profess to pass _a priori_\njudgments upon the real. It will trace the concepts employed to their\ncorresponding faculties, intellectual and sensuous, and will reveal the\nindependence and disparity of sensibility and understanding, and so will\neffectually prevent that false locating of concepts to which\ntranscendental amphiboly is due. Such reflection, he further argues,\nconsists in a _comparison_ of the representations with the faculty to\nwhich they are due, and like ordinary comparison will determine the\nrelations of (1) _identity and difference_, (2) _agreement and\nopposition_, (3) _inner and outer_, (4) _determinable and determining_\n(matter and form). In this arbitrary but ingenious fashion Kant\ncontrives to obtain the four main headings required for his criticism of\nthe Leibnizian ontology.\n\n(1) Under the first heading he deals with the principle of the identity\nof indiscernibles. It is, Kant maintains, a typical example of the\nfallacy of transcendental amphiboly. Leibniz argues that if no\ndifference is discoverable in the concept of things, there can be none\nin the things themselves; things which are identical in conception must\nbe identical in all respects. But this, Kant replies, is true only so\nlong as the concepts abstract from the sensuous conditions of existence.\nThus no two cubic feet of space are alike. They are distinguishable from\none another by their spatial location; and that is a difference which\nconcerns the conditions of intuition; it is not to be discovered in the\npure concept.[1327] Spaces, alike for thought, are distinguishable for\nsense. To take another of Kant\u2019s illustrations: two drops of water, if\nindistinguishable in all their internal properties of quality or\nquantity, are conceptually identical. Through differences of location in\nspace, irrelevant to their conception, they can none the less be\nintuited as numerically different. The principle of indiscernibles is\nnot a law of nature, but only an analytic rule for the comparison of\nthings through mere concepts.[1328]\n\n(2) A second principle of the Leibnizian metaphysics is that realities\ncan never conflict with one another. This is supposed to follow from the\nfact that in pure thought the only form of opposition is logical\nnegation. Realities, being pure affirmations, must necessarily harmonise\nwith one another. This principle ignores the altogether different\nconditions of sense-existence. Space, time, and the resulting\npossibility of dynamical causality supply the conditions for real\nopposition. Two existences, though equally real and positive, may annul\none another. Two forces acting upon a body may neutralise one another.\nFrom the above logical principle Leibniz\u2019s successors[1329] profess to\nobtain the far-reaching metaphysical conclusions, that all realities\nagree with one another, that evil is merely negative, consisting\nexclusively in limitation of existence, and that God, without detriment\nto the unity of his being, can be constituted of all possible realities.\n\n(3) Viewing space and time, which condition external relation, as merely\nconfused forms of apprehension, Leibniz further concluded that the\nreality of substance is purely internal. And ruling out position, shape,\ncontact and motion, all of which involve external relations, he felt\njustified in endowing the monads with the sole remaining form of known\nexistence, namely consciousness. The assertion that the monads are\nincapable of external relation leads to the further conclusion that\nthey are incapable of interaction, and stand in systematic relation to\none another, solely in virtue of a pre-established harmony.\n\n(4) From the point of view of pure thought matter must precede form. The\nuniversal must precede the particular which is a specification of\nit.[1330] Unlimited reality is taken as being the matter of all\npossibility, and its limitation or form as being due to negation.\nSubstances must antecedently exist in order that external relations may\nhave something upon which to ground themselves. Space and time must be\ninterpreted as confused apprehensions of purely intellectual orders,\nspace representing a certain order in the reciprocal (pre-established)\ncorrespondence of substances, and time the dynamical sequence of their\nstates. On the other hand, from the standpoint of sense and its\nintuitional forms the reverse holds. The world of appearance is\nconditioned by the forms of space and time; the objectively possible\ncoincides with the actual; and the _substantia phaenomenon_ has no\nindependent essence, but reduces without remainder to external\nrelations. For pure thought this world of given appearance is an utterly\nparadoxical form of existence; it is the direct opposite of everything\nthat genuine reality ought to be. In this strange conclusion the\nproblems of the _Dialectic_, in one of their most suggestive forms, at\nonce loom up before us. As stated above, this entire discussion is an\nanticipation of questions which cannot be adequately treated within the\nlimits of the _Analytic_.\n\nThe text of this section is highly composite. The entire content of the\n_Appendix_ is twice reintroduced and restated at full length in the\naccompanying _Note_. These successive expositions of one and the same\nargument were doubtless independently written, and then later pieced\ntogether in this external fashion. A 277-8 = B 333-4, on the nature of\nthe _substantia phaenomenon_, would by its references to the\ntranscendental object seem to be of early origin.[1331] It has already\nbeen commented upon.[1332] A 285-9 = B 342-6, on the other hand, which\nsupplements the chapter on _Phenomena and Noumena_,[1333] would seem to\nbe of late origin. It is so dated by Adickes,[1334] owing to the\nreference to schemata in its opening sentence.\n\n=A 289-91 = B 346-9. Table of the division of the conception of\nnothing.=--This curious and ingenious classification of the various\nmeanings of the term \u2018nothing\u2019 is chiefly of interest through its first\ndivision: \u201cempty conception without object, _ens rationis_.\u201d The _ens\nrationis_ can best be defined in its distinction from the fourth\ndivision: \u201cempty object without conception, _nihil negativum_.\u201d The\nformer is a _Gedankending_; the latter is an _Unding_. The former\nindeed, though not contradictory, is mere fiction (_bloss Erdichtung_),\nand consequently must not be taken as falling within the field of the\npossible. The latter is a concept which destroys itself, and which\ntherefore stands in direct conflict with the possible. The _ens\nrationis_ includes, Kant explicitly states,[1335] the conception of\nnoumena, \u201cwhich must not be reckoned among the possibilities, although\nthey must not for that reason be declared to be also impossible.\u201d Kant\nmust here be taking noumena in the positive sense.[1336] As usual Kant\u2019s\nattempt to obtain parallels for the four classes of category breaks\ndown. The so-called _nihil privativum_ and the _ens imaginarium_ do not\nproperly come within the denotation of the term \u2018nothing.\u2019 This is very\nevident in the examples which Kant cites. Cold is as real as the\nopposite with which it is contrasted, while pure space and pure time are\nnot negative even in a conventional sense.\n\n\n\n\nTRANSCENDENTAL LOGIC\n\n\n\nDIVISION II\n\nTHE TRANSCENDENTAL DIALECTIC\n\n\nINTRODUCTORY COMMENT UPON THE COMPOSITE ORIGIN AND CONFLICTING\nTENDENCIES OF THE DIALECTIC.\n\n\nWe have had constant occasion to observe the composite origin and\nconflicting tendencies of the _Analytic_. The _Dialectic_ is hardly less\ncomposite in character, and is certainly not more uniform in its\nfundamental teaching.\n\nThe composite nature of the text, though bewildering to the\nunsophisticated reader, is not, however, without its compensations. The\ntext, as it stands, preserves the record of the manifold influences\nwhich presided over its first inception, and of the devious paths by\nwhich Kant travelled to his later conclusions. It thus enables us to\ndetermine, with considerable accuracy, the successive stages through\nwhich it has passed in the process of settling into its present form. As\nwe shall find, the sections on the antinomies contain the original\nargument, out of which by varied processes of supplementation and\nmodification the other parts have arisen.\n\nThe conflict of doctrine has also its counter-advantages. The problems\nare impartially discussed from opposed standpoints; the difficulties\npeculiar to each of the competing possible solutions are frankly\nrecognised, and indeed insisted upon; and the internal dialectic of\nKant\u2019s own personal thinking obtains dramatic expression. We are thus\nthe better enabled to appreciate the open-minded pertinacity with which\nKant set himself to do justice to every significant aspect of his\nmany-sided problems, and are consequently in less danger of simplifying\nhis argument in any arbitrary manner, or of ignoring the tentative\ncharacter of the solutions at which he arrives.\n\nI shall first define the main lines of conflict, and shall then attempt\nto trace those conflicts to the considerations in which they have their\nsource. The two chief lines of thought traceable throughout the\n_Dialectic_ are represented by its negative and by its positive\ntendencies respectively. From one point of view, Reason is merely the\nunderstanding in its self-limiting, self-regulative employment, and the\nmain purpose of the _Dialectic_ is to guard against the delusive power\nof fictitious principles. From the other point of view, Reason is a\nfaculty distinct from understanding, and its problems run parallel with\nthose of the _Analytic_, forming no less important a subject of\nphilosophical reflection, and no less fruitful a source of positive\nteaching. The one line of argument connects with Kant\u2019s more sceptical\ntendencies, the other with his deep-rooted belief in the ultimate\nvalidity of the absolute claims of pure thought.\n\nWhen we approach the _Dialectic_ from the standpoint of the _Analytic_,\nit is the negative aspect that is naturally most prominent. In the\n_Analytic_ Kant has proved that all knowledge is limited to\nsense-experience, and that a metaphysical interpretation of reality is\naltogether impossible. But as the human mind would seem to be possessed\nby an inborn need of metaphysical construction, this conclusion cannot\nobtain its due influence until the sources of the metaphysical tendency\nhave been detected and laid bare. The _Dialectic_ must yield _a\npsychology of metaphysics_ as well as a _logic of illusion_.\n\nBut when, on the other hand, the problems of the _Dialectic_ are viewed\nin their _distinction_ from those of the _Analytic_, and their\nindependent character is recognised, they appear in a perspective which\nsets them in a very different light. Reason is a faculty co-ordinate\nwith understanding, and yields _a priori_ concepts distinct in function,\nno less than in nature, from the categories. To mark this distinction\nKant entitles the concepts of Reason Ideas. They demand both a\nmetaphysical and a transcendental deduction. These requirements are\nfulfilled through their derivation from the three forms of syllogism,\nand by the proof that they exercise an indispensable function, at once\nlimiting and directing the understanding. As limiting concepts, they\ncondition the consciousness of those Ideal standards through which the\nhuman mind is enabled to distinguish between appearance and things in\nthemselves. As regulative, they prescribe the problems which the\nunderstanding in its search for knowledge is called upon to solve.\n\nThese two tendencies, sceptical and constructive, are never, indeed, in\ncomplete opposition. Common to both, rendering possible the\npsychological explanation of the metaphysical impulse, which even the\nnegative standpoint demands, is the doctrine of the regulative function\nof Ideal principles. This doctrine, which already appears in the\n_Dissertation_ of 1770, was later developed into the Critical theory of\ntranscendental illusion; and by means of that theory Kant succeeded in\nbringing the two standpoints into a very real and vital connection with\none another. At first sight it may seem to achieve their complete\nreconciliation, accounting for their distinction while rendering them\nmutually complementary; and Kant\u2019s teaching may perhaps be so restated\nas to bear out that impression. But the harmony is never completely\nattained by Kant. Here, as in the _Analytic_, there is an equipoise of\ntendencies that persist in opposition.\n\nKant\u2019s mediating doctrine of transcendental illusion may first be\nstated. It rests upon a distinction between appearance and illusion.\nAppearance (_Erscheinung_) is a transcript in phenomenal terms of some\nindependent reality; and of such appearances we can acquire what from\nthe human point of view is genuine knowledge. On the other hand, all\nprofessed insight into the nature of the transcendent or non-empirical\nis sheer illusion (_Schein_), and purely subjective. There are three\nspecies of illusion, logical, empirical and transcendental. Logical\nillusion stands apart by itself. It is due merely to inattention or\nignorance; and vanishes immediately the attention is aroused. Empirical\nand transcendental illusion, on the other hand, have a twofold point of\nagreement, first, in being unavoidable, and secondly, in that they\noriginate in our practical needs. We may know that the moon at its\nrising is no larger than in mid-heavens, that the ocean is no higher in\nthe distance than at the shore; this makes not the least difference in\nthe perceptions as they continue to present themselves. That the\nillusions are adapted to our practical needs, and are consequently\nbeneficial, is less often observed. Changes in the colour, form, and\nsize of objects as they recede from us, the seeing of the parallel sides\nof a street as converging, enable us to achieve what would not otherwise\nbe possible. By their means we acquire the power of compressing a wide\nextent of landscape into a single visual field, of determining distance,\nand the like. Their practical usefulness is in almost exact proportion\nto the freedom with which they depart from the standards of the\nindependently real. Kant argues that, in these respects, transcendental\nillusion is analogous to the empirical. Just as the illusory\ncharacteristics of our perceptions are to be understood only in terms of\ntheir practical function, so the Ideas of pure Reason have always a\npractical bearing, and can only be explained and justified in terms of\nthe needs which they satisfy. As theoretical enquirers, we accept all\nthat affords us orientation in the attainment of knowledge; as moral\nagents, we postulate the conditions which are necessary for the\nrealisation of the moral imperative. And as the Ideals of natural\nscience are found (such is Kant\u2019s contention) to be in general form akin\nto those of the moral consciousness, they thus acquire a twofold footing\nin the mental life, maintaining their place there quite independently of\ntheoretical proof. Though illusory, they are unavoidable; and though\ntheoretically false,[1337] they are from a practical point of view both\nlegitimate and indispensable.\n\nKant, in developing this thesis, might profitably have pointed to still\nanother respect in which the analogy holds between sense-experience and\ntranscendental beliefs. The illusions of sense-perception come in the\nordinary processes of experience to be detected as such by the mind.\nFrom the theoretical standpoint of the outside observer who compares the\nsituation of one percipient with that of another, and so is enabled to\ncancel the differences which variety of situation carries with it, the\nuseful illusions of ordinary experience are reduced to the level of mere\nappearance. In contradicting one another they reveal their subjective\ncharacter, and also at the same time afford data for determining the\nobjective conditions to which their subjectively necessary existence is\ncausally due. In similar fashion the transcendental illusions result in\ncontradictions, which compel the mind to recognise that the Ideals to\nwhich it is committed by its practical needs are of a merely subjective\ncharacter, and may never be legitimately interpreted as representing the\nactual nature of the independently real.\n\nThe chief transcendental illusion, and ultimately the cause of all the\nothers, consists in the belief that the Ideals of explanation which\nsatisfy Reason must in general outline represent the nature of ultimate\nreality. What the individual seeks to discover he naturally believes to\nexist prior to the discovery. As practical beings, we regard the objects\nof sense-experience as absolute realities--they are the realities of\npractical life, and we are practical rather than theoretical beings--and\nthe existing empirical sciences, conceived as Ideally completed, are\ntherefore viewed as yielding an adequate representation of ultimate\nreality. But such a belief involves us in contradictions. The world of\nphenomena in space and time is endlessly relative. It can have no outer\nbounds or first beginning, and no smallest parts; and in the series of\ncausal antecedents there can be no member that is not effect as well as\ncause. Viewed as representing a pre-existent goal, the Ideas of Reason\nare imaginary completions of the intrinsically and merely relative, and\nare in their very notion self-contradictory. All that is definite in\ntheir content conflicts with their absoluteness; and yet, as it would\nseem, only in their empirical reference can they hope for objective\nverification.\n\nSuch are the problems of the _Dialectic_, so far as they can be\nformulated in terms common to the two opposed standpoints. Their deeper\nsignificance, and the grounds of Kant\u2019s alternating treatment of them,\nonly appear when he raises the further questions, what those Ideals of\nexplanation which Reason prescribes really are, and how, if they\nconflict with the content of experience, it is possible that they should\nbe conceived at all. To these questions Kant propounds both a sceptical\nand an Idealist answer. The former, in bare outline, may be stated as\nfollows. The so-called Ideas are based upon experience and are derived\nfrom it. The understanding removes the limitations to which its pure\nconcepts are subject in sense-experience, and proceeds to use them in\ntheir widest possible application, _i.e._ to things in general. As thus\nemployed, they are without real significance, and are indeed\nself-contradictory. To form the Idea of the unconditioned, we have to\nomit all those conditions through which alone anything can be\napprehended, even as possible. To construct the concept of absolute or\nunconditioned necessity, we have similarly to leave aside the conditions\nupon which necessity, as revealed in experience, in all cases depends;\nin eliminating conditions, we eliminate necessity in the only forms in\nwhich it is conceivable by us. Such Ideas are, indeed, simply _schematic\nforms_, whereby we body forth to ourselves, in more or less metaphorical\nterms, the concept of a _maximum_. They are imaginary extensions, in\nIdeal form, of the unity and system which understanding has discovered\nin actual experience, and which, under the inspiration of such Ideals,\nit seeks to realise in ever-increasing degree. If the understanding, as\nthus insisting upon Ideal satisfaction, be entitled Reason, the Ideas\nmust be taken as expressing a subjective interest, and as exhausting\ntheir legitimate employment in the regulation of the understanding.\nTheir transcendental deduction will consist in the proof that they are\nnecessary to the understanding for the _perfecting_ of its experience.\nThey do not justify us in attempting to decide, in anticipation of\nactual experience, how far the contingent collocations and the\ninexhaustible complexities of brute experience are really reducible to\na completely unified system; but they quite legitimately demand that\nthrough all discouragements we persist in the endeavour towards their\nrealisation. In any case, it is by experience that the degree of their\nreality has to be decided. We judge of things by the standard of that\nfor which they exist, and not _vice versa_. As the sole legitimate\nfunction of the Ideas is that of inspiring the understanding in its\nempirical employment, they must never be interpreted as having\nmetaphysical significance. As the Ideas exist solely for the sake of\nexperience, it is they that must be condemned, if the two really\ndiverge. We do not say \u201cthat a man is too long for his coat, but that\nthe coat is too short for the man.\u201d[1338] It is experience, not Ideas,\nwhich forms the criterion alike of truth and of reality.\n\nKant\u2019s teaching, when on Idealist lines, is of a very different\ncharacter. Reason is distinct from understanding, and yet is no less\nindispensably involved in the conditioning of experience. All\nconsciousness is consciousness of a whole which precedes and conditions\nits parts. Such consciousness cannot be accounted for by assuming that\nwe are first conscious of the conditioned, and then proceed through\nomission of its limitations to form to ourselves, by means of the more\npositive factors involved in this antecedent consciousness, an Idea of\nan unconditioned whole. The Idea of the unconditioned is distinct in\nnature from all other concepts, and cannot be derived from them. It must\nbe a pure _a priori_ product of what may be named the faculty of Reason.\nIts uniqueness is what causes its apparent meaninglessness. As it is\ninvolved in all consciousness, it conditions all other concepts; and\ncannot, therefore, be defined in terms of them. Its significance must\nnot be looked for save in that Ideal, to which no experience, and no\nconcept other than itself, can ever be adequate. That in this Ideal form\nit has a very real and genuine meaning is proved by our capacity to\ndistinguish between appearance and reality. For upon it this\ndistinction, in ultimate analysis, is found to rest. Consciousness of\nlimitation presupposes a consciousness of what is beyond the limit;\nconsciousness of the unconditioned is prior to, and renders possible,\nour consciousness of the contingently given. The Idea of the\nunconditioned must therefore be counted as being, like the categories,\nthough in a somewhat different manner, a condition of the possibility of\nexperience. With it our standards both of truth and of reality are\ninextricably bound up.[1339] The Ideas in which it specifies itself, so\nfar from depending upon empirical verification, are the touchstone by\nwhich we detect the unreality of the sensible world, and by which a\ntruer reality, such as would be adequate to the Ideal demands of pure\nReason, is prefigured to the mind.\n\nThese two standpoints are extremely divergent in their consequences.\nEach leads to a very different interpretation of the content of the\nIdeas, of their function in experience, and of their objective validity.\nOn the one view, their content is merely empirical, and sense-experience\nis our sole criterion of truth and reality; on the other, they have to\nbe recognised as containing a pure _a priori_ concept, and are\nthemselves the standards by which even empirical truth can alone be\ndetermined. In the one case, they are Ideals projected by experience for\nits own empirical guidance; they are built upon contingent experience,\nand depend upon it alike for the content which makes them conceivable\nand for their validity. In the other, they are presuppositions of\nexperience, at once conditioning its possibility and revealing its\nmerely phenomenal character. According to the sceptical view, Reason is\nconcerned only with itself and its own subjective demands; on the\nIdealist view, it is a metaphysical faculty, and outlines possibilities\nthat may perhaps be established by practical Reason.\n\nSuch, in broad outline, are the central doctrines of the _Dialectic_.\nThey constitute an extraordinarily stimulating and suggestive body of\nCritical teaching. In no other division of the _Critique_ do the power\nand originality of Kant\u2019s thinking gain such abundant, forceful and\nilluminating expression. The accumulated results of the painstaking\nanalyses of the earlier sections contribute a solidity and fulness of\nmeaning, which render the argument extremely impressive, even to those\nwho are out of sympathy with Kant\u2019s ultimate purposes. Its persistent\ninfluence, on sceptical no less than on Idealist lines, and often\nconveyed by very devious channels, can frequently be detected even in\nthinkers--Herbert Spencer is an instance--who would indignantly\nrepudiate the charge of being indebted to such a source.\n\n\nTHE HISTORY AND DEVELOPMENT OF KANT\u2019S VIEWS IN REGARD TO THE PROBLEMS OF\nTHE DIALECTIC[1340]\n\nWe may now proceed to consider the evidence in support of the early\norigin of the central portions of the _Dialectic_--the sections on the\nantinomies. As Benno Erdmann[1341] has very conclusively shown,\npreoccupation with the problem of antinomy was the chief cause of the\nrevolution which took place in Kant\u2019s views in 1769, and which found\nexpression in his _Dissertation_ of 1770. It was the existence of\nantinomy which led Kant to recognise the subjectivity of space and time.\nThat is to say, it led him to develop that doctrine of transcendental\nidealism which reappears in the concluding sections of the _Aesthetic_,\nand which was recast and developed in the _Analytic_. Already in the\n_Dissertation_ it supplies the key for the solution of the problems\nconcerning infinity. The impossibility of completing the space, time,\nand causal series, and the consequent impossibility of satisfying the\ndemands of the mind for totality, simplicity and unconditionedness, do\nnot, it is there maintained, discredit reason, but only serve to\nestablish the subjectivity of the sensuous forms to which the element of\ninfinitude is in all cases due.\n\nKant\u2019s thinking was, of course, diverted into an entirely new channel\n(as his letter to Herz of February 21, 1772,[1342] shows), when he came\nto realise that the metaphysical validity or invalidity of thought must\nbe decided prior to any attempt to discover a positive solution of such\nproblems as are presented by the antinomies. And when, owing to the\nrenewed influence of Hume, at some time subsequent to the date of the\nletter to Herz, this new problem was recognised as being the problem of\n_a priori synthesis_, all questions regarding the nature of the\nabsolutely real were made to take secondary rank, yielding precedence to\nthose of logical theory. When the antinomy problems re-emerge, their\ndiscussion assumes Critical form.\n\nIn three fundamental respects Kant\u2019s treatment of the antinomies in the\n_Dissertation_ differs from that of the _Critique_. In the first place,\nthe demand for totality or absoluteness is not in the _Dissertation_\nascribed to a separate faculty. Indeed Kant\u2019s words would seem to show\nthat at times he had inclined to ascribe it merely to the free-ranging\nfancy or imagination.[1343] Secondly, as the various antinomies were\ntraced exclusively to the influence of space and time upon pure thought,\nthey were treated together, and no classification of them was attempted.\nAnd lastly, though Kant\u2019s utterances are somewhat ambiguous,[1344] the\nillusory character of the antinomies was in the main viewed as being of\na more or less _logical_ nature. That is to say, it was regarded as\nentirely preventable and as \u201cvanishing like smoke\u201d[1345] upon adoption\nof a true philosophical standpoint.\n\nA number of the _Reflexionen_ reveal the various tentative schemes, by\ntrial of which Kant worked his way toward a more genuinely Critical\ntreatment of the problems of infinity. The intellectual factors receive\nfuller recognition, and as a consequence a definite classification\nresults. At some time prior to the discovery of the table of categories,\nKant adopted a threefold division of what he names first principles or\npresuppositions--principles of substance-accident, of\nground-consequence, and of whole-part. _Reflexion_ ii. 578 is typical.\n\n     \u201cThree _principia_: (1) in the field of the actual there is the\n     relation of substance to accident (_inhaerentia_): (2) of ground to\n     consequence (_dependentia_): (3) of parts and of composition\n     (_compositio_). There are three presuppositions: of the subject, of\n     the ground, and of the parts; of insition [Kant\u2019s own term], of\n     subordination, and of composition; therefore also three first\n     _principia_: (1) subject, which is never a predicate; (2) ground,\n     which is never a consequence; (3) unity, which is not itself\n     composite.\u201d\n\nThere are numerous other _Reflexionen_ to the same effect.[1346] The\nresulting conceptions are defined both as limits[1347] and as absolute\ntotalities, and in _Reflexion_ ii. 1252 are enumerated as follows:\n\n     \u201cThe first subject; the first ground; the first part. The subject\n     which holds everything in itself; the ground which takes everything\n     under itself; the whole which comprehends everything. The\n     _totalitas absoluta_ of reality, of series, of co-ordination.\u201d\n\nThe introduction of the terms \u2018absolute\u2019 and \u2018totality\u2019 indicate that\nKant has also come to recognise the presence of a unique notion\n(equivalent to the \u201cunconditioned\u201d of the _Critique_), distinct in\ncontent from any of the three enumerated _principia_, but common to them\nall. From the very first Kant would seem to have appropriated for it the\ntitle Idea. _Reflexionen_ ii. 1243, 1244, and 124 may be quoted:\n\n     \u201cThe Idea is single (_individuum_), self-sufficient, and eternal.\n     The divinity of our soul is its capacity to form the Idea. The\n     senses give only copies or rather _apparentia_.\u201d \u201cIdea is the\n     representation of the whole in so far as it necessarily precedes\n     the determination of the parts. It can never be empirically\n     represented, for the reason that in experience we proceed from the\n     parts through successive syntheses to the whole. It is the\n     archetype (_Urbild_) of things, for certain objects are only\n     possible through an Idea. Transcendental Ideas are those in which\n     the absolute whole determines the parts in an aggregate or as\n     series.\u201d \u201cMetaphysics proper is the application of transcendental\n     philosophy to concepts supplied by Reason and necessary to it, to\n     which, however, no corresponding objects can be given in\n     experience. The concepts must therefore refer to the supersensible.\n     That, however, can be nothing but the unconditioned, _for that is\n     the sole theoretical Idea of reason_. [Not italicised in the\n     original.] Metaphysics thus relates: (1) to that of which only the\n     whole can be represented as absolutely unconditioned: (2) to things\n     so far as they are in themselves sensuously unconditioned. The\n     first part is cosmology, the second rational doctrine of the soul,\n     pneumatology and theology.\u201d\n\nAt this stage, therefore, Kant would seem to have held that there is but\none Idea strictly so called, and that the above three _principia_ are\nmerely specifications of it in terms of the concepts of\nsubstance-accident, ground-consequence, and whole-part. The\nclassification thus obtained is in certain respects more satisfactory\nthan that which is adopted in the _Critique_. It locates the\ncosmological argument with the causal category, and so would enable the\nconceptions of freedom or _causa sui_, and of Divine Existence, to be\ndealt with in their natural connection with one another. It also\nsupplies, in the category of whole and part, a more fitting heading for\nthose antinomy problems which deal with the unlimited and the limited,\nthe divisible and the indivisible, the complex and the simple. The\nclassification would, however, in separating the problem of the simple\nfrom that of substance, remain open to the same criticism as that of the\n_Critique_.[1348]\n\nThis classification must, as we have stated, be of a date prior to\nKant\u2019s discovery of the table of categories. That is quite clear from\nits ignoring the category of reciprocity, and from its combination of\nthe other two categories of relation with the merely quantitative\ncategory of whole and part. For though the last is also entitled\ncomposition and co-ordination, it is conceived in these particular\n_Reflexionen_ in exclusively quantitative terms. When Kant formulated\nthe \u201cmetaphysical\u201d deduction of the categories he was, of course,\ncompelled to recast the classification, and did so in the only possible\nmanner, consistent with his architectonic, by substituting the category\nof reciprocity for that of whole and part,[1349] and by taking the new\nheading, obtained through combination of reciprocity with the Idea of\nthe unconditioned, as equivalent to the Idea of Divine Existence. But\nthis could not be done without dislocating the entire scheme. The\ncategory of ground and consequence is deprived of its chief application,\nthat expressed in the cosmological argument; and in order to provide a\nnew content for it, Kant is compelled to force upon it the problems\npreviously classified under the displaced category of whole and part.\nEven so, the problem of the _causa sui_ cannot be eliminated, and\nreappears, partly as the problem of freedom, and partly as the modal\nproblem of necessary existence.\n\nThe identification of the theological Idea with the category of\nreciprocity has a further consequence. It carries the problem of Divine\nExistence outside the sphere of the problems of infinity, and\nnecessitates a very different treatment from that which it would\nnaturally have received at Kant\u2019s hands, if developed in its connection\nwith his own Critical teaching. He is driven to expound it in the\nextreme rationalistic form in which it had been formulated by Leibniz\nand Wolff, as a doctrine of the _Ens realissimum_.\n\nPrior to the rearrangement, necessitated by recognition of the category\nof reciprocity, Kant would seem to have expected to bring the entire\nbody of Wolffian metaphysics within the scope of a general doctrine of\nantinomy. The problems of the divisible and the indivisible, of the\nsimple and the complex, leading as they do to discussion of the\npresuppositions underlying the Leibnizian monadology, concern spiritual\nas well as material substance. Similarly, the main problems of theology\nwould have been treated in connection with the cosmological inference to\na first cause, and with the discussion of the possibility of first\nbeginnings in space and time.[1350]\n\nThe sections in the _Critique_ devoted to the antinomies reveal, in many\nways, Kant\u2019s original design. It is especially noticeable in his\ndiscussion of the third and fourth antinomies. The problems of freedom\nand of necessary existence are by no means treated in merely\ncosmological fashion. Indeed Kant makes no pretence of concealing their\npsychological and theological implications. Even the first and second\nantinomies have obvious bearings of a similar character. But it is in\nthe section entitled _The Interest of Reason in this\nSelf-conflict_[1351] that the broader significance of the antinomies\nfinds its fullest expression. In its suggestive contrast of the two\npossible types of philosophy, Epicurean and Platonic, the argument\nentirely transcends the bounds prescribed to it by its cosmological\nsetting. As we follow the comprehensive sweep of its argument, we can\nhardly avoid regretting that Kant failed to keep to his original plan,\nas here unfolded,[1352] of expounding the self-conflict of Reason in the\nform of a broad judicial statement of the grounds and claims of the two\nopposing authorities which divide the allegiance of the human spirit,\nnamely, the intellectual and the moral, science with its cognitive\ndemands on the one hand, the consciousness of duty with its no less\nimperious prescriptions on the other. The materialist philosophies would\nthen have been presented as inevitably arising when intellectual values\nare made supreme; and the Idealist philosophies as equally cogent when\nmoral values are taken as primary and are allowed to determine\nspeculative tenets. Against this background of conflicting dogmatisms\nthe comprehensive and satisfying character of the Critical standpoint\nwould have stood out the more clearly; and its historical affiliations,\nits debt to the sceptics and materialists, no less than to the\nIdealists, would have been depicted in more adequate terms. As it is, in\nthe chapters on the _Paralogisms_ and the _Ideal of Pure Reason_ there\nis almost entire failure to recognise the possibility of a naturalistic\nsolution of the problems with which they deal, and Kant so far succumbs\nto the outworn influences of his day and generation--the very influences\nfrom which the Critical philosophy, consistently developed, is a final\nbreaking away--as to maintain, almost in the manner of the English\nDeists, of Voltaire and Rousseau, that God, Freedom, and Immortality are\nconceptions which the mind must necessarily form, and in the validity of\nwhich it must spontaneously believe. Kant is here, indeed, interpreting\n\u201cnatural reason\u201d in the light of his own personal history. The Christian\nbeliefs, in which he had been nurtured from childhood, and their\nrationalist counterparts in the Wolffian philosophy, had become, as it\nwere, a second nature to him; and the resistance, which in his own\nperson they had offered to the development of Critical teaching, he not\nunnaturally interpreted as evidence of their being imposed by the very\nstructure of reason. He transforms the metaphysical sciences in their\nWolffian form into inevitable illusions of the human mind.[1353]\n\nThere is evidence that the theological problems were the first to be\nwithdrawn from the sphere of the \u201csceptical method,\u201d[1354] peculiar to\nthe antinomies. Thus _Reflexion_ ii. 125[1355] states that \u201cmetaphysics\nproper consists of _cosmologia rationalis_ and _theologia\nnaturalis_\u201d--rational psychology being, as it would seem, still included\nwithin cosmology.[1356] What the considerations were which induced Kant\nto claim similarly independent treatment for rational psychology, we can\nonly conjecture. For a time, while still holding to the bipartite\ndivision, he would seem to have made the further change of also\nseparating psychology from cosmology, classing psychology and theology\ntogether as subdivisions of the rational science of soul.\n\n     \u201d[Metaphysics has two parts]: the first is cosmology, the second\n     rational doctrine of soul, pneumatology and theology.\u201d[1357]\n\nA main factor deciding Kant in favour of a dogmatic, non-sceptical\ntreatment of rational psychology may have been the greater opportunity\nwhich it seemed to afford him of connecting its doctrines with the\nteaching of the _Analytic_, and especially with his central doctrine of\napperception. But to whatever cause the decision was due, it resulted in\nthe impoverishment of the second antinomy, through withdrawal of the\nmore important half of its natural content. This antinomy could no\nlonger be made to comprehend a discussion of the logical bases of\nmonadology, and of its professed proofs of the simplicity and\nimmortality of the soul. Nothing is left to it save the discussion of\nthe monadistic theory of matter (_somatologia pura_).[1358] This change\nhas also, as already noted, the unfortunate effect of precluding Kant\nfrom recognition of the physical application of the category of\nsubstance. By the simple he means the substantial, and yet he may not\nsay so; his architectonic forbids.\n\nI may hazard the further suggestion that Kant\u2019s interpretation of\nrational psychology in terms of the Critical doctrine of apperception is\nof earlier date than his doctrine of transcendental illusion. For the\nchapter on the _Paralogisms_ seems in its first form to have contained\nno reference to that latter doctrine.[1359] The few passages which take\naccount of it, all bear evidence of being later intercalations. This is\nthe more remarkable in that the Paralogisms can easily be shown to be\ntypical examples of transcendental illusion. Indeed, neither the\nantinomies nor the theological Ideal conform to its definition in the\nsame strict fashion.\n\nThe problem as to whether the doctrine of transcendental illusion and\nthe deduction of the Ideas from the three species of syllogism\noriginated early or late, is largely bound up with the question as to\nwhen Kant finally adopted the terms _Analytic_ and _Dialectic_ as titles\nfor the two main divisions of his _Transcendental Logic_. That Kant was\nat first very uncertain as to what the main divisions of his system\nought to be, appears very clearly from the _Reflexionen_.[1360] To his\nteaching as a whole he usually applies the title _Transcendental\nPhilosophy_, and in _Reflexion_ ii. 123 he enumerates the following\nsubdivisions within it: _Aesthetic_, _Logic_, _Critique_, and\n_Architectonic_. By _Critique_ Kant must here mean what in other\n_Reflexionen_ he names _Discipline_, and which he finally named\n_Dialectic_. As thus identified with the _Discipline_, the _Dialectic_\nis at times viewed as a division of a _Methodology_ or _Organon_, whose\nother divisions are entitled _Canon_ and _Architectonic_.[1361] This\nearlier scheme may therefore be represented as follows:\n\n                                       {Aesthetic.\n                 {Doctrine of Elements {Logic.\n  Transcendental {\n  Philosophy     {                     {Critique = Discipline [corresponding\n                 {Doctrine of Methods  {to the Dialectic\n                 {  (Methodology)      {of the _Critique_].\n                                       {Canon.\n                                       {Architectonic.\n\nThe terms _Analytic_ and _Dialectic_ do not occur in these\n_Reflexionen_, and their adoption may therefore be inferred to\nsynchronise with Kant\u2019s later decision to include the treatment of the\nmetaphysical sciences within his _Logic_; and that decision was probably\nan immediate result of his having developed meantime a doctrine of\ntranscendental illusion. The new scheme in its final form is therefore\nas follows:\n\n                 {            {Aesthetic.\n                 {Doctrine of {\n                 { Elements   {      {Analytic {of Concepts.\n                 {            {Logic {         {of Judgement.\n  Transcendental {                   {Dialectic--of Reason.\n  Philosophy     {\n  or Critique of {\n  Pure Reason    {              {Discipline (retained but given a new\n                 {Doctrine of   {and more general content).\n                 {Methods       {Canon.\n                 {(Methodology) {Architectonic.\n                                {History.\n\nIn thus transferring _Dialectic_ from the _Methodology_ to the _Doctrine\nof Elements_, Kant stands committed to the view that it contains\npositive teaching of a character analogous to that of the _Analytic_,\nwith which it is now co-ordinated. As we have already noted, the\nfundamental opposition which runs through the entire _Dialectic_ is due\nto the conflict between the older view of Reason as merely understanding\nin its transcendent employment, and this later view of it as a distinct\nfaculty, yielding concepts with a positive and indispensable function,\ndifferent from, and yet also analogous to, that exercised by the\ncategories of the understanding.\n\nAdickes, to whom students of Kant are indebted for a convincing\ndemonstration of the constant influence of Kant\u2019s logical architectonic\nupon the content of the Critical teaching, would seem at this point to\nrely too exclusively upon that method of explanation. He contends that\nKant\u2019s deduction of the Ideas of Reason from the three species of\nsyllogism is entirely traceable to this source, and is without real\nphilosophical significance. That is perhaps in the main true. But it\nneed not prevent us from appreciating the importance of the doctrines\nwhich Kant contrives to expound under guise of this logical machinery.\nWe have already observed that prior to the discovery of this deduction\nKant had recognised the connection between the concept of the\nunconditioned and the three Ideas through which it finds expression. As\nthe forms of syllogism are differentiated in terms of the three\ncategories of relation, the deduction does not interfere with Kant\u2019s\nretention of this classification of Ideas; while in connecting Reason as\na faculty with reasoning as a logical process, an excellent opportunity\nis found for explaining the grounds and significance of the demand for\nunconditionedness, i.e. for completeness of explanation. This demand, as\nhe has also come to recognise, lies open to question, and therefore\ncalls for more precise definition.\n\nThe artificial character of the metaphysical deduction lies not so much\nin this derivation of the three Ideas of the\nunconditioned--unconditioned substance, unconditioned causality,\nunconditioned system--from the categorical, hypothetical, and\ndisjunctive forms of syllogism respectively, as in the further equating\nof them with the Ideas of the Self, the World, and God. The Idea of\nunconditioned substance has many possible applications besides the use\nto which it is put in rational psychology. The Idea of an unconditioned\ncausality may be conceived in psychological and theological as well as\nin cosmological terms; and as a matter of fact Kant himself frequently\nidentifies it with the concept of freedom, as in the third and fourth\nantinomies, or when he enumerates the Ideas as being those of God,\nFreedom, and Immortality.[1362] Similarly, the Idea of system is the\ninspiring principle of materialism, and also finds in such philosophies\nas that of Spinoza much more adequate expression than in the _Ens\nrealissimum_ of the Wolffian School. But further comment is not, at this\nstage, really profitable. These are questions which can best be\ndiscussed as they emerge in the course of the argument.[1363]\n\nKant carried his logical architectonic one stage further. Not satisfied\nwith connecting the three Ideas of Reason with the categories that\nunderlie the three species of syllogism, he also attempted to organise\nthe various particular applications of each Idea in terms of the\nfourfold division of the table of categories. By the use of his usual\nhigh-handed methods he succeeded in doing so in the case of the\npsychological and cosmological Ideas. There are four paralogisms and\nfour antinomies. But when the attempt failed in regard to the\ntheological Idea, he very wisely abstained from either apology or\nexplanation. That the failure was not due to lack of desire or\nperseverance appears from _Reflexion_ ii. 1573, which would seem to be\nthe record of an unavailing attempt to obtain a satisfactory\narticulation of the theological Ideal. Doubtless, had he been\nsufficiently bent upon it, he could have worked out some sort of\nfourfold division; but there were limits even to Kant\u2019s devotion to the\narchitectonic scheme. It is difficult to see how any such arrangement\ncould have been followed without serious perversion of the argument.\n\nAdickes has suggested[1364] that the distinction between the faculty of\nunderstanding and the faculty of judgment is subsequent to, and\nsuggested by, Kant\u2019s successful tracing of the Ideas to a separate\nfaculty of Reason. Some such distinction was demanded in order that the\nparallelism of transcendental and formal logic might be complete. This\nconjecture of Adickes is probably correct. It would seem to be supported\nby the internal evidence of the _Analytic of Principles_. As we have had\noccasion to note,[1365] the doctrine of schematism, in terms of which\nthe distinction between understanding and judgment is formulated, is\nlate in date of origin.[1366] This distinction is of the same artificial\ncharacter as that between understanding and Reason; and though, like the\nlatter distinction, it supplies Kant with a convenient framework for the\narrangement of genuine Critical material, it also tends to conceal the\nsimpler and more inward bonds of true relationship.\n\n\n\n\nTRANSCENDENTAL DIALECTIC\n\n\n\n\nINTRODUCTION\n\n\n_I. Transcendental Illusion_\n\n=Dialectic is a Logic of Illusion.=[1367]--The meaning which Kant attaches\nto the term dialectic has already been considered. The passage above\nquoted[1368] from his _Logic_ shows the meaning which he supposed the\nterm historically to possess, namely, as being a sophistical art of\ndisputation, presenting false principles in the guise of truth by means\nof a seeming fulfilment of the demands of strict logical proof. The\nincorrectness of this historical derivation hardly needs to be pointed\nout. Kant professes[1369] to be following his contemporaries in thus\nusing the term as a title for the treatment of false reasoning. But even\nthis statement must be challenged. Adickes, after examination of a large\nnumber of eighteenth-century text-books, reports[1370] that in the six\npassages in which alone he has found it to occur it is never so\nemployed. In Meier it is used as a title for the theory of probable\nreasoning,[1371] and in Baumgarten it occurs only in adjectival form as\nequivalent to sophistical. This last is the nearest approach to Kant\u2019s\ndefinition. All historical considerations may therefore be swept aside.\nWe are concerned only with the specific meaning which Kant thought good\nto attach to the term. He adapts it in the freest manner to the needs of\nhis system. In A 61 = B 85, as in his _Logic_, he has defined it in\nmerely negative fashion. He is now careful to specify the more positive\naspects of the problems with which it deals. Though definable as the\nlogic of illusion, the deceptive inferences with which it concerns\nitself are of a quite unique and supremely significant character. They\nmust, as above noted,[1372] be distinguished alike from logical and from\nempirical illusion. They have their roots in the fundamental needs of\nthe human mind, and the recognition of their illusory character does not\nrender unnecessary either a positive explanation of their occurrence or\na Critical valuation of their practical function as regulative ideals.\n\n=A 293 = B 349.=--Regarding the connection between illusion and error cf.\nB 69, and above, pp. 148-53.\n\n=A 295 = B 352.=--Logical, empirical, and transcendental illusion. Cf.\nabove, pp. 13, 427-9, 437.\n\n=A 296 = B 352.=--Kant here defines the terms transcendental and\ntranscendent in a very unusual manner. The two terms are not, he states,\nsynonymous. The principles of pure understanding are of merely empirical\nvalidity, and _consequently_ are not of _transcendental_ employment\nbeyond the limits of experience. A principle is _transcendent_ when it\nnot only removes these limits, but prescribes the overstepping of them.\n\n\n_II. Pure Reason as the Seat of Transcendental Illusion_[1373]\n\n(_a_) _Reason in General_\n\nReason, like understanding, is employed in two ways, formal or logical\nand real. The logical use of Reason consists in mediate inference, the\nreal in the generation of concepts and principles. Reason is thus both a\nlogical and a transcendental faculty, and we may therefore expect that\nits logical functions will serve as a clue to those that are\ntranscendental. The argument which follows is extremely obscure. It is a\nforeshadowing in logical terms of a distinction which, as Kant himself\nindicates, cannot at this stage be adequately stated. The distinction\nmay be extended and paraphrased as follows. Reason, generically taken as\nincluding both activities, is the faculty of principles, in distinction\nfrom understanding which is the faculty of rules.[1374] Principles,\nproperly so-called, are absolutely _a priori_. Universals which imply\nthe element of intuition must not, therefore, be ranked as principles in\nthe strict sense. They are more properly to be entitled rules. A true\nprinciple is one that affords knowledge of the particulars which come\nunder it, and which does so from its own internal resources, that is to\nsay, through pure concepts. In other words, it yields _a priori_\nsynthetic knowledge, and yet does so independently of all given\nexperience. Now, as the _Analytic_ has proved, knowledge obtained\nthrough understanding, whether in mathematical or in physical science,\nis never of this character. Its principles, even though originating in\npure intuition or in the pure understanding, are valid only as\nconditions of possible experience, and are applicable only to such\nobjects as can occur in the context of a sense-perception. That is to\nsay, the understanding can never obtain synthetic knowledge through pure\nconcepts. Though, for instance, it prescribes the principle that\neverything which happens must have a cause, that principle does not\nestablish itself by means of the concepts which it contains, but only as\nbeing a presupposition necessary to the possibility of sense-experience.\nIf, then, principles in the strict sense actually exist, they must be\ndue to a faculty distinct from understanding, and will call for a\ndeduction of a different character from that of the categories.\n\nIn the last paragraph but one of the section Kant indicates the doctrine\nwhich he is foreshadowing. The rules of understanding apply to\nappearances, prescribing the conditions under which the unity necessary\nto any and every experience can alone be attained. The principles of\nReason do not apply directly to appearances, but only to the\nunderstanding, defining the standards to which its activities must\nconform, if a completely unified experience is to be achieved. Whereas\nthe rules of understanding are the conditions of objective existence in\nspace and time, principles in the strict sense are criteria for the\nattainment of such absoluteness and totality as will harmonise Reason\nwith itself. Reason, determined by principles which issue from its own\ninherent nature, prescribes what the actual ought to be; understanding,\nproceeding from rules which express the conditions of possible\nexperience, can yield knowledge only of what is found to exist in the\ncourse of sense-experience. The unity of Reason is Ideal; the unity of\nunderstanding is empirical. Principles are due to the\n_self_-determination of reason; the rules of understanding express the\nnecessitated determinations of sense. The former demand a more perfect\nand complete unity than is ever attainable by means of the latter. Two\npassages from the _Lose Bl\u00e4tter_ will help to define the distinction.\n\n     \u201cThere is a synthesis prototypon and a synthesis ectypon. The one\n... _simpliciter, a termino a priori_, ... the other _secundum\n     quid, a termino a posteriori_.... Reason advances from the\n     universal to the particular, the understanding from the particular\n     to the universal.... The first is absolute and belongs to the free\n     or metaphysical, and also to the moral, employment of\n     Reason.\u201d[1375] \u201cThe principles of the synthesis of pure Reason are\n     all metaphysical.... [They] are principles of the subjective unity\n     of knowledge through Reason, _i.e._ of the agreement of Reason with\n     itself.\u201d[1376]\n\nThe chief interest of this section lies in its clear indication of the\ndual standpoint to which Kant is committing himself by the manner in\nwhich he formulates this distinction between rules and principles. The\nindispensableness of the latter, upon which Kant is prepared to insist,\npoints to the Idealist interpretation of their grounds and validity;\ntheir derivation from mere concepts, without reference to or basis in\nexperience, must, on the other hand, in view of the teaching of the\n_Analytic_, commit Kant to a sceptical treatment of their objective\nvalidity. In the above account, suggestions of the Idealist point of\nview are not entirely absent; but, on the whole, it is the sceptical\nview that is dominant. The Ideas of Reason can be justified as necessary\nonly for the perfecting of experience, not as conditions of experience\nas such. They express a subjective interest in the attainment of unity,\nnot conditions of the possibility of objective existence.\n\n     \u201d[Civil Laws] are only limitations imposed upon our freedom in\n     order that such freedom may completely harmonise with itself; hence\n     they are directed to something which is entirely our own work, and\n     of which we ourselves, through these concepts, can be the cause.\n     But that objects in themselves, the very nature of things, should\n     stand under principles, and should be determined according to mere\n     concepts, is a demand which, if not impossible, is at least quite\n     contrary to common sense [_Widersinnisches_].\u201d[1377]\n\n\n(_b_) _The Logical Use of Reason_[1378]\n\nIn this subsection Kant introduces the distinction between understanding\nand judgment which he has sought to justify in A 130 ff. = B 169 ff. By\nshowing that inference determines the _relation_ between a major premiss\n(due to the understanding) and the condition defined in the minor\npremiss (due to the faculty of judgment), he professes to obtain\njustification for classifying the possible forms of reasoning according\nto the three categories of relation. The general remark is added that\nthe purpose of Reason, in its logical employment as inference, is to\nobtain the highest possible unity, through subsumption of all\nmultiplicity under the smallest possible number of universals.\n\n\n(c) _The Pure Use of Reason_[1379]\n\nKant here states the alternatives between which the _Dialectic_ has to\ndecide. Is Reason merely formal, arranging given material according to\ngiven forms of unity, or is it a source of principles which prescribe\nhigher forms of unity than any revealed by actual experience? Further\nexamination of its formal and logical procedure constrains us, Kant\nasserts, to adopt the latter position; and at the same time indicates\nhow those principles must be interpreted, namely, as subjective laws\nthat apply not to objects but only to the activities of the\nunderstanding.\n\nIn the first place, a syllogism is not directly concerned with\nintuitions, but only with concepts and judgments. This may be taken as\nindicating that pure Reason relates to objects only mediately _by way_\nof understanding and its judgments. The unity which it seeks is higher\nthan that of any possible experience; it is a unity which must be\nconstructed and cannot be given.[1380]\n\nSecondly, Reason in its logical use seeks the _universal_ condition of\nits judgment; and when such is not found in the major premiss proceeds\nto its discovery through a regressive series of prosyllogisms. In so\ndoing it is obviously determined by a principle expressive of the\npeculiar function of Reason in its logical employment, namely, that for\nthe conditioned knowledge of understanding the _unconditioned_ unity in\nwhich that knowledge may find completion must be discovered. Such a\nprinciple is synthetic, since from analysis of the conception of the\nconditioned we can discover its relation to a condition, but never its\nrelation to the _unconditioned_. That is a notion which falls entirely\noutside the sphere of the understanding, and which therefore demands a\nseparate enquiry. How is the above _a priori_ synthetic principle to be\naccounted for, if it cannot be traced to understanding? Has it\nobjective, or has it merely subjective validity? And lastly, what\nfurther synthetic principles can be based upon it? Such are the\nquestions to which Critical _Dialectic_ must supply an answer. This\n_Dialectic_ will be composed of two main divisions, the doctrine of \u201cthe\ntranscendent concepts of pure Reason\u201d and the doctrine of \u201ctranscendent\nand dialectical inferences of Reason.\u201d\n\n\n\n\nBOOK I\n\nTHE CONCEPTS OF PURE REASON[1381]\n\n\nThe distinction here drawn between concepts obtained by reflection and\nconcepts gained by inference is a somewhat misleading mode of stating\nthe fact that, whereas the categories of understanding condition\nexperience and so make possible the unity of consciousness necessary to\nall reflection, or, in other words, are conditions of the material\nsupplied for inference, the concepts of Reason are Ideal constructions\nwhich though in a certain sense resting upon experience none the less\ntranscend it. The function of the Ideas is to organise experience in its\ntotality; that of the categories is to render possible the\nsense-perceptions constitutive of its content. The former refer to the\nunconditioned, and though that is a conception under which everything\nexperienced is conceived to fall, it represents a type of knowledge to\nwhich no actual experience can ever be adequate.\n\n=Conceptus ratiocinati--conceptus ratiocinantes.= When such transcendent\nconcepts possess \u201cobjective validity,\u201d they are correctly inferred, and\nmay be entitled _conceptus ratiocinati_. If, on the other hand, they are\ndue to merely sophistical[1382] reasoning, they are purely fictitious,\n_conceptus ratiocinantes_. This distinction raises many difficulties.\nKant\u2019s intention cannot be to deny that the _conceptus ratiocinati_ are\n\u201cmere Ideas\u201d (_entia rationis_)[1383]--for such is his avowed and\nconstant contention--or that the inference to them is dialectical and is\nbased upon a transcendental illusion. Two alternatives are open. He may\nmean that they are only valid when the results of such inference are\nCritically reinterpreted, and when the function of the Ideas is realised\nto be merely regulative; or his intention may be to mark off the Ideas,\nstrictly so-called, which are inevitable and beneficial products of\nReason, from the many idle and superfluous inventions of speculative\nthought. Kant\u2019s concluding remark, that the questions at issue can be\nadequately discussed only at a later stage, may be taken as in the\nnature of an apology for the looseness of these preliminary statements,\nand as a warning to the reader not to insist upon them too absolutely.\nThe participles _ratiocinati_ and _ratiocinantes_[1384] are of doubtful\nlatinity. The distinction of meaning here imposed upon them has not been\ntraced in any other writer, and is perhaps Kant\u2019s own invention.[1385]\n\n\nSECTION I\n\nIDEAS IN GENERAL[1386]\n\nKant connects his use of the term Idea with the meaning in which it is\nemployed by Plato. He urges upon all true lovers of philosophy the\nimperative need of rescuing from misuse a term so indispensable to mark\na distinction more vital than any other to the very existence of the\nphilosophical disciplines.\n\n     \u201d[For Plato] Ideas are the archetypes of the things themselves, and\n     not, like the categories, merely keys to possible experiences. In\n     his view they issued from the Supreme Reason, and from that source\n     have come to be shared in by human Reason.... He very well realised\n     that our faculty of knowledge feels a much higher need than merely\n     to spell out appearances according to a synthetic unity, in order\n     to read them as experience. He knew that our Reason naturally\n     exalts itself to forms of knowledge which so far transcend the\n     bounds of experience that no given empirical object can ever\n     coincide with them, but which must none the less be recognised as\n     having their own reality and which are by no means mere fictions of\n     the brain.\u201d[1387]\n\nPlato found these ideas chiefly, though not exclusively, in the\npractical sphere. When moral standards are in question, experience is\nthe mother of illusion.\n\n     \u201cFor nothing can be more injurious or more unworthy of a\n     philosopher than the vulgar appeal to so-called adverse experience.\n     Such experience would never have existed at all, if those\n     institutions had been established at the proper time in accordance\n     with Ideas, and if Ideas had not been displaced by crude\n     conceptions which, just because they have been derived from\n     experience, have nullified all good intentions.\u201d[1388]\n\nEven in the natural sphere Ideas which are never themselves adequately\nembodied in the actual must be postulated in order to account for the\nactual. Certain forms of existences \u201care possible only according to\nIdeas.\u201d\n\n     \u201cA plant, an animal, the orderly arrangement of the\n     cosmos--probably, therefore, the entire natural world--clearly show\n     that they are possible only according to Ideas, and that though no\n     single creature in the conditions of its individual existence\n     coincides with the Idea of what is most perfect in its kind--just\n     as little as does any individual man exactly conform to the Idea of\n     humanity, which he actually carries in his soul as the archetype of\n     his actions--yet these Ideas are none the less completely\n     determined in the Supreme Understanding, each as an individual and\n     each as unchangeable, and are the original causes of things. But\n     only the totality of things, in their interconnection as\n     constituting the universe, is completely adequate to the\n     Idea.\u201d[1389]\n\nThough Kant avows the intention of adapting the term Idea freely to the\nneeds of his more Critical standpoint, all these considerations\ncontribute to the rich and varied meanings in which he employs it.\n\n_Reflexionen_ and passages from the _Lectures on Metaphysics_ may be\nquoted to show the thoroughly Platonic character of Kant\u2019s early use of\nthe term, and to illustrate its gradual adjustment to Critical demands.\n\n     \u201cThe Idea is the unity of knowledge, through which the manifold\n     either of knowledge or of the object is possible. In the former,\n     the whole of knowledge precedes its parts, the universal precedes\n     the particular; in the latter, knowledge of the objects precedes\n     their possibility, as _e.g._ in [objects that possess] order and\n     perfection.\u201d[1390] \u201cThat an object is possible only through a form\n     of knowledge is a surprising statement; but all teleological\n     relations are possible only through a form of knowledge [_i.e._ a\n     concept].\u201d[1391] \u201cThe Idea is single (_individuum_),\n     self-sufficient, and eternal. The divinity of our soul is its\n     capacity to form the Idea. The senses give only copies or rather\n     _apparentia_.\u201d[1392] \u201cAs the Understanding of God is the ground of\n     all possibility, archetypes, Ideas, are in God.... The divine\n     _Intuitus_ contains Ideas according to which we ourselves are\n     possible; _cognitio divina est cognitio archetypa_, and His Ideas\n     are archetypes of things. The [corresponding] forms of knowledge\n     possessed by the human understanding we may also entitle (in a\n     comparative sense) archetypes or Ideas. They are those\n     representations of our understanding which serve for judgment upon\n     things.\u201d[1393] \u201cIdea is the representation of the whole in so far\n     as it necessarily precedes the determination of the parts. It can\n     never be empirically represented, because in experience we proceed\n     from the parts through successive synthesis to the whole. It is the\n     archetype of things, for certain objects are only possible through\n     an Idea. Transcendental Ideas are those in which the absolute whole\n     determines the parts in an aggregate or as series.\u201d[1394] \u201cThe pure\n     concepts of Reason have no _exemplaria_; they are themselves\n     archetypes. But the concepts of our pure Reason have as their\n     archetypes this Reason itself and are therefore subjective, not\n     objective.\u201d[1395] \u201cThe transcendental Ideas serve to limit the\n     principles of experience, forbidding their extension to things in\n     themselves, and showing that what is never an object of possible\n     experience is not therefore a non-entity [_Unding_], and that\n     experience is not adequate either to itself or to Reason, but\n     always refers us further to what is beyond itself.\u201d[1396] \u201cThe\n     employment of the concept of understanding was immanent, that of\n     the Ideas as concepts of objects is transcendent. But as regulative\n     principles alike of the completion and of the limitation of our\n     knowledge, they are Critically immanent.\u201d[1397] \u201cThe difficulties\n     of metaphysics all arise in connection with the reconciling of\n     empirical principles with Ideas. The possibility of the latter\n     cannot be denied, but neither can they be made empirically\n     intelligible. The Idea is never a _conceptus dabilis_; it is not an\n     empirically possible conception.\u201d[1398]\n\nKant[1399] appends the following \u2018Stufenleiter\u2019 (ladder-like)\narrangement of titles for the various kinds of representation.\nRepresentation (_Vorstellung_) is the term which he substitutes for the\nCartesian and Lockian employment of the term idea, now reserved for use\nin its true Platonic meaning. To entitle such a representation as that\nof red colour an idea is, in Kant\u2019s view, an intolerable and barbaric\nprocedure; that representation is not even a concept of the\nunderstanding.\n\n  Representation    {\n  (_Vorstellung_,   { Perception =\n  _Repraesentatio_) {  Representation     --->\n  with or without   {  with consciousness\n  consciousness     {\n\n  -------------------------------------------------------------------\n\n                                           { Idea (_Idee_) or concept of\n                                           { Reason [formed, Kant here\n                                           { says, from notions, but\n                                    { Pure { transcending the possibility\n                                    {      { of experience].\n   { Objective =    { Concept       {      { Notion (_notio_) [= category\n   { Knowledge      {  (_Begriff_)  {      { of understanding].\n-->{ (_Erkenntnis_) {               { Empirical.\n   {                { Intuition (_Anschauung_).\n   { Subjective = Sensation (_Empfindung_).\n\n\n\nSECTION II\n\nTHE TRANSCENDENTAL IDEAS[1400]\n\nThis section completes the metaphysical deduction of the Ideas. In the\npreceding sections on the logical and on the pure use of Reason, Kant\nhas pointed out that Reason proceeds in accordance with the principle,\nthat for the conditioned knowledge of understanding the unconditioned,\nin which it finds completion, must be discovered. This principle is\nsynthetic, involving a concept which transcends the understanding; and\nas Reason in its logical use is merely formal, that concept must be due\nto Reason in its creative or transcendental activity. In the section\nbefore us Kant deduces from the three kinds of syllogism the three\npossible forms in which such an Idea of Reason can present itself. The\ndeduction is, as already noted, wholly artificial, and masks Kant\u2019s real\nmethod of obtaining the Ideas, namely, through combination of the unique\nconcept of the unconditioned with the three categories of relation. The\ndeduction is based upon an extremely ingenious analogy between the\nlogical function of Reason in deductive inference and its transcendental\nprocedure in prescribing the Ideal of unconditioned totality. In the\nsyllogism the predicate of the conclusion is shown to be connected with\nits subject in accordance with a condition which is stated in its\nuniversality in the major premiss. Thus if the conclusion be: Caius is\nmortal, in constructing the syllogism, required to establish it, we seek\nfor a conception which contains the condition under which the predicate\nis given--in this case the conception \u201cman\u201d--and we state that condition\nin its universality: All men are mortal. Under this major premiss is\nthen subsumed Caius, the object dealt with: Caius is a man. And so\nindirectly, by reference to the universal condition, we obtain the\nknowledge that Caius is mortal. Universality, antecedently stated, is\nrestricted in the conclusion to a specific object. Now what corresponds\nin the synthesis of intuition to the _universality_ (_universalitas_) of\na logical premiss is _allness_ (_universitas_) or _totality_ of\nconditions. The transcendental concept of Reason, to which the logical\nprocedure is to serve as clue, can therefore be no other than that of\nthe totality of conditions for any given conditioned. And as totality of\nconditions is equivalent to the _unconditioned_, this latter must be\ntaken as the fundamental concept of Reason; the unconditioned is\nconceived as being the ground of the synthesis of everything\nconditioned. But there are three species of relation, and consequently\nthere are three forms in which the concept of Reason seeks to realise\nits demand for the unconditioned: (1) through categorical synthesis in\none subject, (2) through hypothetical synthesis of the members of a\nseries, and (3) through disjunctive synthesis of the parts in one\nsystem. To these three correspond the three species of syllogism,\ncategorical, hypothetical, and disjunctive, in each of which thought\npasses through a regressive series of prosyllogisms back to an\nunconditioned: the first to a concept which stands for what is always a\nsubject and never a predicate; the second to a presupposition which\nitself presupposes nothing further; and the third to such an aggregate\nof the members of the division as will make that division complete. It\nmay be observed that in this proof the threefold specification of the\nconcept of the unconditioned is really obtained directly from the\ncategories of relation, or at least from the judgments of relation, and\nnot from the corresponding species of syllogism.\n\nTotality and unconditionedness, when taken as equivalent, become\nsynonymous with the _absolute_.[1401] This last term, however,\nespecially when taken as defining possibility and necessity, is\nambiguous. The absolutely possible may signify either that which in\nitself, _i.e._ so far as regards its internal content, is possible; or\nelse that which is in every respect and in all relations possible. The\ntwo meanings have come to be connected largely owing to the fact that\nthe internally impossible is impossible in every respect. Otherwise,\nhowever, the two meanings fall completely apart. Absolute necessity and\ninner necessity are quite diverse in character. We must not, for\ninstance, argue that the opposite of what is absolutely necessary must\nbe inwardly impossible, nor consequently that absolute necessity must in\nthe end reduce to an inner necessity. Examination will show that, in\ncertain types of cases, not the slightest meaning can be attached to the\nphrase \u2018inner necessity.\u2019 As we possess the terms inner and logical to\ndenote the first form of necessity, there is no excuse for employing the\nterm absolute in any but the wider sense. That, Kant holds, is its\noriginal and proper meaning. The _absolute_ totality to which the\nconcept of Reason refers is that form of completeness which is in every\nrespect unconditioned.\n\nIn A 326 = B 383 Kant\u2019s mode of statement emphasises the connection of\nthe Ideas with the categories of relation. Reason, he claims, \u201cseeks to\nextend the synthetic unity, which is thought in the category, to the\nabsolutely unconditioned.\u201d Such positive content as the Ideas can\npossess lies in the experience which they profess to unify; in so far as\nthey transcend experience and point to an Ideal completion that is not\nempirically attainable, they refer to things of which the understanding\ncan have no concept. It is necessary, however, that they should present\nthemselves in this absolute and transcendent form, since otherwise the\nunderstanding would be without stimulus and without guidance. Though\nmere Ideas, they are neither arbitrary nor superfluous. They regulate\nthe understanding in its empirical pursuit of that systematic unity\nwhich it requires for its own satisfaction.\n\nIn A 327-8 = B 383-4 one and the same ground is assigned for entitling\nthe Ideas transcendental and also transcendent, namely, that, as they\nsurpass experience, no object capable of being given through the senses\ncorresponds to them. But a difference would none the less seem to be\nimplied in the connotation of the two terms. In being prescribed by the\nvery nature of Reason, they are transcendental; as overstepping the\nlimits of experience, they are transcendent. Kant\u2019s use of the terms\nsubject and object in this passage is also somewhat puzzling. \u2018Object\u2019\nis employed in the metaphysical sense proper only from the pre-Critical\nstandpoint of the _Dissertation_, as meaning an existence apprehended\nthrough pure thought. The term \u2018subject\u2019 receives a correspondingly\nun-Critical connotation. The further phrase \u201cthe merely speculative use\nof Reason\u201d is somewhat misleading, even though we recognise that for\nKant speculative and theoretical are synonymous terms; we should rather\nexpect \u201cReason in its legitimate or Critical or directive function.\u201d\nKant\u2019s intended meaning, however, is sufficiently clear. When we say\nthat a concept of Reason is an Idea merely, we have in mind the degree\nto which it can be _empirically_ verified. We are asserting that it\nprescribes an Ideal to which experience may be made to approach, but\nwhich it can never attain. It defines \u201ca problem to which there is no\nsolution.\u201d In the practical sphere of morals, on the other hand, the\nIdeal of Reason must never be so described. Though only partially\nrealisable, it is genuinely actual. Even those actions which imperfectly\nembody it none the less presuppose it as their indispensable condition.\nIn two respects, therefore, as Kant points out, the statement that the\ntranscendental concepts of Reason are merely Ideas calls for\nqualification. In the first place they are by no means \u201csuperfluous and\nvoid.\u201d They supply a canon for the fruitful employment of understanding.\nAnd secondly, they may perhaps be found to make possible a transition\nfrom natural to moral concepts, and so to bring the Ideas of practical\nReason into connection with the principles of speculative thought. The\nreader may again note the genuinely Platonic character of Kant\u2019s use of\nthe term Idea.\n\nIn A 330-1 = B 386-7 Kant returns to the problem of the metaphysical\ndeduction, and analyses the nature of syllogistic reasoning. The\nanalysis differs from that of A 321 ff. = C 377 ff. only in emphasising\nthat when a conclusion is given as valid the totality of the premisses\nrequired for its establishment can be postulated as likewise given, and\nthat when completely stated in the implied prosyllogisms the premisses\nform a regressive series. In this way Kant contrives to bring the\nlogical process into closer connection with the transcendental\nprinciple, which he now definitively formulates as follows: When the\nconditioned is given, the series of conditions up to the unconditioned\nis likewise given. The series of antecedent conditions may either have a\nfirst term or may be incapable of such. In either case it has to be\nviewed as unconditioned, in the one case in virtue of its unconditioned\nbeginning, in the other in its character as an unending and therefore\nunlimited series. In one or other form Reason demands that the\nunconditioned be recognised as underlying and determining everything\nconditioned.[1402]\n\n\nSECTION III\n\nSYSTEM OF THE TRANSCENDENTAL IDEAS[1403]\n\nThe three Ideas of Reason, as derived from the three kinds of syllogism,\nare now brought into connection with the three possible relations in\nwhich representations are found to stand: first, to the thinking\nsubject; secondly, to objects as appearances; thirdly, to objects of\nthought in general. Kant argues that the completed totalities towards\nwhich Reason strives are likewise three in number. Reason seeks: (1) in\nregard to the subject known, as constituting the fact of inner\nexperience, a representation of the self or soul that will render\ncompletely intelligible what is peculiar to the inner life; (2) in\nregard to the object known, a conception of the completed totality of\nthe world of phenomena, the cosmos; (3) in regard to the ultimate\nsynthesis of the subject known and the object known, such a conception\nof all existing things as will render intelligible the co-operation of\nmind and external nature in one experience. In this way Kant professes\nto obtain transcendental justification for the threefold division of\nmetaphysical science into rational psychology, rational cosmology, and\nrational theology. The absolute unity of the thinking subject is dealt\nwith by psychology, the totality of all appearances by cosmology, and\nthe Being, which contains the condition of the possibility of all that\ncan be thought, by theology.\n\nIn thus proceeding, Kant is assuming that the concepts of unconditioned\nsubstance and of unconditioned necessity can be interpreted only in\nspiritualist and theological terms.[1404] This assumption stands in\ndirect conflict with what the history of philosophy records. The\nAbsolute has frequently been materialistically defined, and, as Kant\nhimself admits, we cannot prove that the thinking subject may not be\nnaturalistically conditioned. Architectonic is again exercising its\nbaleful influence. That the argument is lacking in cogency is indeed so\nevident that Kant takes notice of the deficiency,[1405] and promises\nthat it will be remedied in the sequel. This promise he is unable to\nfulfil. Such further reasons as he is able to offer are of the same\nexternal character.[1406]\n\n\u201cOf these transcendental Ideas, _strictly speaking, no objective\ndeduction, such as we were able to give of the categories, is\npossible_.\u201d[1407] As Kant indicates by use of the phrase \u2018strictly\nspeaking,\u2019 this statement is subject to modification. He himself\nformulates a transcendental deduction of the Ideas, as principles\nregulative of experience.[1408] The deduction from the three forms of\nsyllogism, which Kant here entitles subjective, ought properly to be\nnamed \u2018metaphysical.\u2019[1409]\n\n\n\n\nBOOK II\n\nTHE DIALECTICAL INFERENCES OF PURE REASON[1410]\n\n\n\n\nCHAPTER I\n\nTHE PARALOGISMS OF PURE REASON[1411]\n\n\nAs rational psychology fails to distinguish between appearances and\nthings in themselves, it identifies mere apperception with inner sense;\nthe self in experiencing the succession of its inner states is supposed\nto acquire knowledge of its own essential nature. \u201cI, as thinking, am an\nobject of inner sense, and am entitled soul,\u201d in contrast to the body\nwhich is an object of outer sense. Empirical psychology deals with the\nconcrete detail of inner experience; rational psychology abstracts from\nall such special experiences, indeed from everything empirical,\nprofessing to establish its doctrine upon the single judgment, \u201cI\nthink.\u201d That judgment has already been investigated in its connection\nwith the problem of the possibility, within the field of experience, of\nsynthetic _a priori_ judgments. It has now to be considered as a\npossible basis for knowledge of the self as a thinking being (_ein\ndenkend Wesen_) or soul (_Seele_).\n\nFollowing the guiding thread of the table of categories, but placing\nthem in what he regards as being, in this connection, the most\nconvenient order, Kant obtains a \u201ctopic\u201d or classification of the\npossible rubrics for the doctrines of a rational psychology: (1) the\nsoul is _substance_; (2) is _simple_; (3) is _numerically identical_;\n(4) stands in relation to _possible_ objects in space. Now all those\nfour doctrines are, Kant holds, incapable of demonstration. The proofs\npropounded by rational psychology are logically imperfect, committing\nthe logical fallacy which is technically named paralogism.[1412] The\nfallacy is not, however, of merely logical character. Had that been the\ncase, it could never have gained such general currency. Certainly no\nmetaphysical science, widely accepted by profound thinkers, could ever\nhave come to be based upon it. The paralogism is transcendental in\ncharacter, resting upon a transcendental ground. It represents an\nillusion which from any non-Critical standpoint is altogether\nunavoidable. Its dialectic is a natural dialectic, wrongly interpreted\nby the Schools, but not capriciously invented by them. The key to its\nproper treatment is first supplied by the results of the transcendental\ndeduction. We are now called upon to apply these results in explanation\nof the occurrence of the paralogisms, and in judgment upon their false\nclaims. Little that is really new is to be found in this chapter; but\nmany of the established results of the _Analytic_ receive interesting\nillustration, and are thereby set in a clearer light.\n\nIn rational psychology the \u201cI think\u201d is taken in its universal, or to\nuse Kant\u2019s somewhat misleading term, problematic aspect, that is to say,\nnot as a judgment expressive of the self\u2019s own existence but \u201cin its\nmere possibility,\u201d[1413] as representing the self-consciousness of all\npossible thinking beings. As we cannot gain a representation of thinking\nbeings through outer experience, we are constrained to think them in\nterms of our own self-consciousness. The \u201cI think\u201d is thus taken as a\nuniversal judgment, expressing what belongs to the conception of\nthinking being in general. The judgment is so interpreted by rational\npsychology, \u201cin order to see what predicates applicable to its subject\n(be that subject actually existent or not) may flow from so simple a\njudgment.\u201d\n\nIn summarising what is directly relevant in the argument of the\ntranscendental deduction, Kant emphasises that the I, as representation,\nis altogether empty of content.[1414]\n\n     \u201cWe cannot even say that it is a conception, but only that it is a\n     bare (_blosses_) consciousness which accompanies all conceptions.\n     Through this I or he or it (the thing) which thinks, nothing\n     further is represented than a transcendental subject of the\n     thoughts = _x_....\u201d\n\nIt is apprehended only in its relation to the thoughts which are its\npredicates; apart from them we cannot form any conception whatever of\nit, but can only revolve in a perpetual circle, since any judgment upon\nit has already made use of its representation.[1415]\n\nThe patchwork character of the _Critique_, the artificial nature of the\nconnections between its various parts, is nowhere more evident than in\nthis section on the _Paralogisms_. According to the definition given of\ntranscendental illusion, we naturally expect Kant\u2019s argument to show\nthat the _Paralogisms_ rest upon a failure to distinguish between\nappearance and reality. As a matter of fact, the cause of their fallacy\nis traced in the first three _Paralogisms_ solely to a failure to\ndistinguish between the logical and the real application of the\ncategories. The argument can indeed be restated so as to agree with the\nintroductory sections of the _Dialectic_. But Kant\u2019s manner of\nexpounding the _Paralogisms_ shows that this chapter must originally\nhave been written independently of any intention to develop such\nteaching as that of the sections which in the ultimate arrangement of\nthe _Critique_ are made to lead up to it.[1416]\n\n=First Paralogism: of Substantiality.=[1417]--Save for the phrase \u2018subject\nin itself,\u2019 there is, in Kant\u2019s comment upon this _Paralogism_, not a\nword regarding the necessity of a distinction between appearance and\nreality, but only an insistence that the \u201cI think\u201d yields no knowledge\nof the thinking self. Consciousness of the self and knowledge of its\nunderlying substance are by no means identical. The self, so far as it\nenters into consciousness, is a merely logical subject; the underlying\nsubstrate is that to which this self-consciousness and all other\nthoughts are due. It is in the light of this distinction that Kant\ndiscusses the substantiality of the subject. As expressive of the \u201cI\nthink,\u201d the category of substance and attribute can be employed only to\ndefine the relation in which consciousness stands to its thoughts; it\nexpresses the merely logical relation of a subject to its predicates. It\ntells us nothing regarding the nature of the \u201cI,\u201d save only that it is\nthe invariable centre of reference for all thoughts. In order to know\nthe self as substance, and so as capable of persisting throughout all\nchange, and as surviving even the death of the body, we should require\nto have an intuition of it, and of such intuition there is not the\nslightest trace in the \u201cI think.\u201d It \u201csignifies a substance only in\nIdea, not in reality.\u201d[1418] As Kant adds later,[1419] the permanence\nand self-identity of the representation of the self justifies no\nargument to the permanence and self-identity of its underlying\nconditions. Inference from the nature of representation to the nature of\nthe object represented is entirely illegitimate. In the equating of the\ntwo, and not, as the introduction to the _Dialectic_ would lead us to\nexpect, in a failure to distinguish appearance from reality, consists\nthe paralogistic fallacy of this first syllogism.\n\n=Second Paralogism: of Simplicity.=[1420]--We may follow Adickes[1421] in\nhis analysis of A 351-62. (_a_) The original criticism, parallel to that\nof the first _Paralogism_, would seem to be contained in paragraphs five\nto nine. (_b_) The opening paragraphs, and (_c_) the concluding\nparagraphs, would seem, for reasons stated below, to be independent and\nlater additions.\n\n(_a_) The argument of the central paragraphs runs almost exactly\nparallel with the criticism of the first _Paralogism_, applying the same\nline of thought, in disproof of the assumed argument for the simplicity\nof the soul. It may be noted, in passing, that Kant here departs from\nhis table of categories. There is no category of simplicity. The\nconnection which he seeks to establish between the concept of simplicity\nand the categories of quality is arbitrary. It more naturally connects\nwith the category of unity; but the category of unity is required for\nthe third _Paralogism_. For explanation of the way in which he equates\nthe concept of simplicity with the category of reality Kant is satisfied\nto refer the reader to the section on the second antinomy in which this\nsame identification occurs.[1422] Indeed the simplicity here dwelt upon\nseems hardly distinguishable from substantiality, and therefore it is\nnot surprising that Kant\u2019s criticism of the second _Paralogism_ should\nbe practically identical with that of the first.[1423] Since the \u201cI,\u201d as\nlogical subject of thought, signifies only a something in general, and\nembodies no insight into the constitution of this something, it is for\nthat reason empty of all content, and consequently simple. \u201cThe\nsimplicity of the _representation_ of a subject is not _eo ipso_ a\nknowledge of the simplicity of the subject itself....\u201d The second\n_Paralogism_ thus, in Kant\u2019s view, falsely argues from the merely\nlogical unity of the subject _in representation_ to the actual\nsimplicity of the subject _in itself_.\n\n(_b_) One reason for regarding the first four paragraphs as a later\naddition is their opening reference to the introductory sections of the\n_Dialectic_, of which this chapter otherwise takes little or no account.\nThis _Paralogism_ is, Kant declares, \u201cthe Achilles of all the\ndialectical inferences in the pure doctrine of the soul,\u201d meaning that\nit may well seem a quite invulnerable argument.[1424]\n\n     \u201cIt is no mere sophistical play contrived by a dogmatist in order\n     to impart to his assertions a superficial plausibility (_Schein_),\n     but an inference which appears to withstand even the keenest\n     scrutiny and the most scrupulously exact investigation.\u201d\n\nThe second paragraph is a very pointed restatement of a main supporting\nargument of this second _Paralogism_. This argument well deserves the\neulogy with which Kant has ushered it in. It is as follows. The unity of\nconsciousness can not be explained as due to the co-operative action of\nindependent substances. Such a merely external effect as that of motion\nin a material body may be the resultant of the united motions of its\nparts. But it is otherwise with thought. For should that which thinks be\nviewed as composite, and the different representations, as, for\ninstance, of the single words of a verse, be conceived as distributed\namong the several parts, a multiplicity of separate consciousnesses\nwould result, and the single complex consciousness, that of the verse as\na whole, would be rendered impossible. Consciousness cannot\ntherefore--such is the argument--inhere in the composite. The soul must\nbe a simple substance.[1425]\n\nAs there is no reference in this argument to the \u201cI think,\u201d the\ncriticism cannot be that of the first _Paralogism_, nor that of the\ncentral paragraphs of this second _Paralogism_. Kant\u2019s reply--as given\nin the third and fourth paragraphs--is in effect to refer the reader to\nthe results of the _Analytic_, and is formulated in the manner of his\n_Introduction_ to the _Critique_. The principle that multiplicity of\nrepresentation presupposes absolute unity in the thinking subject can\nneither be demonstrated analytically from mere concepts, nor derived\nfrom experience. Being a synthetic _a priori_ judgment, it can be\nestablished only by means of a transcendental deduction. But in that\nform it will define only a condition required for the possibility of\nconsciousness; it can tell us nothing in regard to the noumenal nature\nof the thinking being. And, as Kant argues in the third\n_Paralogism_,[1426] there may be a possible analogy between thought and\nmotion, though of a different kind from that above suggested.\n\nThe entire absence of all connection between the argument of these\nparagraphs and the argument of those which immediately follow upon\nthem, at least suffices to show that this second _Paralogism_ has not\nbeen written as a continuous whole; and taken together with the fact\nthat the problem is here formulated in terms of the _Introduction_ to\nthe _Critique_, would seem to show that this part of the section is of\ncomparatively late origin.\n\n(_c_) The concluding paragraphs, which are of considerable intrinsic\ninterest, also reflect an independent line of criticism. As the phrase\n\u201cthe above proposition\u201d[1427] seems to indicate, they were not\noriginally composed in this present connection. They give expression to\nKant\u2019s partial agreement with the line of argument followed by the\nrationalists, but also seek to show that, despite such partial validity,\nthe argument does not lend support to any metaphysical extension of our\nempirical knowledge. In A 358 we have what may be a reference to the\nargument of the introductory sections of the _Dialectic_. The argument\nunder criticism is praised as being \u201cnatural and popular,\u201d \u201coccurring\neven to the least sophisticated understanding,\u201d and as leading it to\nview the soul as an altogether different existence from the body. The\nargument is as follows. None of the qualities proper to material\nexistence, such as impenetrability or motion, are to be discovered in\nour inner experience. Nor can feelings, desires, thoughts, etc., be\nexternally intuited. In view of these differences, we seem justified in\nasserting that the soul cannot be an appearance in space, and cannot\ntherefore be corporeal. Kant replies by drawing attention to the\nfundamental Critical distinction between appearances and things in\nthemselves.[1428] If material bodies, as apprehended, were things in\nthemselves, the argument would certainly justify us in refusing to\nregard the soul and its states as of similar nature. But since, as the\n_Aesthetic_ has shown, bodies, as known, are mere appearances of outer\nsense, the real question at issue is not that of the distinction between\nthe soul and bodies in space, but of the distinction between the soul\nand that something which conditions all outer appearances.\n\n     \u201c...this something which underlies the outer appearances and which\n     so affects our sense that it obtains the representations of space,\n     matter, shape, etc., this something, viewed as noumenon (or better,\n     as transcendental object), might yet also at the same time serve as\n     the subject of our thoughts....\u201d[1429]\n\nThus the argument criticised serves only to enforce the very genuine\ndistinction between inner and outer appearances; it justifies no\nassertion, either positive or negative, as to the nature of the soul or\nas to its relation to body in its noumenal aspect. The monadistic,\nspiritualist theory of material existence remains an open possibility,\nthough only as an hypothesis incapable either of proof or of disproof.\nWe cannot obtain, by way of inference from the character of our\napperceptive consciousness, any genuine addition to our speculative\ninsight.\n\n=Third Paralogism: of Personality.=[1430]--Kant\u2019s criticism again runs\nparallel with that of the preceding _Paralogisms_. The fallacy involved\nis traced to a confusion between the numerical identity of the self _in\nrepresentation_ and the numerical identity of the subject _in itself_.\nThe logical subject of knowledge must, as the transcendental deduction\nhas proved, think itself as self-identical throughout all its\nexperiences. This is indeed all that the judgment \u201cI think\u201d expresses.\nIt is mere identity, \u201cI am I.\u201d But from the identity of representation\nwe must not argue to identity of the underlying self. So far as the\nunity of self-consciousness is concerned, there is nothing to prevent\nthe noumenal conditions of the self from undergoing transformation so\ncomplete as to involve the loss of identity, while yet supporting the\nrepresentation of an identical self.\n\n     \u201cAlthough the dictum of certain ancient Schools, that everything in\n     the world is in a flux and nothing permanent and abiding, cannot be\n     reconciled with the admission of substances, it is not refuted by\n     the unity of self-consciousness. For we are unable from our own\n     consciousness to determine whether, as souls, we are permanent or\n     not. Since we reckon as belonging to our identical self only that\n     of which we are conscious, we must necessarily judge that we are\n     one and the same throughout the whole time of which we are\n     conscious. We cannot, however, claim that such a judgment would be\n     valid from the standpoint of an outside observer. As the only\n     permanent appearance which we meet with in the soul is the\n     representation \u2018I\u2019 that accompanies and connects them all, we are\n     unable to prove that this \u2018I,\u2019 a mere thought, may not be in the\n     same state of flux as the other thoughts which are connected\n     together by its means.\u201d[1431]\n\nAnd Kant adds an interesting illustration.[1432]\n\n     \u201cAn elastic ball which impinges on another similar ball in a\n     straight line communicates to the latter its whole motion, and\n     therefore its whole state (_i.e._ if we take account only of the\n     positions in space). If, then, in analogy with such bodies, we\n     postulate substances such that the one communicates to the other\n     representations together with the consciousness of them, we can\n     conceive a whole series of substances of which the first transmits\n     its state together with its consciousness to the second, the second\n     its own state with that of the preceding substance to the third,\n     and this in turn the states of all the preceding substances\n     together with its own consciousness and with their consciousness to\n     another. The last substance would then be conscious of all the\n     states of the substances, which had undergone change before its own\n     change, as being its own states, because they would have been\n     transferred to it together with the consciousness of them. And yet\n     it would not have been one and the same person in all these\n     states.\u201d[1433]\n\nThe perversely Hegelian character of Caird\u2019s and Watson\u2019s manner of\ninterpreting the _Critique_ is especially evident in their treatment of\nthe _Paralogisms_. They make not the least mention of this part of\nKant\u2019s teaching.\n\nKant employs a further argument which would seem to show that at the\ntime when these paragraphs were written the general tendency of his\nthought was predominantly subjectivist in character. There are, he\nimplies, as many different times as there are selves that represent\ntime.[1434] The argument is as follows. As the \u201cI think\u201d is equivalent\nto \u201cI am I,\u201d we may say either that all time of which I am conscious is\nin me, or that I am conscious of myself as numerically identical in each\nand every part of it. In my individual consciousness, therefore,\nidentity of my person is unfailingly present. But an observer, viewing\nme from the outside,[1435] represents me in the time of his own\nconsciousness; and as the time in which he thus sets me is not that of\nmy own thinking, the self-identity of my consciousness, even if he\nrecognises its existence, does not justify him in inferring the\nobjective permanence of my self.\n\nThe two concluding paragraphs seem to have been independently\ncomposed.[1436] They contribute nothing of importance.\n\n=Fourth Paralogism: of Ideality.=[1437]--The main argument of this\n_Paralogism_, which contains the first edition refutation of idealism,\nhas already been considered above.[1438] We require, therefore, only to\ntreat of it in its connection with the other _Paralogisms_, and to note\nsome few minor points that remain for consideration. Its argument\ndiffers from that of the other _Paralogisms_ in that the fallacy\ninvolved is traced, in agreement with the requirements of the\nintroductory sections of the _Dialectic_, to a failure to distinguish\nbetween appearances and things in themselves. Its connection with the\ntable of categories is extremely artificial. In A 344 = B 402 the\ncategory employed is that of possibility, in A 404 and A 344 _n._ that\nof existence.[1439] Kant\u2019s attempt to combine the problem here treated\nwith that of the other _Paralogisms_ can only be explained as due to the\nrequirements of his architectonic.[1440] This _Paralogism_ does not\nconcern itself with the nature of the soul. It refers exclusively to the\nmode of existence to be ascribed to objective appearances. None the\nless, Kant contrives to bring it within the range of rational psychology\nin the following manner. He argues[1441] that rational psychologists are\none and all adherents of empirical idealism. They confound appearances\nin space with things in themselves, and therefore assert that our\nknowledge of their existence is inferential and consequently uncertain.\nThe errors of empirical idealism are thus bound up with the dogmatic\nassumptions of the rationalist position. They are traceable to its\nfailure to distinguish between appearances and things in themselves.\nSuch dogmatism may take the form of materialism or of ontological\ndualism, as well as of spiritualism.[1442] All three, in professing to\npossess knowledge of things in themselves, violate Critical principles.\nIf the chief function of rational psychology consists in securing the\nconception of the soul against the onslaughts of materialism,[1443] that\ncan be much more effectively attained through transcendental idealism.\n\n     \u201cFor, on [Critical] teaching, so completely are we freed from the\n     fear that on the removal of matter all thought, and even the very\n     existence of thinking beings, would be destroyed, that on the\n     contrary it is clearly shown that if I remove the thinking subject\n     the whole corporeal world must at once vanish, since it is nothing\n     save appearance in the sensibility of our subject and a species of\n     its representations.\u201d[1444]\n\nWe do not, indeed, succeed in proving that the thinking self is in its\nexistence independent of the \u201ctranscendental substrate\u201d[1445] of outer\nappearances. But as both possibilities remain open, the admission of our\nignorance leaves us free to look to other than speculative sources for\nproof of the independent and abiding existence of the self.\n\n=Reflection on the Whole of Pure Psychology.=[1446]--This section affords\nKant the opportunity of discussing certain problems which he desires to\ndeal with, but is unable to introduce under the recognised rubrics of\nhis logical architectonic.[1447] There are, Kant says, three other\ndialectical questions, essential to the purposes of rational psychology,\ngrounded upon the same transcendental illusion (confusion of appearances\nwith things in themselves), and soluble in similar fashion: (1) as to\nthe possibility of the communion of soul and body, _i.e._ of the state\nof the soul during the life of the body; (2) as to the beginning of this\nassociation, _i.e._ of the soul in and before birth; (3) as to the\ntermination of this association, _i.e._ of the soul in and after the\ndeath of the body. Kant treats these three problems from the extreme\nsubjectivist standpoint, inner and outer sense being distinguished and\nrelated in the manner peculiar to the first edition. The contrast\nbetween mind and body is a difference solely between the appearances of\ninner and those of outer sense. Both alike exist only in and through the\nthinking subject, though the latter\n\n     \u201c...have this deceptive property that, representing objects in\n     space, they as it were detach themselves from the soul and appear\n     to hover outside it.\u201d[1448]\n\nThe problem, therefore, of the association of soul and body, properly\nunderstood, is not that of the interaction of the soul with other known\nsubstances of an opposite nature, but only\n\n     \u201c...how in a thinking subject outer intuition, namely, that of\n     space, with its filling in of figure and motion, is possible. And\n     that is a question which no human being can possibly answer. The\n     gap in our knowledge ... can only be indicated through the\n     ascription of outer appearances to that transcendental object which\n     is the cause of this species of representations, but of which we\n     can have no knowledge whatsoever and of which we shall never\n     acquire any conception.\u201d[1449]\n\nThe familiar problem of the association of mind and body is thus due to\na transcendental illusion which leads the mind to hypostatise\nrepresentations, viewing them as independent existences that act upon\nthe senses and generate our subjective states. The motions in space,\nwhich are merely the expression in terms of appearance of the influence\nof the transcendental object upon \u201cour senses,\u201d[1450] are thus wrongly\nregarded as the causes of our sensations. They themselves are mere\nrepresentations, and, as Kant implies, are for that reason incapable of\nacting as causes. In this section, it may be noted in passing, there is\nnot the least trace of the phenomenalist teaching, according to which\nspatial objects are viewed as acting upon the bodily sense-organs. Kant\nhere denies all interaction of mind and body, and recognises only the\ninteraction of their noumenal conditions. Appearances as such can never\nhave causal efficacy. The position represented is pure subjectivism, and\nvery significantly goes along with Kant\u2019s earlier doctrine of the\ntranscendental object.[1451]\n\nThe dogmatic character of the interaction theory appears very clearly,\nas Kant proceeds to point out, in the objections which have been made to\nit, whether by those who substitute for it the theories of\npre-established harmony and occasionalism, or by those who adopt a\nsceptical non-committal attitude. Their objections rest upon exactly the\nsame presupposition as the theory which they are attacking. To\ndemonstrate the impossibility of interaction, they must be able to show\nthat the transcendental object is not the cause of outer appearances;\nand owing to the limitations of our knowledge that is entirely beyond\nour powers. Failing, however, to draw a distinction between appearances\nand things in themselves, they have not realised the actual nature of\nthe situation, and accordingly have directed their objections merely to\nshowing that mind and body, taken as independent existences, must not be\nviewed as capable of interaction.\n\nThe Critical standpoint also supplies the proper formulation for the\nother two problems--a formulation which in itself decides the degree and\nmanner of our possible insight in regard to them. The view that the\nthinking subject may be capable of thought prior to all association with\nthe body should be stated as asserting\n\n     \u201c...that prior to the beginning of that species of sensibility in\n     virtue of which something appears to us in space, those\n     transcendental objects, which in our present state appear to us as\n     bodies, could have been intuited in an entirely different\n     manner.\u201d[1452]\n\nThe view that the soul, upon the cessation of all association with the\ncorporeal world, may still continue to think, will similarly consist in\nthe contention\n\n     \u201c...that if that species of sensibility, in virtue of which\n     transcendental objects (which in our present state are entirely\n     unknown) appear to us as a material world, should cease, all\n     intuition of them would not for that reason be removed; but that\n     it would still be possible that those same unknown objects should\n     continue to be known [sic] by the thinking subject, though no\n     longer, indeed, in the quality of bodies.\u201d[1453]\n\nNot the least ground, Kant claims, can be discovered by means of\nspeculation in support of such assertions. Even their bare possibility\ncannot be demonstrated. But it is equally impossible to establish any\nvalid objection to them. Since we cannot pretend to knowledge of things\nin themselves, a modest acquiescence in the limitations of experience\nalone becomes us.\n\nThe remaining paragraphs (A 396-405) contain nothing that is new. They\nmerely repeat points already more adequately stated. A 401-2, which\ndeals with the nature of apperception and its relation to the\ncategories, has been considered above.[1454] The argument that, as the\nself must presuppose the thought of itself in knowing anything, it\ncannot know itself as object, is also commented upon above.[1455]\n\nThe statement[1456] that the determining self (the thinking, _das\nDenken_) is to be distinguished from the determinable self (the thinking\nsubject) as knowledge from its object, should be interpreted in the\nlight of Kant\u2019s argument in the second and third _Paralogisms_, that the\nsimplicity and self-identity of the representation of an object must not\nbe taken as knowledge of simplicity or numerical identity in the object\nrepresented.\n\nThe analysis given in A 402-3 of the fallacy involved in the\n_Paralogisms_ is, as Adickes has pointed out,[1457] confused and\nmisleading. Kant here declares that in the major premiss of each\nsyllogism the assertion is intended in the merely logical sense, and\ntherefore as applicable only to the subject _in representation_, but in\nthe minor premiss and conclusion is asserted of the subject as bearer of\nconsciousness, _i.e. in itself_. But were that so, the minor premiss\nwould be a false assertion, and the false conclusion would not be\ntraceable to logical fallacy. Kant gives the correct statement of his\nposition in B 410-11.[1458] The attempted justification of the fourfold\narrangement of the _Paralogisms_ with which the section concludes\nsuffers from the artificiality of Kant\u2019s logical architectonic.\n\n\nSECOND EDITION STATEMENT OF THE PARALOGISMS[1459]\n\nExcept for the introductory paragraphs, which remain unaltered, the\nchapter is completely recast in the second edition. The treatment of the\nfour _Paralogisms_ which in the first edition occupied thirty-three\npages is reduced to five. The problems of the mutual interaction of mind\nand body, of its prenatal character and of its immortality, the\ndiscussion of which in the first edition required some ten pages, are\nnow disposed of in a single paragraph (B 426-7). The remaining\ntwenty-two pages of the new chapter are almost entirely devoted to more\nor less polemical discussion of criticisms which had been passed upon\nthe first edition. These had been in great part directed against Kant\u2019s\ndoctrine of apperception and of inner sense, and so could fittingly be\ndealt with in connection with the problems of rational psychology. As\nBenno Erdmann has suggested,[1460] B 409-14 and 419-21 would seem to be\ndirected against Ulrichs\u2019[1461] Leibnizian position and especially\nagainst his metaphysical interpretation of apperception. B 428-30 treats\nof the difficulties raised by Pistorius[1462] in regard to the existence\nof the self. B 414-15 is similarly polemical, but in this case Kant\ncites his opponent, Mendelssohn, by name. Throughout, as in the\nalterations made in the chapter on _Phenomena and Noumena_, Kant insists\nmore strongly than in the first edition upon the unknowableness of the\nself, and on the difference between thought and knowledge. The pure\nforms of thought are not, Kant now declares, concepts of objects, that\nis, are not categories,[1463] but \u201cmerely logical functions.\u201d Though\nthis involves no essential doctrinal change, it indicates the altered\nstandpoint from which Kant now regards his problem. Its significance has\nalready been dwelt upon.[1464]\n\nIn formulating the several arguments of the four _Paralogisms_, Kant\ndevelops and places in the forefront a statement which receives only\npassing mention in A 352-3, 362, 366-7, 381-2, namely, that the truths\ncontained in the judgments of rational psychology find expression in\nmerely identical (_i.e._ analytic) propositions. This enables Kant to\nformulate both the _Paralogisms_ and his criticisms thereof in much\nbriefer and more pointed fashion. In each case the _Paralogism_, as he\nshows, substitutes a synthetic _a priori_ judgment, involving an\nextension of our knowledge and a reference to the noumenal self, for the\ngiven judgment which, in so far as it is valid, is always a merely\nanalytic restatement of the purely formal \u201cI think.\u201d From the very\nstart also, Kant introduces the distinctions of his own Critical\nteaching, especially that between thinking and intuiting, and that\nbetween the determining and the determinable self.\n\n=First Paralogism.=--That the I which thinks must always _in thought_ be\nviewed as subject and not as mere predicate, is an identical\nproposition. It must not be taken as meaning that the subject which\nunderlies thought is an abiding substance. This latter proposition is of\nmuch wider scope, and would involve such data (in this case entirely\nlacking) as are required for the establishment of a synthetic _a priori_\njudgment.\n\n=Second Paralogism.=--That the I of _apperception_ and so of all _thought_\nis single and cannot be resolved into a multiplicity of subjects, is\ninvolved in the very conception of thought, and is therefore an analytic\nproposition. It must not be interpreted as signifying that the self is a\nsimple substance. For the latter assertion is again a synthetic\nproposition, and presupposes for its possibility an intuition by the\nself of its own essential nature. As all our intuitions are merely\nsensuous, that cannot be looked for in the \u201cI think.\u201d\n\n     \u201cIt would, indeed, be surprising if what in other cases requires so\n     much labour to discover--namely, what it is, of all that is\n     presented by intuition, that is substance, and further, whether\n     this substance is simple (_e.g._ in the parts of matter)--should be\n     thus directly given me, as if by revelation, in the poorest of all\n     representations.\u201d[1465]\n\nWe may here observe how the practice, adopted by Caird, of translating\n_Anschauung_ by \u2018perception\u2019 has misled him into serious\nmisunderstanding of Kant\u2019s teaching. It has caused him[1466] to\ninterpret Kant as arguing that we have no knowledge of the self because\nwe can have no _sensuous_ perception of it. Kant\u2019s argument rather is\nthat as all human \u201cintuition\u201d is sensuous, we are cut off from all\npossibility of determining our noumenal nature. We are thrown back upon\nmere concepts which, as yielding only analytic propositions, cannot\nextend our insight beyond the limits of sense-experience. The term\n\u2018intuition\u2019 is much broader in meaning than the term \u2018perception\u2019; it\ncan also be employed as equivalent to the phrase \u2018_immediate_\napprehension.\u2019[1467] The grounds for Kant\u2019s contention that we have no\nintuition or immediate knowledge of the self are embodied in, and\ninspire, his doctrine of inner sense.[1468] It may also be noted that\nin B 412 Kant, speaking of the necessity of intuition for knowledge of\nthe self, uses the unusual phrase \u2018a permanent intuition\u2019--a phrase\nwhich, so far as I have observed, he nowhere employs in dealing with the\nintuition that conditions the sense perception of material bodies.[1469]\nIts employment here may perhaps be due to the fact that its implied\nreference is not to a given sensuous manifold but to some form of\nimmediate apprehension, capable of revealing the permanent nature of the\nnoumenal self.\n\n=Third Paralogism.=--That I am identical with myself throughout the\nconsciousness of my manifold experiences, is likewise an analytic\nproposition obtainable by mere analysis of the \u201cI think.\u201d And since that\nform of consciousness, as stated in the criticism of the preceding\n_Paralogism_, is purely conceptual, containing no element of intuition,\nno judgment based solely upon it can ever be taken as equivalent to the\nsynthetic proposition that the self, as thinking being, is an identical\nsubstance.\n\n=Fourth Paralogism.=--This _Paralogism_ is somewhat altered. As noted\nabove,[1470] the problem dealt with in the first edition concerns the\nouter world, and only quite indirectly the nature of the self. In the\nsecond edition that argument is restated,[1471] and is more properly\nlocated within the _Analytic_. The argument which now takes its place\nruns parallel with that of the three preceding _Paralogisms_. The\nassertion that I distinguish my own existence as a thinking being from\nother things outside me, including thereunder my own body, is an\nanalytic proposition, since by _other_ things is meant things which I\nthink as different from myself.\n\n     \u201cBut I do not thereby learn whether this consciousness of myself\n     would be at all possible apart from things outside me through which\n     representations are given to me, and whether, therefore, I can\n     exist merely as thinking being (_i.e._ without existing in human\n     form).\u201d\n\nIn =B 417-18= Kant points out that rational psychology, in asserting that\nthe self can be conscious apart from all consciousness of outer things,\ncommits itself to the acceptance of problematic idealism. If\nconsciousness of outer objects is not necessary to consciousness of\nself, there can be no valid method of proving their existence. In the\nfourth _Paralogism_ of the first edition, the inter-dependence of\nrational psychology and empirical idealism is also dwelt upon, but is\nthere traced to a confusion of appearances with things in\nthemselves.[1472]\n\n=B 410-11.=--The correct formulation is here given of what in the first\nedition[1473] is quite incorrectly stated.[1474] A paralogism is a\nsyllogism which errs in logical form (as contrasted with a syllogism\nerring in matter, _i.e._ the premisses of which are false). In the\nparalogisms of Rational Psychology, the logical fallacy committed is\nthat of ambiguous middle, or as Kant names it, the _sophisma figurae\ndictionis_. In the major premiss the middle term is used as referring to\nreal existence, in the minor only as expressive of the unity of\nconsciousness.\n\n=Refutation of Mendelssohn\u2019s Proof of the Permanence of the\nSoul.=[1475]--Mendelssohn\u2019s argument is that the soul, as it does not\nconsist of parts,[1476] cannot disappear _gradually_ by disintegration\ninto its constituent elements. If, therefore, it perishes, it must pass\nout of existence _suddenly_; at one moment it will exist, at the next\nmoment it will be non-existent. But, Mendelssohn maintains, for three\nclosely connected reasons this would seem to be impossible. In the first\nplace, the immediate juxtaposition of directly opposed states is never\nto be met with in the material world. Complete opposites, such as day\nand night, waking and sleeping, never follow upon one another abruptly,\nbut only through a series of intermediate states.[1477] Secondly, among\nthe opposites which material processes thus bridge over, the opposition\nof being and not-being is never to be found. Only by a miracle can a\nmaterial existence be annihilated.[1478] If, therefore, empirical\nevidence is to be allowed as relevant, we must not assert of the\ninvisible soul what is never known to befall the material existences of\nthe visible world. Thirdly--the only part of Mendelssohn\u2019s argument\nwhich Kant mentions--the sudden cessation of the soul\u2019s existence would\nalso violate the law of the continuity of time.[1479] Between any two\nmoments there is always an intermediate time in which the one moment\npasses continuously into the other.\n\nKant\u2019s reply to this third part of Mendelssohn\u2019s argument is that though\nthe soul must not be conceived as perishing suddenly, it may pass out of\nexistence by a continuous diminution through an infinite number of\nsmaller degrees of intensive reality; and in support of this view he\nmaintains the very doubtful position that clearness and obscurity of\nrepresentation are not features of the contents apprehended, but only of\nthe intensity of the consciousness directed upon them.[1480]\n\n=B 417-22.=--Kant here points out that rational psychology, as above\nexpounded, proceeds synthetically, starting from the assertion of the\nsubstantiality of the soul and proceeding to the proof that its\nexistence is independent of outer things. But it may proceed in the\nreverse fashion, analytically developing the implications supposed to be\ninvolved in the \u201cI think,\u201d viewed as an existential judgment, _i.e._ as\nsignifying \u201cI exist thinking.\u201d Kant restates the argument in this\nanalytic form in order, as it would seem, to secure the opportunity of\nreplying to those criticisms of his teaching in the first edition which\nconcern his doctrine of apperception and his employment of the\ncategories, especially of the category of existence, in relation to the\nself. What is new and important in these pages, and also in the\nconnected passages in B 428-30, has been discussed above.[1481]\n\n=B 419-20.=--After remarking that simplicity or unity is involved in the\nvery possibility of apperception, Kant proceeds to argue that it can\nnever be explained from a strictly materialist standpoint, since nothing\nthat is real in space is ever simple. Points are merely limits, and are\nnot therefore themselves anything that can form part of space. The\npassage as a whole would seem to be directed against the Leibnizian\nteaching of Ulrichs.[1482]\n\n=B 426-7.=--Kant makes a remark to which nothing in his argument yields\nany real support, namely, that the dialectical illusion in rational\npsychology is due to the substitution of an Idea of reason for the quite\nindeterminate concept of a thinking being in general. As is argued\nbelow,[1483] the assumption which he is here making that the concept of\nthe self is an _a priori_ and ultimate Idea of pure Reason, cannot be\nregarded as a genuine part of his Critical teaching.\n\n=B 427-8= touches quite briefly upon questions more fully and adequately\ntreated in the first edition. The scanty treatment here accorded to them\nwould seem to indicate, as Benno Erdmann remarks,[1484] that the problem\nof the interaction of mind and body which so occupied Kant\u2019s mind from\n1747 to 1770 has meantime almost entirely lost interest for him. The\nproblem of immortality remains central, but it is now approached from\nthe ethical side.\n\nIn =B 421= and =B 423-6= Kant draws from his criticism of the _Paralogisms_\nthe final conclusion that the metaphysical problems as to the nature and\ndestiny of the self are essentially _practical_ problems. When\napproached from a theoretical standpoint, as curious questions to be\nsettled by logical dialectic, their speculative proof\n\n     \u201c...so stands upon the point of a hair, that even the schools\n     preserve it from falling only so long as they keep it unceasingly\n     spinning round like a top; even in their own eyes it yields no\n     abiding foundation upon which anything could be built.\u201d[1485]\n     \u201cRational psychology exists not as _doctrine_, ... but only as\n     _discipline_. It sets impassable limits to speculative reason in\n     this field, and thus keeps us, on the one hand, from throwing\n     ourselves into the arms of soulless materialism, or, on the other\n     hand, from losing ourselves in an unsubstantial spiritualism which\n     can have no real meaning for us in this present life. But though it\n     furnishes no positive doctrine, it reminds us that we should regard\n     this refusal of Reason to give satisfying response to our\n     inquisitive probings into what is beyond the limits of this present\n     life as a hint from Reason to divert our self-knowledge from\n     fruitless and extravagant speculation to its fruitful practical\n     employment.\u201d[1486] \u201cThe proofs which are serviceable for the world\n     at large preserve their entire value undiminished, and indeed, upon\n     the surrender of these dogmatic pretensions, gain in clearness and\n     in natural force. For Reason is then located in its own peculiar\n     sphere, namely the order of ends, which is also at the same time an\n     order of nature; and since it is in itself a practical faculty\n     which is not bound down to natural conditions, it is justified in\n     extending the order of ends, and therewith our own existence,\n     beyond the limits of experience and of life.\u201d[1487]\n\nThen follows brief indication of the central teaching of the\n_Metaphysics of Ethics_ and of the two later _Critiques_. Through moral\nvalues that outweigh all considerations of utility and happiness, we\nbecome conscious of an inner vocation which inspires feelings of\nsublimity similar to those which are aroused by contemplation of the\nstarry firmament; and to the verities thus disclosed we can add the less\ncertain but none the less valuable confirmation yielded by natural\nbeauty and design, and by the conformity of nature to our intellectual\ndemands.\n\n     \u201cMan\u2019s natural endowments--not merely his talents and the impulses\n     to employ them, but above all else the Moral Law within him--go so\n     far beyond all utility and advantage which he may derive from them\n     in this present life, that he learns thereby to prize the mere\n     consciousness of a righteous will as being, apart from all\n     advantageous consequences, apart even from the shadowy reward of\n     posthumous fame, supreme over all other values; and so feels an\n     inner call to fit himself, by his conduct in this world, and by the\n     sacrifice of many of its advantages, for being a citizen of a\n     better world upon which he lays hold in Idea. This powerful and\n     incontrovertible proof is reinforced by our ever-increasing\n     knowledge of purposiveness in all that we see around us, and by a\n     glimpse of the immensity of creation, and therefore also by the\n     consciousness of a certain illimitableness in the possible\n     extension of our knowledge and of a striving commensurate\n     therewith. All this still remains to us, though we must renounce\n     the hope of ever comprehending, from the mere theoretical knowledge\n     of ourselves, the necessary continuance of our existence.\u201d[1488]\n\n\nIS THE NOTION OF THE SELF A NECESSARY IDEA OF REASON?\n\nOne point of great importance must be dwelt upon before we pass from the\n_Paralogisms_. Though the negative consequences which follow from the\nteaching of the objective deduction are here developed in the most\nexplicit manner, Kant does not within the limits of this chapter, in\neither edition, make any further reference to the doctrine expounded in\nthe introductory sections of the _Dialectic_,[1489] viz. that the notion\nof the self as an immortal being is a necessary Idea of human Reason.\nThe reader is therefore left under the impression that that doctrine is\nunaffected by the destructive criticism passed upon rational psychology,\nand that it still survives as an essential tenet of the Critical\nphilosophy. And he is confirmed in this view when he finds the doctrine\nreappearing in the _Appendix_ to the _Dialectic_ and in the\n_Methodology_. The Idea of the self is there represented as performing a\nquite indispensable, regulative function in the development of the\nempirical science of psychology. Now it is one thing to maintain the\nexistence of Ideal demands of Reason for unity, system and\nunconditionedness, and to assert that it is in virtue of these demands\nthat we are led, in the face of immense discouragement and seeming\ncontradictions, to reduce the chance collocations and bewildering\ncomplexities of ordinary experience to something more nearly\napproximating to what Reason prescribes. But it is a very different\nmatter when Kant claims that in any one sphere, such as that of\npsychology, the unity and the unconditionedness must necessarily be of\none predetermined type. He is then injecting into the Ideals that\n_specific_ guidance which only the detail of experience is really\ncapable of supplying. He is proving false to his own Critical\nempiricism, in which no function is ascribed to Reason that need in any\nway conflict with the autonomy of specialist research; and he is also\nviolating his fundamental principle that the _a priori_ can never be\nother than purely formal. Indeed, when Kant discloses somewhat more in\ndetail what he means by the regulative function of the Idea of the self,\nthe ambiguity of his statements reveals the unconsidered character of\nthis part of his teaching. It is the expression only of a preconception,\nand has eluded the scrutiny of his Critical method largely because of\nthe protective colouring which its admirable adaptation to the needs of\nhis architectonic confers upon it. If, for instance, we compare the\nthree passages in which it is expounded in the _Appendix_ to the\n_Dialectic_, we find that Kant himself alternates between the\nauthoritative prescription to psychology of a spiritualist hypothesis\nand what in ultimate analysis, when ambiguities of language are\ndiscounted, amounts simply to the demand for the greatest possible\nsimplification of its complex phenomena. The passages are as follows.\n\n     \u201cIn conformity with these Ideas as principles we shall first, in\n     psychology, connect in inner experience all appearances, all\n     actions and receptivity of our mind, as if (_als ob_) the mind were\n     a simple substance which persists with personal identity (in this\n     life at least), while its states, to which those of the body belong\n     only as outer conditions, are in continual change.\u201d[1490]\n\n     \u201c...in the human mind we have sensation, consciousness,\n     imagination, memory, wit, power of discrimination, pleasure,\n     desire, etc. Now, to begin with, a logical maxim requires that we\n     should reduce, so far as may be possible, this seeming diversity,\n     by comparing these with one another and detecting their hidden\n     identity. We have to enquire whether imagination combined with\n     consciousness may not be the same thing as memory, wit, power of\n     discrimination, and perhaps even identical with understanding and\n     Reason. Though logic is not capable of deciding whether a\n     _fundamental power_ actually exists, the Idea of such a power is\n     the problem involved in a systematic representation of the\n     multiplicity of powers. The logical principle of Reason calls upon\n     us to bring about such unity as completely as possible; and the\n     more appearances of this or that power are found to be identical\n     with one another, the more probable it becomes that they are simply\n     different manifestations of one and the same power, which may be\n     entitled, relatively speaking, their _fundamental power_. The same\n     is done with the other powers. The relatively fundamental powers\n     must in turn be compared with one another, with a view to\n     discovering their harmony, and so bringing them nearer to a single\n     radical, _i.e._ absolutely fundamental, power. But this unity of\n     Reason is purely hypothetical. We do not assert that such a power\n     must necessarily be met with, but that we must seek it in the\n     interest of Reason, that is, of establishing certain principles for\n     the manifold rules which experience may supply to us. We must\n     endeavour, wherever possible, to bring in this way systematic unity\n     into our knowledge.\u201d[1491]\n\nIn the third of the _Appendix_ passages these two views are confusedly\ncombined. Kant is insisting that an Idea never asserts, even as an\nhypothesis, the existence of a real thing.\n\n     \u201d[An Idea] is only the schema of the regulative principle by which\n     Reason, so far as lies in its power, extends systematic unity over\n     the whole field of experience. The first object of such an Idea is\n     the \u2018I\u2019 itself, viewed simply as thinking nature or soul. If I am\n     to investigate the properties with which a thinking being exists in\n     itself, I must interrogate experience. I cannot even apply any one\n     of the categories to this object, except in so far as its schema is\n     given in sense intuition. But I never thereby attain to a\n     systematic unity of all appearances of inner sense. Instead, then,\n     of the empirical concept (of that which the soul actually is),\n     which cannot carry us far, Reason takes the concept of the\n     empirical unity of all thought; and by thinking this unity as\n     unconditioned and original, it forms from it a concept of Reason,\n     _i.e._ the Idea of a simple substance, which, unchangeable in\n     itself (personally identical), stands in association with other\n     real things outside it; in a word, the Idea of a simple\n     self-subsisting intelligence. Yet in so doing it has nothing in\n     view save principles of systematic unity in the explanation of the\n     appearances of the soul. It is endeavouring to represent all\n     determinations as existing in a single subject, all powers, so far\n     as possible, as derived from a single fundamental power, all change\n     as belonging to the states of one and the same permanent being, and\n     all appearances in space as completely different from the actions\n     of thought. The simplicity and other properties of substance are\n     intended to be only the schema of this regulative principle, and\n     are not presupposed as the real ground of the properties of the\n     soul. For these may rest on altogether different grounds of which\n     we can know nothing. The soul in itself could not be known through\n     these assumed predicates, not even if we regarded them as\n     absolutely valid in regard to it. For they constitute a mere Idea\n     which cannot be represented _in concreto_. Nothing but advantage\n     can result from the psychological Idea thus conceived, if only we\n     take heed that it is not viewed as more than a mere Idea, and that\n     it is therefore taken as valid only in its bearing on the\n     systematic employment of Reason in determining the appearances of\n     our soul. For no empirical laws of bodily appearances, which are of\n     a totally different kind, will then intervene in the explanation of\n     what belongs exclusively to inner sense. No windy hypotheses of\n     generation, extinction, and palingenesis of souls will be\n     permitted. The consideration of this object of inner sense will\n     thus be kept completely pure and unmixed, without employing\n     heterogeneous properties. Also, Reason\u2019s investigations will be\n     directed to reducing the grounds of explanation in this field, so\n     far as may be possible, to a single principle. All this will be\n     best obtained (indeed is obtainable in no other way) through such a\n     schema, viewed as if (_als ob_) it were a real being. The\n     psychological Idea, moreover, can signify nothing but the schema of\n     a regulative principle. For were I to enquire whether _the soul in\n     itself_ is of spiritual nature, the question would have no meaning.\n     In employing such a concept I not only abstract from corporeal\n     nature, but from nature in general, _i.e._ from all predicates of a\n     possible experience, and therefore from all conditions for thinking\n     an object for such a concept: yet only as related to an object can\n     it be said to have a meaning.\u201d[1492]\n\nThe last passage would seem to indicate that Kant has still another and\nonly partially avowed reason for insisting upon a _special_ and\n_spiritualist_ Idea, as regulative of empirical psychology. It is\nnecessary, he would seem to argue, in order to mark off the peculiar\nnature of its subject matter, and to warn us against attempting to\nexplain its phenomena in the mechanistic manner of physical science. But\nif that is Kant\u2019s intention, he has failed to formulate the position in\nany really tenable way. It is impossible to maintain, as he here does,\nthat \u201cno empirical laws of bodily appearances [can] intervene in the\nexplanation of what belongs exclusively to inner sense.\u201d[1493] Indeed,\nin the immediately following sentences, he very clearly indicates how\ncompletely such a position conflicts with his own real teaching. To\nthink away the corporeal is to think away all experience. Experience is\nnot dualistically divided into separate worlds. It is one and single,\nand the principle of causality rules universally throughout, connecting\ninner experiences of sense, feeling, and desire, with their outer\nconditions, organic and physical.[1494] Thus Kant\u2019s retention of the\nIdea of the self is chiefly of interest as revealing the strength and\ntenacity of his spiritualist leanings. We may judge of the\ndisinterestedness and courage of his thinking by the contrary character\nof his pre-conceptions. For even when they have been shown to be\ntheoretically indemonstrable, they continue to retain by honorific title\nthe dignity from which they have been deposed. The full force of the\nobjections is none the less recognised.\n\n     \u201cThe simplicity of substance ... is not presupposed as the real\n     ground of the properties of the soul. For these may rest on\n     altogether different grounds of which we can know nothing.\u201d\n\nThat, however, is only Kant\u2019s unbiassed estimate of the theoretical\nevidence; it is not an expression of his own personal belief.\n\n\n\n\nCHAPTER II\n\nTHE ANTINOMY OF PURE REASON[1495]\n\n\nThis introduction summarises the preceding argument, and distinguishes\nthe new problems of _Antinomy_ from those of the _Paralogisms_. In\nrational psychology pure Reason attains, as it were, euthanasia; in the\nantinomies an entirely different situation is disclosed. For though\nrational cosmology is able to expound itself in a series of demonstrated\ntheses, its teaching stands in irreconcilable conflict with the actual\nnature of appearances, as expressed through a series of antitheses which\nare demonstrable in an equally cogent manner.\n\n\nSECTION I\n\nSYSTEM OF THE COSMOLOGICAL IDEAS[1496]\n\nThe first eight paragraphs of this section are of great textual\ninterest. They must have been written at a time when Kant still intended\nto expound his entire criticism of metaphysical science in the form of a\ndoctrine of antinomy. For they define the Ideas of Reason as exclusively\ncosmological,[1497] and give a very different explanation of their\norigin from that which has been expounded in the preceding chapters.\nEvidently, therefore, this part of the section must have been written\nprior to Kant\u2019s formulation of the metaphysical deduction from the three\nspecies of syllogism. This is supported by the fact that the argument\nbegins anew, just as if the matter had not previously been discussed;\nand that, though a new view of the nature of Reason is propounded, there\nis not the least mention of the more Idealist view which it displaces.\nReason, Kant here teaches, is not a faculty separate from the\nunderstanding, and does not therefore produce any concept peculiar to\nitself. Reason is simply a name for the understanding in so far as it\nacts independently of sensibility, and seeks, by means of its pure\nforms, in abstraction from all empirical limitations, to grasp the\nunconditioned. \u201cThe transcendental Ideas are in reality nothing but\ncategories extended to the unconditioned.\u201d The intelligible, as thus\nconceived by the understanding, expresses itself, as he later shows, in\na series of theses; while the sensuous expresses its opposite and\nconflicting character in a series of antitheses.\n\nYet not all categories yield a concept of the unconditioned. That is\npossible only to those which concern themselves with a series of members\nconditioning and conditioned, and in reference to which, therefore, the\npostulate of an unconditioned would seem to be legitimate, viz.: (1)\nunconditioned quantity in space and time; (2) unconditioned quality\n(indivisibility and simplicity) of reality in space (matter); (3)\nunconditioned causality of appearances; (4) unconditioned necessity of\nappearances. As this arrangement is determined by the needs of Kant\u2019s\narchitectonic, no detailed comment is here called for. Its consequences\nwe shall have ample opportunity to consider later. As already noted,\nKant\u2019s statement in A 414 = B 441, that \u201cthe category of substance and\naccident does not lend itself to a transcendental Idea,\u201d shows very\nclearly that, at the time when he composed this passage, he had not yet\nbethought himself of placing a separate and independent Idea at the\nbasis of rational psychology. But as Kant here strives to follow the\nfourfold arrangement of the categories, the content of these paragraphs\nmust either have been later recast or have been composed in the interval\nbetween his discovery of the metaphysical deduction of the categories\nand his formulation of the corresponding deduction of the Ideas from the\nthree forms of syllogism. It may also be observed that the derivation of\nthe cosmological Idea from the hypothetical syllogism, which embodies\nonly the category of causality, clashes with the above specification of\nit in terms of all four rubrics of category.\n\nThe remaining paragraphs (ninth to thirteenth) of this section must be\nof later date, as they are developed in view of the independent\ntreatment of the theological Ideal.[1498] (Adickes, in dating the ninth\nand tenth paragraphs with the preceding instead of with the concluding\nparagraphs, would seem to have overlooked this fact.) In order to\njustify the treatment of the Ideas of a first cause and of unconditioned\nnecessity, as _cosmological_, Kant now asserts that the antinomies\nconcern only appearances--\u201cour [cosmical] Ideas being directed only to\nwhat is unconditioned _among the appearances_,\u201d[1499] _and not to\nnoumena_.[1500] His explanation of the nature of transcendental\nillusion, and of the antinomies in particular, as being due to a failure\nto distinguish between appearance and things in themselves, is thus\nruthlessly sacrificed to considerations of architectonic. Kant could\nnot, of course, consistently hold to the position here adopted; but it\ncauses him from time to time, especially in dealing with the third and\nfourth antinomies, to make statements which tend seriously to obscure\nthe argument and to bewilder the careful reader.\n\nKant is far from clear as to the relation in which the concepts of the\ntotality of conditions and of the unconditioned stand to one\nanother.[1501] In A 322 = B 379 they would seem to be taken as exactly\nequivalent concepts. In A 416-17 = B 443-5 they are apparently regarded\nas distinct, the former only leading up to the latter. But discussion of\nthis important point must meantime be deferred.[1502]\n\n\nSECTION II\n\nANTITHETIC OF PURE REASON[1503]\n\n     \u201d[Antithetic] is the conflict between two apparently dogmatic\n     judgments [_Erkenntnisse_] to neither of which can we ascribe any\n     superior claim to acceptance over the other, _i.e._ by Antithetic I\n     mean a _thesis_, together with an _antithesis_.\u201d \u201cTranscendental\n     Antithetic is an investigation of the antinomy of pure Reason, its\n     causes and outcome.\u201d\n\nThe very existence of such antinomy presupposes a twofold condition:\nfirst, that it does not refer to a gratuitous but to an inevitable\nproblem of human Reason, \u201cone which it must necessarily encounter in its\nnatural progress\u201d; and secondly, that the thesis and the antithesis\ntogether generate a \u201cnatural and inevitable illusion,\u201d which continues\nto persist even after its deceptive power has been clearly disclosed.\nSuch conflict is caused by the fact that Reason seeks a unity which\ntranscends the understanding, and which nevertheless is meant to\nconform to the conditions of the understanding. If the unity is adequate\nto the demands of Reason, it is too great for the understanding; if it\nis commensurate with the understanding, it is too small for\nReason.[1504] The theses express the higher unity at which Reason aims;\nthe antitheses are the judgments to which the understanding is\nconstrained by the nature of the appearances with which both it and\nReason profess to deal. If we hold to Reason, we make assertions\ncontradictory of the appearances; while if we place reliance on the\nunderstanding, Reason condemns our conclusions.\n\nThis conflict is limited to those few problems above enumerated in which\nwe are called upon to complete a given series.[1505] Since totality,\nwhether in the form of a first beginning of the series or as an actual\ninfinity of the whole series, can never itself be experienced, these are\nproblems in regard to which experience can be of no assistance to us. It\ncan neither confirm nor refute any particular solution. The only\npossible method of deciding between the competing claims is to watch or\neven to provoke the conflict, in the hope that we may finally be able to\ndetect some misunderstanding, and so to resolve the conflict to the\nsatisfaction of both the litigants. Such is Kant\u2019s description of what\nhe entitles his \u201csceptical method.\u201d[1506]\n\nWithout here attempting a full discussion of the subject, it seems\nadvisable to point out at the very start what Kant\u2019s exposition\nseriously obscures, namely, the real character of the evidence upon\nwhich the theses and the antitheses respectively rest. The latter are\nnot correctly stated as transcending experience, and as therefore\nincapable of confirmation by it. The proofs which Kant offers of them\nare, indeed, of a non-empirical _a priori_ character. They are\nformulated in terms of the dogmatic rationalism of the Leibnizian\nposition, with a constant appeal to abstract principles. But, as a\nmatter of fact, they can be much more adequately established--in so far\nas they can be established at all--through analysis of the spatial and\ntemporal conditions of material existence. As space and time are\ncontinuous and homogeneous, any assertion which is true of a space or\ntime however small is likewise true of a space or time however large.\nAny space consists of spaces, and must be regarded as itself part of a\nlarger whole.[1507] Any time consists of parts which are themselves\ntimes, and is apprehensible only as following upon preceding times. It\nis by such considerations as these that we are led to regard the\nmaterial world as unlimited, as infinitely divisible, and as having no\nfirst state.\n\nKant\u2019s method of demonstrating the theses--that the world is limited, is\nfinitely divisible, and has a first state--is no less misleading. Here\nagain his rationalistic arguments conceal the basis upon which the\nvarious theses really rest. Their true determining ground is the demand\nof Reason for some more satisfactory form of unconditionedness than that\nwhich is found in the actual infinite. It is this demand which has led\nphilosophers to look around for proofs in support of the theses, and to\nelaborate those rationalistic arguments which Kant here reproduces. Thus\nthe grounds of the antitheses are altogether different from those of the\ntheses; and in neither case are they properly represented by the\narguments which Kant employs.[1508]\n\nThe reasons why Kant in his detailed statement of the antinomies has\nomitted, or at least subordinated, the above considerations, are complex\nand various. In the first place, this doctrine of antinomy was in\nseveral of its main features already formulated prior to his development\nof the Critical philosophy. It forms part of his _Dissertation_ of 1770;\nand at that time Kant was still largely in fundamental sympathy with the\nLeibnizian ontology. Secondly, Kant is here professing to criticise the\nscience of rational cosmology, and is therefore bound to expound it in\nmore or less current form. And in the third place, he teaches that _the\nantinomies exist as antinomies only when viewed from the false\nstandpoint of dogmatic rationalism_. Had he eliminated the rationalistic\nproofs, the conflict of the antinomies, in its strictly logical form,\nas the conflict of direct contradictories, would at once have vanished.\nThe general framework of this division of the _Dialectic_ demanded a\nrationalistic treatment of both theses and antitheses, and Kant believed\nthat the rationalistic proofs which he propounds in their support are\nunanswerable, so long as the dogmatic standpoint of ordinary\nconsciousness and of Leibnizian ontology is preserved. But even when\nthat important limitation is kept in view, Kant fails to justify this\ninterpretation of the conflict, and we must therefore be prepared to\nfind that his proofs, _whether of theses or of antitheses_, are in all\ncases inconclusive. I shall append to each of his arguments a statement\nof the reasons which constrain us to reject them as unsound. We shall\nthen be in a position to consider his whole doctrine of antinomy in its\nbroader aspects, and in its connection with the teaching of the other\nmain divisions of the _Dialectic_.\n\n\nFIRST ANTINOMY\n\n=Thesis.=--(_a_) The world has a beginning in time, and (_b_) is also\nlimited in regard to space.\n\n=Thesis= _a._ =Proof.=--If we assume the opposite, namely, that the world\nhas no beginning in time, and if we define the infinite as that which\ncan never be completed by means of a successive synthesis, we must\nconclude that the world-series can never complete itself. But the entire\nseries of past events _elapses_, _i.e._ completes itself at each moment.\nIt cannot therefore be infinite.\n\n=Criticism.=--This argument gains its plausibility from the illegitimate\nuse of the term \u2018elapse\u2019 (_verfliessen_) as equivalent to \u2018complete\nitself.\u2019 If it be really correct to define the infinite as that which\ncan never be completed, the conclusion to be drawn is that the temporal\nseries is always actually infinite, and that no point or event in it is\nnearer to or further from either its beginning or its end.[1509] We may\nselect any point in the series as that from which we propose to begin a\nregress to the earlier members of the series, but if the series is\nactually infinite, it will be a regress without possibility of\ncompletion, and one therefore which removes all justification for\nasserting that at the point chosen a series has completed itself. It has\nno beginning, and has no completion. What it has done at each moment of\nthe past it is still doing at each present moment, namely, coming out of\nan inexhaustible past and passing into an equally inexhaustible future.\nTime is by its given nature capable of being interpreted only as\nactually infinite, alike in its past and in its future. It cannot\ncomplete itself any more than it can begin itself. The one would be as\ngross a violation of its nature as would the other. The present exists\nonly as a species of transition, unique in itself, but analogous in\nnature to the innumerable other times that constitute time past. It is a\ntransition from the infinite through the infinite to the infinite. That\nwe cannot comprehend how, from an infinitude that has no beginning, the\npresent should ever have been reached, is no sufficient reason for\ndenying what by the very nature of time we are compelled to accept as a\ncorrect description of the situation which is being analysed. The actual\nnature of time is such as to rule out from among the possibilities the\nthesis which Kant is here professing to be able to establish; time,\nbeing such as it actually is, can have no beginning.\n\nWhat thus holds of time may likewise hold of events in time. If time is\nactually infinite, no proof can be derived from it in support of the\nassumption that the world has had a beginning in time.\n\nThe phrase \u201cby means of a successive synthesis\u201d gives a needlessly\nsubjectivist colouring to Kant\u2019s method of proof. The antinomy is\nprofessedly being stated from the realist standpoint, and ought not\ntherefore to be complicated by any such reference. This objection\napplies, as we shall find, still more strongly to Kant\u2019s proof of the\nsecond part of the thesis. The latter proof depends upon this\nsubjectivist reference; the present proof does not.\n\nKant limits his problem to the past infinitude of time. The reason for\nthis lies, of course, in the fact that he is concerned with the problem\nof creation. The limitation is, however, misleading.\n\n=Thesis= _b._--The world is limited in regard to space.\n\n=Proof.=--Assume the opposite, namely, that the world is an infinite,\ngiven whole of coexisting parts. A magnitude not given within the\ndeterminate limits of an intuition can only be thought through the\nsynthesis of its parts, and its totality through their completed\nsynthesis. In order, therefore, that we may be able to think as a single\nwhole the world which fills all space, the successive synthesis of the\nparts of an infinite world must be regarded as completed, _i.e._ an\ninfinite must be regarded as having elapsed in the enumeration of all\ncoexisting things. This, however, is impossible. An infinite aggregate\nof actual things cannot therefore be viewed as a given whole, nor as\nbeing given as coexistent. Consequently the world of spatial existences\nmust be regarded as finite.\n\n=Criticism.=--From the impossibility of traversing infinite space in\nthought by the successive addition of part to part, Kant here argues\nthat \u201can infinite aggregate of actual things cannot be viewed as a given\nwhole,\u201d and consequently that the world cannot be infinitely extended in\nspace. That is, from a _subjective impossibility of apprehension_ he\ninfers an _objective impossibility of existence_. But Kant has himself\ndefined the infinite as involving this subjective impossibility; for in\nthe proof of thesis a he has stated that the infinitude of a series\nconsists in the very fact that it can never be completed through\nsuccessive synthesis. Kant is therefore propounding against the\n_existence_ of the infinite the very feature which by definition\nconstitutes its infinitude. The implication would seem to be that the\nconcept of the infinite is the concept of that which _ex definitione_\ncannot exist, and that there is therefore a contradiction in the very\nidea of the actual infinite.\n\nDeferring for a moment the further objections to which such procedure\nlies open, we may observe that Kant, in arguing from a subjective to an\nobjective impossibility, commits the fallacy of _ignoratio elenchi_. For\nwhen the conditions of objective existence are recognised in their\ndistinction from those of mental apprehension, the supposed\ncontradiction vanishes, and the argument ceases to have any cogency. The\nuse of the words \u2018given\u2019 and \u2018whole\u2019 is misleading. If space is\ninfinite, it is without bounds, and cannot therefore exist as a whole in\nany usual meaning of that term. For the same reason it must be incapable\nof being given as a whole. Its infinitude is a presupposition which\nanalysis of actually given portions of it constrains us to postulate,\nand has to be conceived in terms of the definition employed in thesis\n_a_. The given must always be conceived as involving what is not itself\ngiven and what is not even capable of complete construction. In terms of\nthis presupposition an actual infinite, not given and not capable of\nconstruction, can be represented with entire consistency.\n\nBut to return to the main assumption upon which Kant\u2019s proof would seem\nto rest: it is all-important to observe that Kant does not, either in\nthe _Critique_ or in any other of his writings, assert that the concept\nof the actual infinite is inherently self-contradictory. This is a\nmatter in regard to which many of Kant\u2019s critics have misrepresented his\nteaching. Kant\u2019s argument may, as we have just maintained, be found on\nexamination to involve the above assertion; but this, if clearly\nestablished, so far from commending the argument to Kant, would have led\nhim to reject it as invalid. The passage in the _Dissertation_[1510] of\n1770, which contains his most definite utterance on this point,\nrepresents the view from which he never afterwards departed. It may be\nquoted in full.\n\n     \u201cThose who reject the actual mathematical infinite do so in a very\n     casual manner. For they so construct their definition of the\n     _infinite_ that they are able to extract a contradiction from it.\n     The infinite is described by them as _a quantity than which none\n     greater is possible_, and the mathematical infinite as a\n     multiplicity--of an assignable unit--than which none greater is\n     possible. Since they thus substitute _maximum_ for _infinitum_, and\n     a greatest multiplicity is impossible, they easily conclude against\n     this infinite which they have themselves invented. Or, it may be,\n     they entitle an infinite multiplicity an _infinite number_, and\n     point out that such a phrase is meaningless, as is, indeed,\n     perfectly evident. But again they have fought and overthrown only\n     the figments of their own minds. If, however, they had conceived\n     the mathematical infinite as a quantity which, when related to\n     measure, as its unity, is _a multiplicity greater than all number_;\n     and if furthermore, they had observed that _measurability_ here\n     denotes only the relation [of the infinite] to the standards of the\n     human intellect, which is not permitted to attain to a _definite\n     conception of multiplicity_ save by the successive addition of unit\n     to unit, nor to the _sum-total_ (which is called _number_) save by\n     completing this progress in a finite time; they would have\n     perceived clearly that what does not conform to the established law\n     of some subject need not on that account exceed all intellection.\n     An intellect may exist, though not indeed a human intellect, which\n     perceives a multiplicity distinctly in one intuition [_uno obtutu_]\n     without the successive application of a measure.\u201d\n\nThe concluding sentences of this _Dissertation_ passage may be taken as\nKant\u2019s own better and abiding judgment in regard to the question before\nus. We must not argue from the impossibility of mentally traversing the\ninfinite to the impossibility of its existence. Indeed the essentials\nof the above passage are restated in the \u2018Observation\u2019 on this\nthesis.[1511] Thus the concept of the actual infinite is not only, as a\nconcept, perfectly self-consistent, it is also one which, in view of the\nnature of time and of space, we are constrained to accept as a correct\nrepresentation of the actually given. The thesis of this first antinomy\nruns directly counter to admitted facts. That Kant is here arguing in\nrespect to the world, and not merely in respect to space and time, does\nnot essentially alter the situation. For if space and time are\nnecessarily to be viewed as infinite, there can be no _a priori_\nproof--none, at least, of the kind here attempted--that the world-series\nmay not be so likewise.\n\n=Antithesis.=--(_a_) The world has no beginning in time; (_b_) has no\nlimits in space. In both these respects the world is infinite.\n\nIn these antitheses Kant assumes that space and time are actually\ninfinite, and from that assumption advances to the proof that this is\nlikewise true of the world in its spatial and temporal aspects. This, by\nitself, ought to be sufficient evidence that Kant does not regard the\nactual infinite as an inherently impossible conception. As the\nantinomies are avowedly formulated from the realist, dogmatic standpoint\nof ordinary consciousness, Kant is also enabled to assume that if the\nworld begins to be, it must have an antecedent cause determining it to\nexist at that moment rather than at another.\n\n=Antithesis= _a._ =Proof.=--Let us assume the opposite, namely, that the\nworld has a beginning. It will then be preceded by an empty time in\nwhich it was not. But in an empty time no becoming is possible, since in\nsuch a time no part possesses over any other any distinguishing\ncondition of existence rather than of non-existence. The world must\ntherefore be infinite as regards past time.\n\n=Criticism.=--In this argument everything depends upon what is to be meant\nby the term \u2018world.\u2019 If Kant means by it merely the material world, the\nassumption of its non-existence does not leave only empty time and\nspace. Other kinds of existence may be possible, and in these a\nsufficient cause of its first beginning may be found. The nature of\ncreative action will remain mysterious and incomprehensible, but that is\nno sufficient reason for denying its possibility. If, on the other hand,\nKant means by the world \u2018all that is,\u2019 the assumption of its\nnon-existence is likewise the assumption of the non-existence of all\nits possible causes. That, however, is for ordinary consciousness a\nquite impossible assumption, since it runs counter to the causal\nprinciple which is taken as universally valid. From this point of view\nthe argument consists in making an impossible assumption, and in then\npointing out the impossible consequence which must follow. By such a\nmode of argument no conclusion can be reached. Kant\u2019s decision ought\nrather to have been that, as time is actually infinite, the world may be\nso likewise, but that though _reality_ must in some form be eternally\nexistent, the _material world_ cannot be proved to be so by any _a\npriori_ proof of the kind here given.\n\n=Antithesis= _b._ =Proof.=--Let us assume the opposite, namely, that the\nworld is finite, existing in an empty limitless space. There will then\nbe not only a relation of things _in space_, but also of things _to\nspace_. But as the world is a totality outside of which no object of\nintuition can be found, the relation of the world to empty space is a\nrelation to _no object_. Such a relation is nothing. Consequently the\nopposite holds; the world must be infinitely extended.\n\n=Criticism.=--That Kant himself felt the inadequacy of this argument, when\ntaken from the dogmatic standpoint, is indicated by the lengthy note\nwhich he has appended to it, and which develops his own Critical view of\nspace as not a real independent object, but merely the form of external\nintuition. From the standpoint of ordinary consciousness space is a\nself-existent entity, and there is no insuperable difficulty in\nconceiving a relation as holding between it and its contents. The\nintroduction of the opposed standpoint of the _Aesthetic_ therefore runs\ndirectly counter to Kant\u2019s own intention of expounding the antinomies\nfrom the dogmatic standpoint which involves this realist view of space,\nand of showing that they afford, in independence of the arguments of the\n_Aesthetic_, an indirect proof of the untenableness of that\nbelief.[1512] The conclusion which ought to have been drawn is analogous\nto that above suggested for thesis _a._ As space is actually infinite,\nthe material world may be so likewise; but that it actually is so,\ncannot be established by an _a priori_ argument of the kind here\nattempted.\n\n\nSECOND ANTINOMY\n\n=Thesis.=--Every composite substance in the world consists of simple\nparts, and nothing anywhere exists save the simple or what is composed\nof it.\n\n=Proof.=--Let us assume the opposite, namely, that substances do not\nconsist of simple parts. If all composition be then removed in thought,\nno composite part, and (as there are no simple parts) also no simple\npart, and therefore nothing whatsoever, will remain. Consequently no\nsubstance will be given. Either, therefore, it is impossible to remove\nin thought all composition, or after its removal something that exists\nwithout composition, _i.e._ the simple, must remain. In the former case\nthe composite would not itself consist of substances (with them\ncomposition is a merely accidental relation, and they must, as\nself-persisting beings, be able to exist independently of it). As this\ncontradicts our assumption, only the latter alternative remains, namely,\nthat the substantial compounds in the world consist of simple parts.\n\n=Criticism.=--Kant here assumes, by his definition of terms, the point\nwhich he professes to establish by argument. The substance referred to,\nthough never itself mentioned by name, is extended matter. Kant\nidentifies it with \u2018composite substance.\u2019 Substance, he further\ndogmatically decides, is that which is capable of independent existence,\nand to which all relations of composition are therefore merely\naccidental. If these assumptions be granted, it at once follows that\ncomposition cannot be essential to matter, and that when all composition\nis thought away, its reality will be disclosed as consisting in simple\nparts. Kant, however, makes no attempt to prove that extended matter can\nbe defined in any such terms. From the dogmatic point of view of\nordinary consciousness, though not from the sophisticated standpoint of\nLeibniz, extension is of the very essence of matter; and, as Kant\nhimself believed,[1513] the continuity of extension is such as to\nexclude all possibility of elimination of the composite. For he\nmaintains that, however far division be carried, the parts remain no\nless composite than the whole from which the regress has started. On any\nsuch view the extended and the composite are not equivalent terms. The\nopposite of the composite is the simple; the opposite of the extended is\nthe non-extended. Kant is here surreptitiously substituting a Leibnizian\nmetaphysics in place of the empirical reality which is supposed to\nnecessitate the argument.\n\nIn the _Observation_ on this thesis Kant shows consciousness of the\ndefects of his argument. It does not apply to space, time, or change.\n\n     \u201cWe ought not to call space a _compositum_ but a _totum_, because\n     its parts are possible only in the whole, not the whole through the\n     parts.\u201d[1514]\n\nAs Kant further states, he is speaking only of the simples of the\nLeibnizian system. This thesis is \u201cthe dialectical principle of\nmonadology.\u201d Again in the _Observation_ on the antithesis, in commenting\non the mathematical proof of the infinite divisibility of matter, Kant\neven goes so far as to declare that the argument of the thesis is based\non an illegitimate substitution of things in themselves, conceived by\nthe pure understanding, for the appearances with which alone the\nantinomy is concerned.[1515]\n\n     \u201c...it is quite futile to attempt to overthrow, by sophistical\n     manipulation of purely discursive concepts, the manifest,\n     demonstrated truth of mathematics.\u201d\n\n=Antithesis.=--No composite thing in the world consists of simple parts,\nand there nowhere exists in the world anything simple.\n\n=Proof.=--Let us assume the opposite, namely, that a composite thing (as\nsubstance) consists of simple parts. As all external relation, and\ntherefore all composition of substances, is only possible in space,\nspace must consist of as many parts as there are parts of the composite\nthat occupies it. Space, however, does not consist of simple parts, but\nof spaces. The simple must therefore occupy a space. Now as everything\nreal which occupies a space contains in itself a manifold of\nconstituents external to one another, and therefore is composite, and as\na real composite is not composed of accidents (for without substance\naccidents could not be outside one another), but of substances, the\nsimple would be a substantial composite, which is self-contradictory.\n\n=Criticism.=--The Leibnizian standpoint is here completely deserted.\nInstead of proceeding to demonstrate the direct opposite of the thesis,\nKant in this argument deals with the extended bodies of empirical\nintuition. The proof given ultimately reduces to an argument from the\ncontinuous nature of space to the continuous nature of the matter which\noccupies it. But as the thesis and the antithesis thus refer to\ndifferent realities, the former to things in themselves conceived by\npure understanding, and the latter to the sensuous, no antinomy has\nbeen shown to subsist. Antinomy presupposes that both the opposing\nassertions have the same reference. Kant, as already noted, argues in\nthe _Observation_ to this antithesis that all attempts \u201cmade by the\nmonadists\u201d to refute the mathematical proof of the infinite divisibility\nof matter are quite futile, and are due to their forgetting that in this\ndiscussion we are concerned only with appearances.\n\n     \u201cThe monadists have, indeed, been sufficiently acute to seek to\n     avoid this difficulty by not treating space as a condition of the\n     possibility of the objects of outer intuition (bodies), but by\n     taking these and the dynamical relation of substances as the\n     condition of the possibility of space. But we have a concept of\n     bodies only as appearances, and as such they necessarily presuppose\n     space as the condition of the possibility of all outer\n     appearance.\u201d[1516]\n\nHow Kant, after writing these words, should still have left standing the\nproof which he has given of the thesis may be partially explained as due\nto the continuing influence of his earlier view,[1517] according to\nwhich antinomy represents not a conflict between opposing views of the\nworld of ordinary consciousness, but between the demands of pure thought\nand the forms of sensuous existence. That older view of antinomy here\ngains the upper hand, notwithstanding its lack of agreement with the\ngeneral scheme of the _Dialectic_.\n\nThere is a further inconsistency in Kant\u2019s procedure which may perhaps\nbe taken as indicating the early origin of this portion of the\n_Critique_. He presents the mathematical proof of the continuity of\nmatter as conclusive. Yet in the _Metaphysical First Principles of\nNatural Science_ (1786) he most emphatically states that \u201cthe infinite\ndivisibility of matter is very far from being proved through proof of\nthe infinite divisibility of space.\u201d[1518]\n\nRussell,[1519] in discussing the thesis and antithesis on their merits,\nfrom the point of view of certain present-day mathematical theories,\nmakes the following criticism of Kant\u2019s procedure.\n\n     \u201cHere, again, the argument applies to things in space and time, and\n     to all collections, whether existent or not.... And with this\n     extension[1520] the proof of the proposition must, I think, be\n     admitted; only that _terms_ or _concepts_ should be substituted for\n     _substances_, and that, instead of the argument that relations\n     between substances are accidental (_zuf\u00e4llig_), we should content\n     ourselves with saying that relations imply terms and complexity\n     implies relations.\u201d\n\nRussell further argues that Kant\u2019s assumption in the antithesis, that\n\u201cspace does not consist of simple parts, but of spaces,\u201d cannot be\ngranted. It\n\n     \u201c...involves a covert use of the axiom of finitude, _i.e._ the\n     axiom that, if a space does consist of points, it must consist of\n     some finite number of points. When once this is denied, we may\n     admit that no finite number of divisions of a space will lead to\n     points, while yet holding every space to be composed of points. A\n     finite space is a whole consisting of simple parts, but not of any\n     finite number of simple parts. Exactly the same thing is true of\n     the stretch between 1 and 2. Thus the antinomy is not specially\n     spatial, and any answer which is applicable in Arithmetic is\n     applicable here also. The thesis, which is an essential postulate\n     of Logic, should be accepted, while the antithesis should be\n     rejected.\u201d\n\nBut, as above observed,[1521] those mathematicians who adopt this view\nso alter the meaning of the term point that it would perhaps be equally\ntrue to say that the thesis, as thus interpreted by Russell, coincides\nwith what Kant believes himself to be asserting in the antithesis.\n\n\nTHIRD ANTINOMY\n\n=Thesis.=--Causality according to the laws of nature is not the only\ncausality from which the appearances of the world can be deduced. There\nis also required for their explanation another, that of freedom.\n\n=Proof.=--Let us assume the opposite. In that case everything that happens\npresupposes a previous state upon which it follows according to a rule.\nThat previous state is itself caused in similar fashion, and so on _in\ninfinitum_. But if everything thus happens according to the mere laws of\nnature, there can never be a first beginning, and therefore no\ncompleteness of the series on the side of the derivative causes. But the\nlaw of nature is that nothing happens without a cause _sufficiently_\ndetermined _a priori_. If, therefore, all causality is possible only\naccording to the laws of nature, the principle contradicts itself when\ntaken in unlimited universality. Such causality cannot therefore be the\nsole causality possible. We must admit an absolute spontaneity, whereby\na series of appearances, that proceed according to laws of nature,\nbegins by itself.\n\n=Criticism.=--The vital point of this argument lies in the assertion that\nthe principle of causality calls for a _sufficient_ cause for each\nevent, and that such sufficiency is not to be found in natural causes\nwhich are themselves derivative or conditioned. As the antecedent series\nof causes for an event can never be traced back to a first cause, it can\nnever be completed, and can never, therefore, be sufficient to account\nfor the event under consideration. Either, therefore, the principle of\ncausality contradicts itself, or some form of free self-originative\ncausality must be postulated. This argument cannot be accepted as valid.\nEach natural cause is sufficient to account for its effect. That is to\nsay, the causation is sufficient _at each stage_. That the series of\nantecedent causes cannot be completed is due to its actual infinitude,\nnot to any insufficiency in the causality which it embodies.[1522] To\nprove his point, Kant would have to show that the conception of the\nactual infinite is inherently self-contradictory; and that, as we have\nalready noted, he does not mean to assert. His argument here lies open\nto the same criticism as we have already passed upon his argument in\nproof of the thesis of the first antinomy.\n\n=Antithesis.=--There is no freedom; everything in the world proceeds\nsolely in accordance with laws of nature.\n\n=Proof.=--Let us assume the opposite. Free causality, _i.e._ the power of\nabsolute origination, presupposes the possibility of a state of the\ncause which has no causal connection with its preceding state, and which\ndoes not follow from it. But this is opposed to the law of causality,\nand would render unity of experience impossible. Freedom is therefore an\nempty thought-entity (_Gedankending_), and is not to be met with in any\nexperience.\n\n=Criticism.=--We may first observe the strange relation in which the proof\nof the thesis stands to that of the antithesis. According to the former,\nfreedom must be postulated because otherwise the principle of causality\nwould contradict itself. According to the latter, freedom is impossible,\nand for the same reason. Now, as Erhardt has pointed out,[1523] a\nprinciple cannot be reconciled with itself through the making of an\nassumption which contradicts it. That would only be the institution of a\nsecond contradiction, not the removal of the previous conflict. If the\nproof of the thesis be correct, that of the antithesis must be false; if\nthe proof of the antithesis be correct, that of the thesis must be\ninvalid. For though the thesis and the antithesis may themselves\ncontradict one another, such conflict must not exist between the grounds\nupon which they establish themselves. If the reasons cited in their\nsupport are contradictory of one another, the total argument is rendered\nnull and void. The supporting proofs being contradictory of one another,\nnothing whatsoever has been established. There will remain as a pressing\nand immediate problem the task of distinguishing the truth from among\nthe competing alternatives; and until this has been done, the argument\ncannot proceed. The assumption of freedom either does or does not\ncontradict the principle of causality. Antinomy is not the simple\nassertion that both A and not-A are true, but that A and not-A, though\ncontradictory of one another, can both be established by arguments in\nwhich such contradiction does not occur.[1524]\n\nThe proof given of the thesis would seem, as already noted, to be\nuntenable. The principle of natural causality is not self-contradictory.\nWhat now is to be said regarding the proof of the antithesis? If the\nprinciple of natural causality be formulated as asserting that every\nevent has an _antecedent_ cause determining it to exist, then certainly\nfree, spontaneous, or self-originating causality is excluded. Here, as\nin Kant\u2019s proof of the antithesis of the first antinomy, everything\ndepends upon definition of the terms employed. It must be borne in mind\nthat the antinomies are asserted to exist only on the dogmatic level.\nCritical considerations must not, therefore, be allowed to intervene.\nNow for ordinary consciousness the concept of causality has a very\nindefinite meaning, and a very wide application. Causation may be\nspontaneous as well as mechanical, spiritual as well as material. All\npossibilities lie open, and no mere reference to the concept of causal\ndependence suffices to decide between them. Free causality, so far as\n_dogmatic_ analysis of the causal postulate can show to the contrary,\nmay or may not be possible.[1525] Kant has failed to establish the\nantithesis save by the surreptitious introduction of conclusions which\npresuppose the truth of his Critical teaching. This is especially shown\nin the emphasis laid upon \u2018unity of experience.\u2019 The further\nstatement[1526] that freedom means lawlessness is only true if Kant\u2019s\nteaching is mutilated by reduction merely to its assertion of the\nobjective validity of the mechanistic principles of natural science.\nKant is both running with the hare and hunting with the hounds.\n\nThough this antinomy is chiefly concerned with the problem of freedom,\n_i.e._ of spontaneous origination _within_ the world, the proof of the\nthesis refers only to the cosmological problem of a first cause.[1527]\nThe reasons of this oscillation we shall have occasion to consider in\ndealing with the fourth antinomy. The terms world and nature play the\nsame ambiguous part as in the antithesis of the first antinomy; they\ntend to be employed in the narrower, mechanistic sense of Kant\u2019s own\nCritical teaching.\n\n\nFOURTH ANTINOMY\n\nAs the proofs of the thesis and antithesis proceed on lines identical\nwith those of the third antinomy, I shall omit detailed statement of\nthem.[1528] Kant again argues from the fact that every change has a\ncondition which precedes it in time. There is no difference in the\nproofs themselves, but only in the nature of the inference which they\nare made to support. In the third antinomy they lead to the assertion\nand denial of free causality; in the fourth antinomy they lead to the\nassertion and denial of an absolutely necessary being. The assertion is\nrequired in order to save the principle of causality from\nself-contradiction; the denial is also necessary, and for the same\nreason. The illegitimacy of this procedure has already been pointed\nout.[1529] Though the thesis and the antithesis will, if antinomy be\nassumed to represent an actual conflict, contradict one another, no\nsuch conflict is allowable in the grounds which profess to establish\nthem. We must not assert, as _argument_, that both A and not-A are true.\n\nIn the _Observation_ on the antithesis[1530] Kant has himself taken\nnotice of this \u201cstrange\u201d situation.\n\n     \u201cFrom the same ground on which, in the thesis, the existence of an\n     original being was inferred, its non-existence is inferred, and\n     that with equal stringency.\u201d\n\nA necessary being is inferred to exist, because the past series of\nevents cannot contain all the conditions of an event, unless the\nunconditioned is to be found among them. A necessary being is denied to\nexist, because the series of merely conditioned events contains all the\nconditions that there are. Kant\u2019s defence of this procedure is as\nfollows:\n\n     \u201cNevertheless, the method of argument in both cases is entirely in\n     conformity even with ordinary human reason, which frequently falls\n     into conflict with itself from considering its object from two\n     different points of view. M. de Mairan[1531] regarded the\n     controversy between two famous astronomers, which arose from a\n     similar difficulty in regard to choice of standpoint, as a\n     sufficiently remarkable phenomenon to justify his writing a special\n     treatise upon it. The one had argued that _the moon revolves on its\n     own axis_, because it always turns the same side towards the earth.\n     The other drew the opposite conclusion that _the moon does not\n     revolve on its own axis_, because it always turns the same side\n     towards the earth. Both inferences were correct, according to the\n     point of view which each chose in observing the moon\u2019s motion.\u201d\n\nThis example is not really relevant. In spite of Kant\u2019s assertion to the\ncontrary, the point of view is one and the same in thesis and in\nantithesis. In both cases the absolutely necessary being is viewed as\nthe first of the changes in the world of sense. To maintain that when\nthus viewed it both is and is not demanded by the law of causality, is\nas impossible as to assert that in one and the same meaning of our terms\nthe moon both does and does not revolve on its own axis.\n\nThat the proofs of the fourth antinomy are identical with those of the\nthird is due to the fact that Kant, under the stress of his\narchitectonic,[1532] is striving to construct four antinomies while only\nthree are really distinguishable. The third and fourth antinomies\ncoincide as formulations of the problem whether or not the conditioned\nimplies, and originates in, the unconditioned. The precise\ndetermination of this unconditioned, whether as free causality or as a\nnecessary being, or in any other way, is a further problem, and does not\nproperly fall within the scope of the _cosmological_ inquiries, which\nare alone in place in this division of the _Critique_.\n\nThe manner in which Kant, in treating of freedom, makes the\ntransition[1533] from the cosmological (or theological) unconditioned to\nthe psychological is significant. The cosmological unconditioned is\nproved to exist by the argument of the thesis, and its existence is at\nonce interpreted as establishing at least in this one case the actuality\nof free spontaneous causality. Kant remarks that this\n\n     \u201c...transcendental Idea of freedom does not by any means constitute\n     the entire content of the psychological concept of that name, which\n     is mainly empirical, but only that of absolute spontaneity of\n     action.... The necessity of a first beginning, due to freedom, of a\n     series of appearances we have demonstrated only in so far as it is\n     required for the conceivability of an origin of the world.... But\n     as, after all, the power of spontaneously originating a series in\n     time has thus been proved (though not understood), it is now\n     permissible for us to admit within the course of the world\n     different series as capable in their causality of beginning of\n     themselves, and so to attribute to their substances a power of\n     acting from freedom.\u201d\n\nThat each such successive series in the world can only have a relatively\nprimary beginning, and must always be preceded by some other state of\nthings, is no sufficient objection to such causality.\n\n     \u201cFor we are here speaking of an absolutely first beginning not in\n     time, but in causality. If, for instance, I at this moment arise\n     from my chair in complete freedom, without being necessarily\n     determined thereto by the influence of natural causes, a new\n     series, with all its natural consequences _in infinitum_, has its\n     absolute beginning in this event, although the event itself is\n     only, with regard to time, the continuation of a preceding series.\u201d\n\nThus Kant\u2019s proof of freedom in the thesis of the third antinomy is\nmerely a corollary from his proof of the existence of a cosmological or\ntheological unconditioned; and further, this freedom is not, like the\ncosmological unconditioned, proved to exist, but only to be \u201cadmissible\u201d\nas a possibility. Similarly in the antithesis, the only disproof of\nfreedom is the disproof of unconditioned causality _in general_. The\nantinomy deals with the general opposition and relation between the\ncontingent and the unconditioned.\n\nIt is this same opposition exactly which constitutes the subject-matter\nof the fourth antinomy. The terms used are different, but their meanings\nare one and the same. For though Kant substitutes \u2018absolutely necessary\nbeing\u2019 for \u2018unconditioned causality,\u2019 the former is still conceived as\nbelonging to the world of sense, as the unconditioned origin of its\nchanges. And as Kant is careful to add, only the causal, cosmological\nargument can be employed to establish the existence of an absolutely\nnecessary being; nothing can legitimately be inferred from the mere\nIdea. The verbal change is consequently verbal only; the argument of the\nfourth antinomy coincides in result no less than in method of proof with\nthe argument of the third. It is impossible to define the unconditioned\nin any more specific fashion save by an enquiry which entirely\ntranscends the scope of the argument that Kant is here presenting.\nKant\u2019s procedure also lies open to the further objection that the\nconception of an absolutely necessary being, which he here introduces\nwithout preliminary analysis or explanation, is later shown by him[1534]\nto be devoid of significance. He employs it, but precludes himself from\neither investigating it or from drawing any serviceable consequences\nfrom it. The situation is not without the elements of comedy. In order\nto seem to mark a real distinction between the fourth and the third\nantinomies, Kant has perforce to trespass upon the domain of theology;\nbut as he is aware that the trespass is forbidden, he seeks to mitigate\nthe offence by returning from the foray empty-handed. To such unhappy\nstraits is he again reduced by his over-fond devotion to architectonic.\n\n\nSECTION III\n\nTHE INTEREST OF REASON IN THIS SELF-CONFLICT[1535]\n\nThis section, though extremely important, requires no lengthy comment.\nIt is lucid and straightforward. It may be summarised as follows. The\ntheses and the antitheses rest upon diverse and conflicting interests.\nThe theses, though expressed in dry formulas, divested of the empirical\nfeatures through which alone their true grandeur can be displayed,\nrepresent the proud pretensions of dogmatic Reason. The antitheses give\nexpression to principles of pure empiricism. The former are supported by\ninterests of a practical and popular character: upon them morals and\nreligion are based. The latter, while conflicting with our spiritual\ninterests, far exceed the theses in their intellectual advantages. This\nexplains\n\n     \u201c...the zelotic passion of the one party, and the calm assurance of\n     the other, and why the world hails the one with eager approval, and\n     is implacably prejudiced against the other.\u201d\n\nNo legitimate objection could be raised against the principles of the\nempirical philosopher, if he sought only to rebuke the rashness and\npresumption of Reason when it boasts of _knowledge_, and when it\nrepresents as _speculative insight_ that which is grounded only in\n_faith_.\n\n     \u201cBut when empiricism itself, as frequently happens, becomes\n     dogmatic ..., and confidently denies whatever lies beyond the\n     sphere of its intuitive knowledge, it betrays the same lack of\n     modesty; and that is all the more reprehensible owing to the\n     irreparable injury which is thereby caused to the practical\n     interests of Reason.\u201d\n\nEach party asserts more than it knows. The one allows our practical\ninterests to delude Reason as to its inherent powers; the other would so\nextend empirical knowledge as to destroy the validity of our moral\nprinciples. Kant regards the opposition as being historically typified\nby the contrasted systems of Platonism and Epicureanism. It befits us,\nas self-reflecting beings, to free ourselves, at least provisionally,\nfrom the partiality of those divergent interests, and by application of\n\u201cthe sceptical method,\u201d unconcerned about consequences, to penetrate to\nthe primary sources of this perennial conflict. As Kant states in the\nnext section, the conflict is of such a character as to be genuinely\nresolvable.\n\nThis section must have been written, or at least first sketched, at the\ntime when Kant still intended to bring his whole criticism of the\nmetaphysical sciences within the scope of his doctrine of\nantinomy.[1536]\n\n\nSECTION IV\n\nOF THE TRANSCENDENTAL PROBLEMS OF PURE REASON, IN SO FAR AS THEY\nABSOLUTELY MUST BE CAPABLE OF SOLUTION[1537]\n\nThere are sciences the very nature of which requires that every question\nwhich can occur in them must be completely answerable from what can be\npresumed to be known. This is true of the science of ethics. When I ask\nto what course of action I am committed in moral duty, the question must\nbe answerable in terms of the considerations which have led to its being\npropounded. For there can be no moral obligation in regard to that of\nwhich we cannot have knowledge. We must not plead that the problem is\nunanswerable; a solution must be found. Kant proceeds to argue that this\nis no less true of transcendental philosophy.\n\n     \u201c...it is unique among speculative sciences in that no question\n     which concerns an object given to pure Reason is insoluble for this\n     same human Reason, and that no excuse of an unavoidable ignorance,\n     or of the unfathomable depth of the problem, can release us from\n     the obligation to answer it thoroughly and completely. That very\n     concept which enables us to ask the question must also qualify us\n     to answer it, since, as in the case of right and wrong, the object\n     is not to be met with outside the concept.\u201d\n\nThe third and fourth paragraphs would seem to be later interpolations.\nThe section, like Section III., must have been written at the time when\nKant still regarded the doctrine of antinomy as covering the entire\nfield of metaphysics. Transcendental philosophy is identified with\ncosmology, as dealt with in the antinomies. But in the third paragraph\nthe former is taken as a wider term. Also, in the first two paragraphs\nthe problems of pure Reason are regarded as soluble because their\nobjects are not to be met with outside the concepts of them; whereas in\nthe third paragraph they are viewed as soluble because their object is\ngiven empirically. Again, in the second paragraph transcendental\nphilosophy has been taken as unique among speculative [_i.e._\ntheoretical] sciences; in the fourth paragraph mathematics is placed\nalongside it.\n\nExamination of this section as a whole (and the same is true of the\nimmediately following section) justifies the conclusion that at the time\nwhen it was written Kant regarded the Ideas of Reason as having a purely\nand exclusively regulative function, and consequently as exhausting\ntheir inherent meaning in their empirical reference. He regards them as\nentirely lacking in metaphysical significance. They are invented by\nReason for Reason\u2019s own satisfaction, and must therefore yield in their\ninternal content the explanation of their existence, and must also\nsupply a complete and thorough answer to all problems which are\ntraceable to them. A dogmatic (_i.e._ ontological) solution of the\nantinomies is, as we have already found, impossible; the Critical\nsolution considers the question subjectively,\n\n     \u201c...in accordance with the foundation of the knowledge upon which\n     it is based.\u201d[1538] \u201cFor your object is only in your brain, and\n     cannot be given outside it; so that you have only to take care to\n     be at one with yourself, and to avoid the amphiboly which\n     transforms your Idea into a supposed representation of an object\n     which is empirically given and therefore to be known according to\n     the laws of experience.\u201d[1539]\n\nKant\u2019s argument in proof of this purely subjective interpretation of the\nIdeas consists in showing that they are not presented in any given\nappearances, and are not even necessary to explain appearances. The\nunconditioned, whether of quantity, of division, or of origination, has\nnothing to do with any experience, whether actual or possible.\n\n     \u201cYou would not, for instance, in any wise be able to explain the\n     appearances of a body better, or even differently, if you assumed\n     that it consists either of simple or of inexhaustibly composite\n     parts; for neither a simple appearance nor an infinite composition\n     can ever come before you. Appearances demand explanation only in so\n     far as the conditions of their explanation are given in perception,\n     [and the unconditioned can never be so given].\u201d[1540]\n\nThis standpoint, at once sceptical and empirical, is further developed\nin the next section.\n\n\nSECTION V\n\nSCEPTICAL REPRESENTATION OF THE COSMOLOGICAL QUESTIONS[1541]\n\nApplying the \u201csceptical method,\u201d[1542] Kant argues that even supposing\none or other party could conclusively establish itself through final\nrefutation of the other, no advantage of any kind would accrue. The\nvictory would be a fruitless one, and the outcome \u201cmere nonsense.\u201d[1543]\nThe sole validity of the Ideas lies in their empirical reference; and\nyet that reference is one which proves them to be, when objectively\ninterpreted, entirely meaningless. The cosmological Idea is always\neither too large or too small for any concept of the understanding. No\nmatter what view is taken, the only possible object (viz. that yielded\nby experience) will not fit into it. If the world has no beginning, or\nis infinitely divisible, or has no first cause, the regress transcends\nall empirical concepts; while if the world has a beginning, is composed\nof simple parts, and has a first cause, it is too small for the\nconcepts through which alone it can be experienced. In other words, the\ncosmological Ideas are always either too large or too small for the\nempirical regress, and therefore stand condemned by sense-experience,\nwhich can alone impart relation to an object, _i.e._ truth and meaning\nto any concept. For, as Kant explicitly states, _we must not reverse\nthis relation and condemn empirical concepts, as being in the one case\ntoo small, and in the other case too large for the Idea. Experience, not\nIdeas, is the criterion alike of reality and of truth_.\n\n     \u201cThe possible empirical concept is, therefore, the standard by\n     which we must judge whether the Idea is mere Idea and\n     thought-entity (_Gedankending_), or whether it finds its object in\n     the world.\u201d[1544]\n\nWhen two things are compared, that for the sake of which the other\nexists is the sole proper standard. We do not say \u201cthat a man is too\nlong for his coat, but that the coat is too short for the man.\u201d[1545] We\nare thus confirmed in the view that the antinomies rest upon a false\nview of the manner in which the object of the cosmological Ideas can be\ngiven; and are set upon the track, followed out in the next section, of\nthe illusion to which they are due.\n\nThis reduction of the Ideas to mere thought-entities is one of the two\nalternative views which, as we have already stated,[1546] compete with\none another throughout the entire _Dialectic_. We may, for instance,\ncompare the above explanation of the conflict between the Ideas and\nexperience with that given in A 422 = B 450. In the latter passage the\nantinomies are traced to a conflict between Reason and understanding. If\nthe unity is adequate to the demands of Reason, it is too great for the\nunderstanding; if it is adequate to the understanding, it is too small\nfor Reason. Kant does not here allow that the claims of Reason are _ipso\nfacto_ condemned through the incapacity of experience to fulfil them. On\nthe contrary, he implies that it is through the Ideas that we come to\nrealise the merely phenomenal character of everything experienced.\n\nOur task, in this Commentary, is only to distinguish the passages in\nwhich those two conflicting tendencies appear, and to trace the\nconsequences which follow from Kant\u2019s alternation between them.\nDiscussion of their significance had best be deferred to the close of\nthe _Dialectic_, where Kant dwells upon the regulative function of\nReason. At present we need merely note that the main content of the\nabove sections, in which the sceptical view is expounded, is of early\ndate, prior to the working out of the _Paralogisms_ and of the _Ideal_.\n\n\nSECTION VI\n\nTRANSCENDENTAL IDEALISM AS THE KEY TO THE SOLUTION OF THE COSMOLOGICAL\nDIALECTIC[1547]\n\nIn this section subjectivism is dominant. The type of transcendental\nidealism expounded is that earlier and less developed form which\nconnects with the doctrine of the transcendental object.[1548] It shows\nno trace of Kant\u2019s maturer teaching. No distinction is drawn between\nrepresentation and the objects represented. To the transcendental\nobject, the \u201cpurely intelligible cause\u201d of appearances in general, and\nto it alone, Kant ascribes \u201cthe whole extent and connection of our\npossible perceptions.\u201d[1549] Appearances exist only in the degree to\nwhich they are constructed in experience. As they are mere\nrepresentations, they cannot exist outside the mind. Independently of\nsuch construction, they may indeed be said to be given in the\ntranscendental object, but they only become objects to us on the\nsupposition that they can be reached through extension of the series of\nour actual perceptions. It is in this form alone, as conceived in a\nregressive series of possible perceptions, and not as having existed in\nitself, that even the immemorial past course of the world can be\nrepresented as real;\n\n     \u201c...so that all events which have taken place in the immense\n     periods that have preceded my own existence mean really nothing but\n     the possibility of extending the chain of experience from the\n     present perception back to the conditions which determine it in\n     time.\u201d[1550]\n\nA similar interpretation has to be given to all propositions which\nassert the present reality of that which has never been actually\nexperienced.\n\n     \u201cIn outcome it is a matter of indifference whether I say that in\n     the empirical progress in space I can meet with stars a hundred\n     times farther removed than the outermost now perceptible to me, or\n     whether I say that they are perhaps to be met with in cosmical\n     space even though no human being has ever perceived or ever will\n     perceive them. For even if they were given as things in themselves,\n     without relation to possible experience,[1551] they are still\n     nothing for me, and therefore are not objects, save in so far as\n     they are contained in the series of the empirical regress.\u201d[1552]\n\nThe distinction between appearances and things in themselves must\nalways, Kant observes, be borne in mind when we are interpreting the\nmeaning of our empirical concepts; and this is especially necessary when\nthose concepts are brought into connection with the cosmological Idea of\nan unconditioned. The antinomies are due to a failure to appreciate this\nfundamental distinction, and the key to their solution lies in its\nrecognition.\n\n     \u201cIt would be an injustice to ascribe to us that long-decried\n     empirical idealism which, while it admits the genuine actuality of\n     space, denies the existence of the extended beings in it....\u201d[1553]\n\nThis is in line with the passages from the _Prolegomena_ commented upon\nabove.[1554]\n\n\nSECTION VII\n\nCRITICAL DECISION OF THE COSMOLOGICAL CONFLICT OF REASON WITH\nITSELF[1555]\n\nKant\u2019s argument is as follows. The antinomies rest upon the principle\nthat if the conditioned be given, the entire series of all its\nconditions is likewise given. If the objects of the senses were\nindependently real, there would be no escape from this assumption, and\nthe dialectical conflict would consequently be irresolvable.\nTranscendental idealism, as above stated, reveals a way out of the\ndilemma. As appearances are merely representations, their antecedent\nconditions do not exist _as appearances_, save in the degree in which\nthey are mentally constructed. Though the appearances are given, their\n_empirical_ conditions are not thereby given. The most that we can say\nis that a regress to the conditions, _i.e._ a continued empirical\nsynthesis in that direction, is _commanded_ or _required_. The\ncosmological argument can thus be shown to be logically invalid. The\nsyllogism, which it involves, is as follows:\n\nIf the conditioned be given, the entire series of all its conditions is\nlikewise given.\n\nThe objects of the senses are given.\n\nTherefore the entire series of all their conditions is likewise given.\n\nIn the major premiss the concept of the conditioned is employed\ntranscendently (Kant says transcendentally), in the minor empirically.\nBut though the inference thus commits the logical fallacy of _sophisma\nfigurae dictionis_, the ground of its occurrence, and the reason why it\nis not at once detected, lie in a natural and inevitable illusion which\nleads us to accept the sensible world in space as being independently\nreal. Only through Critical investigation can the deceptive power of\nthis illusion be overcome. Owing to its influence, the above fallacy has\nbeen committed by dogmatists and empiricists alike. It can be shown that\nin refuting each other\n\n... \u201cthey are really quarrelling about nothing, and that a certain\n     transcendental illusion has caused them to see a reality where none\n     is to be found.\u201d[1556]\n\nThe existence of antinomy, Kant further argues, presupposes that theses\nand antitheses are contradictory opposites, _i.e._ that no third\nalternative is possible. When opposed assertions are not contradictories\nbut contraries, the opposition, to use Kant\u2019s terms, is not analytical\nbut dialectical. Both may be false; for the one does not merely\ncontradict the other, but makes, in addition, a further statement on its\nown account. Now examination of the illusion above described enables us\nto perceive that the opposites, in reference to which antinomy occurs,\nare of this dialectical character. _Theses and antitheses are alike\nfalse._ Since the world does not exist as a thing in itself, it exists\nneither as an infinite whole nor as a finite whole, but only in the\ndegree in which it is constructed in an empirical regress. We must not\napply \u201c_the Idea of absolute totality, which is valid only as a\ncondition of things in themselves_,\u201d[1557] to appearances. (The words\nwhich I have italicised mark the emergence of Kant\u2019s non-sceptical,\nnon-empirical view of the nature and function of the Ideas of Reason.)\nThus antinomy, rightly understood, does not favour scepticism, but only\nthe \u201csceptical method,\u201d and indeed yields an indirect proof of the\ncorrectness of Critical teaching. This proof may be presented in the\nform of a dilemma. If the world is a whole existing in itself, it is\neither finite or infinite. But the former alternative is refuted by the\nproofs given of the antitheses, and the latter alternative by the proofs\nof the theses. Therefore the world cannot be a whole existing in itself.\nFrom this it follows that appearances are nothing outside our\nrepresentations; and that is what is asserted in the doctrine of\ntranscendental idealism.\n\nIn A 499 = B 527 Kant uses ambiguous language,[1558] which can be\ninterpreted as asserting that in the regress there can be no lack of\ngiven conditions. Such a statement would presuppose positive knowledge\nregarding the unknown transcendental object.[1559] The opposite, more\ncorrect, view is given in A 514-15 = B 542-3 and A 517 ff. = B 545 ff.,\nthough in the latter passage with a reversion to the above\nposition.[1560]\n\nThe earlier manuscripts, which Kant has so far been employing, probably\nterminate either, as Adickes suggests,[1561] at the end of this section,\nor at the close of Section VIII., which is of doubtful date. Section IX.\nis certainly from a later period; it represents a more complex\nstandpoint, in which Reason is no longer viewed as possessing a merely\nempirical function, and in which consequently the theses and antitheses\nare no longer indiscriminately denounced as being alike false. Under the\ninfluence of his later, more Idealistic preoccupations, Kant so far\nmodifies the above solution as to assert _that in the ease of the last\ntwo antinomies both theses and antitheses are true, when properly\ninterpreted_.\n\n\nSECTION VIII\n\nTHE REGULATIVE PRINCIPLE OF PURE REASON IN REGARD TO THE COSMOLOGICAL\nIDEAS[1562]\n\nThe principle of pure Reason, correctly formulated, is that when the\nconditioned is given a regress upon the totality of its conditions is\n_set as a problem_. As such it is valid,\n\n     \u201c...not indeed as an _axiom_ ... but as a _problem_ for the\n     understanding ..., leading it to undertake and to continue,\n     according to the completeness in the Idea, the regress in the\n     series of conditions of any given conditioned.\u201d[1563]\n\nIt does not anticipate, prior to the regress, what actually exists as\nobject, but only postulates, in the form of a rule, how the\nunderstanding ought to proceed. It does not tell us whether or how the\nunconditioned exists, but how the empirical regress is to be carried out\nunder the guidance of a mere Idea. Such a rule can be regulative only,\nand the Idea of totality which it contains must never be invested with\nobjective reality. As the absolutely unconditioned can never be met with\nin experience, we know, indeed, beforehand that in the process of the\nregress the unconditioned will never be reached. But the duty of seeking\nit by way of such regress is none the less prescribed.\n\nKant proceeds to give a somewhat bewildering account of the familiar\ndistinction between _progressus in infinitum_ and _progressus in\nindefinitum_, and to draw a very doubtful distinction between the series\nin division of a given whole and the series in extension of it.[1564]\nThe illustration from the series of human generations is an unfortunate\none; the discovery that it began at some one point in the past would not\nnecessarily violate any demand of Reason. Such a series is not\ncomparable with those of space, time, and causality.[1565] The only\nimportant result of this digression is the conclusion that whatever\ndemand be made, whether of regress _in infinitum_ or of regress _in\nindefinitum_, in neither case can the series of conditions be regarded\nas being given as infinite in the object.\n\n     \u201cThe question, therefore, is no longer how great this series of\n     conditions may be in itself, whether finite or infinite, for it is\n     nothing in itself; but how we are to carry out the empirical\n     regress, and how far we should continue it.\u201d[1566]\n\nWe have already noted[1567] Kant\u2019s ambiguous suggestion in A 499 = B\n527, that in the empirical regress there can be no lack of given\nconditions. The statement, thus interpreted, is illegitimate. The most\nthat he can claim is that, were further sensations not forthcoming, we\nshould still have to conceive those last obtained as being preceded by\nempty space and time, and as lacking in any experienced cause. Under\nsuch circumstances we should experience neither finitude nor\nunconditionedness, but only incapacity to find a content suitable to the\ninexhaustible character of the spatial and temporal conditions of\nexperience, or in satisfaction of our demand for causal antecedents. In\nA 514-15 = B 542-3 Kant shows consciousness of this difficulty, but in\ndealing with it adopts a half-way position which still lies open to\nobjection. He recognises that, since no member of a series can be\nempirically given as absolutely unconditioned, a higher member is always\n_possible_, and that the search for it is therefore prescribed; none the\nless he asserts that in regard to _given wholes_ we are justified in\ntaking up a very different position, namely, that the regress in the\nseries of their internal conditions does not proceed, as in the above\ncase, _in indefinitum_, but _in infinitum_, _i.e._ that in this case\nmore members _exist and are empirically given_ than we can reach through\nthe regress. In given wholes we are commanded to _find_ more members; in\nserial extension we are justified only in _inquiring_ for more. This\nhalf-way position is a makeshift, and is in no respect tenable. The\nevidence for the infinite extensibility of space and time is as\nconclusive as for their infinite divisibility. And when we consider\nsensuous existence under these forms, it is just as possible that the\ntranscendental object may, beyond a certain point, fail to supply\nmaterial for further division, as that it may fail to yield data for\nfurther expansion. What Kant asserts of the latter, that further advance\nmust always remain as a possibility, and for that Reason must always\ncall for the open mind of further inquiry, without any attempted\nanticipatory assertion either _pro_ or _contra_, alone represents the\ntrue Critical standpoint. The cessation of data may really, however, be\ndue to an increase in the subtlety of the conditioning processes that\nincapacitates them from acting upon our senses;[1568] by indirect means\nthis disability may be overcome. Reason, in its conception of an\nunconditioned, prescribes to us a task that is inexhaustible in its\ndemands. We have no right to lay down our intellectual arms before any\nbarrier however baffling, or to despair before any chasm however empty\nand abrupt.\n\n\nSECTION IX\n\nTHE EMPIRICAL EMPLOYMENT OF THE REGULATIVE PRINCIPLE OF REASON IN REGARD\nTO ALL COSMOLOGICAL IDEAS[1569]\n\nSOLUTION OF THE FIRST AND SECOND ANTINOMIES\n\n=Statement.=--The fundamental fact upon which, as Kant has already stated,\nthe regulative principle of Reason is based, is that it is impossible to\nexperience an absolute limit. It is always _possible_ that a still\nhigher member of the series may be found; and that being so, it is our\nduty to search for it. But as we are here dealing with possibilities\nonly, the regress is _in indefinitum_, not _in infinitum_.\n\n     \u201c...we must seek the concept of the quantity of the world only\n     according to the rule which determines the empirical regress in it.\n     This rule says no more than that however far we may have attained\n     in the series of empirical conditions, we should never assume an\n     absolute limit, but should subordinate every appearance, as\n     conditioned, to another as its condition, and that we must then\n     advance to this condition. This is the _regressus in indefinitum_,\n     which, as it determines no quantity in the object, is clearly\n     enough distinguishable from the _regressus in infinitum_.\u201d[1570]\n\nWe are acquainted only with the rule, and not with the whole object. Any\nassertion, therefore, which we can make, must be dictated solely by the\nrule, and be an expression of it. Neither the thesis nor the antithesis\nof the first antinomy is valid; there is a third alternative. The\nsensible world is neither finite nor infinite in extent; it is\ninfinitely extensible, in terms of the rule.\n\nUnfortunately Kant is not content to leave his conclusion in this form.\nHe complicates his argument, and bewilders the reader, by maintaining\nthat this is a virtual acceptance of the antithesis, in that we assert\nnegatively, that an absolute limit in either time or space is\nempirically impossible;[1571] and affirmatively, that the regress goes\non _in indefinitum_, and consequently has no absolute quantity.\n\nKant also repeats the argument of the preceding section in regard to\ngiven wholes.[1572] When the problem is that of subdivision, the regress\nstarts from a given whole, and therefore from a whole whose conditions\n(the parts) are given with it. The division is, therefore, _in\ninfinitum_, and not merely _in indefinitum_. This does not, however, he\nargues, mean that the given whole consists of infinitely many parts. For\nthough the parts are contained in the intuition of the whole, yet the\nwhole division arises only through the regress that generates it. It is\na _quantum continuum_, not a _quantum discretum_.[1573] This argument\nhas been criticised above.[1574] Kant here ignores the possibility that\nthe parts of matter, though extended, may be physically indivisible, or\nthat they may be centres of force which control, but do not occupy, a\ndeterminate space.\n\n\nREMARKS ON THE DISTINCTION BETWEEN THE MATHEMATICAL-TRANSCENDENTAL AND\nTHE DYNAMICAL-TRANSCENDENTAL IDEAS[1575]\n\n=Statement.=--Kant again[1576] introduces the distinction between the\nmathematical and the dynamical. The mathematical Ideas synthesise the\nhomogeneous, the dynamical may connect the heterogeneous. In employing\nthe former we must therefore remain within the phenomenal; through the\nlatter we may be able to transcend it. The way is thus opened for\npropounding, in regard to the third and fourth antinomies, a solution in\nwhich the pretensions of Reason no less than those of understanding may\nfind satisfaction. Whereas both the theses and the antitheses of the\nfirst and second antinomies have to be declared false, those of the\nthird and fourth antinomies _may both be true_--the theses applying to\nthe intelligible realm, and the antitheses to the world of sense.\n\n=Comment.=--When the distinction between the mathematical and the\ndynamical is thus extended from the categories to the Ideas, its\nvalidity becomes highly doubtful. Space and time are certainly\nthemselves homogeneous, and the categories of quality and quantity, in\nso far as they are mathematically employed, may perhaps be similarly\ndescribed. But when the term is still further extended, to cover the\npairs of correlative opposites with which the first two antinomies deal,\nthose, namely, between the limited and the unlimited, the simple and the\ninfinitely divisible, Kant would seem to be making a highly artificial\ndistinction. The first two antinomies deal not with space and time as\nsuch, but with the sensible world in space and time; and within this\nsensible world, even in its quantitative aspects, qualitative\ndifferences have to be reckoned with. Common sense does, indeed, tend to\nassume that the unlimited and the simple must, like that which they\ncondition, be in space and time, and so form with the conditioned a\nhomogeneous series. But this assumption ordinary consciousness is\nequally disposed to make in regard to a first cause and to the\nunconditionally necessary.\n\nKant further attempts[1577] to distinguish between the mathematical and\nthe dynamical by asserting that the dynamical antinomies are not\nconcerned with the quantity of their object, but only with its\n_existence_. He admits, however, that in all four cases a series arises\nwhich is either too large or too small for the understanding; and that\nbeing so, in each case the problem arises as to the _existence_ of an\nunconditioned.\n\nThe artificiality of Kant\u2019s distinction becomes clear when we recognise\nthat the opposed solutions, which he gives of the two sets of\nantinomies, can be mutually interchanged. As the sensible world rests\nupon intelligible grounds, both the theses and the antitheses of the\nfirst two antinomies may be true, the former in the intelligible realm\nand the latter in the sensuous. Similarly, both the theses and\nantitheses of the third and fourth antinomies may be false. In the\nsensible world, about which alone anything can be determined, the series\nof dynamical conditions forms neither a finite nor an infinite series.\nThere is a third alternative, akin to that of the antitheses, but\ndistinct in character from it, namely, that the series is infinitely\nextensible. Kant\u2019s differential treatment of the two sets of antinomies\nis arbitrary, and would seem to be due to his having attempted to\nsuperimpose, with the least possible modification, a later solution of\nthe antinomies upon one previously developed. In the earlier view, as we\nhave already had occasion to observe, Reason has a merely empirical\napplication. Its Ideas are taken as existing \u201conly in the brain.\u201d Only\ntheir empirical reference can substantiate them, or indeed give them the\nleast significance. And as they are by their very nature incapable of\nempirical embodiment, all assertions which involve them must necessarily\nbe false. Later, Kant came to regard Reason as having its own\nindependent rights. Encouraged by his successful establishment of the\nobjective validity of the categories, progressively more and more\nconvinced of the importance of the distinction, which that proof\nreinforced, between appearances and things in themselves, and\npreoccupied with the problems of the spiritual life, his old-time faith\nin the absolute claims of pure thought reasserted itself. Through Reason\nwe realise our kinship with noumenal realities, and through its demands\nthe nature of the unconditioned is foreshadowed to the mind. The theses\nand antitheses, which throughout the entire history of philosophy have\ncompeted with one another, may both be true. Their perennial conflict\ndemonstrates the need for some more catholic standpoint from which the\ntwo great authorities by which human life is controlled and directed,\nthe intellectual and the moral, may be reconciled. Neither can be made\nto yield to the other; each is supreme in its own field. The distinction\nbetween appearances and things in themselves, recognition of which is\nthe first step towards an adequate theory of knowledge, and without\nwhich the nature of the intellectual life remains self-contradictory and\nincomprehensible, itself affords the means of such a reconciliation. The\nunderstanding is the sole key to the world of appearance, the moral\nimperative to the realm of things in themselves. Reason with its demand\nfor the unconditioned mediates between them, and enables us to realise\nour dual vocation.\n\nThis radical alteration of standpoint was bound to make the employment\nof manuscript representing the earlier and more sceptical attitude\naltogether unsatisfactory; and only Kant\u2019s constitutional unwillingness\nto sacrifice what he had once committed to paper can account for his\nretention of the older expositions. He allows his previous treatment of\nthe first two antinomies to remain in its sceptical form, and, by means\nof the distinction between the mathematical and the dynamical, develops\nhis newer, more Idealist view exclusively in reference to the third and\nfourth antinomies. That it is no less applicable to the others, we have\nalready seen.\n\nThough the Idealist view, _as here expounded_, may be thus described,\nrelatively to the sceptical view of Reason, as later, that is not to be\ntaken as meaning that it represents the latest stage in the development\nof Kant\u2019s Critical teaching. It seems to belong to the period prior to\nthat in which the central sections of the _Analytic_ were composed. The\nevidence[1578] for this consists chiefly in its subjectivist references\nto the nature of appearances. It would seem to be contemporary with\nKant\u2019s doctrine of the transcendental object.\n\n\nSOLUTION OF THE THIRD ANTINOMY[1579]\n\n=Statement.=--As appearances are representations only, they must have a\nground which is not itself an appearance;[1580] and though the effects\nof such an intelligible cause appear, and accordingly are determined\nthrough other appearances, its causality is not itself similarly\nconditioned. Both it and its causality lie outside the empirical series;\nonly the effects fall within the realm of experience. And that\ncausality, not being subject to time, does not require to stand under\nanother cause as its effect. In this way Kant derives from his\ntranscendental idealism an explanation of the possibility of an action\nbeing at once free and causally determined. This explanation he takes as\napplying either to a first cause of the whole realm of natural phenomena\nor to a finite being regarded as a free agent. The proof of the\npossibility of this metaphysical, or, as Kant entitles it,\n\u201ctranscendental freedom,\u201d removes what has always been the real\ndifficulty that lay in the way of \u201cpractical freedom.\u201d The conception of\nfreedom is a transcendental Idea which can neither be derived from\nexperience nor verified by it. It is created by Reason for itself;[1581]\nand reveals the _possibility_ that in this third antinomy both thesis\nand antithesis may be true. The alternatives--\u201cevery effect must arise\nfrom nature,\u201d and \u201cevery effect must arise from freedom\u201d--are not\nexclusive of one another. They may be true of one and the same event in\ndifferent relations.[1582] The event may be free in reference to its\nintelligible cause, determined as an existence in space and time. Were\nappearances things in themselves, freedom and causality would\nnecessarily conflict: by means of the above ontological distinction\nfreedom can be asserted without any diminution in the scope allowed to\nthe causal principle. All events, without a single possible exception,\nare subject to the law of natural determination; and yet every event may\n_at the same time_ proceed from a free cause.\n\n\nPOSSIBILITY OF HARMONISING CAUSALITY THROUGH FREEDOM WITH THE UNIVERSAL\nLAW OF NATURAL NECESSITY[1583]\n\n=Statement.=--The above conclusion is so seemingly paradoxical that Kant\ndevotes this and the following section to its further elucidation. How\ncan events be both free and determined? The answer lies in recognition\nof the two-sided character of every natural existence. It is, in one\naspect, mere appearance; in another, it has at its foundation a\ntranscendental object. It is an appearance of the latter, and for its\ncomplete comprehension this latter must be taken into account. Now there\nis nothing to _prevent_ us from attributing to the transcendental object\na causality which is not phenomenal. Such causality may make the\nappearance just that appearance which it is. In the world of sense every\nefficient cause must have a specific empirical character, since only so\ncan it determine one effect rather than another according to the\nuniversal and invariable law expressive of its nature. We must similarly\nallow to the transcendental object an intelligible character, and trace\nto it all those appearances which as members of the empirical series\nstand to one another in unbroken causal connection. This transcendental\nobject, owing to its intelligible character, is not in time. Its act\ndoes not either arise or perish, and is not, therefore, subject to the\nlaw of empirical determination which applies only to the changeable,\n_i.e._ to events subsequent upon previous states. Such supersensuous\ncausality can find no place in the series of empirical conditions, and\nthough it can be conceived only in terms of the empirical character\nwhich is its outcome, the difference between it and natural causality\nmay be as complete as that which subsists between the transcendental and\nthe empirical objects of knowledge. In its empirical character the\naction is a part of nature, and enters into a causal nexus which\nconforms to universal laws.[1584] All its effects are inevitably\ndetermined by antecedent natural conditions. In its intelligible\ncharacter, however, this same active subject must be considered free\nfrom all influence of sensibility and from all determination through\nnatural events. In so far as it is a noumenon, there can be no change in\nit, and therefore nothing which is capable of explanation in terms of\nnatural causes. Even its empirical effects are not traceable to it as\nevents in time. For as events these effects are always the results of\nantecedent empirical causes. What is alone due to noumenal causality is\nthat empirical character in virtue of which appearances are what they\nare, and owing to which they stand in specific and necessary causal\nrelations to one another.\n\n     \u201c...the empirical character is _permanent_, while its effects,\n     according to variation in the concomitant, and in part limiting\n     conditions, appear in changeable forms.\u201d[1585]\n\nEmpirical causality is itself in its specific nature conditioned by an\nintelligible cause.[1586]\n\n\nEXPLANATION OF THE RELATION OF FREEDOM TO NECESSITY OF NATURE[1587]\n\n=Statement.=--No single appearance can be exempted from the law of natural\ncausality. For it would then be placed outside all possible experience,\nand would be for us a fiction of the brain, or rather could not be\nconceived at all. Nothing, therefore, _in nature_ can act freely or\nspontaneously. But while thus recognising that all _events_ without\nexception are empirically conditioned, we may, as already pointed out,\nregard empirical causality as itself an effect of a non-empirical and\nintelligible power.[1588] In events there may be nothing but nature, and\nyet nature itself, or perhaps even some of the existences composing it,\nmay rest upon powers of a noumenal order. Kant proceeds to show that\nsuch an hypothesis is not only allowable, but is indispensable for\nunderstanding the distinguishing features of human life in its practical\naspect.\n\nMan is a natural existence, and his activities are subject to empirical\nlaws. Like all other objects of nature, he has an empirical character,\nand in virtue of it takes his place as an integral part of the system of\nnature. But man is unique among all natural existences in that he not\nonly knows himself as a sensible existence, but also, through pure\napperception, becomes aware of himself as possessing faculties of a\nstrictly intelligible character.[1589] Such are the faculties of\nunderstanding and Reason, especially the latter in its practical\nemployment. The \u201cought\u201d of the moral imperative expresses a kind of\nnecessity and a form of causation which we nowhere find in the world of\nnature. The understanding can know in nature only what actually is, has\nbeen, or will be. Nothing natural can be other than it is in the\nparticular relations in which it is found. Moral action transcends the\nnatural in that it finds its cause, not in an appearance or set of\nappearances, but in an Ideal of pure Reason. Such action must indeed be\npossible under natural conditions, but such conditions do not determine\nits rightness, and consequently cannot determine its causality.\n\n     \u201cReason ... does not here follow the order of things as they\n     present themselves in appearance, but frames to itself with perfect\n     spontaneity an order of its own according to Ideas, to which it\n     adapts the empirical conditions, and according to which it declares\n     actions to be necessary even although they have never yet taken\n     place, and perhaps never will take place. And at the same time it\n     also presupposes that Reason can have causality in regard to all\n     these actions, since otherwise no empirical effects could be\n     expected from its Ideas.\u201d[1590]\n\nIf such action of pure Reason be admitted to be possible, it will have\nto be viewed, purely intelligible though it be, as also possessing an\nempirical character, i.e. as conforming to the system of nature. Its\nempirical consequences will be the effects of antecedent appearances,\nand will empirically determine by natural necessity all subsequent acts.\nIn this empirical character, therefore, there can be no freedom. Were\nour knowledge of the circumstances sufficiently extensive, every human\naction, so far as it is appearance, could be predicted and shown to be\nnecessary. How, then, can we talk of actions as free, when from the\npoint of view of appearances they must in all cases be regarded as\ninevitable? The solution is that which has already been given of the\nbroader issue. The _entire_ empirical character, the whole system of\nnature, is determined by the intelligible character. And the former\nresults from the latter, not empirically, and therefore not according to\nany temporal, causal law. It does not arise or begin at a certain time.\nThe intelligible character conditions the empirical series _as a\nseries_, and not as if it were a first member of it.\n\n     \u201cThus what we have missed in all empirical series is disclosed as\n     possible, namely, that the condition of a successive series of\n     events may itself be empirically unconditioned.\u201d[1591]\n\nThe intelligible character lies _outside_ the series of appearances.\n\u201cReason is the _abiding_ (_beharrliche_) condition of all free\nactions....\u201d[1592] Freedom ought not, therefore, to be conceived only\nnegatively as independence of empirical conditions, but also positively\nas the power of originating a series of events. The empirical series is\nin time. Reason, which is its unconditioned condition, admits of nothing\nantecedent to itself; it knows neither before nor after. The series is\nthe immediate effect of a non-temporal reality.\n\nIn illustration of his meaning, not, as he is careful to add, with the\nprofession of thereby confirming its truth, Kant points out that moral\njudgment upon a vicious action is not determined in view of the\ninheritance, circumstances and past life of the offender, but is passed\njust as if he might in each action be supposed to begin, quite by\nhimself, a new series of effects. This, in Kant\u2019s view, shows that\npractical Reason is regarded as a cause completely capable,\nindependently of all empirical conditions, of determining the act, and\nthat it is present in all the actions of men under all conditions, and\nis always the same. To explain why the intelligible character should in\nany specific case produce just this particular empirical character, good\nor bad,\n\n     \u201c...transcends all the powers of our Reason, indeed all its rights\n     of questioning, just as if we were to ask why the transcendental\n     object of our outer sense-intuition yields intuition in _space_\n     only and no other.\u201d[1593]\n\nIn conclusion Kant states that his intention has not been to establish\nthe reality of freedom, not even to prove its possibility. Freedom has\nbeen dealt with only as a transcendental Idea; and the only point\nestablished is that freedom is, so to speak, a _possible possibility_,\nin that it is _not contradicted_ either by experience or by anything\nthat can be proved to be a presupposition of experience.\n\n=Comment.=--Adequate comment upon this section is difficult for many\nreasons. The section is full of archaic expressions from the earlier\nstages of Kant\u2019s Critical teaching. Secondly, the section anticipates a\nproblem which is first adequately dealt with in the second _Critique_.\nAnd lastly, but not least, the discussion of freedom in connection with\na cosmological antinomy leads Kant to treat it in the same manner as the\ngeneral antinomy, and in so doing to ignore the chief difficulty to\nwhich human freedom, as an independent problem with its own peculiar\ndifficulties, lies open. For it is comparatively easy to reconcile the\nuniversality of the causal principle with the unconditionedness of the\ntranscendental ground upon which nature _as a whole_ is made to rest. It\nis a very different matter to reconcile the spontaneous origination of\n_particular_ causal series, or the freedom of _particular_ existences,\nsuch as human beings, with the singleness and uniformity of a natural\nsystem in which every part is determined by every other.\nSelf-consciousness, with the capacity which it confers of constructing\nrational ideals, certainly, as Kant rightly contends, creates a\nsituation to which mechanical categories are by no means adequate. But\nthe mere reference to the conceivability of distinct causal series,\nhaving each a pure conception as their intelligible ground, does not\nsuffice to meet the fundamental difficulty that, on Kant\u2019s own\nadmission, each such separate series must form an integral part of the\nunitary system of natural law. In only one passage does Kant even touch\nupon this difficulty. Speaking[1594] of Reason\u2019s power of originating a\nseries of events, he adds that while nothing _begins_ in Reason itself\n(as it admits of no conditions antecedent to itself in time), the new\nseries must none the less have a beginning in the natural world. But the\nproviso, which he at once makes, indicates that he is aware that this\nstatement is untenable. For he adds the qualification that though a\nbeginning of the series, it is never an absolutely first beginning. In\nother words, it is not a beginning in any real sense of the term. As the\nargument of his next paragraph shows, it is the entire system of\nnature, and not any one series within it, which can alone account, in\nempirical terms, for any one action.\n\nIt is open to Kant to argue, as he has already done,[1595] that the\ntranscendental object conditions each separate appearance as well as all\nappearances in their totality, and that the specific empirical character\nof each causal series is therefore no less noumenally conditioned than\nis nature as a whole. But this does not suffice to meet the\ndifficulty--how, if all natural phenomena constitute a single closed\nsystem in which everything is determined by everything else, a moral\nagent, acting spontaneously, can be free to originate a genuinely new\nseries of natural events. We seem constrained to conclude that Kant has\nfailed to sustain his position. A solution is rendered impossible by the\nvery terms in which he formulates the problem. If the spiritual and the\nnatural be opposed to one another as the timeless and the temporal, and\nif the natural be further viewed as a unitary system, individual moral\nfreedom is no longer defensible. Only the \u201ctranscendental freedom\u201d of\nthe _cosmological_ argument can be reckoned as among the open\npossibilities.\n\nAs regards the character of the Critical doctrine which underlies this\nsection, we need only note that the statement in A 546-7 = B 574-5, that\nman knows himself through pure apperception as \u201ca purely intelligible\nobject,\u201d[1596] does not conform to Kant\u2019s final teaching. The section\ncan be dated through its unwavering adherence to the subjectivist\ndoctrine of the transcendental object.[1597]\n\n\nSOLUTION OF THE FOURTH ANTINOMY[1598]\n\n=Statement.=--The above solution is adopted. Both thesis and antithesis\nmay be true, the latter of the world of sense and the former of its\nnon-empirical ground. All things sensible are contingent, but the\ncontingent series in its entirety may nevertheless rest upon an\nunconditionally necessary being. The unconditioned, since it is outside\nthe series, does not require that any one link in the series should be\nitself unconditioned. \u201cReason follows its own course in the empirical,\nand again a peculiar course in its transcendental use,\u201d _i.e._ it limits\nitself by the law of causality in dealing with appearances, lest in\nlosing the thread of the empirical conditions it should fall into idle\nand empty speculations; while, on the other hand, it limits that law to\nappearances, lest it should wrongly declare that what is useless for the\nexplanation of appearances is therefore impossible in itself. This does\nnot prove that an absolutely necessary being is really possible, but\nonly that its impossibility must not be concluded from the necessary\ncontingency of all things sensuous.\n\n=Comment.=--Kant\u2019s method of distinguishing[1599] this conclusion from\nthat of the preceding antinomy is again artificial. \u201cNecessary being\u201d is\nnot in conception more _extramundanum_ than \u201cunconditioned cause.\u201d If\nKant\u2019s distinction were valid, the argument of the fourth antinomy would\nno longer be cosmological; it would coincide with the problem of the\n_Ideal of Pure Reason_.\n\n\nCONCLUDING NOTE ON THE WHOLE ANTINOMY OF PURE REASON[1600]\n\n=Statement.=--When we seek the unconditioned entirely beyond experience,\nour Ideas cease to be cosmological; they become transcendent. They\nseparate themselves off from all empirical use of the understanding, and\ncreate to themselves an object, the material of which is not taken from\nexperience, and which is therefore a mere thing of the mind (_blosses\nGedankending_). None the less the cosmological Idea of the fourth\nantinomy impels us to take this step. When sensuous appearances, as\nmerely contingent, require us to look for something altogether distinct\nin nature from them, our only available instruments, in so doing, are\nthose pure concepts of things in general which contingent experience\ninvolves. We use them as instruments in such manner as may enable us to\nform, through analogy, some kind of notion of intelligible things. Taken\nin abstraction from the forms of sense, they yield that notion of an\nabsolutely necessary Being which is equivalent to the concept of the\ntheological Ideal.\n\n\nCONCLUDING COMMENT ON KANT\u2019S DOCTRINE OF THE ANTINOMIES\n\nWe may now, in conclusion, briefly summarise the results obtained in\nthis chapter. Kant fails to justify the assertion that on the dogmatic\nlevel there exist antinomies in which both the contradictory\nalternatives allow of cogent demonstration. His proofs are in every\ninstance invalid. The real nature of antinomy must, as he himself\noccasionally intimates, be defined in a very different manner, namely,\nas a conflict between the demand of Reason for unity and system, and\nthe specific nature of the conditions, especially of the spatial and\ntemporal conditions, under which the sensuous exists. In this wider form\nit constitutes a genuine problem, which demands for its solution the\nfundamental Critical distinction between appearances and\nthings-in-themselves, and also a more thoroughgoing discussion than has\nyet been attempted of the nature of Reason and of the function of its\nIdeas. It is to these connected questions that Kant devotes his main\nattention in the remaining portions of the _Dialectic_, so that in\npassing to the _Ideal of Pure Reason_ he is not proceeding to the\ntreatment of a new set of problems, but to the restatement and to the\nmore adequate solution of the fundamental conflict between understanding\nand Reason.\n\nThe observations which closed our comment upon the _Paralogisms_ are\nthus again in order. The teaching of the sections on the _Antinomies_,\nno less than that of those on the _Paralogisms_, is incomplete, and if\ntaken by itself is bound to mislead. The Ideas of an unconditioned self\nand of an unconditioned ground of nature have thus far been taken as at\nleast conceptually possible, and as signifying what may perhaps be real\nexistences. These Ideas are in certain of the remaining sections of the\n_Dialectic_ called in question. They are there declared to be without\ninherent meaning. They are useful fictions--_heuristische\nFiktionen_--and in their psychological nature are simply _schemata_ of\nregulative principles. Their theoretical significance consists merely in\ntheir regulative and limitative functions. They must not be regarded,\neven hypothetically, as representing real existences. In the practical\n(_i.e._ ethical) sphere they do indeed acquire a very different\nstanding. But with that the _Critique of Pure Reason_ is not directly\nconcerned. The reader may therefore be warned not to omit the chapter on\nthe _Ideal of Pure Reason_, on the supposition that it embodies only a\ncriticism of the Cartesian and teleological proofs of God\u2019s existence.\nIt is an integral part of Critical teaching, and carries Kant\u2019s entire\nargument forward to its final conclusions. Only in view of the new and\ndeeper considerations, which it brings to light, can his treatment even\nof the _Antinomies_ be properly understood. Its main opening section\n(Section II.) is, indeed, among the most scholastically rationalistic in\nthe entire _Critique_; but in the later sections it unfolds, with a\nboldness and consistency to which we find no parallel in the treatment\nof the _Paralogisms_ and of the _Antinomies_, the full consequences of\nthe more sceptical of Kant\u2019s alternating standpoints. It disintegrates\nthe concepts of the unconditioned, which have hitherto been employed\nwithout analysis and without question; and upon their elimination from\namong the legitimate instruments of Reason, the situation undergoes\nentire transformation, the two points of view appearing for the first\ntime in the full extent of their divergence and conflict. For Kant\u2019s\nIdealist view of Reason and of its Ideas still continues to find\noccasional statement, showing that he has not been able decisively to\ncommit himself to this more sceptical interpretation of the function of\nReason; that he is conscious that the Idealist view alone gives adequate\nexpression to certain fundamental considerations which have to be\nreckoned with; and that unless the two views can in some manner be\nreconciled with one another, a really definitive and satisfactory\nsolution of the problem has not been reached. When, therefore, we speak\nof Kant\u2019s final conclusions, we must be taken as referring to the\ntwofold tendencies, sceptical and Idealist, which to the very last\npersist in competition with one another. The greater adequacy of Kant\u2019s\nargument in the chapter on the _Ideal of Pure Reason_ and in the\nimportant _Appendix_ attached to the _Dialectic_ consists in its\nforcible and considered exposition of both attitudes. Most of the\nsections on the _Antinomies_ must, as we have seen, be dated as among\nthe earliest parts of the _Critique_. Their teaching is correspondingly\nimmature. The chapter on the _Ideal_ and the _Appendix_, on the other\nhand, were among the latest to be written, and contain, together with\nthe central portions of the _Analytic_, our most authoritative\nexposition of Kant\u2019s Critical principles.\n\n\n\n\nCHAPTER III\n\nTHE IDEAL OF PURE REASON\n\n\nSECTIONS I and II\n\nTHE TRANSCENDENTAL IDEAL[1601]\n\nThe statements of the first section cannot profitably be commented upon\nat this stage; they are of a merely general character.[1602] I pass at\nonce to Section II., which, as above stated, is quite the most archaic\npiece of rationalistic argument in the entire _Critique_. It is not\nmerely Leibnizian, but Wolffian in character. For Kant the Wolffian\nlogic had an old-time flavour and familiarity that rendered it by no\nmeans distasteful; and he is here, as it were, recalling, not altogether\nwithout sympathy, the lessons of his student years. They enable him to\nrender definite, by way of contrast, the outcome of his own Critical\nteaching.\n\nAs Kant here restates the Wolffian notion of the _Ens realissimum_ in\nsuch fashion as is required to make it conform to his deduction of the\ntheological Idea from the disjunctive syllogism, a preliminary statement\nof the more orthodox formulation will help to set Wolff\u2019s doctrine in a\nclearer light. In so doing, I shall follow Baumgarten, whose\n_Metaphysica_ Kant used as a class text-book. Briefly summarised\nBaumgarten\u2019s statement is as follows.[1603] The _Ens perfectissimum_ is\nthat Being which possesses as many predicates, _i.e._ perfections, as\ncan possibly exist together in a single thing, and in which every one of\nits perfections is as great as is anywhere possible. This most perfect\nBeing must be a real Being, and its reality must be the greatest\npossible. It is that in which the most and the greatest realities are.\nBut all realities are affirmative determinations, and no denial is a\nreality. Accordingly no reality can contradict another reality, and all\nrealities can exist together in the same thing. The _Ens\nperfectissimum_, in possessing all the realities that can exist\ntogether, must therefore possess all realities without exception, and\nevery one of them in the highest degree. The notion of an individual\nexistence that is at once _perfectissimum_ and also _realissimum_ is\nthus determinable by pure Reason from its internal resources. It is the\nground and condition of all other existences; all of them arise through\nlimitation of its purely positive nature.\n\nKant seeks to justify his metaphysical deduction of the _Ideal_ from the\ndisjunctive syllogism, by recasting the above argument in the following\nmanner. Since everything which exists is completely determined, it is\nsubject to the principle of complete determination, according to which\none of each of the possible pairs of contradictory predicates must be\napplicable to it. To be completely determined the thing must be compared\nwith the sum total of all possible predicates. Although this idea of the\nsum total of all possible predicates, through reference to which alone\nany concept can be completely determined, seems itself indeterminate, we\nfind nevertheless on closer examination that it individualises itself _a\npriori_, transforming itself into the concept of an individual existence\nthat is completely determined by the mere Idea, and which may therefore\nbe called an Ideal of pure Reason. That is proved as follows. No one can\ndefinitely think a negation unless he founds it on the opposite\naffirmation. A man completely blind cannot frame the smallest conception\nof darkness, because he has none of light. All negations are therefore\nderivative; it is the realities which contain the material by which a\ncomplete determination of anything becomes possible. The source, from\nwhich all possible predicates may be derived, can be nothing but the sum\ntotal of reality. And this concept of the _omnitudo realitatis_ is the\nIdea of a Being that is single and individual. As all finite beings\nderive the material of their possibility from it, they presuppose it,\nand cannot, therefore, constitute it. They are imperfect copies\n(_ectypa_), of which it is the sole Ideal. The Idea is also individual.\nOut of each possible pair of contradictory predicates, that one which\nexpresses reality belongs to it. By these infinitely numerous positive\npredicates it is determined to absolute concreteness; and as it\ntherefore possesses all that has reality, not only in nature but in man,\nit must be conceived as a personal and intelligent Primordial Being. The\nlogical Ideal, thus determining itself completely by its own concept,\nappears not only as ideal but also as real, not only as logical but also\nas divine.\n\nKant so far anticipates his criticism of the ontological argument as to\ngive, in the remaining paragraphs of this second section, a preliminary\ncriticism of this procedure. For the purpose for which the Ideal is\npostulated, namely, the determination of all finite and therefore\nlimited existences, Reason does not require to presuppose an existence\ncorresponding to it. Its mere Idea will suffice.\n\n     \u201cAll manifoldness of things is only a correspondingly varied mode\n     of limiting the concept of the highest reality which forms their\n     common substratum, just as all figures are only possible as so many\n     different modes of limiting infinite space.\u201d[1604]\n\nThis relation is not, however, that of a real existence to other things\nbut of an Idea to concepts. The Idea is a mere fiction, necessary for\ncomprehending the limited, not a reality that can be asserted, even\nhypothetically,[1605] as given along with the limited. None the less,\nowing to a natural transcendental illusion, the mind inevitably tends to\nhypostatise it, and so generates the object of rational theology.\n\n=Comment.=--The explanation of this illusion, which Kant proceeds to give\nin the two concluding paragraphs, is peculiarly confusing. Though the\nconcept of an all-comprehensive reality may, he argues, be required for\nthe definition of sensible objects, such a concept must not for that\nreason be taken as representing a real existence. The teaching of the\nsection on _Amphiboly_ is here entirely ignored; and the reader is\nbewildered by the assumption, which Kant apparently makes, that\nsomething analogous to the Leibnizian Ideal is a prerequisite of\npossible experience.\n\nThese last remarks indicate the kind of criticism to which the argument\nof this section lays itself open. In expounding the teaching of the\nLeibnizian science of rational theology, Kant strives to represent its\nIdeal as being an inevitable Idea of human Reason; and in order to make\nthis argument at all convincing he is constrained to treat as valid the\npresupposed ontology, though that has already been shown in the\ndiscussion of _Amphiboly_ to be altogether untenable.[1606] Limitation\nis not merely negative; genuine realities may negate one another. Though\nthe objects of sense presuppose the entire system to which they belong,\nthe form of this presupposition is in no respect analogous to that which\nWolff would represent as holding between finite existences and the _Ens\nrealissimum_. The passage in the _Analytic_[1607] in which Kant directly\ncontroverts the above teaching is as follows:\n\n     \u201cThe principle, that realities (as pure assertions) never logically\n     contradict each other ... has not the least meaning either in\n     regard to nature or in regard to any thing-in-itself.... Although\n     Herr von Leibniz did not, indeed, announce this proposition with\n     all the pomp of a new principle, he yet made use of it for new\n     assertions, and his followers expressly incorporated it in their\n     Leibnizian-Wolffian system. According to this principle all evils,\n     for instance, are merely consequences of the limitations of created\n     beings, _i.e._ negations, because negations alone conflict with\n     reality.... Similarly his disciples consider it not only possible,\n     but even natural, to combine all reality, without fear of any\n     conflict, in one being, because the only conflict which they\n     recognise is that of contradiction, whereby the concept of a thing\n     is itself removed. They do not admit the conflict of reciprocal\n     injury in which each of two real grounds destroys the effect of the\n     other--a process which we can represent to ourselves only in terms\n     of conditions presented to us in sensibility.\u201d\n\nThus the Ideal which Kant here declares to be a necessary Idea of Reason\nis denounced in the _Analytic_ as based on false principles peculiar to\nthe Leibnizian philosophy, and as \u201cwithout the least meaning in regard\neither to nature or to any thing in itself.\u201d The teaching of the\n_Analytic_ will no more combine with this scholastic rationalism than\noil with water. The reader may safely absolve himself from the thankless\ntask of attempting to render Kant\u2019s argumentation in these paragraphs\nconsistent with itself. Fortunately, in the next section, Kant returns\nto the standpoint proper to the doctrine he is expounding, and lays\nbare, with remarkable subtlety and in a very convincing manner, the\nconcealed dialectic by which the conclusions of this metaphysical\nscience are really determined.[1608]\n\n\nSECTION III\n\nTHE SPECULATIVE ARGUMENTS IN PROOF OF THE EXISTENCE OF A SUPREME\nBEING[1609]\n\n=Statement.=--Though the Ideal is not arbitrary, but is presupposed in\nevery attempt to define completely a finite concept, Reason would feel\nhesitation in thus transforming what is merely a logical concept into a\nDivine Existence, were it not that it is impelled from another\ndirection to derive reality from such a source. All existences known in\nexperience are contingent, and so lead us (owing to the constitution of\nour Reason) to assume an absolutely necessary Being as their ground and\ncause. Now when we examine our various concepts, to ascertain which will\ncover this notion of necessary existence, we find that there is one that\npossesses outstanding claims, namely, that Idea which contains a\ntherefore for every wherefore, which is in no respect defective, and\nwhich does not permit us to postulate any condition. The concepts of the\nIdeal and of the necessary alone represent the unconditioned; and as\nthey agree in this fundamental respect, they must, we therefore argue,\nbe identical. And to this conclusion we are the more inclined, in that,\nby thus idealising reality, we are at the same time enabled to realise\nour Ideal.\n\nThis line of argument, which starts from the contingent, is as little\nvalid as that which proceeds directly from the Ideal. But since these\narguments express certain tendencies inherent in the human mind, they\nhave a vitality which survives any merely forensic refutation. Though\nthe conclusions to which they lead are false, they are none the less\ninevitably drawn. Our acceptance of them is due to a transcendental\nillusion which may be detected as such, but which, like the ingrained\nillusions of sense-experience, must none the less persist.\n\nThe opening paragraph of Section V[1610] is the natural completion of\nthe above analysis. The ontological argument, in _starting_ from the\nconcept of the _Ens realissimum_, inverts the natural procedure. It is\n\u201ca merely scholastic innovation,\u201d and would never have been attempted\nsave for the need of finding some necessary Being, to which we may\nascend from contingent existence. It maintains that this necessary Being\nmust be unconditioned and _a priori_ certain, and accordingly looks for\na concept capable of fulfilling this requirement. Such a concept is\nsupposed to exist in the Idea of an _Ens realissimum_, and this Idea is\ntherefore used to gain more definite knowledge of that which has been\npreviously and independently recognised, namely, the necessary Being,\n\n     \u201cThis natural procedure of Reason was concealed from view, and\n     instead of ending with this concept, the attempt was made to begin\n     with it, and so to deduce from it that necessity of existence which\n     it was only fitted to complete. Thus arose the unfortunate\n     ontological proof, which yields satisfaction neither to the natural\n     and healthy understanding nor to the more academic demands of\n     strict proof.\u201d[1611]\n\nTo return to Section III.: Kant breaks the continuity of his argument,\nand anticipates his discussion of the cosmological proof, by stopping to\npoint out the illegitimacy of the assumption which underlies the first\nstep in the above argument, namely, that a limited being cannot be\nabsolutely necessary. Though the concept of a limited being does not\ncontain the unconditioned, that does not prove that its existence is\nconditioned. Indeed each and every limited being may, for all their\nconcepts show to the contrary, be unconditionally necessary.[1612] The\nabove argument is consequently inconclusive, and cannot be relied on to\ngive us any concept whatever of the qualities of a necessary Being. But\nthis is a merely logical defect, and, as already noted, it is not really\nupon logical cogency that the persuasive force of the argument depends.\n\nIn conclusion Kant points out that there are only three possible kinds\nof speculative (_i.e._ theoretical) proofs of the existence of God: (1)\nfrom definite experience and the specific nature of the world of sense\nas revealed in experience; (2) from indefinite experience, _i.e._ from\nthe fact that any existence at all is empirically given; (3) the\nnon-empirical _a priori_ proof from mere concepts. The first is the\n_physico-theological_ or _teleological_ argument, the second is the\n_cosmological_, and the third is the _ontological_. Kant finds it\nadvisable to reverse the order of the proofs, and to begin by\nconsideration of the ontological argument. This would seem to indicate\nthat the \u2018scholastic innovation\u2019 to which he traces the origin of the\nontological proof has more justification than his remarks appear to\nallow.\n\n\nSECTION IV\n\nTHE IMPOSSIBILITY OF AN ONTOLOGICAL PROOF[1613]\n\n=Statement.=--Hitherto Kant has employed the concept of an absolutely\nnecessary Being without question. He now recognises that the problem,\nfrom which we ought to start, is not whether the existence of an\nabsolutely necessary Being can be demonstrated, but whether, and how,\nsuch a Being can even be conceived. And upon analysis he discovers that\nthe assumed notion of an absolutely necessary, _i.e._ _unconditioned_\nBeing is entirely lacking in intelligible content. For in eliminating\nall conditioning causes--through which alone the understanding can\nconceive necessity of existence--we also remove this particular kind of\nnecessity. A verbal definition may, indeed, be given of the Idea, as\nwhen we say that it represents something the non-existence of which is\nimpossible. But this yields no insight into the reasons which make its\nnon-existence inconceivable, and such insight is required if anything at\nall is to be thought in the Idea.\n\n     \u201cThe expedient of removing all those conditions which the\n     understanding indispensably requires in order to regard something\n     as necessary, simply through the introduction of the word\n     _unconditioned_, is very far from sufficing to show whether I am\n     still thinking anything, or not rather perhaps nothing at all, in\n     the concept of the unconditionally necessary.\u201d[1614]\n\nThe untenableness of the concept has been in large part concealed\nthrough a confusion between logical and ontological necessity, that is,\nbetween necessity of judgment and necessity of existence. The fact that\nevery proposition of geometry must be regarded as absolutely necessary\nwas supposed to justify this identification. It was not observed that\nlogical necessity refers only to _judgments_, not to _things_ and their\nrelations, and that the absolute necessity of the judgment holds only\nupon the assumption that the conditioned necessity of the thing referred\nto has previously been granted. If there be any such thing as a\ntriangle, the assertion that it has three angles will follow with\nabsolute necessity; but the existence of a triangle or even of space in\ngeneral is contingent. In other words, the asserted necessity is only a\nform of logical sequence, not the unconditioned necessity of existence\nwhich is supposed to be disclosed in the Idea of Reason. All judgments,\nso far as they refer to existence, as distinct from mere possibility,\nare hypothetical, and serve to define a reality that is only\ncontingently given. In adopting this position, Kant is in entire\nagreement with Hume. The contradictory of a matter of fact is always\nthinkable. There has, Kant claims, been no more fruitful source of\nillusion throughout the whole history of philosophy than the belief in\nan absolute necessity that is purely logical.[1615] In the ontological\nargument we have the most striking instance of such rationalistic\nexaggeration of the powers of thought.\n\n=Comment.=--Had this criticism of the Idea of unconditioned necessity been\nintroduced at an earlier stage in Kant\u2019s argument, much confusion would\nhave been avoided. It involves the thorough revisal of his criticism of\nthe third and fourth antinomies, as well as of the whole account\nhitherto given of the function of Reason and of its metaphysical\ndialectic. The principle, that if the conditioned be given, the whole\nseries of conditions up to the unconditioned is likewise given, must no\nlonger be accepted as a basis for argument. Indeed the very terms in\nwhich Reason has so far been defined, as the faculty of the\nunconditioned, become subject to question. In that definition the term\nunconditioned has tacitly been taken as equivalent to the\nunconditionally necessary, and on elimination of the element of\nnecessity, it will reduce merely to the concept of totality, which is a\npure form of the understanding. Those parts of the _Dialectic_, which\nembody the view that Reason is simply the understanding transcendently\nemployed, will thus be confirmed; the alternative view of Reason as a\nseparate faculty will have to be eliminated. But these are questions\nwhich Kant himself proceeds to raise and discuss.[1616] Meantime he\napplies the above results in criticism of the ontological argument.\n\n=Statement.=--In an identical judgment it is contradictory to reject the\npredicate while retaining the subject. But there is no contradiction if\nwe reject subject and predicate alike, for nothing is then left that can\nbe contradicted. If we assume that there is a triangle, we are bound to\nrecognise that it has three angles, but there is no contradiction in\nrejecting the triangle together with its three angles. The same holds\ntrue of an absolutely necessary Being. \u2018God is omnipotent\u2019 is an\nidentical and therefore necessary judgment. But if we say, \u2018There is no\nGod,\u2019 neither the omnipotence nor any other attribute remains; and there\nis therefore not the least contradiction in saying that God does not\nexist. The only way of evading this conclusion is to argue that there\nare subjects which cannot be removed out of existence. That, however,\nwould only be another way of asserting that there exist absolutely\nnecessary subjects, and that is the very assertion which is now in\nquestion, and which the ontological argument undertakes to prove. Our\nsole test of what cannot be removed is the contradiction which would\nthereby result; and the only possible instance which can be cited is the\nconcept of the _Ens realissimum_. It remains, therefore, to establish\nthe above criticism for this specific case.\n\nAt the start Kant points out that absence of internal contradiction,\neven if granted, proves only that the _Ens realissimum_ is a _logically_\npossible concept (as distinguished from the _nihil negativum_[1617]); it\ndoes not suffice to establish the possibility of the object of the\nconcept. But for the sake of argument Kant allows this initial\nassumption to pass. The argument to be disproved is that as reality\ncomprehends existence, existence is contained in the concept of _Ens\nrealissimum_, and cannot therefore be denied of it without removing its\ninternal possibility. The really fundamental assumption of this argument\nis that existence is capable of being included in the concept of a\n_possible_ being. If that were so, the assertion of its existence would\nbe an analytic proposition, and the proof could not be challenged. (The\nassumption is partly concealed by alternation of the terms reality and\nexistence: in their actual employment they are completely synonymous.)\nAs the above assumption thus decides the entire issue, Kant sets himself\nto establish, in direct opposition to it, the thesis, that every\nproposition which predicates existence is synthetic, and that in\nconsequence its denial can never involve a logical contradiction.\nExistence can never form part of the content of a conception, and\ntherefore must not be regarded as a possible predicate. What logically\ncorresponds to it in a judgment is a purely formal factor, namely, the\ncopula. The proposition, \u2018God is omnipotent,\u2019 contains two concepts,\neach of which has its object--God and omnipotence. The word \u2018is\u2019 adds no\nnew predicate, but only serves to posit the predicate in its relation to\nthe subject. Similarly, when we take the subject together with all its\npredicates (including that of omnipotence), and say, \u2018God is\u2019 or \u2018there\nis a God,\u2019 we attach no new predicate to the concept of God, but only\nposit the subject in itself with all its predicates as being an object\nthat stands in relation to our concept. In order that the proposition be\ntrue, the _content_ of the object and of the concept must be one and the\nsame. If the object contained more than the concept, the concept would\nnot express the object, and the proposition would assert a relation that\ndoes not hold. Or to state the same point in another way, the real must\nnot contain more content than the possible. Otherwise it would not be\nthe possible, but something different from the possible, which would\nthen be taken as existing. A hundred real thalers do not contain the\nleast coin more than a hundred possible thalers. Though my financial\nposition is very differently affected by a hundred real thalers than by\nthe thought of them only, a conceived hundred thalers are not in the\nleast increased through acquiring existence outside my concept.\n\nKant presents his argument in still another form. If we think in a thing\nevery kind of reality except one, the missing reality is not supplied by\nmy saying that this defective thing exists. On the contrary, it exists\nwith the same defect with which I have thought it. When, therefore, I\nthink a Being as the highest reality, without any defect, the question\nstill remains whether it exists or not. For though, in my concept,\nnothing may be lacking of the possible real _content_ of a thing in\ngeneral, something is still lacking in its relation to my whole state of\nthinking, namely, knowledge of its existence; and such knowledge can\nnever be obtained save in an _a posteriori_ manner. That is owing to the\nlimitations imposed by the conditions of our sense-experience. We never\nconfound the existence of a sensible object with its mere concept. The\nconcept represents something that may or may not exist: to determine\nexistence we must refer to actual experience. As Kant has already\nstated, the actual is always for us the accidental, and its assertion is\ntherefore synthetic. A possible idea and the idea of a possible thing\nare quite distinct.[1618] A thing is known to be possible only when\npresented in some concrete experience, or when, though not actually\nexperienced, it has been proved to be bound up, according to empirical\nlaws, with given perceptions. It is not, therefore, surprising that if\nwe try, as is done in the ontological argument, to think existence\nthrough the pure category, we cannot mention a single mark\ndistinguishing it from a merely logical possibility. The concept of a\nSupreme Being is, in many respects, a valuable Idea, but just because it\nis an Idea of pure Reason, _i.e._ a mere Idea, we can no more extend our\nknowledge of real existence by means of it, than a merchant can better\nhis position by adding a few noughts to his cash account.\n\nThere are many points of connection between this section and the first\nedition _Introduction_; and in view of these points of contact Adickes\nhas suggested[1619] that the considerations which arose in the\nexamination of the ontological argument may have been what brought Kant\nto realise that the various problems of the _Critique_ can all be traced\nto the central problem of _a priori_ synthesis.\n\n\nSECTION V\n\nTHE IMPOSSIBILITY OF A COSMOLOGICAL PROOF OF THE EXISTENCE OF GOD[1620]\n\n=Statement.=--Kant, as already noted, views the ontological proof as \u2018a\nmere innovation of scholastic wisdom\u2019 which restates, in a quite\nunnatural form, a line of thought much more adequately expressed in the\ncosmological proof. To discover the natural dialectic of Reason we must\ntherefore look to this latter form of argument. It is composed of two\ndistinct stages. In the first stage it makes no use of specific\nexperience: if _anything_ is given us as existing, _e.g._ the self,\nthere must exist an absolutely necessary Being as its cause. Then, in\nthe second stage, it is argued that as such a Being must be altogether\noutside experience, Reason must leave experience entirely aside, and\ndiscover from among pure concepts what properties an absolutely\nnecessary Being ought to possess, _i.e._ which among all possible things\ncontains in itself the conditions of absolute necessity. The requisite\nenlightenment is believed by Reason to be derivable only from the\nconcept of an _Ens realissimum_, and Reason therefore at once concludes\nthat this concept must represent the absolutely necessary Being.\n\nNow in that final conclusion the truth of the ontological argument is\nassumed. If the concept of a Being of the highest reality is so\ncompletely adequate to the concept of necessary existence that they can\nbe regarded as identical, the latter must be capable of being derived\nfrom the former, and that is all that is maintained in the ontological\nproof. To make this point clearer, Kant states it in scholastic form. If\nthe proposition be true, that every absolutely necessary Being is at the\nsame time the most real Being (and this is the _nervus probandi_ of the\ncosmological proof in so far as it is also theological), it must, like\nall affirmative propositions, be capable of conversion, at least _per\naccidens_. This gives us the proposition that some _Entia realissima_\nare at the same time absolutely necessary Beings. One _Ens realissimum_,\nhowever, does not differ from another, and what applies to one applies\nto all. In this case, therefore, we must employ simple conversion, and\nsay that every _Ens realissimum_ is a necessary Being. Thus the\ncosmological proof is not only as illusory as the ontological, but also\nless honest. While pretending to lead us by a new road to a sound\nconclusion, it brings us back, after a short circuit, into the old path.\nIf the ontological argument is correct, the cosmological is superfluous;\nand if the ontological is false, the cosmological cannot possibly be\ntrue.\n\nBut the first stage of the cosmological argument, that by which it is\ndistinguished from the ontological, is itself fallacious. A whole nest\nof dialectical assumptions lies hidden in its apparently simple and\nlegitimate inference from the contingent to the necessary. To advance\nfrom the contingent to the necessary, from the relative to the absolute,\nfrom the given to the transcendent, is just as illegitimate as the\nopposite process of passing from Idea to existence. The necessity of\nthought, which is in both cases the sole ground of the inference, is\nfound on examination to be of merely subjective character. No less than\nthree false assumptions are involved in this inference. In the first\nplace, the principle that everything must have a cause, which can be\nproved to be valid only within the world of sense, is here applied to\nthe sensible world as a whole; and is therefore employed in the wider\nform which coincides with the fundamental principle of the higher\nfaculty of Reason. We assume, that if the conditioned be given, the\ntotality of its conditions _up to the unconditioned_ is given likewise.\nNo such principle can be granted. As it is synthetic, it could be\nestablished only as a condition of the possibility of experience. But no\nsuch proof is offered: the principle is based upon a purely intellectual\nconcept. Secondly, the inference to a first cause rests on the kindred\nassumption that an infinite series of empirical causes is impossible.\nThat conclusion can never be drawn, even within the realm of experience.\nHow, then, can we rely upon it in advancing beyond experience?\nCertainly, no one can prove that the empirical series is infinite, but\njust as little can we establish the opposite. In discussing the third\nand fourth antinomies Kant has shown that the existence of a first cause\nor of an absolutely necessary Being, though possible (or rather,\npossibly possible), is never demonstrable. Thirdly--as has been shown in\nA 592-3 = B 620-1--in inferring to an unconditioned cause, it is blindly\nassumed that the removal of all conditions does not at the same time\nremove the very concept of necessity. Our only notion of necessity is\nderived from experience, and therefore depends on those finite\nconditions which the argument would deny to us. The concept of\nunconditioned necessity is entirely null and void.\n\nThe fourth defect, which Kant enumerates, refers to the second stage of\nthe cosmological argument, and has already been considered. He ought\nalso to have mentioned a still further assumption underlying its first\nstage, namely, that a concept which represents a limited being, as, for\ninstance, that of matter, cannot represent necessary existence. This\nalso is an assumption which it cannot justify. This objection Kant has\nhimself stated in A 586 = B 614 and A 588 = B 616.[1621]\n\n=Comment.=--We are apt to overlook the wider sweep which Kant\u2019s criticism\ntakes in this section, owing to his omission to notify the reader that\nhe is here calling in question a principle which he has hitherto been\ntaking for granted, namely, the principle in terms of which he has in\nthe opening sections of the _Dialectic_ defined the faculty of Reason,\nthat if the conditioned be given the totality of conditions up to the\nunconditioned is given likewise. The first step in his rejection of this\nprinciple occurs as merely incidental to his criticism of the\nontological argument. It is there shown that the concept of the\nunconditionally necessary is without meaning. Now, in this present\nsection, he calls in question the principle itself. It must be rejected\nnot only, as stated in the third of the above objections, because the\nconcept of the unconditioned, which tacitly implies the factor of\nabsolute necessity, is without real significance, but also for two\nfurther reasons--those above cited in the first and second objections.\nHow very differently the problems of the _Dialectic_ appear, and how\nvery differently the Ideas of Reason have to be regarded, when this\nprinciple, and also the concept of the unconditioned of which it is the\napplication, are thus called in question, will be shown in the sequel.\n\nDISCOVERY AND EXPLANATION OF THE TRANSCENDENTAL ILLUSION IN ALL\nTRANSCENDENTAL PROOFS OF THE EXISTENCE OF A NECESSARY BEING[1622]\n\n=Statement.=--We do not properly fulfil the task prescribed by Critical\nteaching in merely disproving the cosmological argument. We must also\nexplain its hold upon the mind. If it is, as Kant insists, more natural\nto the mind than the ontological, and yet, as we have just seen, is more\nfallacious; if it has not been invented by philosophers, but is the\ninstinctive reasoning of the natural man, it must rest, like all\ndialectical illusion, upon a misunderstanding of the legitimate demands\nof pure Reason. _Reason demands the unconditioned, and yet cannot think\nit._\n\n     \u201cUnconditioned necessity, which we so indispensably require as the\n     last bearer of all things, is for human Reason the veritable\n     abyss.... We can neither help thinking, nor can we bear the\n     thought, that a Being--even if it be the one which we represent to\n     ourselves as supreme amongst all Beings--should, as it were, say to\n     itself: \u2018I am from eternity to eternity, and outside me there is\n     nothing save what is through my will; _but whence am I_?\u2019 All\n     support here fails us; and _supreme_ perfection, no less than the\n     _least_ perfection, is unsubstantial and baseless for the merely\n     speculative Reason....\u201d[1623]\n\nWe are obliged to think something as necessary for all existence, and\nyet at the same time are unable to think anything as in itself\nnecessary--God as little as anything else.\n\nThe explanation[1624] of this strange fact must be that which follows as\na corollary from the limitation of our knowledge to sense-experience,\nnamely, that our concepts of necessity and contingency do not concern\nthings in themselves, and cannot therefore be applied to them in\naccordance with either of the two possible alternatives. Each\nalternative must express a subjective principle of Reason; and the two\ntogether (that something exists by necessity, and that everything is\nonly contingent) must form complementary rules for the guidance of the\nunderstanding. These rules will then be purely heuristic and regulative,\nrelating only to the formal interests of Reason, and may well stand side\nby side. For the one tells us that we ought to philosophise as if there\nwere a necessary first ground for everything that exists, _i.e._ that we\nought to be always dissatisfied with relativity and contingency, and to\nseek always for what is unconditionally necessary. The other warns us\nagainst regarding any single determination in things (such, for\ninstance, as impenetrability or gravity) as absolutely necessary, and so\nbids us keep the way always open for further derivation. In other words,\nReason guides the understanding by a twofold command. The understanding\nmust derive phenomena and their existence from other phenomena, just as\nif there were no necessary Being at all; while at the same time it must\nalways strive towards the completeness of that derivation, just as if\nsuch a necessary Being were presupposed. It is owing to a transcendental\nillusion or subreption that we view the latter principle as\nconstitutive, and so think its unity as hypostatised in the form of an\n_Ens realissimum_. The falsity of this substitution becomes evident as\nsoon as we consider _that unconditioned necessity, as a thing in itself,\ncannot even be conceived, and that the \u201cIdea\u201d of it cannot, therefore,\nbe ascribed to Reason save as a merely formal principle, regulative of\nthe understanding in its interpretation of given experience_.[1625]\n\n=Comment.=--The reader may observe that, when Kant is developing this\nsceptical view of the Ideal of Reason, the explanation of dialectical\nillusion in terms of transcendental idealism falls into the background.\nThe illusion is no longer traced to a confusion between appearances and\nthings in themselves, but to the false interpretation of regulative\nprinciples as being constitutive. When it is the cosmological problem\nwith which we are dealing, the two illusions do, indeed, coincide. If we\nview the objects of sense-experience as things in themselves, we are\nbound to regard the Ideal completion of the natural sciences as an\nadequate representation of ultimate reality. But in Rational Theology,\nwhich is professedly directed towards the definition of a Being distinct\nfrom nature and conditioning all finite existence, it is not failure to\ndistinguish between appearance and things in themselves, but the\nmistaking of a merely formal Ideal for a representation of reality, that\nis alone responsible for the conclusions drawn.\n\nIn A 617-18 = B 645-6 Kant makes statements which conflict with the\nteaching of A 586 = B 614 and A 588 = B 616. In the latter passages he\nhas argued that the concept of a limited being may not without specific\nproof be taken as contradictory of absolute necessity. He now\ncategorically declares that the philosophers of antiquity are in error\nin regarding matter as primitive and necessary; and the reason which he\ngives is that the regulative principle of Reason forbids us to view\nextension and impenetrability, \u201cwhich together constitute the concept of\nmatter,\u201d as ultimate principles of experience. But obviously Kant is\nhere going further than his regulative principle will justify. It\ndemands only that we should always look for still higher principles of\nunity, and so keep open the way for possible further derivation; it does\nnot enable us to assert that such will actually be found to exist.\nNotwithstanding the Ideal demands of the regulative principle, matter\nmay be primordial and necessary, and its properties of extension and\nimpenetrability may not be derivable from anything more ultimate.\n\nIn this connection we may raise the more general question, how far the\nIdeal demand for necessity and unity in knowledge and existence can be\nconcretely pictured. Kant gives a varying answer. Sometimes--when he is\nemphasising the limitation of our theoretical knowledge to\nsense-experience--he reduces the speculative Idea of Divine Existence to\na purely abstract maxim for the regulation of natural science. When the\nIdeal occupies the mind on its own account, and so attracts our\nattention away from our sense-knowledge, it is an unreality, and\nperverts the understanding; it yields genuine light and leading only as\na quite general maxim within the sphere of natural science. From this\npoint of view necessary Being, even as an Ideal, can by no means be\nidentified with a personal God. It signifies only the highest possible\nsystem and unity of the endlessly varied natural phenomena in space and\ntime, and can be approximately realised in the most various ways. Its\nsignificance is entirely cosmological. It is an Ideal of positive\nscience, and signifies only systematic unity in the object known. In\nbeing transformed from a scientific ideal into a subject of theological\nenquiry, it has inevitably given rise to dialectical illusion. At other\ntimes,--when he is concerned to defend the concept of Divine Existence\nas at least possible, and so to prepare the way for its postulation as\nimplied in the moral law, or when he is seeking, as in the _Critique of\nJudgment_, to render comprehensible the complete adaptation of\nphenomenal nature in its material aspect to the needs of our\nunderstanding--Kant insists that we are ultimately compelled, by the\nnature of our faculties, to conceive the Ideal of Reason as a personal\nGod, as an Intelligence working according to purposes. Only by such a\npersonal God, he maintains, can the demands of Reason be genuinely\nsatisfied.\n\nThese two interpretations of the Ideal of Reason are in conflict with\none another; and so far as the _Critique of Pure Reason_ is concerned, a\nvery insufficient attempt is made to justify the frequent assertion that\nthe Idea of God is _the_ Ideal of Reason, and not merely one possible,\nand highly problematic, interpretation of it. If the Idea of God is a\nnecessary Idea, it cannot be adequately expressed through any merely\nregulative maxim. It demands not only system in knowledge but also\nperfection in the nature of the known. It is not a merely logical Ideal\nsuch as might be satisfied by any rational system, but an Ideal which\nconcerns matter as well as form, man as well as nature, our moral needs\nas well as our intellectual demands. If Kant is to maintain that the\nonly genuine function of theoretical Reason is to guide the\nunderstanding in its scientific application, he is debarred from\nasserting that a concrete interpretation of its regulative principles is\nunavoidable. And he is also precluded by his own limitation of all\nknowledge to sense-experience from seeking to define by any positive\npredicate the transcendent nature of the thing in itself.\n\nSuch justification as Kant can offer in support of his assertion that\nthe Idea of God, of Intelligent Perfection, is an indispensable Idea of\nhuman Reason, is chiefly based upon the teleological aspect of nature\nwhich is dealt with in the physico-theological proof. Mechanical science\nimplies only the cosmological Idea: teleological unity presupposes the\ntheological Ideal. Further enquiry, then, into the necessity of the Idea\nof God as a regulative principle, and its dangers as a source of\ndialectical illusion, we must defer until we have examined the one\nremaining argument.[1626]\n\n\nSECTION VI\n\nTHE IMPOSSIBILITY OF THE PHYSICO-THEOLOGICAL PROOF[1627]\n\n=Statement.=--The teleological proof starts from our definite knowledge of\nthe order and constitution of the sensible world. The actual world\npresents such immeasurable order, variety, fitness, and beauty, that we\nare led to believe that here at least is sufficient proof of the\nexistence of God. Kant\u2019s attitude towards this argument is at once\nextremely critical and extremely sympathetic. Though he represents it as\nthe oldest, the clearest, and the most convincing, he is none the less\nprepared to show that it contains every one of the fallacies involved in\nthe other two proofs, as well as some false assumptions peculiar to\nitself. It possesses overpowering persuasive force, not because of any\ninherent logical cogency, but because it so successfully appeals to\nfeeling as to silence the intellect. It would, Kant declares, be not\nonly comfortless, but utterly vain to attempt to diminish its influence.\n\n     \u201d[The mind is] aroused from the indecision of all melancholy\n     reflection, as from a dream, by one glance at the wonders of nature\n     and the majesty of the universe....\u201d[1628]\n\nMeantime, however, we are concerned with its merely logical force. We\nhave to decide whether, as theoretical proof, it can claim assent on its\nown merits, requiring no favour, and no help from any other quarter. On\nthe basis of empirical facts the argument makes the following\nassertions. (1) There are everywhere in the world clear indications of\nadaptation to a definite end. (2) As this adaptation cannot be due to\nthe working of blind, mechanical laws, and accordingly cannot be\nexplained as originating in things themselves, it must have been imposed\nupon them from without; and there must therefore exist, apart from the\nsensible world, an intelligent Being who has arranged it according to\nideas antecedently formed. (3) As there is unity in the reciprocal\nrelations of the parts of the universe as portions of a single edifice,\nand as the universe is infinite in extent and inexhaustible in variety,\nits intelligent cause must be single, all-powerful, all-wise, i.e. God.\n\nNow, even granting for the sake of argument the admissibility of these\nassertions, they enable us to infer only an intelligent author of the\npurposive form of nature, not of its matter, only an architect who is\nvery much hampered by the inadaptability of the material in which he\nhas to work, not a Creator to whose will everything is due. To prove the\ncontingency of matter itself, we should have to establish the truth of\nthe cosmological proof.\n\nBut the assumptions implied even in the demonstration that God exists as\na _formative_ power, are by no means beyond dispute. Why may not nature\nbe regarded as giving form to itself by its blindly working forces? Can\nit really be proved that nature is a work of art that demands an\nartificer as certainly as does a house, or a ship, or a clock? Kant\u2019s\nargument is at this point extremely brief, and I shall so far digress\nfrom the statement of it, which he here gives, as to supplement it from\nhis other writings. Even so-called dead matter is not merely inert. By\nits inherent powers of gravity and chemical attraction it spontaneously\ngives rise to the most wonderful forms. When Clarke and Voltaire, in\ntheir first enthusiasm over Newton\u2019s great discovery, asserted that the\nplanetary system must have been divinely created, each planet being\nlaunched in the tangent of its orbit by the finger of God, just as a\nwheel must be fixed into its place by the hand of the mechanician, they\nunder-estimated the organising power of blind inanimate nature. As Kant\nargued in his early treatise,[1629] the planetary system can quite well\nhave arisen, and, as it would seem, actually has come into existence,\nthrough the action of blindly working laws. The mechanical principles\nwhich account for its present maintenance will also account for its\norigin and development. But it is when we turn to animate nature, which\nis the chief source from which arguments for design are derived, that\nthe insufficiency of the teleological argument becomes most manifest. As\nKant points out in the _Critique of Judgment_, the differentia\ndistinguishing the living from the lifeless, is not so much that it is\norganised as that it is self-organising. When, therefore, we treat an\norganism as an analogon of art we completely misrepresent its essential\nnature.[1630] In regarding it as put together by an external agent we\nare ignoring its internal self-developing power. As Hume had previously\nmaintained in his _Dialogues on Natural Religion_,[1631] the facts of\nthe organic world not only agree with the facts of the inorganic world\nin _not_ supporting the argument of the teleological proof, but are in\ndirect conflict with it.\n\nBut to return to Kant\u2019s immediate statement of the argument. Setting the\nabove objection aside, and granting for the present that nature may be\nregarded as the outcome of an external artificer, we can argue only to a\ncause adequate to its production, _i.e._ to an extraordinarily wise and\nwonderfully powerful Being. Even if we ignore the existence of evil and\ndefect in nature, the step from great power to omnipotence, and from\ngreat wisdom to omniscience, is one that can never be justified on\nempirical grounds.[1632] Since the Ideas of Reason, and above all the\ncompletely determined, individual Ideal of Reason, transcend experience,\nexperience can never justify us in inferring their reality. The\nteleological argument can, indeed, only lead us to the point of admiring\nthe greatness, wisdom, and power of the author of the world. In\nproceeding further it abandons experience altogether, and reasons, not\nfrom particular kinds and excellencies of natural design, but from the\ncontingency of all such adaptation to the existence of a necessary\nBeing, exactly in the manner of the cosmological argument. And it ends\nby assuming, in agreement with the ontological proof, that the only\npossible necessary Being is the Ideal of Reason. Thus after committing a\nnumber of fallacies on its own account, the teleological argument itself\nendorses all those that are involved in the more _a priori_ proofs. The\nteleological argument rests on the cosmological, and the cosmological on\nthe ontological, which therefore would be the only proof possible, were\nthe proof of a completely transcendent proposition ever possible at all.\nThe strange fact that the convincing force of the arguments thus varies\ninversely with their validity shows, Kant maintains, that we are correct\nin concluding that they do not really depend upon their logical cogency,\nand merely express, in abstract terms, beliefs deep-rooted in the human\nspirit.\n\n\nSECTION VII\n\nCRITICISM OF ALL THEOLOGY BASED ON SPECULATIVE PRINCIPLES OF\nREASON[1633]\n\n=A 631-3 = B 659-66.=--On the distinction between \u201ctheist\u201d and \u201cdeist,\u201d\ncf. _Encyclop\u00e6dia Britannica_, vii. p. 934:\n\n     \u201cThe later distinction between \u2018theist\u2019 and \u2018deist,\u2019 which stamped\n     the latter word as excluding the belief in providence or in the\n     immanence of God, was apparently formulated in the end of the\n     eighteenth century by those rationalists who were aggrieved at\n     being identified with the naturalists.\u201d\n\n=A 633-4 = B 661-2.=--Kant here does no more than indicate that by way of\npractical Reason it may be possible to postulate, though not\ntheoretically to comprehend, a Supreme Being. On the distinction between\npostulates and hypotheses, cf. A 769 ff. = B 797 ff., and below, p. 543\nff. Cf. also p. 571 ff.\n\n=A 634 = B 662.=--On relative necessity, cf. below, pp. 555, 571 ff.\n\n=A 635-9 = B 663-7= only summarises points already treated.\n\n=A 639-42 = B 667-70.=--Kant concludes by declaring that the Ideal, in\naddition to its regulative function, possesses two further prerogatives.\nIn the first place, it supplies a standard, in the light of which any\nknowledge of Divine Existence, acquired from other sources, can be\npurified and rendered consistent with itself. For it is \u201can Ideal\nwithout a flaw,\u201d the true crown and culmination of the whole of human\nknowledge.\n\n     \u201cIf there should be a moral theology ... transcendental theology\n... will then prove itself indispensable in determining its concept\n     and in constantly testing Reason which is so often deceived by\n     sensibility, and which is frequently out of harmony with its own\n     Ideas.\u201d[1634]\n\nAnd secondly, though the Ideal fails to establish itself theoretically,\nthe arguments given in its support suffice to show the quite\ninsufficient foundations upon which all atheistic, deistic, and\nanthropomorphic philosophies rest.\n\n=Comment.=--These concluding remarks cannot be accepted as representing\nKant\u2019s true teaching. The Ideal, by his own showing, is by no means\nwithout a flaw. In so far as it involves the concept of unconditioned\nnecessity, it is meaningless; it is purely logical, and therefore\ncontains no indication of real content; it embodies a false view of the\nnature of negation, and therefore of the relation of realities to one\nanother. In short, it is constituted in accordance with the false,\nun-Critical principles of Leibnizian metaphysics, and is found on\nexamination to be non-existent even as a purely mental entity. Reduced\nto its proper terms, it becomes a mere schema regulative of the\nunderstanding in the extension of experience, and does not yield even a\nnegative criterion for the testing of our ideals of Divine Existence.\nThe criterion, which Kant really so employs, is not that of an _Ens\nrealissimum_, but the concept of an Intuitive Understanding, which, as\nhe has indicated in the chapter on _Phenomena and Noumena_,[1635] is our\nmost adequate Ideal of completed Perfection. This latter is not itself,\nhowever, a spontaneously formed concept of natural Reason, and does not\njustify the assertion that the Idea of God is a necessary Idea of the\nhuman mind. In attempting to defend such a thesis, Kant is unduly\ninfluenced by the almost universal acceptance of deistic beliefs in the\nEurope of his time.[1636] His criticism of the Ideal of Reason and of\nrational theology is much more destructive, and really allows that\ntheology much less value, even as natural dialectic, than he is willing\nto admit.[1637] Architectonic forbids that the extreme radical\nconsequences of the teaching of the _Analytic_ should be allowed to show\nin their full force. These shortcomings are, however, in great part\nremedied in the elaborate _Appendix_ which Kant has attached to the\n_Dialectic_.\n\n\n\n\nAPPENDIX TO THE TRANSCENDENTAL DIALECTIC\n\n\nTHE REGULATIVE EMPLOYMENT OF THE IDEAS OF PURE REASON[1638]\n\nBefore we proceed to deal with this _Appendix_ it will be of advantage\nto consider the section in the _Methodology_ on _the Discipline of Pure\nReason in regard to Hypotheses_.[1639] That section affords a very\nilluminating introduction to the problems here discussed, and is\nextremely important for understanding Kant\u2019s view of metaphysical\nscience as yielding either complete certainty or else nothing at all.\nThis is a doctrine which he from time to time suggests, to the\nconsiderable bewilderment of the modern reader.[1640] In discussing it\nhe starts from the obvious objection, that though nothing can be known\nthrough Reason in its pure _a priori_ employment, metaphysics may yet be\npossible in an empirical form, as consisting of hypotheses, constructed\nin conjectural explanation of the facts of experience. Kant replies by\ndefining the conditions under which alone hypotheses can be entertained\nas such. There must always be something completely certain, and not only\ninvented or merely \u201copined,\u201d namely, the _possibility_ of the object to\nwhich the hypothesis appeals. Once that is proved, it is allowable, on\nthe basis of experience, to form opinions regarding its reality. Then,\nand only then, can such opinions be entitled hypotheses. Otherwise we\nare not employing the understanding to explain; we are simply indulging\nthe imagination in its tendency to dream. Now since the categories of\nthe pure understanding do not enable us to invent _a priori_ the concept\nof a dynamical connection, but only to apprehend it when presented in\nexperience, we cannot by means of these categories invent a single\nobject endowed with a new quality not empirically given; and cannot,\ntherefore, base an hypothesis upon any such conception.\n\n     \u201cThus it is not permissible to invent any new original powers, as,\n     for instance, an understanding capable of intuiting its objects\n     without the aid of senses; or a force of attraction without any\n     contact; or a new kind of substance existing in space and yet not\n     impenetrable. Nor is it legitimate to postulate any other form of\n     communion of substances than that revealed in experience, any\n     presence that is not spatial, any duration that is not temporal. In\n     a word our Reason can employ as conditions of the possibility of\n     things only the conditions of possible experience; it can never, as\n     it were, _create_ concepts of things, independently of those\n     conditions. Such concepts, though not self-contradictory, would be\n     without an object.\u201d[1641]\n\nThis does not, however, mean that the concepts of pure Reason can have\nno valid employment. They are, it is true, Ideas merely, with no object\ncorresponding to them in any experience; but then it is also true that\nthey are not hypotheses, referring to imagined objects, supposed to be\npossibly real. They are purely problematic. They are heuristic fictions\n(_heuristische Fiktionen_), the sole function of which is to serve as\nprinciples regulative of the understanding in its systematic employment.\nUsed in any other manner they reduce to the level of merely mental\nentities (_Gedankendinge_) whose very possibility is indemonstrable, and\nwhich cannot therefore be employed as hypotheses for the explanation of\nappearances. Given appearances can be accounted for only in terms of\nlaws known to hold among appearances. To explain natural phenomena by a\ntranscendental hypothesis--mental processes by the assumption of the\nsoul as a substantial, simple, spiritual being, or order and design in\nnature by the assumption of a Divine Author--is never admissible.\n\n     \u201c...that would be to explain something, which in terms of known\n     empirical principles we do not understand sufficiently, by\n     something which we do not understand at all.\u201d[1642]\n\nAnd Kant adds that the wildest hypotheses, if only they are physical,\nare more tolerable than a hyperphysical one. They at least conform to\nthe conditions under which alone hypothetical explanation as such is\nallowable. \u201cOutside this field, to form opinions, is merely to play with\nthoughts....\u201d[1643]\n\nA further condition, required to render an hypothesis acceptable, is its\nadequacy for determining _a priori_ all the consequences which are\nactually given. If for that purpose supplementary hypotheses have to be\ncalled in, the force of the main assumption is proportionately\nweakened. Thus we can easily explain natural order and design, if we are\nallowed to postulate a Divine Author who is absolutely perfect and\nall-powerful. But that hypothesis lies open to all the objections\nsuggested by defects and evils in nature, and can only be preserved\nthrough new hypotheses which modify the main assumption. Similarly the\nhypothesis of the human soul as an abiding and purely spiritual being,\nexisting in independence of the body, has to be modified to meet the\ndifficulties which arise from the phenomena of growth and decay. But the\nnew hypotheses, then constructed, derive their whole authority from the\nmain hypothesis which they are themselves defending.\n\nSuch is Kant\u2019s criticism of metaphysics when its teaching is based on\nthe facts of experience hypothetically interpreted. In regard to\ntranscendent metaphysics, there are, in Kant\u2019s view, only two\nalternatives.[1644] Either its propositions must be established\nindependently of all experience in purely _a priori_ fashion, and\ntherefore as absolutely certain; or they must consist in hypotheses\nempirically grounded. The first alternative has in the _Analytic_ and\n_Dialectic_ been shown to be impossible; the second alternative he\nrejects for the above reasons.\n\nBut this does not close Kant\u2019s treatment of metaphysical hypotheses. He\nproceeds to develop a doctrine which, in its fearless confidence in the\ntruth of Critical teaching, is the worthy outcome of his abiding belief\nin the value of a \u201csceptical method.\u201d[1645] As Reason is by its very\nnature dialectical, outside opponents are not those from whom we have\nmost to fear. Their objections are really derived from a source which\nlies in ourselves, and until these have been traced to their origin, and\ndestroyed from the root upwards, we can expect no lasting peace. Our\nduty, therefore, is to encourage our doubts, until by the very\nluxuriance of their growth they enable us to discover the hidden roots\nfrom which they derive their perennial vitality.\n\n     \u201cExternal tranquillity is a mere illusion. The germ of these\n     objections, which lies in the nature of human Reason, must be\n     rooted out. But how can we uproot it, unless we give it freedom,\n     nay, nourishment, to send out shoots so that it may discover itself\n     to our eyes, and that we may then destroy it together with its\n     root? Therefore think out objections which have never yet occurred\n     to any opponent; lend him, indeed, your weapons, or grant him the\n     most favourable position which he could possibly desire. You have\n     nothing to fear in all this, but much to hope for; you may gain for\n     yourselves a possession which can never again be contested.\u201d[1646]\n\nIn this campaign to eradicate doubt by following it out to its\nfurthermost limits, the hypotheses of pure Reason, \u201cleaden weapons\nthough they be, since they are not steeled by any law of experience,\u201d\nare an indispensable part of our equipment. For though hypotheses are\nuseless for the establishment of metaphysical propositions, they are,\nKant teaches, both admirable and valuable for their defence. That is to\nsay, their true metaphysical function is not dogmatic, but polemical.\nThey are weapons of war to which we may legitimately resort for the\nmaintenance of beliefs otherwise established. If, for instance, we have\nbeen led to postulate the immaterial, self-subsistent nature of the\nsoul, and are met by the difficulty that experience would seem to prove\nthat both the growth and the decay of our mental powers are due to the\nbody, we can weaken this objection by formulating the hypothesis that\nthe body is not the cause of our thinking, but only a restrictive\ncondition of it, peculiar to our present state, and that, though it\nfurthers our sensuous and animal faculties, it acts as an impediment to\nour spiritual life. Similarly, to meet the many objections against\nbelief in the eternal existence of a finite being whose birth depends\nupon contingencies of all kinds, such as the food supply, the whims of\ngovernment, or even vice, we can adduce the transcendental hypothesis\nthat life has neither beginning in birth nor ending in death, the entire\nworld of sense being but an image due to our present mode of knowledge,\nan image which like a dream has in itself no objective reality. Such\nhypotheses are not, indeed, even Ideas of Reason, but simply concepts\n_invented_ to show that the objections which are raised depend upon the\nfalse assumption that the possibilities have been exhausted, and that\nthe mere laws of nature comprehend the whole field of possible\nexistences. These hypotheses at least suffice to reveal the uncertain\ncharacter of the doubts which assail us in our practical beliefs.\n\n     \u201d[Transcendental hypotheses] are nothing but private opinions.\n     Nevertheless, we cannot properly dispense with them as weapons\n     against the misgivings which are apt to occur; they are necessary\n     even to secure our inner tranquillity. We must preserve to them\n     this character, carefully guarding against the assumption of their\n     independent authority or absolute validity, since otherwise they\n     would drown Reason in fictions and delusions.\u201d[1647]\n\nWe may now return to A 642-68 = B 670-96. The teaching of this section\nis extremely self-contradictory, wavering between a subjective and an\nobjective interpretation of the Ideas of Reason. The probable\nexplanation is that Kant is here recasting older material, and leaves\nstanding more of his earlier solutions than is consistent with his final\nconclusions. We can best approach the discussion by considering Kant\u2019s\nstatements in A 645 = B 673 and in A 650 ff. = B 678 ff. They expound,\nthough unfortunately in the briefest terms, a point of view which\nIdealism has since adopted as fundamental. Kant himself, very strangely,\nnever develops its consequences at any great length.[1648] The Idea,\nwhich Reason follows in the exercise of its sole true function, the\nsystematising of the knowledge supplied by the understanding, is that of\na unity in which the thought of the whole precedes the knowledge of its\nparts, and contains the conditions according to which the place of every\npart and its relation to the other parts are determined _a priori_. This\nIdea specialises itself in various forms, and in all of them directs the\nunderstanding to a knowledge that will be that of no mere aggregate but\nof a genuine system. Such concepts are not derived from nature; we\ninterrogate nature _according to them_, and consider our knowledge\ndefective so long as it fails to embody them. In A 650 = B 678 Kant\nfurther points out that this Idea of Reason does not merely direct the\nunderstanding to search for such unity, but also claims for itself\nobjective reality. And he adds,\n\n     \u201c...it is difficult to understand how there can be a logical\n     principle by which Reason prescribes the unity of rules, unless we\n     also presuppose a transcendental principle whereby such systematic\n     unity is _a priori_ assumed to be necessarily inherent in the\n     objects.\u201d\n\nFor how could we treat diversity in nature as only disguised unity, if\nwe were also free to regard that unity as contrary to the actual nature\nof the real?\n\n     \u201cReason would then run counter to its own vocation, proposing as\n     its aim an Idea quite inconsistent with the constitution of\n     nature.\u201d[1649]\n\nNor is our knowledge of the principle merely empirical, deduced from the\nunity which we find in contingent experience. On the contrary, there is\nan inherent and necessary law of Reason compelling us, antecedently to\nall specific experience, to look for such unity.\n\n     \u201c...without it we should have no Reason at all, and without Reason\n     no coherent employment of the understanding, and in the absence of\n     this no sufficient criterion of empirical truth. In order,\n     therefore, to secure an empirical criterion we are absolutely\n     compelled to presuppose the systematic unity of nature as\n     objectively valid and necessary.\u201d[1650] \u201cIt might be supposed that\n     this is merely an economical contrivance of Reason, seeking to save\n     itself all possible trouble, a hypothetical attempt, which, if it\n     succeeds, will, through the unity thus attained, impart probability\n     to the presumed principle of explanation. But such a selfish\n     purpose can very easily be distinguished from the Idea. For in the\n     latter we presuppose that this unity of Reason is in conformity\n     with nature itself; and that, although we are indeed unable to\n     determine the limits of this unity, Reason does not here beg but\n     command.\u201d[1651]\n\nThis last alternative, that Reason is here propounding a tentative\nhypothesis, in order by trial to discover how far it can be empirically\nverified--an alternative which Kant in the above passage rejects as\nunduly subjective, and as consequently failing to recognise the\nobjective claims and _a priori_ authority of the Ideas of Reason,--is\nyet a view which he himself adopts and indeed develops at considerable\nlength in this same section. This, as already stated, affords evidence\nof the composite character and varying origins of the material here\npresented.\n\nThe _Dissertation_ of 1770 gives a purely subjectivist interpretation of\nthe regulative principles, among which, from its pre-Critical\nstandpoint, it classes the principle of causality and the principle of\nthe conservation of matter.\n\n     \u201d[We adopt principles] which delude the intellect into mistaking\n     them for arguments derived from the object, whereas they are\n     commended to us only by the peculiar nature of the intellect, owing\n     to their convenience for its free and ample employment. They\n     therefore ... rest on _subjective_ grounds ... namely, on the\n     conditions under which it seems easy and expeditious for the\n     intellect to make use of its insight.... These rules of judging, to\n     which we freely submit and to which we adhere as if they were\n     axioms, solely for the reason that _were we to depart from them\n     almost no judgment regarding a given object would be permissible to\n     our intellect, I entitle principles of convenience_.... [One of\n     these is] the popularly received canon, _principia non esse\n     multiplicanda praeter summam necessitatem_, to which we yield our\n     adhesion, not because we have insight into causal unity in the\n     world either by reason or by experience, but because we seek it by\n     an impulse of the intellect, which seems to itself to have advanced\n     in the explanation of phenomena only in the degree in which it is\n     granted to it to descend from a single principle to the greatest\n     number of consequences.\u201d[1652]\n\nThis, in essentials, is the view which we find developed in A 646-9 = B\n674-8. Reason is the faculty of deducing the particular from the\ngeneral. When the general is admitted only as _problematical_, as a mere\nidea, while the particular is certain, we determine the universality of\nthe rule by applying it to the particulars, and then upon confirmation\nof its validity proceed to draw conclusions regarding cases not actually\ngiven. This Kant entitles the _hypothetical_ use of Reason. Reason must\nnever be employed constitutively. It serves only for the introduction,\nas far as may be found possible, of unity into the particulars of\nknowledge. It seeks to make the rule _approximate_ to\nuniversality.[1653] The unity which it demands\n\n     \u201c...is a _projected_ unity, to be regarded not as given in itself,\n     but as a problem only. This unity aids us in discovering a\n     principle for the manifold and special employment of the\n     understanding, drawing its attention to cases which are not given,\n     and thus rendering it more coherent.\u201d[1654]\n\nThe unity is merely logical, or rather methodological.[1655] To\npostulate, in consequence of its serviceableness, real unity in the\nobjects themselves would be to transform it into a transcendental\nprinciple of Reason, and to render\n\n     \u201c...the systematic unity necessary, not only subjectively and\n     logically, as method, but objectively also.\u201d[1656]\n\nThe above paragraphs are intercalated between A 645 = B 673 and A 650-63\n= B 678-91, in which, as we have already seen, the directly opposite\nview is propounded, namely, that such principles are _not_ merely\nhypothetical, _nor_ merely logical. In all cases they claim reality, and\nrest upon transcendental principles; they condition the very possibility\nof experience; and may therefore be asserted to be _a priori_ necessary\nand to be objectively valid. To quote two additional passages:\n\n     \u201c...we can conclude from the universal to the particular, only if\n     universal qualities are ascribed to things as the foundation upon\n     which the particular qualities rest.\u201d[1657] \u201cThe foundation of\n     these laws [cf. below, pp. 550-1] is not due to any secret design\n     of making an experiment by putting them forward as merely tentative\n     suggestions.... It is easily seen that they contemplate the\n     parsimony of fundamental causes, the manifoldness of effects, and\n     the consequent affinity of the parts of nature, as being in\n     themselves both rational and natural. Hence these principles carry\n     their recommendation directly in themselves, and not merely as\n     methodological devices.\u201d[1658]\n\nThus, in direct opposition to the preceding view of Reason\u2019s function as\nhypothetical, Kant is now prepared to maintain that the maxims of Reason\nare without meaning and without application save in so far as they can\nbe grounded in a transcendental principle.[1659]\n\nLet us follow Kant\u2019s detailed exposition of this last thesis. The\nlogical maxim, to seek for systematic unity, rests upon the\ntranscendental principle that the apparently infinite variety of nature\ndoes not exclude identity of species, that the various species are\nvarieties of a few genera, and these again of still higher genera. This\nis the scholastic maxim: _entia praeter necessitatem non esse\nmultiplicanda_. Upon this principle rests the possibility of concepts,\nand therefore of the understanding itself. It is balanced, however, by a\nsecond principle, no less necessary, the transcendental law of\nspecification, namely, that there must be manifoldness and diversity in\nthings, that every genus must specify itself in divergent species, and\nthese again in sub-species. Or as it is expressed in its scholastic\nform: _entium varietates non temere esse minuendas_. This principle is\nequally transcendental. It expresses a condition no less necessary for\nthe possibility of the understanding, and therefore of experience. As\nthe understanding knows all that it knows by concepts only, however far\nit may carry the division of genera, it can never know by means of pure\nintuition, but always again by lower concepts. If, therefore, there were\nno lower concepts, there could be no higher concepts;[1660] the gap\nexisting between individuals and genera could never be bridged; or\nrather, since neither individuals nor universals could then be\napprehended, neither would exist for the mind. As the higher concepts\nacquire all their content from the lower, they presuppose them for their\nown existence.\n\n     \u201cEvery concept may be regarded as a point which, in so far as it\n     represents the standpoint of a spectator, has its own horizon....\n     This horizon must be capable of containing an infinite number of\n     points, each of which again has its own narrower horizon; that is,\n     every species contains sub-species, according to the principle of\n     specification, and the logical horizon consists exclusively of\n     smaller horizons (sub-species), never of points which possess no\n     extent (individuals).\u201d[1661]\n\nCombining these two principles, that of _homogeneity_ and that of\n_specification_, we obtain a third, that of _continuity_. The logical\nlaw of the _continuum formarum logicarum_ presupposes the transcendental\nlaw, _lex continui in natura_. It provides that homogeneity be combined\nwith the greatest possible diversity by prescribing a continuous\ntransition from every species to every other, or in other words by\nrequiring that between any two species or sub-species, however closely\nrelated, intermediate species be always regarded as possible. (The\nparagraph at the end of A 661 = B 689, with its proviso that we cannot\nmake any definite empirical use of this law, is probably of later\norigin; it connects with the concluding parts of the section.) That this\nthird law is also _a priori_ and transcendental, is shown by the fact\nthat it is not derived from the prior discovery of system in nature, but\nhas itself given rise to the systematised character of our\nknowledge.[1662]\n\nThe psychological, chemical, and astronomical examples which Kant\nemploys to illustrate these laws call for no special comment. They were\ntaken from contemporary science, and in the advance of our knowledge\nhave become more confusing than helpful. The citation in A 646 = B 674\nof the concepts of \u201cpure earth, pure water, pure air\u201d as being \u201cconcepts\nof Reason\u201d is especially bewildering. They are, even in the use which\nKant himself ascribes to them, simply empirical hypotheses, formulated\nfor the purposes of purely physical explanation; they are in no genuine\nsense universal, regulative principles.\n\nIn passing to A 663-8 = B 691-6 we find still another variation in the\nsubstance of Kant\u2019s teaching. He returns, though with a greater maturity\nof statement, and with a very different and much more satisfactory\nterminology, to the more sceptical view of A 646-9 = B 674-7.[1663] The\ninterest of the above principles, Kant continues to maintain, lies in\ntheir transcendentality. Despite the fact that they are mere Ideas for\nthe guidance of understanding, and can only be approached\nasymptotically, they are synthetic _a priori_ judgments, and would seem\nto have an objective, though indeterminate, validity. So far his\nstatements are in line with the preceding paragraphs. But he proceeds to\nadd that this objective validity consists exclusively in their\nheuristic function. They differ fundamentally from the dynamical, no\nless than from the mathematical, principles of understanding, in that no\nschema of sensibility can be assigned to them. In other words, their\nobject can never be exhibited _in concreto_; it transcends all possible\nexperience. For this reason they are incapable of a transcendental\ndeduction.[1664] They are among the conditions indispensably necessary\nto the possibility, not of each and every experience, but only of\nexperience as systematised _in the interest of Reason_. In place of a\nschema they can possess only what may be called the _analogon_ of a\nschema, that is, they represent the Idea of a _maximum_, which the\nunderstanding in the subjective interest of Reason--or, otherwise\nexpressed,[1665] in the interest of a certain possible perfection of our\nknowledge of objects--is called upon to realise _as much as possible_.\nThus they are at once _subjective_ in the source from which they arise,\nand also _indeterminate_ as to the conditions under which, and the\nextent to which, they can obtain empirical embodiment. The fact that in\nthis capacity they represent a _maximum_, does not justify any assertion\neither as to the degree of unity which experience on detailed\ninvestigation will ultimately be found to verify, or as to the noumenal\nreality by which experience is conditioned.\n\nIn A 644-5 = B 672-3 Kant employs certain optical analogies to\nillustrate the illusion which the Ideas, in the absence of Critical\nteaching, inevitably generate. When the understanding is regulated by\nthe Idea of a _maximum_, and seeks to view all the lines of experience\nas converging upon and pointing to it, it necessarily regards it, _focus\nimaginarius_ though it be, as actually existing. The illusion, by which\nobjects are seen behind the surface of a mirror, is indispensably\nnecessary if we are to be able to see what lies behind our backs. The\ntranscendental illusion, which confers reality upon the Ideas of Reason,\nis similarly incidental to the attempt to view experience in its\ngreatest possible extension.\n\n\nON THE FINAL PURPOSE OF THE NATURAL DIALECTIC OF HUMAN REASON[1666]\n\nThis section is thoroughly unified and consistent in its teaching. Its\nrepetitious character is doubtless due to Kant\u2019s personal difficulty\neither in definitively accepting or in altogether rejecting the\nconstructive, Idealist interpretation of the function of Reason. He at\nleast succeeds in formulating a view which, while not asserting anything\nmore than is required in the scientific extension of experience,\nindicates the many possibilities which such experience fails to exclude.\nAs the Ideas of Reason are not merely empty thought-entities (_entia\nrationis ratiocinantis_[1667]), but have a certain kind of objective\nvalidity (_i.e._ are _entia rationis ratiocinatae_[1668]), they demand a\ntranscendental deduction.[1669] What this deduction is, and how it\ndiffers from that of the categories, we must now determine. Its\ndiscovery will, Kant claims, crown and complete our Critical labours.\n\nKant begins by drawing a distinction between representing _an object\nabsolutely_, and representing _an object in the Idea_.\n\n     \u201cIn the former case our concepts are employed to determine the\n     object, in the latter case there is in truth only a schema for\n     which no object, not even a hypothetical one, is directly given,\n     and which only enables us to represent to ourselves indirectly\n     other objects in their systematic unity, by means of their relation\n     to this Idea.\u201d[1670]\n\nAn Idea is only a schema (Kant in terms of A 655 = B 693 ought rather to\nhave said _analogon_ of a schema) whereby we represent to ourselves, as\nfor instance in the concept of a Highest Intelligence, not an objective\nreality but only such perfection of Reason as will tend to the greatest\npossible unity in the empirical employment of understanding.\n\nWith this introduction, Kant ushers in his famous \u201c_als ob_\u201d doctrine.\nWe must view the things of the world _as if_ they derived their\nexistence from a Highest Intelligence. That Idea is heuristic only, not\nexpository. Its purpose is not to enable us to comprehend such a Being,\nor even to think its existence, but only to show us how we should seek\nto determine the constitution and connection of the objects of\nexperience. The three transcendental Ideas do not determine an object\ncorresponding to them, but, _under the presupposition of such an object\nin the Idea_, lead us to systematic unity of _empirical_ knowledge. When\nthey are thus strictly interpreted as merely regulative of empirical\nenquiry, they will always endorse experience and never run counter to\nit. Reason, which seeks completeness of explanation, must therefore\nalways act in accordance with them. Only thereby can experience acquire\nits fullest possible extension. This is the transcendental deduction of\nwhich we are in search. It establishes the indispensableness of the\nIdeas of Reason for the completion of experience, and their legitimacy\nas regulative principles.\n\nWe may here interrupt Kant\u2019s exposition so far as to point out that\nthis argument does not do justice to the full force of his position. The\ntrue Critical contention--and only if we interpret the passage in the\nlight of this contention can the proof be regarded as transcendental in\nthe strict sense--is that the Ideas are necessary to the possibility of\neach and every experience, involved together with the categories as\nconditions of the very existence of consciousness. They are not merely\nregulative, but are regulative of an experience which they also help to\nmake possible.[1671] They express the standards in whose light we\ncondemn all knowledge which does not fulfil them; and we have\nconsequently no option save to endeavour to conform to their demands. In\nother words, they are not derivative concepts obtained by merely\nomitting the restrictions essential to our empirical consciousness, but\nrepresent a presupposition necessarily involved in all consciousness.\nSome such restatement of the argument is demanded by the position which\nKant has himself outlined in A 645 = B 673 and in A 650 ff. = B 678 ff.\nUnfortunately he does not return to it. The more sceptical view which he\nhas meantime been developing remains dominant. The deduction is left in\nthis semi-Critical form.\n\nA 672-6 = B 700-4 give a fuller statement of the \u201c_als ob_\u201d doctrine. In\npsychology we must proceed _as if_ the mind were a simple substance\nendowed with personal identity[1672] (in this life at least), not in\norder to derive explanation of its changing states from the soul so\nconceived, but to derive them from each other in accordance with the\nIdea. In cosmology and theology (we may observe the straits to which\nKant is reduced in his attempt to distinguish them) we ought to consider\nall phenomena both in their series and in their totality _as if_ they\nwere due to a highest and all-sufficient unitary ground. In so doing we\nshall not derive the order and system in the world from the object of\nthe Idea, but only extract from the Idea the rule whereby the\nunderstanding attains the greatest possible satisfaction in the\nconnecting of natural causes and effects.\n\nIn A 676-7 = B 704-5 Kant resorts to still another distinction--between\n_suppositio relativa_ and _suppositio absoluta_. This distinction is\nsuggested by the semi-objectivity of principles that are merely\nregulative. Though we have to recognise them as necessary, such\nnecessity does not justify the assertion of their independent validity.\nWhen we admit a supreme ground as the source of the order and system\nwhich the principles demand, we do so only in order to think the\nuniversality of the principles with greater definiteness. Such\nsupposition is relative to the needs of Reason in its _empirical_\nemployment: not absolute, as pointing to the existence of such a being\nin itself.\n\n     \u201cThis explains why, in relation to what is given to the senses as\n     existing, we require the Idea of a primordial Being necessary in\n     itself, and yet can never form the slightest concept of it or of\n     its absolute necessity.\u201d[1673]\n\nThis last statement leads to the further problem to which Kant here\ngives his final solution, how if, as has been shown in the _Dialectic_,\nthe concepts of absolute necessity and of unconditionedness are without\nmeaning, the Ideas of Reason can be entertained at all, even mentally.\nWhat is their actual content and how is it possible to conceive them?\nKant\u2019s reply is developed in terms of the semi-Critical subjectivist\npoint of view which dominates this section. The Ideas are mere Ideas.\nThey yield not the slightest concept either of the internal possibility\nor of the necessity of any object corresponding to them. They only seem\nto do so, owing to a transcendental illusion. On examination we find\nthat the concepts which we employ in thinking them as independently\nreal, are one and all derived from experience. That is to say, we judge\nof them after the analogy of reality, substance, causality, and\nnecessity in the sensible world.[1674]\n\n     \u201d[They are consequently] _analoga_ only of real things, not real\n     things in themselves. We remove from the object of the Idea the\n     conditions which limit the concept of the understanding, but which\n     at the same time alone make it possible for us to have a\n     determinate concept of anything. What we then think is, therefore,\n     a something of which, as it is in itself, we have no concept\n     whatsoever, but which we none the less represent to ourselves as\n     standing in a relation to the sum-total of appearances analogous to\n     that in which appearances stand to one another.\u201d[1675]\n\nThey do not carry our knowledge beyond the objects of possible\nexperience, but only extend the empirical unity of experience. They are\nthe schemata of regulative principles. In them Reason is concerned with\nnothing but its own inherent demands; and as their unity is the unity of\na system which is to be sought only in experience,[1676] qualities\nderived from the sensible world can quite legitimately be employed in\ntheir specific determination. They are not inherently dialectical; their\ndemands have the rationality which we have a right to expect in the\nIdeals of Reason. When Critically examined, they propound no problem\nwhich Reason is not in itself entirely competent to solve.[1677] It is\nto their misemployment that transcendental illusion is due. In the form\nin which they arise from the natural disposition of our Reason they are\ngood and serviceable.[1678]\n\nTo the question what is the most adequate form in which the regulative\nschema can be represented,[1679] Kant gives an answer which shows how\nvery far he is from regarding the Leibnizian _Ens realissimum_ as the\ntrue expression of the Ideal of Reason. It is through the employment of\nteleological concepts that we can best attain the highest possible form\nof systematic unity.\n\n     \u201cThe highest formal unity ... is the _purposive_ unity of things.\n     The _speculative_ [_i.e._ theoretical] interest of Reason makes it\n     necessary to regard all order in the world _as if_ it had\n     originated in the purpose of a Supreme Reason. Such a principle\n     opens out to our Reason, as applied in the field of experience,\n     altogether new views as to how the things of the world may be\n     connected according to teleological laws, and so enables it to\n     arrive at their greatest systematic unity. The assumption of a\n     Supreme Intelligence, as the one and only cause of the universe,\n     though in the Idea alone, can therefore always benefit Reason and\n     can never injure it.\u201d[1680]\n\nFor so long as this assumption is employed only as a regulative\nprinciple, even error cannot be really harmful. The worst that can\nhappen is that where we expected a teleological connection, a merely\nmechanical or physical one is met with. If, on the other hand, we leave\nthe solid ground of experience, and use the assumption to explain what\nwe are unable to account for in empirical terms, we sacrifice all real\ninsight, and confound Reason by transforming a concept, which is\nanthropomorphically determined for the purposes of empirical\norientation, into a means of explaining order as non-natural and as\nimposed from without on the material basis of things.\n\nThis is a point of sufficient importance to call for more detailed\nstatement. Hume in his _Dialogues_ points out that the main defect in\nthe teleological proof of God\u2019s existence is its assumption that order\nand design are foreign to the inherent constitution of things, and must\nbe of non-natural origin. The argument is therefore weakened by every\nadvance in the natural sciences. It also runs directly counter to the\nvery phenomena, those of animal life, upon which it is chiefly based,\nsince the main characteristic of the organic in its distinction from the\ninorganic is its inner wealth of productive and reproductive powers.\nWith these criticisms Kant is in entire agreement. From them, in the\npassage before us, he derives an argument in support of a strictly\nregulative interpretation of his \u201c_als ob_\u201d doctrine. The avowed\nintention of the teleological argument is to prove _from nature_ the\nexistence of an intelligent supreme cause. If therefore its standpoint\nbe held to with more consistency than its own defenders have hitherto\nshown, it will be found to rest upon the regulative principle, that we\nmust study nature as if an _inherent_ order were _native_ to it, and so\nseek to approach by degrees, in proportion as such _natural unity_ is\nempirically discovered, the absolute perfection which inspires our\nresearches. But if we transform our Ideal into an instrument of\nexplanation, beginning with what ought properly to be only our goal, we\ndelude ourselves with the belief that what can only be acquired through\nthe slow and tentative labours of empirical enquiry is already in our\npossession.\n\n     \u201cIf I begin with a supreme purposive Being as the ground of all\n     things, the unity of nature is really surrendered, as being quite\n     foreign and accidental to the nature of things, and as not to be\n     known from its own general laws. There thus arises a vicious\n     circle: we are assuming just that very point which is mainly in\n     dispute.\u201d[1681]\n\nSuch a method of argument is self-destructive, since if we do not find\norder and perfection in the nature of things, and _therefore in their\ngeneral and necessary laws_, we are not in a position to infer such a\nBeing as the source of all causality.\n\nTo the question whether we may not interpret natural order, once it has\nbeen discovered by empirical investigation, as due to the divine will,\nKant replies that such procedure is allowable only on the condition that\nit is the same to us whether we say that God has wisely willed it or\nthat nature has wisely arranged it. We may admit the Idea of a Supreme\nBeing only in so far as it is required by Reason as the regulative\nprinciple of all investigation of _nature_;\n\n     \u201c...and we cannot, therefore, without contradicting ourselves,\n     ignore the general laws of nature in view of which the Idea was\n     adopted, and look upon the purposiveness of nature as contingent\n     and hyper-physical in its origin. For we were not justified in\n     assuming above nature a Being of those qualities, but only in\n     adopting the Idea of it in order to be able to view the\n     appearances, according to the analogy of a causal determination, as\n     systematically connected with one another.\u201d[1682] \u201cThus pure\n     Reason, which at first seemed to promise nothing less than the\n     extension of knowledge beyond all limits of experience, contains,\n     if properly understood, nothing but regulative\n     principles....\u201d[1683]\n\n\nCONCLUDING COMMENT ON THE DIALECTIC\n\nI may now summarise Kant\u2019s answer to the three main questions of the\n_Dialectic_: (1) Whether, or in what degree, the so-called Ideas of\nReason are concepts due to a faculty altogether distinct from the\nunderstanding, and how far, as thus originating in pure Reason, they\nallow of definition; (2) how far they are capable of a transcendental\ndeduction; (3) what kind of objective validity this deduction proves\nthem to possess.\n\nThese questions are closely interconnected; the solution of any one\ndetermines the kind of solution to be given to all three. Kant, as we\nhave found, develops his final position through a series of very subtle\ndistinctions by which he contrives to justify and retain, though in a\nhighly modified form, the more crudely stated divisions between Ideas\nand categories, between Reason and understanding, upon which the initial\nargument of the _Dialectic_ is based.\n\nThe answer amounts in essentials to the conclusion that understanding,\nin directing itself by means of Ideals, exercises a function so distinct\nfrom that whereby it conditions concrete and specific experience, that\nit may well receive a separate title; that the Ideas in terms of which\nit constructs these Ideals, though schematic (_i.e._ sensuous and\nempirical in content), are not themselves empirical, and so far from\nbeing merely extended concepts of understanding, express transcendental\nconditions upon which all use of the understanding rests.\n\nNow if this position is to be justified, Kant ought to show that the\nfundamental Idea of Reason, that of the unconditioned, is altogether\ndistinct from any concept of the understanding, and in particular that\nit must not be identified with the category of totality, nor be viewed\nas being merely the concept of conditioned existence with its various\nempirical limitations thought away. Needless to say, Kant does not\nfulfil these requirements in any consistent manner. The _Critique_\ncontains the material for a variety of different solutions; it does not\ndefinitively commit itself to any one of them.\n\nIf the argument of A 650 ff. = B 678 ff. were developed we should be in\npossession of what may be called the Idealist solution. It would proceed\nsomewhat as follows. Consciousness as such is always the awareness of a\nwhole which precedes and conditions its parts. Such consciousness cannot\nbe accounted for on the assumption that we are first conscious of the\nconditioned, and then proceed to remove limitations and to form for\nourselves, by means of the more positive factors involved in this\nantecedent consciousness, an Idea of the totality within which the given\nfalls. The Idea of the unconditioned, distinct from all concepts of\nunderstanding, is one of the _a priori_ conditions of possible\nexperience, and is capable of a transcendental deduction of equal\nvalidity with, and of the same general nature as, that of the\ncategories. It is presupposed in the possibility of our contingently\ngiven experience.\n\nAs this Idea conditions all subordinate concepts, it cannot be defined\nin terms of them. That does not, however, deprive it of all meaning; its\nsignificance is of a unique kind; it finds expression in those Ideals\nwhich, while guiding the mind in the construction of experience, also\nserve as the criteria through which experience is condemned as only\nphenomenal.\n\nBut this, as we have found, is not a line of argument which Kant has\ndeveloped in any detail. The passages which point to it occur chiefly in\nthe introductory portions of the _Dialectic_; in its later sections they\nare both brief and scanty. When he sets himself, as in the chapter on\nthe _Ideal of Pure Reason_ and in the subsequent _Appendix_, to define\nhis conclusions, it is a much more empirical, and indeed sceptical, line\nthat he almost invariably follows. There are, he then declares, strictly\nno pure, _a priori_ Ideas. The supposed Ideas of unconditionedness and\nof absolute necessity are discovered on examination to be without the\nleast significance for the mind. The Ideas, properly defined, are merely\nschemata of regulative principles, and their whole content reduces\nwithout remainder to such categories as totality, substance, causality,\nnecessity, transcendently applied. As _Ideas_, they are then without\nreal meaning; but they can be employed by analogy to define an Ideal\nwhich serves an indispensable function in the extension of experience.\nFrom this point of view, the transcendental deduction of the Ideas is\nradically distinct from that of the categories. The proof is not that\nthey are necessary for the possibility of experience, but only that they\nare required for its perfect, or at least more complete, development.\nAnd as Kant is unable to prove that such completion is really possible,\nthe objective validity of the Ideas is left open to question. They\nshould be taken only as heuristic principles; the extent of their truth,\neven in the empirical realm, cannot be determined by the _a priori_\nmethod that is alone proper to a _Critique of Pure Reason_.\n\nThe first view is inspired by the fundamental teaching of the\n_Analytic_, and is the only view which will justify Kant in retaining\nhis distinction between appearance and things in themselves. All that is\npositive in the second view can be combined with the first view; but, on\nthe other hand, the negative implications of the second view are at\nvariance with its own positive teaching. For when the Ideas are regarded\nas empirical in origin no less than in function, their entire authority\nis derived from experience, and cannot be regarded as being\ntranscendental in any valid sense of that term. In alternating between\nthese two interpretations of the function of Reason, Kant is wavering\nbetween the Idealist and the merely sceptical view of the scope and\npowers of pure thought. On the Idealist interpretation Reason is a\nmetaphysical faculty, revealing to us the phenomenal character of\nexperience, and outlining possibilities such as may perhaps be\nestablished on moral grounds. From the sceptical standpoint, on the\nother hand, Reason gives expression to what may be only our subjective\npreference for unity and system in the ordering of experience. According\nto the one, the criteria of truth and reality are bound up with the\nIdeas; according to the other, sense-experience is the standard by which\nthe validity even of the Ideas must ultimately be judged. From the fact\nthat Kant should have continued sympathetically to develop two such\nopposite standpoints, we would seem to be justified in concluding that\nhe discerned, or at least desiderated, some more complete reconciliation\nof their teaching than he has himself thus far been able to achieve, and\nthat no solution which would either subordinate the Ideal demands of\nthought, or ignore the gifts of experience, could ever have been\ndefinitively accepted by him as satisfactorily meeting the issues at\nstake. The Idealist solution is that to which his teaching as a whole\nmost decisively points; but he is as conscious of the difficulties which\nlie in its path as he is personally convinced of its ultimate truth. His\ncontinuing appreciation of the value of sceptical teaching is a tacit\nadmission that the Idealist doctrines, in the form which he has so far\nbeen able to give to them, are not really adequate to the complexity of\nthe problems. As further confirmation of the tentative character of\nKant\u2019s conclusions in the _Critique of Pure Reason_, we have his own\nlater writings. In the _Critique of Judgment_, published nine years\nlater, in teaching less sceptical and more constructive, though still\ndelicately balanced between the competing possibilities, and always,\ntherefore, leaving the final decision to moral considerations, Kant\nventures upon a restatement of the problems of the _Dialectic_. To this\nrestatement both of the above tendencies contribute valuable elements.\n\n\n\n\nAPPENDIX A[1684]\n\nTRANSCENDENTAL DOCTRINE OF METHODS\n\n\n\n\nCHAPTER I\n\nTHE DISCIPLINE[1685] OF PURE REASON\n\n\nKant is neither an intellectualist nor an anti-intellectualist. Reason,\nthe proper duty of which is to prescribe a discipline to all other\nendeavours, itself requires discipline; and when it is employed in the\nmetaphysical sphere, independently of experience, it demands not merely\nthe correction of single errors, but the eradication of their causes\nthrough \u201ca separate negative code,\u201d such as a Critical philosophy can\nalone supply. In the _Transcendental Doctrine of Elements_ this demand\nhas been met as regards the _materials_ or _contents_ of the Critical\nsystem; we are now concerned only with its _methods_ or _formal\nconditions_.[1686]\n\nThis distinction is highly artificial. As already indicated, it is\ndetermined by the requirements of Kant\u2019s architectonic. The entire\nteaching of the _Methodology_ has already been more or less exhaustively\nexpounded in the earlier divisions of the _Critique_.\n\n\nSECTION I\n\nTHE DISCIPLINE OF PURE REASON IN ITS DOGMATIC EMPLOYMENT\n\nIn dealing with the distinction between mathematical and philosophical\nknowledge, Kant is here returning to one of the main points of his\n_Introduction_ to the _Critique_.[1687] His most exhaustive treatment of\nit is, however, to be found in a treatise which he wrote as early as\n1764, his _Enquiry into the Clearness of the Principles of Natural\nTheology and Morals_. The continued influence of the teaching of that\nearly work is obvious throughout this section, and largely accounts for\nthe form in which certain of its tenets are propounded.\n\n     \u201c...one can say with Bishop Warburton that nothing has been more\n     injurious to philosophy than mathematics, that is, than the\n     imitation of its method in a sphere where it is impossible of\n     application....\u201d[1688]\n\nSo far from being identical in general nature, mathematics and\nphilosophy are, Kant declares, fundamentally opposed in all essential\nfeatures. For it is in their methods, and not merely in their\nsubject-matter, that the essential difference between them is to be\nfound.[1689] Philosophical knowledge can be acquired only through\n_concepts_, mathematical knowledge is gained through the _construction_\nof concepts.[1690] The one is discursive merely; the other is intuitive.\nPhilosophy can consider the particular only in the general; mathematics\nstudies the general in the particular.[1691] Philosophical concepts,\nsuch as those of substance and causality, are, indeed, capable of\napplication in transcendental synthesis, but in this employment they\nyield only empirical knowledge of the sensuously given; and from\nempirical concepts the universal and necessary judgments required for\nthe possibility of metaphysical science can never be obtained.\n\nThe exactness of mathematics depends on definitions, axioms, and\ndemonstrations, none of which are obtainable in philosophy. To take each\nin order.\n\nI. =Definitions.=--To define in the manner prescribed by mathematics is to\nrepresent the _complete_ concept of a thing. This is never possible in\nregard to empirical concepts. We are more certain of their denotation\nthan of their connotation; and though they may be _explained_, they\ncannot be defined. Since new observations add or remove predicates, an\nempirical concept is always liable to modification.\n\n     \u201cWhat useful purpose could be served by defining an empirical\n     concept, such, for instance, as that of water? When we speak of\n     water and its properties, we do not stop short at what is thought\n     in the word water, but proceed to experiments. The word, with the\n     few marks which are attached to it, is more properly to be regarded\n     as merely a designation than as a conception. The so-called\n     definition is nothing more than a determining of the word.\u201d[1692]\n\nExact definition is equally impossible in regard to _a priori_ forms,\nsuch as time or causality. Since they are not framed by the mind, but\nare _given_ to it, the completeness of our analysis of them can never be\nguaranteed. Though they are known, they are known only as problems.\n\n     \u201cAs Augustine has said, \u2018I know well what time is, but if any one\n     asks me, I cannot tell.\u2019\u201d[1693]\n\nMathematical definitions _make_ concepts; philosophical definitions only\nexplain them.[1694] Philosophy cannot, therefore, imitate mathematics by\nbeginning with definitions. In philosophy the incomplete exposition must\nprecede the complete; definitions are the final outcome of our enquiry,\nand not as in mathematics the only possible beginning of its proofs.\nIndeed, the mathematical concept may be said to be given by the very\nprocess in which it is constructively defined; and, as thus originating\nin the process of definition, it can never be erroneous.[1695]\nPhilosophy, on the other hand, swarms with faulty definitions, which are\nnone the less serviceable.\n\n     \u201cIn mathematics definition belongs _ad esse_, in philosophy _ad\n     melius esse_. It is desirable to attain it, but often very\n     difficult. Jurists are still without a definition of their concept\n     of Right.\u201d[1696]\n\nII. =Axioms.=--This paragraph is extremely misleading as a statement of\nKant\u2019s view regarding the nature of geometrical axioms. In stating that\nthey are self-evident,[1697] he does not really mean to assert what that\nphrase usually involves, namely, absolute _a priori_ validity. For Kant\nthe geometrical axioms are merely descriptions of certain _de facto_\nproperties of the given intuition of space. They have the merely\nhypothetical validity of all propositions that refer to the\ncontingently given. For even as a pure intuition, space belongs to the\nrealm of the merely factual.[1698] This un-Critical opposition of the\nself-evidence of geometrical axioms to the synthetic character of such\n\u201cphilosophical\u201d truths as the principle of causality is bound up with\nKant\u2019s unreasoned conviction that space in order to be space at all,\nmust be Euclidean.[1699] Kant\u2019s reference in this paragraph to the\npropositions of arithmetic is equally open to criticism. For though he\nis more consistent in recognising their synthetic character, he still\nspeaks as if they could be described as self-evident, _i.e._ as\nimmediately certain. The cause of this inconsistency is, of course, to\nbe found in his intuitional theory of mathematical science. Mathematical\npropositions are obtained through intuition; those of philosophy call\nfor an elaborate and difficult process of transcendental deduction. When\nmodern mathematical theory rejects this intuitional view, it is really\nextending to mathematical concepts Kant\u2019s own interpretation of the\nfunction of the categories. Concepts condition the _possibility_ of\nintuitional experience, and find in this conditioning power the ground\nof their objective validity.[1700] Here, as in the _Aesthetic_,[1701]\nKant fails adequately to distinguish between the problems of pure and\napplied mathematics.\n\n=III. Demonstrations.=--Kant again introduces his very unsatisfactory\ndoctrine of the construction of concepts:[1702] and he even goes so far\nas to maintain, in complete violation of his own doctrine of\ntranscendental deduction, that where there is no intuition, there can be\nno demonstration. Apodictic propositions, he declares, are either\n_dogmata_ or _mathemata_; and the former are beyond the competence of\nthe human mind. But no sooner has he made these statements than he\nvirtually withdraws them by adding that, though apodictic propositions\ncannot be established directly from concepts, they can be indirectly\nproved by reference to something purely contingent, namely, possible\nexperience. Thus the principle of causality can be apodictically proved\nas a condition of possible experience. Though it may not be called a\n_dogma_, it can be entitled a _principle_! In explanation of this\ndistinction, which betrays a lingering regard for the self-evident\nmaxims of rationalistic teaching, Kant adds that the principle of\ncausality, though a principle, has itself to be proved.\n\n     \u201c...it has the peculiarity that it first makes possible its own\n     ground of proof, namely, experience....\u201d[1703]\n\nThis, as we have noted,[1704] is exactly what mathematical axioms must\nalso be able to do, if they are to establish their objective validity.\n\n\nSECTION II\n\nTHE DISCIPLINE OF PURE REASON IN ITS POLEMICAL EMPLOYMENT\n\nThis section contains an admirable defence of the value of scepticism.\n\n     \u201cEven poisons have their use. They serve to counteract other\n     poisons generated in our system, and must have a place in every\n     complete pharmacopeia. The objections against the persuasions and\n     complacency of our purely speculative Reason arise from the very\n     nature of Reason itself, and must therefore have their own good use\n     and purpose, which ought not to be disdained. Why has Providence\n     placed many things which are closely bound up with our highest\n     interests so far beyond our reach, that we are only permitted to\n     apprehend them in a manner lacking in clearness and subject to\n     doubt, in such fashion that our enquiring gaze is more excited than\n     satisfied? It is at least doubtful whether it serves any useful\n     purpose, and whether it is not, indeed, perhaps even harmful to\n     venture upon bold interpretations of such uncertain appearances.\n     But there can be no manner of doubt that it is always best to grant\n     Reason complete liberty, both of enquiry and of criticism, so that\n     it may be without hindrance in attending to its own proper\n     interests. These interests are no less furthered by the limitation\n     than by the extension of its speculations; and they will always\n     suffer when outside influences intervene to divert it from its\n     natural path, and to constrain it by what is irrelevant to its own\n     proper ends.\u201d[1705] \u201cWhenever I hear that a writer of real ability\n     has demonstrated away the freedom of the human will, the hope of a\n     future life, and the existence of God, I am eager to read the book,\n     for I expect him by his talents to increase my insight into these\n     matters.\u201d[1706]\n\n\nSECTION IV[1707]\n\nTHE DISCIPLINE OF PURE REASON IN REGARD TO ITS PROOFS[1708]\n\nThis section merely restates the general nature and requirements of\ntranscendental proof. The exposition is much less satisfactory than that\nalready given in the _Analytic_ and _Dialectic_. The only really new\nfactor is the distinction between apagogical and direct proof. The\nformer may produce conviction, but cannot enable us to comprehend the\ngrounds of the truth of our conviction. Also, outside mathematics, it is\nextremely dangerous to attempt to establish a thesis by showing its\ncontradictory to be impossible.[1709] This is especially true in the\nsphere of our Critical enquiries, since the chief danger to be guarded\nagainst is the confounding of the subjectively necessary with the\nindependently real. In this field of investigation it is never\npermissible to attempt to justify a synthetic proposition by refuting\nits opposite. Such seeming proofs can easily be secured, and have been\nthe favourite weapons of dogmatic thinkers.\n\n     \u201cEach must defend his position directly, by a legitimate proof that\n     carries with it transcendental deduction of the grounds upon which\n     it is itself made to rest. Only when this has been done, are we in\n     a position to decide how far its claims allow of rational\n     justification. If an opponent relies on subjective grounds, it is\n     an easy matter to refute him. The dogmatist cannot, however, profit\n     by this advantage. His own judgments are, as a rule, no less\n     dependent upon subjective influences; and he can himself in turn be\n     similarly cornered. But if both parties proceed by the direct\n     method, _either_ they will soon discover the difficulty, nay, the\n     impossibility, of showing reason for their assertions, and will be\n     left with no resort save to appeal to some form of prescriptive\n     authority; _or_ the _Critique_ will the more easily discover the\n     illusion to which their dogmatic procedure is due; and pure Reason\n     will be compelled to relinquish its exaggerated pretensions in the\n     realm of speculation, and to withdraw within the limits of its\n     proper territory--that of practical principles.\u201d[1710]\n\n\n\n\nCHAPTER II\n\nTHE CANON[1711] OF PURE REASON\n\n\nSECTION I\n\nTHE ULTIMATE END OF THE PURE USE OF OUR REASON[1712]\n\nThe problems of the existence of God, the freedom of the will, and the\nimmortality of the soul have, Kant declares, little _theoretical_\ninterest. For, as he has already argued, even if we were justified in\npostulating God, freedom, and immortality, they would not enable us to\naccount for the phenomena of sense-experience, the only objects of\npossible knowledge. But the three problems are also connected with our\n_practical_ interests, and in that reference they constitute the chief\nsubject of metaphysical enquiry.[1713] The practical is whatever is\npossible through freedom; and the decision as to _what we ought to do_\nis the supreme interest of pure Reason in its highest employment.\n\n     \u201c...the ultimate intention of Nature in her wise provision for us\n     has indeed, in the constitution of our Reason, been directed to our\n     moral interests alone.\u201d[1714]\n\nThis is the position which Kant endeavours to establish in his\n_Foundations of the Metaphysics of Morals_, and in the _Critique of\nPractical Reason_. The very brief outline which he here gives of his\nargument is necessarily incomplete; and is in consequence somewhat\nmisleading. He first disposes of the problem of freedom; and does so in\na manner which shows that he had not, when this section was composed,\ndeveloped his Critical views on the nature of moral freedom. He is for\nthe present content to draw a quite un-Critical distinction between\ntranscendental and practical freedom.[1715] The latter belongs to the\nwill in so far as it is determined by Reason alone, independently of\nsensuous impulses. Reason prescribes objective _laws of freedom_, and\nthe will under the influence of these laws overcomes the affections of\nsense. Such practical freedom can, Kant asserts, be proved by experience\nto be a natural cause. Transcendental freedom,[1716] on the other hand,\n_i.e._ the power of making a new beginning in the series of phenomena,\nis a problem which can never be empirically solved. It is a purely\nspeculative question with which Reason in its practical employment is\nnot in the least concerned. The canon of pure Reason has therefore to\ndeal only with the two remaining problems, God and immortality. Comment\nupon these assertions can best be made in connection with the argument\nof the next section.[1717]\n\n\nSECTION II\n\nTHE IDEAL OF THE HIGHEST GOOD, AS A DETERMINING GROUND OF THE ULTIMATE\nEND OF PURE REASON[1718]\n\nReason in its speculative employment transcends experience, but solely\nfor the sake of experience. In other words, speculative Reason has a\npurely empirical function. (This is the explanation of the somewhat\nparadoxical contention, to which Kant has already committed himself,\nthat the problems of God and immortality, _though seemingly speculative\nin character_, really originate in our practical interests.) But pure\nReason has also a practical use; and it is in this latter employment\nthat it first discloses the genuinely metaphysical character of its\npresent constitution and ultimate aims. The moral consciousness, in\nrevealing to us an Ideal of absolute value, places in our hands the only\navailable key to the mysteries of existence. As this moral consciousness\nrepresents the deepest reality of human life, it may be expected to have\ngreater metaphysical significance than anything else in human\nexperience; and since the ends which it reveals also present themselves\nas _absolute_ in value, and are indeed the only absolute values of which\nwe can form any conception, this conclusion would seem to be confirmed.\n\nHappiness has natural value; morality, _i.e._ the being worthy to be\nhappy, has absolute value. The means of attaining the former obtain\nexpression in prudential or pragmatic laws that are empirically\ngrounded. The conditions of the latter are embodied in a categorical\nimperative of an _a priori_ character. The former _advise_ us how best\nto satisfy our natural desire for happiness; the latter _dictates_ to us\nhow we must behave in order to deserve happiness.\n\nKant\u2019s further argument is too condensed to be really clear, and if\nadequately discussed would carry us quite beyond the legitimate limits\nof this _Commentary_. I shall therefore confine myself to a brief and\nfree restatement of his general position. The Critical teaching can be\ndescribed as resulting in a new interpretation of the function of\nphilosophy.[1719] The task of the philosopher, properly viewed, does not\nconsist in the solution of _speculative_ problems; such problems\ntranscend our human powers. All that philosophy can reasonably attempt\nis to analyse and define the situations, cognitive and practical, in\nwhich, owing to the specific conditions of human existence, we find\nourselves to be placed. Upon analysis of the cognitive situation Kant\ndiscovers that while all possibilities are open, the theoretical data\nare never such as to justify ontological assertions.[1720] When,\nhowever, he passes to the practical situation, wider horizons,\ndefinitely outlined, at once present themselves. The moral consciousness\nis the key to the meaning of the entire universe as well as of human\nlife. Its values are the sole ultimate values, and enable us to\ninterpret in _moral_ terms (even though we cannot comprehend in any\ngenuinely _theoretical_ fashion) the meaning of the dispensation under\nwhich we live. The moral consciousness, like sense-experience, discloses\nupon examination a systematic unity of presupposed conditions. In the\ntheoretical sphere this unity cannot be proved to be more than a\npostulated Ideal of _empirical_ experience; and it is an Ideal which,\neven if granted to have absolute validity, is too indefinite to enable\nus to assert that ultimate reality is spiritual in character, or is\nteleologically ordered. The underlying conditions, on the other hand, of\npractical experience have from the start a purely noumenal reference.\nThey have no other function than to define, in terms of the moral\nconsciousness, the ultimate meaning of reality as a whole. They\npostulate[1721] a universe in which the values of spiritual experience\nare supported and conserved.\n\nBut the main difference in Kant\u2019s treatment of the two situations,\ncognitive and practical, only emerges into view when we recognise the\ndiffering modes in which the transcendental method of proof is applied\nin the two cases. The _a priori_ forms of sensibility, understanding,\nand Reason are proved by reference to possible experience, as being its\nindispensable conditions. In moral matters, however, we must not appeal\nto experience. The actual is no test of the Ideal; \u201cwhat is\u201d is no test\nof what ought to be. And secondly, the moral law, if valid at all, must\napply not merely within the limits of experience, but with absolute\nuniversality to all rational beings. The moral law, therefore, can\nneither be given us in experience, nor be proved as one of the\nconditions necessary to its possibility. Its validity, in other words,\ncan be established neither through experience nor through theoretical\nreason.\n\nThough such is Kant\u2019s own method of formulating the issue, it\nexaggerates the difference of his procedure in the two _Critiques_, and\nis very misleading as a statement of his real position. In one passage,\nin the _Critique of Practical Reason_,[1722] Kant does, indeed, assert\nthat the moral law requires no deduction. It is, he claims, a _fact_ of\nwhich we are _a priori_ conscious: so far from itself requiring proof,\nit enables us to prove the reality of freedom. Yet in the very same\nsection he argues that the deduction of freedom from the moral law is a\ncredential of the latter, and is a sufficient substitute for all _a\npriori_ justification. According to the first statement we have an\nimmediate consciousness of the validity of the moral law; according to\nthe second statement the moral law proves itself indirectly, by serving\nas a principle for the deduction of freedom. The second form of\nstatement alone harmonises with the argument developed in the third\nsection of the _Foundations of the Metaphysics of Morals_, and more\ncorrectly expresses the intention of Kant\u2019s central argument in the\n_Critique of Practical Reason_. For the difference between the two\ntranscendental proofs in the two _Critiques_ does not really consist in\nany diversity of method, but solely in the differing character of the\npremisses from which each starts. The ambiguity of Kant\u2019s argument in\nthe second _Critique_ seems chiefly to be caused by his failure clearly\nto recognise that the moral law, though a form of pure Reason,\nexercises, in the process of its transcendental proof, a function which\nexactly corresponds to that which is discharged by possible experience\nin the first _Critique_. Our consciousness of the moral law is, like\nsense-experience, a given fact. It is _de facto_, and cannot be deduced\nfrom anything more ultimate than itself.[1723] But as given, it enables\nus to deduce its transcendental conditions. This does not mean that our\nimmediate consciousness of it _as given_ guarantees its validity. The\nnature of its validity is established only in the process whereby it\nreveals its necessary implications. The objects of sense-experience are\nassumed by ordinary consciousness to be absolutely real; in the process\nof establishing the transcendental conditions of such experience they\nare discovered to be merely phenomenal. The pure principles of\nunderstanding thus gain objective validity as the conditions of a given\nexperience which reveals only appearances. Ordinary consciousness\nsimilarly starts from the assumption of the absolute validity of the\nmoral law. But in this case the consciousness of the law is discovered\non examination to be explicable, _even as a possibility_, only on the\nassumption that it is due to the autonomous activity of a noumenal\nbeing. By its existence it proves the conditions through which alone it\nis explicable. Its mere existence suffices to prove that its validity is\nobjective in a deeper and truer sense than the principles of\nunderstanding. _The notion of freedom, and therefore all the connected\nIdeas of pure Reason, gain noumenal reality as the conditions of a moral\nconsciousness which is incapable of explanation as illusory or even\nphenomenal._ Since the consciousness of the moral law is thus noumenally\ngrounded, it has a validity with which nothing in the phenomenal world\ncan possibly compare. It is the one form in which noumenal reality\ndirectly discloses itself to the human mind.[1724]\n\nObviously the essential crux of Kant\u2019s argument lies in the proof that\nthe moral consciousness _is_ only explicable in this manner, as the\nself-legislation of a noumenal being. Into the merits of his argument we\ncannot, however, here enter; and I need only draw attention to the\nmanner in which it conflicts with the statement of the preceding\nsection, that the possibility of transcendental freedom is a purely\nspeculative question with which practical Reason is not concerned. The\nreality of freedom, as a form of noumenal activity, is the cardinal fact\nof Kant\u2019s metaphysics of morals. For though our consciousness of the\nmoral law is the _ratio cognoscendi_ of freedom, transcendental freedom\nis the _ratio essendi_ of the moral law.[1725]\n\n     \u201cWith this faculty [of practical Reason], transcendental freedom is\n     also established; freedom, namely, in that absolute sense in which\n     speculative Reason required it, in its use of the concept of\n     causality, in order to escape the antinomy into which it inevitably\n     falls, when in the chain of cause and effect it tries to think the\n     _unconditioned_.... Freedom is the only one of all the Ideas of the\n     speculative Reason of which we _know_ the possibility _a priori_\n     (without, however, understanding it), because it is the condition\n     of the moral law which we know.\u201d[1726] \u201d[Freedom] is the only one\n     of all the Ideas of pure Reason whose object is a thing of fact and\n     to be reckoned among the _scibilia_.\u201d[1727] \u201cIt is thus very\n     remarkable that of the three pure rational Ideas, God, freedom, and\n     immortality, that of freedom is the only concept of the\n     supersensible which (by means of the causality that is thought in\n     it) proves its objective reality in nature by means of the effects\n     it can produce there; and thus renders possible the connection of\n     both the others with nature, and of all three with one another so\n     as to form a Religion.... The concept of freedom (as fundamental\n     concept of all unconditioned practical laws) can extend Reason\n     beyond those bounds within which every natural (theoretical)\n     concept must inevitably remain confined.\u201d[1728]\n\nThus freedom is for Kant a demonstrated fact, and in that respect\ndiffers from the Ideas of God and immortality, which are merely\nproblematic conceptions, and which can be postulated only as articles of\n\u201cpractical faith.\u201d\n\nThis brings us to the final question, upon what grounds Kant ascribes\nvalidity to the Ideas of God and immortality. At this point in his\nargument Kant introduces the conception of the _Summum Bonum_. Reason,\nin prescribing the moral law, prescribes, as the final and complete end\nof all our actions, the _Summum Bonum_, _i.e._ happiness proportioned to\nmoral worth. _Owing to the limitations of our faculties_, the complete\nattainment of this supreme end is conceivable by us only on the\nassumption of a future life wherein perfect worthiness may be attained,\nand of an omnipotent Divine Being who will apportion happiness in\naccordance with merit.\n\n     \u201d[This Divine Being] must be omnipotent, in order that the whole of\n     nature and its relation to morality ... may be subject to his will;\n     omniscient, that he may know our innermost sentiments and their\n     moral worth; omnipresent, that he may be immediately present for\n     the satisfying of every need which the highest good demands;\n     eternal, that this harmony of nature and freedom may never fail,\n     etc.\u201d[1729]\n\nThe moral ideal thus supplies us with a ground[1730] for regarding the\nuniverse as systematically ordered according to moral purposes, and also\nwith a principle that enables us to infer the nature and properties of\nits Supreme Cause. In place of a demonology, which is all that physical\ntheology can establish, we construct upon moral grounds a genuine\ntheology.\n\nThe concepts thus obtained are, however, anthropomorphic; and for that\nreason alone must be denied all speculative value. This is especially\nevident in regard to the Idea of God. Owing to our incapacity to\ncomprehend how moral merit can condition happiness, we conceive them as\n_externally_ combined through the intervention of a supreme Judge and\nRuler. As Kant indicates,[1731] we must not assert that this represents\nthe actual situation. He himself seems to have inclined to a more\nmystical interpretation of the universe, conceiving the relation of\nhappiness to virtue as being grounded in a supersensuous but necessary\norder that may, indeed, be bodied forth in the inadequate symbols of the\ndeistic creed, but which in its true nature transcends our powers of\nunderstanding. So far as the Ideas of God and immortality are necessary\nto define the moral standpoint, they have genuine validity for all moral\nbeings; but if developed on their own account as speculative dogmas,\nthey acquire a definiteness of formulation which is not essential to\ntheir moral function, and which lays them open to suspicion even in\ntheir legitimate use.\n\nThese considerations also indicate Kant\u2019s further reason for entitling\nthe _Summum Bonum_, God and immortality, Ideas of _faith_. Though they\ncan be established as presuppositions of the moral situation in which we\nfind ourselves, such demonstration itself rests upon the acceptance of\nthe moral consciousness as possessing a supersensuous sanction; and that\nin turn is determined by features in the moral situation not deducible\nfrom any higher order of considerations.\n\n     \u201cBelief in matters of faith is a belief in a pure practical point\n     of view, _i.e._ a moral faith, which proves nothing for\n     theoretical, pure, rational cognition, but only for that which is\n     practical and directed to the fulfilment of its duties; it in no\n     way extends speculation.... If the supreme principle of all moral\n     laws is a postulate, the possibility of its highest Object ... is\n     thereby postulated along with it.\u201d[1732] \u201cSo far, as practical\n     Reason has the right to yield us guidance, we shall not look upon\n     actions as obligatory because they are the commands of God, but\n     shall regard them as divine commands because we have an inward\n     obligation to them.... Moral theology is thus of immanent use only.\n     It enables us to fulfil our vocation in this present world by\n     showing us how to adapt ourselves to the system of all ends, and by\n     warning us against the fanaticism and indeed the impiety of\n     abandoning the guidance of a morally legislative Reason in the\n     right conduct of our lives, in order to derive guidance directly\n     from the Idea of the Supreme Being. For we should then be making a\n     transcendent employment of moral theology; and that, like a\n     transcendent use of pure speculation, must pervert and frustrate\n     the ultimate ends of Reason.\u201d[1733]\n\n\nSECTION III\n\nOPINING, KNOWING, AND BELIEVING[1734]\n\nKant first distinguishes between conviction (_Ueberzeugung_) and\npersuasion (_Ueberredung_). A judgment which is objectively grounded,\nand which is therefore valid for all other rational beings, is affirmed\nwith conviction. When the affirmation is due only to the peculiar\ncharacter of the subject, the manner in which it is asserted may be\nentitled persuasion. Persuasion is therefore \u201ca mere illusion.\u201d[1735]\nConviction exists in three degrees, opinion, belief, and knowledge. In\nopinion we are conscious that the judgment is insufficiently grounded,\nand that our conviction is subjectively incomplete. In belief the\nsubjective conviction is complete, but is recognised as lacking in\nobjective justification. In knowledge the objective grounds and the\nsubjective conviction are alike complete.\n\nAfter pointing out that opinion is not permissible in judgments of pure\nReason,[1736] Kant develops the further distinction between _pragmatic\nor doctrinal belief_ and _moral belief_. When a belief is contingent\n(_i.e._ is affirmed with the consciousness that on fuller knowledge it\nmay turn out to be false), and yet nevertheless supplies a ground for\nthe employment of means to certain desired ends, it may be called\n_pragmatic_ belief. Such belief admits of degree, and can be tested by\nwager or by oath.[1737] What may be called _doctrinal_ belief is\nanalogous in character, and is taken by Kant, in somewhat misleading\nfashion, as describing our mode of accepting such doctrines as the\nexistence of God and the immortality of the soul.[1738] They are adopted\nas helpful towards a contingent but important end, the discovery of\norder in the system of nature. This account of the nature of Ideas is in\nline with Kant\u2019s early view of them as _merely_ regulative. Taken in\nconnection with his repeated employment of the term \u2018moral sentiments\u2019\n(_moralische Gesinnungen_), it tends to prove that this section is early\nin date of writing.\n\nIn _moral belief_ the end, the _Summum Bonum_, is absolutely necessary,\nand as there is only one condition under which we can conceive it as\nbeing realised, namely, on the assumption of the existence of God and of\na future life, the belief in God and immortality possesses the same\ncertainty as the moral sentiments.\n\n     \u201cThe belief in a God and another world is so interwoven with my\n     moral sentiment that as there is little danger of my losing the\n     latter, there is equally little cause for fear that the former can\n     ever be taken from me.\u201d[1739]\n\nAs I have just suggested, this basing of moral belief upon subjective\nsentiments, which, as Kant very inconsistently proceeds to suggest, may\npossibly be lacking in certain men, marks this section as being of early\norigin. But in concluding the section, in reply to the objection that,\nin thus tracing such articles of faith to our \u201cnatural interest\u201d in\nmorality, philosophy admits its powerlessness to advance beyond the\nordinary understanding, Kant propounds one of his abiding convictions,\nnamely, that in matters which concern all men without distinction nature\nis not guilty of any partial distribution of her gifts, and that in\nregard to the essential ends of human nature the highest philosophy\ncannot advance beyond what is revealed to the common\nunderstanding.[1740] The reverence which Kant ever cherished for the\nmemory of his parents, and for the religion which was so natural to\nthem, must have predisposed him to a recognition of the widespread\nsources of the spiritual life. But Kant has himself placed on record his\nsense of the great debt which in this connection he also owed to the\nteaching of Rousseau.\n\n     \u201cI am by disposition an enquirer. I feel the consuming thirst for\n     knowledge, the eager unrest to advance ever further, and the\n     delights of discovery. There was a time when I believed that this\n     is what confers real dignity upon human life, and I despised the\n     common people who know nothing. Rousseau has set me right. This\n     imagined advantage vanishes. I learn to honor men, and should\n     regard myself as of much less use than the common labourer, if I\n     did not believe that my philosophy will restore to all men the\n     common rights of humanity.\u201d[1741]\n\nThe sublimity of the starry heavens and the imperative of the moral law\nare ever present influences on the life of man; and they require for\ntheir apprehension no previous initiation through science and\nphilosophy. The naked eye reveals the former; of the latter all men are\nimmediately aware.[1742] In their universal appeal they are of the very\nsubstance of human existence. Philosophy may avail to counteract the\nhindrances which prevent them from exercising their native influence; it\ncannot be a substitute for the inspiration which they alone can yield.\n\n\n\n\nCHAPTER III\n\nTHE ARCHITECTONIC OF PURE REASON[1743]\n\n\nAdickes[1744] very justly remarks that \u201cthis is a section after Kant\u2019s\nown heart, in which there is presented, almost unsought, the\nopportunity, which he elsewhere so frequently creates for himself, of\nindulging in his favourite hobby.\u201d The section is of slight scientific\nimportance, and is chiefly of interest for the light which it casts upon\nKant\u2019s personality. Moreover the distinctions which Kant here draws are\nfor the most part not his own philosophical property, but are taken over\nfrom the Wolffian system.\n\nThe distinctions may be exhibited in tabular form as follows:[1745]\n\n\n                                KNOWLEDGE\n\n  +---------------------------------------------------------------------------+\n  | Historical|                                                               |\n  |    or     |                                                               |\n  | empirical |                 Rational (_cognitio ex principiis_)           |\n  | (_cognitio_ |                                                             |\n  | _ex datis_) |-------------------------------------------------------------|\n  |-----------|Mathe- |Philosophical[1746] (based on pure concepts):--meta-   |\n  |     |     |matics,| physics in the widest sense.                          |\n  |Empi-|Empi-|based  |-------------------------------------------------------|\n  |rical|rical|on the |Prop- |System of pure reason, both practical and       |\n  |phys-|anth-|constr-|aedeu-|speculative: metaphysics in the wider sense.    |\n  |iolo-|ropo-|uction |tic:  |------------------------------------------------|\n  |gy   |logy |of con-|Cri-  |Metaph-|Metaphysics (in a more limited sense) of|\n  |     |     |cepts. |tique.|ysics  | nature,[1747] _i.e_. of all that is.   |\n  |     |     |       |      |of mor-|----------------------------------------|\n  |     |     |       |      |als, _i.e._|Trans- |Physiology of pure reason.  |\n  |     |     |       |      |of all |cend-  |--------------------------------|\n  |     |     |       |      |ought  |ental  |  Immanent. |     Transcendent. |\n  |     |     |       |      |to be. |phil-  |------------+-------------------|\n  |     |     |       |      |       |osophy |Rat- |Rat-  |Cosmology| Theology|\n  |     |     |       |      |       |of pure|ional|ional |    |    |   |     |\n  |     |     |       |      |       |reason:|phy- |psych-|    -----|-----    |\n  |     |     |       |      |       |ontol- |sics.|ology.|Metaphysics in the |\n  |     |     |       |      |       |ogy.[1748]|          most limited sense.|\n  +---------------------------------------------------------------------------+\n\nKant further distinguishes between the \u201cscholastic\u201d and the \u201cuniversal\u201d\nor traditional meaning of the term philosophy.[1749] In the former sense\nphilosophy is viewed from the point of view of its _logical_ perfection,\nand the philosopher appears as an _artist_ of Reason.[1750] Philosophy\nin the broader and higher sense is \u201cthe science of the relation of all\nknowledge to the essential ends of human Reason.\u201d[1751] The philosopher\nthen appears as the _lawgiver_ of human Reason. Of the essential ends,\nthe _ultimate_ end is man\u2019s moral destiny; to this the other essential\nends of human Reason are subordinate means. For though the legislation\nof human Reason concerns nature as well as freedom, and has therefore to\nbe dealt with by a philosophy of nature, _i.e._ of _all that is_, as\nwell as by a philosophy of morals, _i.e._ of _that which ought to be_,\nthe former is subordinate to the latter in the same degree in which in\nhuman life knowledge is subordinate to moral action. Whereas speculative\nmetaphysics serves rather to ward off errors than to extend\nknowledge,[1752] in the metaphysics of morals \u201call culture [_Kultur_] of\nhuman Reason\u201d[1753] finds its indispensable completion.\n\nEmpirical psychology is excluded from the domain of metaphysics. It is\ndestined to form part of a complete system of anthropology, the pendant\nto the empirical doctrine of nature.[1754]\n\n\n\n\nCHAPTER IV\n\nTHE HISTORY OF PURE REASON[1755]\n\n\nThis title, as Kant states, is inserted only to mark the place of the\npresent chapter in a complete system of pure reason. The very cursory\noutline, which alone Kant here attempts to give, merely repeats the main\nhistorical distinctions of which the _Critique_ has made use. The\ncontrast between the sensationalism of Epicurus and the intellectualism\nof Plato has been developed in A 465 ff. = B 493 ff.[1756] The contrast\nbetween Locke and Leibniz is dwelt upon in A 43 ff. = B 60 ff. and A 270\nff. = B 326 ff. Under the title \u2018naturalist of pure Reason\u2019 Kant is\nreferring to the \u2018common sense\u2019 school, which is typically represented\nby Beattie.[1757] In his _Logic_[1758] Kant gives a fuller account of\nhis interpretation of the history of philosophy.\n\n\n\n\nAPPENDIX B\n\nA MORE DETAILED STATEMENT OF KANT\u2019S RELATIONS TO HIS PHILOSOPHICAL\nPREDECESSORS[1759]\n\n\nThe development of philosophy, prior to Kant, had rendered two problems\nespecially prominent--the problem of sense-perception and the problem of\njudgment. The one raises the question of the interrelation of mind\nknowing and objects known; the other treats of the connection holding\nbetween subject and predicate in the various forms of judgment. The one\nenquires how it is possible to know reality; the other seeks to\ndetermine the criterion of truth. These two problems are, as Kant\ndiscovered, inseparable from one another; and the logical is the more\nfundamental of the two. Indeed it was Hume\u2019s analysis of the judgment\ninvolved in the causal principle that enabled Kant to formulate his\nCritical solution of the problem of perception. In this Appendix I\npropose to follow these problems as they rise into view in the systems\nof Descartes and his successors.\n\nGalileo\u2019s revolutionary teaching regarding the nature of motion was the\nimmediate occasion of Descartes\u2019 restatement of the problem of\nperception. That teaching necessitated an entirely new view of the\nnature of matter, and consequently of the interrelation of mind and\nbody. Questions never before seriously entertained now became pressing.\nThe solutions had to be as novel as the situation which they were\ndesigned to meet.\n\nThese new problems arose in the following manner. According to the\nmedieval view, motion may properly be conceived on the analogy of human\nactivity. It comes into being, exhausts itself in exercise, and ceases\nto be. It is a fleeting activity; only its \u201cmaterial\u201d and \u201cformal\u201d\nconditions have any permanence of existence. According to Galileo\u2019s\nteaching, on the other hand, motion is as different from human activity\nas matter is from mind. It is ingenerable and indestructible. We know it\nonly through the effect which in some incomprehensible fashion it\nproduces in those bodies into which it enters, namely, their translation\nfrom one part of space to another. That this translatory motion is\ncalled by the same name as the power which generates it, doubtless in\nsome degree accounts for the fact that our understanding of the one\ntends to conceal from us our entire ignorance of the other.[1760] We\nhave only to reflect, however, in order to realise that motion is\ncompletely mysterious in its intrinsic dynamical nature. We cannot, for\ninstance, profess to comprehend, even in the least degree, how motion,\nthough incapable of existing apart from matter, should yet be\nsufficiently independent to be able to pass from one body to another.\n\nDescartes, following out some of the chief consequences of this new\nteaching, concluded that matter is passive and inert, that it is\ndistinguished neither by positive nor by negative properties from the\nspace which it fills, and that it is to motion that all the articulated\norganisation of animate and inanimate nature is due. Descartes failed,\nindeed, to appreciate the dynamical character of motion, and by\nconstantly speaking as if it were reducible to the translatory motion,\nin which it manifests itself, he represented it as known in all its\nessential features. None the less, the r\u00f4les previously assigned to\nmatter and motion are, in Descartes\u2019 system, completely reversed. Matter\nis subordinated to motion as the instrument to the agency by which it is\ndirected and shaped. On the older view, material bodies had, through the\npossession of formative and vital forces, all manner of intrinsic\npowers. By the new view these composite and nondescript existences are\nresolved into two elements, all the properties of which can be\nquantitatively defined--into a matter which is uniform and homogeneous,\nand into motion whose sole effect is the translation of bodies in space.\nMatter is the passive and inert substance out of which motion, by its\nmere mechanical powers, can produce the whole range of material forms.\n\nThis revolutionary change in the physical standpoint involved\nrestatement of the philosophical issues. But the resulting difficulties\nwere found thoroughly baffling. Though Descartes and his successors were\nwilling to adopt any hypothesis, however paradoxical, which the facts\nmight seem to demand, their theories, however modified and restated, led\nonly deeper into a hopeless _impasse_. The unsolved problems of the\nCartesian systems formed the discouraging heritage to which Kant fell\nheir. If matter is always purely material, and motion is its sole\norganising power, there can be no real kinship between body and mind.\nThe formative and vital forces, which in the Scholastic philosophy and\nin popular thought serve to maintain the appearance of continuity\nbetween matter and mind, can no longer be credited. Motion, which alone\nis left to mediate between the opposites, is purely mechanical, and (on\nDescartes\u2019 view) is entirely lacking in inner or hidden powers. The\nanimal body is exclusively material, and is therefore as incapable of\nfeeling or consciousness as any machine made by human hands. The bodily\nsenses are not \u2018sensitive\u2019; the brain cannot think. Mental experiences\ndo, of course, accompany the brain-motions. But why a sensation should\nthus arise when a particular motion is caused in the brain, or how a\nmental resolution can be followed by a brain state, are questions to\nwhich no satisfactory answer can be given. The mental and the material,\nthe spiritual and the mechanical, fall entirely apart.\n\nThe difficulties arising out of this incomprehensibility of the causal\ninterrelations of mind and body are not, however, in themselves a valid\nargument against a dualistic interpretation of the real. The\ndifficulties of accounting for the causal relation are, in essential\nrespects, equally great even when the interaction is between homogeneous\nexistences. The difficulties are due to the nature of causal action as\nsuch, not to the character of the bodies between which it holds. This,\nindeed, was clearly recognised by Descartes, and was insisted upon by\nhis immediate successors. The transference of motion by impact is no\nless incomprehensible than the interaction of soul and body. If motion\ncan exist only in matter, there is no possible method of conceiving how\nit can make the transition from one discrete portion of matter to\nanother. Causal action is thus a problem which no philosophy can pretend\nto solve, and which every philosophy, whether monistic or dualistic,\nmust recognise as transcending the scope of our present knowledge.\n\nIt is in another and more special form that Descartes\u2019 dualism first\nreveals its fatal defects, namely, in its bearing upon the problem of\nsense-perception. Descartes can solve the problem of knowledge only by\nfirst postulating the doctrine of representative perception. That\ndoctrine is rendered necessary by the dualism of mind and body. Objects\ncan be known only mediately by means of their action upon the\nsense-organs, and through the sense-organs upon the brain. The resulting\nbrain states are in themselves merely forms of motion. They lead,\nhowever, in a manner which Descartes never professes to explain,[1761]\nto the appearance of sensations in the mental field. Out of these\nsensations the mind then constructs mental images of the distant bodies;\nand it is these mental images alone which are directly apprehended.\nMaterial bodies are invisible and intangible; they are knowable only\nthrough their mental duplicates. Thus, according to the doctrine of\nrepresentative perception, each mind is segregated in a world apart. It\nlooks out upon a landscape which is as mental and as truly inward as are\nits feelings and desires. The apparently ultimate relation of mind\nknowing and object known is rendered complex and problematic through the\ndistinction between mental objects and real things. Mental objects are\nin all cases images merely. They exist only so long as they are\napprehended; and they are numerically and existentially distinct in each\nindividual mind. Real things are not immediately perceived; they are\nhypothetically inferred. To ordinary consciousness the body which acts\non the sense-organ is the object known; when reflective consciousness is\nphilosophically enlightened, the object immediately known is recognised\nas a merely mental image, and the external object sinks to the level of\nan assumed cause.\n\nThe paradoxical character of this doctrine is accentuated by Galileo\u2019s\ndistinction between primary and secondary qualities.[1762] Those\nphysical processes, which are entitled light and heat, bear no\nresemblance to the sensations through which they become known. The\nmany-coloured world of ordinary consciousness is an illusory appearance\nwhich can exist only in the human mind. We must distinguish between the\nsensible world which, though purely mental, appears, through an\nunavoidable illusion, to be externally real, and that very different\nworld of matter and motion which reveals its independent nature only to\nreflective thinking. In the latter world the rich variety of sensuous\nappearance can find no place. There remain only the quantitative,\nmechanical properties of extension, figure and motion; and even these\nhave to be interpreted in the revolutionary fashion of physical science.\n\nThe doctrine of representative perception cannot, however, defend\nsuccessfully the positions which it thus involves. It wavers in unstable\nequilibrium. The facts, physical and physiological, upon which it is\nbased, are in conflict with the conclusions in which it results. This\nhas been very clearly demonstrated by many writers in recent\ntimes.[1763] The conflict manifested itself in the period between\nDescartes and Kant only through the uneasy questionings of Locke and\nBerkeley. The problem, fundamental though it be, is almost completely\nignored by Spinoza, Leibniz, and Wolff.\n\nStated in modern terms, the inherently contradictory character of the\ndoctrine consists in its unavoidable alternation between the realist\nattitude to which it owes its origin, and the idealist conclusion in\nwhich it issues. Such oscillation is due to the twofold simultaneous\nrelation in which it regards ideas as standing to the objects that they\nare supposed to represent. The function of sensations is cognitive;\ntheir origin is mechanical. As cognitive they stand to objects in a\nrelation of inclusion; they reveal the objects, reduplicating them in\nimage before the mind. Yet in their origin they are effects,\nmechanically generated by the action of material bodies upon the\nsense-organs and brain. As they are effects mechanically generated,\nthere is no guarantee that they resemble their causes; and if we may\nargue from other forms of mechanical causation, there is little\nlikelihood that they do. They stand to their first causes in a relation\nof exclusion, separated from them by a large number of varying\nintermediate processes. There is thus a conflict between the function of\nsensations and their origin. Their origin in the external objects is\nsupposed to confer upon them a representative power; and yet the very\nnature of this origin invalidates any such claim.\n\nThis irreconcilability of the subjectivist consequences of the doctrine\nwith its realist basis was seized upon by Berkeley. To remove the\ncontradiction, he denied the facts from which the doctrine had been\ndeveloped. That is to say, starting from its results he disproved its\npremisses. Arguing from the physical and physiological conditions of\nperception Descartes had concluded that only sensations can be directly\napprehended by the mind. Berkeley starts from this conclusion, and\nvirtually adopts it as an assumption which cannot be questioned, and\nwhich does not call for proof. Since, he contends, we know only\nsensations, the assertion that they are due to material causes is mere\nhypothesis, and is one for which there may be no valid grounds. As\nDescartes himself had already suggested, there is a second possible\nmethod of interpreting the relevant facts. There may exist an\nall-powerful Being who produces the sensations in our minds from moment\nto moment; and provided that they are produced in the same order as now,\nthe whole material world might be annihilated without our being in the\nleast aware that so important an event had taken place. Since we can\nexperience only sensations, any hypothesis which will account for the\norder of their happening is equally legitimate. The whole question\nbecomes one of relative simplicity in the explanation given. The simpler\nanalysis, other things being equal, must hold the field.\n\nBerkeley reinforces this argument by pointing to the many embarrassing\nconsequences to which Descartes\u2019 dualism must lead. We postulate bodies\nin order to account for the origin of our sensations, and yet are unable\nto do so by their means. The dualistic theory creates more difficulties\nthan it solves, without a single counter-advantage, save perhaps--so\nBerkeley argues--that it seems to harmonise better with the traditional\nprejudices of the philosophic consciousness.\n\nIf we grant Berkeley his premisses, the main lines of his argument are\nfairly cogent, however unconvincing may be his own positive views. The\ncrux, however, of the Berkeleian idealism lies almost exclusively in the\nestablishment of its fundamental assumption, that only ideas (_i.e._\nimages) can be known by the mind. This assumption Berkeley, almost\nwithout argument, takes over from his predecessors. It was currently\naccepted, and from it, therefore, he believed that he could safely\nargue. It rests, however, upon the assumption of facts which he himself\nquestions. In rejecting the Cartesian dualism he casts down the ladder\nby which alone it is possible to climb into his position. For save\nthrough the facts of physics and physiology there seems to be no\npossible method of disproving the belief of ordinary consciousness, that\nin perception we apprehend independent material bodies. And until that\nbelief can be shown to be false and ungrounded, the Berkeleian idealism\nis without support. It cannot establish the fundamental assumption upon\nwhich its entire argument proceeds. Thus, though Berkeley convincingly\ndemonstrates the internal incoherence of the doctrine of representative\nperception--the inconsistency of its conclusions with the physical and\nphysiological facts upon which alone it can be based--he cannot himself\nsolve the problem in answer to which that doctrine was propounded. His\nservices, like those of so many other reformers, were such as he did not\nhimself foresee. In simplifying the problem, he prepared the way for the\nmore sceptical treatment of its difficult issues by Hume.\n\nAt this point, in the philosophy of Hume, the problem of perception\ncomes into the closest possible connection with the logical problem,\nreferred to above. The question, how mind knowing is related to the\nobjects known, is found to depend upon the question, how in certain\ncrucial cases predicates may legitimately be referred to their subject.\nThis logical problem arises in two forms, a narrower and a wider. The\nnarrower issue concerns only the principle of causality. With what right\ndo we assert that every event must have a cause? What is the ground\nwhich justifies us in thus predicating of events a causal character?\nObviously, this logical question is fundamental, and must be answered\nbefore we can hope to solve the more special problem, as to our right to\ninterpret sensations as effects of material bodies. Hume was the first\nto emphasise the vital interconnection of these two lines of enquiry.\n\nThe wider issue is the generating problem of Kant\u2019s _Critique_: How in a\njudgment can a predicate be asserted of a subject in which it is not\nalready involved? In other words, what is it that in such a case\njustifies us in connecting the predicate with the subject? Though this\nproblem was never directly raised by any pre-Kantian thinker, not even\nby Hume, it is absolutely vital to all the pre-Kantian systems. Thus\nDescartes\u2019 philosophy is based upon a distinction, nowhere explicitly\ndrawn but everywhere silently assumed, between abstract and fruitful\nideas. The former contain just so much content and no more; this content\nmay be explicitly unfolded in a series of judgments, but no addition is\nthereby made to our knowledge. The latter, on the other hand, are\nendowed with an extraordinary power of inner growth. To the attentive\nmind they disclose a marvellous variety of inner meaning. The chief\nproblem of scientific method consists, according to Descartes, in the\ndiscovery of these fruitful ideas, and in the separation of them from\nthe irrelevant accompaniments which prevent them from unfolding their\ninner content. Once they are discovered, the steady progress of\nknowledge is assured. They are the springs of knowledge, and from them\nwe have only to follow down the widening river of truth.\n\nDescartes professed to give a complete list of the possible fruitful\nideas. They are, he claimed, better known than any other concepts. They\nlie at the basis of all experience, and no one can possibly be ignorant\nof them; though, owing to their simplicity and omnipresence, their\nphilosophical importance has been overlooked. When, however, Descartes\nproceeded to classify them, he found that while such ideas as space,\ntriangle, number, motion, contain an inexhaustible content that is\nprogressively unfolded in the mathematical sciences, those ideas, on the\nother hand, through which we conceive mental existences,--the notions of\nmind, thought, self--do not by any means prove fruitful upon attentive\nenquiry. As Malebranche later insisted, we can define mind only in\nnegative terms; its whole meaning is determined through its opposition\nto the space-world, which alone is truly known. Though it is the\nfunction of mind to know, it cannot know itself. And when we remove from\nour list of ideas those which are not really fruitful, we find that only\nmathematical concepts remain.[1764] They alone have this apparently\nmiraculous property of inexhaustibly developing before the mind.\nScientific knowledge is limited to the material world; and even there,\nthe limits of our mathematical insight are the limits of our knowledge.\n\nMalebranche believed no less thoroughly than Descartes in the asserted\npower and fruitfulness of mathematical concepts. Under the influence of\nthis belief, he developed, as so many other thinkers from Plato onwards\nhave done, a highly mystical theory of scientific knowledge. It is a\nrevelation of eternal truth, and yet is acquired by inner reflection,\nnot laboriously built up by external observation. It comes by searching\nof the mind, not by exploration of the outer world. But Malebranche was\nnot content, like Descartes, merely to accept this type of knowledge. He\nproceeded to account for it in metaphysical terms. The fruitfulness of\nmathematical ideas is due, he claimed, to the fundamental concept of\nextension in which they all share. This idea, representing, as it does,\nan infinite existence, is too great to be contained within the finite\nmind. Through it the mind is widened to the apprehension of something\nbeyond itself; we know it through consciousness of its archetype in the\nmind of God. It is the one point at which consciousness transcends its\nsubjective limits. Its fruitfulness is due to, and is the manifestation\nof, this divine source. The reason why we are condemned to remain\nignorant of everything beyond the sphere of quantity is that extension\nalone holds this unique position. It is the only fruitful idea which the\nmind possesses, and other concepts, such as triangle, circle, or number,\nare fruitful only in proportion as they share in it. We can acquire no\ngenuine knowledge even of the nature of the self. Being ignorant of\nmind, we cannot comprehend the self which is one of its modes. It is as\nif we sought to comprehend the nature of a triangle, in the absence of\nany conception of space. Were we in possession of the archetypal idea of\nmind, we should not only be able to deduce from it those various\nfeelings and emotions which we have already experienced, and those\nsensations of the secondary qualities which we falsely ascribe to the\ninfluence of external objects, but we should also be able to discover by\npure contemplation innumerable other emotions and qualities, which\nentirely transcend our present powers. And all of these would then be\nexperienced in their ideal nature, and not, as now, merely through\nfeeble and confused feeling. If mathematicians destroy their bodily\nhealth through absorption in the progressive clarification of the\nmysteries of space, what might not happen if the archetypal idea of mind\nwere revealed to us? Could we attend to the preservation of a body which\nwould incessantly distract us from the infinite and overwhelming\nexperiences of our divine destiny?\n\nThis romantic conception of the possibilities of rational science\nreveals more clearly than any other Cartesian doctrine the real bearing\nand perverse character of the rationalistic preconceptions which\nunderlie the Cartesian systems. The Cartesians would fain make rational\nscience, conceived on the analogy of the mathematical disciplines,\ncoextensive with the entire realm of the real. This grotesque enterprise\nis conceived as abstractly possible even by so cautious a thinker as\nJohn Locke. His reason for condemning the physical sciences as logically\nimperfect is that they fail to conform to this rationalistic ideal.\nHence those sentences which sound so strangely in the mouth of Locke,\nthe sensationalist.\n\n     \u201cIt is the contemplation of our abstract ideas that alone is able\n     to afford us general knowledge.\u201d[1765] \u201cThe true method of\n     advancing knowledge is by considering our abstract ideas.\u201d[1766]\n     \u201d[Did we know the real essence of gold] it would be no more\n     necessary that gold should exist, and that we should make\n     experiments upon it, than it is necessary for the knowing of the\n     properties of a triangle, that a triangle should exist in any\n     matter: the idea in our minds would serve for the one as well as\n     for the other.\u201d[1767] \u201cIn the knowledge of bodies, we must be\n     content to glean what we can from particular experiments, since we\n     cannot, from a discovery of their real essences, grasp at a time\n     whole sheaves, and in bundles comprehend the nature and the\n     properties of whole species together.\u201d[4]\n\nLocke\u2019s empirical doctrine of knowledge is thus based upon a\nrationalistic theory of the real. It is not, he holds, the constitution\nof reality, but the _de facto_ limitations of our human faculties which\nmake empirical induction the only practicable mode of discovery in\nnatural science. Indeed, Locke gives more extreme expression than even\nDescartes does, to the mystically conceived mathematical method. Being\nignorant of mathematics, and not over well-informed even in the physical\nsciences, Locke was not checked by any too close acquaintance with the\nreal character and necessary limits of this method; and he accordingly\nmakes statements in that unqualified fashion which seldom fails to\nbetray the writer who is expounding views which he has not developed for\nhimself by first-hand study of the relevant facts.\n\nBut though the unique character of mathematical knowledge thus forced\nitself upon the attention of all the Cartesian thinkers, and in the\nabove manner led even the most level-headed of Descartes\u2019 successors to\ndream strange dreams, no real attempt was made (save in the neglected\nwritings of Leibniz) to examine, in a sober spirit, the grounds and\nconditions of its possibility. In the English School, Locke\u2019s eulogy of\nabstract ideas served only to drive his immediate successors to an\nopposite extreme. Both Berkeley and Hume attempted to explain away, in\nan impossible manner, those fundamental differences, which, beyond all\nquestioning, profoundly differentiate mathematical from empirical\njudgments.[1768] It is not surprising that Kant, who had no direct\nacquaintance with Hume\u2019s _Treatise_, should have asserted that had Hume\nrealised the bearing of his main teaching upon the theory of\nmathematical science, he would have hesitated to draw his sceptical\nconclusions. Such, however, was not the case. Hume\u2019s theory of\nmathematical reasoning undoubtedly forms the least satisfactory part of\nhis philosophy. He did, however, perceive the general bearing of his\ncentral teaching. It was in large degree his ignorance of the\nmathematical disciplines that concealed from him the thorough\nunsatisfactoriness of his general position, and which prevented him from\nformulating the logical problem in its full scope--the problem, namely,\nhow judgments which make additions to our previous knowledge, and yet do\nnot rest upon mere sensation, are possible. He treated it only as it\npresents itself in those judgments which involve the concept of\ncausality.[1769] But this analysis of causal judgments awoke Kant from\nhis dogmatic slumber, and so ultimately led to the raising of the\nlogical problem in its widest form:--how synthetic _a priori_ judgments,\nwhether mathematical, physical, or metaphysical, are possible.\n\nHume discussed the causal problem both in regard to the general\nprinciple of causality and in its bearing upon our particular judgments\nof causal relation. The problems concerned in these two discussions are\nessentially distinct. The first involves immensely wider issues, and so\nfar as can be judged from the existing circumstantial evidence,[1770] it\nwas this first discussion, not as has been so often assumed by Kant\u2019s\ncommentators the second and more limited problem, which exercised so\nprofound an influence upon Kant at the turning-point of his\nspeculations. In stating it, it will be best to take Hume\u2019s own words.\n\n     \u201cTo begin with the first question concerning the necessity of a\n     cause: \u2018Tis a general maxim in philosophy, that _whatever begins to\n     exist_, must have a cause of existence. This is commonly taken for\n     granted in all reasonings, without any proof given or demanded.\n     \u2018Tis supposed to be founded on intuition, and to be one of those\n     maxims, which though they may be deny\u2019d with the lips, \u2018tis\n     impossible for men in their hearts really to doubt of. But if we\n     examine this maxim by the idea of knowledge above explain\u2019d we\n     shall discover in it no mark of any such intuitive certainty; but\n     on the contrary shall find, that \u2018tis of a nature quite foreign to\n     that species of conviction.\u201d[1771]\n\nThe principle that every event must have a cause, is neither intuitively\nnor demonstratively certain. So far from there existing a _necessary_\nconnection between the idea of an event as something happening in time\nand the idea of a cause, no connection of any kind is discoverable by\nus. We can conceive an object to be non-existent at this moment, and\nexistent the next, without requiring to conjoin with it the altogether\ndifferent idea of a productive source.\n\nThis had been implicitly recognised by those few philosophers who had\nattempted to give demonstrations of the principle. By so doing, however,\nthey only reinforce Hume\u2019s contention that it possesses no rational\nbasis. When Hobbes argues that as all the points of time and place in\nwhich we can suppose an object to begin to exist, are in themselves\nequal, there must be some cause determining an event to happen at one\nmoment rather than at another, he is assuming the very principle which\nhe professes to prove. There is no greater difficulty in supposing the\ntime and place to be fixed without a cause, than in supposing the\nexistence to be so determined. If the denial of a cause is not\nintuitively absurd in the one case, it cannot be so in the other. If the\nfirst demands a proof, so likewise must the second. Similarly with the\narguments advanced by Locke and Clarke. Locke argues that if anything is\nproduced without a cause, it is produced by nothing, and that that is\nimpossible, since nothing can never be a cause any more than it can be\nsomething, or equal to two right angles. Clarke\u2019s contention that if\nanything were without a cause, it would produce _itself_, _i.e._ exist\nbefore it existed, is of the same character. These arguments assume the\nonly point which is in question.\n\n     \u201cWhen we exclude all causes we really do exclude them, and neither\n     suppose nothing nor the object itself to be the causes of the\n     existence, and consequently can draw no argument from the absurdity\n     of these suppositions to prove the absurdity of that\n     exclusion.\u201d[1772]\n\nThe remaining argument, that every effect must have a cause, since this\nis implied in the very idea of an effect, is \u201cstill more frivolous.\u201d\n\n     \u201cEvery effect necessarily presupposes a cause; effect being a\n     relative term, of which cause is the correlative. But this does not\n     prove that every being must be preceded by a cause; no more than it\n     follows, because every husband must have a wife, that therefore\n     every man must be married.\u201d[1773]\n\nThe far-reaching conclusion, that the principle of causality has no\npossible rational basis, Hume extends and reinforces through his other\ndoctrines, viz. that synthetic reason[1774] is merely generalised\nbelief, and that belief is in all cases due to the ultimate instincts\nand propensities which _de facto_ constitute our human nature. The\nsynthetic principles which lie at the basis of our experience are\nnon-rational in character. Each is due to a \u2018blind and powerful\ninstinct,\u2019 which, demanding no evidence, and ignoring theoretical\ninconsistency for the sake of practical convenience, necessitates\nbelief.\n\n     \u201cNature by an absolute and uncontrollable necessity has determined\n     us to judge as well as to breathe and feel.\u201d[1775] \u201cAll these\n     operations are a species of natural instincts, which no reasoning\n     or process of the thought and understanding is able either to\n     produce or to prevent.\u201d[1776]\n\nReason is \u201cnothing but a wonderful and unintelligible instinct in our\nsouls.\u201d[1777] It justifies itself by its practical uses, but can afford\nno standard to which objective reality must conform.\n\nIt is from this point of view that Hume states his answer to the problem\nof perception. Our natural belief in the permanence and identity of\nobjects, as expressed through the principle of substance and attribute,\nleads us to interpret the objects of sense-perception as independent\nrealities. We interpret our subjective sensations as being qualities of\nindependent substances. Our other natural belief, in the dynamical\ninterdependence of events, as expressed through the principle of\ncausality, leads, however, to the opposite conclusion, that the known\nobjects are merely mental. For by it we are constrained to interpret\nsensations, not as objective qualities, but only as subjective effects,\nexpressive of the reactions of our psycho-physical organism. The\nCartesian problems owe their origin to the mistaken attempt to\nharmonise, in a theoretical fashion, these two conflicting principles.\nThe conflict is inevitable and the antinomy is insoluble, so long as the\ntwo principles are regarded as objectively valid. The only satisfactory\nsolution comes through recognition that reason is unable to account,\nsave in reference to practical ends, even for its own inevitable\ndemands. The principle of substance and attribute and the principle of\ncausality co-operate in rendering possible such organisation of our\nsense-experience as is required for practical life. But when we carry\nthis organisation further than practical life itself demands, the two\nprinciples at once conflict.\n\nKant shows no interest in this constructive part of Hume\u2019s philosophy;\nand must, indeed, have been almost entirely ignorant of it, since it\nfinds only very imperfect expression in the _Enquiry_, and is ignored in\nBeattie\u2019s _Nature of Truth_. Accordingly, Kant does not regard Hume as\noffering a positive explanation of knowledge, but rather as representing\nthe point of view of thoroughgoing scepticism. But even had he been\nacquainted at first hand with Hume\u2019s _Treatise_, he would undoubtedly\nhave felt little sympathy with Hume\u2019s naturalistic view of the function\nof reason. His training in the mathematical sciences would have enabled\nhim to detect the inadequacy of Hume\u2019s treatment of mathematical\nknowledge, and his strong moral convictions would have led him to rebel\nagainst the naturalistic assumptions which underlie Hume\u2019s entire\nposition. The Berkeley-Hume comedy is thus repeated with reversed r\u00f4les.\nJust as Berkeley\u2019s anti-materialistic philosophy was mainly influential\nas a step towards the naturalism of Hume, and as such still survives in\nthe philosophies of John Stuart Mill, Herbert Spencer, Huxley, Mach and\nKarl Pearson, so in turn Hume\u2019s anti-metaphysical theory of knowledge\nwas destined to be one of the chief contributory sources of the German\nspeculative movement.\n\nWe may now turn to Hume\u2019s treatment of the narrower problem--that of\njustifying our _particular_ causal judgments. Hume\u2019s attitude towards\nthis question is predetermined by the more fundamental argument, above\nstated, which precedes it in the _Treatise_, but which is entirely\nomitted from the corresponding chapters of the _Enquiry_. As the general\nprinciple of causality is of an irrational character, the same must be\ntrue of those particular judgments which are based upon it. Much of\nHume\u2019s argument on this question is, indeed, merely a restatement of\nwhat had already been pointed out by his predecessors. There is no\nnecessary connection discoverable between _any_ cause and its effect.\nThis is especially evident as regards the connection between brain\nstates and mental experiences. No explanation can be given why a motion\nin the brain should produce sensations in the mind, or why a mental\nresolution should produce movements in the body. Such sequences may be\nempirically verified; they cannot be rationally understood. That this\nlikewise holds, though in less obvious fashion, of the causal\ninterrelations of material bodies, had been emphasised by Geulincx,\nMalebranche, Locke, and Berkeley. The fact that one billiard ball should\ncommunicate motion to another by impact is, when examined, found to be\nno less incomprehensible than the interaction of mind and body. Hume, in\nthe following passage, is only reinforcing this admitted fact, in terms\nof his own philosophy.\n\n     \u201cWe fancy that were we brought on a sudden into this world we could\n     at first have inferred that one billiard ball would communicate\n     motion to another upon impulse; and that we needed not to have\n     waited for the event, in order to pronounce with certainty upon it.\n     Such is the influence of custom, that, where it is strongest, it\n     not only covers our natural ignorance, but even conceals itself,\n     and seems not to take place merely because it is found in the\n     highest degree.\u201d[1778]\n\nNor are we conscious of any causal power within the self. When Berkeley\nclaims that mind has the faculty of producing images at will, he is\nreally ascribing to it creative agency. And such creation, as\nMalebranche had already pointed out, is not even conceivable.\n\n     \u201cI deny that my will produces in me my ideas, for I cannot even\n     conceive how it could produce them, since my will, not being able\n     to act or will without knowledge, presupposes my ideas and does not\n     make them.\u201d[1779] \u201cIs there not here,\u201d Hume asks, \u201ceither in a\n     spiritual or material substance, or both, some secret mechanism or\n     structure of parts, upon which the effect depends, and which, being\n     entirely unknown to us, renders the power or energy of the will\n     equally unknown and incomprehensible?\u201d[1780]\n\nBut the fact that Hume thus restates conclusions already emphasised by\nhis predecessors will not justify us in contending (as certain\nhistorians of philosophy seem inclined to do) that in his treatment of\nthe causal problem he failed to make any important advance upon the\nteaching of the Occasionalists. Hume was the first to perceive the\nessential falsity of the Cartesian, rationalistic view of the causal\nnexus. For Descartes, an effect is that which can be deduced with\nlogical necessity from the concept of its cause. The Occasionalists\nsimilarly argued that because natural events can never be deduced from\none another they must in all cases be due to supernatural agency; like\nDescartes, they one and all failed to comprehend that since by an effect\nwe mean that which follows _in time_ upon its cause, and since,\ntherefore, the principle of causality is the law of _change_, the nature\nof causality cannot be expressed in logical terms. Hume was the first to\nappreciate the significance of this fundamental fact; and an entirely\nnew set of problems at once came into view. If causal connection is not,\nas previous thinkers had believed, logical in character, if it does not\nsignify logical dependence of the so-called effect upon its cause, its\ntrue connotation must lie elsewhere; and until this has been traced to\nits hidden source, any attempted solution of metaphysical problems is\ncertain to involve many false assumptions. The answer that is given to\nthe problem of the origin and content of the causal concept must\ndetermine our interpretation alike of sense-experience and of pure\nthought.\n\nThe problem presents on examination, however, a most paradoxical aspect.\nAs Hume has already shown, every effect is an event distinct from its\ncause, and there is never any connection, beyond that of mere sequence,\ndiscoverable between them. We observe only sequence; we assert necessary\nconnection. What, then, is in our minds when this latter assertion is\nmade? And how, if the notion of necessitated connection cannot be gained\nthrough observation of the external events, is it acquired by us? Hume\nagain propounds a naturalistic solution. Causation, _i.e._\n_necessitated_ sequence in time, is not in any sense a conception; it is\nnot a comprehended relation between events, but a misunderstood feeling\nin our minds. We cannot form any, even the most remote, conception of\nhow one event can produce another. Neither imagination nor pure thought,\nhowever freely they may act, are capable of inventing any such notion.\nBut nature, by the manner in which it has constituted our minds, deludes\nus into the belief that we are in actual possession of this idea. The\nrepeated sequence of events, in fixed order, generates in us the feeling\nof a tendency to pass from the perception or idea of the one to the idea\nof the other. This feeling, thus generated by custom, and often in\nsomewhat confused fashion combined with the feeling of \u2018animal nisus,\u2019\nwhich is experienced in bodily effort, is mistaken by the mind for a\ndefinite concept of force, causality, necessary connection. As mere\nfeeling it can afford no insight into the relation holding between\nevents, and as merely subjective can justify no inference in regard to\nthat relation. The terms force, causality, necessitated sequence in\ntime, have a practical value, as names for our instinctive, natural\nexpectations; but when employed as instruments for the _theoretical_\ninterpretation of experience, they lead us off on a false trail.\n\nThis is one of the fundamental points upon which Hume reveals a deeper\nspeculative insight than either Malebranche, Geulincx, or Locke. Though\nthese latter insist upon our ignorance of the relation holding between\nevents, they still assume that causation and natural necessity are\nconcepts which have a quite intelligible meaning; and in consequence\nthey fail to draw the all-important conclusion, that the general\nprinciple of causality has neither intuitive nor demonstrative validity.\nFor that is the revolutionary outcome of Hume\u2019s analysis of the notion\nof necessitated connection. The principle of causality is a synthetic\njudgment in which no connection is discoverable between its subject and\nits predicate. That is the reason why it is neither self-evident nor\ncapable of being established upon more ultimate grounds.\n\nAs has already been stated, the wider problem concerning the _principle_\nof causality is developed only in the _Treatise_; the problem regarding\nthe _concept_ of causality is discussed both in the _Treatise_ and in\nthe _Enquiry_. An appreciation of the wider problem is required,\nhowever, in order to set this second problem in its true light, for it\nis only through its connection with the wider issue that Hume\u2019s\nreduction of the concept of causality to a merely instinctive,\nnon-rational expectation acquires its full significance. Hume\u2019s analysis\nthen amounts, as Kant was the first to realise, to an attack upon the\nobjective validity of all constructive thinking. Not only rationalism,\nbut even such metaphysics as may claim to base its conclusions upon the\nteaching of experience, is thereby rendered altogether impossible. The\nissue is crucial, and must be honestly faced, before metaphysical\nconclusions, no matter what their specific character may be, whether _a\npriori_ or empirical, can legitimately be drawn. If we may not assert\nthat an event must have some cause, even the right to enquire for a\ncause must first be justified. And if so fundamental a principle as that\nof causality is not self-evident, are there any principles which can\nmake this claim?\n\nThe account which we have so far given of Hume\u2019s argument covers only\nthat part of it which is directed against the rationalist position, and\nwhich was therefore so influential in turning Kant on to the line of his\nCritical speculations. But Hume attacked with equal vigour the\nempiricist standpoint; and as this aspect of his teaching, constituting\nas it did an integral part of Kant\u2019s own philosophy, must undoubtedly\nhave helped to confirm Kant in his early rationalist convictions, we may\nprofitably dwell upon it at some length. In opposition to the\nempiricists, Hume argues that experience is incapable of justifying any\ninference in regard to matters of fact. It cannot serve as a basis from\nwhich we can inductively extend our knowledge of facts beyond what the\nsenses and memory reveal. Inductive inference, when so employed,\nnecessarily involves a _petitio principii_; we assume the very point we\nprofess to have proved.\n\nThe argument by which Hume establishes this important contention is as\nfollows. All inductive reasoning from experience presupposes the\nvalidity of belief in causal connection. For when we have no knowledge\nof causes, we have no justification for asserting the continuance of\nuniformities. Now it has been shown that we have no experience of any\nnecessary relation between so-called causes and their effects. The most\nthat experience can supply are sequences which repeat themselves. In\nregarding the sequences as causal, and so as universally constant, we\nmake an assertion for which experience gives no support, and to which no\namount of repeated experience, recalled in memory, can add one jot of\nreal evidence. To argue that because the sequences have remained\nconstant in a great number of repeated experiences, they are therefore\nmore likely to remain constant, is to assume that constancy in the past\nis a ground for inferring it in the future; and that is the very point\nwhich demands proof. In drawing the conclusion we virtually assume that\nthere is a necessary connection, _i.e._ an absolutely constant\nrelation, between events. But since no _single_ experience of causal\nsequence affords ground for inferring that the sequence will continue in\nthe future, no number of repeated experiences, recalled in memory, can\ncontribute to the strengthening of the inference. It is meaningless to\ntalk even of likelihood or probability. The fact that the sun has\nwithout a single known exception arisen each day in the past does not\n(if we accept the argument disproving all knowledge of necessary\nconnection) constitute _proof_ that it will rise to-morrow.\n\n     \u201cNone but a fool or a madman will be unaffected in his expectations\n     or natural beliefs by this constancy, but he is no philosopher who\n     accepts this as in the nature of evidence.\u201d[1781]\n\nSince, for all that we know to the contrary, bodies may change their\nnature and mode of action at any moment, it is vain to pretend that we\nare scientifically assured of the future because of the past.\n\n     \u201cMy practice, you say, refutes my doubts.[1782] But you mistake the\n     purport of my question. As an agent, I am quite satisfied in the\n     point; but as a philosopher, who has some share of curiosity, I\n     will not say scepticism, I want to learn the foundation of this\n     inference. No reading, no enquiry has yet been able to remove my\n     difficulty or give me satisfaction in a matter of such importance.\n     Can I do better than propose the difficulty to the public, even\n     though, perhaps, I have small hopes of obtaining a solution? We\n     shall at least, by this means, be sensible of our ignorance, if we\n     do not augment our knowledge.\u201d[1783]\n\nKant was the first, after thirty years, to take up this challenge.\nExperience is no source of evidence until the causal postulate has been\n_independently_ proved. Only if the principle of causality can be\nestablished prior to all specific experience, only if we can\npredetermine experience as necessarily conforming to it, are empirical\narguments valid at all. Hume\u2019s enquiry thus directly leads to the later,\nno less than to the earlier form of Kant\u2019s epoch-making question.[1784]\nIn its earlier formulation it referred only to _a priori_ judgments; in\nits wider application it was found to arise with equal cogency in\nconnection with empirical judgments. And as thus extended, it generated\nthe problem: How is sense-experience, regarded as a form of _knowledge_,\npossible at all?[1785] By showing that the principle of causality has\nneither intuitive nor demonstrative validity, Hume cuts the ground from\nunder the rationalists; by showing that sense-experience cannot by\nitself yield conclusions which are objectively valid, he at the same\ntime destroys the empiricist position. In this latter contention Kant\nstands in complete agreement with Hume. That the sensuously given is\nincapable of grounding even probable inferences, is a fundamental\npresupposition (never discussed, but always explicitly assumed) of the\nCritical philosophy. It was by challenging the sufficiency of Hume\u2019s\nother line of argument, that which is directed against the rationalists,\nthat Kant discovered a way of escape from the sceptical dilemma. The\nconditions of experience can be proved by a transcendental method,\nwhich, though _a priori_ in character, does not lie open to Hume\u2019s\nsceptical objections. Each single experience involves rational\nprinciples, and consequently even a single empirical observation may\nsuffice to justify an inductive inference. Experience conforms to the\ndemands of pure _a priori_ thought; and can legitimately be construed in\naccordance with them.\n\nWe may now pass to the philosophy in which Kant was educated. It gave to\nhis thinking that rationalist trend, to which, in spite of all\ncounter-influences, he never ceased to remain true.[1786] It also\ncontributed to his philosophy several of its constructive principles.\nOnly two rationalist systems need be considered, those of Leibniz and of\nWolff. Kant, by his own admission,[1787] had been baffled in his\nattempts (probably not very persevering) to master Spinoza\u2019s philosophy.\nIt was with Wolff\u2019s system that he was most familiar; but both directly\nand indirectly, both in his early years and in the \u2019seventies, the\nincomparably deeper teaching of Leibniz must have exercised upon him a\nprofoundly formative influence. In defining the points of agreement and\nof difference between Hume and Leibniz,[1788] we have already outlined\nLeibniz\u2019s general view of the nature and powers of pure thought, and may\ntherefore at once proceed to the relevant detail of his main tenets.\n\nUpon two fundamental points Leibniz stands in opposition to Spinoza. He\nseeks to maintain the reality of the contingent or accidental. These\nterms are indeed, as he conceives them, synonymous with the actual.\nNecessity rules only in the sphere of the possible. Contingency or\nfreedom is the differentiating characteristic of the real. This point of\nview is bound up with his second contention, namely, that the real is a\nkingdom of ends. It is through divine choice of the best among the\npossible worlds that the actual present order has arisen. There are thus\ntwo principles which determine the real: the principle of contradiction\nwhich legislates with absolute universality, and the principle of the\nbest, or, otherwise formulated, of sufficient reason, which\ndifferentiates reality from truth, limiting thought, in order that,\nwithout violating logic, it may freely satisfy the moral needs. Leibniz\nthus vindicates against Spinoza the reality of freedom and the existence\nof ends.\n\nThough Leibniz agrees with Spinoza that the philosophically perfect\nmethod would be to start from an adequate concept of the Divine Being,\nand to deduce from His attributes the whole nature of finite reality, he\nregards our concept of God as being too imperfect to allow of such\nprocedure. We are compelled to resort to experience, and by analysis to\nsearch out the various concepts which it involves. By the study of these\nconcepts and their interrelations, we determine, in obedience to the law\nof contradiction, the nature of the possible. The real, in\ncontradistinction from the possible, involves, however, the notion of\nends. The existence of these ends can never be determined by logical,\nbut only by moral considerations. The chief problem of philosophical\nmethod is, therefore, to discover the exact relation in which the\nlogical and the teleological, the necessary and the contingent, stand to\none another.\n\nThe absence of contradiction is in itself a sufficient guarantee of\npossibility, _i.e._ even of the possibility of real existence. How very\nfar Leibniz is willing to go on this line is shown by his acceptance of\nthe ontological argument. The whole weight of his system rests, indeed,\nupon this proof. The notion of God is, he maintains, the sole concept\nwhich can determine itself in a purely logical manner not only as\npossible but also as real. If we are to avoid violating the principle of\ncontradiction, the _Ens perfectissimum_ must be regarded as possessing\nthe perfection of real existence. And since God is perfect in moral as\nin all other attributes, His actions must be in conformity with moral\ndemands. In creating the natural order God must therefore have chosen\nthat combination of possibilities which constitutes the best of all\npossible worlds. By means of this conceptual bridge we are enabled to\npass by pure _a priori_ thinking from the logically possible to the\nfactually real.\n\nPure logical thinking is thus an instrument whereby ultimate reality can\nbe defined in a valid manner. _Pure thought is speculative and\nmetaphysical in its very essence._ It uncovers to us what no experience\ncan reveal, the wider universe which exists eternally in the mind of\nGod. Every concept (whether mathematical, dynamical, or moral), provided\nonly that it is not self-contradictory, is an eternal essence, with the\nintrinsic nature of which even God must reckon in the creation of\nthings. When, therefore, we are determining the unchanging nature of the\neternally possible, there is no necessary reference to Divine existence.\nThe purely logical criterion suffices as a test of truth. Every judgment\nwhich is made in regard to such concepts must express only what their\ncontent involves. All such judgments must be analytic in order to be\ntrue.\n\nWhen, however, we proceed from the possible to the real, that is to say,\nfrom the necessary to the contingent, the logical test is no longer\nsufficient; and only by appeal to the second principle, that of\nsufficient reason, can judgments about reality be logically justified.\nWhether or not the principle of sufficient reason is deducible, as Wolff\nsought to maintain, from the principle of contradiction, is a point of\nquite secondary importance. That is a question which does not deserve\nthe emphasis which has been laid upon it. What is chiefly important is\nthat for Leibniz, as for Wolff, both principles are principles of\nanalysis. The principle of sufficient reason is not an instrument for\ndetermining necessary relations between independent substances. The\nsufficient ground of a valid predicate must in all cases be found in the\nconcept of the subject to which it is referred. The difference between\nthe two principles lies elsewhere, namely, in the character of the\nconnection established between subject and predicate. In the one case\nthe denial of the proposition involves a direct self-contradiction. In\nthe other the opposite of the judgment is perfectly conceivable; our\nreason for asserting it is a moral (employing the term in the\neighteenth-century sense), not a logical ground. The subject is so\nconstituted, that in the choice of ends, in pursuit of the good, it must\nby its very nature so behave. The principle of sufficient reason, which\nrepresents in our finite knowledge the divine principle of the best,\ncompels us to recognise the predicate as involved in the subject--as\ninvolved through a ground which inclines without necessitating. Often\nthe analysis cannot be carried sufficiently far to enable us thus to\ntransform a judgment empirically given into one which is adequately\ngrounded. None the less, in recognising it as true, we postulate that\nthe predicate is related to the subject in this way. There are not for\nLeibniz two methods of establishing truth, sense-perception to reveal\ncontingent fact, and general reasoning to establish necessary truth. A\nproposition can be accepted as true only in so far as we can at least\n_postulate_, through absence of contradiction and through sufficient\nreason, its analytic character. It must express some form of identity.\nThe proposition, Caesar crossed the Rubicon, is given us as historical\nfact. The more complete our knowledge of Caesar and of his time, the\nfurther we can carry the analysis; and that analysis if completely\nexecuted would displace the merely factual validity of the judgment by\ninsight into its metaphysical truth. Thus experience, with its\nassertions of the here and now about particulars inexhaustibly concrete,\nsets to rational science an inexhaustible task. We can proceed in our\nanalysis indefinitely, pushing out the frontiers of thought further and\nfurther into the empirical realm. Only by the Divine Mind can the task\nbe completed, and all things seen as ordered in complete obedience to\nthe two principles of thought.\n\nLeibniz, in propounding this view, develops a genuinely original\nconception of the relation holding between appearance and reality. Only\nmonads, that is, spiritual beings, exist. Apart from the representative\nactivity of the monads there are no such existences as space and time,\nas matter and motion. The mathematical and physical sciences, in their\npresent forms, therefore, cannot be interpreted as revealing absolute\nexistences. But, if ideally developed, they would emancipate themselves\nfrom mechanical and sensuous notions; and would consist of a body of\ntruths, which, as thus perfected, would be discovered to constitute the\nvery being of thought. Pure thought or reason consists in the\napprehension of such truths. To discover and to prove them thought does\nnot require to issue out beyond itself. It creates this conceptual world\nin the very act of apprehending it; and as this realm of truth thus\nexpresses the necessary character of all thought, whether divine or\nhuman, it is universal and unchanging. Each mind apprehends the same\neternal truth; but owing to imperfection each finite being apprehends it\nwith some degree of obscurity and confusion, fragmentarily, in terms of\nsense, and so falls prey to the illusion that the self stands in\nmechanical relations to a spatial and temporal world of matter and\nmotion.\n\nLeibniz supports this doctrine by his theory of sense-experience as\noriginating spontaneously from within the individual mind. Thereby he is\nonly repeating that pure thought generates its whole content from within\nitself. Sense-experience, in its intrinsic nature, is nothing but pure\nthought. Such thought, owing to the inexhaustible wealth of its\nconceptual significance, so confuses the mind which thus generates it,\nthat only by prolonged analysis can larger and larger portions of it be\nconstrued into the conceptual judgments which have all along constituted\nits sole content. And in the process, space, time, and motion lose all\nsensuous character, appearing in their true nature as orders of relation\nwhich can be adequately apprehended only in conceptual terms. They\nremain absolutely real as objects of thought, though as sensible\nexistences they are reduced to the level of mere appearance. Such is the\nview of thought which is unfolded in Leibniz\u2019s writings, in startling\ncontrast to the naturalistic teaching of his Scotch antagonist.\n\nAs already indicated, Kant\u2019s first-hand knowledge of Leibniz\u2019s teaching\nwas very limited. He was acquainted with it chiefly through the\ninadequate channel of Wolff\u2019s somewhat commonplace exposition of its\nprinciples. But even from such a source he could derive what was most\nessential, namely, Leibniz\u2019s view of thought as absolute in its powers\nand unlimited in its claims. How closely Wolff holds to the main tenet\nof Leibniz\u2019s system appears from his definition of philosophy as \u201cthe\nscience of possible things, so far as they are possible.\u201d He thus\nretains, though without the deeper suggestiveness of Leibniz\u2019s\nspeculative insight, the view that thought precedes reality and\nlegislates for it. By the possible is not meant the existentially or\npsychologically possible, but the conceptually necessary, that which,\nprior to all existence, has objective validity, sharing in the universal\nand necessary character of thought itself.\n\nAs Riehl has very justly pointed out,[1789] Wolff\u2019s philosophy had,\nprior even to the period of Kant\u2019s earliest writings, been displaced by\nempirical, psychological enquiries and by eclectic, popular philosophy.\nOwing to the prevailing lack of thoroughness in philosophical thinking,\n\u201cProblemlosigkeit\u201d characterised the whole period. The two exclusively\nalternative views of the function of thought stood alongside one another\nwithin each of the competing systems, quite unreconciled and in their\nmutual conflict absolutely destructive of all real consistency and\nthoroughness of thought. It was Kant who restored rationalism to its\nrightful place. He reinvigorated the flaccid tone of his day by adopting\nin his writings, both early and late, the strict method of rational\nscience, and by insisting that the really crucial issues be boldly\nfaced. In essentials Kant holds to Wolff\u2019s definition of philosophy as\n\u201cthe science of possible things, so far as they are possible.\u201d As I have\njust remarked, the possible is taken in an objective sense, and the\ndefinition consequently gives expression to the view of philosophy upon\nwhich Kant so frequently insists, as lying wholly in the sphere of pure\n_a priori_ thought. Its function is to determine prior to specific\nexperience what experience must be; and obviously that is only possible\nby means of an _a priori_, purely conceptual method. His _Critique_, as\nits title indicates, is a criticism of pure reason by pure reason.\nNothing which escapes definition through pure _a priori_ thinking can\ncome within its sphere. The problem of the \u201cpossibility of experience\u201d\nis the problem of discovering the conditions which _necessarily_\ndetermine experience to be what it is. Kant, of course, radically\ntransforms the whole problem, in method of treatment as well as in\nresults, when in defining the subject-matter of enquiry he substitutes\nexperience for things absolutely existent. This modification is\nprimarily due to the influence of Hume. But the constant occurrence in\nKant\u2019s philosophy of the term \u201cpossibility\u201d marks his continued belief\nin the Idealist view of thought. Though pure thought never by itself\namounts to knowledge--therein Kant departs from the extreme rationalist\nposition--only through it is any knowledge, empirical or _a priori_,\npossible at all. Philosophy, in order to exist, must be a system of _a\npriori_ rational principles. Nothing empirical or hypothetical can find\nany place in it.[1790] Yet at the same time it is the system of the _a\npriori_ conditions only of experience, not of ultimate reality. Such is\nthe twofold relation of agreement and difference in which Kant stands to\nhis rationalist predecessors.\n\n\n\n\nINDEX\n\n\nAbsolute. _See_ Unconditioned\n\nAbsolutist aspect of human consciousness, xxx-xxxiii, liii, liv-lv, lvi-lvii,\n    270-1, 274, 282, 285-7, 331 _n._, 423 _n._\n\nActuality, 391 ff.\n\nAdamson, R., 38, 311, 314\n\nAddison, 156\n\nAdickes, E., xx-xxi, 76, 166, 200, 215 _n._, 233 _n._, 234, 304, 363,\n     376, 397, 406 _n._, 423, 439-40, 441, 464 _n._, 466, 479,\n     579 _n._, 601 _n._\n\nAffinity, objective, 224, 253-7, 266-7\n\n\u201c_Als ob_\u201d doctrine, 524, 553 ff.\n\nAnalogy, Kant\u2019s use of the term, 356-8\n\nAnalytic and synthetic judgment, xxv ff., xxxv-xxxvi, xxxviii,\n    28 ff., 37 ff., 59-60, 65;\n  existential judgment, 530-1;\n  distinction perhaps suggested by examination of ontological argument, 531.\n  _See_ Judgment\n\nAnalytic and synthetic methods, 44 ff., 111, 117 _n._\n  _See_ Transcendental method\n\n_Analytic_, distinguished from the _Dialectic_, 172-4, 438-42\n\n_Anthropologie_, Kant\u2019s, 81 _n._, 100 _n._\n\nAntinomies, lii, liii, 432, 478 ff., 519-20\n\nAppearance, Kant\u2019s views regarding, xxxvii, xlvi-xlvii, liii-liv, 18-22, 83-5,\n    120-2, 147 ff., 205 ff., 215 ff., 279-284, 293 ff., 301 ff., 312 ff., 321\n    ff., 330-1, 372-3, 404 ff., 427 ff.;\n criticism of Leibnizian view of, 143-6;\n criticism of Locke\u2019s view of, 146-7;\n ideality of, 147 ff.;\n outer and inner appearances reduce to relations, 147-8;\n appearance and illusion, 148 ff.;\n causal efficacy of appearances, 216, 217-18, 351, 373-4;\n distinction between appearance and reality based not on\n    categories of understanding but on Ideas of Reason,\n    liii-liv, 217-18, 326 _n._, 331, 390-1, 414-17, 426-31,\n    473-7, 511-12, 519-21, 541-2, 558-61\n\nApperception, and memory, 251;\n  in what sense original, xxxiv, xliii-xlv, l-lii, 260-3, 461-2, 472-7;\n  transcendental unity of, l-lii, 207 ff., 212, 250-3, 260-3, 270, 277-9, 322\n    ff., 455 ff., 473-7;\n  absent from the animal mind, xlvii-l;\n  objective unity of, 270-1, 274, 282, 285-7;\n  and inner sense, 295-8, 321 ff., 512 _n._\n  _See_ Self\n\n_A priori_, Kant\u2019s views regarding the, xxvi-xxviii, xxxiii-xxxvi,\n    lii-lv, 1-2, 39-40, 42, 54 ff.;\n  problem of _a priori_ synthetic judgment, 26 ff., 39-40, 43 ff.;\n  its validity merely _de facto_, xxxv-xxxvi, xliv, 30, 57, 118, 142, 185-6,\n    257-9, 291, 391-2, 393, 400-1, 411;\n  the faculties in which it originates, xliii-xlv, l-lii, 1-2, 50-1, 237-8,\n    263 ff., 391-2, 393, 398, 563-5;\n  semi-Critical view of the, 188-9, 232, 263-4.\n  _See_ Understanding, Reason\n\nAquinas, 73\n\nArchitectonic, xxii, 100, 184, 332-6, 340-1, 342, 343, 345, 347, 390, 392, 394,\n    419 ff., 434, 437, 440, 439-40, 454, 463, 464, 474, 479-80, 496, 498, 542,\n    563, 579\n\nAristotle, xlv, 196, 198, 390. _See_ Logic\n\nArithmetic, 32, 40-1, 65-6, 128 ff., 337-8, 347, 566\n\nAssociation, and judgment, xxxiv-xxxv, xlviii-l;\n  and consciousness, xli-xlii; rests on objective affinity, 253-7, 266-7\n\nAttributive judgment, Kant\u2019s exclusive emphasis upon, 37-8, 180-1, 197\n\nAugustine, St., 73, 110, 565\n\nAvenarius, 587 _n._\n\nAxioms, Kant\u2019s view of, 50, 127, 348, 565-7\n\n\nBacon, Francis, 4-5, 74\n\nBain, A., 86 _n._\n\nBalfour, A. J., 314\n\nBaumgarten, 192-3, 441, 522\n\n_Beantwortung der Frage: Was heisst Aufkl\u00e4rung?_ Kant\u2019s, 15\n\nBeattie, James, xxviii-xxix, xxxi _n._, 207, 582, 595, 600 _n._\n\nBeck, 80\n\nBelief, Kant\u2019s view of, lv ff., 576-7\n\nBeloselsky, F\u00fcrst von, xlix\n\nBergson, 86, 142, 359-60 _n._, 587 _n._\n\nBerkeley, xxxii, xl, xlvi, 112, 153-4, 155 ff., 272, 298 ff., 587-8, 592, 595,\n    596\n\nBorowski, 63, 156\n\nBosanquet, B., 36, 181, 197\n\nBradley, F. G., 36, 181, 197\n\nBruno, Giordano, 74\n\nB\u00fclffinger, 155\n\n\nCaird, E., xx, 1 _n._, 23, 51, 102 _n._, 114, 117, 183, 194,\n     195, 262, 296, 314, 328, 340, 357 _n._, 359 _n._, 373, 378, 399, 462, 468\n\nCampanella, 74\n\nCanon, 72, 169-70, 174, 332-3, 438, 569 ff.\n\nCassirer, E., 132\n\nCategorical imperative, xxxvi, lvi-lviii, 571 ff.\n\nCategories, distinction from generic concept, 178 ff.;\n  _de facto_ nature of the, xxxv-xxxvi, xxxviii, xliv, 30, 57, 185-6, 257-8,\n    291, 391-2, 398, 400-1, 411;\n  definition of the, 195-6, 198, 339-42, 404-5;\n  semi-Critical view of the, 188-9, 217-18, 232, 263-4;\n  merely logical forms, xxxv-xxxvi, xxxviii, 30 _n._, 32, 39, 108, 176 ff.,\n     185-6, 191, 195-196, 257-8, 290-1, 325 ff., 339-40, 398, 404-5, 409-10,\n     413-14, 467;\n  valid only for appearances, 259-60;\n  and schemata, 195-6, 311, 333, 339-342, 467 _n._;\n  metaphysical deduction of the, 183 ff., 192 ff., 287-8;\n  transcendental deduction of the, 234 ff., 287-8;\n  all categories involved in every act of consciousness, xli-xlv, liii-liv\n     199-200, 356, 368, 370, 377, 387-91;\n  have wider scope than the forms of sense, lv-lvi, 20, 25, 290-1,\n     331, 404 ff.;\n  restricted by time and space, 342, 357;\n  in relation to outer and inner experience, 311-12;\n  how far predicable of the \u2018I think,\u2019 325 ff.;\n  how far applicable to sensations, desires, etc., xlvi _n._, 275-6, 279 ff.,\n    312 ff., 384-5, 476;\n  proof of specific, 242-3, 252-3, 258-9, 287-8, 333, 343-4;\n  determinate and indeterminate application of the, 325 ff., 405 ff.;\n  may be intrinsically inapplicable to things in themselves, 290,\n     409-10, 413-14;\n  category of existence, 322, 415 _n._;\n  category of totality and Idea of the unconditioned, 199-200, 433, 451, 480,\n    529;\n  mathematical and dynamical, 198, 345-7, 510-11.\n  See _A priori_, Understanding\n\nCatharticon, 169, 174\n\nCausality, Kant influenced by Hume\u2019s teaching regarding, xxv ff.,\n     61-4, 364 ff., 593-600;\n  Kant\u2019s treatment of the principle of, 363 ff.;\n  Kant\u2019s subjectivist and phenomenalist views of, 216, 217-18,\n     318-21, 351, 373-4;\n  sensations, feelings, etc. subject to principle of, xlvi _n._, 275, 279-82,\n    312, 384-5;\n  category of, involved in consciousness of time, liii-liv, 365 ff., 377 ff.,\n    387;\n  and freedom, 492 ff.\n  _See_ Hume\n\nClarke, 140, 539, 594\n\nCohen, H., 51, 102 _n._, 195, 262, 340\n\nCoherence theory of truth, xxxvi-xxxix, 36 ff., 173 _n._;\n  criterion of truth bound up with the Ideas of Reason, 217-18, 326 _n._, 331,\n    390-1, 414-17, 426-31, 473-7, 511-12, 519-21, 541-2, 558-61\n\nConcept, Kant\u2019s generic or class view of the, 99-100, 105-7, 118-19, 126, 132-3\n    , 177-84, 338-9, 370-1, 377-84, 390-1;\n  intuition and conception, 38-42, 93-4, 105-9, 118-20, 126, 128-134, 165-6,\n    167-8, 194, 370, 390-1, 564-6;\n  construction of concepts, 41, 131-3, 338-9, 418 ff., 564-6;\n  concepts and images, 337-9;\n  Kant\u2019s doctrine of the pure concept, xxxix, 394-400, 418 ff.\n  _See_ Understanding\n\n_Concerning the Advances, made, etc._\n  See _Fortschritte_\n\nConsciousness, Kant\u2019s views regarding, xxxiv-xxxv, xxxix-xlv,\n     xliii-xlvii, l-lii;\n  and the animal mind, xlvii-l;\n  may be a resultant, xxxiv, xliii-xlv, l-lii, 261-3, 277-9, 327,\n     459-62, 473-7;\n  no immediate consciousness of mind\u2019s own activities, xliii-xlv, l-lii, 263\n    ff., 273 ff., 293, 295 ff., 322 ff.;\n  consciousness of time Kant\u2019s datum, xxxiv, 120, 241 ff., 365 ff., 381 ff.;\n  absolutist aspect of, xxx-xxxiii, liii, lvi-lvii, 270-1, 274,\n     282, 285-7, 331 _n._\n  See Apperception, Judgment\n\nContingency, assertion of, 39 ff., 55, 286-9\n\nContinuity, Kant\u2019s views regarding, 352-355, 488 ff., 509;\n  principle of, 380-1;\n  transcendental principle of, 551\n\nCopernicus, 18-19, 22-5\n\nCosmological Argument, 531 ff.\n\nCriterion of truth. _See_ Coherence theory of truth\n\nCriticism, Kant\u2019s use of term, 1, 9, 13-14, 21;\n  Age of, 15\n\n_Critique of Practical Reason_, lvi, lvii, lx, 77-8, 569 ff., 572\n\n_Critique of Judgment_, lxi, 77, 83, 97-8, 191, 265, 537, 539, 561, 569 _n._,\n    574, 575 _n._, 576, 577 _n._\n\nCrusius, xxviii, xxxii, 47\n\nCurtius, E., 336\n\n\nDeduction of categories, distinction between subjective and objective,\n     xliv _n._, 235 ff.;\n  subjective, 245 ff., 263 ff.; objective, 248 ff.;\n  metaphysical, 175 ff., 192 ff.;\n  stages in Kant\u2019s development of metaphysical, 186 ff.\n  _See_ Transcendental method of proof\n\nDeduction of Ideas, metaphysical, 426, 433 ff., 450-4, 478-80, 522-3;\n  transcendental, 426, 430, 436, 454, 552-4, 572 ff.\n  _See_ Ideas of Reason\n\nDefinition, Kant\u2019s view of, 564-5\n\nDeist, as contrasted with Theist, 541;\n  Kant\u2019s deistic interpretation of the Ideas of Reason, 418, 436, 439-40, 454,\n    473-7, 520-1, 537, 542, 575.\n  _See_ Idealist view of Reason.\n\nDemocritus, 354 _n._\n\nDemonstration, Kant\u2019s view of, 566-7\n\nDescartes, xxxi, xxxix-xliii, xlvi, 155, 157, 272-3, 279 ff., 298 ff.,\n     354 _n._, 421, 449, 583-7, 589-90, 597\n\nDesires, Kant\u2019s view of the, xlvi _n._, 276, 279-82, 312, 384-5\n\nDewey, J., 36\n\n_Dialectic_, distinguished from the _Analytic_, 172-4, 438-42;\n  the problems of the, 425 ff.;\n  development of Kant\u2019s views regarding the, 431 ff.\n\n_Dilucidatio Principiorum primorum_, etc., Kant\u2019s, 155, 299\n\nDiscipline, 170, 174, 438, 563 ff.\n\n_Dissertation_, Kant\u2019s Inaugural, xx, 26, 40, 46, 81, 86, 87, 89 ff., 96, 99,\n    101, 117, 123, 128, 131, 135, 137, 140-1, 144-5, 147, 159-60, 163-5, 185,\n    186-9, 208, 260, 263, 299, 382, 419, 427, 432, 482, 486, 489 _n._, 548\n\nDivine Existence, in relation to space and time, 159-61;\n  and intuitive understanding, 160;\n  Idea of, 434-7;\n  how far can be concretely pictured, 536-7, 541-2, 556 ff.\n  _See_ God\n\nDogmatism, as distinguished from Criticism, 9, 13-14, 21\n\n_Dreams of a Visionseer_, Kant\u2019s, 155 _n._, 299\n\nDuns Scotus, 73-4\n\n\nEberhard, Kant\u2019s reply to, 90 ff., 143 _n._\n\nEgo, transcendental. _See_ Apperception\n\nEleatics, the, 159\n\nEmotions, Kant\u2019s view of the, xlvi _n._, 276, 279-82, 312, 384-5\n\nEmpirical, relation to the _a priori_, 36 ff.;\n  problem of empirical knowledge, 39-40, 53;\n  empirical object intermediate between subjective representations and thing\n    in itself, 206 ff., 223, 270 ff., 308 ff.\n  _See_ Experience\n\n_Enquiry into the Clearness of the Principles of Natural Theology and Morals_,\n    Kant\u2019s, 15, 40, 563 ff.\n\n_Ens realissimum_, 522 ff., 529-30, 532, 534, 541-2, 556\n\nEpicurus, lix, 436, 499, 582\n\nErdmann, B., xx, xxviii _n._, 46, 142 _n._, 158, 161, 163, 200-1,\n     208 _n._, 294 _n._, 314, 373, 382 _n._, 412, 431-2, 471, 601 _n._\n\nErhardt, F., 484 _n._, 494\n\nError, _See_ Appearance, Illusion\n\nEuler, 162\n\nExistence, and the \u201cI think,\u201d 322 ff.;\n  judgment of, always synthetic, 527 ff.;\n  necessary existence, 533-7\n\nExperience, proof by reference to the possibility of, xxxvi, xxxvii-xxxviii,\n    45, 238-9, 241-3, 259-60, 344, 426, 430, 454, 552-4, 572 ff.;\n  meaning of term, 52; problem of, 57-8;\n  as datum is equivalent to consciousness of time, xxxiv, 120,\n     241 ff., 365 ff., 381 ff.\n\nExposition, Kant\u2019s use of term, 109-10\n\n\nFaith, Kant\u2019s view of, lv-lvi, lxi, 571 ff., 575-6\n\nFeeling, Kant\u2019s use of term, 82-3;\n  Kant\u2019s view of, xlvi _n._, 276, 279-82, 312, 384-5\n\nFichte, l\n\nFischer, K., 46, 75, 113-14, 140, 601 _n._\n\nForm and matter, importance of distinction between, xxxiii-xxxiv, xxxvi, 85 ff.\n\nForms of the understanding. _See_ Categories\n\n_Fortschritte, Welches sind die wirklichen_, etc., Kant\u2019s, li _n._, 59, 60, 84,\n    578 _n._, 580 _n._\n\n_Foundations of the Metaphysics of Morals_, Kant\u2019s, lviii, lix, 569, 572\n\n_Fragmente aus dem Nachlasse_, Kant\u2019s, lvii, 578\n\nFreedom of the will, problem of, 20-1, 435, 512 ff., 569-70;\n  and causality, 492 ff.;\n  transcendental and practical freedom, 497, 512-13, 517-18, 569-70, 573-4\n\n\nGalileo, 18, 583-4, 586\n\nGarve, xix, 150; Garve-Feder review, 158\n\n_Gedanken von der wahren Sch\u00e4tzung der lebendigen Kr\u00e4fte_, Kant\u2019s, 117, 161-2\n\nGeometry, the fundamental mathematical science, 96 _n._;\n  pure and applied, 111-12, 147, 349, 565-6;\n  Kant\u2019s attitude to modern, 117 ff.\n\nGeulincx, 596, 598\n\nGod, ontological proof of existence of, 527 ff.;\n  cosmological proof, 531 ff.;\n  physico-theological proof, 538 ff.;\n  problem of God\u2019s existence, 569 ff.;\n  how far an indispensable Idea of Reason, 439-40, 536-7, 541-2, 556 ff.\n\nGreen, T. H., l _n._, 23, 36\n\nGroos, K., xxviii _n._\n\n\nHamann, 157, 539-40 _n._;\n  describes Kant as \u201ca Prussian Hume,\u201d 305\n\nHegel, xxxvii, xlv, l, 36, 190, 194, 274, 554 _n._\n\nHerbart, 86 _n._, 124\n\nHerz, Marcus, xxii-xxiii, xxix, xlix, 6, 26, 28, 46, 51, 114 _n._, 138, 187,\n    189, 198, 206-7, 219-22, 432\n\nHicks, G. Dawes, 415 _n._\n\nHobbes, 593\n\nH\u00f6ffding, H., 23\n\nHome, Henry, 1\n\nHomogeneity, transcendental principle of, 550-1\n\nHume, date of first influence on Kant, xx, xxviii;\n  Kant\u2019s relation to, xxv-xxxiii, xxxv, xxxvii, xlvi;\n  his view of consciousness, xl-xliii;\n  anticipates Kant\u2019s phenomenalism, 21-2;\n  maintains that experience cannot prove universality or necessity, 27, 57-8;\n  shows causal axiom to be synthetic, 30-1;\n  Hume\u2019s problem a deepening of Kant\u2019s earlier problem, 46;\n  Kant\u2019s relation to, 61-4;\n  on the self, 207 _n._;\n  his subjectivism, 272-3, 284, 300;\n  Kant \u201ca Prussian Hume,\u201d 305;\n  much of Hume\u2019s teaching in regard to causality accepted by Kant, 364;\n  Kant\u2019s reply to Hume, 369-71;\n  Hume\u2019s philosophy the perfected expression of the empirical\n     and sceptical position, 421;\n  influence on Kant, 432;\n  on existential judgment, 528;\n  influence on Kant of Hume\u2019s _Dialogues on Natural Religion_, 539-540, 557,\n    567 _n._;\n  influence on Kant, 583;\n  the philosophical teaching of, 588-601;\n  influence on Kant, 606\n\nHumility, lvi, lviii-lix, 554 _n._\n\nHypotheses, and postulates, xxxvii-xxxviii, 541, 543 ff., 571 ff.;\n  how far valid in metaphysics, lxi, 9-12, 543 ff.\n\nHypothetical employment of Reason, 549-50\n\n\nIdealism, objective or Critical, 274;\n  Kant\u2019s refutations of subjective idealism. 298 ff., 462-3;\n  transcendental idealism as key to solution of the antinomies, 503 ff.\n  _See_ Phenomenalism and subjectivism\n\nIdeal of Reason, 522 ff., 536-7, 541-2, 554 _n._, 556-61\n\nIdealist view of Reason, xxxviii-xxxix, xliv, liii, 97-8, 102, 331-2, 390-1,\n    414-17, 426 ff., 433 ff., 447 ff., 473-7, 478 ff., 500-6, 511-12,\n     519-21, 547 ff., 552 ff., 558-61\n\nIdeality, of space and time, 76, 111, 116-17, 138, 147, 154, 308\n\nIdeas of Reason, Kant\u2019s sceptical and Idealist views of the, xxxviii-xxxix,\n     xliii, xliv, lii-lv, lvi ff., 330-1, 390-1, 414-17. 426 ff., 433 ff.,\n     446 ff., 473-7, 478 ff., 500-6, 511-12, 520-1, 547 ff., 558-61;\n  involved in consciousness of space and time, liii-liv, 96-8, 102 _n._, 165-6,\n    390-1;\n  Kant\u2019s deistic interpretation of the, 418, 436, 439-40, 454, 473-477, 520-1,\n    537, 575;\n  as limiting concepts, 408, 413-17, 426 ff.;\n  as regulative, xxxviii-xxxix, xliii, liii, 473-7, 500 ff., 547 ff.;\n  and categories of relation, 451-2;\n  distinction between mathematical and dynamical, 510-11;\n  Kant\u2019s criticism of Idea of unconditioned necessity, 527 ff., 533-7, 541-2;\n  metaphysical and practical validity of the Ideas, 570-6;\n  concluding comments on Kant\u2019s views of the, 558-61;\n  condition distinction between appearance and reality, liii-liv, 217-18, 326\n    _n._, 331, 391, 414-17, 426-31, 473-7, 511-12, 519-21, 541-2, 558-61.\n  _See_ Deduction of Ideas\n\nIllusion, and appearance, 148 ff.;\n  Berkeley regards objects of outer sense as, 157, 307-8;\n  inner experience not illusory, 323-4;\n  transcendental, 13, 427-9, 437, 456 ff., 480, 552, 555\n\nImagination, may be the common root of sensibility and understanding, 77, 225,\n    265;\n  productive, 224 ff., 264 ff., 337, 348, 375-6\n\nImmanent and transcendent metaphysics. _See_ Metaphysics\n\nImmortality, problem of, 569 ff.\n\nIncongruous counterparts, 161 ff.\n\nInfinitude, of space, 105 ff.;\n  of time, 124 ff.;\n  Kant\u2019s view of, 483 ff.;\n  distinction between _in infinitum_ and _in indefinitum_, 507 ff.\n\nInner Sense, xliii _n._, 148, 291 ff., 360, 464, 468-9;\n  and apperception, 321 ff.\n\nIntuition, Kant\u2019s doctrine of pure, 40 ff., 79-80, 118-20, 128 ff.,\n     167-8, 468-9;\n  intuition and conception, 38-42, 93-98, 105-9, 118-20, 126, 128-34, 165-166,\n    167-8, 194, 390-1, 564-6;\n  formal intuition and form of intuition, 109, 114-16\n\nIntuitive understanding, Kant\u2019s view of, 160, 291, 408 ff., 468 _n._ 542\n\n\nJacobi, 300\n\nJakob, xxviii _n._\n\nJames, W., 86, 277-8, 459 _n._, 461 _n._\n\nJanitsch, 155, 156\n\nJones, Sir Henry, 36\n\nJudgment, Kant\u2019s doctrine of the, xxxiv-xxxv, xxxviii, xli-xliv, xlviii-l, 177\n    ff., 192 ff., 286 ff.;\n  the fundamental activity of the understanding, xxxiv-xxxv, xxxviii, xli-xlii,\n    133, 181-2, 288, 332, 370;\n  _a priori_ and empirical, 27-8;\n  analytic and synthetic, xxv ff., 28 ff., 37 ff., 59-60;\n  judgment 7 + 5 = 12, 65;\n  relational types ignored by Kant, 37 ff.;\n  Kant\u2019s attributive view of, 37-38, 180-1, 197;\n  as assertion of contingency, 39 ff., 55, 286-9;\n  Kant\u2019s distinction between judgments of perception and judgments of\n     experience, 288-9;\n  existential, 527-31\n\nKnowing and thinking, distinction between, lv-lvi, 20, 25, 290-1, 331, 404 ff.\n  _See_ Categories\n\nKnowledge, the narrow meaning assigned to term by Kant, lv-lvi, lxi, 25\n\nKn\u00fctzen, 161\n\n\nLambert, xx, xxviii, xxxii, 74, 138, 150, 193\n\nLange, F., 23\n\n_Lectures on Metaphysics_, Kant\u2019s, 261, 275 _n._, 299, 448-9, 475 _n._\n\n_Lectures on the Philosophy of Religion_ Kant\u2019s, 261\n\nLeibniz, Kant\u2019s relation to, xxx-xxxiii, xxxv, xxxvii, xlvi, l, lvii;\n  his absolutist view of thought, xxx-xxxii;\n  anticipates Kant\u2019s phenomenalism, 21-2;\n  his rejection of empiricism, 27, 58;\n  his pre-established harmony, 28;\n  regards synthetic judgments as always empirical, 30;\n  his conceptual atomism, 38;\n  Kant probably influenced by the _Nouveaux Essais_ of, 92, 186;\n  referred to by Kant, 112;\n  Kant\u2019s relation to, 140-1;\n  Kant\u2019s criticism of his interpretation of sensibility and appearance, 143-6;\n  his view of space, 161 ff.;\n  Kant influenced by the spiritualism of, 208-9, 243, 260-1, 263;\n  his subjectivism and doctrine of _petites perceptions_, 272-3, 298-9, 306;\n  his alternative views of the reality of the material world, 298-9;\n  continuing influence of his rationalism on Kant, 394-5, 398-9, 418 ff.;\n  his view of the possible as wider than the actual, 401-2;\n  antinomies formulated by Kant from the standpoint of the Leibnizian\n     rationalism, 481 ff.;\n  Kant\u2019s formulation of the ontological argument Leibnizian, 522 ff., 556;\n  contrast between Locke and, 146-7, 421, 582;\n  on mathematical method, 592;\n  the philosophical teaching of, 601-6;\n  on the nature of sense-experience, 604-5;\n  influence on Kant, 605-6\n\nLimiting concepts, Ideas as, 408, 413-17, 426 ff.\n  _See_ Ideas of Reason\n\nLocke, xxxii, xl, xlvi, 15;\n  Kant\u2019s criticism of his view of appearance, 146-7;\n  Kant\u2019s restatement of his distinction between primary and secondary\n     qualities, 120-2, 146, 149 ff., 306;\n  subjectivism of, 272, 306;\n  on inner sense, 148, 292-3;\n  contrast between Leibniz and, 146-7, 421, 582;\n  his use of term idea rejected by Kant, 449;\n  on primary and secondary qualities, 586 _n._;\n  rationalism of, 591-2;\n  his proof of causal axiom, 594;\n  on the causal relation, 596, 598\n\nLogic, Kant\u2019s contribution to the science of, xxxvi-xxxix;\n  Kant\u2019s view of the traditional, 10, 21, 33-6, 100, 181, 183, 184-6,\n     259, 332;\n  the various kinds of, 167 ff.;\n  distinction between general and transcendental, xxxix, 170 ff.,\n     176 ff., 178\n    _n._, 181, 183, 184-5, 194-5, 196, 335\n\n_Logic_, Kant\u2019s, 1, 110, 170 ff., 180-1, 576 _n._, 577 _n._,\n     580 _n._, 581 _n._, 582\n\n_Lose Bl\u00e4tter aus Kant\u2019s Nachlass_, xx-xxi, 112 _n._, 202-3,\n     209, 211 _n._, 232-4, 261\n\nLotze, 1 _n._, 36, 181\n\n\nMach, E., 596\n\nMairan, J. J. Dortous de, 496\n\nMalebranche, xxxi, xxxii, xliii _n._, 15, 28, 47;\n  Kant\u2019s phenomenalism anticipated by, 21-2;\n  rationalism of, 590-1;\n  on the causal relation, 596-8\n\nManifolds, of appearance, 84-5;\n  empirical, 267, 274 ff.;\n  pure _a priori_, 88-90, 92 ff., 95, 96-7, 134, 142 _n._, 148 _n._,\n     171, 194-5, 226, 228-9, 267, 269-70, 289, 337, 344, 375, 385 _n._\n\nMathematics, methods of, 17-18;\n  judgments in, not all synthetic, 64;\n  principle of contradiction in mathematical reasoning, 60, 64-5, 344;\n  Kant\u2019s intuitional view of, 40-1, 65-6;\n  distinction between mathematical and philosophical knowledge, 15, 563 ff.;\n  pure and applied, 68, 111-12, 114-15, 140, 166, 566;\n  use of schemata in, 337-9.\n  _See_ Arithmetic, Geometry\n\nMatter, Kant\u2019s dynamical theory of, 354-5;\n  principle of conservation of, 361-2\n\nMeier, 441\n\nMendelssohn, Moses, xix, xxxii, 6, 11, 58, 138 _n._, 139 _n._,\n     150, 153, 160-1,\n    458-9 _n._, 467, 470-1\n\n_Metaphysical First Principles of Natural Science_, Kant\u2019s, 56 _n._, 66, 97,\n    127-8, 164-5 _n._, 312 _n._, 354 _n._, 361 _n._, 380-1,\n     384 _n._, 491, 579 _n._\n\nMetaphysics, distinction between immanent and transcendent,\n     liv-lv, 15, 19, 22,\n    26-7, 33, 50, 52, 53, 55-6, 58-9, 66-70, 244-5, 257-8,\n     545, 580-1; in disrepute, 8-9;\n  Kant professes to establish a quite final, 10, 35, 543 ff.;\n  \u201cCopernican hypothesis\u201d and, 18 ff.;\n  as natural disposition, 12-13, 68 ff.;\n  as science, 68 ff.;\n  hypotheses not valid in, 543 ff.;\n  the problems of, 569-76, 579-81\n\nMethod, the sceptical, 545-6;\n  mathematical, 563-7.\n  _See_ Analytic and Synthetic Methods\n\nMill, J. S., 86, 364-5, 377, 596\n\nMind, Kant\u2019s use of term, 81\n\n_Mistaken Subtlety of the Four Syllogistic Figures_, Kant\u2019s, 181-2\n\nModality, 391 ff.\n\n_Monadologia physica_, Kant\u2019s, 354\n\nMoral attitude, the, xxxvi, xlv, lv ff., 515-16, 571 ff.\n\nMoral belief, lvi ff., 577\n\nMoral law, consciousness of the, _de facto_, xxxvi, xlv, 572-3\n\nMotion, doctrine of, 127-9, 133;\n  Galileo\u2019s revolutionary doctrine of, 583-4\n\nM\u00fcller, Max, 75\n\n\nNatural Science, pure, 66-8;\n  and immanent metaphysics, 70. _See_ Metaphysics\n\nNature, means \u201call that is,\u201d 16\n\nNecessity, and universality, 56-7;\n  definition of, 391 ff.;\n  of thought and of existence, 402-3, 527, 533, 536;\n  limited being may exist by unconditioned, 527, 533, 536;\n  absolute necessity not purely logical, 528;\n  unconditioned, Idea of, 527 ff., 533-7, 541-2, 555, 558-61;\n  and contingency, concepts of, not applicable to things in themselves, 535;\n  relative, 541, 555, 571 ff.\n\n_Negative Quantity_, Kant\u2019s essay on, 381, 403 _n._\n\n_New Doctrine of Motion and Rest_, Kant\u2019s, 354, 381 _n._\n\nNewton, his influence on Kant, lv-lvi, 96 _n._, 140-2, 161 ff., 354 _n._;\n  Kant modifies Newton\u2019s cosmology, 539\n\nNoumenon, positive and negative conception of, 408 ff., 413.\n  _See_ Appearance\n\nNumber, schema of, 347-8. _See_ Arithmetic\n\n\nObject, Kant\u2019s use of term, 79-81, 167 _n._, 174;\n  transcendental, 203 ff.;\n  empirical, 206 ff., 223, 270 ff., 308 ff.\n\nObjective, not the opposite of the subjective, 279 ff., 313-14;\n  validity of Ideas, 558-61\n\nOccasionalism, 465, 596-7, 598\n\n_On the Radical Evil in Human Nature_, Kant\u2019s treatise, lviii, lix\n\nOntological argument, 527 ff.\n\nOpinion, Kant\u2019s use of term, 543, 576-7\n\nOrganon, 71-2, 169-70, 174\n\nOswald, xxviii _n._\n\nOuter Sense, 147, 276, 293 ff., 360\n\n\nParalogisms, 455 ff.;\n  nature of fallacy of the, 466, 470\n\nPaulsen, 46-7, 64, 373, 601 _n._\n\nPearson, K., 596\n\n_Perpetual Peace_, Kant\u2019s treatise on, lvii _n._\n\nPhenomenalism and subjectivism, xxxix ff.,\n     xlv-xlvii, 82-4, 120-2, 136-8, 138-9, 140, 150-4,\n     155-9, 223, 227, 270 ff., 312 ff., 349-51,\n     357-8, 373-4, 407 _n._, 414-17\n\nPhenomenon, distinction between appearance and, 83.\n  _See_ Appearance\n\nPhilosophy, causes of failure of, 59;\n  Kant reinterprets its function and aims, lvi, 571-6, 577-8;\n  the domains of, 579-81;\n  Kant\u2019s view of history of, 582\n\nPhysico-theological argument, 538 ff.\n\nPhysics, method of, 17-18;\n  Kant\u2019s views regarding, 354-5, 361-2, 379-81\n\nPistorius, 305, 307-8, 323, 467\n\nPlato, xlv, 47, 158, 301, 390, 436, 496, 582\n\nPope, 156\n\nPossibility, Kant\u2019s definition of, 391 ff.\n\nPostulates. _See_ Hypotheses\n\nPractical employment of Reason, lvi-lix, 569 ff.\n\nPragmatic belief, lvi, 577\n\nPrantl, 73\n\nPre-established harmony, 28, 47, 114, 141-2, 267-8, 290, 465, 590\n\nPriestley, J., xxviii _n._, 11, 567 _n._\n\nPrimary and secondary qualities, 82, 120-2, 146, 149 ff., 306\n\nPrinciples never self-evident, xxvi-xxviii,\n  xxxv-xxxviii, 36 ff., 53, 185-6, 340.\n  See _A priori_\n\nProbabilities, inference from. _See_ Hypotheses\n\n_Prolegomena to Every Future Metaphysics_, Kant\u2019s, xxv,\n     xxviii _n._, xxix _n._,\n    12, 13, 46, 47, 49, 59-60, 61-4, 65, 66-7, 68 ff.,\n     80, 84, 91, 106, 109-11, 116, 121, 129, 146, 149,\n     152, 153, 155, 156, 158, 159, 161, 163, 165, 178-9,\n     184, 188, 234, 288-9, 299, 300-1, 305-8, 346,\n     361 _n._, 376-7\n\nPsychology, Kant\u2019s views regarding, xliii-xlvii, 50-1,\n     235 ff., 263 ff., 269-270, 311 _n._, 312 _n._, 384-5,\n     455 ff., 473-7, 580-1\n\nPure, Kant\u2019s use of term, 1-2, 54-6, 64\n\n\nQuality, and intensive magnitude, 352.\n  _See_ Primary and secondary qualities\n\n\nRationalism, Kant\u2019s type of, xxxv-xxxvi, 257-8;\n  relation to the rationalism of Leibniz, 418 ff.\n  See _A priori_\n\nReason, meanings of the term, liii-lv, lvi, 2-3, 71, 426 ff., 520-1, 558-61;\n  ineradicably metaphysical, liii-lv, lvi, 8;\n  condition of free actions, 515-16;\n  as practical, lvi ff., 515-17;\n  as causing antinomy, liii, 519 ff.;\n  hypothetical employment of, 549-50;\n  Ideal of, 556 ff.\n  _See_ Idealist and sceptical views of Reason, Ideas of Reason\n\n_Rechtslehre_, Kant\u2019s, 190\n\nReciprocity, category of, 197, 381 ff., 434-5, 439-40, 451-4\n\n_Reflexionen Kants zur Kritik der reinen Vernunft_, xx,\n     xxiii, xlix, lv, 85,\n    86, 106, 127, 182, 188 ff., 196, 197, 198, 200-1, 202-3\n    , 208, 231-2, 261, 334 _n._, 399, 433-40, 448-9, 543\n\nRegulative. _See_ Ideas of Reason\n\nReicke, xx. See _Lose Bl\u00e4tter_\n\nReimarus, 193\n\nRepresentation, Kant\u2019s use of term, 81, 104;\n  distinction between representation and its object, 135,\n     136-7, 272 ff., 308\n    ff., 317-18, 365.\n  _See_ Phenomenalism and subjectivism\n\nRepresentative perception, doctrine of, xxxix-xliii, xlvi, 272 ff., 298 ff.,\n    585-8\n\nRiehl, A., xliv _n._, 46, 51, 88, 102 _n._, 195, 303-4,\n     317-18, 340-1, 342, 357-8, 372 _n._, 373, 601 _n._, 605\n\nRousseau, lvii, lviii-lix, 436, 567 _n._, 578\n\nRule, two kinds of, 372\n\nRussell, B., 491-2, 568 _n._\n\n\nSceptical method, Kant\u2019s, 481, 545-6\n\nSceptical view of Reason, 481, 500-3, 511-12, 519-21, 528-9, 533-7, 541-2, 547\n    ff., 558-61\n\nScepticism, 9, 13-15, 545-6, 567\n\nSchematism, 195-6, 265-7, 289, 311, 333, 334 ff., 467 _n._;\n  and images, 337-9\n\nSchopenhauer, 75, 197, 315-16, 365-7, 377-9, 387-9, 407 _n._, 482-3 _n._, 493\n    _n._, 495 _n._\n\nSchulze, Johann, 129-31, 138, 198-9, 480 _n._\n\nSch\u00fctz, 153\n\nSegner, 66\n\nSelf, Kant\u2019s semi-Critical, spiritualist view of the, l,\n     207-9, 212, 243, 260-3, 327-8, 473-7, 515;\n  may not be an ultimate form of existence, l-lii, 260-3,\n     277-9, 327, 459-62,\n    473-7;\n  Idea of the, 439-40, 455-62, 471, 472-7, 554;\n  Kant\u2019s view of nature and destiny of, 472-7.\n  _See_ Apperception, Soul\n\nSelf-consciousness. _See_ Apperception, Consciousness\n\nSelf-evidence, Kant\u2019s rejection of, xxvi-xxviii, xxxv-xxxviii,\n     36, 53, 118, 142, 185-6, 563-4, 565-6.\n  See _A priori_\n\nSensation, Kant\u2019s views of, 81-2, 84-8, 274-7, 349-52;\n  non-spatial, 85-8, 100-1, 105;\n  required for determining actuality, 391 ff.;\n  sensations, feelings, etc., subject to law of causality,\n     xlvi _n._, 275, 279-82, 311-12, 313-14, 384-5\n\nSensibility, may have a common root with understanding, 77;\n  definition of, 81, 167-8;\n  as a limitation, 116;\n  criticism of Leibniz\u2019s view of, 143-6;\n  Kant\u2019s view of, 274-7\n\n_Seven Small Papers_, Kant\u2019s, 298\n\nSidgwick, H., 314\n\nSigsbee, R. A., 11\n\nSigwart, 36, 181, 197\n\nSimultaneity. _See_ Time\n\nSoul, and body, Kant\u2019s view of their relation, 275-6,\n     279-84, 312 ff., 384-5,\n    464-6, 467, 471, 476.\n  _See_ Apperception, Self\n\nSpace, Kant\u2019s views of, xxxv-xxxvi, lii, 85 ff., 188;\n  involves an Idea of Reason, liii-liv, 96-8, 102 _n._, 165-6, 390-1;\n  metaphysical exposition of, 99 ff., 109-10, 112 ff., 134 ff.;\n  transcendental exposition of, 109 ff., 344-5;\n  not a property of things in themselves, 112 ff.;\n  is the form of outer sense, 114-16;\n  transcendental ideality of, 76, 116-17;\n  uniform for all human beings, 116-18, 120, 241-2, 257;\n  possibility of other spaces, 117 ff.;\n  criticism of Newtonian and Leibnizian views of, 140-2;\n  merely _de facto_ character of, 57, 118, 142, 185-6, 257;\n  as _Unding_, 154;\n  in relation to Divine Existence, 159-61;\n  and incongruous counterparts, 161 ff.;\n  involved in consciousness of time, 309 ff., 384-6, 390-1;\n  ignored by Kant in doctrine of schematism, 341, 348, 360;\n  involves category of reciprocity, liii-liv, 385-7, 390-1;\n  and antinomy, 480 ff.\n  _See_ Geometry\n\nSpecification, transcendental principle of, 501-2\n\nSpencer, Herbert, 87, 584 _n._, 596\n\nSpinoza, 74, 273 _n._, 440, 587, 601-2\n\nStadler, 197, 378-9, 389 _n._\n\nStirling, J. Hutchison, 23, 75, 366 _n._, 377\n\nStout, G. F., 87, 367 _n._, 387\n\nSubconscious, the, Kant\u2019s view of, 263 ff., 273-4\n\nSubjectivism, in Kant\u2019s doctrine of the transcendental\n     object, xxxix ff., xlv-xlvii, 206 ff., 217-18.\n  _See_ Phenomenalism and subjectivism, Idealism\n\nSubstance and attribute, category of, 362-3\n\nSulzer, xxviii _n._\n\nSummum Bonum, 575, 577\n\nSwedenborg, 155 _n._, 158 _n._, 299\n\nSwift, Benjamin, 74\n\nSynthetic, problem of knowledge _a priori_ and, xxv ff.,\n     xxxv-xxxvi, xxxviii,\n    28 ff., 37 ff., 59 ff.;\n  knowledge from mere concepts, 64;\n  decomposing synthesis, 95;\n  ambiguities in Kant\u2019s formulation of problem of _a priori_\n     synthetic judgments, 43 ff.;\n  processes, xliii-xlv, l-lii, 245-8, 261-2, 263 ff.,\n     277-8, 293, 295 ff., 322,\n    327 ff.\n  _See_ Analytic and synthetic judgments, and methods\n\nSystem of pure reason, 71-3, 579-80\n\n\nTeleological argument, 536-7, 538-42, 556-8\n\nTerrasson, Jean, 15\n\nTetens, 82, 148, 294 _n._, 475\n\nThales, 18\n\nTheist, as contrasted with deist, 541\n\nThings in themselves, Kant\u2019s first use of phrase, 112 _n._;\n  transcendental object equivalent to thing in itself, 204 ff.\n  _See_ Appearance\n\nThinking, discursive and creative. _See_ Understanding\n\nThomasius, 193\n\nTime, consciousness of, Kant\u2019s datum, xxxiv, 120, 241-2;\n  metaphysical exposition of, 123 ff.;\n  as infinite, 125;\n  transcendental exposition of, 126 ff., 344-5;\n  as form of inner sense, 134-5, 293 ff.; axioms of, 127;\n  not a determination of outer appearance, 134 ff.;\n  merely _de facto_ character of, xxxv-xxxvi, lii, 142, 565-7;\n  simultaneity not a mode of, 135 ff., 356, 358-9;\n  and simultaneous apprehension, 135-6, 348, 358-9, 367-8, 371-2, 381-2;\n  and reality of inner changes, 138-40;\n  transcendental ideality of, 76, 138;\n  Kant\u2019s view of, not a mere hypothesis, 147;\n  space involved in consciousness of, 134-6, 309 ff., 341, 347-8;\n  subjective and objective order of, 358 ff., 365 ff., 381 ff.;\n  time relations determined by the given, 34-5, 267-8, 367, 370, 371-2, 377;\n  does not itself change, 142, 359-60;\n  category of causality involved in consciousness of, liii-liv, 365 ff., 377\n    ff., 387;\n  cannot be experienced in and by itself, 375-6;\n  category of reciprocity involved in consciousness of, 381-91;\n  Idea of Reason involved in consciousness of, liii-liv, 96-8, 390-1;\n  infinitude and infinite divisibility of, 390-1, 481, 483 ff.\n\nTotality. _See_ Unconditioned\n\nTranscendent. _See_ Transcendental and Metaphysics\n\nTranscendental, meaning of term, 73-6, 116-17, 302; illusion, 13, 427-9, 552,\n    555;\n  method of proof, xxxv, xxxvii-xxxviii, 45, 238-9, 241-3, 259-60, 344, 568,\n    572 ff.;\n  ideality of space and time, 76, 116-17, 138;\n  exposition of space and time, 109 ff., 126 ff., 344-5;\n  object, Kant\u2019s doctrine of, xlvi _n._, 203, 204 ff., 322, 328, 371-3, 406-7,\n    412, 414, 415, 513-14, 518;\n  unity of apperception, Kant\u2019s pre-Critical view of, 207 ff., 212;\n  unity of apperception, Kant\u2019s doctrine of, l-lii, 250-3, 260-3, 270, 277-9,\n    322 ff., 455 ff., 473-7;\n  psychology, xliii-xlvii, l-lii, 50-1, 235 ff., 253, 263 ff.; Ideal, 522 ff.;\n  principles of Reason, 550-1;\n  illusion, 13, 427-9, 437, 456 ff., 480, 552, 555.\n  _See_ Deduction of Categories and of Ideas\n\n_Transition from the Metaphysical First Principles of Natural\n     Science to Physics_, Kant\u2019s treatise, 275 _n._, 283 _n._,\n     482 _n._, 514 _n._\n\nTrendelenburg, 113-14, 140\n\nTruth. _See_ Coherence theory of truth\n\n\n_Ueber das Organ der Seele_, Kant\u2019s, 81 _n._, 275 _n._\n\n_Ueber eine Entdeckung_, etc., Kant\u2019s reply to Eberhard, 90 ff., 143 _n._\n\n_Ueber Philosophie \u00fcberhaupt_, Kant\u2019s, 83 _n._, 128\n\nUlrichs, 467, 471\n\nUnconditioned, Idea of the, its relation to category of totality,\n     199-200, 433-4, 451, 480, 529, 559-60;\n  our awareness of the conditioned presupposes the, 416-17, 429 ff.;\n  in connection with Kant\u2019s view of the self, 473-7;\n  Kant\u2019s criticism of the Idea of the, 498, 527 ff., 533-7,\n     541-2, 555, 558-61.\n  _See_ Idealist and Sceptical views of Reason\n\nUnderstanding, and Reason, lii-lv, 2, 52;\n  defined, 81;\n  may have common root with sensibility, 77;\n  distinction between its discursive and its originative\n     activities, 172, 176\n    ff., 182-3, 263 ff., 277-8, 334-5, 370, 377;\n  viewed by Kant as a unity, 174 ff., 185-6;\n  its primary function, xxxiv-xxxv, xxxviii, xli-xlii, 93-94,\n     133, 181-2, 288-9, 332, 370, 377;\n  as intuitive, 160, 291, 408 ff., 468 _n._, 542.\n  _See_ Concept\n\n_Universal Natural History and Theory of the Heavens_, Kant\u2019s, 539\n\n_Untersuchung \u00fcber die Deutlichkeit der Grunds\u00e4tze_, Kant\u2019s, 110 _n._, 131\n\n\nVaihinger, Hans, xx, xxv, xxviii _n._, xliv _n._, 2, 13, 23, 43,\n     45 ff., 52,\n    53, 59, 60, 64, 65, 66, 81, 87, 104, 105, 109, 112, 113, 117,\n    127, 130, 139, 140, 143, 147, 148, 156, 161, 162, 202 ff.,\n     261, 268-9, 298-9, 301, 314-315, 579 _n._, 601 _n._\n\nValue, problems of, lvi, lx-lxi\n\nVoid, Kant\u2019s doctrine of the, 354-5\n\nVoltaire, xxxi, 436, 539\n\n_Von dem ersten Grunde des Unterschiedes der Gegenden im Raume_, Kant\u2019s, 40,\n    86 _n._, 140, 162\n\n\nWatson, J., 1 _n._, 23 _n._, 75, 102 _n._, 117, 183, 195, 196, 198, 262, 328,\n    462, 468 _n._, 564\n\nWindelband, 46\n\nWolff, 192-3, 272, 436, 440, 522, 579, 587, 601-6\n\n\nZedlitz, Freiherr von, 6-7\n\nTHE END\n\n_Printed by_ R. & R. CLARK, LIMITED, _Edinburgh_.\n\n       *       *       *       *       *\n\nFOOTNOTES:\n\n[1] Headings not in Kant\u2019s Table of Contents are printed in italics.\n\n[2] _W._ x. p. 323.\n\n[3] _W._ x. p. 316.\n\n[4] Cf. Kant\u2019s letter to Lambert, September 2, 1770: _W._ x. p. 93.\n\n[5] Embodied in his edition of the _Kritik_ (1889).\n\n[6] From letter to Marcus Herz, June 7, 1777: _W._ x. pp. 116-17.\n\n[7] From letter to Marcus Herz, February 21, 1772: _W._ x. p. 127.\n\n[8] _Reflexionen_ ii. 5.\n\n[9] These passages are by no means unambiguous, and are commented upon\nbelow, p. 61 ff.\n\n[10] For justification of this interpretation of Hume I must refer the\nreader to my articles on \u201cThe Naturalism of Hume\u201d in _Mind_, vol. xiv.\nN.S. pp. 149-73, 335-47.\n\n[11] To this fact Kant himself draws attention: \u201cBut the perpetual hard\nfate of metaphysics would not allow Hume to be understood. We cannot\nwithout a certain sense of pain consider how utterly his opponents,\nReid, Oswald, Beattie, and even Priestley, missed the point of the\nproblem. For while they were ever assuming as conceded what he doubted,\nand demonstrating with eagerness and often with arrogance what he never\nthought of disputing, they so overlooked his inclination towards a\nbetter state of things, that everything remained undisturbed in its old\ncondition.\u201d--_Prolegomena_, p. 6; Mahaffy and Bernard\u2019s trans. p. 5.\n\n[12] Sulzer\u2019s translation of Hume\u2019s _Essays_ (including the _Enquiries_)\nappeared in 1754-56.\n\n[13] The word which Kant uses is _Erinnerung_ (cf. below, p. xxix, _n._\n4). There are two main reasons for believing that Kant had not himself\nread the _Treatise_. He was imperfectly acquainted with the English\nlanguage, and there was no existing German translation. (Jakob\u2019s\ntranslation did not appear till 1790-91. On Kant\u2019s knowledge of English,\ncf. Erdmann: _Archiv f\u00fcr Geschichte der Philosophie_, Bd. i. (1888) pp.\n62 ff., 216 ff.; and K. Groos: _Kant-Studien_, Bd. v. (1900) p. 177 ff.:\nand below, p. 156.) And, secondly, Kant\u2019s statements reveal his entire\nignorance of Hume\u2019s view of mathematical science as given in the\n_Treatise_.\n\n[14] Cf. Vaihinger, _Commentary_, i. p. 344 ff. Beattie does, indeed,\nrefer to Hume\u2019s view of mathematical science as given in the _Treatise_,\nbut in so indirect and casual a manner that Kant could not possibly\ngather from the reference any notion of what that treatment was. Cf.\nBeattie\u2019s _Essay on the Nature and Immutability of Truth_ (sixth\nedition), pp. 138, 142, 269.\n\n[15] These Hume had himself pointed out both in the _Treatise_ and in\nthe _Enquiry_; and because of them he rejects scepticism as a feasible\nphilosophy of life. Kant\u2019s statement above quoted that Hume\u2019s critics\n(among whom Beattie is cited) \u201cwere ever assuming what Hume doubted, and\ndemonstrating with eagerness and _often with arrogance_ what he never\nthought of disputing,\u201d undoubtedly refer in a quite especial degree to\nBeattie.\n\n[16] _Werke_, x. p. 123 ff. It is dated February 21, 1772. Cf. below,\npp. 219-20.\n\n[17] In _Prolegomena_, p. 6 (above quoted, p. xxviii, _n._ 1), and p. 8\n(trans. p. 6): \u201cI should think Hume might fairly have laid as much claim\nto sound sense as Beattie, and besides to a critical understanding (such\nas the latter did not possess).\u201d\n\n[18] Cf. _Prolegomena_, p. 8: \u201cI honestly confess that my recollection\nof David Hume\u2019s teaching (_die Erinnerung des David Hume_) was the very\nthing which many years ago [Kant is writing in 1783] first interrupted\nmy dogmatic slumber, and gave my investigations in the field of\nspeculative philosophy quite a new direction.\u201d Kant\u2019s employment of the\nterm _Erinnerung_ may perhaps be interpreted in view of the indirect\nsource of his knowledge of Hume\u2019s main position. He would bring to his\nreading of Beattie\u2019s quotations the memory of Hume\u2019s other sceptical\ndoctrines as expounded in the _Enquiry_.\n\n[19] Kant, it should be noted, classifies philosophies as either\ndogmatic (= rationalistic) or sceptical. Empiricism he regards as a form\nof scepticism.\n\n[20] Quoted by Beattie (_op. cit._, sixth edition, p. 295), who, however\nincapable of appreciating the force of Hume\u2019s arguments, was at least\nawake to certain of their ultimate consequences.\n\n[21] For a more detailed statement of Kant\u2019s relation to his\nphilosophical predecessors, cf. below, Appendix B, p. 583 ff.\n\n[22] The term \u201crecognition\u201d is employed by Kant in its widest sense, as\ncovering, for instance, recognition of the past as past, or of an object\nas being a certain kind of object.\n\n[23] Consciousness of time, consciousness of objects in space,\nconsciousness of self, are the three modes of experience which Kant\nseeks to analyse. They are found to be inseparable from one another and\nin their union to constitute a form of conscious experience that is\nequivalent to an act of judgment--_i.e._ to be a form of awareness that\ninvolves relational categories and universal concepts.\n\n[24] As we have noted (above, pp. xxvi-xxvii), it was Hume\u2019s insistence\nupon the synthetic, non-self-evident character of the causal axiom that\nawakened Kant from his dogmatic slumber. Cf. below, pp. 61 ff., 593 ff.\n\n[25] Cf. below, pp. lvi ff., 571 ff.\n\n[26] Cf. below, pp. 36-7.\n\n[27] Cf. below, p. 543 ff.\n\n[28] Cf. below, pp. liii-iv.\n\n[29] Cf. below, pp. 45, 238-43.\n\n[30] Cf. below, pp. 33-6, 181, 183-6.\n\n[31] Cf. below, pp. 33-42, 394-5, 398.\n\n[32] With the sole exception of Malebranche, who on this point\nanticipated Kant.\n\n[33] This is the position that Kant endeavours to expound in the very\nunsatisfactory form of a doctrine of \u201cinner sense.\u201d Cf. below, pp. l-ii,\n291 ff.\n\n[34] This was Kant\u2019s chief reason for omitting the so-called \u201csubjective\ndeduction of the categories\u201d from the second edition. The teaching of\nthe subjective deduction is, however, preserved in almost unmodified\nform throughout the _Critique_ as a whole, and its \u201ctranscendental\npsychology\u201d forms, as I shall try to show, an essential part of Kant\u2019s\ncentral teaching. In this matter I find myself in agreement with\nVaihinger, and in complete disagreement with Riehl and the majority of\nthe neo-Kantians. The neo-Kantian attempt to treat epistemology in\nindependence of all psychological considerations is bound to lead to\nvery different conclusions from those which Kant himself reached. Cf.\nbelow, pp. 237 ff., 263-70.\n\n[35] This subjectivism finds expression in Kant\u2019s doctrine of the\n\u201ctranscendental object\u201d which, as I shall try to prove, is a doctrine of\nearly date and only semi-Critical. That doctrine is especially prominent\nin the section on the Antinomies. See below p. 204 ff.\n\n[36] Cf. pp. 270 ff., 298 ff., 308-21, 373-4, 414-17.\n\n[37] That this statement holds of feelings and desires, and therefore of\nall the emotions, as well as of our sense-contents, is emphasised by\nKant in the _Critique of Practical Reason_. Cf. below, pp. 276, 279-80,\n312, 384-5.\n\n[38] The connection of this teaching with Kant\u2019s theory of consciousness\nmay be noted. If consciousness in all its forms, however primitive, is\n_already_ awareness of meaning, its only possible task is to define,\nmodify, reconstruct, and develop such meaning, never to obtain for bare\ncontents or existences objective or other significance. Cf. above, pp.\nxli-ii, xliv.\n\n[39] _Reflexionen zur Anthropologie_, 207.\n\n[40] In sketch of a letter (summer 1792) to F\u00fcrst von Beloselsky (_W._\nxi. p. 331).\n\n[41] May 26, 1789 (_W._ xi. p. 52).\n\n[42] That Kant has not developed a terminology really adequate to the\nstatement of his meaning, is shown by a parenthesis which I have omitted\nfrom the above quotation.\n\n[43] This interpretation of Kant appears in a very crude form in James\u2019s\nreferences to Kant in his _Principles of Psychology_. It appears in a\nmore subtle form in Lotze and Green. Caird and Watson, on the other\nhand, have carefully guarded themselves against this view of Kant\u2019s\nteaching, and as I have maintained (pp. xliii-v), lie open to criticism\nonly in so far as they tend to ignore those aspects of Kant\u2019s teaching\nwhich cannot be stated in terms of logical implication.\n\n[44] It may be objected that this is virtually what Kant is doing when\nhe postulates synthetic activities as the source of the categories. Kant\nwould probably have replied that he has not attempted to define these\nactivities save to the extent that is absolutely demanded by the known\ncharacter of their products, and that he is willing to admit that many\ndifferent explanations of their nature are possible. They _may_ be due\nto some kind of personal or spiritual agency, but also they may not. On\nthe whole question of the legitimacy of Kant\u2019s general method of\nprocedure, cf. below, pp. 235-9, 263 ff., 273-4, 277 ff., 461-2, 473-7.\n\n[45] Cf. _Concerning the Advances made by Metaphysics since Leibniz and\nWolff_ (_Werke_ (Hartenstein), viii. 530-1): \u201cI am conscious to myself\nof myself--this is a thought which contains a twofold I, the I as\nsubject and the I as object. How it should be possible that I, the I\nthat thinks, should be an object ... to myself, and so should be able to\ndistinguish myself from myself, it is altogether beyond our powers to\nexplain. It is, however, an undoubted fact ... and has as a consequence\nthe complete distinguishing of us off from the whole animal kingdom,\nsince we have no ground for ascribing to animals the power to say I to\nthemselves.\u201d\n\n[46] Cf. above, p. xxxiv; below, pp. 250-3, 260-3, 285-6.\n\n[47] Cf. A 651 = B 679: \u201cThe law of Reason, which requires us to seek\nfor this unity, is a necessary law, as without it we should have no\nReason at all, and without Reason no coherent employment of the\nunderstanding, and in the absence of this no sufficient criterion of\nempirical truth.\u201d Cf. also below, pp. 390-1, 414-17, 429-31, 519-21,\n558-61.\n\n[48] Regarding a further complication, due to the fact that the\n_Dialectic_ was written before the teaching of the _Analytic_ was\nproperly matured, cf. above, p. xxiv.\n\n[49] Cf. below, pp. 331, 390-1, 414-17.\n\n[50] Cf. below, pp. 22, 33, 56, 66 ff.\n\n[51] _Reflexionen_ (B. Erdmann\u2019s edition) ii. 204.\n\n[52] For an alternative and perhaps more adequate method of describing\nKant\u2019s general position, cf. below, p. 571 ff.\n\n[53] Above, pp. xxxviii-ix, xlii, xliv.\n\n[54] Cf. below, p. 577.\n\n[55] _Critique of Practical Reason_, W. v. p. 32; Abbott\u2019s trans. pp.\n120-1.\n\n[56] _Op. cit._ p. 86; Abbott\u2019s trans. p. 180.\n\n[57] _Fragmente aus dem Nachlasse (Werke_ (Hartenstein), viii. p. 624).\nCf. below, pp. 577-8. Kant claims for all men equality of political\nrights, and in his treatise on _Perpetual Peace_ maintains that wars are\nnot likely to cease until the republican form of government is\nuniversally adopted. He distinguishes, however, between republicanism\nand democracy. By the former he means a genuinely representative system;\nthe latter he interprets as being the (in principle) unlimited despotism\nof majority rule. Kant accordingly contends that the smaller the staff\nof the executive, and the more effective the representation of\nminorities, the more complete will be the approximation to the ideal\nconstitution. In other words, the less government we can get along with,\nthe better.\n\n[58] _On the Radical Evil in Human Nature, W._ vi. p. 20; Abbott\u2019s\ntrans. p. 326. \u201cThis opinion [that the world is constantly advancing\nfrom worse to better] is certainly not founded on experience if what is\nmeant is _moral_ good or evil (not civilisation), for the history of all\ntimes speaks too powerfully against it. Probably it is merely a\ngood-natured hypothesis ... designed to encourage us in the unwearied\ncultivation of the germ of good that perhaps lies in us....\u201d\n\n[59] _Foundations of the Metaphysics of Morals, W._ iv. p. 407; Abbott\u2019s\ntrans. p. 24.\n\n[60] _Critique of Practical Reason, W._ v. pp. 84-5; Abbott\u2019s trans. pp.\n178-9.\n\n[61] Cf. _Foundations of the Metaphysics of Morals, W._ iv. p. 463;\nAbbott\u2019s trans. p. 84: \u201cWhile we do not comprehend the practical\nunconditional necessity of the moral imperative, we yet comprehend its\n_incomprehensibility_, and this is all that can be fairly demanded of a\nphilosophy which strives to carry its principles up to the very limit of\nhuman reason.\u201d\n\n[62] _On the Radical Evil in Human Nature, W._ vi. pp. 49-50; Abbott\u2019s\ntrans. pp. 357-8.\n\n[63] Cf. Pringle-Pattison: _The Idea of God in the Light of Recent\nPhilosophy_, p. 25 ff.\n\n[64] _Einleitung_, i.\n\n[65] Henry Home, Lord Kames, published his _Elements of Criticism_ in\n1762.\n\n[66] _W._ ii. p. 311. In referring to his course in logic, Kant states\nthat he will consider the training of the power of sound judgment in\nordinary life, and adds that \u201cin the _Kritik der Vernunft_ the close\nkinship of subject-matter gives occasion for casting some glances upon\nthe _Kritik des Geschmacks, i.e._ upon _Aesthetics_.\u201d This passage\nserves to confirm the conjecture that the term _Kritik_ was borrowed\nfrom the title of Home\u2019s work.\n\n[67] For Kant\u2019s other uses of the term _pure_, cf. below, p. 55.\n\n[68] _Commentar zu Kants Kritik der reinen Vernunft_, i. pp. 117-20.\n\n[69] For a definition, less exclusively titular, and more adequate to\nthe actual scope of the _Critique_, cf. below, p. 56. Reason, when\ndistinguished from understanding, I shall hereafter print with a capital\nletter, to mark the very special sense in which it is being employed.\n\n[70] _Philosophical Works of Francis Bacon_ (edited by J. M. Robertson,\n1905), p. 247.\n\n[71] For Zedlitz\u2019s severe strictures (Dec. 1775) upon the teaching in\nK\u00f6nigsberg University, and his incidental appreciative reference to\nKant, cf. Schubert\u2019s edition of Kant\u2019s _Werke_, xi. pt. ii. pp. 59-61.\n\n[72] Cf. _W._ x. p. 207.\n\n[73] _Op. cit._ pp. 212-13.\n\n[74] Cf. _op. cit._ pp. 208-9.\n\n[75] _Op. cit._ p. 219.\n\n[76] A v.-vi.\n\n[77] A v. _n._\n\n[78] Cf. above on title, pp. 2-3.\n\n[79] Cf. below, pp. 543, 576-7.\n\n[80] A vii.-viii.\n\n[81] A xiv.\n\n[82] Cf. below, pp. 543 ff.\n\n[83] Cf. A 86 = B 118-19.\n\n[84] Morgenstunden; _Gesammelte Schriften_, 1863 edition, ii. pp. 246,\n288. Cf. below, pp. 160-1.\n\n[85] Cited by R. A. Sigsbee, _Philosophisches System Joseph Priestleys_\n(1912), p. 33.\n\n[86] A v. _n._\n\n[87] A viii.\n\n[88] _Prolegomena_, _Anhang_, Trans. of Mahaffy and Bernard, p. 147.\n\n[89] A 1.\n\n[90] B 21. Cf. _Prolegomena_, \u00a7 60 ff., and below, pp. 427-9, 552.\n\n[91] A 297-8 = B 353-5. Cf. below, pp. 427-9.\n\n[92] A iii.\n\n[93] i. p. 50.\n\n[94] P. 9.\n\n[95] This statement, as we shall find, calls for modification. Kant\u2019s\nCritical position is more correctly described as phenomenalism than as\nsubjectivism. Cf. above, pp. xlv-vii; below, p. 270 ff.\n\n[96] A 769 = B 797.\n\n[97] A 761 = B 789-90. Cf. Sections I.-III. in the _Methodology_.\n\n[98] A iii.\n\n[99] A v. _n._\n\n[100] A v. _n._\n\n[101] Cf. Kant\u2019s _Beantwortung der Frage: Was heist Aufkl\u00e4rung?_ 1784.\n\n[102] A v.\n\n[103] Cf. above, pp. 2-3.\n\n[104] Cf. above, pp. xliv-v; below, pp. 19, 33, 56, 66 ff.\n\n[105] A ix.\n\n[106] A x.-xi.\n\n[107] A xii.-xiii.\n\n[108] A xv.\n\n[109] A xv. Cf. below, pp. 66-7.\n\n[110] B xiv.\n\n[111] B ix.\n\n[112] B xi.\n\n[113] Cf. below, pp. 22-5.\n\n[114] Cf. above, p. lvi; below, p. 571 ff.\n\n[115] _Dissertation_, \u00a7 7.\n\n[116] All these assertions call for later modification and restatement.\n\n[117] B xxx.\n\n[118] B xxxii.\n\n[119] B xxxvii.\n\n[120] B xxxviii.\n\n[121] B vii.\n\n[122] B viii.\n\n[123] B xvi.\n\n[124] Cf. above, pp. xxvi-vii; below, pp. 594-5.\n\n[125] Cf. \u201cMalebranche\u2019s Theory of the Perception of Distance and\nMagnitude,\u201d in _British Journal of Psychology_ (1905), i. pp. 191-204.\n\n[126] Cf. below, pp. 143 ff., 604.\n\n[127] B xviii.-xix.\n\n[128] Cf. below, pp. 33, 56, 66 ff.\n\n[129] B xx.\n\n[130] B xxii.\n\n[131] B xvi.; B xxii. _n._\n\n[132] Watson\u2019s _The Philosophy of Kant Explained_ (p. 37) is the only\nwork in which I have found correct and unambiguous indication of the\ntrue interpretation of Kant\u2019s analogy.\n\n[133] _Prolegomena to Ethics_, bk i. ch. i. \u00a7 11.\n\n[134] _Text-Book to Kant_ (1881), p. 29.\n\n[135] _History of Materialism_, Eng. transl., ii. pp. 156, 158, 237.\n\n[136] _Geschichte der neueren Philosophie_ (1896), ii. p. 64.\n\n[137] _Hibbert Journal_, October 1910, p. 49.\n\n[138] _De Revolutionibus_, I. v.\n\n[139] _Ibid._ I. x.\n\n[140] B xxii. _n._\n\n[141] Cf. below, p. 543 ff.\n\n[142] B xxvi. Cf. above, pp. lv-vi, 20; below, pp. 290-1, 331, 342, 404\nff.\n\n[143] This restatement will continue up to p. 33. In pp. 33-43 I shall\nthen give general comment on the _Introduction_ as a whole. In p. 43 ff.\nI add the necessary detailed treatment of special points.\n\n[144] Cf. below, p. 219 ff.\n\n[145] Cf. above, p. xxv ff.; below, pp. 61 ff., 593 ff.\n\n[146] This statement is first made in the _Introduction_ to the second\nedition. It is really out of keeping with the argument of the\n_Introduction_ in either edition. Cf. below, pp. 39-40, 57, 85, 168,\n222, 245 ff. (especially pp. 278, 288).\n\n[147] This is the argument of the _Introduction_ to the second edition.\nIn the first edition Kant assumes without question the existence of the\n_a priori_. He enquires only whether it is also valid in its\nmetaphysical employment beyond the field of possible experience.\n\n[148] The argument of the first edition, though briefer, is\nsubstantially the same.\n\n[149] Quoted below, pp. 219-20.\n\n[150] Cf. below, pp. 114, 290, 590.\n\n[151] A 6 = B 10. I here follow the wording of the second edition.\n\n[152] Kant\u2019s view of the _a priori_ differs from that of Leibniz in\n_two_ respects. For Kant _a priori_ concepts are merely logical\nfunctions, _i.e._ empty; and secondly, are always synthetic. Cf. above,\npp. xxxiii-vi, 186, 195-6, 257-8, 290-1, 404 ff.\n\n[153] Cf. above, pp. xxv-vii; below, pp. 61 ff., 593 ff.\n\n[154] B 24.\n\n[155] Cf. above, pp. xliv-xlv, 22; below, pp. 52-3, 55-6, 66 ff.\n\n[156] Needless to say, this \u201cAristotelian\u201d logic, in the traditional\nform in which alone Kant was acquainted with it, diverges very widely\nfrom Aristotle\u2019s actual teaching.\n\n[157] Cf. above, pp. xxxvi-ix; below, pp. 36, 181, 184-6.\n\n[158] A vii.\n\n[159] B xxiii-iv.\n\n[160] Above, pp. xxv-vii, 26; below, p. 593 ff.\n\n[161] Cf. above, p. xxxvi ff.\n\n[162] Cf. above, pp. xxxvii-viii; below, pp. 238-42.\n\n[163] Cf. below, pp. 176 ff., 181, 191, 257.\n\n[164] A 6 = B 10.\n\n[165] Leibniz\u2019s interpretation of the judgment seems to result in an\natomism which is the conceptual counterpart of his metaphysical monadism\n(cf. Adamson, _Development of Modern Philosophy_, i. p. 77 ff.; and my\n_Studies in the Cartesian Philosophy_, p. 160 ff.; also below, p. 603).\nEach concept is regarded as having exclusive jurisdiction, so to speak,\nover a content wholly internal to itself. The various concepts are like\nsovereign states with no mediating tribunals capable of prescribing to\nthem their mutual dealings. Cf. below, pp. 394-400, 418 ff.\n\n[166] A 9 = B 13.\n\n[167] _Erste Betrachtung_, \u00a7\u00a7 2, 3; _dritte Betrachtung_, \u00a7 1.\n\n[168] Cf. below, p. 162.\n\n[169] \u00a7 12, 15 C.\n\n[170] Cf. B 15-16.\n\n[171] Cf. below, p. 128 ff., on Kant\u2019s views regarding arithmetical\nscience.\n\n[172] Cf. below, p. 117 ff., on Kant and modern geometry, and p. 128\nff., on Kant\u2019s views regarding arithmetical science.\n\n[173] Cf. below, pp. 131-3, 338-9, 418 ff.\n\n[174] That certain parts of the _Introduction_ were written at different\ndates is shown below, pp. 71-2. That other parts may be of similarly\ncomposite origin is always possible. There is, however, no sufficient\nevidence to establish this conclusion. Adickes\u2019 attempt to do so (_K._\npp. 35-7 _n._) is not convincing.\n\n[175] Cf. above, pp. xxxiii ff., 1-2, 26 ff.\n\n[176] i. pp. 317 and 450 ff.\n\n[177] i. p. 412 ff.; cf. p. 388 ff.\n\n[178] Cf. below, pp. 219-20.\n\n[179] Cf. Vaihinger, i. p. 394. Cf. above, p. 28.\n\n[180] Cf. Vaihinger, i. pp. 415-17.\n\n[181] Paulsen objects that if synthetic _a priori_ judgments are valid\nwithout explanation, they do not need it. For two reasons the objection\ndoes not hold. (_a_) Without this explanation it would be impossible to\nrepel the pretensions of transcendent metaphysics (cf. A 209 = B 254-5;\nA 283 = B 285). (_b_) This solution of the theoretical problem has also,\nas above stated, its own intrinsic interest and value. Without such\nexplanation the validity of these judgments might be granted, but could\nnot be understood. (Cf. _Prolegomena_, \u00a7\u00a7 4-5 and \u00a7 12 at the end. Cf.\nVaihinger, i. p. 394.)\n\n[182] Cf. Vaihinger, ii. p. 336. The argument of the _Analytic_, which\nis still more complicated, will be considered later.\n\n[183] Cf. A 46-9 = B 64-6. The corresponding sections of the\n_Prolegomena_, Vaihinger contends, were developed from this first\nedition passage, and the transcendental exposition of space in the\nsecond edition from the argument of the _Prolegomena_.\n\n[184] The synthetic method of argument is, as we shall see later,\nfurther extended in the _Analytic_ by being connected with the problem\nof the validity of ordinary experience. But as the mathematical sciences\nare proved to have the same conditions as--neither more nor less\nthan--the consciousness of time, this also allows of a corresponding\nextension of the analytic method. The mathematical sciences can be\nsubstituted for the _de facto_ premiss by which these conditions are\nproved.\n\n[185] Cf. above, p. 43.\n\n[186] What follows should be read along with p. 235 ff. below, in which\nthis distinction between the \u201csubjective\u201d and \u201cobjective\u201d deductions is\ndiscussed in greater detail.\n\n[187] A x-xi.\n\n[188] This is a criticism to which Cohen, Caird, and Riehl lay\nthemselves open.\n\n[189] Cf. below, pp. 219-20.\n\n[190] Cf. above, pp. 49-50.\n\n[191] Cf. Vaihinger, i. p. 405. The existing sciences can, as Vaihinger\nsays, be treated _en bloc_, whereas each of the principles of the new\nphilosophy must be separately established.\n\n[192] A 1.\n\n[193] A 1-2.\n\n[194] B 6 = A 2.\n\n[195] A 2.\n\n[196] Cf. above, pp. xxxv, 36 ff.; below, pp. 565-7.\n\n[197] A 2.\n\n[198] A 2.\n\n[199] B 1.\n\n[200] Cf. below, p. 55.\n\n[201] B 1.\n\n[202] Cf. below, p. 54.\n\n[203] B 2-3.\n\n[204] Cf. below, p. 55.\n\n[205] B 1.\n\n[206] B 1.\n\n[207] B 1.\n\n[208] Cf. below, p. 88 ff.\n\n[209] Cf. below, p. 237 ff.\n\n[210] B 1.\n\n[211] Cf. below, pp. 55-6.\n\n[212] B 2.\n\n[213] Cf. above, p. 27 _n._\n\n[214] B 2.\n\n[215] Cf. above, pp. 39 ff., 53; below, pp. 57-8, 222 ff., 241, 286-9.\n\n[216] B 2-3.\n\n[217] Cf. above, p. 53.\n\n[218] A 9-10 = B 13.\n\n[219] Cf. above, p. 39 ff., and below, pp. 286-9.\n\n[220] P. 53; cf. also pp. 1-2.\n\n[221] Cf. also above, pp. 2-3.\n\n[222] B 3.\n\n[223] Cf. _Metaphysische Anfangsgr\u00fcnde_, Hauptst\u00fcck ii. Lehrs. 8, Zus.\n2, in which elasticity and gravity are spoken of as the only universal\nproperties of matter which can be apprehended _a priori_.\n\n[224] B 3-4.\n\n[225] Cf. above, p. 27 ff.\n\n[226] B 4.\n\n[227] Cf. above, pp. xxxiii-iv, 27, 599 ff.\n\n[228] Cf. above, pp. xxxv-vi, 30; below, pp. 185-6, 257-9.\n\n[229] _Loc. cit._\n\n[230] B 5.\n\n[231] Cf. above, pp. xxx, 599 ff.\n\n[232] Cf. above, pp. 39, 54.\n\n[233] A 2 = B 6.\n\n[234] B 7.\n\n[235] Cf. _Kritik der Urtheilskraft_, \u00a7 91, _W._ v. p. 473.\n_Fortschritte_, _Werke_ (Hartenstein), viii. pp. 572-3.\n\n[236] Cf. above, pp. 22, 49-50, 52.\n\n[237] Cf. _Prolegomena_, \u00a7 40; _Fortschritte_, pp. 577-8.\n\n[238] i. p. 238.\n\n[239] P. 579.\n\n[240] A 712 ff. = B 740 ff.; cf. also _Fortschritte_, p. 522.\n\n[241] A 4 = B 8; cf. below, p. 563 ff.\n\n[242] A 4 = B 8.\n\n[243] A 5 = B 9.\n\n[244] Cf. B 18.\n\n[245] Cf. above, p. 29.\n\n[246] A 6 ff. = B 10 ff.\n\n[247] _Prolegomena_, \u00a7 2, _b_, _c_; Eng. trans, pp. 15-16. On the\nconnection of mathematical reasoning with the principle of\ncontradiction, cf. below, pp. 64-5.\n\n[248] P. 582; cf. _Logik_, \u00a7 37.\n\n[249] ii. p. 257.\n\n[250] _Prolegomena_, \u00a7 4.\n\n[251] Cf. B 290.\n\n[252] \u00a7 2, _c_.\n\n[253] B 161.\n\n[254] B 218.\n\n[255] A 9 = B 13.\n\n[256] Cf. above, pp. xxv ff., 26; below, p. 593 ff.; cf. Vaihinger, i.\np. 340 ff.\n\n[257] A 9 = B 13, B 11, B 19.\n\n[258] In A 9 = B 13, B 11, B 19.\n\n[259] Cf. Borowski\u2019s _Darstellung des Lebens und Charakters Im. Kants_\n(Hoffmann\u2019s edition, 1902), p. 252. The German translation of Hume\u2019s\n_Enquiry concerning the Human Understanding_ appeared in 1755, and Kant\nprobably made his first acquaintance with Hume through it. Cf. above, p.\nxxviii; below, p. 156.\n\n[260] Cf. below, _Appendix B_, p. 593 ff.\n\n[261] A 9 = B 13.\n\n[262] A 733 = B 761.\n\n[263] A 737 = B 764.\n\n[264] i. p. 291.\n\n[265] B 14.\n\n[266] B 14. Cf. above, pp. 59-60.\n\n[267] i. p. 294.\n\n[268] B 15.\n\n[269] B 15. Cf. above, p. 41.\n\n[270] Cf. Vaihinger, i. p. 296.\n\n[271] A 164.\n\n[272] A 164.\n\n[273] In _Prolegomena_ and in second edition.\n\n[274] B 15.\n\n[275] \u00a7 2 _c_.\n\n[276] Cf. below, p. 128 ff.\n\n[277] Cf. A 713 = B 741.\n\n[278] A 140 = B 179. Cf. below, p. 337 ff.\n\n[279] B 15.\n\n[280] B 17.\n\n[281] i. p. 304 ff.\n\n[282] \u00a7 15.\n\n[283] This latter Kant developed in his _Metaphysische Anfangsgr\u00fcnde_\n(1786).\n\n[284] Cf. A 840 = B 869. \u201cNature\u201d means, in the Kantian terminology,\n\u201call that is.\u201d\n\n[285] Cf. above, pp. xliv-v, 19, 22, 33, 52-3, 55-6.\n\n[286] \u00a7 4.\n\n[287] The propositions of pure natural science are not separately\ntreated in \u00a7 4 of the _Prolegomena_, though the subsequent argument\nimplies that this has been done. Vaihinger\u2019s inference (i. p. 310) that\na paragraph, present in Kant\u2019s manuscript, has been dropped out in the\nprocess of printing the fourth section (the section which contains the\nparagraphs transposed from the end of \u00a7 2) seems unavoidable. The\nmissing paragraph was very probably that which is here given in B 17.\n\n[288] B 18.\n\n[289] In \u00a7 4 (at end of paragraphs transposed from \u00a7 2).\n\n[290] B 19.\n\n[291] B 19.\n\n[292] B 20.\n\n[293] B 20.\n\n[294] Cf. B 17.\n\n[295] B 20.\n\n[296] B 21.\n\n[297] B 22.\n\n[298] Vaihinger\u2019s analysis (i. p. 371 ff.) is invaluable. I follow it\nthroughout.\n\n[299] When corrected as above, pp. 51-2, 66-7.\n\n[300] Cf. above, p. 38 ff.\n\n[301] By J. Erdmann (cited by Vaihinger, i. p. 371).\n\n[302] By B. Erdmann, _Kriticismus_, p. 183.\n\n[303] As above noted, pp. 66-7.\n\n[304] Above, p. 66.\n\n[305] A 11.\n\n[306] A 11 = B 24.\n\n[307] Cf. _Dissertation_, \u00a7 23: _usus logicus_--_usus realis_.\n\n[308] Cf. above, p. 2.\n\n[309] A 11 = B 24.\n\n[310] i. p. 459 ff.\n\n[311] _Entwickelungsgeschichte der Kantischen Erkenntnistheorie_, p.\n113.\n\n[312] Cf. A 795 = B 823. Cf. below, pp. 170, 174.\n\n[313] Cf. A 796 = B 824.\n\n[314] Cf. Vaihinger, i. pp. 461-2 for the very varied meanings in which\nKant \u201ccapriciously\u201d employs the terms _Organon_, _Canon_, _Doctrine_,\nand _Discipline_.\n\n[315] A 11 = B 25.\n\n[316] _Erkl\u00e4rung in Beziehung auf Fichte\u2019s Wissenschaftslehre_ (1799),\n_Werke_ (Hartenstein), viii. p. 600.\n\n[317] B 25.\n\n[318] Cf. A xv.\n\n[319] Cf. B xxiv.\n\n[320] A 11 = B 25.\n\n[321] _De vera religione_, 72; _De civitate Dei_, viii. 6. Cited by\nEisler, _W\u00f6rterbuch_, p. 1521.\n\n[322] Cf. Prantl, _Geschichte der Logik im Abendlande_, iii. pp. 114,\n244-5.\n\n[323] _Ethica_ (Vloten and Land), ii. prop. xl. schol. 1.\n\n[324] _Principles of Human Knowledge_, cxviii. The above citations are\nfrom Eisler, _loc. cit._ pp. 1524-5. I have also myself come upon the\nterm in Swift\u2019s _Gulliver\u2019s Travels_ (Dent, 1897, p. 166): \u201cAnd as to\n\u2018ideas, entities, abstractions, and transcendentals,\u2019 I could never\ndrive the least conception into their heads.\u201d\n\n[325] _Organon_, i. 484, cited by Eucken in _Geschichte der\nphilosophischer Terminologie_, p. 205.\n\n[326] A 11 = B 25, A 56 = B 80.\n\n[327] Cited by Vaihinger, i. p. 468.\n\n[328] Cf. _Text-Book to Kant_, p. 13.\n\n[329] Cf. _Kant Explained_, p. 89.\n\n[330] Cf. below, p. 238.\n\n[331] Cf. below, pp. 116-17, 302.\n\n[332] Adickes has taken the liberty in his edition of the _Critique_ of\nsubstituting in A 297 = B 354 transcendental for transcendent. The\nBerlin edition very rightly retains the original reading.\n\n[333] B 27.\n\n[334] A vi.\n\n[335] A 14-15 = B 28. Cf. below, p. 570_n._\n\n[336] This alteration is not given in Max M\u00fcller\u2019s translation.\n\n[337] Cf. the corresponding alteration made in the second edition at end\nof note to A 21 = B 35.\n\n[338] A 15 = B 29.\n\n[339] _Loc. cit._\n\n[340] _Loc. cit._ Cf. A 835 = B 863.\n\n[341] Cf. A 124, B 151-2, and below, pp. 225, 265.\n\n[342] Cf. A 141 = B 180-1. Cf. _Critique of Judgment_, \u00a7 57: \u201cThus here\n[in the _Critique of Aesthetic Judgment_], as also in the _Critique of\nPractical Reason_, the antinomies force us against our will to look\nbeyond the sensible and to seek in the supersensible the point of union\nfor all our _a priori_ faculties; because no other expedient is left to\nmake our Reason harmonious with itself.\u201d Cf. also below, p. 473 ff., in\ncomment on A 649 = B 677.\n\n[343] A 16 = B 30.\n\n[344] Introduction (_W._ v. p. 16). Cf. below, p. 438.\n\n[345] Cf. also above, p. 25.\n\n[346] Cf. A 51 = B 75.\n\n[347] That thought finds in intuition its sole possible content is, of\ncourse, a conclusion first established in the _Analytic_. Kant is here\ndefining his terms in the light of his later results.\n\n[348] A 51 = B 75.\n\n[349] Cf. _Prolegomena_, \u00a7 12, Remark ii. at the beginning.\n\n[350] Cf. below, p. 88 ff.; B 146-7.\n\n[351] _Prolegomena_, \u00a7 8 (Eng. trans. p. 33).\n\n[352] Quoted by Vaihinger, ii. p. 4.\n\n[353] Cf. _Ueber das Organ der Seele_ (1796) and _Anthropologie_, \u00a7 22.\n\n[354] \u00a7 3.\n\n[355] A 27 = B 43, A 34 = B 51, A 42 = B 59, A 51 = B 75.\n\n[356] Cf. B 72.\n\n[357] In the second paragraph, A 20 = B 34.\n\n[358] _Dissertation_, \u00a7 4.\n\n[359] A 320 = B 376.\n\n[360] _Dissertation_, \u00a7 4.\n\n[361] A 50 = B 74.\n\n[362] This view, as I shall endeavour to show, is only semi-Critical,\nand is profoundly modified by the more revolutionary conclusions to\nwhich Kant finally worked his way. Cf. below, p. 274 ff.\n\n[363] In this he was anticipated by Tetens, _Philosophische Versuche\n\u00fcber die menschliche Natur_, Bd. i. (1777), _Versuch_ X. v. Cf. below,\np. 294.\n\n[364] _Critique of Judgment_, \u00a7 3 (Eng. trans, p. 49). Kant was the\nfirst to adopt the threefold division of mental powers--\u201cthe faculty of\nknowledge, the feeling of pleasure and pain, and the faculty of desire.\u201d\nThis threefold division is first given in his _Ueber Philosophie\n\u00fcberhaupt_ (Hartenstein, vi. p. 379), which was written some time\nbetween 1780 and 1790, being originally designed as an Introduction to\nthe _Critique of Judgment_.\n\n[365] A 248 (occurs in a lengthy section omitted in B).\n\n[366] This distinction between intuition and appearance practically\ncoincides with that above noted between intuition and its object.\n\n[367] For statement of the precise meaning in which these terms are here\nemployed, cf. above, pp. xlv-vii; below, pp. 270 ff., 312 ff.\n\n[368] This would harmonise with the view developed in A 166 (in its\nformulation of the principle of the _Anticipations_), A 374 ff., B 274\nff., A 723 = B 751.\n\n[369] Cf. A 50 = B 74: \u201cWe may name sensation the matter of sensuous\nknowledge.\u201d Similarly in A 42 = B 59; _Prolegomena_, \u00a7 11;\n_Fortschritte_, (Hartenstein, viii. p. 527).\n\n[370] Cf. below, p. 274 ff.\n\n[371] Cf. below, pp. 366-7, 370-2, 377.\n\n[372] ii. p. 59.\n\n[373] A 42 = B 60.\n\n[374] Cf. _Reflexionen_, ii. note to 469; also note to 357.\n\n[375] Cf. above, p. xxxiii ff.\n\n[376] A 266 = B 322.\n\n[377] In discussing _a_ and _b_ we may for the present identify form\nwith space. The problem has special complications in reference to time.\n\n[378] Cf. B 207.\n\n[379] Herbart\u2019s doctrine of space, Lotze\u2019s local sign theory, also the\nempiricist theories of the Mills and Bain, all rest upon this same\nassumption. It was first effectively called in question by William\nJames. Cf. Bergson: _Les Donn\u00e9es imm\u00e9diates_, pp. 70-71, Eng. trans. pp.\n92-3: \u201cThe solution given by Kant does not seem to have been seriously\ndisputed since his time: indeed, it has forced itself, sometimes without\ntheir knowledge, on the majority of those who have approached the\nproblem anew, whether nativists or empiricists. Psychologists agree in\nassigning a Kantian origin to the nativistic explanation of Johann\nM\u00fcller; but Lotze\u2019s hypothesis of local signs, Bain\u2019s theory, and the\nmore comprehensive explanation suggested by Wundt, may seem at first\nsight quite independent of the _Transcendental Aesthetic_. The authors\nof these theories seem indeed to have put aside the problem of the\nnature of space, in order to investigate simply by what process our\nsensations come to be situated in space and to be set, so to speak,\nalongside one another: but this very question shows that they regard\nsensations as inextensive, and make a radical distinction, just as Kant\ndid, between the matter of representation and its form. The conclusion\nto be drawn from the theories of Lotze and Bain, and from Wundt\u2019s\nattempt to reconcile them, is that the sensations by means of which we\ncome to form the notion of space are themselves unextended and simply\nqualitative: extensity is supposed to result from their synthesis, as\nwater from the combination of two gases. The empirical or genetic\nexplanations have thus taken up the problem of space at the very point\nwhere Kant left it: Kant separated space from its contents: the\nempiricists ask how these contents, which are taken out of space by our\nthought, manage to get back again.\u201d Bergson proceeds to argue that the\nanalogy of chemical combination is quite inapplicable, and that some\n\u201cunique act very like what Kant calls an _a priori_ form\u201d must still be\nappealed to. With the Kantian standpoint in this matter Bergson does\nnot, of course, agree. He is merely pointing out what the consequences\nmust be of this initial assumption of inextensive sensations.\n\n[380] Cf. _Von dem ersten Grunde des Unterschiedes der Gegenden im\nRaume_, in its penultimate paragraph.\n\n[381] Cf. _Dissertation_, last sentence of \u00a7 4, quoted below, p. 87.\n\n[382] A 291 = B 347; A 429 = B 457.\n\n[383] _Reflexionen_, ii. 334.\n\n[384] ii. p. 73.\n\n[385] Cf. Stout: _Manual of Psychology_ (3rd edition), pp. 465-6. \u201cWe\nfind that the definite apprehension of an order of coexistence, as such,\narises and develops only in connection with that peculiar aspect of\nsense-experience which we have called _extensity_, and more especially\nthe extensity of sight and touch. Two sounds or a sound and a smell may\nbe presented as coexistent in the sense of being simultaneous; but taken\nby themselves apart from association with experiences of touch and\nsight, they are not apprehended as spatially juxtaposed or separated by\na perceived spatial interval or as having perceived spatial direction\nand distance relatively to each other. Such relations can only be\nperceived or imagined, except perhaps in a very rudimentary way, when\nthe external object is determined for us as an extensive whole by the\nextensity of the same presentation through which we apprehend it.\u201d\n\n[386] _Principles of Psychology_, \u00a7 399, cited by Vaihinger.\n\n[387] \u00a7 4.\n\n[388] _Sich ordnen_ has here, in line with common German usage, the\nforce of a passive verb.\n\n[389] Riehl: _Kriticismus_ (1876-1879) ii. _Erster Theil_, p. 104. As\nalready noted, Kant tacitly admits this in regard to time relations of\ncoexistence and sequence. He continues, however, to deny it in regard to\nspace relations.\n\n[390] Cf. below, pp. 101-2, 105.\n\n[391] A 20 = B 34.\n\n[392] A 20 = B 34.\n\n[393] A 42 = B 60. Cf. _Dissertation_, \u00a7 12: [\u201cSpace and time, the\nobjects of pure mathematics,] are not only formal principles of all\nintuition, but themselves original intuitions.\u201d\n\n[394] A 196 = B 241; A 293 = B 349.\n\n[395] That is to say, in his published writings. It finds expression in\none, and only one, of the _Reflexionen_ (ii. 410: \u201cBoth space and time\nare nothing but combinations of sensuous impressions\u201d).\n\n[396] \u00a7 15, Coroll. at the end.\n\n[397] Cf. \u00a7 12, quoted above, p. 89 _n._ 2.\n\n[398] There also Kant teaches that the representation of space is gained\nfrom the space-endowed objects of experience.\n\n[399] Cf. B 1.\n\n[400] \u00a7 43.\n\n[401] _Ueber eine Entdeckung nach der alle neue Kritik der reinen Kritik\ndurch eine \u00e4ltere entbehrlich gemacht werden soll._\n\n[402] _Op. cit. W._ viii. pp. 221-2.\n\n[403] _Loc. cit._ p. 222.\n\n[404] Especially those which he had offered in support of the contention\nthat pure mathematical science is intuitive, not merely conceptual.\n\n[405] Cf. below, p. 291 ff., on Kant\u2019s reasons for developing his\ndoctrine of inner sense.\n\n[406] As no one passage can be regarded as quite decisively proving\nKant\u2019s belief in a pure manifold of intuition, the question can only be\ndecided by a collation of all the relevant statements in the light of\nthe general tendencies of Kant\u2019s thinking.\n\n[407] This at least would seem to be implied in the wording of his later\npositions; it is not explicitly avowed.\n\n[408] Cf. A 76-7 = B 102.\n\n[409] Cf. above, pp. xlii, 38-42; below, pp. 118-20, 128-34.\n\n[410] The last statement may be more freely translated: \u201cOnly in this\nway can I get the intuition before me in visible form.\u201d Cf. below, pp.\n135-6, 347-8, 359.\n\n[411] B 202-3.\n\n[412] Cf. _Reflexionen_, ii. 393, 409, 465, 630, 649.\n\n[413] This, indeed, is Kant\u2019s reason for describing space as an Idea of\nreason. Cf. below, pp. 97-8.\n\n[414] Geometry is for Kant the fundamental and chief mathematical\nscience (cf. A 39 = B 56 and _Dissertation_, \u00a7 15 _c_). In this respect\nhe is a disciple of Newton, not a follower of Leibniz. His neglect to\ntake adequate account of arithmetic and algebra is due to this cause.\nJust as in speaking of the manifold of sense he almost invariably has\nsight alone in view, so in speaking of mathematical science he usually\nrefers only to geometry and the kindred discipline of pure mechanics.\n\n[415] A 76-7 = B 102. Cf. B 160-1 _n._\n\n[416] Cf. above, pp. 90, 92 ff.; below, pp. 171, 226-9, 267-70, 337.\n\n[417] Cf. B 160.\n\n[418] _Metaphysical First Principles_, _W._ iv. p. 559, cf. p. 481.\n\n[419] _Op. cit._ p. 560.\n\n[420] _Critique of Judgment_, \u00a7\u00a7 26-7, Eng. trans. pp. 115-16 and 121.\n\n[421] Cf. below, pp. 102 _n._, 165-6, 390-1.\n\n[422] The title of this section, and the points raised in the opening\nparagraph, are commented upon below. Cf. pp. 110, 114-15, 134 ff. I pass\nat once to the first space argument.\n\n[423] Added in second edition.\n\n[424] This argument is an almost verbal repetition of the first argument\non space in the _Dissertation_, \u00a7 15.\n\n[425] Cf. below, pp. 106-7, 126, 132-3, 177-84, 338-9.\n\n[426] Cf. above, p. 37 ff.; below, p. 178 ff.\n\n[427] That is particularly obvious in Kant\u2019s formulation of his problem\nin the _Introduction_. For that is the assumption which underlies his\nmode of distinguishing between analytic and synthetic judgments. Cf.\nabove, p. 37.\n\n[428] Cf. above, p. xxii.\n\n[429] Cf. especially, pp. 184, 332-6, 419, 474, 479.\n\n[430] I here use the more modern terms. Kant, in _Anthropologie_, \u00a7 14,\ndistinguishes between them as _Organenempfindungen_ and\n_Vitalempfindungen_.\n\n[431] ii. p. 165.\n\n[432] Cf. above, pp. 85-8.\n\n[433] Cf. _Dissertation_, \u00a7 15 D: \u201cSpace is not anything objective and\nreal. It is neither substance, nor accident, nor relation, but is\nsubjective and ideal, proceeding by a fixed law from the nature of the\nmind, and being, as it were, a schema for co-ordinating, in the manner\nwhich it prescribes, all external sensations whatsoever.\u201d And \u00a7 15,\ncorollary at end: \u201cAction of the mind co-ordinating its sensations in\naccordance with abiding laws.\u201d\n\n[434] Especially in view of the third and fourth arguments on space, and\nof Kant\u2019s teaching in the transcendental exposition.\n\n[435] _E.g._ Cohen, Riehl, Caird, Watson.\n\n[436] Cf. Watson, _The Philosophy of Kant explained_, p. 83: \u201cKant,\ntherefore, concludes from the logical priority of space that it is _a\npriori_.\u201d\n\n[437] Upon it Kant bases the assertion that space is an Idea of reason;\ncf. above, pp. 96-8, and below, pp. 165-6, 390-1.\n\n[438] This second argument is not in the _Dissertation_.\n\n[439] Cf. Vaihinger, ii. pp. 196-7. The corresponding argument on time,\nin the form in which it is given in the second edition, is, as we shall\nfind, seriously misleading. It has caused Herbart and others to\nmisinterpret the connection in which this corollary stands to the main\nthesis. Herbart\u2019s interpretation is considered below, p. 124.\n\n[440] Cf. Vaihinger, ii. p. 220.\n\n[441] _Reflexionen_, ii. 403.\n\n[442] \u201cThat in space there are no more than three dimensions, that\nbetween two points there can be but one straight line, that in a plane\nsurface from a given point with a given straight line a circle is\ndescribable, cannot be inferred from any universal notion of space, but\ncan only be discerned in space as in the concrete.\u201d Cf. also\n_Prolegomena_, \u00a7 12.\n\n[443] In the second edition, the third.\n\n[444] For a different view cf. Vaihinger, ii. p. 233.\n\n[445] Cf. above, pp. 99-100; below, pp. 126, 180-1, 184, 338-9.\n\n[446] Cf. below, p. 180.\n\n[447] Cf. above, p. xxxvi; below, pp. 176 ff., 191, 195-6, 257, 290-1,\n404 ff., 413.\n\n[448] This statement occurs in a parenthesis; it has already been dwelt\nupon in the fourth (third) argument.\n\n[449] It has led Kant to substitute _er\u00f6rtern_ for _betrachten_ in A 23\n= B 38.\n\n[450] Cf. Vaihinger, ii. p. 151.\n\n[451] \u00a7 1 (Eng. trans, p. 13). Cf. above, p. 64.\n\n[452] This is, no doubt, one reason why Kant employs, in reference to\nspace, the unfortunate and confusing term concept (_Begriff_) in place\nof the wider term representation (_Vorstellung_). Cf. B 37, and above,\np. 64.\n\n[453] Cf. A 729 = B 757: \u201cIn place of the term _definition_ I should\nprefer to employ the term exposition. For that is a more guarded\nexpression, the claims of which the critic may allow as being in a\ncertain degree valid even though he entertain doubts as to the\ncompleteness of the analysis.\u201d Cf. _Logic_, \u00a7\u00a7 99 ff., 105. Cf. also\n_Untersuchung \u00fcber die Deutlichkeit der Grunds\u00e4tze, W._ ii. pp. 183-4:\n\u201cAugustine has said, \u2018I know well what time is, but if any one asks me,\nI cannot tell.\u2019\u201d\n\n[454] For explanation of the phrase \u201cconstruction of concepts\u201d cf.\nbelow, pp. 132-3.\n\n[455] Cf. below, p. 117. ff.\n\n[456] Cf. conclusion of fourth argument on space.\n\n[457] _A priori_ is here employed in its ambiguous double sense, as _a\npriori_ in so far as it precedes experience (as a _representation_), and\nin so far as it is valid independently of experience (as a\n_proposition_). Cf. Vaihinger, ii. p. 268.\n\n[458] Cf. below, p. 114 ff.\n\n[459] Cf. below, pp. 115-16.\n\n[460] Cf. _Lose Bl\u00e4tter_, i. p. 18: \u201cThis is a proof (_Beweis_) that\nspace is a subjective condition. For its propositions are synthetic and\nthrough them objects can be known _a priori_. This would be impossible\nif space were not a subjective condition of the representation of these\nobjects.\u201d Cf. _Reflexionen_, ii. p. 396, in which this _direct_ proof of\nthe ideality of space is distinguished from the _indirect_ proof by\nmeans of the antinomies.\n\n[461] By \u201cconcepts\u201d Kant seems to mean the five arguments, though as a\nmatter of fact other conclusions and presuppositions are taken into\naccount, and quite new points are raised.\n\n[462] This, according to Vaihinger (ii. p. 287), is the first occurrence\nof the phrase _Dinge an sich_ in Kant\u2019s writings.\n\n[463] Cf. Vaihinger\u2019s analysis of this discussion, ii. pp. 290-313.\n\n[464] ii. pp. 289-90.\n\n[465] Cf. below, pp. 415 ff., 515 ff., 558 ff.\n\n[466] In B 166 ff.\n\n[467] This is likewise true of the references in the letter to Herz,\n21st Feb. 1772. Cf. below, pp. 219-20.\n\n[468] _The Critical Philosophy of Kant_, i. pp. 306-9.\n\n[469] Cf. letter to Herz, _W._ x. p. 126. It is, Kant there says, the\nmost absurd explanation which can be offered of the origin and validity\nof our knowledge, involving an illegitimate _circulus in probando_, and\nalso throwing open the door to the wildest speculations. Cf. above, p.\n28; below, pp. 141-2, 290, 590.\n\n[470] Cf. B 167-8.\n\n[471] That is, in the first edition. Cf. above, p. 85 ff.; and below, p.\n116.\n\n[472] Above, pp. 111-12.\n\n[473] ii. p. 335.\n\n[474] \u00a7\u00a7 6-11.\n\n[475] This identification of the two is especially clear in A 39 = B 56.\n\n[476] A 27 = B 43.\n\n[477] Cf. above, p. xxxv; below, pp. 117-20, 142, 185-6, 241-2, 257,\n290-1.\n\n[478] A 28 = B 44, cf. A 35 = B 52.\n\n[479] Cf. Vaihinger, i. pp. 351-4; and above, p. 76; below, p. 302. Cf.\nCaird, _The Critical Philosophy_, i. pp. 298-9, 301; and Watson, _Kant\nExplained_, p. 91.\n\n[480] _Gedanken von der wahren Sch\u00e4tzung der lebendigen Kr\u00e4fte_ (1747),\n\u00a7 10.\n\n[481] This important and far-reaching assertion we cannot at this point\ndiscuss. Kant\u2019s reasoning is really circular in the bad sense. Kant may\nlegitimately argue from the _a priori_ character of space to the\napodictic character of _pure_ mathematical science; but when he proceeds\nsimilarly to infer the apodictic character of _applied_ mathematics, he\nis constrained to make the further assumption that space is a fixed and\nabsolutely uniform mode in which alone members of the human species can\nintuit objects. That, as we point out below (p. 120), is an assumption\nwhich Kant does not really succeed in proving. In any case the\nrequirements of the strict synthetic method preclude him from arguing,\nas he does both in the _Dissertation_ (\u00a7 15) and in the third space\nargument of the first edition, that the _a priori_ certitude of applied\nmathematics affords proof of the necessary uniformity of all space.\n\n[482] \u00a7 15 D.\n\n[483] Cf. above, p. 111.\n\n[484] Cf. above, pp. 40-2, 93-4; below, pp. 131-3, 338-9, 418 ff.\n\n[485] A 99-100.\n\n[486] A 78 = B 104. Cf. A 159 = B 198, B 147.\n\n[487] \u00a7 38, Eng. trans, p. 81.\n\n[488] Cf. p. 241 ff.\n\n[489] A 28-9. Cf. B 1; _Prolegomena_, \u00a7 13, Remark II. at the end:\n\u201cCinnabar excites the sensation of red in me.\u201d Cf. above, pp. 80-8;\nbelow, pp. 146 ff., 274 ff.\n\n[490] Kant continues the discussion of this general problem in A 45 ff.\n= B 62 ff.\n\n[491] Kant himself again uses the confusing term conception.\n\n[492] \u00a7 14, 1.\n\n[493] Herbart, _Werke_, ii. 30. Quoted by Vaihinger, iii. p. 198.\n\n[494] The third argument on time will be considered below in its\nconnection with the transcendental exposition.\n\n[495] The chief omission goes, as we shall see, to form the concluding\nargument on time.\n\n[496] In the second edition, the third.\n\n[497] In the second edition, the third.\n\n[498] In the second edition, the fourth.\n\n[499] Cf. Vaihinger, ii. pp. 380-1.\n\n[500] Cf. second part of fourth (third) argument on space.\n\n[501] Kant\u2019s _Logik, Einleitung_, \u00a7 8, Eng. trans, p. 49.\n\n[502] Cf. above, pp. 99-100.\n\n[503] These axioms are: (1) time has only one dimension; (2) different\ntimes are not simultaneous but successive. In the fourth argument the\nsynthetic character of these axioms is taken as further evidence of the\nintuitive nature of time. This passage also is really part of the\ntranscendental exposition. That exposition has to account for the\nsynthetic character of the axioms as well as for their apodictic\ncharacter; and as a matter of fact the intuitive and consequent\nsynthetic character of the _a priori_ knowledge which arises from time\nis much more emphasised in the transcendental exposition than its\napodictic nature.\n\n[504] Cf. _Reflexionen_, ii. 374 ff.\n\n[505] Vaihinger, ii. p. 387.\n\n[506] Cf. A 41 = B 58: \u201cMotion which combines both [space and time]\npresupposes something empirical.\u201d\n\n[507] _W._ iv. p. 471.\n\n[508] _Ueber Philosophie \u00fcberhaupt_ (Hartenstein, vi. p. 395).\n\n[509] \u00a7 12.\n\n[510] _Loc. cit._\n\n[511] A 78 = B 104.\n\n[512] A 142-3 = B 182. It should be observed that in Kant\u2019s view\nschemata \u201cexist nowhere but in thought\u201d (A 141 = B 180). It may also be\nnoted that time is taken as conditioning the schemata of _all_ the\ncategories.\n\n[513] A 717 ff. = B 745 ff.\n\n[514] \u00a7 10.\n\n[515] _Erl\u00e4uterungen \u00fcber des Herrn Professor Kant Critik der reinen\nVernunft_ (K\u00f6nigsberg, 1784), p. 24. Johann Schulze (or Schultz) was\nprofessor of mathematics in K\u00f6nigsberg. He was also _Hofprediger_, and\nis frequently referred to as Pastor Schulze. Kant has eulogised him\n(_W._ x. p. 128) as \u201cthe best philosophical head that I am acquainted\nwith in our part of the world.\u201d In preparing the _Erl\u00e4uterungen_, which\nis a paraphrase or simplified statement of the argument of the\n_Critique_, with appended comment, Schulze had the advantage of Kant\u2019s\nadvice in all difficulties. Kant also read his manuscript, and suggested\na few modifications (_op. cit._ pp. 329, 343).\n\n[516] Cf. Vaihinger, ii. pp. 388-9.\n\n[517] _Werke_ (Frauenst\u00e4dt\u2019s ed., 1873), i. p. 133.\n\n[518] P. 129.\n\n[519] _W._ x. p. 530. Italics not in Kant.\n\n[520] _Untersuchung \u00fcber die Deutlichkeit der Grunds\u00e4tze: Erste\nBetrachtung_, \u00a7\u00a7 2, 3; _dritte Betrachtung_, \u00a7 1; _Dissertation_, \u00a7\u00a7 12,\n15 C.\n\n[521] P. 128.\n\n[522] _Dissertation_, \u00a7 15 C.\n\n[523] Cf. above, pp. 40-2, 118-20; below, pp. 338-9.\n\n[524] _Kant und die moderne Mathematik in Kant-Studien_, xii. (1907) p.\n34 _n._\n\n[525] Cf. A 713 ff. = B 741 ff.; A 4 = B 8; B 15-16; A 24; A 47-8 = B\n64-5.\n\n[526] Cf. below, pp. 337-8.\n\n[527] Cf. above, pp. 112 n. 4.\n\n[528] The content of the second _Conclusion_ in regard to space.\n\n[529] This expresses the matter a little more clearly than Kant himself\ndoes. The term representation is ambiguous. In the first paragraph it is\nmade to cover the appearances as well as their representation.\n\n[530] Cf. _Dissertation_, \u00a7 15 Coroll.: \u201cSpace properly concerns the\nintuition of the _object_; time the _state_, especially the\n_representative_ state.\u201d\n\n[531] Cf. below, pp. 309 ff., 347-8, 359.\n\n[532] Cf. _Reflexionem_, ii. 365 ff.\n\n[533] \u00a7 14, 5 and note to 5.\n\n[534] The opposite is, however, asserted in B 67.\n\n[535] Cf. A 427-8 _n._ = B 456 _n._\n\n[536] A 99. Cf. A 162 = B 203: \u201cI cannot represent to myself a line,\nhowever small, without drawing it in thought, _i.e._ generating from a\npoint all its parts one after another.\u201d Cf. pp. 94, 347-8.\n\n[537] Cf. _Lose Bl\u00e4tter_, i. 54: \u201cWithout space time itself would not be\nrepresented as quantity (_Gr\u00f6sse_), and in general this conception would\nhave no object.\u201d Cf. _Dissertation_, \u00a7 14. 5.\n\n[538] Cf. below, p. 365 ff.\n\n[539] In the _Dissertation_ time is treated before space.\n\n[540] Cf. above, pp. xxxiv, 120; below, pp. 241-2, 365, 367-70, 390-1.\n\n[541] Cf. _Dissertation_, \u00a7 15 C.\n\n[542] Cf. below, pp. 272 ff., 294-5, 308 ff., 365 ff.\n\n[543] A 23 = B 37.\n\n[544] They correspond to the third paragraph dealing with space. Cf.\nabove, p. 116.\n\n[545] Cf. above, pp. 116-17.\n\n[546] Cf. _W._ x. p. 102. Mendelssohn had also protested; cf. _op. cit._\nx. p. 110.\n\n[547] _W._ x. pp. 128-9. Italics not in Kant. Kant is entirely justified\nin protesting against the view that in denying things in themselves to\nbe in time he is asserting that they remain eternally the same with\nthemselves. To make a dancer preserve one and the same posture is not to\ntake him out of time, but to bring home to him the reality of time in an\nextremely unpleasant manner. Duration is one of the modes of time.\n\n[548] This is Kant\u2019s reply to Mendelssohn\u2019s objection (December 1770,\n_W._ x. p. 110): \u201cSuccession is at least a necessary condition of the\nrepresentations of finite spirits. Now the finite spirits are not only\nsubjects but also objects of representations, both for God and for our\nfellow-men. The succession must therefore be regarded as something\nobjective.\u201d\n\n[549] Cf. A 277 = B 333: \u201cIt is not given to us to observe even our own\nmind with any intuition but that of our inner sense.\u201d\n\n[550] Quoted by Vaihinger, ii. p. 406.\n\n[551] In the fourth _Paralogism_, A 366, and in the _Refutation of\nIdealism_, B 274.\n\n[552] Cf. A 42 = B 59.\n\n[553] Above, pp. 113-14.\n\n[554] Cf. Vaihinger, ii. p. 114.\n\n[555] The date of Kant\u2019s _Von dem ersten Grunde des Unterschiedes der\nGegenden im Raume_.\n\n[556] Cf. below, p. 161 ff.\n\n[557] Cf. _Dissertation_, \u00a7 15 D: \u201cThose who defend the reality of space\nconceive it either as an absolute and immense _receptacle_ of possible\nthings--a view which appeals not only to the English [thinkers] but to\nmost geometricians--or they contend that it is _nothing but a relation_\nholding between existing things, which must vanish when the things are\nremoved, and which is thinkable only in actual things. This latter is\nthe teaching of Leibniz and of most of our countrymen.\u201d That the account\nof Leibniz\u2019s teaching given in the paragraphs under consideration is not\naltogether accurate, need hardly be pointed out. Kant, following his\nusual method in the discussion of opposing systems, is stating what he\nregards as being the logical consequences of certain of Leibniz\u2019s\ntenets, rather than his avowed positions.\n\n[558] Cf. A 275-6 = B 331-2: \u201cLeibniz conceived space as a certain order\nin the community of substances, and time as the dynamical sequence of\ntheir states. But that which both seem to possess as proper to\nthemselves, in independence of things, he ascribed to the confused\ncharacter of their concepts, asserting this confusion to be the reason\nwhy what is a mere form of dynamical relations has come to be regarded\nas a special intuition, self-subsistent and antecedent to the things\nthemselves. Thus space and time were [for Leibniz] the intelligible form\nof the connection of things (substances and their states) in\nthemselves.\u201d Cf. also _Prolegomena_, \u00a7 13, _Anm._ i.\n\n[559] Kant has stated that both views conflict with \u201cthe principles of\nexperience.\u201d But his criticisms are not altogether on that line. The\nstatement strictly applies only to his criticism of the Leibnizian view.\nCf. _Dissertation_, \u00a7 15 D: \u201cThat first inane invention of reason,\nassuming as it does the existence of true infinite relations in the\nabsence of all interrelated entities, belongs to the realm of fable. But\nthose who adopt the other view fall into a much worse error. For whereas\nthe former place an obstacle in the way only of certain rational\nconcepts, _i.e._ concepts that concern noumena, and which also in\nthemselves are extremely obscure bearing upon questions as to the\nspiritual world, omnipresence, etc., the latter set themselves in direct\nantagonism to the phenomena themselves and to geometry, the most\nfaithful interpreter of all phenomena. For--not to dwell upon the\nobvious circle in which they necessarily become involved in defining\nspace--they cast geometry down from its position at the highest point of\ncertitude, and throw it back into the class of those sciences the\nprinciples of which are empirical. For if all modifications of space are\nderived only through experience from external relations, geometrical\naxioms can have only comparative universality, like that acquired\nthrough induction, in other words, such as extends only as far as\nobservation has gone. They cannot lay claim to any necessity save that\nof being in accordance with the established laws of nature, nor to any\nprecision except of the artificial sort, resting upon assumptions. And\nas happens in matters empirical, the possibility is not excluded that a\nspace endowed with other original modifications, and perhaps even a\nrectilineal figure enclosed by two lines, may sometime be discovered.\u201d\nCf. above, p. 114; below, p. 290.\n\n[560] In B 155 _n._ Kant distinguishes between motion of an object in\nspace, and motion as generation of a geometrical figure. The former\nalone involves experience; the latter is a pure act of the productive\nimagination, and belongs not only to geometry but also to transcendental\nphilosophy. This note, as Erdmann has pointed out (_Kriticismus_, pp.\n115, 168), was introduced by Kant into the second edition as a reply to\na criticism of Sch\u00fctz. The distinction as thus drawn is only tenable on\nthe assumption of a pure manifold distinct from the manifold of sense.\n\n[561] A 230 = B 283. Cf. above, pp. 57, 118; below, pp. 185-6, 257.\n\n[562] A 41 = B 58.\n\n[563] Cf. below, pp. 359-60.\n\n[564] _Les Donn\u00e9es Imm\u00e9diates_, p. 75.\n\n[565] ii. p. 446.\n\n[566] \u00a7\u00a7 4 and 27.\n\n[567] Cf. _Ueber eine Entdeckung_, etc.: _W._ viii. p. 220.\n\n[568] A 44 = B 61.\n\n[569] A 277 = B 334. Cf. A 278-9 = B 335-6.\n\n[570] When Kant says that the distinction is not logical (that of\nrelative clearness and obscurity) but transcendental, the latter term is\ntaken as practically equivalent to epistemological. It does not mean\n\u2018relating to the _a priori_,\u2019 but relating to transcendental philosophy,\njust as logical here means relating to logic. Cf. Vaihinger, ii. p. 452.\n\n[571] Cf. A 270 ff. = B 326 ff.\n\n[572] \u00a7 7 (I read _autem_ for _autor_). Cf. below, p. 187.\n\n[573] Cf. _Prolegomena_, \u00a7 13, Remark II.\n\n[574] Cf. above, pp. 120-1.\n\n[575] Cf. A 257 = B 313.\n\n[576] A 46 = B 63. This is the first occurrence in the _Critique_ of the\nphrase transcendental object. Transcendental is employed as synonymous\nwith transcendent. Cf. below, p. 204 ff.\n\n[577] Cf. above, pp. 120-2.\n\n[578] A 271 = B 327.\n\n[579] A 46-9 = B 63-6.\n\n[580] A 48 = B 65-6. Vaihinger (ii. pp. 470-2) gives what appears to be\na sufficient explanation of what Kant had in mind in its employment.\n\n[581] A 46 = B 64. Cf. _Dissertation_, \u00a7 15 C. In the concluding\nsentence of the first edition\u2019s _Aesthetic_, Kant for the first time\nuses the singular _Ding an sich_ in place of the more usual _Dinge an\nsich_ and also refers to it in problematic terms as what _may_ underlie\nappearances.\n\n[582] B 66-73.\n\n[583] _a_ does not contain anything not to be found elsewhere in the\nfirst edition. It is a restatement of A 265 ff. = B 321 ff., A 274 = B\n330, A 277 ff. = B 333 ff., A 283-5 = B 339-41.\n\n[584] An assertion, it may be noted, which conflicts with Kant\u2019s view of\nit as a pure manifold.\n\n[585] Kant was probably influenced by Tetens. Cp. below, p. 294.\n\n[586] Cf. below, p. 291 ff. _b_ together with B 152-8 is a more explicit\nstatement of the doctrine of inner sense than Kant had given in the\nfirst edition.\n\n[587] Vaihinger (ii. p. 486 ff.), who has done more than any other\ncommentator to clear up the ambiguities of this passage, distinguishes\nonly two views.\n\n[588] A 38 = B 55.\n\n[589] Cf. _Prolegomena_, _W._ iv. p. 376 _n._, Eng. trans. p. 149: \u201cThe\nreviewer often fights his own shadow. When I oppose the truth of\nexperience to dreaming, he never suspects that I am only concerned with\nthe _somnium objective sumtum_ of Wolff\u2019s philosophy, which is merely\nformal, and has nothing to do with the distinction of dreaming and\nwaking, which indeed has no place in any transcendental philosophy.\u201d\n\n[590] Cf. below, p. 270 ff.\n\n[591] B 69. For explanation of the references to time and\nself-consciousness, cf. below, pp. 308, 323.\n\n[592] This view of illusion likewise appears in A 293 = B 349, A 377-8,\nA 396, and _Prolegomena_, \u00a7 13, III., at the beginning.\n\n[593] _Prolegomena_, _loc. cit._\n\n[594] Cf. in the 1863 edition, Bd. ii. 267 ff. The examples of illusion\nemployed by Mendelssohn are reflection in a mirror and the rainbow.\n\n[595] _W._ x. p. 405.\n\n[596] _Schein_ is so used by Kant himself (_W._ x. p. 105) in a letter\nto Lambert in 1770.\n\n[597] A 38.\n\n[598] Cf. above, A 39 = B 57. This is, however, merely asserted by\nimplication; it is not proved. As already noted, Kant does not really\nshow that space and time, viewed as absolute realities, are\n\u201cinconsistent with the principles of experience.\u201d Nor does Kant here\nsupply sufficient grounds for his description of space and time as\n_Undinge_. Kant, it must be observed, does not regard the conception of\nthe actual infinite as in itself self-contradictory. Cf. below, p. 486.\n\n[599] B 275.\n\n[600] Cf. below, p. 298 ff., on Kant\u2019s _Refutations of Idealism_. This\nis also the meaning in which Kant employs the term in his pre-Critical\nwritings. Cf. _Dilucidatio_ (1755), prop. xii. _usus_; _Tr\u00e4ume eines\nGeistersehers_ (1766), ii. 2, _W._ ii. p. 364. These citations are given\nby Janitsch (_Kant\u2019s Urtheile \u00fcber Berkeley_, 1879, p. 20), who also\npoints out that the term is already used in this sense by B\u00fclffinger as\nearly as 1725, _Dilucidationes philos._ This is also the meaning in\nwhich the term is employed in B xxxiv. Cf. A 28 = B 44.\n\n[601] _Prolegomena_; _Anhang. W._ iv. pp. 374-5.\n\n[602] In his _Kleine Aufs\u00e4tze_ (3. _Refutation of Problematic Idealism_,\nHartenstein, v. p. 502) Kant would seem very inconsistently to accuse\nBerkeley of maintaining a solipsistic position. \u201cBerkeley denies the\nexistence of all things save that of the being who asserts them.\u201d This\nis probably, however, merely a careless formulation of the statement\nthat thinking beings alone exist. Cf. _Prolegomena_, \u00a7 13, Anm. ii.\n\n[603] _Prolegomena_, _W._ iv. p. 375; Eng. trans. p. 148.\n\n[604] Borowski (_Darstellung des Lebens und Charakters Immanuel Kant_,\nin Hoffman\u2019s ed. 1902, p. 248 ff.) gives a list of English writers with\nwhom Kant was acquainted. They were, according to Janitsch (_loc. cit._\np. 35), accessible in translation. Cf. above, pp. xxviii _n._ 3, 63 _n._\n1.\n\n[605] Cf. _W._ i. pp. 318, 322. When Kant cites Hume in the\n_Prolegomena_ (Introduction), the reference is to the German\ntranslation.\n\n[606] This was the first of Berkeley\u2019s writings to appear in German. The\ntranslation was published in Leipzig in 1781.\n\n[607] Cf. below, pp. 307-8. The opposite view has, however, been\ndefended by Vaihinger: _Philos. Monatshefte_, 1883, p. 501 ff.\n\n[608] _Enquiry Concerning the Human Understanding_ (sec. xii. pt. ii. at\nthe end).\n\n[609] Sixth edition, pp. 132, 214, 243 ff.\n\n[610] A 38.\n\n[611] A 377.\n\n[612] A 377-8. Though Kant here distinguishes between perceptions and\ntheir \u201couter objects,\u201d the latter are none the less identified with\nmental representations.\n\n[613] Cf. below, p. 305 ff.\n\n[614] _Prolegomena_, \u00a7 13, _Remark_ III.; and _Anhang_ (_W._ iv. p.\n374).\n\n[615] Kant\u2019s description of Berkeley\u2019s idealism as visionary and\nmystical is doubtless partly due to the old-time association of idealism\nin Kant\u2019s mind with the spiritualistic teaching of Swedenborg (_W._ ii.\np. 372). This association of ideas was further reinforced owing to his\nhaving classed Berkeley along with Plato.\n\n[616] _Prolegomena_, _Anhang_, _W._ iv. p. 374; Eng. trans. p. 147.\n\n[617] Cf. above, pp. 140-1.\n\n[618] \u00a7 27. In translating Kant\u2019s somewhat difficult Latin I have found\nhelpful the English translation of the _Dissertation_ by W. J. Eckoff\n(New York, 1894).\n\n[619] Besides the internal evidence of the passage before us, we also\nhave Kant\u2019s own mention of Mendelssohn in this connection in notes (to A\n43 and A 66) in his private copy of the first edition of the _Critique_.\nCf. Erdmann\u2019s _Nachtr\u00e4ge zu Kant\u2019s Kritik_, xx. and xxxii.; and above,\np. 11.\n\n[620] Cf. _Morgenstunden_, Bd. ii. of _Gesammelte Schriften_ (1863), pp.\n246, 288.\n\n[621] Cf. above, p. 116.\n\n[622] B 72.\n\n[623] Upon this subject cf. Vaihinger\u2019s exhaustive discussion in ii. p.\n518 ff.\n\n[624] _Gedanken von der wahren Sch\u00e4tzung der lebendigen Kr\u00e4fte_ (1747).\n\n[625] _Op. cit._ \u00a7 10. Cf. above, p. 117 ff.\n\n[626] _Von dem ersten Grunde des Unterschiedes der Gegenden im Raume._\n\n[627] Euler, _R\u00e9flexions sur l\u2019espace et le temps_ (1748). Vaihinger\n(ii. p. 530) points out that Kant may also have been here influenced by\ncertain passages in the controversy between Leibniz and Clarke.\n\n[628] _Loc. cit._, at the end.\n\n[629] In the Dorpater manuscript, quoted by Erdmann in his edition of\nthe _Prolegomena_, p. xcvii _n._\n\n[630] \u00a7 15 C.\n\n[631] So also in the _Metaphysical First Principles of Natural Science_\n(1786), _Erstes Hauptst\u00fcck_, _Erkl\u00e4rung 2_, _Anmerkung 3_.\n\n[632] Cf. above, p. 105.\n\n[633] A 289 = B 345.\n\n[634] More exactly between the writing of the _Metaphysical First\nPrinciples_ (in which as above noted the argument of the _Prolegomena_\nis endorsed) and 1787.\n\n[635] Cf. A 260 ff. = B 316 ff. on the _Amphiboly of Reflective\nConcepts_.\n\n[636] The _Dissertation_ cites the argument only with this purpose in\nview. And yet it is only from the _Dissertation_ standpoint that the\nwider argument of the _Prolegomena_ can be legitimately propounded.\n\n[637] Above, pp. 96-8, 102 _n._ 4; below, pp. 390-1.\n\n[638] B 73.\n\n[639] A 50 = B 74.\n\n[640] Cf. below, p. 176 _n._ 1.\n\n[641] K. p. 99 _n._\n\n[642] A 64 = B 89.\n\n[643] The definition of intuition given in A 19 = B 33 also applies only\nto empirical intuition.\n\n[644] For discussion of Kant\u2019s view of sensation as the matter of\nsensuous intuition, cf. above, p. 80 ff.\n\n[645] Second paragraph, A 51 = B 75.\n\n[646] Object (_Gegenstand_) is here used in the strict sense and no\nlonger as merely equivalent to content (_Inhalt_).\n\n[647] Cf. above, p. 79 ff.\n\n[648] P. 85.\n\n[649] Third paragraph, A 52 = B 77.\n\n[650] K. p. 100.\n\n[651] Kant\u2019s _Logik: Einleitung_, i. (Abbott\u2019s trans. p. 4).\n\n[652] Cf. A 796 = B 824; A 130 = B 169; also above, pp. 71-2.\n\n[653] A 709 = B 737.\n\n[654] Logik: _Einleitung_, i. (Eng. trans. p. 3).\n\n[655] Cf. below, p. 194.\n\n[656] _Einleitung_, i. (Eng. trans. p. 4).\n\n[657] Cf. A 796 = B 824.\n\n[658] _Logik_: _Einleitung_, i. (Eng. trans. p. 5).\n\n[659] A 57 = B 81.\n\n[660] _Einleitung_, vii. (Eng. trans. p. 40 ff.).\n\n[661] Kant might have added that transcendental logic defines further\nconditions, those of possible experience, and that by implication it\nrefers us to coherence as the ultimate test even of material truth.\n\n[662] A 60-2 = B 84-6.\n\n[663] _Einleitung_, ii. (Eng. trans. pp. 6-7).\n\n[664] Cf. above, pp. 71-2, 170; below, pp. 438, 563.\n\n[665] Cf. below, p. 425 ff.\n\n[666] Kant employs _Gegenstand_ and _Object_ as synonymous terms.\n\n[667] Cf. below, p. 426.\n\n[668] A 64 = B 89.\n\n[669] A 65 = B 90.\n\n[670] The opening statement, A 67 = B 92, that hitherto understanding\nhas been defined only negatively, is not correct, and would seem to\nprove that this section was written prior to the introduction to the\n_Analytic_, cf. above, p. 167.\n\n[671] See above, pp. 170-1.\n\n[672] A 79 = B 105. \u2018Element\u2019 translates the misleading term \u2018Inhalt.\u2019\n\n[673] Kant\u2019s definition of transcendental logic as differing from\ngeneral logic in that it does not abstract from _a priori_ content must\nnot be taken as implying that the categories of understanding are\ncontents, though of _a priori_ nature. As we shall find, though that is\nKant\u2019s view of the forms of sense, it is by no means his view of the\ncategories. They are, he repeatedly insists, merely functions, and are\nquite indeterminate in meaning save in so far as a content is yielded to\nthem by sense. In A 76-7 = B 102, in distinguishing between the two\nlogics, Kant is careful to make clear that the _a priori_ content of\ntranscendental logic consists exclusively of the _a priori_ manifolds of\nsense.\n\n[674] \u00a7 20, Eng. trans. p. 58.\n\n[675] The view of the two as co-ordinate reappears in the _Prolegomena_\n(\u00a7 20) in a section the general tendency of which runs directly counter\nto any such standpoint.\n\n[676] A 78 = B 103.\n\n[677] Cf. below, pp. 196, 204, 226.\n\n[678] _Einleitung_, viii., Eng. trans. p. 48.\n\n[679] Cf. above, pp. 37-8.\n\n[680] _The Mistaken Subtlety of the Four Syllogistic Figures_ (1762).\n_W._ ii. p. 47, Eng. trans. p. 79.\n\n[681] Cf. above, pp. 99-100, 106-7.\n\n[682] _W._ ii. pp. 58-9, Eng. trans. pp. 92-3.\n\n[683] Cf. _Reflexionen_, ii. 599.\n\n[684] Below, pp. 185-6.\n\n[685] The same indefiniteness of statement is discernible in Caird\u2019s (i.\np. 322 ff.) and Watson\u2019s (_Kant Explained_, pp. 121-2) discussions of\nthe principle supposed to be involved.\n\n[686] Cf. A 80 = B 106.\n\n[687] \u00a7 39.\n\n[688] P. 176 ff.\n\n[689] B 145-6. Cf. above, pp. xxxv-vi, xliv, 57, 142; below, pp. 257,\n291.\n\n[690] \u00a7\u00a7 5-6.\n\n[691] \u00a7\u00a7 7-8. Cf. above, pp. 144-5.\n\n[692] W. x. p. 126. Italics not in Kant.\n\n[693] The relevant _Reflexionen_ have been carefully discussed by\nAdickes (_Kant\u2019s Systematik_, p. 21 ff.). In what follows I have made\nextensive use of his results, though not always arriving at quite the\nsame conclusions.\n\n[694] \u00a7 15, Coroll.\n\n[695] In his later writings Kant recognises that the representations of\nspace and time involve an Idea of Reason. Cf. above, pp. 97-8; below,\npp. 390-1.\n\n[696] \u00a7 39.\n\n[697] _Reflexionen_, ii. 513, cf. 502, 525-7.\n\n[698] _Op. cit._ ii. 513.\n\n[699] Cf. _op. cit._ ii. 537.\n\n[700] Cf. above, p. 90 ff.\n\n[701] Only in one passage, _Rechtslehre_, i., Anhang 3, 2, cited by\nAdickes, _op. cit._ p. 13, does Kant so far depart from his own\northodoxy as to speak of the possibility of an _a priori_ tetrachotomy.\nBut he never wavers in the view that the completeness of a division\ncannot be guaranteed on empirical grounds.\n\n[702] Introduction, \u00a7 9 _n._ Eng. trans. p. 41.\n\n[703] \u00a7\u00a7 4-6, 9.\n\n[704] A 70-6 = B 95-101.\n\n[705] Cf. Adickes, Kant\u2019s _Systematik_, p. 36 ff.\n\n[706] Cf. Adickes, _op. cit._ p. 89 ff.\n\n[707] _Organon_, \u00a7 137. Cited by Adickes.\n\n[708] i. p. 343 ff.\n\n[709] Cf. below, p. 391 ff.\n\n[710] A 76-79 = B 102-5.\n\n[711] Cf. above, p. 171.\n\n[712] A 55.\n\n[713] Cf. also B 160.\n\n[714] _Kant\u2019s Theorie der Erfahrung_, 2nd ed. p. 257 ff.\n\n[715] _The Critical Philosophy of Kant_, i. p. 327 ff.\n\n[716] _Philosophischer Kriticismus_, 2nd ed. i. p. 484 ff.\n\n[717] _Kant explained_, p. 124 ff.\n\n[718] Cf. below, p. 198.\n\n[719] P. 226.\n\n[720] A 79 = B 105.\n\n[721] Cf. above, p. 177 ff.\n\n[722] A 79-80.\n\n[723] Cf. above, p. 192.\n\n[724] Cf. B 111.\n\n[725] Cf. Adickes, _Systematik_, pp. 42-3.\n\n[726] _Kant Explained_, p. 128.\n\n[727] Cf. above, p. 37.\n\n[728] Cf. _Dissertation_, \u00a7\u00a7 16 to 28, and below, p. 381 ff.\n\n[729] Cf. _Reflexionen_, ii. 795.\n\n[730] B 111-13.\n\n[731] _World as Will and Idea, Werke_ (Frauenst\u00e4dt), ii. p. 544; Eng.\ntrans. ii. p. 61.\n\n[732] Cf. Stadler, _Grunds\u00e4tze der reinen Erkenntnisstheorie_(1876), p.\n122. Cf. also below, pp. 387-9.\n\n[733] Cf. below, p. 391 ff.\n\n[734] A 81.\n\n[735] Cf. _Prolegomena_, \u00a7 39.\n\n[736] Cf. above, p. 186 ff.\n\n[737] _Kant Explained_, p. 120.\n\n[738] A 81.\n\n[739] A 82.\n\n[740] A 82.\n\n[741] B 110.\n\n[742] B 110-11.\n\n[743] Cf. below, pp. 199-200.\n\n[744] _W._ x. p. 344-5.\n\n[745] Cf. below, p. 382 ff.\n\n[746] Cf. below, pp. 433-4, 451, 480, 529, 559-60.\n\n[747] B 114.\n\n[748] Cf. 904-5.\n\n[749] Cf. 907-10.\n\n[750] Cf. B. Erdmann, _Mittheilungen_ in _Phil. Monatshefte_, 1884, p.\n80, and Adickes, _Systematik_, pp. 55-9.\n\n[751] _Reflexionen_, ii. p. 252 _n._\n\n[752] \u201cDie transcendentale Deduktion der Kategorien\u201d in the\n_Gedenkschrift f\u00fcr Rudolf Haym_. Published separately in 1902.\n\n[753] Readers who are not immediately interested in the analysis of the\ntext or in the history of Kant\u2019s earlier semi-Critical views may omit\npp. 203-34, with exception of pp. 204-19, on Kant\u2019s doctrine of the\ntranscendental object, which should be read.\n\n[754] The reader is recommended to mark off the passages in a copy of\nthe _Critique_.\n\n[755] Its first occurrence in the _Critique_ is in the _Aesthetic_ A 46\n= B 63. It there signifies the thing in itself.\n\n[756] A 104.\n\n[757] A 109.\n\n[758] A 109.\n\n[759] A 104.\n\n[760] Cf. above, p. 28; below, pp. 219-20.\n\n[761] _W._ x. pp. 124-5.\n\n[762] Cf. below, pp. 209-10.\n\n[763] Hume\u2019s view of the self is not developed in the _Enquiry_, and is\nnot mentioned by Beattie.\n\n[764] A 107.\n\n[765] Cf. _Reflexionen_, ii. 952 (belonging, as Erdmann notes, to the\nearliest Critical period): \u201cAppearances are representations whereby we\nare affected. The representation of our free self-activity\n(_Selbstt\u00e4tigkeit_) does not involve affection, and accordingly is not\nappearance, but apperception.\u201d Cf. below, p. 296.\n\n[766] \u00a7 8. Cf. above, pp. l-ii; below, pp. 243, 260-3, 272-3, 327-8,\n473-7, 515.\n\n[767] A 107. It is significant that Kant in A 107 uses, in reference to\napperception, the very unusual phrase, \u201c_unwandelbares Bewusstsein_.\u201d\n\n[768] A 108.\n\n[769] A 107.\n\n[770] Reicke, _Lose Bl\u00e4tter_, p. 19. The bearing and date of this\npassage is discussed below, p. 233.\n\n[771] _Op. cit._ p. 20.\n\n[772] _Op. cit._ p. 22 (written on a letter dated May 20, 1775).\n\n[773] This last statement cannot possibly be taken literally. In view of\nthe manner in which the transcendental object is spoken of elsewhere in\nthis section, and also in the _Dialectic_, we must regard it as standing\nfor an independent existence, and the relation of representations to it\nas being, therefore, something else than simply the unity of\nconsciousness.\n\n[774] It may be observed that when Kant in A 107, quoted above, refers\nto \u201c_a priori_ concepts,\u201d he adds in explanation, and within brackets,\n\u201cspace and time.\u201d\n\n[775] A 105-6.\n\n[776] The actual nature of Kant\u2019s teaching as to the origin and\nconstitution of the notion of the transcendental object is largely\nmasked by the fact that he places this proof of its validity so\nprominently in the foreground. The general nature of this proof is, of\ncourse, identical with that of his later positions.\n\n[777] A 109.\n\n[778] As in the _Lose Bl\u00e4tter_. Cf. below, p. 233.\n\n[779] Cf. below, pp. 227, 233-4, 268-9.\n\n[780] Cf. below, pp. 322-8; also pp. 260-3.\n\n[781] As above noted (p. 204 _n._) it also occurs in the _Aesthetic_ (A\n46 = B 63), as signifying the thing in itself.\n\n[782] A 238 = B 298.\n\n[783] Cf. below, p. 238.\n\n[784] A 238 ff. = B 298 ff.\n\n[785] A 250-1.\n\n[786] Cf. below, p. 407 ff.\n\n[787] A 253.\n\n[788] A 191 = B 236.\n\n[789] A 277-8 = B 333-4; A 288 = B 344.\n\n[790] Cf. _mundus phaenomenon_ in A 272 = B 328.\n\n[791] It is so dated by Adickes (K. p. 272 _n._), owing to a single\nreference to schemata in A 286 = B 342.\n\n[792] A 358 and A 361 (cf. A 355); A 366; A 372 and A 379-80; A 390-1, A\n393, and A 394.\n\n[793] A 366.\n\n[794] \u201cTranscendental\u201d here means \u201ctranscendent.\u201d Cf. A 379.\n\n[795] A 372; so also in A 613-14 = B 641-2.\n\n[796] The passage in A 393 is given below, p. 464.\n\n[797] A 494 = B 522. Cf. A 492 = B 521: \u201cThe true self (_das eigentliche\nSelbst_) as it exists in itself, _i.e._ the transcendental subject.\u201d\n\n[798] A 495 = B 523.\n\n[799] A 496 = B 524.\n\n[800] _Loc. cit._\n\n[801] Cf. also A 538 = B 566; A 540 = B 568; A 557 = B 585; A 564 = B\n592; A 565-6 = B 593-4; A 613 = B 641-2.\n\n[802] Above, p. 206.\n\n[803] Cf. above, pp. liii-v; below, pp. 280, 331, 373-4, 390-1, 414-17,\n429-31, 558-61.\n\n[804] Viz. the first layer of the deduction of the first edition, the\nrelevant sections in the chapter on phenomena and noumena (A 250 ff.),\nand the _Paralogisms_ with the subsequent _Reflection_.\n\n[805] Viz. the _Note on Amphiboly_, the chapter on the _Antinomies_, and\nthe chapter on the _Ideal_.\n\n[806] To the statement that the alterations in the second edition cease\nat the close of the chapter on the _Paralogisms_, there is only one\nsingle exception, namely, the very brief note appended to A 491 = B 519.\nThis exception, however, supports our general thesis. It is of polemical\norigin, referring to the nature of the distinction between\ntranscendental and subjective idealism, and was demanded by the new\n_Refutation of Idealism_ which in the second edition he had attached to\nthe _Postulates_.\n\n[807] It follows immediately upon the passage quoted above, p. 206.\n\n[808] _W._ x. pp. 125-6.\n\n[809] A 89 = B 121. I adopt B. Erdmann\u2019s reading of _auf_ for _als_.\n\n[810] A 88 = B 120.\n\n[811] A 90.\n\n[812] As we have already found (above, p. 27 _n._ 1), it had not been\nattained at the time when the _Introduction_ to the first edition was\nwritten.\n\n[813] A 95-96.\n\n[814] A 97.\n\n[815] A. 95; cf. A 96.\n\n[816] A 111.\n\n[817] _Loc. cit._\n\n[818] A 112.\n\n[819] A 111.\n\n[820] A 112-14.\n\n[821] A 110-14.\n\n[822] I. \u00a7 14 C Vaihinger regards as intermediate in date, but it is a\ncomparatively unimportant paragraph, and may for the present be left out\nof account. Cf. below, pp. 225-6.\n\n[823] A 118-19.\n\n[824] A 124.\n\n[825] _Loc. cit._\n\n[826] A 115.\n\n[827] A 76-9 = B 102-4. Not yet commented upon.\n\n[828] Cf. Vaihinger, _loc. cit._ p. 63.\n\n[829] A 77 = B 102. Cf. above, pp. 96-7.\n\n[830] _Loc. cit._\n\n[831] For explanation of the exact meaning in which these terms are\nemployed and for discussion of the complicated issues involved, cf.\nbelow, p. 270 ff.\n\n[832] Cf. A 118.\n\n[833] A 102.\n\n[834] A 99-100.\n\n[835] A 102.\n\n[836] A 100.\n\n[837] A 101. Cf. below, p. 255.\n\n[838] Pp. 238, 263 ff.\n\n[839] Cf. above, p. 211.\n\n[840] For Vaihinger\u2019s own statement of it, cf. _op. cit._ pp. 79-98.\n\n[841] Nos. 64-5, 117, 140-5.\n\n[842] No. 146.\n\n[843] Nos. 41, 81.\n\n[844] No. 104.\n\n[845] Cf. Nos. 964-5.\n\n[846] No. 947.\n\n[847] No. 948.\n\n[848] No. 949.\n\n[849] No. 952.\n\n[850] This is Erdmann\u2019s reading. Vaihinger substitutes _allgemein_ for\n_allein_, but without reason given.\n\n[851] No. 935. The translation is literal. Kant in the last sentence\nchanges from singular to plural.\n\n[852] No. 964.\n\n[853] Cf. also Nos. 957, 961. The latter shows how Kant already\nconnected the categories of relation with the logical functions of\njudgment.\n\n[854] Reicke, Nos. 7, 8, 10-18 (pp. 16-26, 29-49).\n\n[855] The chief relevant passages have been quoted above, p. 209.\n\n[856] The letter is given in _W._ x. p. 173.\n\n[857] Reicke, pp. 113-16.\n\n[858] According to Adickes the _Critique_ was \u201cbrought to completion\u201d in\nthe first half of 1780; in Vaihinger\u2019s view, on the other hand, Kant was\noccupied with it from April to September. Cf. above, p. xx.\n\n[859] In two respects, however, fragment B 12 anticipates the teaching\nof the fourth stage: (_a_) in suggesting (p. 114) the necessity of a\npure synthesis of pure intuition, and (_b_) in equating (p. 115)\nsynthesis of apprehension with synthesis of imagination.\n\n[860] Pp. 231-3.\n\n[861] Cf. below, pp. 268-9.\n\n[862] In B 160 Kant states that the synthesis of apprehension is only\nempirical; and in B 152 we find the following emphatic sentence: \u201cIn so\nfar as the faculty of imagination is spontaneously active I sometimes\nalso name it [i.e. in addition to entitling it transcendental and\nfigurative] _productive_, and thereby distinguish it from the\n_reproductive_ imagination whose synthesis is subject only to empirical\nlaws, i.e. those of association, and which therefore contributes nothing\nin explanation of the possibility of _a priori_ knowledge. Hence it\nbelongs, not to transcendental philosophy, but to psychology.\u201d Cf. the\ndirectly counter statement in A 102: \u201cThe reproductive synthesis of the\nfaculty of imagination must be counted among the transcendental actions\nof the mind.\u201d\n\n[863] Though, as we shall find, the deduction of the second edition is\nin certain respects more mature, it is in other respects less complete.\n\n[864] A 314 = B 370.\n\n[865] A x-xi. Cf. above, pp. 50-1.\n\n[866] Cf. below, pp. 543 ff., 576-7.\n\n[867] Whether it was the chief reason is decidedly open to question. The\nun-Critical character of its teaching as regards the function of\nempirical concepts and of the transcendental object, and the\nunsatisfactoriness of its doctrine of a threefold synthesis, would of\nthemselves account for the omission. The passage in the chapter on\nphenomena and noumena (A 250 ff.) in which the doctrine of the\ntranscendental object is again developed was likewise omitted in the\nsecond edition.\n\n[868] Cf. below, pp. 238, 263 ff.\n\n[869] Cf. also in _Methodology_, below, p. 543 ff.\n\n[870] Cf. above, pp. xxxvi, xxxvii-viii, 36; below, pp. 241-3.\n\n[871] Cf. below, p. 543 ff.\n\n[872] A xi.\n\n[873] A 100-1.\n\n[874] Kant\u2019s failure either to distinguish or to connect the two\ndeductions in any really clear and consistent manner is a defect which\nis accentuated rather than diminished in the second edition. Though the\nsections devoted to the subjective enquiry are omitted, and the argument\nof the objective deduction is so recast as to increase the emphasis laid\nupon its more strictly logical aspects, the teaching of the subjective\ndeduction is retained and influences the argument at every point. For\nthe new deduction, no less than that of the first edition, rests\nthroughout upon the initial assumption that though connection or\nsynthesis can never be given, it is yet the generative source of all\nconsciousness of order and relation.\n\n[875] It appears most clearly in Kant\u2019s proof of the category of\ncausality in the second _Analogy_. Cf. below, p. 364 ff.\n\n[876] Cf. below, pp. 252-3, 258, 287, 333, 343.\n\n[877] Cf. above, p. 208 ff.\n\n[878] Cf. above, pp. l-ii, 207-12; below, pp. 260-3, 272-3, 327-8,\n473-7, 515.\n\n[879] P. 239.\n\n[880] A 99.\n\n[881] A 102.\n\n[882] A 103.\n\n[883] _Loc. cit._\n\n[884] _Loc. cit._\n\n[885] A 100-1.\n\n[886] A 106.\n\n[887] A 101.\n\n[888] Such statements are in direct conflict with his own repeated\nassertions in other passages that reproduction and recognition are\nalways merely empirical. Cf. above, pp. 227-31, and below, pp. 264,\n268-9.\n\n[889] B 139-40.\n\n[890] In the first edition the subjective and objective deductions shade\ninto one another; and this question is raised in the section on\nsynthesis of recognition (A 104), where, as above noted (p. 204 ff.),\nKant\u2019s argument is largely pre-Critical, empirical concepts exercising\nthe functions which Kant later ascribed to the categories. But as we\nhave already considered the resulting doctrine of the transcendental\nobject both in its earlier and in its subsequent form, we may at once\npass to the more mature teaching of the other sections.\n\n[891] Cf. above, p. 204 ff.\n\n[892] Memory is only one particular mode in which recognition presents\nitself in our experience; Kant\u2019s purpose is to show that it is not more\nfundamental, nor more truly constitutive of apperception, than is\nrecognition in any of its other manifestations. Indeed the central\ncontention of the objective deduction is that it is through\nconsciousness of _objects_, _i.e._ through consciousness of _objective_\nmeanings, that self-consciousness comes to be actualised at all. Only in\ncontrast with, and through relation to, an objective system is\nconsciousness of inner experience, past or present, and therefore\nself-consciousness in its contingent empirical forms, possible to the\nmind. Cf. above, pp. li-ii; below, pp. 260-3.\n\n[893] B 134.\n\n[894] A 116.\n\n[895] Cf. above, p. 242; below, pp. 258, 332-3.\n\n[896] Cf. A 111.\n\n[897] A 117 _n._\n\n[898] This transcendental psychology is considered below (p. 263 ff.),\nin its connection with the later stages of the subjective deduction. Cf.\nabove, p. 238.\n\n[899] A 113-14.\n\n[900] Cf. above, p. 229.\n\n[901] Cf. A 100-1.\n\n[902] A 122-3.\n\n[903] Cf. B 140-3; B 151-2; B 164-5 5 and below, p. 286.\n\n[904] Here again the second edition text is more explicit than the\nfirst: \u201cThis peculiarity of our understanding, that it can produce _a\npriori_ unity of apperception solely by means of the categories, and\nonly by such and so many, is as little capable of further explanation as\nwhy we have just these and no other functions of judgment, or why space\nand time are the sole forms of our possible intuition.\u201d--B 145-6. Cf.\nabove, pp. xxxiii-vi, xliv, 57, 142, 186; below, pp. 291, 411.\n\n[905] Cf. above, pp. 252-3.\n\n[906] The second _Analogy_ embodies the argument which is implied in,\nand necessary to, the establishment of the assertions dogmatically made\nin A 111-12.\n\n[907] A 119.\n\n[908] Cf. A 128. On this whole question cf. above, p. 242; below, pp.\n287-8.\n\n[909] Cf. A 113, 125-9.\n\n[910] A 107, 111.\n\n[911] The explanation given in the second edition (B 132) is artificial,\nand does not reveal Kant\u2019s real reasons. It is also obscure owing to its\nemployment of dynamical terms to denote the relation of apperception to\nself-consciousness.\n\n[912] Cf. above, pp. 251-3.\n\n[913] Cf. A 112, 113, 128.\n\n[914] A 114.\n\n[915] A 94, 115, 118. Cf. also end of note to B 134.\n\n[916] Cf. above, pp. lii, 207-12, 243; below, pp. 327-8, 473-7, 515.\n\n[917] This is shown, not only by Kant\u2019s ethical writings, but also by\nhis less formal utterances, especially in his _Lectures on Metaphysics_\nand on _Religion_, in his _Reflexionen_, and in his _Lose Bl\u00e4tter_.\n\n[918] Cf. below, pp. 277-8.\n\n[919] Cf. above, pp. l-lii; below, pp. 277 ff., 461-2, 473-7.\n\n[920] In note to B 162 they are indeed identified.\n\n[921] Kant\u2019s vacillating attitude appears in the added phrase \u201cof whose\nactivity we are _hardly ever_ conscious.\u201d Cf. A 78: it is a \u201cblind\u201d\npower.\n\n[922] Cf. above, p. 225; below, p. 337.\n\n[923] A 138 = B 177.\n\n[924] A 118.\n\n[925] Cf. above, p. 253 ff.\n\n[926] A 123.\n\n[927] A 121-3.\n\n[928] A 125-6.\n\n[929] Above, pp. 74 ff., 238, 252.\n\n[930] Cf. above, pp. 96-7.\n\n[931] Cf. below, pp. 367, 371-2.\n\n[932] Cf. above, pp. 211, 227, 233-4.\n\n[933] In direct contradiction of his previous view of transcendental\nimagination as purely productive, it is now stated that it is\nreproductive. Cf. A 102.\n\n[934] Cf. above, pp. 225 ff., 264.\n\n[935] It must be remembered that this was also rendered necessary by the\narchaic character of their teaching in regard to the transcendental\nobject and the function of empirical concepts.\n\n[936] Cf. B 151-2. There is no mention, however, of objective affinity.\n\n[937] B 160-1. Cf. above, pp. 226-9.\n\n[938] In what follows I make use of an article, entitled \u201cThe Problem of\nKnowledge,\u201d which I have contributed to the _Journal of Philosophy,\nPsychology, and Scientific Methods_ (1912), vol. ix. pp. 113-28.\n\n[939] The same wide sense in which Kant employs \u201cempirical idealism.\u201d\n\n[940] Cf. above, pp. xliii-v, 208; below, pp. 295-6, 298 ff. Hume and\nSpinoza are the only pre-Kantian thinkers of whose position the last\nstatement is not strictly descriptive, but even they failed to escape\nits entangling influence.\n\n[941] Cf. A 28-9; also _Lectures on Metaphysics_ (P\u00f6litz\u2019s edition,\n1821), p. 188 ff. In Kant\u2019s posthumously published work, his _Transition\nfrom the Metaphysical First Principles of Natural Science to Physics_,\nit is asserted in at least twenty-six distinct passages that sensations\nare due to the action of \u201cthe moving forces of matter\u201d upon the\nsense-organs. Cf. below, p. 283 _n_. 2. In his _Ueber das Organ der\nSeele_ (1796) (Hartenstein, vi. p. 457 ff.), Kant agrees with S\u00f6mmerring\nin holding that the soul has virtual, _i.e._ dynamical, though not\nlocal, presence in the fluid contained in the cavity of the brain.\n\n[942] Cf. _Critique of Practical Reason_, Bk. i. ch. i. \u00a7 iii.\n\n[943] Cf. below, pp. 279 ff., 293-6, 312 ff., 321, 361 _n._ 3, 384-5,\n464-5, 476.\n\n[944] Cf. below, pp. 279-80, and pp. 293-4, on inner sense.\n\n[945] i. p. 339: \u201cEach pulse of cognitive consciousness, each Thought,\ndies away and is replaced by another.... Each later Thought, knowing and\nincluding thus the Thoughts which went before, is the final\nreceptacle--and appropriating them is the final owner--of all that they\ncontain and own. Each Thought is thus born an owner, and dies owned,\ntransmitting whatever it realized as its Self to its own later\nproprietor. As Kant says [cf. below, pp. 461-2], it is as if elastic\nballs were to have not only motion but knowledge of it, and a first ball\nwere to transmit both its motion and its consciousness to a second,\nwhich took both up into _its_ consciousness and passed them to a third,\nuntil the last ball held all that the other balls had held, and realized\nit as its own.\u201d\n\n[946] I here use \u201cobjective\u201d in its modern meaning: I am not concerned\nwith the special meaning which Descartes himself attached to the terms\n_objective_ and _formaliter_.\n\n[947] Pp. 277-8.\n\n[948] On this whole matter cf. above, p. xlv; below, pp. 312-21 on\nKant\u2019s _Refutation of Idealism_; pp. 373-4 on the _Second Analogy_; pp.\n407 ff., 414 ff. on _Phenomena and Noumena_; p. 461 ff. on the\nParalogisms; and p. 546. Cf. also A 277-8 = B 334.\n\n[949] P. 267 ff.\n\n[950] Though the posthumously published work of Kant\u2019s old age, his\n_Transition from the Metaphysical First Principles of Natural Science to\nPhysics_, bears the marks of weakening powers, and is much too\nincomplete and obscure to allow of any very assured deductions from its\nteaching, it is none the less significant that it is largely occupied in\nattempting to define the relation in which the objective world of\nphysical science stands to the sensible world of ordinary consciousness.\nAs above noted (p. 275 _n._), it is there asserted in at least\ntwenty-six distinct passages that sensations are due to the action of\n\u201cthe moving forces of matter\u201d upon the sense-organs. What is even more\nsignificant is the adoption and frequent occurrence (_Altpreussische\nMonatsschrift_ (1882), pp. 236, 287, 289, 290, 292, 294, 295-6, 300,\n308, 429, 436, 439) of the phrase \u201c_Erscheinung von der Erscheinung_.\u201d\nKant would seem to mean by _\u201cErscheinung vom ersten Range\u201d_ (_op. cit._\np. 436) (_i.e._ appearance as such), the objective world as determined\nby physical science; and by \u201c_Erscheinung vom zweiten Range_\u201d (_i.e._\nappearance of the appearance), this same objective world as known in\nterms of the sensations which material bodies generate by acting on the\nsense-organs. Kant adds that the former is known directly, and the\nlatter indirectly--meaning, apparently, that the former is known through\n_a priori_ forms native to the understanding, and the latter only in\nterms of sense-data which are mechanically conditioned (cf. _loc. cit._\npp. 286, 292, and 444 _n._ The terms latter and former on p. 300 have\ngot transposed).\n\n[951] Cf. below pp. 312-21, 373-4, 414 ff., 425 ff., 558 ff.\n\n[952] B 129.\n\n[953] B 161 _n._\n\n[954] B 130-1.\n\n[955] B 131.\n\n[956] B 131-4.\n\n[957] B 131.\n\n[958] Cf. B 138.\n\n[959] B 135.\n\n[960] B 136-40.\n\n[961] B 140-2.\n\n[962] B 143.\n\n[963] \u00a7\u00a7 21-27.\n\n[964] Above, pp. 252-3, 258, 287.\n\n[965] _Prolegomena_, \u00a7 18.\n\n[966] _Op. cit._ \u00a7 20.\n\n[967] _Op. cit._ \u00a7\u00a7 18-19; Eng. trans. pp. 54-5.\n\n[968] Cf. above, pp. 39-40, 286-7.\n\n[969] Cf. below, p. 370.\n\n[970] _Op. cit._ \u00a7 22. Cf. below, p. 311 n. 4.\n\n[971] \u00a7\u00a7 21-7.\n\n[972] B 143.\n\n[973] This leads on in the second paragraph of \u00a7 21 to further\nstatements, already commented upon above, pp. 186, 257-8. Cf. also \u00a7 23.\n\n[974] Cf. also \u00a7 24.\n\n[975] Cf. above, pp. 90 ff., 171, 226-9, 267-70; below, p. 337.\n\n[976] Cf. above, pp. 28, 47, 114, 141-2.\n\n[977] Cf. \u00a7 21, second paragraph.\n\n[978] Cf. above, pp. 160, 186, 257, and below, pp. 325-6, 330-1, 390-1,\n404 ff.\n\n[979] Cf. below, pp. 324, 329.\n\n[980] Above, p. 148.\n\n[981] Cf. above, pp. xliii-v, l-ii, 238, 261-2, 263 ff., 273 ff.; below,\npp. 295 ff., 322 ff.\n\n[982] Cf. B 67-8; A 33 = B 49.\n\n[983] B 67.\n\n[984] B xxxix _n._\n\n[985] Kant very probably arrived at this view of inner sense under the\ninfluence of Tetens who teaches a similar doctrine in his\n_Philosophische Versuche \u00fcber die menschliche Natur und ihre\nEntwickelung_. Cf. Bd. i.; _Versuch_ i. 7, 8. The first volume of\nTetens\u2019 work was published in 1777 (re-issued by the _Kantgesellschaft_\nin 1913), and had been carefully read by Kant prior to the final\npreparation of the _Critique_. Cf. B. Erdmann, _Kriticismus_, p. 51.\n\n[986] Cf. A. 128-9.\n\n[987] As just noted, it is in the second edition that the above view of\nthe content of inner sense is first definitely formulated.\n\n[988] A 33 = B 49-50.\n\n[989] A 34 = B 50.\n\n[990] Cf. above, pp. 208-9, 251-2, 260-4; below, 311 _n._ 4. It may be\nobserved that Caird (i. pp. 625-7) interprets inner sense as equivalent\nto inner _reflection_. This is one of the respects in which Caird\u2019s\nHegelian standpoint has led him to misrepresent even Kant\u2019s most central\ndoctrines.\n\n[991] Cf. below, pp. 399-400, and A 277-8 = B 333-4.\n\n[992] Above, p. 292.\n\n[993] Cf. above, p. 155.\n\n[994] Cf. Vaihinger in _Strassburger Abhandlungen zur Philosophie_\n(1884), p. 106 ff.\n\n[995] Section III., Prop. _XII Usus_.\n\n[996] Theil II. Hauptst\u00fcck II. _W._ ii. p. 364.\n\n[997] \u00a7 11.\n\n[998] P\u00f6litz\u2019s edition (1821), pp. 100-2.\n\n[999] _W._ iv. p. 373 ff.\n\n[1000] It may be noted that in the _Aesthetic_ (A 38 = B 55) Kant\nemploys the term idealism, without descriptive epithet, in the same\nmanner as in his pre-Critical writings, as signifying a position that\nmust be rejected.\n\n[1001] Cf. below, p. 301 ff.\n\n[1002] Pp. 307-8.\n\n[1003] Cf. A 368-9 and 372.\n\n[1004] A 377: a passage which bears signs of being a later\ninterpolation.\n\n[1005] B 274.\n\n[1006] A 368-9.\n\n[1007] A 369.\n\n[1008] A 28 = B 44. Cf. above, pp. 76, 116-17.\n\n[1009] A 370.\n\n[1010] _Loc. cit._\n\n[1011] A 372.\n\n[1012] A 373: _Weil indessen_, etc.\n\n[1013] Adickes regards them as _later_ additions. To judge by their\ncontent (cf. above, pp. 204 ff., 215-16, on Kant\u2019s doctrine of the\ntranscendental object), they are more probably of quite early origin.\n\n[1014] A 377-8.\n\n[1015] Adickes argues that this paragraph is subsequent to the main body\nof the _Analytic_, but that is in keeping with the tendency which he\nseems to show of dating passages, which cannot belong to the \u201cBrief\nOutline,\u201d later rather than earlier.\n\n[1016] A 375.\n\n[1017] The remaining passages in the fourth _Paralogism_, together with\nthe corresponding passages in B 274 ff., in Kant\u2019s note to B xxxix, and\nin B 291-3, are separately dealt with below, pp. 308 ff., 322 ff.,\n462-3.\n\n[1018] A 377.\n\n[1019] _Loc. cit._\n\n[1020] _E.g._ Garve.\n\n[1021] \u00a7 13, _W._ iv. pp. 288-9: Eng. trans. p. 42.\n\n[1022] _Loc. cit._\n\n[1023] _Op. cit._ pp. 289-90: Eng. trans. pp. 43-4.\n\n[1024] In Note II.\n\n[1025] \u00a7 49, _W._ iv. 336: Eng. trans. p. 99.\n\n[1026] _Anhang_, _W._ iv. p. 375 n.\n\n[1027] _W._ iv. p. 374: Eng. trans. p. 147.\n\n[1028] Cf. above, p. 155 ff.\n\n[1029] _W._ iv. p. 375.\n\n[1030] _W._ iv. p. 375: Eng. trans. p. 147-8.\n\n[1031] Cf. above, p. 156.\n\n[1032] As already noted above, p. 299, it is employed by Kant in his\nlectures on Metaphysics.\n\n[1033] Kant\u2019s phrase \u201cin space outside me\u201d is on Kant\u2019s principles\nreally pleonastic. Cf. _Prolegomena_, \u00a7 49; Eng. trans, p. 101: \u201cthe\nnotion \u2018outside me\u2019 only signifies existence in space.\u201d Cf. A 373.\n\n[1034] Cf. text as altered by note to B xxxviii.\n\n[1035] B xxxix.\n\n[1036] B 291-2. The remaining points in B 274 ff. as well as in B xxxix\n_n._ are separately dealt with below, p. 322 ff.\n\n[1037] The nearest approach to such teaching in the first edition is in\nA 33 = B 50. Cf. above, pp. 135-8.\n\n[1038] Cf. below, pp. 333, 341, 360, 384-5.\n\n[1039] Adamson (_Development of Modern Philosophy_, i. p. 241) takes the\nopposite view as to what is Kant\u2019s intended teaching, but remarks upon\nits inconsistency with Kant\u2019s own fundamental principles. \u201cNow, in\ntruth, Kant grievously endangers his own doctrine by insisting on the\nabsence of _a priori_ elements from our apprehension of the mental life;\nfor it follows from that, if taken rigorously, that according to Kant\nsense and understanding are not so much sources which unite in producing\nknowledge, as, severally, sources of distinct kinds of apprehension. If\nwe admit at all, in respect to inner sense, that there is some kind of\napprehension without the work of understanding, then it has been\nacknowledged that sense is _per se_ adequate to furnish a kind of\napprehension.\u201d As pointed out above (p. 296), by the same line of\nreasoning Kant is disabled from viewing inner consciousness as merely\nreflective. In other words it can neither be more immediate nor less\nsensuous than outer perception. Cf. below, pp. 361, _n._ 3, 384-5.\n\n[1040] Above, pp. xlvi, 275-82; below, pp. 313-14, 384-5.\n\n[1041] Above, pp. 276, 279-80; below, pp. 312, 384-5.\n\n[1042] Cf. below, p. 361.\n\n[1043] Cf. _Metaphysical First Principles of Natural Science_ (1786),\n_W._ iv. pp. 470-1. It should be observed, however, that the reasons\nwhich Kant gives in this treatise for denying that psychology can ever\nbecome more than a merely historical or descriptive discipline are not\nthat the objects of inner sense fall outside the realm of mechanically\ndetermined existence. Kant makes no assertion that even distantly\nimplies any such view. His reasons are--(1) that, as time has only one\ndimension, the main body of mathematical science is not applicable to\nthe phenomena of inner sense and their laws; (2) that such phenomena are\ncapable only of a merely ideal, not of an experimental, analysis; (3)\nthat, as the objects of inner sense do not consist of parts outside each\nother, their parts are not substances, and may therefore be conceived as\ndiminishing in intensity or passing out of existence without prejudice\nto the principle of the permanence of substance (_op. cit._ p. 542,\nquoted below, p. 361, _n._ 2); (4) that inner observation is limited to\nthe individual\u2019s own existence; (5) that the very act of introspection\nalters the state of the object observed.\n\n[1044] A 370.\n\n[1045] B 275. These two sentences are cited in this connection by\nVaihinger: _Strassburger Abhandlungen zur Philosophie_ (1884), p. 131.\n\n[1046] Above, pp. xlv-vii, 279 ff.\n\n[1047] Cf. also above, pp. 275-7.\n\n[1048] \u00a7 13, _Anmerkung II_.\n\n[1049] _Kriticismus_, p. 197 ff.\n\n[1050] _Mind_ (1879), iv. p. 408 ff.; (1880), v. p. 111.\n\n[1051] _A Defence of Philosophic Doubt_ (1879), p. 107 ff.; _Mind_\n(1878), iii. p. 481; iv. p. 115; vi. p. 260.\n\n[1052] _Op. cit._ p. 128 ff.\n\n[1053] _Critical Philosophy_, i. 632 ff.; _Mind_ (1879), iv. pp. 112,\n560-1; v. p. 115.\n\n[1054] _The Philosophy of Kant_, p. 249 ff.\n\n[1055] The one fundamental question to which Erdmann would seem to allow\nthat Kant gives conflicting answers is as to whether or not categories\ncan be transcendently employed. The assumption of a uniform teaching is\nespecially obvious in Sidgwick\u2019s comments; cf. _Mind_ (1880), v. p. 113;\n_Lectures on the Philosophy of Kant_ (1905), p. 28.\n\n[1056] Cf. above, pp. 303-4.\n\n[1057] A 491 = B 520.\n\n[1058] A 225-6 = B 273.\n\n[1059] A 495-6 = B 523-4.\n\n[1060] Cf. below, p. 506.\n\n[1061] Viz. A 225-6 = B 273.\n\n[1062] B 277.\n\n[1063] Above, p. 208 ff.\n\n[1064] A 346 = B 404.\n\n[1065] A 224-5 = B 272-3.\n\n[1066] Cf. B 277.\n\n[1067] Quoted by B. Erdmann: _Kriticismus_, p. 107.\n\n[1068] B xxxix _n._, 67-8, 70, 157-8 with appended note, 276-8, 422\n_n._, 427-9.\n\n[1069] B 70, 157, 428.\n\n[1070] B 157.\n\n[1071] B 157 _n._ Regarding the un-Critical character of Kant\u2019s language\nin this passage, and of the tendencies which inspire it, cf. below, p.\n329.\n\n[1072] B 157.\n\n[1073] B 429.\n\n[1074] Cf. B 277-8 and B 157.\n\n[1075] B 278.\n\n[1076] B 420 and B 422 _n._\n\n[1077] B 422 _n._\n\n[1078] Cf. above, pp. 204 ff., 404 ff.\n\n[1079] Cf. above, p. 204 ff.\n\n[1080] B 406.\n\n[1081] B 422 _n._ Though both concepts are denoted by the same term,\nthey may not--such is the implication--be for that reason identified.\n\n[1082] B 429. Kant does not, however, even in the second edition, hold\nconsistently to this position. In the sentence immediately preceding\nthat just quoted he equates the transcendental self with the notion of\n\u201cobject in general.\u201d \u201cI represent myself to myself neither as I am nor\nas I appear to myself, but think myself only as I do any object in\ngeneral from whose mode of intuition I abstract.\u201d\n\n[1083] The broader bearing of this view may be noted. If consistently\ndeveloped, it must involve the assertion that noumenal reality is\napprehended in terms of the Ideas of reason, for these are the only\nother concepts at the disposal of the mind. Cf. above, pp. liii-v,\n217-18; below, pp. 331, 390-1, 414-17, 426 ff., 558-61.\n\n[1084] A 402.\n\n[1085] It is doubtful whether A 401-2 represents a genuinely Critical\nposition. Several of its phrases seem reminiscent of Kant\u2019s\nsemi-Critical view of the nature of apperception. This is especially\ntrue of the assertion that self-consciousness is \u201citself unconditioned.\u201d\n\n[1086] A 346 = B 404. Cf. below, pp. 456, 461-2.\n\n[1087] Cf. A 345.\n\n[1088] That he does not really do so is clear from the context and also\nfrom the manner in which he restated this argument in the second edition\n(B 421-2).\n\n[1089] A 401-2, B 421-2; below, pp. 461-2.\n\n[1090] A 402; cf. B 407.\n\n[1091] Cf. B 421-2.\n\n[1092] Cf. above, pp. l-ii, 208-9, 260-3.\n\n[1093] Cf. above, _loc. cit._\n\n[1094] Cf. B 157-8 and 157 _n._, B 278, B 428-9.\n\n[1095] Above, pp. 295-6, 311 _n._ 4.\n\n[1096] There is this difference between the category of existence and\nthe categories of relation, namely, that it would seem to be impossible\nto distinguish between a determinate and an indeterminate application of\nit. Either we assert existence or we do not; there is no such third\nalternative as in the case of the categories of substance and causality.\nThe category of substance, determinately used, signifies material\nexistence in space and time; indeterminately applied it is the purely\nproblematic and merely logical notion of something that is always a\nsubject and never a predicate. The determinate category of causality is\nthe conception of events conditioning one another in time;\nindeterminately employed it signifies only the quite indefinite notion\nof a ground or condition. Also, Kant\u2019s explicit teaching (A 597 ff. = B\n625 ff.) is that the notion of existence stands in an altogether\ndifferent position from other predicates. It is not an attribute\nconstitutive of the concept of the subject to which it is applied, but\nis simply the positing of the content of that concept as a whole. Nor,\nagain, is it a relational form for the articulation of content. These\nwould seem to be the reasons why no distinction is possible between a\ndeterminate and an indeterminate application of the notion of existence,\nand why, therefore, Kant, in defending the possible dual employment of\nit, has difficulty in holding consistently to the doctrines expounded in\nthe _Postulates_. He is, by his own explicit teaching, interdicted from\ndeclaring that the notion of existence is both a category and not a\ncategory, or, in other words, that it may vary in meaning according as\nempirical or noumenal reality is referred to, and that only in the\nformer case is it definite and precise. Yet such a view would, perhaps,\nbetter harmonise with certain other lines of thought which first obtain\nstatement in the _Dialectic_. For though it is in the _Dialectic_ that\nKant expounds his grounds for holding that existence and content are\nseparate and independent, it is there also that he first begins to\nrealise the part which the Ideas of Reason are called upon to play in\nthe drawing of the distinction between appearance and reality.\n\n[1097] In the _Fortschritte_ (_Werke_ (Hartenstein), viii. p. 548 ff.)\nthis final step is quite definitely taken. Cf. below, pp. 390-1, 414-17,\n426 ff., 558-61. We have, as we shall find, to recognise a second\nfundamental conflict in Kant\u2019s thinking, additional to that between\nsubjectivism and phenomenalism. He alternates between what may be\nentitled the sceptical and the Idealist views of the function of Reason\nand of its relation to the understanding, or otherwise stated, between\nthe regulative and the absolutist view of the nature of thought. But\nthis conflict first gains explicit expression in the _Dialectic_.\n\n[1098] For Kant\u2019s use of the terms \u2018canon\u2019 and \u2018dialectic\u2019 cf. above,\npp. 72, 77-8, 173-4, and below, p. 425 ff.\n\n[1099] Above, pp. 181-2.\n\n[1100] As we shall have occasion to observe below (p. 336), when Kant\ndefines judgment as \u201cthe faculty of subsumption under rules,\u201d he is\nreally defining it in terms of the process of reasoning, and thus\nviolating the principle which he is professedly following in dividing\nthe _Transcendental Logic into the Analytic of Concepts_, the _Analytic\nof Judgment_, and the _Dialectic of Reasoning_.\n\n[1101] A 132 = B 171.\n\n[1102] Pp. 252-3, 258-9, 287-8.\n\n[1103] The passages that have gone to constitute this chapter are\nprobably quite late in date of writing. This would seem to be proved by\nthe view taken of productive imagination, and also by the fact that in\nthe _Reflexionen_ there is no mention of schematism.\n\n[1104] Cf. above, p. 176 ff.\n\n[1105] Cf. A 137 = B 176. \u201cThe empirical concept of a plate is\nhomogeneous with the pure geometrical concept of a circle, since the\nroundness which is thought in the former can be intuited in the latter.\u201d\n\n[1106] A 138 = B 177.\n\n[1107] Above, p. 334.\n\n[1108] Cf. E. Curtius, _Das Schematismuskapitel in der Kritik der reinen\nVernunft (Kantstudien_, Bd. xix. p. 348 ff.).\n\n[1109] _Op. cit._ \u00a7 58.\n\n[1110] A 138 = B 177. Cf. above, pp. 96-7.\n\n[1111] A 141 = B 180.\n\n[1112] Cf. above, pp. 268-9.\n\n[1113] Cf. above, pp. 133-4.\n\n[1114] A 140 = B 179.\n\n[1115] _Loc. cit._\n\n[1116] Cf. E. Curtius, _op. cit._ p. 356.\n\n[1117] Kant\u2019s other definition of the schema as \u201ca rule for the\ndetermination of our intuition in accordance with a certain universal\nconcept\u201d (A 141 = B 180) is open to similar objections. When, however,\nKant states that \u201cschemata, and not images, _underlie_ our pure sensuous\nconcepts,\u201d he seems to be inclining to the truer view that the schema\n_is_ the concept.\n\n[1118] Above, pp. 131-3.\n\n[1119] Cf. Riehl, _Philos. Krit._ 2nd ed. i. pp. 488, 533. Cf. above,\npp. 195-6, 198; below, pp. 404-5.\n\n[1120] _Critical Philosophy_, i. bk. i. chap. v., especially pp. 437 and\n440.\n\n[1121] _Theorie der Erfahrung_, second edition, p. 384.\n\n[1122] _Op. cit._ p. 532.\n\n[1123] Cf. above, pp. 240-3.\n\n[1124] For comment upon the definition of number, which Kant takes as\nbeing the schema of quantity, and upon the view of arithmetic which this\ndefinition may seem to imply, cf. above, p. 128 ff.\n\n[1125] Cf. above, p. 192.\n\n[1126] Cf. above, pp. 339-40, and below, pp. 357, 404 ff.\n\n[1127] Cf. above, pp. 20, 25, 290-1; below, pp. 407, 412, 414-17.\n\n[1128] _E.g._ Riehl, _Philos. Krit._ 2nd ed. i. pp. 535-6.\n\n[1129] Above, pp. 258, 332-3.\n\n[1130] A 148 = B 188.\n\n[1131] A 156 = B 195.\n\n[1132] A 157 = B 196.\n\n[1133] A 24.\n\n[1134] \u00a7 13, _Anmerkung_ i.\n\n[1135] B 40-1.\n\n[1136] B 110.\n\n[1137] A 160 = B 199-200.\n\n[1138] Cf. above, pp. 288-9.\n\n[1139] A 161-2 = B 201-2.\n\n[1140] Cf. below, pp. 510-11.\n\n[1141] A 178-9 = B 221.\n\n[1142] Cf. above, pp. 94-5.\n\n[1143] Cf. below, pp. 358-9, 367-8, 371-2, 381-2.\n\n[1144] Above, p. 309 ff.\n\n[1145] Cf. Adickes, _K._ p. 190 _n._\n\n[1146] Cf. above, p. 127 ff.\n\n[1147] That is to say, in the first edition.\n\n[1148] The phrase is followed, it may be observed, by a verb in the\nirregular.\n\n[1149] A 143 = B 182.\n\n[1150] _Loc. cit._ in the chapter on _Schematism_.\n\n[1151] _Loc. cit._ Italics not in Kant.\n\n[1152] Cf. A 175 = B 217. Cf. above, pp. 350-1.\n\n[1153] B 217-18.\n\n[1154] Cf. above, pp. 192, 341.\n\n[1155] A 169-70 = B 211-12. For comment upon Kant\u2019s view of the point as\na limit, cf. below, p. 489 ff.\n\n[1156] Though Kant maintains in A 171 = B 212-13 that owing to our\ndependence upon empirical data and our necessary ignorance of the nature\nof the causal relation we cannot similarly demonstrate the principle of\nthe continuity of change, he has himself, in characteristically\ninconsistent fashion, given three such demonstrations. Cf. below, pp.\n380-1.\n\n[1157] Cf. Kant\u2019s _Monadologia physica_ (1756), and _New Doctrine of\nMotion and Rest_ (1758). Kant\u2019s final statement of this dynamical theory\nis given in his _Metaphysical First Principles of Natural Science_\n(1786).\n\n[1158] In this matter Kant regards himself as defending the Newtonian\ntheory of an attractive gravitational force. The mechanistic view admits\nonly one form of action, viz. transference of motion through impact and\npressure. \u201cFrom ... Democritus to Descartes, indeed up to our own day,\nthe mechanistic method of explanation ... has, under the title of\natomism or corpuscular philosophy, maintained its authority with but\nslight modification; and has continued to exercise its influence upon\nthe principles of natural science. Its essential teaching consists in\nthe assumption of the _absolute impenetrability_ of primitive matter, in\nthe _absolute homogeneity_ of its constitution (difference of shape\nbeing the sole remaining difference), and in the _absolutely\nindestructible coherence_ of matter in its fundamental corpuscles\u201d\n(_Metaphysical First Principles_, _W._ vol. iv. p. 533; ii. Allgemeine\nAnmerkung, 4).\n\n[1159] This is additional to its other correlative assumption of the\nabsolute void. \u201cThe absolute void and the absolutely full are in the\ndoctrine of nature very much what blind chance and blind fate are in\nmetaphysical cosmology, namely, a barrier to the enquiring reason, which\neither causes its place to be taken by arbitrary fictions, or lays it to\nrest on the pillow of obscure qualities\u201d (_Metaphysical First\nPrinciples_, _W._ vol. iv. p. 532 (I read _forschende_ for\n_herrschende_)). \u201cThere are only two methods of procedure ...: the\nmechanistic, through combination of the absolutely full with the\nabsolute void, or an opposite dynamical method, that of explaining all\nmaterial differences through mere differences in the combination of the\noriginal forces of repulsion and attraction\u201d (_loc. cit._).\n\n[1160] In the first edition Kant formulates this principle in the light\nof his extremely misleading distinction between mathematical and\ndynamical principles (cf. above, pp. 345-7): \u201cAll appearances, _as\nregards their existence_, are subject _a priori_ to rules determining\ntheir relation to one another in one time.\u201d\n\n[1161] Cf. below, p. 358.\n\n[1162] In A 182 = B 225 the stronger term change (_Wechsel_) is\nemployed.\n\n[1163] A 178-80 = B 221-3 (on the distinction between mathematical and\ndynamical principles) has been commented upon above, pp. 345-7.\n\n[1164] _Philos. Krit._ 2nd ed. i. p. 545. Caird adopts a similar view,\ni. pp. 540, 580.\n\n[1165] A 181 = B 224.\n\n[1166] Cf. below, pp. 373-4.\n\n[1167] That is to say, in the first edition.\n\n[1168] Cf. above, pp. 332-3, 343-4.\n\n[1169] Cf. above, p. 348; below, pp. 367-8, 371-2, 381-2.\n\n[1170] Cf. above, pp. 94, 135-8, 309 ff., 347-8.\n\n[1171] That is to say, in the first edition.\n\n[1172] The new proof added in the second edition calls for no special\ncomment. In all essentials it agrees with this second proof of the first\nedition. It differs only in such ways as are called for by the mode of\nformulating the principle in the second edition.\n\n[1173] This statement, as Caird has pointed out (i. p. 541), is\nextremely questionable. \u201cIt may be objected that to say that \u2018time\nitself does not change\u2019 is like saying that passing away does not itself\npass away. So far the endurance of time and the permanence of the\nchanging might even seem to mean only that the moments of time never\ncease to pass away, and the changing never ceases to change. A perpetual\nflux would therefore sufficiently \u2018represent\u2019 all the permanence that is\nin time.\u201d This is not, however, in itself a vital objection to Kant\u2019s\nargument. For he is here stating more than his argument really requires.\nEvents are dated in a _single_ time, not in an unchanging time. Kant\u2019s\nstatement betrays the extent to which, as Bergson has very justly\npointed out, Kant spatialises time, _i.e._ interprets it on the analogy\nof space. It is based on \u201cthe mixed idea of a measurable time, which is\nspace in so far as it is homogeneity, and duration in so far as it is\nsuccession; that is to say, at bottom, the contradictory idea of\nsuccession in simultaneity\u201d (_Les Donn\u00e9es imm\u00e9diates_, p. 173, Eng.\ntrans. p. 228).\n\n[1174] Cf. A 184 = B 227: \u201cthe proposition, that substance is permanent,\nis tautological.\u201d\n\n[1175] Cf. A 188 = B 231.\n\n[1176] Above, p. 341.\n\n[1177] Cf. above, p. 309 ff.\n\n[1178] B. Erdmann\u2019s edition of the _Nachtr\u00e4ge_, lxxx. p. 32. Cited by\nCaird, i. pp. 541-2.\n\n[1179] _Op. cit._ pp. 33-4.\n\n[1180] That Kant does not mean to imply that the category of substance\nhas no application to the contents of inner sense is made clear by a\ncurious argument in the _Metaphysical First Principles of Natural\nScience_ (1786), _W._ iv. p. 542: \u201cWhat in this proof essentially\ncharacterises substance, which is possible only in space and under\nspatial conditions, and therefore only as object of the outer senses, is\nthat its quantity cannot be increased or diminished without substance\ncoming into being or ceasing to be. For the quantity of an object which\nis possible only in space must consist of parts which are external to\none another, and these, therefore, if they are real (something movable),\nmust necessarily be substances. On the other hand, that which is viewed\nas object of inner sense can, _as substance_, have a quantity which does\nnot consist of parts external to one another. Its parts are therefore\nnot substances, and their coming into being and ceasing to be must not\nbe regarded as creation or annihilation of a substance. Their increase\nor diminution is therefore possible without prejudice to the principle\nof the permanence of substance.\u201d (Italics not in Kant.) Cf. also\n_Prolegomena_, \u00a7 49, and below, pp. 367, 377 _n._ 3.\n\n[1181] A 187 = B 230.\n\n[1182] _K._ p. 211 _n._\n\n[1183] C. A 205-7 = B 252.\n\n[1184] _Werke_ (Frauenst\u00e4dt, 1873), i. p. 85 ff.\n\n[1185] As evidence of this failure I may cite Schopenhauer\u2019s comment\nupon A 371 and 372: \u201cFrom these passages it is quite clear that for Kant\nthe perception of outer things in space is antecedent to all application\nof the causal law, and that this law does not therefore enter into it as\nits element and condition: mere sensation amounts in Kant\u2019s view to\nperception\u201d (_Werke_, i. p. 81). Even when, as in the passages referred\nto, Kant is speaking in his most subjectivist vein, he gives no\njustification for any such assertion. Schopenhauer, notwithstanding his\nsincere admiration for Kant--\u201cI owe what is best in my own system to the\nimpression made upon me by the works of Kant, by the sacred writings of\nthe Hindoos, and by Plato\u201d (_World as Will and Idea, Werke_, ii. p. 493,\nEng. trans. ii. p. 5)--is one of the most unreliable of Kant\u2019s critics.\nHis comments are extremely misleading, and largely for the reason that\nhe was interested in Kant only as he could obtain from him confirmation\nof his own philosophical tenets. Several of these tenets he certainly\nderived directly from the _Critique_; but they are placed by him in so\nentirely different a setting that their essential meaning is greatly\naltered. We have already noted (above, p. 41) Schopenhauer\u2019s exaggerated\nstatement of Kant\u2019s intuitive theory of mathematics. Kant\u2019s subjectivism\nis similarly expounded in a one-sided and quite unrepresentative manner\n(cf. below, p. 407 _n._). Hutchison Stirling\u2019s criticisms of Kant in his\n_Text Book to Kant_ are vitiated by a similar failure to recognise the\ncompletely un-Critical character of the occasional passages in which\nKant admits a distinction between \u201cjudgments of perception\u201d and\n\u201cjudgments of experience\u201d (cf. above, pp. 288-9). Stirling (cf. below,\np. 377) has amplified his criticism of Kant in _Princeton Review_ (Jan.\n1879, pp. 178-210), _Fortnightly Review_ (July 1872), and in _Mind_\n(ix., 1884, p. 531, and x., 1885, p. 45).\n\n[1186] Cf. above, pp. 240-2, 365, and below, p. 377.\n\n[1187] Cf. Stout, _Manual of Psychology_, third edition, pp. 444-6:\n\u201cUnless we assume from the outset that the primitive mind treats a\nperceived change which challenges its interest and attention, not as\nsomething self-existent in isolation, but as something conditioned by\nand conditioning other changes, it seems hopeless to attempt to show how\nthis causal point of view could have arisen through any extension of\nknowledge in accordance with ascertained psychological laws and\nconditions.... There is good reason for denying that customary\nrepetition is even required to furnish a first occasion or opportunity\nfor the first emergence of the apprehension of causal relations. For, as\nwe have already insisted, the process of learning by experience is from\nthe first experimental.... Regularities are only found because they are\nsought. But it is in the seeking that the category of causal unity is\nprimarily involved.\u201d Cf. below, pp. 371-2.\n\n[1188] A 193 = B 238.\n\n[1189] A 191 = B 236.\n\n[1190] By an \u201carbitrary\u201d order Kant does not, of course, mean an order\nof succession that is not determined, but only one that is determined by\nsubjectively conditioned direction of attention. Cf. below, p. 377.\n\n[1191] Cf. A 199 = B 244, and above, pp. 133, 288-9; below, p. 377.\n\n[1192] Cf. A 195-6 = B 240-1, and above, pp. 172, 176 ff., 182-3, 263\nff., 277-8.\n\n[1193] A 736-7 = B 765. Italics of last sentence not in Kant.\n\n[1194] A 189-94 = B 234-9: first to fourth paragraphs (first edition).\n\n[1195] Cf. above, pp. 348, 358.\n\n[1196] Cf. A 192-3 = B 238-9.\n\n[1197] Cf. Riehl, _Philosophischer Kriticismus_ (second edition), i. pp.\n551-2. While recognising the above main point, Riehl seems to assert\nthat empirical sequence determines the _application_ of the causal\nconcept. It would be truer, and more in accordance with the position\nwhich Kant is endeavouring to establish, to assert that appeal to\nconstancy of sequence enables us to determine which antecedents of any\ngiven event are causal conditions. The principle of causality is already\napplied when the sequent experiences are apprehended as sequent events.\nThis ambiguity, however, would seem to be due only to Riehl\u2019s mode of\nexpression. For, as he himself says (p. 551), the law of causality is a\nground of experience, and cannot therefore be derived from it. Cf.\nabove, pp. 267-8, 367.\n\n[1198] Pp. 365-71, 377.\n\n[1199] A 191 = B 236. Cf. above, pp. 216-18.\n\n[1200] As pointed out above, this is really a secondary meaning which\nKant reads into the term analogy; it is not the true explanation of his\nchoice of the term.\n\n[1201] _Critical Philosophy of Kant_, vol. i. pp. 540, 580.\n\n[1202] _Kant_, p. 198: trans. by Creighton and Lefevre, p. 196.\n\n[1203] Cf. above, pp. 270 ff., 313-21.\n\n[1204] Kant, of course, recognises that we cannot make any such positive\nassertion; to do so would be to transcend the limits imposed by Critical\nprinciples. Cf. below, p. 382.\n\n[1205] A 194-6 = B 239-41: fifth to seventh paragraphs (first edition).\n\n[1206] A 196-9 = B 241-4: eighth to tenth paragraphs (first edition).\n\n[1207] A 199-201 = B 244-6: eleventh to thirteenth paragraphs (first\nedition).\n\n[1208] Cf. above, pp. 224 ff., 264 ff.; below, 377.\n\n[1209] A 201-2 = B 246-7: fourteenth paragraph (first edition).\n\n[1210] B 233-4: second paragraph (second edition).\n\n[1211] B 233-4.\n\n[1212] From A 202 = B 247 to the end.\n\n[1213] Kant\u2019s phenomenalist substitute for the Cartesian subjectivism\n(cf. above, pp. 270 ff., 312 ff.) enables him to develop this thesis in\na consistent and thoroughgoing manner. The subjective is a subspecies\nwithin the class of what is determined by natural law; and the principle\nof causality is therefore applicable to subjective change in the same\nrigorous fashion as to the objectively sequent.\n\n[1214] A 204 = B 249.\n\n[1215] _W._ i. pp. 87-92.\n\n[1216] _Grunds\u00e4tze der reinen Erkenntniss-Theorie_, p. 151. Quoted and\ntranslated by Caird, i. p. 572. Caird sums up the matter in a sentence\n(p. 571): \u201cKant is showing, not that objective succession is always\ncausal, but that the determination of a succession of perceptions as\nreferring to a succession of states in an object, involves the principle\nof causality.\u201d\n\n[1217] _Loc. cit._\n\n[1218] The connected question how we can determine the ball and the\ncushion as objectively coexistent is the problem of the third _Analogy_.\n\n[1219] III. _Erkl\u00e4rung_ 1 and 2, _Lehrsatz_ 1 (especially _Anmerkung_\nthereto). Cf. also II. _Erkl\u00e4rung_ 1 and 5, and the last pages of the\n_Allgemeine Anmerkung_.\n\n[1220] Pp. 351, 373-4. Cf. pp. 318-21.\n\n[1221] A 170-1 = B 212-13, above, p. 353, _n._ 2.\n\n[1222] A 208 = B 253-4.\n\n[1223] _Metaphysical First Principles, II. Lehrsatz 4, Anmerkung 2._\n\n[1224] A 209-10 = B 255-6.\n\n[1225] A 210 = B 256.\n\n[1226] _W._ ii. p. 22.\n\n[1227] _W._ ii. p. 168.\n\n[1228] _Loc. cit._\n\n[1229] For lack of a more suitable English equivalent I have translated\n_Gemeinschaft_ as \u201ccommunion.\u201d As Kant points out in A 213 = B 260, the\nGerman term is itself ambiguous, signifying _commercium_ (_i.e._\ndynamical interaction) as well as _communio_.\n\n[1230] Cf. above, pp. 348, 358-9, 367-8, 371-2.\n\n[1231] \u00a7 17 ff. Cf. _Nachtr\u00e4ge zu Kants Kritik_, lxxxvi, with B.\nErdmann\u2019s comment, p. 35.\n\n[1232] A 211-12 = B 258. Cf. A 211 = B 257.\n\n[1233] A 211 = B 257.\n\n[1234] A 211-13 = B 258-60: first three paragraphs (first edition).\n\n[1235] A 213-14 = B 260-1: fourth paragraph (first edition).\n\n[1236] A 214-15 = B 261-2: fifth paragraph (first edition).\n\n[1237] Cf. above, pp. 189-90, 208 ff.\n\n[1238] B 257-8: first paragraph (second edition).\n\n[1239] Cf. B 291-3, partially quoted above, pp. 310-11. In the\n_Metaphysical First Principles_ (_III. Lehrsatz_, 4) the principle that\naction and reaction are always equal is similarly limited to the outer\nrelations of material bodies in space, and Kant adds that all change in\nbodies is motion. Cf. _W._ xi. p. 234; and above, p. 147.\n\n[1240] Above pp. 311-12; below, pp. 473-7.\n\n[1241] A 213 = B 260.\n\n[1242] The inconsistency of Kant\u2019s view of pure manifolds of time and\nspace with the argument of the _Analytic of Principles_ is too obvious\nto call for detailed comment.\n\n[1243] Cf. B 257.\n\n[1244] A 213-14 = B 260-1.\n\n[1245] B 257.\n\n[1246] Third edition, p. 438.\n\n[1247] Stout does not himself offer it as complete.\n\n[1248] _World as Will and Idea, W._ ii. pp. 544-5: Eng. trans. ii. pp.\n61-3.\n\n[1249] Cf. above, p. 197.\n\n[1250] A 212-13 = B 259.\n\n[1251] B 258.\n\n[1252] _Op. cit._ pp. 545-6: Eng. trans. p. 63.\n\n[1253] _Op. cit._ pp. 546-7: Eng. trans. pp. 63-5.\n\n[1254] Cf. above, p. 379.\n\n[1255] Cf. Stadler, _Grunds\u00e4tze_, p. 124.\n\n[1256] _Op. cit._ p. 546: Eng. trans. p. 63.\n\n[1257] Cf. above, pp. 97-8, 102 _n._, 165-6; below, pp. 429 ff., 447\nff., 547 ff.\n\n[1258] Cf. Adickes, K. p. 233 _n._\n\n[1259] A 222 = B 269. Cf. A 220 = B 268.\n\n[1260] A 148 ff. = B 187 ff.\n\n[1261] A 219 = B 266.\n\n[1262] This, by Kant\u2019s own account (A 232-4 = B 285-7), is what led him\nto adopt the title \u2018postulates.\u2019 A geometrical postulate does not add\nanything to the concept of its object but only defines the conditions of\nits production.\n\n[1263] Cf. above, pp. 38-9; below, pp. 398-9, 418 ff.\n\n[1264] Cf. A 220-3 = B 267-71.\n\n[1265] A 223 = B 270.\n\n[1266] Cf. A 220 = B 268.\n\n[1267] A 223 = B 270-1.\n\n[1268] K. p. 223 _n._\n\n[1269] A 224 = B 272.\n\n[1270] Cf. above, pp. 288-9.\n\n[1271] A 225 = B 272. Cf. above, pp. 394-6.\n\n[1272] A 225 = B 273. Italics not in Kant.\n\n[1273] _The Critical Philosophy of Kant_, i. p. 591.\n\n[1274] _Op. cit._ p. 595.\n\n[1275] A 226 = B 273-4.\n\n[1276] A 226 ff. = B 279 ff.\n\n[1277] A 218 = B 281.\n\n[1278] A 232 = B 284.\n\n[1279] A 231 = B 284.\n\n[1280] Cf. above, p. 309 ff.\n\n[1281] Kant\u2019s argument in the note to B 290 is that of his early essay\non _Negative Quantity_. Cf. below, pp. 527 ff., 533 ff., 536.\n\n[1282] A 236 = B 295.\n\n[1283] Cf. above, pp. xxxv-vi, xxxviii, 185-6, 191, 195-6, 257-8, 290-1,\n325 ff., 339.\n\n[1284] The mathematical illustrations which Kant proceeds to give (A 239\n= B 299) are peculiarly crude and off-hand in manner of statement. Cf.\n_per contra_ A 140 = B 179 for Kant\u2019s real view of the distinction\nbetween image, schema, and concept.\n\n[1285] Cf. above, pp. 195-6, 198, 339-42.\n\n[1286] A 243 = B 301.\n\n[1287] A 242 = B 302.\n\n[1288] Cf. A 248 = B 305.\n\n[1289] A 246-7 = B 303-4. A 247-8 = B 304-5 (beginning \u201cThought is the\naction,\u201d etc.) is merely a repetition of the preceding argument, and\nprobably represents a later intercalation.\n\n[1290] Beginning \u201cAppearances, so far as ...,\u201d which was omitted in the\nsecond edition. It probably constitutes, as Adickes maintains (_K._ p.\n254 _n._), the original beginning of this chapter. The \u201cas we have\nhitherto maintained\u201d of its second paragraph, which obviously cannot\napply to the pages which precede it in its present position, must refer\nto the argument of the _Analytic_.\n\n[1291] A 249, 251.\n\n[1292] Above, p. 204 ff.\n\n[1293] In large part it represents the Critical position as understood\nby Schopenhauer, who never succeeded in acquiring any genuine\nunderstanding of Kant\u2019s more mature teaching (cf. above, p. 366 _n._).\nSchopenhauer is correct in maintaining that one chief ground of Kant\u2019s\nbelief in the existence of things in themselves lies in his initial\nassumption that they must be postulated in order to account for the\ngiven manifold. Schopenhauer is also justified in stating that Kant,\nthough starting from the dualistic Cartesian standpoint, so far modified\nit as to conclude that the origin of this manifold must be \u201c_objective,\nsince there is no ground for regarding it as subjective_\u201d (_Parerga und\nParalipomena_, 1851 ed., p. 74 ff.). But for two reasons this is a very\nincomplete, and therefore extremely misleading, account of Kant\u2019s final\nteaching. In the first place, Schopenhauer fails to take account of\nKant\u2019s implied distinction between the sensations of the special senses\nand the manifold of outer sense. When Kant recognises that the\nsensations of the special senses are empirically conditioned, he is\nconstrained in consistency to distinguish between them and the manifold\nwhich constitutes the matter of all experiences (cf. above, p. 275 ff.).\nThings in themselves, in accounting for the latter, account also, but in\nquite indirect fashion, for the former. Though sensations are\nempirically conditioned, the _entire_ natural world is noumenally\ngrounded. Secondly, Kant\u2019s subjectivism undergoes a similar\ntransformation on its inner or mental side. The analysis of\nself-consciousness, which is given both in the _Deductions_ and in the\n_Paralogisms_, indicates with sufficient clearness Kant\u2019s recognition\nthat the form of experience is as little self-explanatory as its\ncontent, and that it must not, without such proof as, owing to the\nlimitations of our experience, we are debarred from giving, be regarded\nas more ultimate in nature. The realities which constitute and condition\nour mental processes are not apprehended in any more direct manner than\nthe thing in itself. When, therefore, Schopenhauer asserts in the _World\nas Will and Idea_ (_Werke, Frauenst\u00e4dt_, ii. p. 494, Eng. trans, ii. p.\n6) that Kant proves the world to be merely phenomenal by demonstrating\nthat it is conditioned by the intellect, he is emphasising what is least\ncharacteristic in Kant\u2019s teaching. Schopenhauer\u2019s occasional\nidentification of the intellect with the brain--the nearest\napproximation in his writings to what may be described as\nphenomenalism--itself suffices to show how entirely he is lacking in any\nfirm grasp of Critical principles.\n\n[1294] As we have noted (above, p. 204 ff.), the doctrine of the\ntranscendental object was entirely eliminated from those main sections\nthat were rewritten or substantially altered in the second edition,\nnamely, the chapters on the _Transcendental Deduction_, on _Phenomena\nand Noumena_, and on the _Paralogisms_. That it remained in the section\non _Amphiboly_, in the _Second Analogy_, and in the _Antinomies_ is\nsufficiently explained by Kant\u2019s unwillingness to make the very\nextensive alterations which such further rewriting would have involved.\n\n[1295] A 251.\n\n[1296] Not even, as Kant teaches in his doctrine of inner sense, in the\ninner world of apperception, cf. above, p. 295 ff.\n\n[1297] Kant claims in the _Dialectic_ that this process is also\nunavoidable, constituting what he calls \u201ctranscendental illusion.\u201d\n\n[1298] A 254-7 = B 310-12.\n\n[1299] A 255 = B 310-11.\n\n[1300] Cf. below, p. 412 ff.\n\n[1301] A 256 = B 312. For A 257 = B 312 on the empirical manner of\ndistinguishing between the sensuous and the intelligible, cf. above, pp.\n143 ff., 149 ff.\n\n[1302] Cf. above, pp. 143-4, 147, 214-15, 291 ff.\n\n[1303] Kant here (A 286 = B 342) speaks of this concept of the noumenon\n_as an object of non-sensuous intuition_ as being \u201cmerely negative.\u201d\nThis is apt to confuse the reader, as he usually comes to it after\nhaving read the passage introduced into the chapter on _Phenomena and\nNoumena_ in the second edition, in which, as above noted (p. 409), Kant\ndescribes this meaning of the term as positive, in distinction from its\nmore negative meaning as signifying a thing merely so far as it is not\nan object of _our_ sense-intuition. Cf. below, p. 413.\n\n[1304] Kant\u2019s meaning here is not quite clear. He may mean either that\nthe categories as such are inapplicable to things in themselves, or\nthat, as this form of intuition is altogether different from our own, it\nwill not help in giving meaning to the categories. What follows would\nseem to point to the former view.\n\n[1305] A 286 = B 343.\n\n[1306] A 287-8 = B 344.\n\n[1307] A 288 = B 345.\n\n[1308] A 288 = B 344. Kant allowed the section within which this passage\noccurs to remain, without the least modification, in the second edition.\n\n[1309] Benno Erdmann\u2019s explanation (_Kriticismus_, p. 194) of Kant\u2019s\nomission of all references to the transcendental object, namely, because\nof their being likely to conduce to a mistaken idealistic interpretation\nof his teaching, we cannot accept. As already argued (above, p. 204\nff.), they represent a view which he had quite definitely and\nconsciously outgrown.\n\n[1310] B 306. Cf. above, pp. 290-1.\n\n[1311] B 308. This, it may be noted, is in keeping with the passages\nabove quoted from the section on _Amphiboly_.\n\n[1312] A 255 = B 311.\n\n[1313] Cf. above, p. 404 ff., especially pp. 409-10; also above, p. 331.\n\n[1314] In order to form an adequate judgment upon Kant\u2019s justification\nfor distinguishing between appearance and reality the reader must bear\nin mind (1) the results obtained in the _Transcendental Deduction_\n(above, p. 270 ff.); (2) the discussions developed in the _Paralogisms_\n(below, p. 457 ff.); (3) the treatment of noumenal causality, that is of\nfreedom, in the _Third_ and _Fourth Antinomies_; (4) the many connected\nissues raised in the _Ideal_ (below, pp. 534-7, 541-2), and in the\n_Appendix to the Dialectic_ (below, p. 543 ff.). Professor Dawes Hicks\nis justified in maintaining in his book, _die Begriffe Ph\u00e4nomenon und\nNoumenon in ihrem Verh\u00e4ltniss zu einander bei Kant_ (Leipzig, 1897, p.\n167)--a work which unfortunately is not accessible to the English\nreader--that \u201cthe thing in itself is by no means a mere excrescence or\naddendum of the Kantian system, but forms a thoroughly necessary\ncompletion to the doctrine of appearances. At every turn in Kant\u2019s\nthought the doctrine of the noumenon, in one form or another, plays an\nessential part.\u201d Indeed it may be said that to state Kant\u2019s reasons for\nasserting the existence of things in themselves, is to expound his\nphilosophy as a whole. Upon this question there appears in Kant the same\nalternation of view as in regard to his other main tenets. On Kant\u2019s\ndiscussion of the applicability of the category of existence to things\nin themselves, cf. above, p. 322 ff. Also, on Kant\u2019s extension of the\nconcepts possibility and actuality to noumena, cf. above, pp. 391 ff.,\n401-3.\n\n[1315] \u2018Ideal\u2019 and \u2018Idealist\u2019 are printed with capitals, to mark the\nvery special sense in which these terms are being used. As already noted\n(above, p. 3), the same remark applies to the term \u2018Reason.\u2019\n\n[1316] Cf. above, pp. xli-ii, xliv, liii-v, 331.\n\n[1317] A 260 ff. = B 316 ff.\n\n[1318] Cf. above, pp. 38-9, 119, 131-3, 338-9, 394-400.\n\n[1319] Above, p. xxx ff., and below, p. 601 ff.\n\n[1320] Cf. A 267 = B 323.\n\n[1321] Cf. Adickes\u2019 _Systematik_, pp. 60, 70, 72, and 111-12.\n\n[1322] A 270 = B 326.\n\n[1323] Cf. A 264 = B 319, and A 266 = B 322.\n\n[1324] Cf. below, pp. 563-5, 589 ff., 601 ff.\n\n[1325] I have dwelt upon this at length in my _Studies in the Cartesian\nPhilosophy_.\n\n[1326] A 271 = B 327.\n\n[1327] The un-Critical character of Kant\u2019s doctrine of the pure concept\nhas already been noted (above, pp. 418-19), and need not be further\ndiscussed.\n\n[1328] A 272 = B 328.\n\n[1329] A 273 = B 329.\n\n[1330] This is Leibniz\u2019s mode of stating the absolutist view of thought\n(cf. above, p. xxx ff.) to which, as we shall find, Kant gives much more\nadequate and incomparably deeper formulation in the _Dialectic_. Cf. pp.\n430, 547 ff., 558 ff.\n\n[1331] Adickes, _K._ p. 272 _n._, allows that the passage may be of\nearlier origin than the passages which precede and follow it.\n\n[1332] Pp. 214-15.\n\n[1333] As such it is commented on above, p. 410 ff.\n\n[1334] _Loc. cit._\n\n[1335] A 290 = B 347.\n\n[1336] Cf. above, p. 409 ff.\n\n[1337] Kant\u2019s commentators have frequently misrepresented this aspect of\nhis teaching. Cf. below, pp. 498, 520-1, 527-37, 541-2, 543 ff., 555,\n558-61.\n\n[1338] A 490 = B 518.\n\n[1339] Cf. above, pp. 416-17.\n\n[1340] Those readers who are not already well acquainted with the\nargument of the _Dialectic_ may be recommended to pass at once to p.\n441. What here follows presupposes acquaintance with the nature and\npurposes of the main divisions of the _Dialectic_.\n\n[1341] Introd. to _Reflexionen_, Bd. ii.\n\n[1342] _W._ x. p. 123 ff. Cf. above, pp. 219-20.\n\n[1343] Cf. _Dissertation_, \u00a7 27 _n._\n\n[1344] _Op. cit._ Cf. \u00a7 24 with \u00a7 27.\n\n[1345] _Op. cit._ \u00a7 27.\n\n[1346] Cf. ii. 567, 571, 584, 585.\n\n[1347] Cf. ii. 1251 and 586.\n\n[1348] Cf. below, pp. 458, 488 ff.\n\n[1349] In _Reflexionen_ ii. 573, 576, and 582 we find Kant in the very\nact of so doing. _Compositio_, _co-ordinatio_, and _commercium_ are\ntreated as synonymous terms.\n\n[1350] The problem of freedom is first met with in Kant\u2019s _Lectures on\nMetaphysics_ (P\u00f6litz, edition of 1821, pp. 89, 330), but is not there\ngiven as an antinomy, and is treated as falling within the field of\ntheology. In _Reflexion_ ii. 585, also, it is equated in terms of the\ncategory of ground and consequence, with the concept of Divine\nExistence, the \u201c_absolute_ or _primum contingens_ (_libertas_).\u201d Upon\nelimination of theology, and therefore of the cosmological argument,\nfrom the sphere of antinomy, Kant raised freedom to the rank of an\nindependent problem.\n\n[1351] A 462 = B 490.\n\n[1352] Cf. below, pp. 498-9, 571 ff.\n\n[1353] Cf. below, p. 454, with references in _n._ 1.\n\n[1354] A 507 = B 535. Cf. below, pp. 481, 545-6.\n\n[1355] Cf. ii. 93, 94, 95, 1233, 1247.\n\n[1356] This is the view represented in _Reflexionen_ ii. 94, 95.\n\n[1357] Cf. _Reflexionen_ ii. 124.\n\n[1358] Cf. _Reflexionen_ ii. 95.\n\n[1359] Cf. below, p. 457.\n\n[1360] Cf. ii. 86 ff.\n\n[1361] Cf. _Reflexionen_ ii. 114-15.\n\n[1362] B 394 _n._ Immortality is here taken as representing the Idea of\nthe soul as unconditioned substance.\n\n[1363] Cf. below, p. 454, with further references in _n._ 1.\n\n[1364] _Systematik_, pp. 115-16.\n\n[1365] Above, p. 334.\n\n[1366] This conclusion is supported by the evidence of the\n_Reflexionen_: they contain not a single reference to schematism.\n\n[1367] A 293 = B 349.\n\n[1368] Pp. 173-4.\n\n[1369] Cf. A 61 = B 85.\n\n[1370] Adickes, _Systematik_, p. 77 ff.\n\n[1371] Cf. Kant\u2019s _caveat_ in A 293 = B 349 against identifying\ndialectic with the doctrine of probable reasoning.\n\n[1372] Pp. 427-8.\n\n[1373] A 298 = B 355.\n\n[1374] Cf. above, p. 332.\n\n[1375] Reicke, i. p. 105.\n\n[1376] _Op. cit._ i. pp. 109-10.\n\n[1377] A 301-2 = B 358.\n\n[1378] A 303 = B 359.\n\n[1379] A 305 = B 362.\n\n[1380] The wording of the concluding sentence of the third paragraph (A\n307 = B 363-4) is so condensed as to be misleading. \u201cIt [viz. the\nprinciple of causality] makes the unity of experience possible, and\nborrows nothing from the Reason. The latter, if it were not for this\n[its indirect] reference [through mediation of the understanding] to\npossible experience, could never [of itself], from mere concepts, have\nimposed a synthetic unity of that kind.\u201d\n\n[1381] A 310 = B 366.\n\n[1382] _Schein des Schliessens_ would seem to be here used in that\nsense.\n\n[1383] Cf. above, p. 424.\n\n[1384] Cf. also A 669 = B 697; A 680 = B 709.\n\n[1385] Cf. Vaihinger, \u201cKant--ein Metaphysiker?\u201d in _Philosophische\nAbhandlungen_ (Sigwart _Gedenkschrift_), p. 144.\n\n[1386] A 312 = B 368.\n\n[1387] A 313 = B 370.\n\n[1388] A 316-17 = B 373. The context of this passage is a defence of\nPlato\u2019s _Republic_ against the charge that it is Utopian, because\nunrealisable.\n\n[1389] A 317-18 = B 374-5.\n\n[1390] _Reflexionen_ ii. 1240. Cf. Schopenhauer: _World as Will and\nIdea_ (_Werke_, ii. p. 277: Eng. trans. i. p. 303): \u201cThe _Idea_ is the\nunity that falls into multiplicity on account of the temporal and\nspatial form of our intuitive apprehension; the _concept_, on the\ncontrary, is the unity reconstructed out of multiplicity by the\nabstraction of our reason; the latter may be defined as _unitas post\nrem_, the former as _unitas ante rem_.\u201d\n\n[1391] _Lectures on Metaphysics_ (P\u00f6litz, 1821), p. 79.\n\n[1392] _Reflexionen_ ii. 1243.\n\n[1393] _Lectures on Metaphysics_, pp. 308-9.\n\n[1394] _Reflexionen_ ii. 1244.\n\n[1395] _Reflexionen_ ii. 1254.\n\n[1396] _Reflexionen_ ii. 1258.\n\n[1397] _Reflexionen_ ii. 1259.\n\n[1398] _Reflexionen_ ii. 1260.\n\n[1399] A 320 = B 376-7.\n\n[1400] A 321 = B 377.\n\n[1401] A 323-4 = B 380-1. Cf. below, pp. 480, 529, 559-60.\n\n[1402] Regarding the _progressive series_ from the conditioned to its\nconsequences, cf. A 336-7 = B 393-4, A 410-11 = B 437-8, A 511 = B 539.\n\n[1403] A 333 = B 390.\n\n[1404] Cf. above, pp. 418, 436, 439-40; below, pp. 473-7, 520-1, 537,\n543 ff., 575.\n\n[1405] Cf. A 335.\n\n[1406] Cf. A 337-8 = B 394-6 and note appended to B 394.\n\n[1407] A 336 = B 393.\n\n[1408] Cf. A 671 = B 699; above, pp. 426, 430, 436; below, pp. 552-4,\n572 ff.\n\n[1409] On the difference between the ascending and the descending\nseries, cf. A 331-2 = B 338 and A 410-11 = B 437-8.\n\n[1410] The questions raised in the two introductory paragraphs (A 336-40\n= B 396-8) as to the content of the Ideas, their problematic character,\nand their possibility as concepts, are first adequately discussed in\nlater chapters. The three new terms here introduced, Paralogism,\nAntinomy, and Ideal, can also best be commented upon in their own\nspecial context.\n\n[1411] A 341 = B 399.\n\n[1412] Cf. below, pp. 466, 470.\n\n[1413] A 347.\n\n[1414] A 345-6 = B 403-4.\n\n[1415] Cf. A 354-5.\n\n[1416] Cf. above, p. 437.\n\n[1417] A 348.\n\n[1418] A 351.\n\n[1419] A 363-4.\n\n[1420] A 351.\n\n[1421] _K._ 688 _n._\n\n[1422] A similar criticism holds true of the conception of identity\nemployed in the third _Paralogism_, and arbitrarily equated with the\ncategories of quantity.\n\n[1423] Cf. A 355-6.\n\n[1424] It is very forcibly developed in Mendelssohn\u2019s \u201cPh\u00e4don\u201d (1767)\n(_Gesammelte Schriften_, 1843, ii. p. 151 ff.). This is a work with\nwhich Kant was familiar. Cf. below, p. 470.\n\n[1425] This is the argument which William James has expounded in his\ncharacteristically picturesque style. \u201cTake a sentence of a dozen words,\nand take twelve men and tell to each one word. Then stand the men in a\nrow or jam them in a bunch, and let each think of his word as intently\nas he will; nowhere will there be a consciousness of the whole sentence\u201d\n(_Principles of Psychology_, i. p. 160).\n\n[1426] A 363 _n._ Cf. below, pp. 461-2.\n\n[1427] A 356. Cf. Adickes, _K._ p. 688 _n._\n\n[1428] The argument is here in harmony with Kant\u2019s definition of\ntranscendental illusion.\n\n[1429] A 358.\n\n[1430] A 361.\n\n[1431] A 364.\n\n[1432] William James\u2019s psychological description of self-consciousness\nis simply an extension of this illustration. Cf. _Principles of\nPsychology_, i. p. 339; quoted above, p. 278 _n._\n\n[1433] A 363 _n._\n\n[1434] A 362-3 and A 364. We must also, however, bear in mind that in\nthis chapter Kant occasionally argues in _ad hominem_ fashion from the\npoint of view of the position criticised.\n\n[1435] Cf. A 353-4.\n\n[1436] Cf. Adickes, _K._ p. 695 _n._\n\n[1437] A 366.\n\n[1438] P. 301 ff.\n\n[1439] The note to A 344 has evidently got displaced; it must, as\nAdickes points out, belong to A 404.\n\n[1440] Cf. above, pp. 320, 455.\n\n[1441] A 371-2.\n\n[1442] A 380-1.\n\n[1443] Cf. A 383.\n\n[1444] A 383.\n\n[1445] A 383.\n\n[1446] A 381.\n\n[1447] The first four paragraphs are probably a later intercalation\n(Adickes, _K._ p. 708 _n._), since they connect both with the\nintroductory sections of the _Dialectic_ and with the _Introduction_ to\nthe _Critique_. Also, the opening words of the fifth paragraph seem to\nrefer us not to anything antecedent in this section, but directly to the\nconcluding passages of the fourth _Paralogism_.\n\n[1448] A 385.\n\n[1449] A 393.\n\n[1450] A 387.\n\n[1451] Cf. above, pp. 215-16.\n\n[1452] A 393-4.\n\n[1453] A 394.\n\n[1454] Pp. 326-7.\n\n[1455] Pp. 327-8.\n\n[1456] A 402. Cf. B 407.\n\n[1457] _K._ p. 717 _n._\n\n[1458] Cf. below, p. 470.\n\n[1459] B 406 ff.\n\n[1460] _Kriticismus_, p. 227, cf. p. 106 ff.\n\n[1461] A. H. Ulrichs, _Institutiones logicae et metaphysicae_ (1785).\n\n[1462] In his review of Kant\u2019s _Prolegomena_ in the _Allgemeine Deutsche\nBibliothek_ (1784).\n\n[1463] Obviously by categories Kant here really means schemata. Cf. A\n348, where Kant states that \u201cpure categories ... have in themselves no\nobjective meaning.... Apart from intuition they are merely functions of\na judgment, without content.\u201d\n\n[1464] Above, pp. 404 ff., 413 ff.\n\n[1465] B 408.\n\n[1466] _Critical Philosophy_, ii. p. 34. So also in Watson\u2019s _Kant\nExplained_, p. 244.\n\n[1467] Caird (_op. cit._ p. 35) takes account of Kant\u2019s conception of a\npossible intuitive understanding, but illegitimately assumes that by it\nhe must mean a _creative_ understanding.\n\n[1468] Cf. above, p. 295 ff.\n\n[1469] Cf. B 415 _n._ In B xxxix. _n._ (at the end), quoted above pp.\n309-10, Kant is careful to point out that the representation of\nsomething permanent is by no means identical with permanent\nrepresentation.\n\n[1470] P. 463.\n\n[1471] Namely, as _Refutation of Idealism_, B 274 ff. Cf. above, p. 308\nff.\n\n[1472] Cf. above, pp. 457, 462-3.\n\n[1473] A 402.\n\n[1474] Cf. above, p. 466.\n\n[1475] B 413-15.\n\n[1476] _Gesammelte Schriften_, ii. p. 151 ff.\n\n[1477] _Op. cit._ p. 121 ff.\n\n[1478] _Op. cit._ pp. 128 ff., 168.\n\n[1479] _Op. cit._ p. 125 ff.\n\n[1480] Regarding the value of the hypotheses propounded by Kant in his\nnote to B 415, cf. below, p. 543 ff.\n\n[1481] P. 321 ff.\n\n[1482] Cf. above, p. 467.\n\n[1483] Pp. 473-7.\n\n[1484] _Kriticismus_, p. 226.\n\n[1485] B 424.\n\n[1486] B 421.\n\n[1487] B 424-5.\n\n[1488] B 425-6. Cf. above, pp. lvi-lxi; below, p. 570 ff.\n\n[1489] The only approach to such a reference is in B 426-7, noted above,\np. 471.\n\n[1490] A 672 = B 700. Cf. below, p. 554.\n\n[1491] A 649 = B 677-8. Tetens in his _Philosophische Versuche_ (1777)\nhad devoted an entire chapter to this question. His term _Grundkraft_ is\nthat which Kant here employs. Cf. _Philosophische Versuche_, Bd. i.,\n_Elfter Versuch_: \u201cConcerning the fundamental power of the human soul.\u201d\nIncidentally Tetens discusses Rousseau\u2019s suggestion that this\nfundamental power consists in man\u2019s capacity for perfecting himself. Cf.\nKant\u2019s _Lectures on Metaphysics_ (P\u00f6litz, 1821, p. 192 ff.).\n\n[1492] A 682-4 = B 710-12. A 771-2 = B 799 in the _Methodology_ is\nsimilarly ambiguous, though tending to the spiritualist mode of\nformulation.\n\n[1493] Cf. above, pp. 275-6, 279 ff., 312 ff., 384-5, 464-5.\n\n[1494] Cf. end of B xxxix. _n._, quoted above, pp. 309-10.\n\n[1495] A 405 = B 432.\n\n[1496] A 408 = B 435.\n\n[1497] Cf. A 414 = B 441, where it is stated that there is no\ntranscendental Idea of the substantial.\n\n[1498] Cf. above, p. 434 ff.\n\n[1499] A 419 = B 447.\n\n[1500] A 420 = B 447.\n\n[1501] A very curious sentence in Kant\u2019s letter to Schulze (_W._ x. pp.\n344-5, quoted above, p. 199) seems to be traceable to this source.\n\n[1502] Cf. below, pp. 529, 559-60, and above, pp. 199-200, 433-4, 451.\nFor A 410-11 = B 439-40 on the difference between the ascending and\ndescending series, cf. A 331-2 = B 387-8 and A 336-7 = B 393-4.\n\n[1503] A 420 = B 448.\n\n[1504] Cf. _per contra_ A 486 = B 514.\n\n[1505] The limitation of Kant\u2019s discussion to space, time, and causality\nis, of course, due to his acceptance of the current view that the\nconcepts of infinity and continuity are derived from our intuitions of\nspace and time. As we have already noted in discussing his intuitional\ntheory of mathematical reasoning (above, pp. 40-1, 117 ff., 128 ff.), he\nfails to extend to mathematical concepts his own \u201ctranscendental\u201d view\nof the categories, namely, as _conditioning_ the possibility of\nintuitional experience. Such concepts as order, plurality, whole and\npart, continuity, infinity, are prior to time and space in the logical\norder of thought; and to be adequately treated must be considered in\ntheir widest application.\n\n[1506] Cf. A 507 = B 535, and above, p. 431 ff.; below, pp. 501, 545-6.\n\n[1507] Cf. Kant\u2019s posthumously published _Transition from the\nMetaphysical First Principles of Natural Science to Physics_\n(_Altpreussische Monatsschrift_, 1882), pp. 279-80: \u201cIf we take in\nregard to space, not its definition, but only an _a priori_ proposition,\n_e.g._ that space is a whole which must be thought only as part of a\nstill greater whole, it is clear ... that it is an irrational magnitude,\nmeasurable indeed, but in its comparison with unity transcending all\nnumber.\u201d \u201cIf space is something objectively existent, it is a magnitude\nwhich can exist only as part of another given magnitude.\u201d\n\n[1508] Cf. Schopenhauer, _World as Will and Idea_ (_Werke, Frauenst\u00e4dt_,\nii. pp. 585-6; Eng. trans, ii. pp. 107-8). \u201cI find and assert that the\nwhole antinomy is a mere delusion, a sham fight. Only the assertions of\nthe antitheses really rest upon the forms of our faculty of knowledge,\n_i.e._ if we express it objectively, on the necessary, _a priori_\ncertain, most universal laws of nature. Their proofs alone are t\nherefore drawn from objective grounds. On the other\nhand, the assertions and proofs of the theses have no other than a\nsubjective ground, rest solely on the weakness of the reasoning\nindividual; for his imagination becomes tired with an endless\nregression, and therefore he puts an end to it by arbitrary assumptions,\nwhich he tries to smooth over as well as he can; and his judgment,\nmoreover, is in this case paralysed by early and deeply imprinted\nprejudices. On this account the proof of the thesis in all the four\nconflicts is throughout a mere sophism, while that of the antithesis is\na necessary inference of the reason from the laws of the world as idea\nknown to us _a priori_. It is, moreover, only with great pains and skill\nthat Kant is able to sustain the thesis, and make it appear to attack\nits opponent, which is endowed with native power.... I shall show that\nthe proofs which Kant adduces of the individual theses are sophisms,\nwhile those of the antitheses are quite fairly and correctly drawn from\nobjective grounds.\u201d]\n\n[1509] Cf. F. Erhardt\u2019s _Kritik der Kantischen Antinomienlehre_ (1888),\na brief but excellent analysis of this section of the _Critique_.\n\n[1510] \u00a7 1 _n._\n\n[1511] Cf. A 431-2 = B 460-1: \u201c...the concept [of the infinite] is not\nthe concept of a maximum; by it we think only its relation to any\nassignable unit, in respect to which it is greater than all number.\u201d\n\n[1512] Cf. Kant\u2019s statement in the _Observation_ to this antithesis, A\n431-3 = B 459-61.\n\n[1513] Kant regarded the point as a limit, _i.e._ as a boundary\n(_Dissertation_, \u00a7 14, 4; \u00a7 15, C: \u201cThe simple in space is not a part\nbut a limit\u201d; A 169-70 = B 211); whereas certain modern mathematicians\ntake the point as one of the undefined elements. When the point is\nregarded in this latter manner, space may perhaps be satisfactorily\ndefined as a set of points. In arguing for the antithesis, and in the\npassages just cited, Kant also assumes that, in the case of space, the\nproperties of the class are determined by the properties of its\nelements. This questionable assumption is involved in his assertion that\nspace can consist only of spaces.\n\n[1514] A 438 = B 466.\n\n[1515] A 439-41 = B 467-9.\n\n[1516] A 441 = B 469.\n\n[1517] Developed in the _Dissertation_ (1770).\n\n[1518] Zweites Hauptst\u00fcck, Lehrsatz 4, Anmerkung 1. Cf. also Anmerkung\n2.\n\n[1519] _Principles of Mathematics_, i. p. 460.\n\n[1520] Cf. above, p. 481 _n._ 2.\n\n[1521] P. 489 _n._\n\n[1522] Cf. Schopenhauer, _World as Will and Idea_ (_Werke_,\n_Frauenst\u00e4dt_, ii. p. 590; Eng. trans. ii. pp. 111-12). \u201cThe argument\nfor the third thesis is a very fine sophism, and is really Kant\u2019s\npretended principle of pure reason itself entirely unadulterated and\nunchanged. It tries to prove the finiteness of the series of causes by\nsaying that, in order to be _sufficient_, a cause must contain the\ncomplete sum of the conditions from which the succeeding state, the\neffect, proceeds. For the completeness of the determinations present\n_together_ in the state which is the cause, the argument now substitutes\nthe completeness of the series of causes by which that state itself was\nbrought to actuality; and because completeness presupposes the condition\nof being rounded off or closed in, and this again presupposes\nfiniteness, the argument infers from this a first cause, closing the\nseries and therefore unconditioned. But the juggling is obvious. In\norder to conceive the state A as the sufficient cause of the state B, I\nassume that it contains the sum of the necessary determinations from the\ncoexistence of which the state B inevitably follows. Now by this my\ndemand upon it as a _sufficient_ cause is entirely satisfied, and has no\ndirect connection with the question how the state A itself came to be;\nthis rather belongs to an entirely different consideration, in which I\nregard the said state A no more as cause, but as itself an effect; in\nwhich case another state again must be related to it, just as it was\nrelated to B. The assumption of the finiteness of the series of causes\nand effects, and accordingly of a first beginning, appears nowhere in\nthis as necessary, any more than the presentness of the present moment\nrequires us to assume a beginning of time itself.\u201d\n\n[1523] _Op. cit._ p. 24.\n\n[1524] For comment upon Kant\u2019s defence of his procedure cf. below, p.\n496.\n\n[1525] Cf. Kant\u2019s _Observation_ on the thesis.\n\n[1526] A 451 = B 479.\n\n[1527] Cf. also A 451 = B 479.\n\n[1528] Cf. Schopenhauer, _op. cit._ p. 591; Eng. trans. p. 113. \u201cThe\nfourth conflict is ... really tautological with the third; and the proof\nof the thesis is also essentially the same as that of the preceding one.\nKant\u2019s assertion that every conditioned presupposes a complete series of\nconditions, and therefore a series which ends with an unconditioned, is\na _petitio principii_ which must simply be denied. Everything\nconditioned presupposes nothing but its condition; that this is again\nconditioned raises a new consideration which is not directly contained\nin the first.\u201d\n\n[1529] Above, p. 494.\n\n[1530] A 459 = B 487.\n\n[1531] Jean Jacques Dortous de Mairan (1678-1771), physicist and\nmathematician. In 1740 he succeeded Fontenelle as perpetual Secretary of\nthe French Academy of Sciences.\n\n[1532] Cf. above, pp. 435, 495 _n._ 4.\n\n[1533] A 448-50 = B 476-8.\n\n[1534] Cf. above, p. 427 ff.; below, pp. 520-1, 527-37, 541 ff.\n\n[1535] A 462 = B 490.\n\n[1536] Cf. above, pp. 434 ff., 479.\n\n[1537] A 476 = B 504.\n\n[1538] A 484 = B 512.\n\n[1539] _Ibid._\n\n[1540] A 483 = B 511.\n\n[1541] A 485 = B 513.\n\n[1542] Cf. above, p. 481; below, pp. 545-6.\n\n[1543] Kant is here playing on the double meaning of the German\n\u201csinnleeres\u201d--\u201cempty of sense\u201d and \u201cnon-sense.\u201d\n\n[1544] A 489 = B 517.\n\n[1545] A 490 = B 518.\n\n[1546] Above, p. 426 ff.\n\n[1547] A 490 = B 518.\n\n[1548] Cf. above p. 204 ff.\n\n[1549] A 494 = B 522-3.\n\n[1550] A 495 = B 523.\n\n[1551] Cf. A 494 = B 522-3: \u201c...we can say of the transcendental object\nthat it is given in itself prior to all experience.\u201d\n\n[1552] A 496 = B 524.\n\n[1553] A 491 = B 519.\n\n[1554] Pp. 306-7.\n\n[1555] A 497 = B 525.\n\n[1556] A 501-2 = B 529-30.\n\n[1557] A 506 = B 534.\n\n[1558] Cf. end of passage: \u201cThere can be no lack of conditions that are\ngiven through this regress.\u201d\n\n[1559] Cf. below, pp. 507-8.\n\n[1560] Cf. below, pp. 507-9.\n\n[1561] _K._ p. 414 _n._ The two last paragraphs of Section VII., which\ncorrect its argument, that of the Transcendental Aesthetic, are probably\nlater additions.\n\n[1562] A 508 = B 536.\n\n[1563] _Loc. cit._\n\n[1564] As to the distinction between the ascending and the descending\nseries, cf. above, pp. 453 _n._, 484.\n\n[1565] Cf. A 522 = B 549-50.\n\n[1566] A 514 = B 542.\n\n[1567] Above, p. 506.\n\n[1568] Cf. A 522 = B 550.\n\n[1569] A 515 = B 543.\n\n[1570] A 519-20 = B 547-8.\n\n[1571] When Kant adds (A 521 = B 549), \u201cand therefore absolutely also,\u201d\nhe inconsistently reverts to the position ambiguously suggested in A 499\n= B 527. Cf. above, p. 506.\n\n[1572] A 523-6 = B 551-4.\n\n[1573] The assertion of infinite divisibility is not applicable, Kant\nstates (A 526-7 = B 554-5), to bodies as organised, but only to bodies\nas mere occupants of space. Organisation involves distinction of parts,\nand therefore discreteness. How far organisation can go in organised\nbodies, experience alone can show us.\n\n[1574] P. 508.\n\n[1575] A 528 = B 556.\n\n[1576] Cf. above, pp. 345-7.\n\n[1577] A 535-6 = B 563-4.\n\n[1578] Cf. A 537 = B 564-5; also A 546 = B 574-5, in which Kant asserts\nthat man knows himself not only through the senses but \u201calso through\npure apperception, and indeed in actions and inner determinations which\ncannot be reckoned as impressions of the senses.\u201d Such statements would\nseem to show that, at the time of writing, Kant had not yet developed\nhis doctrine of inner sense.\n\n[1579] A 532 = B 560.\n\n[1580] A 536-7 = B 564-5.\n\n[1581] A 533 = B 561.\n\n[1582] A 536 = B 564.\n\n[1583] A 538 = B 566.\n\n[1584] Cf. Kant\u2019s _Uebergang von der metaph. Anfangsgr\u00fcnde der\nNaturwissenschaft zur Physik (Altpreussische Monatsschrift_ (1882), pp.\n272-3).\n\n[1585] A 549 = B 577. Italics not in Kant.\n\n[1586] In A 540 = B 568 a different and less satisfactory view finds\nexpression.\n\n[1587] A 542 = B 570.\n\n[1588] A 544 = B 572.\n\n[1589] A 546-7 = B 574-5.\n\n[1590] A 548 = B 576.\n\n[1591] A 552 = B 580.\n\n[1592] A 553 = B 581.\n\n[1593] A 557 = B 585.\n\n[1594] A 553-4 = B 581-2.\n\n[1595] Cf. A 537-41 = B 565-9 and A 544 = B 572.\n\n[1596] Cf. A 566 = B 594.\n\n[1597] Cf. above, p. 204 ff.\n\n[1598] A 559 = B 587.\n\n[1599] A 561 = B 589.\n\n[1600] A 565 = B 593.\n\n[1601] A 567 = B 593.\n\n[1602] For Kant\u2019s comparison of his Ideas with those of Plato, cf.\nabove, pp. 447-9.\n\n[1603] \u00a7\u00a7 803 ff. in 5th edition (Halle, 1763).\n\n[1604] A 578 = B 606.\n\n[1605] A 580 = B 608.\n\n[1606] Cf. above, p. 418 ff.\n\n[1607] A 272-4 = B 328-30.\n\n[1608] Cf. Kant\u2019s distinction between distributive and collective unity\nin A 582-3 = B 610 with A 644 = B 672.\n\n[1609] A 583 = B 611.\n\n[1610] A 603 = B 631.\n\n[1611] A 603-4 = B 631-2.\n\n[1612] Cf. below, pp. 533, 536.\n\n[1613] A 592 = B 620.\n\n[1614] A 593 = B 621.\n\n[1615] Cf. A 4-5 = B 8-9; A 735-8 = B 763-6.\n\n[1616] Cf. above, pp. 427-8, and references there given.\n\n[1617] Cf. above, p. 424.\n\n[1618] Cf. above, p. 392 ff.\n\n[1619] _K._ p. 475 _n._\n\n[1620] A 603 = B 631.\n\n[1621] Cf. above, p. 527. The concluding paragraphs A 613-14 = B 641-2\ncan best be treated later in another connection. Cf. below, p. 536.\n\n[1622] A 614 = B 642.\n\n[1623] A 613 = B 641.\n\n[1624] A 616 = B 644.\n\n[1625] A 619-20 = B 647-8.\n\n[1626] Cf. below, pp. 541-2, 552 ff.\n\n[1627] A 620 = B 648.\n\n[1628] A 624 = B 652.\n\n[1629] _Universal Natural History and Theory of the Heavens_ (1755).\n\n[1630] _Critique of Judgment_, \u00a7\u00a7 64, 65.\n\n[1631] Hamann completed his translation of Hume\u2019s _Dialogues on Natural\nReligion_ on August 7, 1780 (cf. Hamann\u2019s _Werke_, vi. 154 ff.): and\nKant, notwithstanding his being occupied in finishing the _Critique_,\nread through the manuscript. It is highly likely that this first perusal\nof Hume\u2019s _Dialogues_ not only confirmed Kant in his negative attitude\ntowards natural theology, but also enabled him to define more clearly\nthan he otherwise would have done, the negative consequences of his own\nCritical principles. The chapter on the _Ideal_, as we have already\nobserved (above, pp. 434-5, 527-9, 531), was probably one of the last\nparts of the _Critique_ to be brought into final form. It does not seem\npossible, however, to establish in any specific manner the exact\ninfluence which Hume\u2019s _Dialogues_ may thus have exercised upon the\nargument of this portion of the _Critique_. When Schreiter\u2019s translation\nof the _Dialogues_ appeared in 1781, Hamann, not unwilling to escape the\nnotoriety of seeming to father so sceptical a work, withdrew his own\ntranslation.\n\n[1632] This is the main point of Hume\u2019s argument in Section XI. of his\n_Enquiry concerning the Human Understanding_.\n\n[1633] A 631 = B 659.\n\n[1634] A 641 = B 669.\n\n[1635] Cf. above, p. 407 ff., and below, p. 552 ff.\n\n[1636] Cf. above, p. 454, with further references in _n._ 1.\n\n[1637] Cf. above, pp. 536-7.\n\n[1638] A 642 = B 670.\n\n[1639] A 769-82 = B 797-810.\n\n[1640] A xiv, B xxiii-iv, and _Reflexionen_ ii. 1451: \u201cIn metaphysics\nthere can be no such thing as uncertainty.\u201d Cf. above, pp. 10, 35.\n\n[1641] A 770-1 = B 798-9.\n\n[1642] A 772 = B 800.\n\n[1643] A 775 = B 803.\n\n[1644] Cf. A 781-2 = B 809-10.\n\n[1645] Cf. above, pp. 481, 501.\n\n[1646] A 777-8 = B 805-6.\n\n[1647] A 782 = B 810.\n\n[1648] Cf. above, pp. 97-8, 102, 390-1, 426 ff., 447 ff.\n\n[1649] A 651 = B 679.\n\n[1650] _Loc. cit._\n\n[1651] A 653 = B 681.\n\n[1652] _Dissertation_, \u00a7 30.\n\n[1653] The extremely un-Critical reason which Kant here (A 647 = B 675)\ngives for its necessarily remaining hypothetical is the \u201cimpossibility\nof knowing all possible consequences.\u201d This use of the term hypothetical\nis also confusing in view of Kant\u2019s criticism of the hypothetical\nemployment of Reason in A 769 ff. = B 797 ff.\n\n[1654] A 647 = B 675.\n\n[1655] _Loc. cit._ and A 649 = B 677.\n\n[1656] A 648 = B 676.\n\n[1657] A 652 = B 680.\n\n[1658] A 660-1 = B 688-9.\n\n[1659] A 656 = B 684.\n\n[1660] A 656 = B 684.\n\n[1661] A 658 = B 686.\n\n[1662] A 660 = B 688.\n\n[1663] The opening paragraphs of the section, A 642-5 = B 670-3, may be\nof the same date as the concluding paragraphs.\n\n[1664] Cf. _per contra_ A 669-70 = B 697-8.\n\n[1665] A 666 = B 694.\n\n[1666] A 669 = B 697.\n\n[1667] Cf. above, pp. 446-7.\n\n[1668] Cf. A 681 = B 709.\n\n[1669] Cf. _per contra_ A 663-4 = B 691-2.\n\n[1670] A 670 = B 698.\n\n[1671] I may here guard against misunderstanding. Though the Ideas of\nReason condition the experience which they regulate, this must not be\ntaken as nullifying Kant\u2019s fundamental distinction between the\nregulative and the constitutive. Even when he is developing his less\nsceptical view, he adopts, in metaphysics as in ethics, a position which\nis radically distinct from that of Hegel. Though the moral ideal\nrepresents reality of the highest order, it transcends all possible\nrealisation of itself in human life. Though it conditions all our\nmorality, it at the same time condemns it. The Christian virtue of\nhumility defines the only attitude proper to the human soul. In an\nexactly similar manner, the fact that the Ideas of Reason have to be\nregarded as conditioning the possibility of sense-experience need not\nprevent us from also recognising that they likewise make possible our\nconsciousness of its limitations.\n\n[1672] Cf. above, pp. 473-7.\n\n[1673] A 679 = B 707.\n\n[1674] A 678 = B 706.\n\n[1675] A 674 = B 702. Cf. A 678-9 = B 706-7.\n\n[1676] A 680 = B 708.\n\n[1677] As above noted (pp. 499 ff.), when we find Kant thus insisting\nupon the completely soluble character of all problems of pure Reason,\nthe sceptical, subjectivist tendency is dominant.\n\n[1678] A 669 = B 697.\n\n[1679] Cf. above, pp. 536-7, 541-2.\n\n[1680] A 686-7 = B 714-15.\n\n[1681] A 693 = B 721.\n\n[1682] A 699-700 = B 727-8.\n\n[1683] A 701 = B 729.\n\n[1684] Nearly all the important points raised in the _Methodology_, and\nseveral of its chief sections, I have commented upon in their connection\nwith the earlier parts of the _Critique_. Also, the _Methodology_ is\nextremely diffuse. For these reasons I have found it advisable to give\nsuch additional comment as seems necessary in the form of this Appendix.\n\n[1685] On Kant\u2019s use of the terms \u2018discipline\u2019 and \u2018canon,\u2019 cf. above,\npp. 71-2, 170, 174, 438.\n\n[1686] Cf. above, p. 438.\n\n[1687] A 4-5 = B 8-9.\n\n[1688] _Untersuchung: Zweite Betrachtung, W._ ii. p. 283.\n\n[1689] Kant here disavows the position of the _Untersuchung_ in which\n(_Erste Betrachtung_, \u00a7 4) he had asserted that mathematics deals with\nquantity and philosophy with qualities.\n\n[1690] For comment upon this distinction, cf. above, pp. 131-3, 338-9.\n\n[1691] _Untersuchung: Erste Betrachtung_, \u00a7 2.\n\n[1692] A 728 = B 756.\n\n[1693] _Untersuchung: Zweite Betrachtung, W._ ii. p. 283.\n\n[1694] _Untersuchung: Erste Betrachtung_, \u00a7 1, _W._ ii. p. 276:\n\u201cMathematics proceeds to all its definitions by a _synthetic_ procedure,\nphilosophy by an analytic procedure.\u201d\n\n[1695] In the _Untersuchung_ Kant\u2019s statements are more cautious, and\nalso more adequate. Cf. _Erste Betrachtung_, \u00a7 3, _W._ ii. p. 279: \u201cIn\nmathematics there are only a few but in philosophy there are innumerable\nirresolvable concepts....\u201d\n\n[1696] A 731 _n._ = B 759 _n._\n\n[1697] The phrases which Kant employs (A 732-3 = B 760-1) are:\n\u201c_unmittelbargewiss_,\u201d \u201c_evident_,\u201d \u201c_augenscheinlich_.\u201d Cf. above, pp.\nxxxv-vi, 36 ff., 53.\n\n[1698] Cf. above, pp. 118, 142, 185-6.\n\n[1699] Cf. above, p. 117 ff.\n\n[1700] Cf. above, pp. 38-42, 93-4, 118-20, 133.\n\n[1701] Cf. above, pp. 111-12, 114-15.\n\n[1702] Cf. above, p. 131 ff.\n\n[1703] A 737 = B 765.\n\n[1704] Cf. above, pp. 36 ff., 117 ff., 128 ff., 565-6.\n\n[1705] A 743-4 = B 771-2.\n\n[1706] A 753 = B 781. In A 745 = B 773 Kant\u2019s mention of Hume can hardly\nrefer to Hume\u2019s _Dialogues_ (cf. above, pp. 539-40 n.). Kant probably\nhas in mind Section XI. of the _Enquiry_. The important discussion of\nHume\u2019s position in A 760 ff. = B 788 ff. has been commented upon above,\np. 61 ff. With Priestley\u2019s teaching (A 745-6 = B 773-4) Kant probably\nbecame acquainted through some indirect source. The first of Priestley\u2019s\nphilosophical writings to appear in German was his _History of the\nCorruptions of Christianity_. The translation was published in 1782. In\nA 747-8 = B 775-6 Kant quite obviously has Rousseau in mind.\n\n[1707] Section III., on _The Discipline of Pure Reason in Regard to\nHypotheses_, has been commented on above, pp. 543-6.\n\n[1708] A 782 = B 810.\n\n[1709] Even in mathematics the indirect method is not always available.\nCf. Russell, _Principles of Mathematics_, i. p. 15.\n\n[1710] A 794 = B 822.\n\n[1711] Cf. above, p. 563 _n._ 2.\n\n[1712] A 797 = B 825.\n\n[1713] Cf. _Critique of Judgment, W._ v. p. 473; Bernard\u2019s trans. p.\n411: \u201c_God_, _freedom_, and _immortality_ are the problems at the\nsolution of which all the preparations of Metaphysics aim, as their\nultimate and unique purpose.\u201d\n\n[1714] A 800-1 = B 829.\n\n[1715] The statement in A 801 = B 829 that morals is a subject foreign\nto transcendental philosophy is in line with that of A 14-15 = B 28, and\nconflicts with the position later adopted in the _Critique of Practical\nReason_. Cf. above, p. 77.\n\n[1716] A 803 = B 831-2.\n\n[1717] Cf. below, pp. 571-5.\n\n[1718] A 804 = B 832.\n\n[1719] Cf. above, p. lvi.\n\n[1720] These statements are subject to modification, if the distinction\n(not clearly recognised by Kant, but really essential to his position)\nbetween immanent and transcendent metaphysics is insisted upon. Cf.\nabove, pp. liv-v, 22, 56, 66-70.\n\n[1721] Cf. above, p. 541.\n\n[1722] _W._ v. pp. 47-8; Abbott\u2019s trans. (3rd edition) p. 136.\n\n[1723] Cf. _Critique of Practical Reason, W._ v. pp. 31-7; Abbott\u2019s\ntrans. p. 120.\n\n[1724] Cf. _Critique of Practical Reason, W._ v. p. 43; Abbott\u2019s trans.\np. 132: \u201cThe moral law, although it gives no _view_, yet gives us a fact\nabsolutely inexplicable from any data of the sensible world, or from the\nwhole compass of our theoretical use of reason, a fact which points to a\npure world of the understanding, nay, even defines it _positively_, and\nenables us to know something of it, namely, a law.\u201d\n\n[1725] Cf. _Critique of Practical Reason_, in note to Preface.\n\n[1726] _Op. cit._, Preface, at the beginning, Abbott\u2019s trans. pp. 87-8.\nCf. also the concluding pages of Book I., _W._ v. pp. 103-6, Abbott, pp.\n197-200.\n\n[1727] _Critique of Judgment, W._ v. p. 468; Bernard\u2019s trans. p. 406.\n\n[1728] _Op. cit._ p. 474; Bernard\u2019s trans. p. 413.\n\n[1729] A 815 = B 843.\n\n[1730] Cf. _Critique of Practical Reason, W._ v. pp. 143-4 _n._;\nAbbott\u2019s trans. p. 242: \u201cIt is a duty to realise the _Summum Bonum_ to\nthe utmost of our power, therefore it must be possible, consequently it\nis unavoidable for every rational being in the world to assume what is\nnecessary for its objective possibility. The assumption is as necessary\nas the moral law, in connexion with which alone it is valid.\u201d\n\n[1731] Cf. _Critique of Practical Reason, W._ v. p. 142 ff.; Abbott\u2019s\ntrans. p. 240 ff.; _Critique of Judgment, W._ v. pp. 469-70; Bernard\u2019s\ntrans. pp. 406-8.\n\n[1732] _Critique of Judgment, W._ v. pp. 369-72; Bernard\u2019s trans. pp.\n407-10. Cf. note in same section: \u201cIt is a trust in the promise of the\nmoral law; not, however, such as is contained in it, but such as I put\ninto it, and that on morally adequate grounds.\u201d\n\n[1733] A 819 = B 847.\n\n[1734] A 820 = B 848.\n\n[1735] The distinction is less harshly drawn in Kant\u2019s _Logic,\nEinleitung_, ix. (Hartenstein), viii. p. 73; Eng. trans, p. 63:\n\u201cConviction is opposed to persuasion. Persuasion is an assent from\ninadequate reasons, in respect to which we do not know whether they are\nonly subjective or are also objective. Persuasion often precedes\nconviction.\u201d\n\n[1736] Cf. above, pp. 10, 543. Cf. _Fortschritte; Werke_ (Hartenstein),\nviii. p. 561.\n\n[1737] Cf. _Logic, loc. cit._ Cf. _Foundations of the Metaphysics of\nMorals, W._ iv. pp. 416-17: Abbott\u2019s trans. pp. 33-34.\n\n[1738] Regarding Kant\u2019s distinction in A 827 = B 855 between Ideas and\nhypotheses cf. above, p. 543 ff. Cf. also _Critique of Judgment, W._ v.\npp. 392 ff., 461 ff.; Bernard\u2019s trans. pp. 302 ff., 395 ff.\n\n[1739] A 829 = B 857.\n\n[1740] Cf. Kant\u2019s Preface to the _Critique of Practical Reason, W._ v.\np. 8 _n._; Abbott\u2019s trans. p. 93 n. \u201cA reviewer who wanted to find some\nfault with this work--[the _Foundations of the Metaphysics of\nMorals_]--has hit the truth better, perhaps, than he thought, when he\nsays that no new principle of morality is set forth in it, but only a\n_new formula_. But who would think of introducing a new principle of all\nmorality, and making himself as it were the first inventor of it, just\nas if all the world before him were ignorant what duty was, or had been\nin thorough-going error? But whoever knows of what importance to a\nmathematician a formula is, which defines accurately what is to be done\nto work out a problem, will not think that a formula is insignificant\nand useless which does the same for all duty in general.\u201d Cf.\n_Fortschritte_, _Werke_ (Hartenstein), viii. p. 563.\n\n[1741] _Fragmente aus dem Nachlasse, Werke_ (Hartenstein), viii. p. 624,\nalready quoted above, p. lvii. Cf. also _op. cit._ p. 630.\n\n[1742] Cf. _Critique of Practical Reason_, Conclusion, _W._ v. pp.\n161-2; Abbott\u2019s trans. p. 260.\n\n[1743] A 832 = B 860.\n\n[1744] _K._ p. 633 _n._ Cf. above, p. xxii.\n\n[1745] Cf. Adickes, _K._ p. 635 _n._, and Vaihinger, i. p. 306. In this\ntable _Critique_ is distinguished from the _System_ of pure Reason (cf.\nabove, pp. 71-2). The transcendental philosophy of pure Reason of this\ntable corresponds to the _Analytic_ of the _Critique_, and to \u201cpure\nnatural science\u201d in the absolute sense (cf. above, pp. 66-7). The\nrational physics of this table corresponds to the _Metaphysical First\nPrinciples of Natural Science_.\n\n[1746] When Kant in A 840 = B 868 takes philosophy as including\nempirical knowledge he contradicts the spirit, though not the letter of\nhis own preceding statements. In his _Introduction to Logic_\n(Hartenstein, viii. p. 22, Abbott\u2019s trans. p. 12) the empirical is\nidentified with the historical.\n\n[1747] _Fortschritte, Werke_ (Hartenstein), viii. p. 554.\n\n[1748] _Op. cit_. p. 520.\n\n[1749] _I.e._ between the conception of philosophy as _Schulbegriff_ and\nas _Weltbegriff_ (_conceptus cosmicus_). He explains in a note to A 839\n= B 868 that he employs these latter terms as indicating that philosophy\nin the traditional or humanistic sense is concerned with \u201cthat which\nmust necessarily interest every one.\u201d I have translated _Weltbegriff_ as\n\u2018_universal concept_.\u2019 By _conceptus cosmicus_ Kant means \u2018concept\nshared by the whole world,\u2019 or \u2018common to all mankind.\u2019\n\n[1750] Cf. Kant\u2019s _Logic, Introduction_, \u00a7 iii.: Abbott\u2019s trans. pp.\n14-15: \u201cIn this scholastic signification of the word, philosophy aims\nonly _at skill_; in reference to the higher concept common to all\nmankind, on the contrary, it aims at _utility_. In the former aspect,\ntherefore, it is a doctrine of skill; in the latter a doctrine of\nwisdom; it is the lawgiver of reason; and hence the philosopher is _not_\na _master of the art of reason_, but a _lawgiver_. The master of the art\nof reason, or as Socrates calls him, the _philodoxus_, strives merely\nfor speculative knowledge, without concerning himself how much this\nknowledge contributes to the ultimate end of human reason: he gives\nrules for the use of reason for all kinds of ends. The practical\nphilosopher, the teacher of wisdom by doctrine and example, is the true\nphilosopher. For philosophy is the Ideal of a perfect wisdom, which\nshows us the ultimate ends of all human reason.\u201d\n\n[1751] A 839 = B 867.\n\n[1752] A 851 = B 879.\n\n[1753] A 850 = B 878.\n\n[1754] A 848-9 = B 876-7. Cf. above, pp. 237, 311 _n._, 312 _n._, 384-5,\n473-7, 554.\n\n[1755] A 852 = B 880.\n\n[1756] Cf. A 313 ff. = B 370 ff., above, pp. 498-9.\n\n[1757] Cf. above, pp. xxviii-xxix.\n\n[1758] _Einleitung_, \u00a7 iv.: Abbott\u2019s trans, pp. 17-23.\n\n[1759] Supplementary to pp. xxv-xxxiii. Throughout I shall make use of\nmy _Studies in the Cartesian Philosophy_, and may refer the reader to\nthem for further justification of the positions adopted.\n\n[1760] For recognition of this distinction, cf. Herbert Spencer,\n_Principles of Psychology_, vol. i., 3rd ed., pp. 620-3.\n\n[1761] Cf. _Studies in the Cartesian Philosophy_, pp. 80-2, 106-7.\n\n[1762] This distinction is due to Galileo, though the terms \u201cprimary\u201d\nand \u201csecondary\u201d were first employed by Locke.\n\n[1763] I have dealt with Avenarius\u2019 criticism in \u201cAvenarius\u2019 Philosophy\nof Pure Experience\u201d (_Mind_, vol. xv. N.S., pp. 13-31, 149-160); with\nBergson\u2019s criticism in \u201cSubjectivism and Realism in Modern Philosophy\u201d\n(_Philosophical Review_, vol. xvii. pp. 138-148); and with the general\nissue as a whole in \u201cThe Problem of Knowledge\u201d (_Journal of Philosophy_,\nvol. ix. pp. 113-128).\n\n[1764] On Descartes\u2019 failure to distinguish between the mathematical and\nthe dynamical aspects of motion, cf. above, p. 584.\n\n[1765] _Essay concerning Human Understanding_, IV. vi. 16.\n\n[1766] _Op. cit._ IV. xii. 7.\n\n[1767] _Op. cit._ IV. vi. 11.\n\n[1768] Cf. above, pp. 27-8.\n\n[1769] Though the concept of substance is also discussed by Hume, his\ntreatment of it is quite perfunctory.\n\n[1770] Cf. above, pp. xxv ff., 61 ff.\n\n[1771] _Treatise on Human Nature_ (Green and Grose), i. p. 380.\n\n[1772] _Op. cit._ p. 383.\n\n[1773] _Loc. cit._\n\n[1774] For justification of the phrase \u201csynthetic reason,\u201d I must refer\nto my articles in _Mind_, vol. xiv. N.S. pp. 149-73, 335-47, on \u201cThe\nNaturalism of Hume.\u201d\n\n[1775] _Treatise_ (Green and Grose), i. pp. 474-5.\n\n[1776] _Enquiry concerning Human Understanding_ (Green and Grose), p.\n40.\n\n[1777] _Treatise_, p. 471.\n\n[1778] _Enquiry_ (Green and Grose), pp. 25-6.\n\n[1779] _\u00c9claircissement_ sur chap. iii. pt. ii. liv. vi. _de la\nRecherche_: tome iv. (1712) p. 381.\n\n[1780] _Enquiry_, p. 57.\n\n[1781] _Enquiry_, p. 32.\n\n[1782] This is the objection upon which Beattie chiefly insists.\n\n[1783] _Op. cit._ pp. 33-4.\n\n[1784] Cf. above, pp. 39 ff., 54, 222 ff., 241, 286-9.\n\n[1785] How far Hume\u2019s criticism of empiricism really influenced Kant in\nhis appreciation of this deeper problem, it seems impossible to decide.\nVery probably Kant proceeded to it by independent development of his own\nstandpoint, after the initial impulse received on the more strictly\nlogical issue.\n\n[1786] The assertion, by Kuno Fischer and Paulsen, of an empirical\nperiod in Kant\u2019s development, has been challenged by Adickes, B.\nErdmann, Riehl, and Vaihinger.\n\n[1787] Cf. B. Erdmann\u2019s _Kriticismus_, p. 147; _Critique of Judgment,\nW._ v. p. 391 (Bernard\u2019s trans, p. 301).\n\n[1788] Above, pp. xxx-iii.\n\n[1789] _Philosophischer Kriticismus_, 2nd ed. p. 209.\n\n[1790] Cf. above, pp. lv-vi, lxi, 543 ff.\n\n       *       *       *       *       *\n\n\nTypographical errors corrected by the etext transcriber:\n\nThe transcenedntal doctrine=> The transcendental doctrine {pg 77}\n\nnon-commital=> non-committal {pg 122}\n\nwidersinnisches=> Widersinnisches {pg 444}\n\nErkenntniss=> Erkenntnis {pg 449}\n\nthemelves=> themselves {pg 505}\n\nwhich contain the the material=> which contain the material {pg 523}\n\nit as valid=> it is valid {pg 575}","title":"A Commentary to Kant's 'Critique of Pure Reason'","tradition":"philosophy-ethics","translated_text":null,"word_count":282050}