{"collection":"philosophy-ethics","full_text_available":true,"graph_url":"https://unchartedknowledge.5gfusion.net/v2/api/page/gutenberg_32701/graph","html_url":"https://unchartedknowledge.5gfusion.net/v2/page/gutenberg_32701","id":"gutenberg_32701","license":"Public Domain (US)","rights_basis":"gutenberg_metadata_copyright_false","source":"gutenberg","source_url":"https://www.gutenberg.org/ebooks/32701","summary":null,"tags":["Category: Philosophy & Ethics","Kant, Immanuel, 1724-1804","Knowledge, Theory of"],"text":"E-text prepared by Meredith Bach, lizardcry, and the Project Gutenberg\nOnline Distributed Proofreading Team (http://www.pgdp.net) from page\nimages generously made available by Internet Archive/Canadian Libraries\n(http://www.archive.org/details/toronto)\n\n\n\nNote: Images of the original pages are available through\n      Internet Archive/Canadian Libraries. See\n      http://www.archive.org/details/kantknowledge00pricuoft\n\n\nTranscriber's note:\n\n   1. Text in italics is enclosed by underscores (_italics_).\n\n   2. The original text includes Greek characters. For this text\n   version these letters have been replaced with transliterations\n   represented within square brackets [Greek: ]. Also greek\n   letters alpha and beta are represented as [alpha] and [beta]\n   in this text.\n\n   3. A subscript is indicated by an underscore followed by the\n   subscript in curly braces. For example, a_{2} indicates a with\n   subscript 2.\n\n   4. Footnotes have been moved to the end of the paragraph\n   wherein they have been referred to.\n\n   5. Other than the corrections listed above, printer's\n   inconsistencies in spelling, punctuation and hyphenation have\n   been retained.\n\n\n\n\n\nKANT'S THEORY OF KNOWLEDGE\n\nby\n\nH. A. PRICHARD\n\nFellow of Trinity College, Oxford\n\n\n\n\n\n\n\nOxford\nAt the Clarendon Press\n1909\n\nHenry Frowde, M.A.\nPublisher to the University of Oxford\nLondon, Edinburgh, New York\nToronto and Melbourne\n\n\n\n\nPREFACE\n\n\nThis book is an attempt to think out the nature and tenability of\nKant's Transcendental Idealism, an attempt animated by the conviction\nthat even the elucidation of Kant's meaning, apart from any criticism,\nis impossible without a discussion on their own merits of the main\nissues which he raises.\n\nMy obligations are many and great: to Caird's _Critical Philosophy of\nKant_ and to the translations of Meiklejohn, Max Mueller, and Professor\nMahaffy; to Mr. J. A. Smith, Fellow of Balliol College, and to Mr. H.\nW. B. Joseph, Fellow of New College, for what I have learned from them\nin discussion; to Mr. A. J. Jenkinson, Fellow of Brasenose College,\nfor reading and commenting on the first half of the MS.; to Mr. H. H.\nJoachim, Fellow of Merton College, for making many important\nsuggestions, especially with regard to matters of translation; to Mr.\nJoseph, for reading the whole of the proofs and for making many\nvaluable corrections; and, above all, to my wife for constant and\nunfailing help throughout, and to Professor Cook Wilson, to have been\nwhose pupil I count the greatest of philosophical good fortunes. Some\nyears ago it was my privilege to be a member of a class with which\nProfessor Cook Wilson read a portion of Kant's _Critique of Pure\nReason_, and subsequently I have had the advantage of discussing with\nhim several of the more important passages. I am especially indebted\nto him in my discussion of the following topics: the distinction\nbetween the Sensibility and the Understanding (pp. 27-31, 146-9,\n162-6), the term 'form of perception' (pp. 37, 40, 133 fin.-135), the\n_Metaphysical Exposition of Space_ (pp. 41-8), Inner Sense (Ch. V,\nand pp. 138-9), the _Metaphysical Deduction of the Categories_ (pp.\n149-53), Kant's account of 'the reference of representations to an\nobject' (pp. 178-86), an implication of perspective (p. 90), the\nimpossibility of a 'theory' of knowledge (p. 245), and the points\nconsidered, pp. 200 med.-202 med., 214 med.-215 med., and 218. The\nviews expressed in the pages referred to originated from Professor\nCook Wilson, though it must not be assumed that he would accept them\nin the form in which they are there stated.\n\n\n\n\nCONTENTS\n\n\n  CHAPTER I                                                     PAGE\n  THE PROBLEM OF THE _Critique_                                    1\n\n  CHAPTER II\n  THE SENSIBILITY AND THE UNDERSTANDING                           27\n\n  CHAPTER III\n  SPACE                                                           36\n\n  CHAPTER IV\n  PHENOMENA AND THINGS IN THEMSELVES                              71\n\n  NOTE\n  THE FIRST ANTINOMY                                             101\n\n  CHAPTER V\n  TIME AND INNER SENSE                                           103\n\n  CHAPTER VI\n  KNOWLEDGE AND REALITY                                          115\n\n  CHAPTER VII\n  THE METAPHYSICAL DEDUCTION OF THE CATEGORIES                   140\n\n  CHAPTER VIII\n  THE TRANSCENDENTAL DEDUCTION OF THE CATEGORIES                 161\n\n  CHAPTER IX\n  GENERAL CRITICISM OF THE TRANSCENDENTAL DEDUCTION OF THE\n  CATEGORIES                                                     214\n\n  CHAPTER X\n  THE SCHEMATISM OF THE CATEGORIES                               246\n\n  CHAPTER XI\n  THE MATHEMATICAL PRINCIPLES                                    260\n\n  CHAPTER XII\n  THE ANALOGIES OF EXPERIENCE                                    268\n\n  CHAPTER XIII\n  THE POSTULATES OF EMPIRICAL THOUGHT                            308\n\n  NOTE\n  THE REFUTATION OF IDEALISM                                     319\n\n\n\n\nREFERENCES\n\n\n  A     = First edition of the _Critique of Pure Reason_.\n  B     = Second edition of the _Critique of Pure Reason_.\n  Prol. = Kant's _Prolegomena to any future Metaphysic_.\n  M     = Meiklejohn's Translation of the _Critique of Pure Reason_.\n  Mah.  = Mahaffy. Translation of Kant's _Prolegomena to any future\n          Metaphysic_. (The pages referred to are those of the first\n          edition; these are also to be found in the text of the\n          second edition.)\n  Caird = Caird's _Critical Philosophy of Kant_.\n\n\n\n\nCHAPTER I\n\nTHE PROBLEM OF THE _CRITIQUE_\n\n\nThe problem of the _Critique_ may be stated in outline and\napproximately in Kant's own words as follows.\n\nHuman reason is called upon to consider certain questions, which it\ncannot decline, as they are presented by its own nature, but which it\ncannot answer. These questions relate to God, freedom of the will, and\nimmortality. And the name for the subject which has to deal with these\nquestions is metaphysics. At one time metaphysics was regarded as the\nqueen of all the sciences, and the importance of its aim justified the\ntitle. At first the subject, propounding as it did a dogmatic system,\nexercised a despotic sway. But its subsequent failure brought it into\ndisrepute. It has constantly been compelled to retrace its steps;\nthere has been fundamental disagreement among philosophers, and no\nphilosopher has successfully refuted his critics. Consequently the\ncurrent attitude to the subject is one of weariness and indifference.\nYet humanity cannot really be indifferent to such problems; even those\nwho profess indifference inevitably make metaphysical assertions; and\nthe current attitude is a sign not of levity but of a refusal to put\nup with the illusory knowledge offered by contemporary philosophy. Now\nthe objects of metaphysics, God, freedom, and immortality, are not\nobjects of experience in the sense in which a tree or a stone is an\nobject of experience. Hence our views about them cannot be due to\nexperience; they must somehow be apprehended by pure reason, i. e. by\nthinking and without appeal to experience. Moreover, it is in fact by\nthinking that men have always tried to solve the problems concerning\nGod, freedom, and immortality. What, then, is the cause of the\nunsatisfactory treatment of these problems and men's consequent\nindifference? It must, in some way, lie in a failure to attain the\nsure scientific method, and really consists in the neglect of an\ninquiry which should be a preliminary to all others in metaphysics.\nMen ought to have begun with a critical investigation of pure reason\nitself. Reason should have examined its own nature, to ascertain in\ngeneral the extent to which it is capable of attaining knowledge\nwithout the aid of experience. This examination will decide whether\nreason is able to deal with the problems of God, freedom, and\nimmortality at all; and without it no discussion of these problems\nwill have a solid foundation. It is this preliminary investigation\nwhich the _Critique of Pure Reason_ proposes to undertake. Its aim is\nto answer the question, 'How far can reason go, without the material\npresented and the aid furnished by experience?' and the result\nfurnishes the solution, or at least the key to the solution, of all\nmetaphysical problems.\n\nKant's problem, then, is similar to Locke's. Locke states[1] that his\npurpose is to inquire into the original, certainty, and extent of\nhuman knowledge; and he says, \"If, by this inquiry into the nature of\nthe understanding I can discover the powers thereof; how far they\nreach, to what things they are in any degree proportionate, and where\nthey fail us; I suppose it may be of use to prevail with the busy mind\nof man, to be more cautious in meddling with things exceeding its\ncomprehension; to stop when it is at the utmost extent of its tether;\nand to sit down in a quiet ignorance of those things, which, upon\nexamination, are found to be beyond the reach of our capacities.\"\nThus, to use Dr. Caird's analogy,[2] the task which both Locke and\nKant set themselves resembled that of investigating a telescope,\nbefore turning it upon the stars, to determine its competence for the\nwork.\n\n    [1] Locke's _Essay_, i, 1, Secs. 2, 4.\n\n    [2] Caird, i, 10.\n\nThe above outline of Kant's problem is of course only an outline. Its\ndefinite formulation is expressed in the well-known question, 'How are\n_a priori_ synthetic judgements possible?'[3] To determine the meaning\nof this question it is necessary to begin with some consideration of\nthe terms '_a priori_' and 'synthetic'.\n\n    [3] B. 19, M. 12.\n\nWhile there is no difficulty in determining what Kant would have\nrecognized as an _a priori_ judgement, there is difficulty in\ndetermining what he meant by calling such a judgement _a priori_. The\ngeneral account is given in the first two sections of the\nIntroduction. An _a priori_ judgement is introduced as something\nopposed to an _a posteriori_ judgement, or a judgement which has its\nsource in experience. Instances of the latter would be 'This body is\nheavy', and 'This body is hot'. The point of the word 'experience' is\nthat there is direct apprehension of some individual, e. g. an\nindividual body. To say that a judgement has its source in experience\nis of course to imply a distinction between the judgement and\nexperience, and the word 'source' may be taken to mean that the\njudgement depends for its validity upon the experience of the\nindividual thing to which the judgement relates. An _a priori_\njudgement, then, as first described, is simply a judgement which is\nnot _a posteriori_. It is independent of all experience; in other\nwords, its validity does not depend on the experience of individual\nthings. It might be illustrated by the judgement that all three-sided\nfigures must have three angles. So far, then, no positive meaning has\nbeen given to _a priori_.[4]\n\n    [4] Kant is careful to exclude from the class of _a priori_\n    judgements proper what may be called relatively _a priori_\n    judgements, viz. judgements which, though not independent of\n    all experience, are independent of experience of the facts to\n    which they relate. \"Thus one would say of a man who\n    undermined the foundations of his house that he might have\n    known _a priori_ that it would fall down, i. e. that he did\n    not need to wait for the experience of its actual falling\n    down. But still he could not know this wholly _a priori_, for\n    he had first to learn through experience that bodies are\n    heavy and consequently fall, if their supports are taken\n    away.\" (B. 2, M. 2.)\n\nKant then proceeds, not as we should expect, to state the positive\nmeaning of _a priori_; but to give tests for what is _a priori_. Since\na test implies a distinction between itself and what is tested, it is\nimplied that the meaning of _a priori_ is already known.[5]\n\n    [5] It may be noted that in this passage (Introduction, Secs.\n    1 and 2) Kant is inconsistent in his use of the term 'pure'.\n    Pure knowledge is introduced as a species of _a priori_\n    knowledge: \"_A priori_ knowledge, if nothing empirical is\n    mixed with it, is called pure\". (B. 3, M. 2, 17.) And in\n    accordance with this, the proposition 'every change has a\n    cause' is said to be _a priori_ but impure, because the\n    conception of change can only be derived from experience. Yet\n    immediately afterwards, pure, being opposed in general to\n    empirical, can only mean _a priori_. Again, in the phrase\n    'pure _a priori_' (B. 4 fin., M. 3 med.), the context shows\n    that 'pure' adds nothing to '_a priori_', and the proposition\n    'every change must have a cause' is expressly given as an\n    instance of pure _a priori_ knowledge. The inconsistency of\n    this treatment of the causal rule is explained by the fact\n    that in the former passage he is thinking of the conception\n    of change as empirical, while in the latter he is thinking of\n    the judgement as not empirical. At bottom in this passage\n    'pure' simply means _a priori_.\n\nThe tests given are necessity and strict universality.[6] Since\njudgements which are necessary and strictly universal cannot be based\non experience, their existence is said to indicate another source of\nknowledge. And Kant gives as illustrations, (1) any proposition in\nmathematics, and (2) the proposition 'Every change must have a cause'.\n\n    [6] In reality, these tests come to the same thing, for\n    necessity means the necessity of connexion between the\n    subject and predicate of a judgement, and since empirical\n    universality, to which strict universality is opposed, means\n    numerical universality, as illustrated by the proposition\n    'All bodies are heavy', the only meaning left for strict\n    universality is that of a universality reached not through an\n    enumeration of instances, but through the apprehension of a\n    necessity of connexion.\n\nSo far Kant has said nothing which determines the positive meaning of\n_a priori_. A clue is, however, to be found in two subsequent phrases.\nHe says that we may content ourselves with having established as a\nfact the pure use of our faculty of knowledge.[7] And he adds that not\nonly in judgements, but even in conceptions, is an _a priori_ origin\nmanifest.[8] The second statement seems to make the _a priori_\ncharacter of a judgement consist in its origin. As this origin cannot\nbe experience, it must, as the first statement implies, lie in our\nfaculty of knowledge. Kant's point is that the existence of universal\nand necessary judgements shows that we must possess a faculty of\nknowledge capable of yielding knowledge without appeal to experience.\nThe term _a priori_, then, has some reference to the existence of this\nfaculty; in other words, it gives expression to a doctrine of 'innate\nideas'. Perhaps, however, it is hardly fair to press the phrase\n'_test_ of _a priori_ judgements'. If so, it may be said that on the\nwhole, by _a priori_ judgements Kant really means judgements which are\nuniversal and necessary, and that he regards them as implying a\nfaculty which gives us knowledge without appeal to experience.\n\n    [7] B. 5, M. 4.\n\n    [8] Ibid.\n\nWe may now turn to the term 'synthetic judgement'. Kant distinguishes\nanalytic and synthetic judgements thus. In any judgement the predicate\nB either belongs to the subject A, as something contained (though\ncovertly) in the conception A, or lies completely outside the\nconception A, although it stands in relation to it. In the former case\nthe judgement is called analytic, in the latter synthetic.[9] 'All\nbodies are extended' is an analytic judgement; 'All bodies are heavy'\nis synthetic. It immediately follows that only synthetic judgements\nextend our knowledge; for in making an analytic judgement we are only\nclearing up our conception of the subject. This process yields no new\nknowledge, for it only gives us a clearer view of what we know\nalready. Further, all judgements based on experience are synthetic,\nfor it would be absurd to base an analytical judgement on experience,\nwhen to make the judgement we need not go beyond our own conceptions.\nOn the other hand, _a priori_ judgements are sometimes analytic and\nsometimes synthetic. For, besides analytical judgements, all\njudgements in mathematics and certain judgements which underlie\nphysics are asserted independently of experience, and they are\nsynthetic.\n\n    [9] B. 10, M. 7.\n\nHere Kant is obviously right in vindicating the synthetic character of\nmathematical judgements. In the arithmetical judgement 7 + 5 = 12, the\nthought of certain units as a group of twelve is no mere repetition of\nthe thought of them as a group of five added to a group of seven.\nThough the same units are referred to, they are regarded differently.\nThus the thought of them as twelve means either that we think of them\nas formed by adding one unit to a group of eleven, or that we think of\nthem as formed by adding two units to a group of ten, and so on. And\nthe assertion is that the same units, which can be grouped in one way,\ncan also be grouped in another. Similarly, Kant is right in pointing\nout that the geometrical judgement, 'A straight line between two\npoints is the shortest,' is synthetic, on the ground that the\nconception of straightness is purely qualitative,[10] while the\nconception of shortest distance implies the thought of quantity.\n\n    [10] Straightness means identity of direction.\n\nIt should now be an easy matter to understand the problem expressed\nby the question, 'How are _a priori_ synthetic judgements possible?'\nIts substance may be stated thus. The existence of _a posteriori_\nsynthetic judgements presents no difficulty. For experience is\nequivalent to perception, and, as we suppose, in perception we are\nconfronted with reality, and apprehend it as it is. If I am asked,\n'How do I know that my pen is black or my chair hard?' I answer that\nit is because I see or feel it to be so. In such cases, then, when my\nassertion is challenged, I appeal to my experience or perception of\nthe reality to which the assertion relates. My appeal raises no\ndifficulty because it conforms to the universal belief that if\njudgements are to rank as knowledge, they must be made to conform to\nthe nature of things, and that the conformity is established by appeal\nto actual experience of the things. But do _a priori_ synthetic\njudgements satisfy this condition? Apparently not. For when I assert\nthat every straight line is the shortest way between its extremities,\nI have not had, and never can have, experience of all possible\nstraight lines. How then can I be sure that all cases will conform to\nmy judgement? In fact, how can I anticipate my experience at all? How\ncan I make an assertion about any individual until I have had actual\nexperience of it? In an _a priori_ synthetic judgement the mind in\nsome way, in virtue of its own powers and independently of experience,\nmakes an assertion to which it claims that reality must conform. Yet\nwhy should reality conform? _A priori_ judgements of the other\nkind, viz. analytic judgements, offer no difficulty, since they\nare at bottom tautologies, and consequently denial of them is\nself-contradictory and meaningless. But there is difficulty where a\njudgement asserts that a term B is connected with another term A, B\nbeing neither identical with nor a part of A. In this case there is no\ncontradiction in asserting that A is not B, and it would seem that\nonly experience can determine whether all A is or is not B. Otherwise\nwe are presupposing that things must conform to our ideas about them.\nNow metaphysics claims to make _a priori_ synthetic judgements, for it\ndoes not base its results on any appeal to experience. Hence, before\nwe enter upon metaphysics, we really ought to investigate our right to\nmake _a priori_ synthetic judgements at all. Therein, in fact, lies\nthe importance to metaphysics of the existence of such judgements in\nmathematics and physics. For it shows that the difficulty is not\npeculiar to metaphysics, but is a general one shared by other\nsubjects; and the existence of such judgements in mathematics is\nspecially important because there their validity or certainty has\nnever been questioned.[11] The success of mathematics shows that at any\nrate under certain conditions _a priori_ synthetic judgements are\nvalid, and if we can determine these conditions, we shall be able to\ndecide whether such judgements are possible in metaphysics. In this\nway we shall be able to settle a disputed case of their validity by\nexamination of an undisputed case. The general problem, however, is\nsimply to show what it is which makes _a priori_ synthetic judgements\nas such possible; and there will be three cases, those of mathematics,\nof physics, and of metaphysics.\n\n    [11] Kant points out that this certainty has usually been\n    attributed to the analytic character of mathematical\n    judgements, and it is of course vital to his argument that\n    he should be successful in showing that they are really\n    synthetic.\n\nThe outline of the solution of this problem is contained in the\nPreface to the Second Edition. There Kant urges that the key is to be\nfound by consideration of mathematics and physics. If the question be\nraised as to what it is that has enabled these subjects to advance, in\nboth cases the answer will be found to lie in a change of method.\n\"Since the earliest times to which the history of human reason\nreaches, mathematics has, among that wonderful nation the Greeks,\nfollowed the safe road of a science. Still it is not to be supposed\nthat it was as easy for this science to strike into, or rather to\nconstruct for itself, that royal road, as it was for logic, in which\nreason has only to do with itself. On the contrary, I believe that it\nmust have remained long in the stage of groping (chiefly among the\nEgyptians), and that this change is to be ascribed to a _revolution_,\ndue to the happy thought of one man, through whose experiment the path\nto be followed was rendered unmistakable for future generations, and\nthe certain way of a science was entered upon and sketched out once\nfor all.... A new light shone upon the first man (Thales, or whatever\nmay have been his name) who demonstrated the properties of the\nisosceles triangle; for he found that he ought not to investigate\nthat which he saw in the figure or even the mere conception of the\nsame, and learn its properties from this, but that he ought to produce\nthe figure by virtue of that which he himself had thought into it _a\npriori_ in accordance with conceptions and had represented (by means\nof a construction), and that in order to know something with certainty\n_a priori_ he must not attribute to the figure any property other than\nthat which necessarily follows from that which he has himself\nintroduced into the figure, in accordance with his conception.\"[12]\n\n    [12] B. x-xii, M. xxvi.\n\nHere Kant's point is as follows. Geometry remained barren so long as\nmen confined themselves either to the empirical study of individual\nfigures, of which the properties were to be discovered by observation,\nor to the consideration of the mere conception of various kinds of\nfigure, e. g. of an isosceles triangle. In order to advance, men had\nin some sense to produce the figure through their own activity, and in\nthe act of constructing it to recognize that certain features were\nnecessitated by those features which they had given to the figure in\nconstructing it. Thus men had to make a triangle by drawing three\nstraight lines so as to enclose a space, and then to recognize that\nthree angles must have been made by the same process. In this way the\nmind discovered a general rule, which must apply to all cases, because\nthe mind itself had determined the nature of the cases. A property B\nfollows from a nature A; all instances of A must possess the property\nB, because they have solely that nature A which the mind has given\nthem and whatever is involved in A. The mind's own rule holds good in\nall cases, because the mind has itself determined the nature of the\ncases.\n\nKant's statements about physics, though not the same, are analogous.\nExperiment, he holds, is only fruitful when reason does not follow\nnature in a passive spirit, but compels nature to answer its own\nquestions. Thus, when Torricelli made an experiment to ascertain\nwhether a certain column of air would sustain a given weight, he had\npreviously calculated that the quantity of air was just sufficient to\nbalance the weight, and the significance of the experiment lay in his\nexpectation that nature would conform to his calculations and in the\nvindication of this expectation. Reason, Kant says, must approach\nnature not as a pupil but as a judge, and this attitude forms the\ncondition of progress in physics.\n\nThe examples of mathematics and physics suggest, according to Kant,\nthat metaphysics may require a similar revolution of standpoint, the\nlack of which will account for its past failure. An attempt should\ntherefore be made to introduce such a change into metaphysics. The\nchange is this. Hitherto it has been assumed that our knowledge must\nconform to objects. This assumption is the real cause of the failure\nto extend our knowledge _a priori_, for it limits thought to the\nanalysis of conceptions, which can only yield tautological judgements.\nLet us therefore try the effect of assuming that objects must conform\nto our knowledge. Herein lies the Copernican revolution. We find that\nthis reversal of the ordinary view of the relation of objects to the\nmind enables us for the first time to understand the possibility of _a\npriori_ synthetic judgements, and even to demonstrate certain laws\nwhich lie at the basis of nature, e. g. the law of causality. It is\ntrue that the reversal also involves the surprising consequence that\nour faculty of knowledge is incapable of dealing with the objects of\nmetaphysics proper, viz. God, freedom, and immortality, for the\nassumption limits our knowledge to objects of possible experience. But\nthis very consequence, viz. the impossibility of metaphysics, serves\nto test and vindicate the assumption. For the view that our knowledge\nconforms to objects as things in themselves leads us into an insoluble\ncontradiction when we go on, as we must, to seek for the\nunconditioned; while the assumption that objects must, as phenomena,\nconform to our way of representing them, removes the contradiction[13].\nFurther, though the assumption leads to the denial of speculative\nknowledge in the sphere of metaphysics, it is still possible that\nreason in its practical aspect may step in to fill the gap. And the\nnegative result of the assumption may even have a positive value. For\nif, as is the case, the moral reason, or reason in its practical\naspect, involves certain postulates concerning God, freedom, and\nimmortality, which are rejected by the speculative reason, it is\nimportant to be able to show that these objects fall beyond the scope\nof the speculative reason. And if we call reliance on these\npostulates, as being presuppositions of morality, faith, we may say\nthat knowledge must be abolished to make room for faith.\n\n    [13] Cf. pp. 101-2.\n\nThis answer to the main problem, given in outline in the Preface, is\nundeniably plausible. Yet examination of it suggests two criticisms\nwhich affect Kant's general position.\n\nIn the first place, the parallel of mathematics which suggests the\n'Copernican' revolution does not really lead to the results which Kant\nsupposes. Advance in mathematics is due to the adoption not of any\nconscious assumption but of a certain procedure, viz. that by which we\ndraw a figure and thereby see the necessity of certain relations\nwithin it. To preserve the parallel, the revolution in metaphysics\nshould have consisted in the adoption of a similar procedure, and\nadvance should have been made dependent on the application of an at\nleast quasi-mathematical method to the objects of metaphysics.\nMoreover, since these objects are God, freedom, and immortality, the\nconclusion should have been that we ought to study God, freedom, and\nimmortality by somehow constructing them in perception and thereby\ngaining insight into the necessity of certain relations. Success or\nfailure in metaphysics would therefore consist simply in success or\nfailure to see the necessity of the relations involved. Kant, however,\nmakes the condition of advance in metaphysics consist in the adoption\nnot of a method of procedure but of an assumption, viz. that objects\nconform to the mind. And it is impossible to see how this assumption\ncan assist what, on Kant's theory, it ought to have assisted, viz. the\nstudy of God, freedom, and immortality, or indeed the study of\nanything. In geometry we presuppose that individual objects conform to\nthe universal rules of relation which we discover. Now suppose we\ndescribe a geometrical judgement, e. g. that two straight lines cannot\nenclose a space, as a mental law, because we are bound to think it\ntrue. Then we may state the presupposition by saying that objects,\ne. g. individual pairs of straight lines, must conform to such a mental\nlaw. But the explicit recognition of this presupposition and the\nconscious assertion of it in no way assist the solution of particular\ngeometrical problems. The presupposition is really a condition of\ngeometrical thinking at all. Without it there is no geometrical\nthinking, and the recognition of it places us in no better position\nfor the study of geometrical problems. Similarly, if we wish to think\nout the nature of God, freedom, and immortality, we are not assisted\nby assuming that these objects must conform to the laws of our\nthinking. We must presuppose this conformity if we are to think at\nall, and consciousness of the presupposition puts us in no better\nposition. What is needed is an insight similar to that which we have\nin geometry, i. e. an insight into the necessity of the relations\nunder consideration such as would enable us to see, for example, that\nbeing a man, as such, involves living for ever.\n\nKant has been led into the mistake by a momentary change in the\nmeaning given to 'metaphysics'. For the moment he is thinking of\nmetaphysics, not as the inquiry concerned with God, freedom, and\nimmortality, but as the inquiry which has to deal with the problem as\nto how we can know _a priori_. This problem is assisted, at any rate\nprima facie, by the assumption that things must conform to the mind.\nAnd this assumption can be said to be suggested by mathematics,\ninasmuch as the mathematician presupposes that particular objects must\ncorrespond to the general rules discovered by the mind. From this\npoint of view Kant's only mistake, if the parallelism is to be\nmaintained, is that he takes for an assumption which enables the\nmathematician to advance a metaphysical presupposition of the advance,\non which the mathematician never reflects, and awareness of which\nwould in no way assist his mathematics.\n\nIn the second place the 'Copernican' revolution is not strictly the\nrevolution which Kant supposes it to be. He speaks as though his aim\nis precisely to reverse the ordinary view of the relation of the mind\nto objects. Instead of the mind being conceived as having to conform\nto objects, objects are to be conceived as having to conform to the\nmind. But if we consider Kant's real position, we see that these views\nare only verbally contrary, since the word object refers to something\ndifferent in each case. On the ordinary view objects are something\noutside the mind, in the sense of independent of it, and the ideas,\nwhich must conform to objects, are something within the mind, in the\nsense of dependent upon it. The conformity then is of something within\nthe mind to something outside it. Again, the conformity means that one\nof the terms, viz. the object, exists first and that then the other\nterm, the idea, is fitted to or made to correspond to it. Hence the\nreal contrary of this view is that ideas, within the mind, exist first\nand that objects outside the mind, coming into existence afterwards,\nmust adapt themselves to the ideas. This of course strikes us as\nabsurd, because we always think of the existence of the object as the\npresupposition of the existence of the knowledge of it; we do not\nthink the existence of the knowledge as the presupposition of the\nexistence of the object. Hence Kant only succeeds in stating the\ncontrary of the ordinary view with any plausibility, because in doing\nso he makes the term object refer to something which like 'knowledge'\nis within the mind. His position is that objects within the mind must\nconform to our general ways of knowing. For Kant, therefore, the\nconformity is not between something within and something without the\nmind, but between two realities within the mind, viz. the individual\nobject, as object of perception, i. e. a phenomenon, and our general\nways of perceiving and thinking. But this view is only verbally the\ncontrary of the ordinary view, and consequently Kant does not succeed\nin reversing the ordinary view that we know objects independent of or\noutside the mind, by bringing our ideas into conformity with them. In\nfact, his conclusion is that we do not know this object, i. e. the\nthing in itself, at all. Hence his real position should be stated by\nsaying not that the ordinary view puts the conformity between mind and\nthings in the wrong way, but that we ought not to speak of conformity\nat all. For the thing in itself being unknowable, our ideas can never\nbe made to conform to it. Kant then only reaches a conclusion which is\napparently the reverse of the ordinary view by substituting another\nobject for the thing in itself, viz. the phenomenon or appearance of\nthe thing in itself to us.\n\nFurther, this second line of criticism, if followed out, will be found\nto affect his statement of the problem as well as that of its\nsolution. It will be seen that the problem is mis-stated, and that the\nsolution offered presupposes it to be mis-stated. His statement of the\nproblem takes the form of raising a difficulty which the existence of\n_a priori_ knowledge presents to the ordinary view, according to which\nobjects are independent of the mind, and ideas must be brought into\nconformity with them. In a synthetic _a priori_ judgement we claim to\ndiscover the nature of certain objects by an act of our thinking, and\nindependently of actual experience of them. Hence if a supporter of\nthe ordinary view is asked to justify the conformity of this judgement\nor idea with the objects to which it relates, he can give no answer.\nThe judgement having _ex hypothesi_ been made without reference to the\nobjects, the belief that the objects must conform to it is the merely\narbitrary supposition that a reality independent of the mind must\nconform to the mind's ideas. But Kant, in thus confining the\ndifficulty to _a priori_ judgements, implies that empirical judgements\npresent no difficulty to the ordinary view; since they rest upon\nactual experience of the objects concerned, they are conformed to the\nobjects by the very process through which they arise. He thereby fails\nto notice that empirical judgements present a precisely parallel\ndifficulty. It can only be supposed that the conformity of empirical\njudgements to their objects is guaranteed by the experience upon which\nthey rest, if it be assumed that in experience we apprehend objects as\nthey are. But our experience or perception of individual objects is\njust as much mental as the thinking which originates _a priori_\njudgements. If we can question the truth of our thinking, we can\nlikewise question the truth of our perception. If we can ask whether\nour ideas must correspond to their objects, we can likewise ask\nwhether our perceptions must correspond to them. The problem relates\nsolely to the correspondence between something within the mind and\nsomething outside it; it applies equally to perceiving and thinking,\nand concerns all judgements alike, empirical as well as _a priori_.\nKant, therefore, has no right to imply that empirical judgements raise\nno problem, if he finds difficulty in _a priori_ judgements. He is\nonly able to draw a distinction between them, because, without being\naware that he is doing so, he takes account of the relation of the\nobject to the subject in the case of an _a priori_ judgement, while in\nthe case of an empirical judgement he ignores it. In other words, in\ndealing with the general connexion between the qualities of an\nobject, he takes into account the fact that we are thinking it, but,\nin dealing with the perception of the coexistence of particular\nqualities of an object, he ignores the fact that we are perceiving it.\nFurther, that the real problem concerns all synthetic judgements alike\nis shown by the solution which he eventually reaches. His conclusion\nturns out to be that while both empirical and _a priori_ judgements\nare valid of phenomena, they are not valid of things in themselves;\ni. e. that of things in themselves we know nothing at all, not even\ntheir particular qualities. Since, then, his conclusion is that even\nempirical judgements are not valid of things in themselves, it shows\nthat the problem cannot be confined to _a priori_ judgements, and\ntherefore constitutes an implicit criticism of his statement of the\nproblem.\n\nMust there not, however, be some problem peculiar to _a priori_\njudgements? Otherwise why should Kant have been led to suppose that\nhis problem concerned them only? Further consideration will show that\nthere is such a problem, and that it was only owing to the mistake\nindicated that Kant treated this problem as identical with that of\nwhich he actually offered a solution. In the universal judgements of\nmathematics we apprehend, as we think, general rules of connexion\nwhich must apply to all possible cases. Such judgements, then,\npresuppose a conformity between the connexions which we discover and\nall possible instances. Now Kant's treatment of this conformity as a\nconformity between our ideas and things has two implications. In the\nfirst place, it implies, as has been pointed out, that relation to the\nsubject, as thinking, is taken into account in the case of the\nuniversal connexion, and that relation to the subject, as perceiving,\nis ignored in the case of the individual thing. In the second place,\nit implies that what is related to the subject as the object of its\nthought must be subjective or mental; that because we have to think\nthe general connexion, the connexion is only our own idea, the\nconformity of things to which may be questioned. But the treatment, to\nbe consistent, should take account of relation to the subject in both\ncases or in neither. If the former alternative be accepted, then the\nsubjective character attributed by Kant in virtue of this relation to\nwhat is object of thought, and equally attributable to what is object\nof perception, reduces the problem to that of the conformity in\ngeneral of all ideas, including perceptions, within the mind to things\noutside it; and this problem does not relate specially to _a priori_\njudgements. To discover the problem which relates specially to them,\nthe other alternative must be accepted, that of ignoring relation to\nthe subject in both cases. The problem then becomes 'What renders\npossible or is presupposed by the conformity of individual things to\ncertain laws of connexion?' And, inasmuch as to deny the conformity is\nreally to deny that there are laws of connexion,[14] the problem\nreduces itself to the question, 'What is the presupposition of the\nexistence of definite laws of connexion in the world?' And the only\nanswer possible is that reality is a system or a whole of connected\nparts, in other words, that nature is uniform. Thus it turns out that\nthe problem relates to the uniformity of nature, and that the\nquestion 'How are _a priori_ synthetic judgements possible?' has in\nreality nothing to do with the problem of the relation of reality to\nthe knowing subject, but is concerned solely with the nature of\nreality.\n\n    [14] To object that the laws in question, being laws which we\n    have thought, may not be the true laws, and that therefore\n    there may still be other laws to which reality conforms, is\n    of course to reintroduce relation to the thinking subject.\n\nFurther, it is important to see that the alternative of ignoring\nrelation to the subject is the right one, not only from the point of\nview of the problem peculiar to _a priori_ judgements, but also from\nthe point of view of the nature of knowledge in general. Perceiving\nand thinking alike presuppose that reality is immediately object of\nthe mind, and that the act of apprehension in no way affects or enters\ninto the nature of what we apprehend about reality. If, for instance,\nI assert on the strength of perception that this table is round, I\nimply that I see the table, and that the shape which I judge it to\nhave is not affected by the fact that I am perceiving it; for I mean\nthat the table really is round. If some one then convinces me that I\nhave made a mistake owing to an effect of foreshortening, and that the\ntable is really oval, I amend my assertion, not by saying that the\ntable is round but only to my apprehension, but by saying that it\nlooks round. Thereby I cease to predicate roundness of the table\naltogether; for I mean that while it still looks round, it is not\nreally so. The case of universal judgements is similar. The statement\nthat a straight line is the shortest distance between its extremities\nmeans that it really is so. The fact is presupposed to be in no way\naltered by our having apprehended it. Moreover, reality is here just\nas much implied to be directly object of the mind as it is in the case\nof the singular judgement. Making the judgement consists, as we say,\nin _seeing_ the connexion between the direction between two points\nand the shortest distance between them. The connexion of real\ncharacteristics is implied to be directly object of thought.[16] Thus\nboth perceiving and thinking presuppose that the reality to which they\nrelate is directly object of the mind, and that the character of it\nwhich we apprehend in the resulting judgement is not affected or\naltered by the fact that we have had to perceive or conceive the\nreality.[17]\n\n    [15] Cf. Bosanquet, _Logic_, vol. ii, p. 2.\n\n    [16] In saying that a universal judgement is an immediate\n    apprehension of fact, it is of course not meant that it can\n    be actualized by itself or, so to say, _in vacuo_. Its\n    actualization obviously presupposes the presentation of\n    individuals in perception or imagination. Perception or\n    imagination thus forms the necessary occasion of a universal\n    judgement, and in that sense mediates it. Moreover, the\n    universal judgement implies an act of abstraction by which\n    we specially attend to those universal characters of the\n    individuals perceived or imagined, which enter into the\n    judgement. But, though our apprehension of a universal\n    connexion thus implies a process, and is therefore mediated,\n    yet the connexion, when we apprehend it, is immediately our\n    object. There is nothing between it and us.\n\n    [17] For a fuller discussion of the subject see Chh. IV and\n    VI.\n\nKant in the formulation of his problem implicitly admits this\npresupposition in the case of perception. He implies that empirical\njudgements involve no difficulty, because they rest upon the\nperception or experience of the objects to which they relate. On the\nother hand, he does not admit the presupposition in the case of\nconception, for he implies that in _a priori_ judgements we are not\nconfronted with reality but are confined to our own ideas. Hence we\nought to ask why Kant is led to adopt an attitude in the latter case\nwhich he does not adopt in the former. The answer appears to be\ntwofold. In the first place, there is an inveterate tendency to think\nof universals, and therefore of the connexions between them, as being\nnot objective realities[18] but mere ideas. In other words, we tend to\nadopt the conceptualist attitude, which regards individuals as the\nonly reality, and universals as mental fictions. In consequence, we\nare apt to think that while in perception, which is of the individual,\nwe are confronted by reality, in universal judgements, in which we\napprehend connexions between universals, we have before us mere ideas.\nKant may fairly be supposed to have been unconsciously under the\ninfluence of this tendency. In the second place, we apprehend a\nuniversal connexion by the operation of thinking. Thinking is\nessentially an activity; and since activity in the ordinary sense in\nwhich we oppose action to knowledge originates something, we tend to\nthink of the activity of thinking as also originating something, viz.\nthat which is our object when we think. Hence, since we think of what\nis real as independent of us and therefore as something which we may\ndiscover but can in no sense make, we tend to think of the object of\nthought as only an idea. On the other hand, what is ordinarily called\nperception, though it involves the activity of thinking, also involves\nan element in respect of which we are passive. This is the fact\npointed to by Kant's phrase 'objects are _given_ in perception'. In\nvirtue of this passive element we are inclined to think that in\nperception we simply stand before the reality in a passive attitude.\nThe reality perceived is thought to be, so to say, there, existing\nindependently of us; relation to the subject is unnoticed because of\nour apparently wholly passive attitude. At times, and especially when\nhe is thinking of the understanding as a faculty of spontaneity, Kant\nseems to have been under the influence of this second tendency.\n\n    [18] i. e. as not having a place in the reality which, as we\n    think, exists independently of the mind.\n\nThe preceding summary of the problem of the _Critique_ represents the\naccount given in the two Prefaces and the Introduction. According to\nthis account, the problem arises from the unquestioned existence of _a\npriori_ knowledge in mathematics and physics and the problematic\nexistence of such knowledge in metaphysics, and Kant's aim is to\ndetermine the range within which _a priori_ knowledge is possible.\nThus the problem is introduced as relating to _a priori_ knowledge as\nsuch, no distinction being drawn between its character in different\ncases. Nevertheless the actual discussion of the problem in the body\nof the _Critique_ implies a fundamental distinction between the nature\nof _a priori_ knowledge in mathematics and its nature in physics, and\nin order that a complete view of the problem may be given, this\ndistinction must be stated.\n\nThe 'Copernican' revolution was brought about by consideration of the\nfacts of mathematics. Kant accepted as an absolute starting-point the\nexistence in mathematics of true universal and necessary judgements.\nHe then asked, 'What follows as to the nature of the objects known in\nmathematics from the fact that we really know them?' Further, in his\nanswer he accepted a distinction which he never examined or even\nquestioned, viz. the distinction between things in themselves and\nphenomena.[19] This distinction assumed, Kant inferred from the truth\nof mathematics that things in space and time are only phenomena.\nAccording to him mathematicians are able to make the true judgements\nthat they do make only because they deal with phenomena. Thus Kant in\nno way sought to _prove_ the truth of mathematics. On the contrary, he\nargued from the truth of mathematics to the nature of the world which\nwe thereby know. The phenomenal character of the world being thus\nestablished, he was able to reverse the argument and to regard the\nphenomenal character of the world as _explaining_ the validity of\nmathematical judgements. They are valid, because they relate to\nphenomena. And the consideration which led Kant to take mathematics\nas his starting-point seems to have been the self-evidence of\nmathematical judgements. As we directly apprehend their necessity,\nthey admit of no reasonable doubt.\n\n    [19] Cf. Ch. IV. This distinction should of course have been\n    examined by one whose aim it was to determine how far our\n    knowledge can reach.\n\n    [20] For the self-evidence of mathematics to Kant compare B.\n    120, M. 73 and B. 200, M. 121.\n\nOn the other hand, the general principles underlying physics, e. g.\nthat every change must have a cause, or that in all change the quantum\nof matter is constant, appeared to Kant in a different light. Though\ncertainly not based on experience, they did not seem to him\nself-evident.[21] Hence,[22] in the case of these principles, he\nsought to give what he did not seek to give in the case of\nmathematical judgements, viz. a proof of their truth.[23] The nerve of\nthe proof lies in the contention that these principles are involved\nnot merely in any general judgement in physics, e. g. 'All bodies are\nheavy,' but even in any singular judgement, e. g. 'This body is\nheavy,' and that the validity of singular judgements is universally\nconceded. Thus here the fact upon which he takes his stand is not the\nadmitted truth of the universal judgements under consideration, but\nthe admitted truth of any singular judgement in physics. His\ntreatment, then, of the universal judgements of mathematics and that\nof the principles underlying physics are distinguished by the fact\nthat, while he accepts the former as needing no proof, he seeks to\nprove the latter from the admitted validity of singular judgements in\nphysics. At the same time the acceptance of mathematical judgements\nand the proof of the _a priori_ principles of physics have for Kant a\ncommon presupposition which distinguishes mathematics and physics from\nmetaphysics. Like universal judgements in mathematics, singular\njudgements in physics, and therefore the principles which they\npresuppose, are true only if the objects to which they relate are\nphenomena. Both in mathematics and physics, therefore, it is a\ncondition of _a priori_ knowledge that it relates to phenomena and not\nto things in themselves. But, just for this reason, metaphysics is in\na different position; since God, freedom, and immortality can never be\nobjects of experience, _a priori_ knowledge in metaphysics, and\ntherefore metaphysics itself, is impossible. Thus for Kant the very\ncondition, the realization of which justifies the acceptance of\nmathematical judgements and enables us to prove the principles of\nphysics, involves the impossibility of metaphysics.\n\n    [21] This is stated B. 200, M. 121. It is also implied B.\n    122, M. 75, B. 263-4, M. 160, and by the argument of the\n    _Analytic_ generally.\n\n    [22] This appears to be the real cause of the difference of\n    treatment, though it is not the reason assigned by Kant\n    himself, cf. B. 120, M. 73-4.\n\n    [23] His remarks about pure natural science in B. 20, M. 13\n    and Prol. Sec. 4 sub fin., do not represent the normal attitude\n    of the _Critique_.\n\nFurther, the distinction drawn between _a priori_ judgements in\nmathematics and in physics is largely responsible for the difficulty\nof understanding what Kant means by _a priori_. His unfortunate\ntendency to explain the term negatively could be remedied if it could\nbe held either that the term refers solely to mathematical judgements\nor that he considers the truth of the law of causality to be\napprehended in the same way that we see that two and two are four. For\nan _a priori_ judgement could then be defined as one in which the\nmind, on the presentation of an individual in perception or\nimagination, and in virtue of its capacity of thinking, apprehends the\nnecessity of a specific relation. But this definition is precluded by\nKant's view that the law of causality and similar principles, though\n_a priori_, are not self-evident.\n\n\n\n\nCHAPTER II\n\nTHE SENSIBILITY AND THE UNDERSTANDING\n\n\nThe distinction between the sensibility and the understanding[1] is\nto Kant fundamental both in itself and in relation to the conclusions\nwhich he reaches. An outline, therefore, of this distinction must\nprecede any statement or examination of the details of his position.\nUnfortunately, in spite of its fundamental character, Kant never\nthinks of questioning or criticizing the distinction in the form in\nwhich he draws it, and the presence of certain confusions often\nrenders it difficult to be sure of his meaning.\n\n    [1] Cf. B. 1, 29, 33, 74-5, 75, 92-4; M. 1, 18, 21, 45-46,\n    57.\n\nThe distinction may be stated in his own words thus: \"There are two\nstems of human knowledge, which perhaps spring from a common but to us\nunknown root, namely sensibility and understanding.\"[2] \"Our knowledge\nsprings from two fundamental sources of the mind; the first receives\nrepresentations[3] (receptivity for impressions); the second is the\npower of knowing an object by means of these representations\n(spontaneity of conceptions). Through the first an object is _given_\nto us; through the second the object is _thought_ in relation to the\nrepresentation (which is a mere determination of the mind). Perception\nand conceptions constitute, therefore, the elements of all our\nknowledge, so that neither conceptions without a perception in some\nway corresponding to them, nor perception without conceptions can\nyield any knowledge.... Neither of these qualities has a preference\nover the other. Without sensibility no object would be given to us,\nand without understanding no object would be thought. Thoughts without\ncontent are empty, perceptions without conceptions are blind. Hence it\nis as necessary for the mind to make its conceptions sensuous (i. e.\nto add to them the object in perception) as to make its perceptions\nintelligible (i. e. to bring them under conceptions). Neither of these\npowers or faculties can exchange its function. The understanding\ncannot perceive, and the senses cannot think. Only by their union can\nknowledge arise.\"[4]\n\n    [2] B. 29, M. 18\n\n    [3] For the sake of uniformity _Vorstellung_ has throughout\n    been translated by 'representation', though sometimes, as in\n    the present passage, it would be better rendered by\n    'presentation'.\n\n    [4] B. 74-5, M. 45-6.\n\nThe distinction so stated appears straightforward and, on the\nwhole,[5] sound. And it is fairly referred to by Kant as the\ndistinction between the faculties of perceiving and conceiving or\nthinking, provided that the terms perceiving and conceiving or\nthinking be taken to indicate a distinction within perception in the\nordinary sense of the word. His meaning can be stated thus: 'All\nknowledge requires the realization of two conditions; an individual\nmust be presented to us in perception, and we as thinking beings must\nbring this individual under or recognize it as an instance of some\nuniversal. Thus, in order to judge 'This is a house' or 'That is red'\nwe need the presence of the house or of the red colour in perception,\nand we must 'recognize' the house or the colour, i. e. apprehend the\nindividual as a member of a certain kind. Suppose either condition\nunrealized. Then if we suppose a failure to conceive, i. e. to\napprehend the individual as a member of some kind, we see that our\nperception--if it could be allowed to be anything at all--would be\nblind i. e. indeterminate, or a mere 'blur'. What we perceived would\nbe for us as good as nothing. In fact, we could not even say that we\nwere perceiving. Again, if we suppose that we had merely the\nconception of a house, and neither perceived nor had perceived an\nindividual to which it applied, we see that the conception, being\nwithout application, would be neither knowledge nor an element in\nknowledge. Moreover, the content of a conception is derived from\nperception; it is only through its relation to perceived individuals\nthat we become aware of a universal. To know the meaning of 'redness'\nwe must have experienced individual red things; to know the meaning of\n'house' we must at least have had experience of individual men and of\ntheir physical needs. Hence 'conceptions' without 'perceptions' are\nvoid or empty. The existence of conceptions presupposes experience of\ncorresponding individuals, even though it also implies the activity of\nthinking in relation to these individuals.'[6]\n\n    [5] Cf. p. 29, note 1.\n\n    [6] Kant's account implies that he has in view only empirical\n    knowledge; in any case it only applies to empirical\n    conceptions.\n\nFurther, it is true to say that as perceiving we are passive; we do\nnot do anything. This, as has been pointed out, is the element of\ntruth contained in the statement that objects are _given_ to us. On\nthe other hand, it may be truly said that as conceiving, in the sense\nof bringing an individual under a universal, we are essentially\nactive. This is presupposed by the notice or attention involved in\nperception ordinarily so called, i. e. perception in the full sense\nin which it includes conceiving as well as perceiving.[7] Kant,\ntherefore, is justified in referring to the sensibility as a\n'receptivity' and to the understanding as a 'spontaneity'.\n\n    [7] This distinction within perception is of course\n    compatible with the view that the elements so distinguished\n    are inseparable.\n\nThe distinction, so stated, appears, as has been already said,\nintelligible and, in the main[8], valid. Kant, however, renders the\nelucidation of his meaning difficult by combining with this view of\nthe distinction an incompatible and unwarranted theory of perception.\nHe supposes,[9] without ever questioning the supposition, that\nperception is due to the operation of things outside the mind, which\nact upon our sensibility and thereby produce sensations. On this\nsupposition, what we perceive is not, as the distinction just stated\nimplies, the thing itself, but a sensation produced by it.\nConsequently a problem arises as to the meaning on this supposition of\nthe statements 'by the sensibility objects are given to us' and 'by\nthe understanding they are thought'. The former statement must mean\nthat when a thing affects us there is a sensation. It cannot mean that\nby the sensibility we know that there exists a thing which causes the\nsensation, for this knowledge would imply the activity of thinking;\nnor can it mean that in virtue of the sensibility the thing itself is\npresented to us. The latter statement must mean that when sensation\narises, the understanding judges that there is something causing it;\nand this assertion must really be _a priori_, because not dependent\nupon experience. Unfortunately the two statements so interpreted are\nwholly inconsistent with the account of the functions of the\nsensibility and the understanding which has just been quoted.\n\n    [8] See p. 29, note 1.\n\n    [9] Cf. B. 1, M. 1.\n\nFurther, this theory of perception has two forms. In its first form\nthe theory is physical rather than metaphysical, and is based upon\nour possession of physical organs. It assumes that the reality to be\napprehended is the world of space and time, and it asserts that by the\naction of bodies upon our physical organs our sensibility is affected,\nand that thereby sensations are originated in us. Thereupon a problem\narises. For if the contribution of the sensibility to our knowledge\nof the physical world is limited to a succession of sensations,\nexplanation must be given of the fact that we have succeeded with\nan experience confined to these sensations in acquiring knowledge\nof a world which does not consist of sensations.[10] Kant, in fact,\nin the _Aesthetic_ has this problem continually before him, and\ntries to solve it. He holds that the mind, by means of its forms of\nperception and its conceptions of the understanding, superinduces upon\nsensations, as data, spatial and other relations, in such a way that\nit acquires knowledge of the spatial world.\n\n    [10] Cf. B. 1 init., M. 1 init.; B. 34, M. 21 sub fin.\n\nAn inherent difficulty, however, of this 'physical' theory of\nperception leads to a transformation of it. If, as the theory\nsupposes, the cause of sensation is outside or beyond the mind, it\ncannot be known. Hence the initial assumption that this cause is the\nphysical world has to be withdrawn, and the cause of sensation comes\nto be thought of as the thing in itself of which we can know nothing.\nThis is undoubtedly the normal form of the theory in Kant's mind.\n\nIt may be objected that to attribute to Kant at any time the physical\nform of the theory is to accuse him of an impossibly crude confusion\nbetween things in themselves and the spatial world, and that he can\nnever have thought that the cause of sensation, being as it is outside\nthe mind, is spatial. But the answer is to be found in the fact that\nthe problem just referred to as occupying Kant's attention in the\n_Aesthetic_ is only a problem at all so long as the cause of sensation\nis thought of as a physical body. For the problem 'How do we,\nbeginning with mere sensation, come to know a spatial and temporal\nworld?' is only a problem so long as it is supposed that the cause of\nsensation is a spatial and temporal world or a part of it, and that\nthis world is what we come to know. If the cause of sensation, as\nbeing beyond the mind, is held to be unknowable and so not known to be\nspatial or temporal, the problem has disappeared. Corroboration is\ngiven by certain passages[11] in the _Critique_ which definitely\nmention 'the senses', a term which refers to bodily organs, and by\nothers[12] to which meaning can be given only if they are taken to\nimply that the objects which affect our sensibility are not unknown\nthings in themselves, but things known to be spatial. Even the use of\nthe plural in the term 'things in themselves' implies a tendency to\nidentify the unknowable reality beyond the mind with bodies in space.\nFor the implication that different sensations are due to different\nthings in themselves originates in the view that different sensations\nare due to the operation of different spatial bodies.\n\n    [11] E. g. B. 1 init., M. 1 init., and B. 75 fin., M. 46,\n    lines 12, 13 [for 'the sensuous faculty' should be\n    substituted 'the senses'].\n\n    [12] E. g. B. 42, lines 11, 12; M. 26, line 13; A. 100, Mah.\n    195 ('even in the absence of the object'). Cf. B. 182-3, M.\n    110-1 (see pp. 257-8, and note p. 257), and B. 207-10, M.\n    126-8 (see pp. 263-5).\n\nIt is now necessary to consider how the distinction between the\nsensibility and the understanding contributes to articulate the\nproblem 'How are _a priori_ synthetic judgements possible?' As has\nbeen pointed out, Kant means by this question, 'How is it possible\nthat the mind is able, in virtue of its own powers, to make universal\nand necessary judgements which anticipate its experience of objects?'\nTo this question his general answer is that it is possible and only\npossible because, so far from ideas, as is generally supposed, having\nto conform to things, the things to which our ideas or judgements\nrelate, viz. phenomena, must conform to the nature of the mind. Now,\nif the mind's knowing nature can be divided into the sensibility and\nthe understanding, the problem becomes 'How is it possible for the\nmind to make such judgements in virtue of its sensibility and its\nunderstanding?' And the answer will be that it is possible because the\nthings concerned, i. e. phenomena, must conform to the sensibility and\nthe understanding, i. e. to the mind's perceiving and thinking nature.\nBut both the problem and the answer, so stated, give no clue to the\nparticular _a priori_ judgements thus rendered possible nor to the\nnature of the sensibility and the understanding in virtue of which we\nmake them. It has been seen, however, that the judgements in question\nfall into two classes, those of mathematics and those which form the\npresuppositions of physics. And it is Kant's aim to relate these\nclasses to the sensibility and the understanding respectively. His\nview is that mathematical judgements, which, as such, deal with\nspatial and temporal relations, are essentially bound up with our\nperceptive nature, i. e. with our sensibility, and that the principles\nunderlying physics are the expression of our thinking nature, i. e. of\nour understanding. Hence if the vindication of this relation between\nour knowing faculties and the judgements to which they are held to\ngive rise is approached from the side of our faculties, it must be\nshown that our sensitive nature is such as to give rise to\nmathematical judgements, and that our understanding or thinking nature\nis such as to originate the principles underlying physics. Again, if\nthe account of this relation is to be adequate, it must be shown to be\nexhaustive, i. e. it must be shown that the sensibility and the\nunderstanding give rise to no other judgements. Otherwise there may be\nother _a priori_ judgements bound up with the sensibility and the\nunderstanding which the inquiry will have ignored. Kant, therefore, by\nhis distinction between the sensibility and the understanding, sets\nhimself another problem, which does not come into sight in the first\nformulation of the general question 'How are _a priori_ synthetic\njudgements possible?' He has to determine what _a priori_ judgements\nare related to the sensibility and to the understanding respectively.\nAt the same time the distinction gives rise to a division within the\nmain problem. His chief aim is to discover how it is that _a priori_\njudgements are universally applicable. But, as Kant conceives\nthe issue, the problem requires different treatment according\nas the judgements in question are related to the sensibility or\nto the understanding. Hence arises the distinction between the\n_Transcendental Aesthetic_ and the _Transcendental Analytic_, the\nformer dealing with the _a priori_ judgements of mathematics, which\nrelate to the sensibility, and the latter dealing with the _a priori_\nprinciples of physics, which originate in the understanding. Again,\nwithin each of these two divisions we have to distinguish two\nproblems, viz. 'What _a priori_ judgements are essentially related to\nthe faculty in question?' and 'How is it that they are applicable to\nobjects?'\n\nIt is important, however, to notice that the distinction between the\nsensibility and the understanding, in the form in which it serves as a\nbasis for distinguishing the _Aesthetic_ and the _Analytic_, is not\nidentical with or even compatible with the distinction, as Kant states\nit when he is considering the distinction in itself and is not\nthinking of any theory which is to be based upon it. In the latter\ncase the sensibility and the understanding are represented as\ninseparable faculties involved in _all_ knowledge.[13] Only from the\nunion of both can knowledge arise. But, regarded as a basis for the\ndistinction between the _Aesthetic_ and the _Analytic_, they are\nimplied to be the source of different kinds of knowledge, viz.\nmathematics and the principles of physics. It is no answer to this to\nurge that Kant afterwards points out that space as an object\npresupposes a synthesis which does not belong to sense. No doubt this\nadmission implies that even the apprehension of spatial relations\ninvolves the activity of the understanding. But the implication is\nreally inconsistent with the existence of the _Aesthetic_ as a\ndistinct part of the subject dealing with a special class of _a\npriori_ judgements.\n\n    [13] B. 74-5, M. 45-6; cf. pp. 27-9.\n\n    [14] B. 160 note, M. 98 note.\n\n\n\n\nCHAPTER III\n\nSPACE\n\n\nIt is the aim of the _Aesthetic_ to deal with the _a priori_ knowledge\nwhich relates to the sensibility. This knowledge, according to Kant,\nis concerned with space and time. Hence he has to show _firstly_ that\nour apprehension of space and time is _a priori_, i. e. that it is not\nderived from experience but originates in our apprehending nature; and\n_secondly_ that within our apprehending nature this apprehension\nbelongs to the sensibility and not to the understanding, or, in his\nlanguage, that space and time are forms of perception or sensibility.\nFurther, if his treatment is to be exhaustive, he should also show\n_thirdly_ that space and time are the only forms of perception. This,\nhowever, he makes no attempt to do except in one passage,[1] where the\nargument fails. The first two points established, Kant is able to\ndevelop his main thesis, viz. that it is a condition of the validity\nof the _a priori_ judgements which relate to space and time that these\nare characteristics of phenomena, and not of things in themselves.\n\n    [1] B. 58, M. 35.\n\nIt will be convenient to consider his treatment of space and time\nseparately, and to begin with his treatment of space. It is necessary,\nhowever, first of all to refer to the term 'form of perception'. As\nKant conceives a form of perception, it involves three antitheses.\n\n(1) As a _form_ of perception it is opposed, as a way or mode of\nperceiving, to particular perceptions.\n\n(2) As a form or mode of _perception_ it is opposed to a form or mode\nof _conception_.\n\n(3) As a form of _perception_ it is also opposed, as a way in which we\napprehend things, to a way in which things are.\n\nWhile we may defer consideration of the second and third antitheses,\nwe should at once give attention to the nature of the first, because\nKant confuses it with two other antitheses. There is no doubt that in\ngeneral a _form_ of perception means for Kant a general capacity of\nperceiving which, as such, is opposed to the actual perceptions in\nwhich it is manifested. For according to him our spatial perceptions\nare not foreign to us, but manifestations of our general perceiving\nnature; and this view finds expression in the assertion that space is\na form of perception or of sensibility.[2]\n\n    [2] Cf. B. 43 init., M. 26 med.\n\nUnfortunately, however, Kant frequently speaks of this form of\nperception as if it were the same thing as the actual perception of\nempty space.[3] In other words, he implies that such a perception is\npossible, and confuses it with a potentiality, i. e. the power of\nperceiving that which is spatial. The confusion is possible because it\ncan be said with some plausibility that a perception of empty\nspace--if its possibility be allowed--does not inform us about actual\nthings, but only informs us what must be true of things, if there\nprove to be any; such a perception, therefore, can be thought of as a\npossibility of knowledge rather than as actual knowledge.\n\n    [3] e. g. B. 34, 35, M. 22; B. 41, M. 25; _Prol._ Secs. 9-11.\n    The commonest expression of the confusion is to be found in\n    the repeated assertion that space is a pure perception.\n\nThe second confusion is closely related to the first, and arises from\nthe fact that Kant speaks of space not only as a form of _perception_,\nbut also as the form of _phenomena_ in opposition to sensation as\ntheir matter. \"That which in the phenomenon corresponds to[4] the\nsensation I term its matter; but that which effects that the manifold\nof the phenomenon can be arranged under certain relations I call the\nform of the phenomenon. Now that in which alone our sensations can be\narranged and placed in a certain form cannot itself be sensation.\nHence while the matter of all phenomena is only given to us _a\nposteriori_, their form [i. e. space] must lie ready for them all\ntogether _a priori_ in the mind.\"[5] Here Kant is clearly under the\ninfluence of his theory of perception.[6] He is thinking that, given\nthe origination of sensations in us by the thing in itself, it is the\nbusiness of the mind to arrange these sensations spatially in order to\nattain knowledge of the spatial world.[7] Space being, as it were,\na kind of empty vessel in which sensations are arranged, is said\nto be the form of phenomena.[8] Moreover, if we bear in mind that\nultimately bodies in space are for Kant only spatial arrangements\nof sensations,[9] we see that the assertion that space is the form\nof phenomena is only Kant's way of saying that all bodies are\nspatial.[10] Now Kant, in thus asserting that space is the form of\nphenomena, is clearly confusing this assertion with the assertion that\nspace is a form of perception, and he does so in consequence of the\nfirst confusion, viz. that between a capacity of perceiving and an\nactual perception of empty space. For in the passage last quoted he\ncontinues thus: \"I call all representations[11] _pure_ (in the\ntranscendental sense) in which nothing is found which belongs to\nsensation. Accordingly there will be found _a priori_ in the mind the\npure form of sensuous perceptions in general, wherein all the manifold\nof phenomena is perceived in certain relations. This pure form of\nsensibility will also itself be called _pure perception_. Thus,\nif I abstract from the representation of a body that which the\nunderstanding thinks respecting it, such as substance, force,\ndivisibility, &c., and also that which belongs to sensation, such as\nimpenetrability, hardness, colour, &c., something is still left over\nfor me from this empirical perception, viz. extension and shape. These\nbelong to pure perception, which exists in the mind _a priori_, even\nwithout an actual object of the senses or a sensation, as a mere form\nof sensibility.\" Here Kant has passed, without any consciousness of a\ntransition, from treating space as that in which the manifold of\nsensation is arranged to treating it as a capacity of perceiving.\nMoreover, since Kant in this passage speaks of space as a perception,\nand thereby identifies space with the perception of it,[12] the\nconfusion may be explained thus. The form of phenomena is said to be\nthe space in which all sensations are arranged, or in which all bodies\nare; space, apart from all sensations or bodies, i. e. empty, being\nthe object of a pure perception, is treated as identical with a pure\nperception, viz. the perception of empty space; and the perception of\nempty space is treated as identical with a capacity of perceiving that\nwhich is spatial.[13]\n\n    [4] 'Corresponds to' must mean 'is'.\n\n    [5] B. 34, M. 21.\n\n    [6] Cf. pp. 30-2.\n\n    [7] It is impossible, of course, to see how such a process\n    can give us knowledge of the spatial world, for, whatever\n    bodies in space are, they are not arrangements of sensations.\n    Nevertheless, Kant's theory of perception really precludes\n    him from holding that bodies are anything else than\n    arrangements of sensations, and he seems at times to accept\n    this view explicitly, e. g. B. 38, M. 23 (quoted p. 41),\n    where he speaks of our representing sensations as external\n    to and next to each other, and, therefore, as in different\n    places.\n\n    [8] It may be noted that it would have been more natural to\n    describe the particular shape of the phenomenon (i. e. the\n    particular spatial arrangement of the sensations) rather than\n    space as the form of the phenomenon; for the matter to which\n    the form is opposed is said to be sensation, and that of\n    which it is the matter is said to be the phenomenon, i. e.\n    a body in space.\n\n    [9] Cf. note 4, p. 38.\n\n    [10] Cf. _Prol._ Sec. 11 and p. 137.\n\n    [11] Cf. p. 41, note 1.\n\n    [12] Cf. p. 51, note 1.\n\n    [13] The same confusion (and due to the same cause) is\n    implied _Prol_. Sec. 11, and B. 42 (b), M. 26 (b) first\n    paragraph. Cf. B. 49 (b), M. 30 (b).\n\nThe existence of the confusion, however, is most easily realized by\nasking, 'How did Kant come to think of space and time as the _only_\nforms of perception?' It would seem obvious that the perception of\n_anything_ implies a form of perception in the sense of a mode or\ncapacity of perceiving. To perceive colours implies a capacity for\nseeing; to hear noises implies a capacity for hearing. And these\ncapacities may fairly be called forms of perception. As soon as this\nis realized, the conclusion is inevitable that Kant was led to think\nof space and time as the only forms of perception, because in this\nconnexion he was thinking of each as a form of phenomena, i. e. as\nsomething in which all bodies or their states are, or, from the point\nof view of our knowledge, as that in which sensuous material is to be\narranged; for there is nothing except space and time in which such\narrangement could plausibly be said to be carried out.\n\nAs has been pointed out, Kant's argument falls into two main parts,\none of which prepares the way for the other. The aim of the former is\nto show _firstly_ that our apprehension of space is _a priori_, and\n_secondly_ that it belongs to perception and not to conception. The\naim of the latter is to conclude from these characteristics of our\napprehension of space that space is a property not of things in\nthemselves but only of phenomena. These arguments may be considered in\nturn.\n\nThe really valid argument adduced by Kant for the _a priori_ character\nof our apprehension of space is based on the nature of geometrical\njudgements. The universality of our judgements in geometry is not\nbased upon experience, i. e. upon the observation of individual things\nin space. The necessity of geometrical relations is apprehended\ndirectly in virtue of the mind's own apprehending nature.\nUnfortunately in the present context Kant ignores this argument and\nsubstitutes two others, both of which are invalid.\n\n1. \"Space is no empirical conception[14] which has been derived from\nexternal[15] experiences. For in order that certain sensations may be\nrelated to something external to me (that is, to something in a\ndifferent part of space from that in which I am), in like manner, in\norder that I may represent them as external to and next to each other,\nand consequently as not merely different but as in different places,\nthe representation of space must already exist as a foundation.\nConsequently, the representation of space cannot be borrowed from the\nrelations of external phenomena through experience; but, on the\ncontrary, this external experience is itself first possible only\nthrough the said representation.\"[16] Here Kant is thinking that in\norder to apprehend, for example, that A is to the right of B we must\nfirst apprehend empty space. He concludes that our apprehension of\nspace is _a priori_, because we apprehend empty space _before_ we\nbecome aware of the spatial relations of individual objects in it.\n\n    [14] _Begriff_ (conception) here is to be understood loosely\n    not as something opposed to _Anschauung_ (perception), but as\n    equivalent to the genus of which _Anschauung_ and _Begriff_\n    are species, i. e. _Vorstellung_, which maybe rendered by\n    'representation' or 'idea', in the general sense in which\n    these words are sometimes used to include 'thought' and\n    'perception'.\n\n    [15] The next sentence shows that 'external' means, not\n    'produced by something external to the mind', but simply\n    'spatial'.\n\n    [16] B. 38, M. 23-4.\n\nTo this the following reply may be made. (_a_) The term _a priori_\napplied to an apprehension should mean, not that it arises prior to\nexperience, but that its validity is independent of experience. (_b_)\nThat to which the term _a priori_ should be applied is not the\napprehension of empty space, which is individual, but the apprehension\nof the nature of space in general, which is universal. (_c_) We do not\napprehend empty space before we apprehend individual spatial relations\nof individual bodies or, indeed, at any time. (_d_) Though we come to\napprehend _a priori_ the nature of space in general, the apprehension\nis not prior but posterior in time to the apprehension of individual\nspatial relations. (_e_) It does not follow from the temporal priority\nof our apprehension of individual spatial relations that our\napprehension of the nature of space in general is 'borrowed from\nexperience', and is therefore not _a priori_.\n\n2. \"We can never represent to ourselves that there is no space, though\nwe can quite well think that no objects are found in it. It must,\ntherefore, be considered as the condition of the possibility of\nphenomena, and not as a determination dependent upon them, and it is\nan _a priori_ representation, which necessarily underlies external\nphenomena.\"[17]\n\n    [17] B. 38, M. 24.\n\nHere the premise is simply false. If 'represent' or 'think' means\n'believe', we can no more represent or think that there are no\nobjects in space than that there is no space. If, on the other hand,\n'represent' or 'think' means 'make a mental picture of', the assertion\nis equally false. Kant is thinking of empty space as a kind of\nreceptacle for objects, and the _a priori_ character of our\napprehension of space lies, as before, in the supposed fact that\nin order to apprehend objects in space we must begin with the\napprehension of empty space.\n\nThe examination of Kant's arguments for the _perceptive_ character\nof our apprehension of space is a more complicated matter. By way of\npreliminary it should be noticed that they presuppose the possibility\nin general of distinguishing features of objects which belong to the\nperception of them from others which belong to the conception of them.\nIn particular, Kant holds that our apprehension of a body as a\nsubstance, as exercising force and as divisible, is due to our\nunderstanding as conceiving it, while our apprehension of it as\nextended and as having a shape is due to our sensibility as perceiving\nit.[18] The distinction, however, will be found untenable in\nprinciple; and if this be granted, Kant's attempt to distinguish in\nthis way the extension and shape of an object from its other features\ncan be ruled out on general grounds. In any case, it must be conceded\nthat the arguments fail by which he seeks to show that space in\nparticular belongs to perception.\n\n    [18] B. 35, M. 22 (quoted p. 39). It is noteworthy (1) that\n    the passage contains no _argument_ to show that extension and\n    shape are not, equally with divisibility, _thought_ to belong\n    to an object, (2) that impenetrability, which is here said\n    to belong to sensation, obviously cannot do so, and (3) that\n    (as has been pointed out, p. 39) the last sentence of the\n    paragraph in question presupposes that we have a perception\n    of empty space, and that this is a _form_ of perception.\n\nThere appears to be no way of distinguishing perception and conception\nas the apprehension of different realities[19] except as the\napprehension of the individual and of the universal respectively.\nDistinguished in this way, the faculty of perception is that in virtue\nof which we apprehend the individual, and the faculty of conception is\nthat power of reflection in virtue of which a universal is made the\nexplicit object of thought.[20] If this be granted, the only test for\nwhat is perceived is that it is individual, and the only test for what\nis conceived is that it is universal. These are in fact the tests\nwhich Kant uses. But if this be so, it follows that the various\ncharacteristics of objects cannot be divided into those which are\nperceived and those which are conceived. For the distinction between\nuniversal and individual is quite general, and applies to all\ncharacteristics of objects alike. Thus, in the case of colour, we can\ndistinguish colour in general and the individual colours of individual\nobjects; or, to take a less ambiguous instance, we can distinguish a\nparticular shade of redness and its individual instances. Further, it\nmay be said that perception is of the individual shade of red of the\nindividual object, and that the faculty by which we become explicitly\naware of the particular shade of red in general is that of conception.\nThe same distinction can be drawn with respect to hardness, or shape,\nor any other characteristic of objects. The distinction, then,\nbetween perception and conception can be drawn with respect to any\ncharacteristic of objects, and does not serve to distinguish one from\nanother.\n\n    [19] And _not_ as mutually involved in the apprehension\n    of any individual reality.\n\n    [20] This distinction is of course different to that\n    previously drawn _within_ perception in the full sense\n    between perception in a narrow sense and conception\n    (pp. 28-9).\n\nKant's arguments to show that our apprehension of space belongs to\nperception are two in number, and both are directed to show not, as\nthey should, that space is a _form_ of perception, but that it is a\n_perception_.[21] The first runs thus: \"Space is no discursive, or, as\nwe say, general conception of relations of things in general, but a\npure perception. For, in the first place, we can represent to\nourselves only one space, and if we speak of many spaces we mean\nthereby only parts of one and the same unique space. Again, these\nparts cannot precede the one all-embracing space as the component\nparts, as it were, out of which it can be composed, but can be thought\nonly in it. Space is essentially one; the manifold in it, and\nconsequently the general conception of spaces in general, rests solely\nupon limitations.\"[22]\n\n    [21] Kant uses the phrase 'pure perception'; but 'pure' can\n    only mean 'not containing sensation', and consequently adds\n    nothing relevant.\n\n    [22] B. 39, M. 24. The concluding sentences of the paragraph\n    need not be considered.\n\nHere Kant is clearly taking the proper test of perception. Its object,\nas being an individual, is unique; there is only one of it, whereas\nany conception has a plurality of instances. But he reaches his\nconclusion by supposing that we first perceive empty space and then\nbecome aware of its parts by dividing it. Parts of space are\nessentially limitations of the one space; therefore to apprehend them\nwe must first apprehend space. And since space is _one_, it must be\nobject of perception; in other words, space, in the sense of the one\nall-embracing space, i. e. the totality of individual spaces, is\nsomething perceived.\n\nThe argument appears open to two objections. In the _first_ place, we\ndo _not_ perceive space as a whole, and then, by dividing it, come to\napprehend individual spaces. We perceive individual spaces, or,\nrather, individual bodies occupying individual spaces.[23] We then\napprehend that these spaces, as spaces, involve an infinity of other\nspaces. In other words, it is reflection on the general nature of\nspace, the apprehension of which is involved in our apprehension of\nindividual spaces or rather of bodies in space, which gives rise to\nthe apprehension of the totality[24] of spaces, the apprehension being\nan act, not of perception, but of thought or conception. It is\nnecessary, then, to distinguish (_a_) individual spaces, which we\nperceive; (_b_) the nature of space in general, of which we become\naware by reflecting upon the character of perceived individual spaces,\nand which we conceive; (_c_) the totality of individual spaces, the\nthought of which we reach by considering the nature of space in\ngeneral.\n\n    [23] This contention is not refuted by the objection that our\n    distinct apprehension of an individual space is always bound\n    up with an indistinct apprehension of the spaces immediately\n    surrounding it. For our indistinct apprehension cannot be\n    supposed to be of the whole of the surrounding space.\n\n    [24] It is here assumed that a whole or a totality can be\n    infinite. Cf. p. 102.\n\nIn the _second_ place, the distinctions just drawn afford no ground\nfor distinguishing space as something perceived from any other\ncharacteristic of objects as something conceived; for any other\ncharacteristic admits of corresponding distinctions. Thus, with\nrespect to colour it is possible to distinguish (_a_) individual\ncolours which we perceive; (_b_) colouredness in general, which we\nconceive by reflecting on the common character exhibited by individual\ncolours and which involves various kinds or species of colouredness;\n(_c_) the totality of individual colours, the thought of which is\nreached by considering the nature of colouredness in general.[25]\n\n    [25] For a possible objection and the answer thereto, see\n    note, p. 70.\n\nBoth in the case of colour and in that of space there is to be found\nthe distinction between universal and individual, and therefore also\nthat between conception and perception. It may be objected that after\nall, as Kant points out, there is only one space, whereas there are\nmany individual colours. But the assertion that there is only one\nspace simply means that all individual bodies in space are related\nspatially. This will be admitted, if the attempt be made to think of\ntwo bodies as in different spaces and therefore as not related\nspatially. Moreover, there is a parallel in the case of colour, since\nindividual coloured bodies are related by way of colour, e. g. as\nbrighter and duller; and though such a relation is different from a\nrelation of bodies in respect of space, the difference is due to the\nspecial nature of the universals conceived, and does not imply a\ndifference between space and colour in respect of perception and\nconception. In any case, space as a whole is not object of perception,\nwhich it must be if Kant is to show that space, as being one, is\nperceived; for space in this context must mean the totality of\nindividual spaces.\n\nKant's second argument is stated as follows: \"Space is represented as\nan infinite _given_ magnitude. Now every conception must indeed be\nconsidered as a representation which is contained in an infinite\nnumber of different possible representations (as their common mark),\nand which therefore contains these _under itself_, but no conception\ncan, as such, be thought of as though it contained _in itself_ an\ninfinite number of representations. Nevertheless, space is so\nconceived, for all parts of space _ad infinitum_ exist simultaneously.\nConsequently the original representation of space is an _a priori\nperception_ and not a _conception_.\" In other words, while a\nconception implies an infinity of individuals which come under it, the\nelements which constitute the conception itself (e. g. that of\ntriangularity or redness) are not infinite; but the elements which go\nto constitute space are infinite, and therefore space is not a\nconception but a perception.\n\nThough, however, space in the sense of the infinity of spaces may be\nsaid to contain an infinite number of spaces if it be meant that it\n_is_ these infinite spaces, it does not follow, nor is it true, that\nspace in this sense is object of perception.\n\nThe aim of the arguments just considered, and stated in Sec. 2 of\nthe _Aesthetic_, is to establish the two characteristics of our\napprehension of space,[26] from which it is to follow that space is a\nproperty of things only as they appear to us and not as they are in\nthemselves. This conclusion is drawn in Sec. 4. Secs. 2 and 4 therefore\ncomplete the argument. Sec. 3, a passage added in the second edition\nof the _Critique_, interrupts the thought, for ignoring Sec. 2, it once\nmore establishes the _a priori_ and perceptive character of our\napprehension of space, and independently draws the conclusion drawn in\nSec. 4. Since, however, Kant draws the final conclusion in the same way\nin Sec. 3 and in Sec. 4, and since a passage in the _Prolegomena_,[27]\nof which Sec. 3 is only a summary, gives a more detailed account of\nKant's thought, attention should be concentrated on Sec. 3, together\nwith the passage in the _Prolegomena_.\n\n    [26] viz. that it is _a priori_ and a pure perception.\n\n    [27] Secs. 6-11.\n\nIt might seem at the outset that since the arguments upon which Kant\nbases the premises for his final argument have turned out invalid, the\nfinal argument itself need not be considered. The argument, however,\nof Sec. 3 ignores the preceding arguments for the _a priori_ and\nperceptive character of our apprehension of space. It returns to the\n_a priori_ synthetic character of geometrical judgements, upon which\nstress is laid in the Introduction, and appeals to this as the\njustification of the _a priori_ and perceptive character of our\napprehension of space.\n\nThe argument of Sec. 3 runs as follows: \"Geometry is a science which\ndetermines the properties of space synthetically and yet _a priori_.\nWhat, then, must be the representation of space, in order that such a\nknowledge of it may be possible? It must be originally perception, for\nfrom a mere conception no propositions can be deduced which go beyond\nthe conception, and yet this happens in geometry. But this perception\nmust be _a priori_, i. e. it must occur in us before all\nsense-perception of an object, and therefore must be pure, not\nempirical perception. For geometrical propositions are always\napodeictic, i. e. bound up with the consciousness of their necessity\n(e. g. space has only three dimensions), and such propositions cannot\nbe empirical judgements nor conclusions from them.\"\n\n\"Now how can there exist in the mind an external perception[28] which\nprecedes[29] the objects themselves, and in which the conception of\nthem can be determined _a priori_? Obviously not otherwise than in so\nfar as it has its seat in the subject only, as the formal nature of\nthe subject to be affected by objects and thereby to obtain _immediate\nrepresentation_, i. e. _perception_ of them, and consequently only as\nthe form of the external sense in general.\"[30]\n\n    [28] 'External perception' can only mean perception of what\n    is spatial.\n\n    [29] _Vorhergeht._\n\n    [30] 'Formal nature _to be affected by objects_' is not\n    relevant to the context.\n\nHere three steps are taken. From the _synthetic_ character of\ngeometrical judgements it is concluded that space is not something\nwhich we _conceive_, but something which we _perceive_. From their _a\npriori_ character, i. e. from the consciousness of necessity involved,\nit is concluded that the perception of space must be _a priori_ in a\nnew sense, that of taking place _before_ the perception of objects in\nit.[31] From the fact that we perceive space before we perceive\nobjects in it, and thereby are able to anticipate the spatial\nrelations which condition these objects, it is concluded that space is\nonly a characteristic of our perceiving nature, and consequently that\nspace is a property not of things in themselves, but only of things as\nperceived by us.[32]\n\n    [31] Cf. B. 42, M. 26 (a) fin., (b) second sentence.\n\n    [32] Cf. B. 43, M. 26-7.\n\nTwo points in this argument are, even on the face of it, paradoxical.\nFirstly, the term _a priori_, as applied not to geometrical judgements\nbut to the perception of space, is given a temporal sense; it means\nnot something whose validity is independent of experience and which is\nthe manifestation of the nature of the mind, but something which takes\nplace before experience. Secondly, the conclusion is not that the\nperception of space _is the manifestation of_ the mind's perceiving\nnature, but that it _is_ the mind's perceiving nature. For the\nconclusion is that space[33] is the formal nature of the subject to be\naffected by objects, and therefore the form of the external sense in\ngeneral. Plainly, then, Kant here confuses an actual perception and a\nform or way of perceiving. These points, however, are more explicit in\nthe corresponding passage in the _Prolegomena_.[34]\n\n    [33] Kant draws no distinction between space and the\n    perception of space, or, rather, habitually speaks of space\n    as a perception. No doubt he considers that his view that\n    space is only a characteristic of phenomena justifies\n    the identification of space and the perception of it.\n    Occasionally, however, he distinguishes them. Thus he\n    sometimes speaks of the representation of space (e. g.\n    B. 38-40, M. 23-4); in _Prol._, Sec. 11, he speaks of a pure\n    perception of space and time; and in B. 40, M. 25, he says\n    that our representation of space must be perception. But this\n    language is due to the pressure of the facts, and not to his\n    general theory; cf. pp. 135-6.\n\n    [34] Secs. 6-11.\n\nIt begins thus: \"Mathematics carries with it thoroughly apodeictic\ncertainty, that is, absolute necessity, and, therefore, rests on no\nempirical grounds, and consequently is a pure product of reason, and,\nbesides, is thoroughly synthetical. How, then, is it possible for\nhuman reason to accomplish such knowledge entirely _a priori_?... But\nwe find that all mathematical knowledge has this peculiarity, that it\nmust represent its conception previously in _perception_, and indeed\n_a priori_, consequently in a perception which is not empirical but\npure, and that otherwise it cannot take a single step. Hence its\njudgements are always _intuitive_.... This observation on the nature\nof mathematics at once gives us a clue to the first and highest\ncondition of its possibility, viz. that there must underlie it _a pure\nperception_ in which it can exhibit or, as we say, _construct_ all its\nconceptions in the concrete and yet _a priori_. If we can discover\nthis pure perception and its possibility, we may thence easily explain\nhow _a priori_ synthetical propositions in pure mathematics are\npossible, and consequently also how the science itself is possible.\nFor just as empirical perception enables us without difficulty to\nenlarge synthetically in experience the conception which we frame of\nan object of perception through new predicates which perception itself\noffers us, so pure perception also will do the same, only with the\ndifference that in this case the synthetical judgement will be _a\npriori_ certain and apodeictic, while in the former case it will be\nonly _a posteriori_ and empirically certain; for the latter [i. e. the\nempirical perception on which the _a posteriori_ synthetic judgement\nis based] contains only that which is to be found in contingent\nempirical perception, while the former [i. e. the pure perception on\nwhich the _a priori_ synthetic judgement is based] contains that which\nis bound to be found in pure perception, since, as _a priori_\nperception, it is inseparably connected with the conception _before\nall experience_ or individual sense-perception.\"\n\nThis passage is evidently based upon the account which Kant gives in\nthe _Doctrine of Method_ of the method of geometry.[35] According to\nthis account, in order to apprehend, for instance, that a three-sided\nfigure must have three angles, we must draw in imagination or on paper\nan individual figure corresponding to the conception of a three-sided\nfigure. We then see that the very nature of the act of construction\ninvolves that the figure constructed must possess three angles as well\nas three sides. Hence, perception being that by which we apprehend the\nindividual, a perception is involved in the act by which we form a\ngeometrical judgement, and the perception can be called _a priori_, in\nthat it is guided by our _a priori_ apprehension of the necessary\nnature of the act of construction, and therefore of the figure\nconstructed.\n\n    [35] B. 740 ff., M. 434 ff. Compare especially the following:\n    \"_Philosophical_ knowledge is _knowledge of reason_ by means\n    of _conceptions_; mathematical knowledge is knowledge by\n    means of the _construction_ of conceptions. But the\n    _construction_ of a conception means the _a priori_\n    presentation of a perception corresponding to it. The\n    construction of a conception therefore demands a\n    _non-empirical_ perception, which, therefore, as a\n    perception, is an _individual_ object, but which none the\n    less, as the construction of a conception (a universal\n    representation), must express in the representation universal\n    validity for all possible perceptions which come under that\n    conception. Thus I construct a triangle by presenting the\n    object corresponding to the conception, either by mere\n    imagination in pure perception, or also, in accordance with\n    pure perception, on paper in empirical perception, but in\n    both cases completely _a priori_, without having borrowed\n    the pattern of it from any experience. The individual drawn\n    figure is empirical, but nevertheless serves to indicate the\n    conception without prejudice to its universality, because in\n    this empirical perception we always attend only to the act of\n    construction of the conception, to which many determinations,\n    e. g. the magnitude of the sides and of the angles, are\n    wholly indifferent, and accordingly abstract from these\n    differences, which do not change the conception of the\n    triangle.\"\n\nThe account in the _Prolegomena_, however, differs from that of the\n_Doctrine of Method_ in one important respect. It asserts that the\nperception involved in a mathematical judgement not only may, but\nmust, be pure, i. e. must be a perception in which no spatial object\nis present, and it implies that the perception must take place\n_before_ all experience of actual objects.[36] Hence _a priori_,\napplied to perception, has here primarily, if not exclusively, the\ntemporal meaning that the perception takes place _antecedently to all\nexperience_.[37]\n\n    [36] This becomes more explicit in Sec. 8 and ff.\n\n    [37] This is also, and more obviously, implied in Secs. 8-11.\n\nThe thought of the passage quoted from the _Prolegomena_ can be stated\nthus: 'A mathematical judgement implies the perception of an\nindividual figure antecedently to all experience. This may be said to\nbe the first condition of the possibility of mathematical judgements\nwhich is revealed by reflection. There is, however, a prior or higher\ncondition. The perception of an individual figure involves as its\nbasis another pure perception. For we can only construct and therefore\nperceive an individual figure in empty space. Space is that _in which_\nit must be constructed and perceived. A perception[38] of empty space\nis, therefore, necessary. If, then, we can discover how this\nperception is possible, we shall be able to explain the possibility of\n_a priori_ synthetical judgements of mathematics.'\n\n    [38] _Pure_ perception only means that the space perceived is\n    empty.\n\nKant continues as follows: \"But with this step the difficulty seems to\nincrease rather than to lessen. For henceforward the question is '_How\nis it possible to perceive anything a priori?_' A perception is such a\nrepresentation as would immediately depend upon the presence of the\nobject. Hence it seems impossible _originally_ to perceive _a priori_,\nbecause perception would in that case have to take place without an\nobject to which it might refer, present either formerly or at the\nmoment, and accordingly could not be perception.... How can\n_perception_ of the object precede the object itself?\"[39] Kant here\nfinds himself face to face with the difficulty created by the\npreceding section. Perception, as such, involves the actual presence\nof an object; yet the pure perception of space involved by\ngeometry--which, as pure, is the perception of empty space, and which,\nas the perception of empty space, is _a priori_ in the sense of\ntemporally prior to the perception of actual objects--presupposes that\nan object is not actually present.\n\n    [39] _Prol._ Sec. 8.\n\nThe solution is given in the next section. \"Were our perception\nnecessarily of such a kind as to represent things _as they are in\nthemselves_, no perception would take place _a priori_, but would\nalways be empirical. For I can only know what is contained in the\nobject in itself, if it is present and given to me. No doubt it is\neven then unintelligible how the perception of a present thing should\nmake me know it as it is in itself, since its qualities cannot migrate\nover into my faculty of representation; but, even granting this\npossibility, such a perception would not occur _a priori_, i. e.\nbefore the object was presented to me; for without this presentation,\nno basis of the relation between my representation and the object can\nbe imagined; the relation would then have to rest upon inspiration. It\nis therefore possible only in one way for my perception to precede the\nactuality of the object and to take place as _a priori_ knowledge,\nviz. _if it contains nothing but the form of the sensibility, which\nprecedes in me, the subject, all actual impressions through which I am\naffected by objects_. For I can know _a priori_ that objects of the\nsenses can only be perceived in accordance with this form of the\nsensibility. Hence it follows that propositions which concern merely\nthis form of sensuous perception will be possible and valid for\nobjects of the senses, and in the same way, conversely, that\nperceptions which are possible _a priori_ can never concern any things\nother than objects of our senses.\"\n\nThis section clearly constitutes the turning-point in Kant's argument,\nand primarily expresses, in an expanded form, the central doctrine of\nSec. 3 of the _Aesthetic_, that an external perception anterior to\nobjects themselves, and in which our conceptions of objects can be\ndetermined _a priori_, is possible, if, and only if, it has its seat\nin the subject as its formal nature of being affected by objects, and\nconsequently as the form of the external sense in general. It argues\nthat, since this is true, and since geometrical judgements involve\nsuch a perception anterior to objects, space must be only the[40] form\nof sensibility.\n\n    [40] _The_ and not _a_, because, for the moment, time is\n    ignored.\n\nNow why does Kant think that this conclusion follows? Before we can\nanswer this question we must remove an initial difficulty. In this\npassage Kant unquestionably identifies a form of perception with an\nactual perception. It is at once an actual perception and a capacity\nof perceiving. This is evident from the words, \"It is possible only in\none way for my perception to precede the actuality of the object ...\nviz. _if it contains nothing but the form of the sensibility_.\"[41]\nThe identification becomes more explicit a little later. \"A pure\nperception (of space and time) can underlie the empirical perception\nof objects, because it is nothing but the mere form of the\nsensibility, which precedes the actual appearance of the objects, in\nthat it in fact first makes them possible. Yet this faculty of\nperceiving _a priori_ affects not the matter of the phenomenon, i. e.\nthat in it which is sensation, for this constitutes that which is\nempirical, but only its form, viz. space and time.\"[42] His argument,\nhowever, can be successfully stated without this identification. It is\nonly necessary to re-write his cardinal assertion in the form 'the\nperception of space must be nothing but the _manifestation_ of the\nform of the sensibility'. Given this modification, the question\nbecomes, 'Why does Kant think that the perception of empty space,\ninvolved by geometrical judgements, can be only a manifestation of our\nperceiving nature, and not in any way the apprehension of a real\nquality of objects?' The answer must be that it is because he thinks\nthat, while in empirical perception a real object is present, in the\nperception of empty space a real object is not present. He regards\nthis as proving that the latter perception is only of something\nsubjective or mental. \"Space and time, by being pure _a priori_\nperceptions, prove that they are mere forms of our sensibility which\nmust precede all empirical perception, i. e. sense-perception of\nactual objects.\"[43] His main conclusion now follows easily enough. If\nin perceiving empty space we are only apprehending a manifestation of\nour perceiving nature, what we apprehend in a geometrical judgement is\nreally a law of our perceiving nature, and therefore, while it _must_\napply to our perceptions of objects or to objects as perceived, it\n_cannot_ apply to objects apart from our perception, or, at least,\nthere is no ground for holding that it does so.\n\n    [41] _Prol._, Sec. 9.\n\n    [42] _Prol._, Sec. 11.\n\n    [43] _Prol._, Sec. 10.\n\nIf, however, this fairly represents Kant's thought, it must be allowed\nthat the conclusion which he should have drawn is different, and even\nthat the conclusion which he does draw is in reality incompatible with\nhis starting-point.\n\nHis starting-point is the view that the truth of geometrical\njudgements presupposes a perception of empty space, in virtue of which\nwe can discover rules of spatial relation which must apply to all\nspatial objects subsequently perceived. His problem is to discover the\npresupposition of this presupposition. The proper answer must be, not\nthat space is a form of sensibility or a way in which objects appear\nto us, but that space is the form of all objects, i. e. that all\nobjects are spatial.[44] For in that case they must be subject to the\nlaws of space, and therefore if we can discover these laws by a study\nof empty space, the only condition to be satisfied, if the objects of\nsubsequent perception are to conform to the laws which we discover, is\nthat all objects should be spatial. Nothing is implied which enables\nus to decide whether the objects are objects as they are in themselves\nor objects as perceived; for in either case the required result\nfollows. If in empirical perception we apprehend things only as they\nappear to us, and if space is the form of them as they appear to us,\nit will no doubt be true that the laws of spatial relation which we\ndiscover must apply to things as they appear to us. But on the other\nhand, if in empirical perception we apprehend things as they are, and\nif space is their form, i. e. if things are spatial, it will be\nequally true that the laws discovered by geometry must apply to things\nas they are.\n\n    [44] Kant expresses the assertion that space is the form of\n    all objects by saying that space is the form of _phenomena_.\n    This of course renders easy an unconscious transition from\n    the thesis that space is the form of objects to the quite\n    different thesis that space is the form of sensibility; cf.\n    p. 39.\n\nAgain, Kant's starting-point really commits him to the view that space\nis a characteristic of things as they are. For--paradoxical though it\nmay be--his problem is to explain the possibility of _perceiving a\npriori_, i. e. of _perceiving_ the characteristics of an object\nanterior to the actual presence of the object in perception.[45] This\nimplies that _empirical_ perception, which involves the actual\npresence of the object, involves no difficulty; in other words, it is\nimplied that empirical perception is of objects as they are. And we\nfind Kant admitting this to the extent of allowing _for the sake of\nargument_ that the perception of a present thing can make us know the\nthing as it is in itself.[46] But if empirical perception gives us\nthings as they are, and if, as is the case, and as Kant really\npresupposes, the objects of empirical perception are spatial, then,\nsince space is their form, the judgements of geometry must relate to\nthings as they are. It is true that on this view Kant's first\npresupposition of geometrical judgements has to be stated by saying\nthat we are able to perceive a real characteristic of things in space,\nbefore we perceive the things; and, no doubt, Kant thinks this\nimpossible. According to him, when we perceive empty space no object\nis present, and therefore what is before the mind must be merely\nmental. But no greater difficulty is involved than that involved in\nthe corresponding supposition required by Kant's own view. It is\nreally just as difficult to hold that we can perceive a characteristic\nof things as they appear to us _before_ they appear, as to hold that\nwe can perceive a characteristic of them as they are in themselves\n_before_ we perceive them.\n\n    [45] Cf. _Prol._, Section 8.\n\n    [46] _Prol._, Sec. 9 (cf. p. 55).\n\nThe fact is that the real difficulty with which Kant is grappling in\nthe _Prolegomena_ arises, not from the supposition that spatial bodies\nare things in themselves, but from the supposed presupposition of\ngeometry that we must be able to perceive empty space before we\nperceive bodies in it. It is, of course, impossible to defend the\nperception of empty space, but _if_ it be maintained, the space\nperceived must be conceded to be not, as Kant thinks, something mental\nor subjective, but a real characteristic of things. For, as has been\npointed out, the paradox of pure perception is reached solely through\nthe consideration that, while in empirical perception we perceive\nobjects, in pure perception we do not, and since the objects of\nempirical perception are spatial, space must be a real characteristic\nof them.\n\nThe general result of the preceding criticism is that Kant's\nconclusion does not follow from the premises by which he supports it.\nIt should therefore be asked whether it is not possible to take\nadvantage of this hiatus by presenting the argument for the merely\nphenomenal character of space without any appeal to the possibility of\nperceiving empty space. For it is clear that what was primarily before\nKant, in writing the _Critique_, was the _a priori_ character of\ngeometrical judgements themselves, and not the existence of a\nperception of empty space which they were held to presuppose.[47]\n\n    [47] The difficulty with which Kant is struggling in the\n    _Prolegomena_, Secs. 6-11, can be stated from a rather different\n    point of view by saying that the thought that geometrical\n    judgements imply a perception of empty space led him to apply\n    the term '_a priori_' to perception as well as to judgement.\n    The term, _a priori_, applied to judgements has a valid\n    meaning; it means, not that the judgement is made prior to\n    all experience, but that it is not based upon experience,\n    being originated by the mind in virtue of its own powers of\n    thinking. Applied to perception, however, '_a priori_' must\n    mean prior to all experience, and, since the object of\n    perception is essentially individual (cf. B. 741, M. 435),\n    this use of the term gives rise to the impossible task of\n    explaining how a perception can take place prior to the\n    actual experience of an individual in perception (cf.\n    _Prol._, Sec. 8).\n\nIf, then, the conclusion that space is only the form of sensibility\ncan be connected with the _a priori_ character of geometrical\njudgements without presupposing the existence of a perception of empty\nspace, his position will be rendered more plausible.\n\nThis can be done as follows. The essential characteristic of a\ngeometrical judgement is not that it takes place prior to experience,\nbut that it is not based upon experience. Thus a judgement, arrived at\nby an activity of the mind in which it remains within itself and does\nnot appeal to actual experience of the objects to which the judgement\nrelates, is implied to hold good of those objects. If the objects were\nthings as they are in themselves, the validity of the judgement could\nnot be justified, for it would involve the gratuitous assumption that\na necessity of thought is binding on things which _ex hypothesi_ are\nindependent of the nature of the mind. If, however, the objects in\nquestion are things as perceived, they will be through and through\nconditioned by the mind's perceiving nature; and, consequently, if a\ngeometrical rule, e. g. that a three-sided figure must have three\nangles, is really a law of the mind's perceiving nature, all\nindividual perceptions, i. e. all objects as perceived by us, will\nnecessarily conform to the law. Therefore, in the latter case, and in\nthat only, will the universal validity of geometrical judgements be\njustified. Since, then, geometrical judgements are universally valid,\nspace, which is that of which geometrical laws are the laws, must be\nmerely a form of perception or a characteristic of objects as\nperceived by us.\n\nThis appears to be the best form in which the substance of Kant's\nargument, stripped of unessentials, can be stated. It will be\nnecessary to consider both the argument and its conclusion.\n\nThe argument, so stated, is undeniably plausible. Nevertheless,\nexamination of it reveals two fatal defects. In the first place, its\nstarting-point is false. To Kant the paradox of geometrical judgements\nlies in the fact that they are not based upon an appeal to experience\nof the things to which they relate. It is implied, therefore, that\njudgements which are based on experience involve no paradox, and for\nthe reason that in experience we apprehend things as they are.[48] In\ncontrast with this, it is implied that in geometrical judgements the\nconnexion which we apprehend is not real, i. e. does not relate to\nthings as they are. Otherwise, there would be no difficulty; if in\ngeometry we apprehended rules of connexion relating to things as they\nare, we could allow without difficulty that the things must conform to\nthem. No such distinction, however, can be drawn between _a priori_\nand empirical judgements. For the necessity of connexion, e. g.\nbetween being a three-sided figure and being a three-angled figure, is\nas much a characteristic of things as the empirically-observed shape\nof an individual body, e. g. a table. Geometrical judgements,\ntherefore, cannot be distinguished from empirical judgements on the\nground that in the former the mind remains within itself, and does not\nimmediately apprehend fact or a real characteristic of reality.[49]\nMoreover, since in a geometrical judgement we do in fact think that we\nare apprehending a real connexion, i. e. a connexion which applies to\nthings and to things as they are in themselves, to question the\nreality of the connexion is to question the validity of thinking\naltogether, and to do this is implicitly to question the validity of\nour thought about the nature of our own mind, as well as the validity\nof our thought about things independent of the mind. Yet Kant's\nargument, in the form in which it has just been stated, presupposes\nthat our thought is valid at any rate when it is concerned with our\nperceptions of things, even if it is not valid when concerned with the\nthings as they are in themselves.\n\n    [48] Cf. p. 17.\n\n    [49] For the reasons which led Kant to draw this distinction\n    between empirical and _a priori_ judgements, cf. pp. 21-2.\n\nThis consideration leads to the second criticism. The supposition that\nspace is only a form of perception, even if it be true, _in no way\nassists_ the explanation of the universal validity of geometrical\njudgements. Kant's argument really confuses a _necessity_ of relation\nwith the _consciousness of a necessity_ of relation. No doubt, if it\nbe a law of our perceiving nature that, whenever we perceive an object\nas a three-sided figure, the object as perceived contains three\nangles, it follows that any object as perceived will conform to this\nlaw; just as if it be a law of things as they are in themselves that\nthree-sided figures contain three angles, all three-sided figures will\nin themselves have three angles. But what has to be explained is the\nuniversal applicability, not of a law, but of a judgement about a law.\nFor Kant's real problem is to explain why _our judgement_ that a\nthree-sided figure must contain three angles must apply to all\nthree-sided figures. Of course, if it be granted that in the judgement\nwe apprehend the true law, the problem may be regarded as solved. But\nhow are we to know that what we judge _is_ the true law? The answer is\nin no way facilitated by the supposition that the judgement relates to\nour perceiving nature. It can just as well be urged that what we think\nto be a necessity of our perceiving nature is not a necessity of it,\nas that what we think to be a necessity of things as they are in\nthemselves is not a necessity of them. The best, or rather the only\npossible, answer is simply that that of which we apprehend the\nnecessity must be true, or, in other words, that we _must_ accept the\nvalidity of thought. Hence nothing is gained by the supposition that\nspace is a form of sensibility. If what we judge to be necessary is,\nas such, valid, a judgement relating to things in themselves will be\nas valid as a judgement relating to our perceiving nature.[50]\n\n    [50] The same criticism can be urged against Kant's appeal to\n    the necessity of _constructing_ geometrical figures. The\n    conclusion drawn from the necessity of construction is stated\n    thus: \"If the object (the triangle) were something in itself\n    without relation to you the subject, how could you say that\n    that which lies necessarily in your subjective conditions of\n    constructing a triangle must also necessarily belong to the\n    triangle in itself?\" (B. 65, M. 39). Kant's thought is that\n    the laws of the mind's constructing nature must apply to\n    objects, if, and only if, the objects are the mind's own\n    construction. Hence it is open to the above criticism if, in\n    the criticism, 'construct' be substituted for 'perceive'.\n\nThis difficulty is concealed from Kant by his insistence on the\n_perception_ of space involved in geometrical judgements. This leads\nhim at times to identify the judgement and the perception, and,\ntherefore, to speak of the judgement as a perception. Thus we find him\nsaying that mathematical judgements are always _perceptive_,[51] and\nthat \"It is only possible for my perception to precede the actuality\nof the object and take place as _a priori_ knowledge, if &c.\"[52]\nHence, if, in addition, a geometrical judgement, as being a judgement\nabout a necessity, be identified with a necessity of judging, the\nconformity of things to these universal judgements will become the\nconformity of things to rules or necessities of our judging, i. e. of\nour perceiving nature, and Kant's conclusion will at once follow.[53]\nUnfortunately for Kant, a geometrical judgement, however closely\nrelated to a perception, must itself, as the apprehension of what is\nnecessary and universal, be an act of thought rather than of\nperception, and therefore the original problem of the conformity of\nthings to our mind can be forced upon him again, even after he thinks\nthat he has solved it, in the new form of that of the conformity\nwithin the mind of perceiving to thinking.\n\n    [51] _Prol._, Sec. 7.\n\n    [52] _Prol._, Sec. 9.\n\n    [53] Cf. (_Introduction_, B. xvii, M. xxix): \"But if the\n    object (as object of the senses) conforms to the nature of\n    our faculty of perception, I can quite well represent to\n    myself the possibility of _a priori_ knowledge of it [i. e.\n    mathematical knowledge].\"\n\nThe fact is simply that the universal validity of geometrical\njudgements can in no way be 'explained'. It is not in the least\nexplained or made easier to accept by the supposition that objects\nare 'phenomena'. These judgements must be accepted as being what we\npresuppose them to be in making them, viz. the direct apprehension of\nnecessities of relation between real characteristics of real things.\nTo explain them by reference to the phenomenal character of what is\nknown is really--though contrary to Kant's intention--to throw doubt\nupon their validity; otherwise, they would not need explanation. As a\nmatter of fact, it is _impossible_ to question their validity. In the\nact of judging, doubt is impossible. Doubt can arise only when\nwe subsequently reflect and temporarily lose our hold upon the\nconsciousness of necessity in judging.[54] The doubt, however, since\nit is non-existent in our geometrical consciousness, is really\ngroundless,[55] and, therefore, the problem to which it gives rise is\nunreal. Moreover if, _per impossibile_, doubt could be raised, it\ncould not be set at rest. No vindication of a judgement in which we\nare conscious of a necessity could do more than take the problem a\nstage further back, by basing it upon some other consciousness of a\nnecessity; and since this latter judgement could be questioned for\nprecisely the same reason, we should only be embarking upon an\ninfinite process.\n\n    [54] Cf. Descartes, _Princ. Phil._ i. Sec. 13, and _Medit._ v\n    sub fin.\n\n    [55] The view that kinds of space other than that with which\n    we are acquainted are possible, though usually held and\n    discussed by mathematicians, belongs to them _qua_\n    metaphysicians, and not _qua_ mathematicians.\n\nWe may now consider Kant's conclusion in abstraction from the\narguments by which he reaches it. It raises three main difficulties.\n\nIn the first place, it is not the conclusion to be expected from\nKant's own standpoint. The phenomenal character of space is inferred,\nnot from the fact that we make judgements at all, but from the fact\nthat we make judgements of a particular kind, viz. _a priori_\njudgements. From this point of view empirical judgements present no\ndifficulty. It should, therefore, be expected that the qualities which\nwe attribute to things in empirical judgements are not phenomenal, but\nbelong to things as they are. Kant himself implies this in drawing his\nconclusion concerning the nature of space. \"Space does not represent\nany quality of things in themselves or things in relation to one\nanother; that is, it does not represent any determination of things\nwhich would attach to the objects themselves and would remain, even\nthough we abstracted from all subjective conditions of perception. For\nneither absolute nor relative[56] determinations of objects can be\nperceived prior to the existence of the things to which they belong,\nand therefore not _a priori_.\"[57] It is, of course, implied that in\nexperience, where we do not discover determinations of objects prior\nto the existence of the objects, we do apprehend determinations of\nthings as they are in themselves, and not as they are in relation to\nus. Thus we should expect the conclusion to be, not that all that we\nknow is phenomenal--which is Kant's real position--but that spatial\n(and temporal) relations alone are phenomenal, i. e. that they alone\nare the result of a transmutation due to the nature of our perceiving\nfaculties.[58] This conclusion would, of course, be absurd, for what\nKant considers to be the empirically known qualities of objects\ndisappear, if the spatial character of objects is removed. Moreover,\nKant is prevented by his theory of perception from seeing that this is\nthe real solution of his problem, absurd though it may be. Since\nperception is held to arise through the origination of sensations by\nthings in themselves, empirical knowledge is naturally thought of as\nknowledge about sensations, and since sensations are palpably within\nthe mind, and are held to be due to things in themselves, knowledge\nabout sensations can be regarded as phenomenal.\n\n    [56] The first sentence shows that 'relative determinations'\n    means, not 'determinations of objects in relation to us', but\n    'determinations of objects in relation to one another.' Cf.\n    B. 37, M. 23; and B. 66 fin., 67 init., M. 40 (where these\n    meanings are confused).\n\n    [57] B. 42, M. 26.\n\n    [58] This conclusion is also to be expected because,\n    inconsistently with his real view, Kant is here (B. 41-2, M.\n    25-6) under the influence of the presupposition of our\n    ordinary consciousness that in perception we are confronted\n    by things in themselves, known to be spatial, and not by\n    appearances produced by unknown things in themselves. Cf. (B.\n    41, M. 25) \"and thereby of obtaining immediate representation\n    of them [i. e. objects];\" and (B. 42, M. 26) \"the receptivity\n    of the subject to be affected by objects necessarily precedes\n    all perceptions of these objects.\" These sentences identify\n    things in themselves and bodies in space, and thereby imply\n    that in empirical perception we perceive things in themselves\n    and as they _are_.\n\nOn the other hand, if we consider Kant's conclusion from the point of\nview, not of the problem which originates it, but of the distinction\nin terms of which he states it, viz. that between things as they are\nin themselves and things as perceived by us, we are led to expect the\ncontrary result. Since perception is the being affected by things, and\nsince the nature of the affection depends upon the nature of our\ncapacity of being affected, in _all_ perception the object will become\ndistorted or transformed, as it were, by our capacity of being\naffected. The conclusion, therefore, should be that in all judgements,\nempirical as well as _a priori_, we apprehend things only as\nperceived. The reason why Kant does not draw this conclusion is\nprobably that given above, viz. that by the time Kant reaches the\nsolution of his problem empirical knowledge has come to relate to\nsensation only; consequently, it has ceased to occur to him that\nempirical judgements could possibly give us knowledge of things as\nthey are. Nevertheless, Kant should not have retained in his\nformulation of the problem a distinction irreconcilable with his\nsolution of it; and if he had realized that he was doing so he might\nhave been compelled to modify his whole view.\n\nThe second difficulty is more serious. If the truth of geometrical\njudgements presupposes that space is only a property of objects as\nperceived by us, it is a paradox that geometricians should be\nconvinced, as they are, of the truth of their judgements. They\nundoubtedly think that their judgements apply to things as they are in\nthemselves, and not merely as they appear to us. They certainly do not\nthink that the relations which they discover apply to objects only as\nperceived. Not only, therefore, do they not think that bodies in space\nare phenomena, but they do not even leave it an open question whether\nbodies are phenomena or not. Hence, if Kant be right, they are really\nin a state of illusion, for on his view the true geometrical judgement\nshould include in itself the phenomenal character of spatial\nrelations; it should be illustrated by expressing Euclid I. 5 in\nthe form that the equality of the angles at the base of an isosceles\ntriangle belongs to objects as perceived. Kant himself lays this down.\n\"The proposition 'all objects are beside one another in space'\nis valid under[59] the limitation that these things are taken as\nobjects of our sensuous perception. If I join the condition to the\nperception, and say 'all things, as external phenomena, are beside\none another in space', the rule is valid universally, and without\nlimitation.\"[60] Kant, then, is in effect allowing that it is possible\nfor geometricians to make judgements, of the necessity of which\nthey are convinced, and yet to be wrong; and that, therefore, the\napprehension of the necessity of a judgement is no ground of its\ntruth. It follows that the truth of geometrical judgements can no\nlonger be accepted as a starting-point of discussion, and, therefore,\nas a ground for inferring the phenomenal character of space.\n\n    [59] A. reads 'only under'\n\n    [60] B. 43, M. 27.\n\nThere seems, indeed, one way of avoiding this consequence, viz. to\nsuppose that for Kant it was an absolute starting-point, which nothing\nwould have caused him to abandon, that only those judgements of which\nwe apprehend the necessity are true. It would, of course, follow that\ngeometricians would be unable to apprehend the necessity of\ngeometrical judgements, and therefore to make such judgements, until\nthey had discovered that things as spatial were only phenomena. It\nwould not be enough that they should think that the phenomenal or\nnon-phenomenal character of things as spatial must be left an open\nquestion for the theory of knowledge to decide. In this way the\nnecessity of admitting the illusory character of geometry would be\navoided. The remedy, however, is at least as bad as the disease.\nFor it would imply that geometry must be preceded by a theory of\nknowledge, which is palpably contrary to fact. Nor could Kant accept\nit; for he avowedly bases his theory of knowledge, i. e. his view\nthat objects as spatial are phenomena, upon the truth of geometry;\nthis procedure would be circular if the making of true geometrical\njudgements was allowed to require the prior adoption of his theory of\nknowledge.\n\nThe third difficulty is the most fundamental. Kant's conclusion (and\nalso, of course, his argument) presupposes the validity of the\ndistinction between phenomena and things in themselves. If, then, this\ndistinction should prove untenable in principle, Kant's conclusion\nwith regard to space must fail on general grounds, and it will even\nhave been unnecessary to consider his arguments for it. The importance\nof the issue, however, requires that it should be considered in a\nseparate chapter.\n\n    NOTE to page 47.\n\n    The argument is not affected by the contention that, while\n    the totality of spaces is infinite, the totality of colours\n    or, at any rate, the totality of instances of some other\n    characteristic of objects is finite; for this difference\n    will involve no difference in respect of perception and\n    conception. In both cases the apprehension that there is a\n    totality will be reached in the same way, i. e. through the\n    _conception_ of the characteristic in general, and the\n    apprehension in the one case that the totality is infinite\n    and in the other that it is finite will depend on the\n    apprehension of the special nature of the characteristic in\n    question.\n\n\n\n\nCHAPTER IV\n\nPHENOMENA AND THINGS IN THEMSELVES\n\n\nThe distinction between phenomena and things in themselves can be best\napproached by considering Kant's formulation of the alternative views\nof the nature of space and time. \"What are space and time? Are they\nreal existences? Or are they merely determinations or relations of\nthings, such, however, as would also belong to them in themselves,\neven if they were not perceived, or are they attached to the form of\nperception only, and consequently to the subjective nature of our\nmind, without which these predicates can never be attributed to any\nthing?\"[1]\n\n    [1] B. 37, M. 23.\n\nOf these three alternatives, the first can be ignored. It is opposed\nto the second, and is the view that space and time are things rather\nthan relations between things. This opposition falls within the first\nmember of the wider opposition between things as they are in\nthemselves and things as they are as perceived, and Kant, and indeed\nany one, would allow that if space and time belong to things as they\nare in themselves and not to things only as perceived, they are\nrelations between things rather than things. The real issue,\ntherefore, lies between the second and third alternatives. Are space\nand time relations between things which belong to them both in\nthemselves and also as perceived by us, or are they relations which\nbelong to things only as perceived?\n\nTo this question we may at once reply that, inasmuch as it involves an\nimpossible antithesis, it is wholly unreal. The thought of a property\nor a relation which belongs to things as perceived involves a\ncontradiction. To take Plato's example, suppose that we are looking at\na straight stick, partially immersed in water. If we have not\npreviously seen the stick, and are ignorant of the laws of refraction,\nwe say that the stick is bent. If, however, we learn the effect of\nrefraction, and observe the stick from several positions, we alter our\nassertion. We say that the stick is not really bent, but only looks or\nappears bent to us. But, if we reflect at all, we do not express our\nmeaning by saying that the stick _is_ bent to us as perceiving, though\nnot in reality.[2] The word 'is' essentially relates to what really\nis. If, therefore, the phrase 'to us as perceiving' involves an\nopposition to the phrase 'in reality', as it must if it is to be a\nreal qualification of 'is', it cannot rightly be added to the word\n'is'. To put the matter more explicitly, the assertion that something\n_is_ so and so implies that it is so and so in itself, whether it be\nperceived or not, and therefore the assertion that something is so and\nso to us as perceiving, though not in itself, is a contradiction in\nterms. The phrase 'to us as perceiving', as a restriction upon the\nword 'is', merely takes back the precise meaning of the word 'is'.\nThat to which the phrase can be added is not the word 'is', but the\nword 'looks' or 'appears'. We can rightly say that the stick looks or\nappears bent to us as perceiving. But even then the addition only\nhelps to make explicit the essential meaning of 'appears', for\n'appears' really means 'appears to us', and 'as perceiving' only\nrepeats the meaning of 'appears' from the side of the perceiving\nsubject as opposed to that of the object perceived. The essential\npoint, however, is thereby brought out that the phrase 'to us as\nperceiving' essentially relates not to what a thing is, but to what it\nlooks or appears to us.\n\n    [2] Similarly, we do not say--if we mean what we say--of a\n    man who is colour blind that an object which others call blue\n    _is_ pink to him or to his perception, but that it _looks_\n    pink to him.\n\nWhat, then, is the proper statement of Kant's view that space is a\ndetermination of things only as they appear to us, and not as they are\nin themselves? It should be said that things are not in reality\nspatial, but only look or appear spatial to us. It should not be said\nthat they _are_ spatial for our perception, though not in themselves.\nThus the view properly stated implies that space is an illusion,\ninasmuch as it is not a real property of things at all. This\nimplication, however, is precisely the conclusion which Kant wishes to\navoid. He takes infinite trouble to explain that he does not hold\nspace and time to be illusions.[3] Though _transcendentally ideal_\n(i. e. though they do not belong to things in themselves), they are\n_empirically real_. In other words, space and time are real relations\nof _something_, though not of things in themselves.\n\n    [3] B. 44, 52, 53-4, 62-3, 69-70; M. 27, 31-2, 37-8, 41-2;\n    _Prol._, Sec. 13, Remark iii.\n\nHow, then, does Kant obtain something of which space and time can be\nregarded as really relations? He reaches it by a transition which at\nfirst sight seems harmless. In stating the fact of perception he\nsubstitutes for the assertion that things appear so and so to us the\nassertion that things produce appearances in us. In this way, instead\nof an assertion which relates to the thing and states what it is not\nbut only appears, he obtains an assertion which introduces a second\nreality distinct from the thing, viz. an appearance or phenomenon, and\nthereby he gains something other than the thing to which space can be\nattached as a real predicate. He thus gains something in respect of\nwhich, with regard to spatial relations we can be said to have\n_knowledge_ and not illusion. For the position now is that space,\nthough not a property of things in themselves, _is_ a property of\nphenomena or appearances; in other words, that while things in\nthemselves are not spatial, phenomena and appearances _are_ spatial.\nAs evidence of this transition, it is enough to point out that, while\nhe states the _problem_ in the form 'Are things in themselves spatial\nor are they only spatial as appearing to us?'[4] he usually states the\n_conclusion_ in the form 'Space is the form of phenomena', i. e.\nphenomena are spatial. A transition is thereby implied from 'things as\nappearing' to 'appearances'. At the same time, it is clear that Kant\nis not aware of the transition, but considers the expressions\nequivalent, or, in other words, fails to distinguish them. For both\nmodes of stating the conclusion are to be found even in the same\nsentence. \"This predicate [space] is applied to things only in so far\nas they appear to us, i. e. are objects of sensibility [i. e.\nphenomena].\"[5] Again, the common phrase 'things as phenomena' implies\nthe same confusion. Moreover, if Kant had realized that the transition\nwas more than one of phraseology he must have seen that it was\nnecessary to recast his argument.\n\n    [4] This is Kant's way of putting the question which should\n    be expressed by asking, 'Are things spatial, or do they only\n    look spatial?'\n\n    [5] B. 43, M. 26. Cf. _Prol._, Sec. 9 fin. with Sec. 10 init.\n\nIt may be said, then, that Kant is compelled to end with a different\ndistinction from that with which he begins. He begins with the\ndistinction between things as they are in themselves and things as\nthey appear to us, the distinction relating to one and the same\nreality regarded from two different points of view. He ends with the\ndistinction between two different realities, things-in-themselves,[6]\nexternal to, in the sense of independent of, the mind, and phenomena\nor appearances within it. Yet if his _argument_ is to be valid, the\ntwo distinctions should be identical, for it is the first distinction\nto which the argument appeals.[7] In fact, we find him expressing what\nis to him the same distinction now in the one way and now in the other\nas the context requires.\n\n    [6] It should be noticed that 'things-in-themselves' and\n    'things as they are in themselves' have a different meaning.\n\n    [7] Cf. p. 55 and ff.\n\nThe final form of Kant's conclusion, then, is that while things in\nthemselves are not, or, at least, cannot be known to be spatial,\n'phenomena,' or the appearances produced in us by things in\nthemselves, are spatial. Unfortunately, the conclusion in this form is\nno more successful than it is in the former form, that things are\nspatial only as perceived. Expressed by the formula 'phenomena are\nspatial', it has, no doubt, a certain plausibility; for the word\n'phenomena' to some extent conceals the essentially mental character\nof what is asserted to be spatial. But the plausibility disappears on\nthe substitution of 'appearances'--the true equivalent of Kant's\n_Erscheinungen_--for 'phenomena'. Just as it is absurd to describe the\nfact that the stick only looks bent by saying that, while the stick is\nnot bent, the appearance which it produces is bent, so it is, even on\nthe face of it, nonsense to say that while things are not spatial, the\nappearances which they produce in us are spatial. For an 'appearance',\nbeing necessarily something mental, cannot possibly be said to be\nextended. Moreover, it is really an abuse of the term 'appearance' to\nspeak of appearances _produced by_ things, for this phrase implies a\nfalse severance of the appearance from the things which appear. If\nthere are 'appearances' at all, they are appearances _of_ things and\nnot appearances _produced by_ them. The importance of the distinction\nlies in the difference of implication. To speak of appearances\nproduced by things is to imply that the object of perception is merely\nsomething mental, viz. an appearance. Consequently, access to a\nnon-mental reality is excluded; for a perception of which the object\nis something belonging to the mind's own being cannot justify an\ninference to something beyond the mind, and the result is inevitably\nsolipsism. On the other hand, the phrase 'appearances of things',\nwhatever defects it may have, at least implies that it is a non-mental\nreality which appears, and therefore that in perception we are in\ndirect relation to it; the phrase, therefore, does not imply from the\nvery beginning that the apprehension of a non-mental reality is\nimpossible.\n\nThe objection will probably be raised that this criticism is much too\nsummary. We do, it will be said, distinguish in ordinary consciousness\nbetween appearance and reality. Consequently there must be some form\nin which Kant's distinction between things in themselves and phenomena\nand the conclusion based upon it are justified. Moreover, Kant's\nreiterated assertion that his view does not imply that space is an\nillusion, and that the distinction between the real and the illusory\nis possible _within_ phenomena, requires us to consider more closely\nwhether Kant may not after all be entitled to hold that space is not\nan illusion.[8]\n\n    [8] Cf. p. 93 and ff.\n\nThis objection is, of course, reasonable. No one can satisfy himself\nof the justice of the above criticisms until he has considered the\nreal nature of the distinction between appearance and reality. This\ndistinction must, therefore, be analysed. But before this is done it\nis necessary, in order to discover the real issue, to formulate the\nlines on which Kant may be defended. 'The reality,' it may be urged,\n'which ideally we wish to know must be admitted to exist _in itself_,\nin the sense of independently of the perception, and consequently its\nnature must be admitted to be independent of perception. Ideally,\nthen, our desire is to know things[9] as they are in themselves, a\ndesire sufficiently expressed by the assertion that we desire to know\nthings, for to know them is to know them as they are, i. e. as they\nare independently of perception. Again, since the reality which we\ndesire to know consists of individuals, and since the apprehension of\nan individual implies perception, knowledge of reality requires\nperception. If in perception we apprehended reality as it is, no\ndifficulty would arise. But we do not, for we are compelled to\ndistinguish what things are, and what they look or appear; and what\nthey appear essentially relates to perception. We perceive them as\nthey look or appear and, therefore, not as they are, for what they\nlook and what they are are _ex hypothesi_ distinguished. And this fact\nconstitutes a fatal obstacle to knowledge in general. We cannot know\nanything as it _is_. At least the negative side of Kant's position\nmust be justified. We never can know things as they are in themselves.\nWhat then do we know? Two alternative answers may be given. It may be\nheld that the positive side of Kant's position, though indefensible in\nthe form that we know things as they appear to us, is valid in the\nform that we know what things look or appear. This, no doubt, implies\nthat our ordinary beliefs about reality are illusory, for what things\nlook is _ex hypothesi_ different from what they are. But the\nimplication does not constitute an important departure from Kant's\nview. For in any case only that is knowledge proper which relates to\nthings as they are, and therefore the supposed knowledge of things as\nthey appear may be discarded without serious loss. On the other hand,\nit may be held that the positive side of Kant's position can be\nvindicated in the form that, while we do not know things in\nthemselves,[10] we do know the appearances which they produce in us.\nIt is true that this view involves the difficulty of maintaining that\nappearances are spatial, but the difficulty is not insuperable.\nMoreover, in this form the doctrine has the advantage that, unlike the\nformer, it does not imply that the knowledge which we have is only of\nillusions, for instead of implying that our knowledge is merely\nknowledge of what things look but really are not, it implies that we\nknow the real nature of realities of another kind, viz. of\nappearances. Again, in this form of the view, it may be possible to\nvindicate Kant's doctrine that the distinction between the real and\nthe illusory is tenable within what we know, for it may be possible to\ndistinguish within appearances between a 'real' appearance[11] and an\n'illusory' appearance.[12]'\n\n    [9] 'Things' is substituted for 'the reality which we believe\n    to exist independently of perception' in order to conform\n    to Kant's language. The substitution, of course, has the\n    implication--which Kant took for granted--that the reality\n    consists of a plurality of individuals.\n\n    [10] 'Things in themselves' has here to be substituted for\n    'things as they are in themselves' in the statement of the\n    negative side of the position, in order to express the proper\n    antithesis, which is now that between two things, the one\n    known and the other unknown, and not that between two points\n    of view from which one and the same thing is known and not\n    known respectively.\n\n    [11] _Erscheinung._\n\n    [12] _Schein._\n\nAn implication of this defence should be noticed. The issue relates\nto the nature of space[13], and may be stated in terms of it. For,\nsince space is a presupposition of all other properties which the\nnon-philosophical consciousness attributes to physical things, it\nmakes no difference whether we say that things _only appear_ heavy,\nhard, in motion, &c., or whether we say that things _only appear_\nspatial. In the same way it is a matter of indifference whether we say\nthat, though things are not heavy, hard, &c., their appearances are\nso, or whether we say that, though things are not spatial, their\nappearances are so. The issue, then, concerns the possibility of\nmaintaining either that things only appear spatial, or that the\nappearances which they produce are spatial, while the things\nthemselves are not, or, at least cannot be known to be, spatial.\n\n    [13] We might add time also; but, for a reason which will\n    appear later (p. 139), it can be neglected.\n\nThe tenability of these alternative positions has to be considered\napart from the argument of the _Aesthetic_, for this, as we have seen,\nbreaks down. At the outset it is important to realize that these\npositions are the product of philosophical reflection, and constitute\ngeneral theories of knowledge. As has been pointed out, the\ndistinction between appearance and reality first arises in our\nordinary or scientific consciousness.[14] In this consciousness we are\ncompelled to distinguish between appearance and reality with respect\nto the details of a reality which, as a whole, or, in principle, we\nsuppose ourselves to know. Afterwards in our philosophical\nconsciousness we come to reflect upon this distinction and to raise\nthe question whether it is not applicable to reality as a whole. We\nask with respect to knowledge in general, and not merely with respect\nto certain particular items of knowledge, whether we know or can know\nreality, and not merely appearance. The two positions just stated are\nalternative ways of answering the question in the negative. They are,\nthen, philosophical views based upon a distinction found in our\nordinary consciousness. Consequently, in order to decide whether the\ndistinction will bear the superstructure placed upon it by the\nphilosophical consciousness, it is necessary to examine the\ndistinction as it exists in our ordinary consciousness.\n\n    [14] I. e. the consciousness for which the problems are those\n    of science as opposed to philosophy.\n\nThe distinction is applied in our ordinary consciousness both to the\nprimary and to the secondary qualities of matter, i. e. to the size,\nshape, position and motion of physical bodies, and to their colour,\nwarmth, &c. We say, for instance, that the moon looks[15] or appears\nas large as the sun, though really it is much smaller. We say that\nrailway lines, though parallel, look convergent, just as we say that\nthe straight stick in water looks bent. We say that at sunset the sun,\nthough really below the horizon, looks above it. Again, we say that to\na person who is colour blind the colour of an object looks different\nto what it really is, and that the water into which we put our hand\nmay be warmer than it appears to our touch.\n\n    [15] 'Looks' means 'appears to sight', and 'looks' is\n    throughout used as synonymous with 'appear', where the\n    instance under discussion relates to visual perception.\n\nThe case of the primary qualities may be considered first. Since the\ninstances are identical in principle, and only differ in complexity,\nit will be sufficient to analyse the simplest, that of the apparent\nconvergence of the railway lines.\n\nTwo points at once force themselves upon our notice. In the first\nplace, we certainly suppose that we perceive the reality which we wish\nto know, i. e. the reality which, as we suppose, exists independently\nof our perception, and not an 'appearance' of it. It is, as we say,\nthe real lines which we see. Even the term 'convergent', in the\nassertion that the lines look convergent, conveys this implication.\nFor 'convergent' is essentially a characteristic not of an appearance\nbut of a reality, in the sense in which something independent of\nperception may be opposed as a reality to an 'appearance', which, as\nsuch, presupposes perception. We can say neither that an appearance is\nconvergent, nor that the appearance of the lines is convergent. Only a\nreality similar to the lines, e. g. two roads, can be said to be\nconvergent. Our ordinary thought, therefore, furnishes no ground for\nthe view that the object of perception is not the thing, but merely an\nappearance of or produced by it. In the second place, the assertion\nthat the lines _look_ convergent implies considerable knowledge of the\nreal nature of the reality to which the assertion relates. Both the\nterms 'lines' and 'convergent' imply that the reality _is_ spatial.\nFurther, if the context is such that we mean that, while the lines\nlook convergent, we do not know their real relation, we imply that\nthe lines really possess some characteristic which falls within the\ngenus to which convergence belongs, i. e. we imply that they are\nconvergent, divergent, or parallel. If, on the other hand, the context\nis such that we mean that the lines only look convergent, we imply\nthat the lines are parallel, and therefore presuppose complete\nknowledge in respect of the very characteristic in regard to which we\nstate what is only appearance. The assertion, then, in respect of a\nprimary quality, that a thing looks so and so implies knowledge of its\ngeneral character as spatial, and ignorance only of a detail; and the\nassertion that a thing only looks or appears so and so implies\nknowledge of the detail in question.\n\nAttention may now be drawn to a general difficulty which may be raised\nwith respect to the use of the terms 'looks' and 'appears'. It may be\nstated thus: 'If the lines are not convergent, how is it possible even\nto say that they _look_ convergent? Must it not be implied that at\nleast under _certain_ circumstances we should perceive the lines as\nthey are? Otherwise, why should we use the words 'look' or 'appear' at\nall? Moreover, this implication can be pushed further; for if we\nmaintain that we perceive the real lines, we may reasonably be asked\nwhether we must not under _all_ circumstances perceive them as they\nare. It seems as though a reality cannot be perceived except as it\nis.' It is the view to which this difficulty gives rise which is\nmainly responsible for the doctrine that the object of perception is\nnot the reality, but an appearance. Since we do distinguish between\nwhat things look and what they are, it would seem that the object of\nperception cannot be the thing, but only an appearance produced by\nit. Moreover, the doctrine gains in plausibility from the existence of\ncertain illusions in the case of which the reality to which the\nillusion relates seems non-existent. For instance, if we look steadily\nat the flame of a candle, and then press one eyeball with a finger, we\nsee, as we say, two candles;[16] but since _ex hypothesi_ there is\nonly one candle, it seems that what we see must be, not the candle,\nbut two images or appearances produced by it.\n\n    [16] Cf. Dr. Stout, on 'Things and Sensations' (_Proceedings\n    of the British Academy_, vol. ii).\n\nThis difficulty is raised in order to draw attention to the fact that,\nin the case of the railway lines, where it can be met on its own\nground[17], this is because, and only because, we believe space to be\n'real', i. e. to be a characteristic of reality, and because we\nunderstand its nature. The distinction between the actual and the\napparent angle made by two straight lines presupposes a limiting case\nin which they coincide. If the line of sight along which we observe\nthe point of intersection of two lines is known to be at right angles\nto both lines, we expect, and rightly expect, to see the angle of\nintersection as it is. Again, if we look at a short portion of two\nrailway lines from a point known to be directly above them, and so\ndistant that the effects of perspective are imperceptible, we can say\nthat the lines look what they are, viz. parallel. Thus, from the point\nof view of the difficulty which has been raised, there is this\njustification in general for saying that two lines _look_ parallel or\n_look_ at right angles, that we know that in certain cases what they\nlook is identical with what they are. In the same way, assertions of\nthe type that the moon _looks_ as large as the sun receive\njustification from our knowledge that two bodies of equal size and\nequally distant from the observer _are_ what they look, viz. of the\nsame size. And in both cases the justification presupposes knowledge\nof the reality of space and also such insight into its nature as\nenables us to see that in certain cases there must be an identity\nbetween what things look and what they are in respect of certain\nspatial relations. Again, in such cases we see that so far is it from\nbeing necessary to think that a thing must be perceived as it is, that\nit is not only possible but necessary to distinguish what a thing\nlooks from what it is, and precisely in consequence of the nature of\nspace. The visual perception of spatial relations from its very nature\npresupposes a particular point of view. Though the perception itself\ncannot be spatial, it presupposes a particular point in space as a\nstandpoint or point of view,[18] and is therefore subject to\nconditions of perspective. This is best realized by considering the\nsupposition that perfect visual powers would enable us to see the\nwhole of a body at once, and that this perception would be possible if\nwe had eyes situated all round the body. The supposition obviously\nbreaks down through the impossibility of combining two or more points\nof view in one perception. But if visual perception is necessarily\nsubject to conditions of perspective, the spatial relations of bodies\ncan never look what they are except in the limiting case referred to.\nMoreover, this distinction is perfectly intelligible, as we should\nexpect from the necessity which we are under of drawing it. We\nunderstand perfectly why it is that bodies must, in respect of their\nspatial relations, look different to what they are, and we do so\nsolely because we understand the nature of space, and therefore also\nthe conditions of perspective involved in the perception of what is\nspatial. It is, therefore, needless to make the assertion 'Two lines\nappear convergent' intelligible by converting the verb 'appears' into\na substantive, viz. an 'appearance', and then making the assertion\nrelate to an 'appearance'. For--apart from the fact that this would\nnot achieve the desired end, since no suitable predicate could be\nfound for the appearance--the assertion that the lines _look_ or\n_appear_ convergent is perfectly intelligible in itself, though not\ncapable of being stated in terms of anything else.[19] If we\ngeneralize this result, we may say that the distinction between\nappearance and reality, drawn with regard to the primary qualities of\nbodies, throughout presupposes the reality of space, and is made\npossible, and indeed necessary, by the nature of space itself.\n\n    [17] Cf., however, p. 87 and pp. 89-91.\n\n    [18] This is, of course, not refuted by the reminder that we\n    see with two eyes, and that these are in different places.\n\n    [19] It is important to notice that the proper formula to\n    express what is loosely called 'an appearance' is 'A looks or\n    appears B', and that this cannot be analysed into anything\n    more simple and, in particular, into a statement about\n    'appearances'. Even in the case of looking at the candle,\n    there is no need to speak of two 'appearances' or 'images'.\n    Before we discover the truth, the proper assertion is 'The\n    body which we perceive looks as if it were two candles', and,\n    after we discover the truth, the proper assertion is 'The\n    candle looks as if it were in two places'.\n\nWe may now turn to the way in which we draw the distinction with\nrespect to the secondary qualities of physical things. It must, it\nseems, be admitted that in our ordinary consciousness we treat these\nqualities as real qualities of bodies. We say that a bell is noisy;\nthat sugar is sweet; that roses smell; that a mustard plaster is hot;\nthat the sky is blue. It must also be admitted that in our ordinary\nconsciousness we draw a distinction between appearance and reality\n_within_ these qualities, just as we do _within_ the primary\nqualities. Just as we speak of the right or real shape of a body, so\nwe speak of its right or real colour, taste, &c., and distinguish\nthese from its apparent colours, taste, &c., to some individual. We\nthereby imply that these qualities are real qualities of bodies, and\nthat the only difficulty is to determine the particular character of\nthe quality in a given case. Yet, as the history of philosophy shows,\nit takes but little reflection to throw doubt on the reality of these\nqualities. The doubt arises not merely from the apparent impossibility\nof finding a principle by which to determine the right or real quality\nin a given case, but also and mainly from misgivings as to the\npossible reality of heat, smell, taste, noise, and colour apart from a\npercipient. It must also be admitted that this misgiving is well\nfounded; in other words, that these supposed real qualities do\npresuppose a percipient, and therefore cannot be qualities of things,\nsince the qualities of a thing must exist independently of the\nperception of the thing.[20] This will readily be allowed in the case\nof all the secondary qualities except colour. No one, it may\nreasonably be said, who is familiar with and really faces the issue,\nwill maintain that sounds, smells, tastes, and sensations of touch\nexist apart from a sensitive subject. So much is this the case, that\nwhen once the issue is raised, it is difficult and, in the end,\nimpossible to use the word 'appear' in connexion with these qualities.\nThus it is difficult and, in the end, impossible to say that a bell\n_appears_ noisy, or that sugar _appears_ sweet. We say, rather, that\nthe bell and the sugar produce certain sensations[21] in us.\n\n    [20] Cf. pp. 72-3, and 91.\n\n    [21] _Not_ 'appearances'.\n\nThe case of colour, however, is more difficult. From the closeness of\nits relation to the shape of bodies, it seems to be a real quality of\nbodies, and not something relative to a sensitive subject like the\nother secondary qualities. In fact, so intimate seems the relation of\ncolour to the shape of bodies, that it would seem--as has, of course,\noften been argued--that if colour be relative to a sensitive subject,\nthe primary qualities of bodies must also be relative to a sensitive\nsubject, on the ground that shape is inseparable from colour.[22] Yet\nwhether this be so or not, it must, in the end, be allowed that colour\ndoes presuppose a sensitive subject in virtue of its own nature, and\nquite apart from the difficulty--which is in itself insuperable--of\ndetermining the right colour of individual bodies. It must, therefore,\nbe conceded that colour is not a quality of bodies. But if this be\ntrue, the use of the term 'look' or 'appear' in connexion with colour\ninvolves a difficulty which does not arise when it is used in\nconnexion with the primary qualities. Bodies undoubtedly look or\nappear coloured. Now, as has already been suggested,[23] the term\n'look' seems to presuppose some identity between what a thing is and\nwhat it looks, and at least the possibility of cases in which they are\nwhat they look--a possibility which, as we have seen, is realized in\nthe case of the primary qualities. Yet, if colour is not a quality of\nbodies, then, with respect to colour, things look what they never are,\nor, in other words, are wholly different from what they look;[24] and\nsince it seems impossible to hold that colour is really a property of\nbodies, this conclusion must, in spite of its difficulty, be admitted\nto be true.\n\n    [22] Cf. p. 91 note.\n\n    [23] Cf. p. 82.\n\n    [24] It is assumed that there is not even plausibility in the\n    supposition of continuity or identity between colour proper\n    and its physical conditions in the way of light vibrations.\n\nThere remain, however, to be noticed two respects in which assertions\nconcerning what things look in respect of colour agree with\ncorresponding assertions in respect of the primary qualities. They\nimply that what we perceive is a reality, in the sense already\nexplained.[25] Thus the assertion that the grass looks green implies\nthat it is a reality which looks green, or, in other words, that the\nobject of perception is a reality, and not an 'appearance'. Again,\nsuch assertions imply that the reality about which the assertion is\nmade is spatial. The term 'grass' implies extension, and only what is\nextended can be said to look coloured. If it be urged that what looks\ncoloured need only _look_ extended, it may be replied that the two\nconsiderations which lead us to think that things only _look_ coloured\npresuppose that they _are_ spatial. For the two questions, the\nconsideration of which leads to this conclusion, are, 'What is the\nright or real colour of an individual thing?' and 'Has it really any\ncolour at all, or does it only look coloured?' and neither question is\nsignificant unless the thing to which it refers is understood to be\nspatial.\n\n    [25] I. e. in the sense of something which exists\n    independently of perception.\n\nWe may now return to the main issue. Is it possible to maintain either\n(1) the position that only appearances are spatial and possess all the\nqualities which imply space, or (2) the position that things only\nappear spatial and only appear or look as if they possessed the\nqualities which imply space? It may be urged that these questions have\nalready been implicitly answered in the negative. For the division of\nthe qualities of things into primary and secondary is exhaustive, and,\nas has been shown, the distinction between 'appearance' and 'reality',\nwhen drawn with respect to the primary qualities and to colour--the\nonly secondary quality with respect to which the term 'appears' can\nproperly be used[26]--presupposes the reality of space. Consequently,\nsince we do draw the distinction, we must accept the reality of that\nwhich is the condition of drawing it at all. But even though this be\nconceded--and the concession is inevitable--the problem cannot be\nregarded as solved until we have discovered what it is in the nature\nof space which makes both positions untenable. Moreover, the admission\nthat in the case of colour there is no identity between what things\nlook and what they are removes at a stroke much of the difficulty of\none position, viz. that we only know what things look or appear, and\nnot what they are. For the admission makes it impossible to maintain\nas a general principle that there must be some identity between what\nthey look and what they are. Consequently, it seems _possible_ that\nthings should be wholly different from what they appear, and, if so,\nthe issue cannot be decided on general grounds. What is in substance\nthe same point may be expressed differently by saying that just as\nthings only _look_ coloured, so things may only _look_ spatial. We are\nthus again[27] led to see that the issue really turns on the nature of\nspace and of spatial characteristics in particular.\n\n    [26] Cf. pp. 86-7.\n\n    [27] Cf. p. 79.\n\nIn discussing the distinction between the real and the apparent shape\nof bodies, it was argued that while the nature of space makes it\nnecessary to distinguish in general between what a body looks and what\nit is, yet the use of the term _look_ receives justification from the\nexistence of limiting cases in which what a thing looks and what it is\nare identical. The instances considered, however, related to qualities\ninvolving only two dimensions, e. g. convergence and bentness, and it\nwill be found that the existence of these limiting cases is due solely\nto this restriction. If the assertion under consideration involves a\nterm implying three dimensions, e. g. 'cubical' or 'cylindrical',\nthere are no such limiting cases. Since our visual perception is\nnecessarily subject to conditions of perspective, it follows that\nalthough we can and do see a cube, we can never see it as it _is_.\nIt _is_, so to say, in the way in which a child draws the side of a\nhouse, i. e. with the effect of perspective eliminated; but it never\ncan be seen in this way. No doubt, our unreflective knowledge of the\nnature of perspective enables us to allow for the effect of\nperspective, and to ascertain the real shape of a solid object from\nwhat it looks when seen from different points. In fact, the habit of\nallowing for the effect of perspective is so thoroughly ingrained in\nhuman beings that the child is not aware that he is making this\nallowance, but thinks that he draws the side of the house as he sees\nit. Nevertheless, it is true that we never see a cube as it is, and if\nwe say that a thing looks cubical, we ought only to mean that it looks\nprecisely what a thing looks which is a cube.\n\nIt is obvious, however, that two dimensions are only an abstraction\nfrom three, and that the spatial relations of bodies, considered\nfully, involve three dimensions; in other words, spatial\ncharacteristics are, properly speaking, three-dimensional. It follows\nthat terms which fully state spatial characteristics can never\nexpress what things look, but only what they are. A body may be\ncylindrical, and we may see a cylindrical body; but such a body can\nnever, strictly speaking, _look_ cylindrical. The opposition, however,\nbetween what a thing _is_ and what it _looks_ implies that what it\n_is_ is independent of a percipient, for it is precisely correlation\nto a percipient which is implied by 'looking' or 'appearing'. In fact,\nit is the view that what a thing really is it is, independently of\na percipient, that forms the real starting-point of Kant's thought.\nIt follows, then, that the spatial characteristics of things, and\ntherefore space itself, must belong to what they are in themselves\napart from a percipient, and not to what they look.[28] Consequently,\nit is so far from being true that we only know what things look and\nnot what they are, that in the case of spatial relations we actually\nknow what things are, even though they never look what they are.\n\n    [28] This consideration disposes of the view that, if colour\n    is relative to perception, the primary qualities, as being\n    inseparable from colour, must also be relative to perception;\n    for it implies that the primary qualities cannot from their\n    very nature be relative to perception. Moreover, if the\n    possibility of the separation of the primary qualities from\n    colour is still doubted, it is only necessary to appeal to\n    the blind man's ability to apprehend the primary qualities,\n    though he may not even know what the word 'colour' means. Of\n    course, it must be admitted that some sensuous elements are\n    involved in the apprehension of the primary qualities, but\n    the case of the blind man shows that these may relate to\n    sight instead of to touch. Moreover, it, of course, does not\n    follow from the fact that sensuous elements are inseparable\n    from our perception of bodies that they belong to, and are\n    therefore inseparable from, the bodies perceived.\n\nThis conclusion, however, seems to present a double difficulty. It is\nadmitted that we perceive things as they look, and not as they are.\nHow, then, is it possible for the belief that things _are_ spatial to\narise? For how can we advance from knowledge of what they look to\nknowledge of what they are but do not look? Again, given that the\nbelief has arisen, may it not after all be illusion? No vindication\nseems possible. For how can it be possible to base the knowledge of\nwhat things are, independently of perception, upon the knowledge of\nwhat they look? Nevertheless, the answer is simple. In the case of the\nperception of what is spatial there is no transition _in principle_\nfrom knowledge of what things look to knowledge of what things are,\nthough there is continually such a transition _in respect of details_.\nIt is, of course, often necessary, and often difficult, to determine\nthe precise position, shape, &c., of a thing, and if we are to come to\na decision, we must appeal to what the thing looks or appears under\nvarious conditions. But, from the very beginning, our consciousness of\nwhat a thing appears in respect of spatial characteristics implies the\nconsciousness of it as spatial and therefore also as, in particular,\nthree-dimensional. If we suppose the latter consciousness absent, any\nassertion as to what a thing appears in respect of spatial\ncharacteristics loses significance. Thus, although there is a process\nby which we come to learn that railway lines are really parallel,\nthere is no process by which we come to learn that they are really\nspatial. Similarly, although there is a process by which we become\naware that a body is a cube, there is no process by which we become\naware that it has a solid shape of some kind; the process is only\nconcerned with the determination of the precise shape of the body.\nThe second difficulty is, therefore, also removed. For if assertions\nconcerning the apparent shape, &c. of things presuppose the\nconsciousness that the things _are_ spatial, to say that this\nconsciousness may be illusory is to say that all statements concerning\nwhat things _appear_, in respect of spatial relations, are equally\nillusory. But, since it is wholly impossible to deny that we can and\ndo state what things appear in this respect, the difficulty must fall\nto the ground.\n\nThere remains to be answered the question whether Kant's position is\ntenable in its other form, viz. that while we cannot say that reality\nis spatial, we can and must say that the appearances which it produces\nare spatial. This question, in view of the foregoing, can be answered\nas soon as it is stated. We must allow that reality is spatial, since,\nas has been pointed out, assertions concerning the apparent shape of\nthings presuppose that they are spatial. We must equally allow that an\nappearance cannot be spatial. For on the one hand, as has just been\nshown, space and spatial relations can only qualify something the\nexistence of which is not relative to perception, since it is\nimpossible to perceive what is spatial as it is; and on the other hand\nan appearance, as being _ex hypothesi_ an appearance to some one,\ni. e. to a percipient, must be relative to perception.\n\nWe may say, then, generally, that analysis of the distinction between\nappearance and reality, as it is actually drawn in our ordinary\nconsciousness, shows the falsity of both forms of the philosophical\nagnosticism which appeals to the distinction. We know things; not\nappearances. We know what things are; and not merely what they appear\nbut are not. We may also say that Kant cannot possibly be successful\nin meeting, at least in respect of space, what he calls 'the easily\nforeseen but worthless objection that the ideality of space and of\ntime would turn the whole sensible world into pure illusion'.[29] For\nspace, according to him, is not a property of things in themselves; it\ncannot, as has been shown, be a property of appearances; to say that\nit is a property of things as they appear to us is self-contradictory;\nand there is nothing else of which it can be said to be a property.\n\n    [29] _Prol_., Sec. 13, Remark iii. (Cf. p. 100 note.) Cf. the\n    confused note B. 70, M. 42. (See Dr. Vaihinger's Commentary\n    on the _Critique_, ii, 488 ff.)\n\nIn conclusion, it may be pointed out that the impossibility that\nspace[30] and spatial characteristics should qualify appearances\nrenders untenable Kant's attempt to draw a distinction between reality\nand appearance _within_ 'phenomena' or 'appearances'. The passage in\nwhich he tries to do so runs as follows:\n\n    [30] The case of time can be ignored, since, as will be seen\n    later (pp. 112-14), the contention that space is 'ideal'\n    really involves the admission that time is real.\n\n\"We generally indeed distinguish in appearances that which essentially\nbelongs to the perception of them, and is valid for every human\nsense in general, from that which belongs to the same perception\naccidentally, as valid not for the sensibility in general, but for a\nparticular state or organization of this or that sense. Accordingly,\nwe are accustomed to say that the former is knowledge which represents\nthe object itself, whilst the latter represents only the appearance\nof the same. This distinction, however, is only empirical. If we stop\nhere (as is usual) and do not again regard that empirical perception\nas itself a mere phenomenon (as we ought to do), in which nothing\nwhich concerns a thing in itself is to be found, our transcendental\ndistinction is lost; and in that case we are after all believing that\nwe know things in themselves, although in the world of sense,\ninvestigate its objects as profoundly as we may, we have to do with\nnothing but appearances. Thus we call the rainbow a mere appearance\nduring a sunny shower, but the rain the thing in itself; and this is\nright, if we understand the latter conception only physically as that\nwhich in universal experience and under all different positions with\nregard to the senses is in perception so and so determined and not\notherwise. But if we consider this empirical element[31] in general,\nand inquire, without considering its agreement with every human sense,\nwhether it represents an object in itself (not the raindrops, for\ntheir being phenomena by itself makes them empirical objects), the\nquestion of the relation of the representation to the object is\ntranscendental; and not only are the raindrops mere appearances, but\neven their circular form, nay, even the space in which they fall,\nare nothing in themselves but mere modifications or fundamental\ndispositions of our sensuous perception; the transcendental object,\nhowever, remains unknown to us.\"[32]\n\n    [31] _Dieses Empirische._\n\n    [32] B. 62-3, M. 37-8. _Erscheinung_ is here translated\n    'appearance'.\n\nKant's meaning is plain. He is anxious to justify the physical\ndistinction made in our ordinary or non-philosophical consciousness\nbetween a thing in itself and a mere appearance,[33] but at the same\ntime to show that it falls within appearances, in respect of the\nphilosophical distinction between things in themselves and appearances\nor phenomena. The physical distinction is the first of which we become\naware, and it arises through problems connected with our senses.\nOwing, presumably, to the contradictions which would otherwise ensue,\nthe mind is forced to distinguish between things and the 'appearances'\nwhich they produce, and to recognize that they do not correspond. The\ndiscrepancy is due to the fact that our perceptions are conditioned by\nthe special positions of our physical organs with regard to the object\nof perception, and we discover its real nature by making allowance for\nthese special positions. We thereby advance in knowledge to the extent\nof overcoming an obstacle due to the nature of our senses. But, this\nobstacle overcome, philosophical reflection forces upon us another.\nThe thing which we distinguish in our ordinary consciousness from its\nappearances is, after all, only another appearance; and although the\nphysical problem is solved concerning its accordance with our special\nsenses, there remains the philosophical problem as to whether this\nappearance need correspond to what in the end is the real thing, viz.\nthat which exists in itself and apart from all perception. The only\npossible answer is that it need not. We therefore can only know\nappearances and not reality; in other words, we cannot have knowledge\nproper. At the same time, our knowledge of appearances is objective to\nthe extent that the appearances in question are the same for every\none, and for us on various occasions; for the effects due to special\npositions of our senses have been removed. If, therefore, we return to\nthe physical distinction, we see that the 'things' to which it refers\nare only a special kind of appearance, viz. that which is the same for\nevery one, and for us at all times. The physical distinction, then,\nbeing a distinction between one kind of appearance and another, falls\nwithin 'phenomena' or 'appearances'.\n\n    [33] It should be noticed that the passage is, in the main,\n    expressed in terms of the distinction between 'things' and\n    'appearances', and not, as it should be, in terms of the\n    distinction between what things are and what things appear\n    or look.\n\nNow the obvious objection to this line of thought is that the result\nof the second or metaphysical application of the distinction between\nreality and appearance is to destroy or annul the first or physical\napplication of it. To oppose the rain, i. e. the raindrops as the\nthing in itself to the rainbow as a mere appearance is to imply that\nthe rain is not an appearance. For though what is opposed to a _mere_\nappearance may still be an appearance, it cannot be called an\nappearance at all if it be described as the thing in itself. If it be\nonly another appearance, it is the same in principle as that to which\nit is opposed, and consequently cannot be opposed to it. Thus, if Kant\nmeans by the rain, in distinction from the rainbow, the appearance\nwhen, as we say, we see the circular raindrops, the title of this\nappearance to the term thing in itself is no better than that of the\nrainbow; it is, in fact, if anything, worse, for the appearance is\nactual only under exceptional circumstances. We may never see the\nraindrops thus, or in Kant's language, have this 'appearance'; and\ntherefore, in general, an appearance of this kind is not actual but\nonly possible. The truth is that we can only distinguish something as\nthe thing in itself from an appearance, so long as we mean by the\nthing in itself what Kant normally means by it, viz. something which\nexists independently of perception and is not an appearance at\nall.[34] That of which Kant is really thinking, and which he _calls_\nthe appearance which is the thing, in distinction from a mere\nappearance, is not an appearance; on the contrary, it is the raindrops\nthemselves, which he describes as circular and as falling through\nspace, and which, as circular and falling, must exist and have these\ncharacteristics in themselves apart from a percipient. Kant's formula\nfor an empirical thing, i. e. a thing which is an appearance, viz.\n'that which in universal experience and under all different positions\nwith regard to the senses is in perception so and so determined', is\nmerely an attempt to achieve the impossible, viz. to combine in one\nthe characteristics of a thing and an appearance. While the reference\nto _perception_ and to _position with regard to the senses_ implies\nthat what is being defined is an appearance, the reference to\n_universal_ experience, to _all_ positions with regard to the senses,\nand to that which _is so and so determined_ implies that it is a\nthing. But, plainly, mention of position with regard to the senses, if\nintroduced at all, should refer to the _differences_ in perception due\nto the different position of the object in particular cases. There is\nnothing of which it can be said that we perceive it in the same way or\nthat it looks the same from _all_ positions. When Kant speaks of that\nwhich under _all_ different positions with regard to the senses is so\nand so determined, he is really referring to something in the\nconsideration of which all reference to the senses has been discarded;\nit is what should be described as that which _in reality and apart\nfrom_ all positions with regard to the senses is so and so determined;\nand this, as such, cannot be an appearance. Again, the qualification\nof 'is so and so determined' by 'in perception' is merely an attempt\nto treat as relative to perception, and so as an appearance, what is\nessentially independent of perception.[35] Kant, no doubt, is thinking\nof a real presupposition of the process by which we distinguish\nbetween the real and the apparent qualities of bodies, i. e. between\nwhat they are and what they appear. We presuppose that that quality is\nreally, and not only apparently, a quality of a body, which we and\nevery one, judging from what it looks under various conditions (i. e.\n'in universal experience'), must believe it to possess in itself and\nindependently of all perception. His mistake is that in formulating\nthis presupposition he treats as an appearance, and so as relative to\nperception, just that which is being distinguished from what, as an\nappearance, is relative to perception.\n\n    [34] Hence Kant's protest (B. 45, M. 27), against\n    illustrating the ideality of space by the 'inadequate'\n    examples of colour, taste, &c., must be unavailing. For his\n    contention is that, while the assertion that space is not a\n    property of things means that it is not a property of things\n    in themselves, the assertion that colour, for example, is not\n    a property of a rose only means that it is not a property of\n    a thing in itself in an empirical sense, i. e. of an\n    appearance of a special kind.\n\n    [35] Cf. pp. 72-3.\n\nUnderlying the mistake is the identification of perception with\njudgement. Our apprehension of what things _are_ is essentially a\nmatter of thought or judgement, and not of perception. We do not\n_perceive_[36] but _think_ a thing as it is. It is true that we can\nfollow Kant's language so far as to say that our judgement that the\nportion of the great circle joining two points on the surface of a\nsphere is the shortest way between them _via_ the surface belongs\nessentially to the thinking faculty of every intelligent being, and\nalso that it is valid for all intelligences, in the sense that they\nmust all hold it to be true; and we can contrast this judgement with a\nperception of the portion of the great circle as something which,\nthough it cannot be said to be invalid, still differs for different\nbeings according to the position from which they perceive it. Kant,\nhowever, treats the judgement as a _perception_; for if we apply his\ngeneral assertion to this instance, we find him saying that what we\njudge the portion of the great circle to be essentially belongs to the\n_perception_ of it, and is valid for the _sensuous_ faculty of every\nhuman being, and that thereby it can be distinguished from what\nbelongs to the same perception of a great circle accidentally, e. g.\nits apparent colour, which is valid only for a particular organization\nof this or that sense.[37] In this way he correlates what the great\ncircle really is, as well as what it looks, with perception, and so is\nable to speak of what it is for perception. But, in fact, what the\ngreat circle is, is correlated with thought, and not with perception;\nand if we raise Kant's transcendental problem in reference not to\nperception but to thought, it cannot be solved in Kant's agnostic\nmanner. For it is a presupposition of thinking that things are in\nthemselves what we think them to be; and from the nature of the case a\npresupposition of thinking not only cannot be rightly questioned, but\ncannot be questioned at all.\n\n    [36] Cf. pp. 72-3.\n\n    [37] In the _Prol._, Sec. 13, Remark iii, Kant carefully\n    distinguishes judgement from perception, but destroys the\n    effect of the distinction by regarding judgement as referring\n    to what is relative to perception, viz. appearances.\n\n\n\n\nNOTE ON THE FIRST ANTINOMY\n\n\nKant holds that the antinomy or contradiction which arises when we\nconsider the character of the world as spatial and temporal, viz. that\nwe are equally bound to hold that the world is infinite in space and\ntime, and that it is finite in space and time, is due to regarding the\nworld as a thing in itself. He holds that the contradiction\ndisappears, as soon as it is recognized that the world is only a\nphenomenon, for then we find that we need only say that the world is\n_capable_ of being extended infinitely in respect of time and\nspace.[1] Objects in space and time are only phenomena, and, as such,\nare actual only in perception. When we say that a past event, or that\na body which we do not perceive, is real, we merely assert the\npossibility of a 'perception'. \"All events from time immemorial prior\nto my existence mean nothing else than the possibility of prolonging\nthe chain of experience from the present perception upwards to the\nconditions which determine this perception according to time.\"[2]\n\"That there may be inhabitants of the moon, although no one has ever\nseen them, must certainly be admitted, but this assertion only means\nthat we could come upon them in the possible progress of\nexperience.\"[3] The contradictions, therefore, can be avoided by\nsubstituting for the actual infinity of space and time, as relating to\nthings in themselves, the possible infinity of a series of\n'perceptions'.\n\n    [1] B. 532-3, M. 315.\n\n    [2] B. 523, M. 309.\n\n    [3] B. 521, M. 308.\n\nThis contention, if successful, is clearly important. If it could be\nshown that the treatment of the world as a thing in itself is the\nsource of a contradiction, we should have what at least would seem a\nstrong, if not conclusive, ground for holding that the world is a\nphenomenon, and, consequently, that the distinction between phenomena\nand things in themselves is valid.\n\nProfessor Cook Wilson has, however, pointed out that Kant's own\ndoctrine does not avoid the difficulty. For, though, according to\nKant, the infinity of actual representations of spaces and times is\nonly possible, yet the possibilities of these representations will be\nthemselves infinite, and, as such, will give rise to contradictions\nsimilar to those involved in the infinity of space and time. Moreover,\nas Professor Cook Wilson has also pointed out, there is no\ncontradiction involved in the thought of the world as spatial and\ntemporal; for, as we see when we reflect, we always presuppose that\nspace and time are infinite, and we are only tempted to think that\nthey must be finite, because, when maintaining that the world must be\na whole, we are apt to make the false assumption, without in any way\nquestioning it, that any whole must be finite.\n\n\n\n\nCHAPTER V\n\nTIME AND INNER SENSE\n\n\nThe arguments by which Kant seeks to show that time is not a\ndetermination of things in themselves but only a form of perception\nare, _mutatis mutandis_, identical with those used in his treatment of\nspace.[1] They are, therefore, open to the same criticisms, and need\nno separate consideration.\n\n    [1] Cf. B. 46-9, Secs. 4, 5 and 6 (a), M. 28-30, Secs. 5, 6 and\n    7 (a) with B. 38-42, Sec. 2 (1-4), and Sec. (3) to (a) inclusive,\n    M. 23-6, Secs. 2, 3, and 4 (a). The only qualification needed is\n    that, since the parts of time cannot, like those of space,\n    be said to exist simultaneously, B. Sec. 4 (5), M. Sec. 5, 5 is\n    compelled to appeal to a different consideration from that\n    adduced in the parallel passage on space (B. Sec. 2 (4), M. Sec.\n    2, 4). Since, however, B. Sec. 4 (5), M. Sec. 5, 5 introduces no\n    new matter, but only appeals to the consideration already urged\n    (B. Sec. 4, 4, M. Sec. 5, 4), this difference can be neglected.\n    B. Sec. 5, M. Sec. 6 adds a remark about change which does not\n    affect the main argument.\n\nTime, however, according to Kant, differs from space in one important\nrespect. It is the form not of outer but of inner sense; in other\nwords, while space is the form under which we perceive things, time is\nthe form under which we perceive ourselves. It is upon this difference\nthat attention must be concentrated. The existence of the difference\nat all is upon general grounds surprising. For since the arguments by\nwhich Kant establishes the character of time as a form of perception\nrun _pari passu_ with those used in the case of space, we should\nexpect time, like space, to be a form under which we perceive things;\nand, as a matter of fact, it will be found that the only _argument_\nused to show that time is the form of inner, as opposed to outer,\nsense is not only independent of Kant's general theory of forms of\nsense, but is actually inconsistent with it.[2] Before, however, we\nattempt to decide Kant's right to distinguish between inner and outer\nsense, we must consider the facts which were before Kant's mind in\nmaking the distinction.\n\n    [2] B. 49 (b), M. 30 (b). See pp. 109-12.\n\nThese facts and, to a large extent, the frame of mind in which Kant\napproached them, find expression in the passage in Locke's _Essay_,\nwhich explains the distinction between 'ideas of sensation' and 'ideas\nof reflection'.\n\n\"Whence has it [i. e. the mind] all the materials of reason and\nknowledge? To this I answer, in one word, from experience.... Our\nobservation, employed either about external, sensible objects, or\nabout the internal operations of our minds, perceived and reflected\non, by ourselves, is that which supplies our understandings with all\nthe materials of thinking. These two are the fountains of\nknowledge....\"\n\n\"First, Our senses, conversant about particular sensible objects, do\nconvey into the mind several distinct perceptions of things, according\nto those various ways, wherein those objects do affect them: and thus\nwe come by those ideas we have of Yellow, White, Heat, Cold, Soft,\nHard, Bitter, Sweet, and all those, which we call sensible qualities;\nwhich, when I say the senses convey into the mind, I mean, they, from\nexternal objects, convey into the mind what produces there those\nperceptions. This great source of most of the ideas we have, depending\nwholly upon our senses, and derived by them to the understanding, I\ncall _sensation_.\"\n\n\"Secondly, The other fountain, from which experience furnisheth the\nunderstanding with ideas, is the perception of the operations of our\nown mind within us, as it is employed about the ideas it has got;\nwhich operations, when the soul comes to reflect on, and consider, do\nfurnish the understanding with another set of ideas, which could not\nbe had from things without; and such are Perception, Thinking,\nDoubting, Believing, Reasoning, Knowing, Willing, and all the\ndifferent actings of our own minds; which we being conscious of, and\nobserving in ourselves, do, from these, receive into our\nunderstandings as distinct ideas, as we do from bodies affecting our\nsenses. This source of ideas every man has wholly in himself; and\nthough it be not sense as having nothing to do with external objects,\nyet it is very like it, and might properly enough be called internal\nsense. But, as I call the other sensation, so I call this\n_reflection_; the ideas it affords being such only as the mind gets,\nby reflecting on its own operations within itself.\"[3]\n\n    [3] Locke, _Essay_, ii, 1, Secs. 2-4.\n\nHere Locke is thinking of the distinction between two attitudes of\nmind, which, however difficult it may be to state satisfactorily, must\nin some sense be recognized. The mind, undoubtedly, in virtue of its\npowers of perceiving and thinking--or whatever they may be--becomes\nthrough a temporal process aware of a spatial world in its varied\ndetail. In the first instance, its attention is absorbed in the world\nof which it thus becomes aware; subsequently, however, it is in some\nway able to direct its attention away from this world to the\nactivities in virtue of which it has become aware of this world, and\nin some sense to make itself its own object. From being conscious it\nbecomes self-conscious. This process by which the mind turns its\nattention back upon itself is said to be a process of 'reflection'.\nWhile we should say that it is by perception that we become aware of\nthings in the physical world, we should say that it is by reflection\nthat we become aware of our activities of perceiving, thinking,\nwilling, &c. Whatever difficulties the thought of self-consciousness\nmay involve, and however inseparable, and perhaps even temporally\ninseparable, the attitudes of consciousness and self-consciousness may\nturn out to be, the distinction between these attitudes must be\nrecognized. The object of the former is the world, and the object of\nthe latter is in some sense the mind itself; and the attitudes may be\ndescribed as that of our ordinary, scientific, or unreflecting\nconsciousness and that of reflection.\n\nThe significance of Locke's account of this distinction lies for our\npurposes in its anticipation of Kant. He states the second attitude,\nas well as the first, in terms of sense. Just as in our apprehension\nof the world things external to, in the sense of existing\nindependently of, the mind are said to act on our physical organs or\n'senses', and thereby to produce 'perceptions' in the mind, so the\nmind is said to become conscious of its own operations by 'sense'. We\nshould notice, however, that Locke hesitates to use the word 'sense'\nin the latter case, on the ground that it involves no operation of\nexternal things (presumably upon our physical organs), though he\nthinks that the difficulty is removed by calling the sense in question\n'internal'.\n\nKant is thinking of the same facts, and also states them in terms of\nsense, though allowance must be made for the difference of standpoint,\nsince for him 'sense', in the case of the external sense, refers not\nto the affection of our physical organs by physical bodies, but to the\naffection of the mind by things in themselves. Things in themselves\nact on our minds and produce in them appearances, or rather\nsensations, and outer sense is the mind's capacity for being so\naffected by outer things, i. e. things independent of the mind. This\nis, in essentials, Kant's statement of the attitude of consciousness,\ni. e. of our apprehension of the world which exists independently of\nthe mind, and which, for him, is the world of things in themselves. He\nalso follows Locke in giving a parallel account of the attitude of\nself-consciousness. He asks, 'How can the subject perceive itself?'\nPerception _in man_ is essentially passive; the mind must be\n_affected_ by that which it perceives. Consequently, if the mind is to\nperceive itself, it must be affected by its own activity; in other\nwords, there must be an inner sense, i. e. a capacity in virtue of\nwhich the mind is affected by itself.[4] Hence Kant is compelled to\nextend his agnosticism to the knowledge of ourselves. Just as we do\nnot know things, but only the appearances which they produce in\nus,[5] so we do not know ourselves, but only the appearances which we\nproduce in ourselves; and since time is a mode of relation of these\nappearances, it is a determination not of ourselves, but only of the\nappearances due to ourselves.\n\n    [4] Cf. B. 67 fin., M. 41 init.\n\n    [5] It is here assumed that this is Kant's normal view of the\n    phenomenal character of our knowledge. Cf. p. 75.\n\nThe above may be said to represent the train of thought by which Kant\narrived at his doctrine of time and the inner sense. It was reached by\ncombining recognition of the fact that we come to be aware not only of\nthe details of the physical world, but also of the successive process\non our part by which we have attained this knowledge, with the view\nthat our apprehension of this successive process is based on 'sense',\njust as is our apprehension of the world. But the question remains\nwhether Kant is, on his own principles, entitled to speak of an inner\nsense at all. According to him, knowledge begins with the production\nin us of sensations, or, as we ought to say in the present context,\nappearances by the action of things in themselves. These sensations or\nappearances can reasonably be ascribed to external sense. They may be\nascribed to sense, because they arise through our being _affected_ by\nthings in themselves. The sense may be called external, because the\nobject affecting it is external to the mind, i. e. independent of it.\nIn conformity with this account, internal sense must be the power of\nbeing affected by something internal to the mind, i. e. dependent upon\nthe mind itself, and since being affected implies the activity of\naffecting, it will be the power of being affected by the mind's own\nactivity.[6] The activity will presumably be that of arranging\nspatially the sensations or appearances due to things in\nthemselves.[7] This activity must be said to produce an affection in\nus, the affection being an appearance due to ourselves. Lastly, the\nmind must be said to arrange these appearances temporally. Hence it\nwill be said to follow that we know only the appearances due to\nourselves and not ourselves, and that time is only a determination of\nthese appearances.[8]\n\n    [6] B. 68 init., M. 41 init.\n\n    [7] The precise nature of the activity makes no difference to\n    the argument.\n\n    [8] In B. 152 fin., M. 93 fin. Kant expresses his conclusion\n    in the form that we know ourselves only as we appear to\n    ourselves, and not as we are in ourselves (cf. p. 75). The\n    above account, and the criticism which immediately follows,\n    can be adapted, _mutatis mutandis_, to this form of the view.\n\nThe weakness of the position just stated lies on the surface. It\nprovides no means of determining whether any affection produced in us\nis produced by ourselves rather than by the thing in itself;\nconsequently we could never say that a given affection was an\nappearance due to _ourselves_, and therefore to _inner_ sense. On the\ncontrary, we should ascribe all affections to things in themselves,\nand should, therefore, be unable to recognize an _inner_ sense at all.\nIn order to recognize an inner sense we must know that certain\naffections are due to _our_ activity, and, to do this, we must know\nwhat the activity consists in--for we can only be aware that we are\nactive by being aware of an activity of ours of a particular\nkind--and, therefore, we must know ourselves. Unless, then, we know\nourselves, we cannot call any affections internal.\n\nIf, however, the doctrine of an internal sense is obviously untenable\nfrom Kant's own point of view, why does he hold it? The answer is\nthat, inconsistently with his general view, he continues to think of\nthe facts as they really are, and that he is deceived by an ambiguity\ninto thinking that the facts justify a distinction between internal\nand external sense.\n\nHe brings forward only one argument to show that time is the form of\nthe internal sense. \"Time is nothing else than the form of the\ninternal sense, i. e. of the perception of ourselves and our inner\nstate. For time cannot be any determination of external phenomena; it\nhas to do neither with a shape nor a position; on the contrary, it\ndetermines the relation of representations in our internal state.\"[9]\n\n    [9] B. 49 (b), M. 30 (b).\n\nTo follow this argument it is first necessary to realize a certain\nlooseness and confusion in the expression of it. The term 'external',\napplied to phenomena, has a double meaning. It must mean (1) that of\nwhich the parts are external to one another, i. e. spatial; for the\nground on which time is denied to be a determination of external\nphenomena is that it has nothing to do with a shape or a position. It\nmust also mean (2) external to, in the sense of independent of, the\nmind; for it is contrasted with our internal state, and if 'internal',\napplied to 'our state', is not to be wholly otiose, it can only serve\nto emphasize the contrast between our state and something external to\nin the sense of independent of us. Again, 'phenomena,' in the phrase\n'external phenomena', can only be an unfortunate expression for\nthings independent of the mind, these things being here called\nphenomena owing to Kant's view that bodies in space are phenomena.\nOtherwise, 'phenomena' offers no contrast to 'our state' and to\n'representations'. The passage, therefore, presupposes a distinction\nbetween states of ourselves and things in space, the former being\ninternal to, or dependent upon, and the latter external to, or\nindependent of, the mind.\n\nIt should now be easy to see that the argument involves a complete\n_non sequitur_. The conclusion which is justified is that time is a\nform not of things but of our own states. For the fact to which he\nappeals is that while things, as being spatial, are not related\ntemporally, our states are temporally related; and if 'a form' be\nunderstood as a mode of relation, this fact can be expressed by the\nformula 'Time is a form not of things but of our own states', the\ncorresponding formula in the case of space being 'Space is a form not\nof our states but of things'. But the conclusion which Kant desires\nto draw--and which he, in fact, actually draws--is the quite different\nconclusion that time is a form of _perception_ of our states, the\ncorresponding conclusion in the case of space being that space is a\nform of perception of things. For time is to be shown to be the form\nof inner sense, i. e. the form of the perception of what is internal\nto ourselves, i. e. of our own states.[10] The fact is that the same\nunconscious transition takes place in Kant's account of time which, as\nwe saw,[11] takes place in his account of space. In the case of space,\nKant passes from the assertion that space is a form of things, in the\nsense that all things are spatially related--an assertion which he\nexpresses by saying that space is the form of phenomena--to the quite\ndifferent assertion that space is a form of perception, in the sense\nof a way in which we perceive things as opposed to a way in which\nthings are. Similarly, in the case of time, Kant passes from the\nassertion that time is the form of our internal states, in the sense\nthat all our states are temporally related, to the assertion that time\nis a way in which we perceive our states as opposed to a way in which\nour states really are. Further, the two positions, which he thus fails\nto distinguish, are not only different, but incompatible. For if space\nis a form of things, and time is a form of our states, space and time\ncannot belong only to our mode of perceiving things and ourselves\nrespectively, and not to the things and ourselves; for _ex hypothesi_\nthings _are_ spatially related, and our states _are_ temporally\nrelated.\n\n    [10] Cf. B. 49 (b) line 2, M. 30 (b) line 2\n\n    [11] Cf. pp. 38-40.\n\nKant's procedure, therefore, may be summed up by saying that he\nformulates a view which is true but at the same time inconsistent with\nhis general position, the view, viz. that while things in space are\nnot temporally related, the acts by which we come to apprehend them\nare so related; and further, that he is deceived by the verbally easy\ntransition from a legitimate way of expressing this view, viz. that\ntime is the form of our states, to the desired conclusion that time is\nthe form of inner sense.\n\nThe untenable character of Kant's position with regard to time and the\nknowledge of ourselves can be seen in another way. It is not difficult\nto show that, in order to prove that we do not know _things_, but only\nthe appearances which they produce, we must allow that we do know\n_ourselves_, and not appearances produced by ourselves, and,\nconsequently, that time is real and not phenomenal. To show this, it\nis only necessary to consider the objection which Kant himself quotes\nagainst his view of time. The objection is important in itself, and\nKant himself remarks that he has heard it so unanimously urged by\nintelligent men that he concludes that it must naturally present\nitself to every reader to whom his views are novel. According to Kant,\nit runs thus: \"Changes are real (this is proved by the change of our\nown representations, even though all external phenomena, together with\ntheir changes, be denied). Now changes are only possible in time;\ntherefore time is something real.\"[12] And he goes on to explain why\nthis objection is so unanimously brought, even by those who can bring\nno intelligible argument against the ideality of space. \"The reason is\nthat men have no hope of proving apodeictically the absolute reality\nof space, because they are confronted by idealism, according to which\nthe reality of external objects is incapable of strict proof, whereas\nthe reality of the object of our internal senses (of myself and my\nstate) is immediately clear through consciousness. External objects\nmight be mere illusion, but the object of our internal senses is to\ntheir mind undeniably something real.\"[13]\n\n    [12] B. 53, M. 32.\n\n    [13] B. 55, M. 33.\n\nHere, though Kant does not see it, he is faced with a difficulty from\nwhich there is no escape. On the one hand, according to him, we do not\nknow things in themselves, i. e. things independent of the mind. In\nparticular, we cannot know that they are spatial; and the objection\nquoted concedes this. On the other hand, we do know phenomena or\nthe appearances produced by things in themselves. Phenomena or\nappearances, however, as he always insists, are essentially states or\ndeterminations of the mind. To the question, therefore, 'Why are we\njustified in saying that we do know phenomena, whereas we do not know\nthe things which produce them?' Kant could only answer that it is\nbecause phenomena are dependent upon the mind, as being its own\nstates.[14] As the objector is made to say, 'the reality of the object\nof our internal senses (of myself and my state) is immediately clear\nthrough consciousness.' If we do not know things in themselves,\nbecause they are independent of the mind, we only know phenomena\nbecause they are dependent upon the mind. Hence Kant is only justified\nin denying that we know things in themselves if he concedes that we\nreally know our own states, and not merely appearances which they\nproduce.\n\n    [14] Cf. p. 123.\n\nAgain, Kant must allow--as indeed he normally does--that these states\nof ours are related by way of succession. Hence, since these states\nare really our states and not appearances produced by our states,\nthese being themselves unknown, time, as a relation of these states,\nmust itself be real, and not a way in which we apprehend what is real.\nIt must, so to say, be really in what we apprehend about ourselves,\nand not put into it by us as perceiving ourselves.\n\nThe objection, then, comes to this. Kant must at least concede that\nwe undergo a succession of changing states, even if he holds that\n_things_, being independent of the mind, cannot be shown to undergo\nsuch a succession; consequently, he ought to allow that time is not a\nway in which we apprehend ourselves, but a real feature of our real\nstates. Kant's answer[15] does not meet the point, and, in any case,\nproceeds on the untenable assumption that it is possible for the\ncharacteristic of a thing to belong to it as perceived, though not\nin itself.[16]\n\n    [15] B. 55, M. 33 med.\n\n    [16] Cf. pp. 71-3.\n\n\n\n\nCHAPTER VI\n\nKNOWLEDGE AND REALITY\n\n\nKant's theory of space, and, still more, his theory of time, are\nbewildering subjects. It is not merely that the facts with which he\ndeals are complex; his treatment of them is also complicated by his\nspecial theories of 'sense' and of 'forms of perception'. Light,\nhowever, may be thrown upon the problems raised by the _Aesthetic_,\nand upon Kant's solution of them, in two ways. In the first place, we\nmay attempt to vindicate the implication of the preceding criticism,\nthat the very nature of knowledge presupposes the independent\nexistence of the reality known, and to show that, in consequence, all\nidealism is of the variety known as subjective. In the second place,\nwe may point out the way in which Kant is misled by failing to realize\n(1) the directness of the relation between the knower and the reality\nknown, and (2) the impossibility of transferring what belongs to one\nside of the relation to the other.\n\nThe question whether any reality exists independently of the knowledge\nof it may be approached thus. The standpoint of the preceding\ncriticism of Kant may be described as that of the plain man. It is the\nview that the mind comes by a temporal process to apprehend or to know\na spatial world which exists independently of it or of any other mind,\nand that the mind knows it as it exists in the independence. 'Now this\nview,' it may be replied, 'is exposed to at least one fatal objection.\nIt presupposes the possibility of knowing the thing in itself, i. e.\nsomething which exists independently of the mind which comes to know\nit. Whatever is true, this is not. Whatever be the criticism to which\nKant's doctrine is exposed in detail, it contains one inexpugnable\nthesis, viz. that the thing in itself cannot be known. Unless the\nphysical world stands in essential relation to the mind, it is\nimpossible to understand how it can be known. This position being\nunassailable, any criticism of an idealistic theory must be compatible\nwith it, and therefore confined to details. Moreover, Kant's view can\nbe transformed into one which will defy criticism. Its unsatisfactory\ncharacter lies in the fact that in regarding the physical world as\ndependent on the mind, it really alters the character of the world by\nreducing the world to a succession of 'appearances' which, as such,\ncan only be mental, i. e. can only belong to the mind's own being.\nBodies, as being really appearances in the mind, are regarded as on\nthe level of transitory mental occurrences, and as thereby at least\nresembling feelings and sensations. This consequence, however, can be\navoided by maintaining that the real truth after which Kant was\ngroping was that knower and known form an inseparable unity, and that,\ntherefore, any reality which is not itself a knower, or the knowing of\na knower, presupposes a mind which knows it. In that case nothing is\nsuggested as to the special nature of the reality known, and, in\nparticular, it is not implied to be a transitory element of the mind's\nown being. The contention merely attributes to any reality, conceived\nto have the special nature ordinarily attributed to it, the additional\ncharacteristic that it is known. Consequently, on this view, the\nphysical world can retain the permanence ordinarily attributed to it.\nTo the objection that, at any rate, _our_ knowledge is transitory,\nand that if the world is relative to it the world also must be\ntransitory, it may be replied--though with some sense of\nuneasiness--that the world must be considered relative not to us as\nknowers, but to a knower who knows always and completely, and whose\nknowing is in some way identical with ours. Further, the view so\ntransformed has two other advantages. In the first place, it renders\nit possible to dispense with what has been called the Mrs. Harris of\nphilosophy, the thing in itself. As Kant states his position, the\nthing in itself must be retained, for it is impossible to believe that\nthere is no reality other than what is mental. But if the physical\nworld need not be considered to be a succession of mental occurrences,\nit can be considered to be the reality which is not mental. In the\nsecond place, knowledge proper is vindicated, for on this view we do\nnot know 'only' phenomena; we know the reality which is not mental,\nand we know it as it is, for it is as object of knowledge.'\n\n'Moreover, the contention must be true, and must form the true basis\nof idealism. For the driving force of idealism is furnished by the\nquestion, 'How can the mind and reality come into the relation which\nwe call knowledge?' This question is unanswerable so long as reality\nis thought to stand in no essential relation to the knowing mind.\nConsequently, in the end, knowledge and reality must be considered\ninseparable. Again, even if it be conceded that the mind in some way\ngains access to an independent reality, it is impossible to hold that\nthe mind can really know it. For the reality cannot in the relation of\nknowledge be what it is apart from this relation. It must become in\nsome way modified or altered in the process. Hence the mind cannot on\nthis view know the reality as it is. On the other hand, if the reality\nis essentially relative to a knower, the knower knows it as it is, for\nwhat it is is what it is in this relation.'\n\nThe fundamental objection, however, to this line of thought is that it\ncontradicts the very nature of knowledge. Knowledge unconditionally\npresupposes that the reality known exists independently of the\nknowledge of it, and that we know it as it exists in this\nindependence. It is simply _impossible_ to think that any reality\ndepends upon our knowledge of it, or upon any knowledge of it. If\nthere is to be knowledge, there must first _be_ something to be known.\nIn other words, knowledge is essentially discovery, or the finding of\nwhat already is. If a reality could only be or come to be in virtue of\nsome activity or process on the part of the mind, that activity or\nprocess would not be 'knowing', but 'making' or 'creating', and to\nmake and to know must in the end be admitted to be mutually\nexclusive.[1]\n\n    [1] Cf. pp. 235-6.\n\nThis presupposition that what is known exists independently of being\nknown is quite general, and applies to feeling and sensation just as\nmuch as to parts of the physical world. It must in the end be conceded\nof a toothache as much as of a stone that it exists independently of\nthe knowledge of it. There must be a pain to be attended to or\nnoticed, which exists independently of our attention or notice. The\ntrue reason for asserting feeling and sensation to be dependent on the\nmind is that they presuppose not a knowing, but a feeling and a\nsentient subject respectively. Again, it is equally presupposed that\nknowing in no way alters or modifies the thing known. We can no more\nthink that in apprehending a reality we do not apprehend it as it is\napart from our knowledge of it, than we can think that its existence\ndepends upon our knowledge of it. Hence, if 'things in themselves'\nmeans 'things existing independently of the knowledge of them',\nknowledge is essentially of 'things in themselves'. It is, therefore,\nunnecessary to consider whether idealism is assisted by the\nsupposition of a non-finite knowing mind, correlated with reality as a\nwhole. For reality must equally be independent of it. Consequently, if\nthe issue between idealism and realism is whether the physical world\nis or is not dependent on the mind, it cannot turn upon a dependence\nin respect of knowledge.\n\nThat the issue does not turn upon knowledge is confirmed by our\ninstinctive procedure when we are asked whether the various realities\nwhich we suppose ourselves to know depend upon the mind. Our natural\nprocedure is not to treat them simply as realities and to ask whether,\nas realities, they involve a mind to know them, but to treat them as\nrealities of the particular kind to which they belong, and to consider\nrelation to the mind of some kind other than that of knowledge. We\nshould say, for instance, that a toothache or an emotion, as being a\nfeeling, presupposes a mind capable of feeling, whose feeling it is;\nfor if the mind be thought of as withdrawn, the pain or the feeling\nmust also be thought of as withdrawn. We should say that an act of\nthinking presupposes a mind which thinks. We should, however,\nnaturally deny that an act of thinking or knowing, in order to be,\npresupposes that it is known either by the thinker whose act it is, or\nby any other mind. In other words, we should say that knowing\npresupposes a mind, not as something which _knows_ the knowing, but\nas something which _does_ the knowing. Again, we should naturally say\nthat the shape or the weight of a stone is _not_ dependent on the mind\nwhich perceives the stone. The shape, we should say, would disappear\nwith the disappearance of the stone, but would not disappear with the\ndisappearance of the mind which perceives the stone. Again, we should\nassert that the stone itself, so far from depending on the mind which\nperceives it, has an independent being of its own. We might, of\ncourse, find difficulty in deciding whether a reality of some\nparticular kind, e. g. a colour, is dependent on a mind. But, in any\ncase, we should think that the ground for decision lay in the special\ncharacter of the reality in question, and should not treat it merely\nas a reality related to the mind as something known. We should ask,\nfor instance, whether a colour, as a colour, involves a mind which\nsees, and not whether a colour, as a reality, involves its being\nknown. Our natural procedure, then, is to divide realities into two\nclasses, those which depend on a mind, and may therefore be called\nmental, and those which do not, and to conclude that some realities\ndepend upon the mind, while others do not. We thereby ignore a\npossible dependence of realities on their being known; for not only is\nthe dependence which we recognize of some other kind, e. g. in respect\nof feeling or sentience, but if the dependence were in respect of\nknowledge, we could not distinguish in respect of dependence between\none reality and another.\n\nFurther, if reality be allowed to exist independently of knowledge, it\nis easy to see that, from the idealist's point of view, Kant's\nprocedure was essentially right, and that all idealism, when pressed,\nmust prove subjective; in other words, that the idealist must hold\nthat the mind can only know what is mental and belongs to its own\nbeing, and that the so-called physical world is merely a succession of\nappearances. Moreover, our instinctive procedure[2] is justified. For,\nin the first place, since it is impossible to think that a reality\ndepends for its existence upon being known, it is impossible to reach\nan idealistic conclusion by taking into account relation by way of\nknowledge; and if this be the relation considered, the only conclusion\ncan be that all reality is independent of the mind. Again, since\nknowledge is essentially of reality as it is apart from its being\nknown, the assertion that a reality is dependent upon the mind is an\nassertion of the kind of thing which it is in itself, apart from its\nbeing known.[3] And when we come to consider what we mean by saying of\na reality that it depends upon the mind, we find we mean that it is in\nits own nature of such a kind as to disappear with the disappearance\nof the mind, or, more simply, that it is of the kind called mental.\nHence, we can only decide that a particular reality depends upon the\nmind by appeal to its special character. We cannot treat it simply as\na reality the relation of which to the mind is solely that of\nknowledge. And we can only decide that all reality is dependent upon\nthe mind by appeal to the special character of all the kinds of\nreality of which we are aware. Hence, Kant in the _Aesthetic_, and\nBerkeley before him, were essentially right in their procedure. They\nboth ignored consideration of the world simply as a reality, and\nappealed exclusively to its special character, the one arguing that in\nits special character as spatial and temporal it presupposed a\npercipient, and the other endeavouring to show that the primary\nqualities are as relative to perception as the secondary.\nUnfortunately for their view, in order to think of bodies in space as\ndependent on the mind, it is necessary to think of them as being in\nthe end only certain sensations or certain combinations of sensations\nwhich may be called appearances. For only sensations or combinations\nof them can be thought of as at once dependent on the mind, and\ncapable with any plausibility of being identified with bodies in\nspace. In other words, in order to think of the world as dependent on\nthe mind, we have to think of it as consisting only of a succession of\nappearances, and in fact Berkeley, and, at certain times, Kant, did\nthink of it in this way.\n\n    [2] Cf. p. 119.\n\n    [3] Though not apart from relation to the mind of some other\n    kind.\n\nThat this is the inevitable result of idealism is not noticed, so long\nas it is supposed that the essential relation of realities to the mind\nconsists in their being known; for, as we have seen, nothing is\nthereby implied as to their special nature. To say of a reality that\nit is essentially an object of knowledge is merely to add to the\nparticular nature ordinarily attributed to the existent in question\nthe further characteristic that it must be known.[4] Moreover,\nsince in fact, though contrary to the theory, any reality exists\nindependently of the knowledge of it, when the relation thought of\nbetween a reality and the mind is _solely_ that of knowledge, the\nrealities can be thought of as independent of the mind. Consequently,\nthe physical world can be thought to have that independence of the\nmind which the ordinary man attributes to it, and, therefore, need not\nbe conceived as only a succession of appearances. But the advantage of\nthis form of idealism is really derived from the very fact which it\nis the aim of idealism in general to deny. For the conclusion that the\nphysical world consists of a succession of appearances is only avoided\nby taking into account the relation of realities to the mind by way of\nknowledge, and, then, without being aware of the inconsistency, making\nuse of the independent existence of the reality known.\n\n    [4] Cf. p. 116.\n\nAgain, that the real contrary to realism is _subjective_ idealism is\nconfirmed by the history of the theory of knowledge from Descartes\nonwards. For the initial supposition which has originated and\nsustained the problem is that in knowledge the mind is, at any rate in\nthe first instance, confined within itself. This supposition granted,\nit has always seemed that, while there is no difficulty in\nunderstanding the mind's acquisition of knowledge of what belongs to\nits own being, it is difficult, if not impossible, to understand how\nit can acquire knowledge of what does not belong to its own being.\nFurther, since the physical world is ordinarily thought of as\nsomething which does not belong to the mind's own being, the problem\nhas always been not 'How is it possible to know anything?' but 'How is\nit possible to know a particular kind of reality, viz. the physical\nworld?' Moreover, in consequence of the initial supposition, any\nanswer to this question has always presupposed that our apprehension\nof the physical world is indirect. Since _ex hypothesi_ the mind is\nconfined within itself, it can only apprehend a reality independent of\nit through something within the mind which 'represents' or 'copies'\nthe reality; and it is perhaps Hume's chief merit that he showed that\nno such solution is possible, or, in other words, that, on the given\nsupposition, knowledge of the physical world is impossible.\n\nNow the essential weakness of this line of thought lies in the initial\nsupposition that the mind can only apprehend what belongs to its own\nbeing. It is as much a fact of our experience that we directly\napprehend bodies in space, as that we directly apprehend our feelings\nand sensations. And, as has already been shown,[5] what is spatial\ncannot be thought to belong to the mind's own being on the ground that\nit is relative to perception. Further, if it is legitimate to ask,\n'How can we apprehend what does not belong to our being?' it is\nequally legitimate to ask, 'How can we apprehend what does belong to\nour own being?' It is wholly arbitrary to limit the question to the\none kind of reality. If a question is to be put at all, it should take\nthe form, 'How is it possible to apprehend anything?' But this\nquestion has only to be put to be discarded. For it amounts to a\ndemand to _explain_ knowledge; and any answer to it would involve the\nderivation of knowledge from what was not knowledge, a task which must\nbe as impossible as the derivation of space from time or of colour\nfrom sound. Knowledge is _sui generis_, and, as such, cannot be\nexplained.[6]\n\n    [5] Cf. pp. 89-91.\n\n    [6] This assertion, being self-evident, admits of no direct\n    proof. A 'proof' can only take the form of showing that\n    any supposed 'derivation' or 'explanation' of knowledge\n    presupposes knowledge in that from which it derives it.\n    Professor Cook Wilson has pointed out that we must understand\n    what knowing is in order to explain anything at all, so\n    that any proposed explanation of knowing would necessarily\n    presuppose that we understood what knowing is. For the\n    general doctrine, cf. p. 245.\n\nMoreover, it may be noted that the support which this form of idealism\nsometimes receives from an argument which uses the terms 'inside' and\n'outside' the mind is unmerited. At first sight it seems a refutation\nof the plain man's view to argue thus: 'The plain man believes the\nspatial world to exist whether any one knows it or not. Consequently,\nhe allows that the world is outside the mind. But, to be known, a\nreality must be inside the mind. Therefore, the plain man's view\nrenders knowledge impossible.' But, as soon as it is realized that\n'inside the mind' and 'outside the mind' are metaphors, and,\ntherefore, must take their meaning from their context, it is easy to\nsee that the argument either rests on an equivocation or assumes the\npoint at issue. The assertion that the world is outside the mind,\nbeing only a metaphorical expression of the plain man's view, should\nonly mean that the world is something independent of the mind, as\nopposed to something inside the mind, in the sense of dependent upon\nit, or mental. But the assertion that, to be known, a reality must be\ninside the mind, if it is to be incontestably true, should only mean\nthat a reality, to be apprehended, must really be object of\napprehension. And in this case 'being inside the mind', since it only\nmeans 'being object of apprehension', is not the opposite of 'being\noutside the mind' in the previous assertion. Hence, on this\ninterpretation, the second assertion is connected with the first only\napparently and by an equivocation; there is really no argument at all.\nIf, however, the equivocation is to be avoided, 'inside the mind' in\nthe second assertion must be the opposite of 'outside the mind' in the\nfirst, and consequently the second must mean that a reality, to be\nknown, must be dependent on the mind, or mental. But in that case the\nobjection to the plain man's view is a _petitio principii_, and not an\nargument.\n\nNevertheless, the tendency to think that the only object or, at least,\nthe only direct object of the mind is something mental still requires\nexplanation. It seems due to a tendency to treat self-consciousness as\nsimilar to consciousness of the world. When in reflection we turn our\nattention away from the world to the activity by which we come to know\nit, we tend to think of our knowledge of the world as a reality to be\napprehended similar to the world which we apprehended prior to\nreflection. We thereby implicitly treat this knowledge as something\nwhich, like the world, merely _is_ and is not the knowledge of\nanything; in other words, we imply that, so far from being knowledge,\ni. e. the knowing of a reality, it is precisely that which we\ndistinguish from knowledge, viz. a reality to be known,\nalthough--since knowledge must be mental--we imply that it is a\nreality of the special kind called mental. But if the knowledge upon\nwhich we reflect is thus treated as consisting in a mental reality\nwhich merely _is_, it is implied that in this knowledge the world is\nnot, at any rate directly, object of the mind, for _ex hypothesi_ a\nreality which merely _is_ and is not the knowledge of anything has no\nobject. Hence it comes to be thought that the only object or, at\nleast, the only direct object of the mind is this mental reality\nitself, which is the object of reflection; in other words, that the\nonly immediate object of the mind comes to be thought of as its own\nidea. The root of the mistake lies in the initial supposition--which,\nit may be noted, seems to underlie the whole treatment of knowledge by\nempirical psychology--that knowledge can be treated as a reality to be\napprehended, in the way in which any reality which is not knowledge is\na reality to be apprehended.\n\nWe may now revert to that form of idealism which maintains that the\nessential relation of reality to the mind is that of _being known_,\nin order to consider two lines of argument by which it may be\ndefended.\n\nAccording to the first of these, the view of the plain man either is,\nor at least involves, materialism; and materialism is demonstrably\nabsurd. The plain man's view involves the existence of the physical\nworld prior to the existence of the knowledge of it, and therefore\nalso prior to the existence of minds which know it, since it is\nimpossible to separate the existence of a knowing mind from its actual\nknowledge. From this it follows that mere matter, having only the\nqualities considered by the physicist, must somehow have originated or\nproduced knowing and knowing minds. But this production is plainly\nimpossible. For matter, possessing solely, as it does, characteristics\nbound up with extension and motion, cannot possibly have originated\nactivities of a wholly different kind, or beings capable of exercising\nthem.\n\nIt may, however, be replied that the supposed consequence, though\nabsurd, does not really follow from the plain man's realism.\nDoubtless, it would be impossible for a universe consisting solely of\nthe physical world to originate thought or beings capable of thinking.\nBut the real presupposition of the coming into existence of human\nknowledge at a certain stage in the process of the universe is to be\nfound in the pre-existence, not of a mind or minds which always\nactually knew, but simply of a mind or minds in which, under certain\nconditions, knowledge is necessarily actualized. A mind cannot be the\nproduct of anything or, at any rate, of anything but a mind. It cannot\nbe a new reality introduced at some time or other into a universe of\nrealities of a wholly different order. Therefore, the presupposition\nof the present existence of knowledge is the pre-existence of a mind\nor minds; it is not implied that its or their knowledge must always\nhave been actual. In other words, knowing implies the ultimate or\nunoriginated existence of beings possessed of the capacity to know.\nOtherwise, knowledge would be a merely derivative product, capable of\nbeing stated in terms of something else, and in the end in terms of\nmatter and motion. This implication is, however, in no wise traversed\nby the plain man's realism. For that implies, not that the existence\nof the physical world is prior to the existence of a mind, but only\nthat it is prior to a mind's actual knowledge of the world.\n\nThe second line of thought appeals to the logic of relation. It may be\nstated thus. If a term is relative, i. e. is essentially 'of' or\nrelative to another, that other is essentially relative to it. Just as\na doctor, for instance, is essentially a doctor of a patient, so a\npatient is essentially the patient of a doctor. As a ruler implies\nsubjects, so subjects imply a ruler. As a line essentially has points\nat its ends, so points are essentially ends of a line. Now knowledge\nis essentially 'of' or relative to reality. Reality, therefore, is\nessentially relative to or implies the knowledge of it. And this\ncorrelativity of knowledge and reality finds linguistic confirmation\nin the terms 'subject' and 'object'. For, linguistically, just as a\nsubject is always the subject of an object, so an object is always the\nobject of a subject.\n\nNevertheless, further analysis of the nature of relative terms, and in\nparticular of knowledge, does not bear out this conclusion. To take\nthe case of a doctor. It is true that if some one is healing, some one\nelse is receiving treatment, i. e. is being healed; and 'patient'\nbeing the name for the recipient of treatment, we can express this\nfact by saying that a doctor is essentially the doctor of a patient.\nFurther, it is true that a recipient of treatment implies a giver of\nit, as much as a giver of it implies a recipient. Hence we can truly\nsay that since a doctor is the doctor of a patient, a patient is the\npatient of a doctor, meaning thereby that since that to which a doctor\nis relative is a patient, a patient must be similarly relative to a\ndoctor. There is, however, another statement which can be made\nconcerning a doctor. We can say that a doctor is a doctor of a human\nbeing who is ill, i. e. a sick man. But in this case we cannot go on\nto say that since a doctor is a doctor of a sick man, a sick man\nimplies or is relative to a doctor. For we mean that the kind of\nreality capable of being related to a doctor as his patient is a sick\nman; and from this it does not follow that a reality of this kind does\nstand in this relation. Doctoring implies a sick man; a sick man does\nnot imply that some one is treating him. We can only say that since a\ndoctor is the doctor of a sick man, a sick man implies the possibility\nof doctoring. In the former case the terms, viz. 'doctor' and\n'patient', are inseparable because they signify the relation in\nquestion in different aspects. The relation is one fact which has two\ninseparable 'sides', and, consequently, the terms must be inseparable\nwhich signify the relation respectively from the point of view of the\none side and from the point of view of the other. Neither term\nsignifies the nature of the elements which can stand in the relation.\nIn the latter case, however, the terms, viz. 'doctor' and 'sick man',\nsignify respectively the relation in question (in one aspect), and the\nnature of one of the elements capable of entering into it;\nconsequently they are separable.\n\nNow when it is said that knowledge is essentially knowledge of\nreality, the statement is parallel to the assertion that a doctor is\nessentially the doctor of a sick man, and not to the assertion that a\ndoctor is essentially the doctor of a patient. It should mean that\nthat which is capable of being related to a knower as his object is\nsomething which is or exists; consequently it cannot be said that\nsince knowledge is of reality, reality must essentially be known. The\nparallel to the assertion that a doctor is the doctor of a patient is\nthe assertion that knowledge is the knowledge of an object; for just\nas 'patient' means that which receives treatment from a doctor, so\n'object' means that which is known. And here we _can_ go on to make\nthe further parallel assertion that since knowledge is essentially the\nknowledge of an object, an object is essentially an object of\nknowledge. Just as 'patient' means a recipient of treatment, or, more\naccurately, a sick man under treatment, so 'object' means something\nknown, or, more accurately, a reality known. And 'knowledge' and\n'object of knowledge', like 'doctor' and 'patient', indicate the same\nrelation, though from different points of view, and, consequently,\nwhen we can use the one term, we can use the other. But to say that an\nobject (i. e. a reality known) implies the knowledge of it is not to\nsay that reality implies the knowledge of it, any more than to say\nthat a patient implies a doctor is to say that a sick man implies a\ndoctor.\n\nBut a doctor, it might be objected, is not a fair parallel to\nknowledge or a knower. A doctor, though an instance of a relative\nterm, is only an instance of one kind of relative term, that in which\nthe elements related are capable of existing apart from the relation,\nthe relation being one in which they can come to stand and cease to\nstand. But there is another kind of relative term, in which the\nelements related presuppose the relation, and any thought of these\nelements involves the thought of the relation. A universal, e. g.\nwhiteness, is always the universal of certain individuals, viz.\nindividual whites; an individual, e. g. this white, is always an\nindividual of a universal, viz. whiteness. A genus is the genus of a\nspecies, and vice versa. A surface is the surface of a volume, and a\nvolume implies a surface. A point is the end of a line, and a line is\nbounded by points. In such cases the very being of the elements\nrelated involves the relation, and, apart from the relation,\ndisappears. The difference between the two kinds of relative terms can\nbe seen from the fact that only in the case of the former kind can two\nelements be found of which we can say significantly that their\nrelation is of the kind in question. We can say of two men that they\nare related as doctor and patient, or as father and son, for we can\napprehend two beings as men without being aware of them as so related.\nBut of no two elements is it possible to say that their relation is\nthat of universal and individual, or of genus and species, or of\nsurface and volume; for to apprehend elements which are so related we\nmust apprehend them so related.[7] To apprehend a surface is to\napprehend a surface of a volume. To apprehend a volume is to apprehend\na volume bounded by a surface. To apprehend a universal is to\napprehend it as the universal of an individual, and vice versa.[8] In\nthe case of relations of this kind, the being of either element which\nstands in the relation is relative to that of the other; neither can\nbe real without the other, as we see if we try to think of one without\nthe other. And it is at least possible that knowledge and reality or,\nspeaking more strictly, a knower and reality, are related in this way.\n\n    [7] It is, of course, possible to say significantly that two\n    elements, A and B, are related as universal and individual,\n    or as surface and volume, if we are trying to explain what we\n    mean by 'universal and individual' or 'surface and volume';\n    but in that case we are elucidating the relationship through\n    the already known relation of A and B, and are not giving\n    information about the hitherto unknown relation of A and B.\n\n    [8] Professor Cook Wilson has pointed out that the\n    distinction between these two kinds of relation is marked\n    in language in that, for instance, while we speak of the\n    'relation _of_ universal _and_ individual', we speak of\n    'the relation _between_ one man _and_ another', or of 'the\n    relation _of_ one man _to_ another', using, however, the\n    phrase 'the relation _of_ doctor _and_ patient', when we\n    consider two men only as in that relation.\n\n    I owe to him recognition of the fact that the use of the word\n    'relation' in connexion with such terms as 'universal and\n    individual' is really justified.\n\nWhat is, however, at least a strong presumption against this view is\nto be found in the fact that while relations of the second kind are\nessentially non-temporal, the relation of knowing is essentially\ntemporal. The relation of a universal and its individuals, or of a\nsurface and the volume which it bounds, does not either come to be, or\npersist, or cease. On the other hand, it is impossible to think of a\nknowing which is susceptible of no temporal predicates and is not\nbound up with a process; and the thought of knowing as something which\ncomes to be involves the thought that the elements which become thus\nrelated exist independently of the relation. Moreover, the real\nrefutation of the view lies in the fact that, when we consider what we\nreally think, we find that we think that the relation between a knower\nand reality is not of the second kind. If we consider what we mean by\n'a reality', we find that we mean by it something which is not\ncorrelative to a mind knowing it. It does not mean something the\nthought of which disappears with the thought of a mind actually\nknowing it, but something which, though it can be known by a mind,\nneed not be actually known by a mind. Again, just as we think of a\nreality as something which _can_ stand as object in the relation of\nknowledge, without necessarily being in this relation, so, as we see\nwhen we reflect, we think of a knowing mind as something which _can_\nstand as subject in this relation without necessarily being in the\nrelation. For though we think of the capacities which constitute the\nnature of a knowing mind as only recognized through their\nactualizations, i. e. through actual knowing, we think of the mind\nwhich is possessed of these capacities as something apart from their\nactualization.\n\nIt is now possible to direct attention to two characteristics of\nperception and knowledge with which Kant's treatment of space and time\nconflicts, and the recognition of which reveals his procedure in its\ntrue light.\n\nIt has been already urged that both knowledge and perception--which,\nthough not identical with knowledge, is presupposed by it--are\nessentially of _reality_. Now, in the _first_ place, it is thereby\nimplied that the relation between the mind and reality in knowledge or\nin perception is essentially direct, i. e. that there is no _tertium\nquid_ in the form of an 'idea' or a 'representation' between us as\nperceiving or knowing and what we perceive or know. In other words, it\nis implied that Locke's view is wrong in principle, and, in fact, the\ncontrary of the truth. In the _second_ place, it is implied that while\nthe whole fact of perception includes the reality perceived and the\nwhole fact of knowledge includes the reality known, since both\nperception and knowledge are 'of', and therefore inseparable from a\nreality, yet the reality perceived or known is essentially distinct\nfrom, and cannot be stated in terms of, the perception or the\nknowledge. Just as neither perception nor knowledge can be stated in\nterms of the reality perceived or known from which they are\ndistinguished, so the reality perceived or known cannot be stated in\nterms of the perception or the knowledge. In other words, the terms\n'perception' and 'knowledge' ought to stand for the activities of\nperceiving and knowing respectively, and not for the reality perceived\nor known. Similarly, the terms 'idea' and 'representation'--the latter\nof which has been used as a synonym for Kant's _Vorstellung_--ought to\nstand not for something thought of or represented, but for the act of\nthinking or representing.\n\nFurther, this second implication throws light on the proper meaning of\nthe terms 'form of perception' and 'form of knowledge or of thought'.\nFor, in accordance with this implication, a 'form of perception' and a\n'form of knowledge' ought to refer to the nature of our acts of\nperceiving and knowing or thinking respectively, and not to the nature\nof the realities perceived or known. Consequently, Kant was right in\nmaking the primary antithesis involved in the term 'form of\nperception' that between a way in which we perceive and a way in which\nthings are, or, in other words, between a characteristic of our\nperceiving nature and a characteristic of the reality perceived.\nMoreover, Kant was also right in making this distinction a real\nantithesis and not a mere distinction within one and the same thing\nregarded from two points of view. That which is a form of perception\ncannot also be a form of the reality and vice versa. Thus we may\nillustrate a perceived form of perception by pointing out that our\napprehension of the physical world (1) is a temporal process, and (2)\nis conditioned by perspective. Both the succession and the conditions\nof perspective belong to the act of perception, and do not form part\nof the nature of the world perceived. And it is significant that in\nour ordinary consciousness it never occurs to us to attribute either\nthe perspective or the time to the reality perceived. Even if it be\ndifficult in certain cases, as in that of colour, to decide whether\nsomething belongs to our act of perception or not, we never suppose\nthat it can be _both_ a form of perception _and_ a characteristic of\nthe reality perceived. We think that if it be the one, it cannot be\nthe other.\n\nMoreover, if we pass from perception to knowledge or thought--which in\nthis context may be treated as identical--and seek to illustrate a\nform of knowledge or of thought, we may cite the distinction of\nlogical subject and logical predicate of a judgement. The distinction\nas it should be understood--for it does not necessitate a difference\nof grammatical form--may be illustrated by the difference between the\njudgements 'Chess is the _most trying of games_' and '_Chess_ is the\nmost trying of games'. In the former case 'chess' is the logical\nsubject, in the latter case it is the logical predicate. Now this\ndistinction clearly does not reside in or belong to the reality about\nwhich we judge; it relates solely to the order of our approach in\nthought to various parts of its nature. For, to take the case of the\nformer judgement, in calling 'chess' its subject, and 'most trying of\ngames' its predicate, we are asserting that in this judgement we begin\nby apprehending the reality of which we are thinking as chess, and\ncome to apprehend it as the most trying of games. In other words, the\ndistinction relates solely to the order of our apprehension, and not\nto anything in the thing apprehended.\n\nIn view of the preceding, it is possible to make clear the nature of\ncertain mistakes on Kant's part. In the first place, space, and time\nalso, so far as we are thinking of the world, and not of our\napprehension of it, as undergoing a temporal process, are essentially\ncharacteristics not of perception but of the reality perceived, and\nKant, in treating space, and time, so regarded, as forms of\nperception, is really transferring to the perceiving subject that\nwhich in the whole fact 'perception of an object' or 'object\nperceived' belongs to the object.\n\nAgain, if we go on to ask how Kant manages to avoid drawing the\nconclusion proper to this transference, viz. that space and time are\nnot characteristics of any realities at all, but belong solely to the\nprocess by which we come to apprehend them, we see that he does so\nbecause, in effect, he contravenes both the characteristics of\nperception referred to. For, in the first place, although in\nconformity with his theory he almost always _speaks_ of space and time\nin terms of perception,[9] he consistently _treats_ them as features\nof the reality perceived, i. e. of phenomena. Thus in arguing that\nspace and time belong not to the understanding but to the sensibility,\nalthough he uniformly speaks of them as perceptions, his argument\nimplies that they are objects of perception; for its aim, properly\nstated, is to show that space and time are not objects of thought but\nobjects of perception. Consequently, in his treatment of space and\ntime, he refers to what are both to him and in fact objects of\nperception in terms of perception, and thereby contravenes the second\nimplication of perception to which attention has been drawn. Again, in\nthe second place, if we go on to ask how Kant is misled into doing\nthis, we see that it is because he contravenes the first implication\nof perception. In virtue of his theory of perception[10] he interposes\na _tertium quid_ between the reality perceived and the percipient, in\nthe shape of an 'appearance'. This _tertium quid_ gives him something\nwhich can plausibly be regarded as at once a perception and something\nperceived. For, though from the point of view of the thing in itself\nan appearance is an appearance or a perception of it, yet, regarded\nfrom the point of view of what it is in itself, an appearance is a\nreality perceived of the kind called mental. Hence space and time,\nbeing characteristics of an appearance, can be regarded as at once\ncharacteristics of our perception of a reality, viz. of a thing in\nitself, and characteristics of a reality perceived, viz. an\nappearance. Moreover, there is another point of view from which the\ntreatment of bodies in space as appearances or phenomena gives\nplausibility to the view that space, though a form of perception, is a\ncharacteristic of a reality. When Kant speaks of space as the form of\nphenomena the fact to which he refers is that all bodies are\nspatial.[11] He means, not that space is a way in which we perceive\nsomething, but that it is a characteristic of things perceived, which\nhe _calls_ phenomena, and which _are_ bodies. But, since in his\nstatement of this fact he substitutes for bodies phenomena, which to\nhim are perceptions, his statement can be put in the form 'space is\n_the form of perceptions_'; and the statement in this form is verbally\nalmost identical with the statement that space is _a form of\nperception_. Consequently, the latter statement, which _should_ mean\nthat space is a way in which we perceive things, is easily identified\nwith a statement of which the meaning is that space is a\ncharacteristic of something perceived.[12]\n\n    [9] Cf. p. 51, note 1.\n\n    [10] Cf. p. 30 and ff.\n\n    [11] Cf. p. 39.\n\n    [12] It can be shown in the same way, _mutatis mutandis_ (cp.\n    p. 111), that the view that time, though the form of inner\n    perception, is a characteristic of a reality gains\n    plausibility from Kant's implicit treatment of our states as\n    appearances due to ourselves.\n\nAgain, Kant's account of time will be found to treat something\nrepresented or perceived as also a perception. We find two consecutive\nparagraphs[13] of which the aim is apparently to establish the\ncontrary conclusions: (1) that time is only the form of our internal\nstate and not of external phenomena, and (2) that time is the formal\ncondition of all phenomena, external and internal.\n\n    [13] B. 49-50 (b) and (c), M. 30 (b) and (c).\n\nTo establish the first conclusion, Kant argues that time has nothing\nto do with shape or position, but, on the contrary, determines the\nrelation of representations in our internal state. His meaning is that\nwe have a succession of perceptions or representations of bodies in\nspace,[14] and that while the bodies perceived are not related\ntemporally, our perceptions or representations of them are so\nrelated. Here 'representations' refers to our apprehension, and\nis distinguished from what is represented, viz. bodies in space.\n\n    [14] Kant here refers to bodies by the term 'phenomena', but\n    their character as phenomena is not relevant to his argument.\n\nHow, then, does Kant reach the second result? He remembers that bodies\nin space are 'phenomena', i. e. representations. He is, therefore,\nable to point out that all representations belong, as determinations\nof the mind, to our internal state, whether they have external things,\ni. e. bodies in space, for their objects or not, and that,\nconsequently, they are subject to time. Hence time is concluded to be\nthe form of all phenomena. In this second argument, however, it is\nclear that Kant has passed from his previous treatment of bodies in\nspace as something represented or perceived to the treatment of them\nas themselves representations or perceptions.[15]\n\n    [15] It may be noted that Kant's assertion (B. 50, M. 31)\n    that time is the immediate condition of internal phenomena,\n    and thereby also mediately the condition of external\n    phenomena, does not help to reconcile the two positions.\n\nIn conclusion, we may point out an insoluble difficulty in Kant's\naccount of time. His treatment of space and time as the forms of outer\nand inner sense respectively implies that, while spatial relations\napply to the realities which we perceive, temporal relations apply\nsolely to our perceptions of them. Unfortunately, however, as Kant in\ncertain contexts is clearly aware, time also belongs to the realities\nperceived. The moon, for instance, moves round the earth. Thus there\nare what may be called real successions as well as successions in our\nperception. Further, not only are we aware of this distinction in\ngeneral, but in particular cases we succeed in distinguishing a\nsuccession of the one kind from a succession of the other. Yet from\nKant's standpoint it would be impossible to distinguish them in\nparticular cases, and even to be aware of the distinction in general.\nFor the distinction is possible only so long as a distinction is\nallowed between our perceptions and the realities perceived. But for\nKant this distinction has disappeared, for in the end the realities\nperceived are merely our perceptions; and time, if it be a\ncharacteristic of anything, must be a characteristic only of our\nperceptions.\n\n\n\n\nCHAPTER VII\n\nTHE METAPHYSICAL DEDUCTION OF THE CATEGORIES\n\n\nThe aim of the _Aesthetic_ is to answer the first question of the\n_Critique_ propounded in the Introduction, viz. 'How is pure\nmathematics possible?'[1] The aim of the _Analytic_ is to answer the\nsecond question, viz. 'How is pure natural science possible?' It has\npreviously[2] been implied that the two questions are only verbally of\nthe same kind. Since Kant thinks of the judgements of mathematics as\nself-evident, and therefore as admitting of no reasonable doubt[3],\nhe takes their truth for granted. Hence the question, 'How is pure\nmathematics possible?' means 'Granted the truth of mathematical\njudgements, what inference can we draw concerning the nature of the\nreality to which they relate?'; and the inference is to proceed from\nthe truth of the judgements to the nature of the reality to which\nthey relate. Kant, however, considers that the principles underlying\nnatural science, of which the law of causality is the most prominent,\nare not self-evident, and consequently need proof.[4] Hence, the\nquestion, 'How is pure natural science possible?' means 'What\njustifies the assertion that the presuppositions of natural science\nare true?' and the inference is to proceed from the nature of the\nobjects of natural science to the truth of the _a priori_ judgements\nwhich relate to them.\n\n    [1] B. 20, M. 13.\n\n    [2] pp. 23-5.\n\n    [3] Cf. p. 24, note 1.\n\n    [4] Cf. p. 24, notes 2 and 3.\n\nAgain, as Kant rightly sees, the vindication of the presuppositions\nof natural science, to be complete, requires the discovery upon a\ndefinite principle of _all_ these presuppositions. The clue to this\ndiscovery he finds in the view that, just as the perceptions of space\nand time originate in the sensibility, so the _a priori_ conceptions\nand laws which underlie natural science originate in the\nunderstanding; for, on this view, the discovery of all the conceptions\nand laws which originate in the understanding will be at the same time\nthe discovery of all the presuppositions of natural science.\n\nKant therefore in the _Analytic_ has a twofold problem to solve.\nHe has firstly to discover the conceptions and laws which belong to\nthe understanding as such, and secondly to vindicate their application\nto individual things. Moreover, although it is obvious that the\nconceptions and the laws of the understanding must be closely\nrelated,[5] he reserves them for separate treatment.\n\n    [5] E. g. the conception of 'cause and effect', and the law\n    that 'all changes take place according to the law of the\n    connexion between cause and effect'.\n\nThe _Analytic_ is accordingly subdivided into the _Analytic of\nConceptions_ and the _Analytic of Principles_. The _Analytic of\nConceptions_, again, is divided into the _Metaphysical Deduction of\nthe Categories_, the aim of which is to discover the conceptions\nof the understanding, and the _Transcendental Deduction of the\nCategories_, the aim of which is to vindicate their validity,\ni. e. their applicability to individual things.\n\nIt should further be noticed that, according to Kant, it is the\nconnexion of the _a priori_ conceptions and laws underlying natural\nscience with the _understanding_ which constitutes the main difficulty\nof the vindication of their validity, and renders necessary an answer\nof a different kind to that which would have been possible, if the\nvalidity of mathematical judgements had been in question.\n\n\"We have been able above, with little trouble, to make comprehensible\nhow the conceptions of space and time, although _a priori_ knowledge,\nmust necessarily relate to objects and render possible a synthetic\nknowledge of them independently of all experience. For since an object\ncan appear to us, i. e. be an object of empirical perception, only by\nmeans of such pure forms of sensibility, space and time are pure\nperceptions, which contain _a priori_ the condition of the possibility\nof objects as phenomena, and the synthesis in space and time has\nobjective validity.\"\n\n\"On the other hand, the categories of the understanding do not\nrepresent the conditions under which objects are given in perception;\nconsequently, objects can certainly appear to us without their\nnecessarily being related to functions of the understanding, and\ntherefore without the understanding containing _a priori_ the\nconditions of these objects. Hence a difficulty appears here, which\nwe did not meet in the field of sensibility, viz. how _subjective\nconditions of thought_ can have _objective validity_, i. e. can\nfurnish conditions of the possibility of all knowledge of objects;\nfor phenomena can certainly be given us in perception without the\nfunctions of the understanding. Let us take, for example, the\nconception of cause, which indicates a peculiar kind of synthesis in\nwhich on A something entirely different B is placed[6] according to a\nlaw. It is not _a priori_ clear why phenomena should contain something\nof this kind ... and it is consequently doubtful _a priori_, whether\nsuch a conception is not wholly empty, and without any corresponding\nobject among phenomena. For that objects of sensuous perception must\nconform to the formal conditions of the sensibility which lie _a\npriori_ in the mind is clear, since otherwise they would not be\nobjects for us; but that they must also conform to the conditions\nwhich the understanding requires for the synthetical unity of thought\nis a conclusion the cogency of which it is not so easy to see. For\nphenomena might quite well be so constituted that the understanding\ndid not find them in conformity with the conditions of its unity, and\neverything might lie in such confusion that, e. g. in the succession\nof phenomena, nothing might present itself which would offer a rule of\nsynthesis, and so correspond to the conception of cause and effect, so\nthat this conception would be quite empty, null, and meaningless.\nPhenomena would none the less present objects to our perception, for\nperception does not in any way require the functions of thinking.\"[7]\n\n    [6] _Gesetzt._\n\n    [7] B. 121-3, M. 75-6.\n\nThis passage, if read in connexion with that immediately preceding\nit,[8] may be paraphrased as follows: 'The argument of the _Aesthetic_\nassumes the validity of mathematical judgements, which as such relate\nto space and time, and thence it deduces the phenomenal character of\nspace and time, and of what is contained therein. At the same time the\npossibility of questioning the validity of the law of causality, and\nof similar principles, may lead us to question even the validity of\nmathematical judgements. In the case of mathematical judgements,\nhowever, in consequence of their relation to perception, an answer is\nreadily forthcoming. We need only reverse the original argument and\nappeal directly to the phenomenal character of space and time and of\nwhat is contained in them. Objects in space and time, being\nappearances, must conform to the laws according to which we have\nappearances; and since space and time are only ways in which we\nperceive, or have appearances, mathematical laws, which constitute the\ngeneral nature of space and time, are the laws according to which we\nhave appearances. Mathematical laws, then, constitute the general\nstructure of appearances, and, as such, enter into the very being\nof objects in space and time. But the case is otherwise with the\nconceptions and principles underlying natural science. For the law of\ncausality, for instance, is a law not of our perceiving but of our\nthinking nature, and consequently it is not presupposed in the\npresentation to us of objects in space and time. Objects in space\nand time, being appearances, need conform only to the laws of our\nperceiving nature. We have therefore to explain the possibility of\nsaying that a law of our thinking nature must be valid for objects\nwhich, as conditioned merely by our perceiving nature, are independent\nof the laws of our thinking; for phenomena might be so constituted as\nnot to correspond to the necessities of our thought.'\n\n    [8] B. 120-1, M. 73-4.\n\nNo doubt Kant's _solution_ of this problem in the _Analytic_ involves\nan emphatic denial of the central feature of this statement of it,\nviz. that phenomena may be given in perception without any help from\nthe activity of the understanding.[9] Hence it may be urged that this\npassage merely expresses a temporary aberration on Kant's part, and\nshould therefore be ignored. Nevertheless, in spite of this\ninconsistency, the view that phenomena may be given in perception\nwithout help from the activity of the understanding forms the basis\nof the difference of treatment which Kant thinks necessary for the\nvindication of the judgements underlying natural science and for that\nof the judgements of mathematics.\n\n    [9] Cf. B. 137-8, M. 85, and B. 160 note, M. 98 note.\n\nWe may now consider how Kant 'discovers' the categories or conceptions\nwhich belong to the understanding as such.[10] His method is sound in\nprinciple. He begins with an account of the understanding in general.\nHe then determines its essential differentiations. Finally, he argues\nthat each of these differentiations involves a special conception,\nand that therefore these conceptions taken together constitute an\nexhaustive list of the conceptions which belong to the understanding.\n\n    [10] B. 91-105, M. 56-63.\n\nHis account of the understanding is expressed thus: \"The understanding\nwas explained above only negatively, as a non-sensuous faculty of\nknowledge. Now, independently of sensibility, we cannot have any\nperception; consequently, the understanding is no faculty of\nperception. But besides perception there is no other kind of\nknowledge, except through conceptions. Consequently, the knowledge of\nevery understanding, or at least of every human understanding, is a\nknowledge through conceptions,--not perceptive, but discursive. All\nperceptions, as sensuous, depend on affections; conceptions,\ntherefore, upon functions. By the word function, I understand the\nunity of the act of arranging different representations under one\ncommon representation. Conceptions, then, are based on the spontaneity\nof thinking, as sensuous perceptions are on the receptivity of\nimpressions. Now the understanding cannot make any other use of these\nconceptions than to judge by means of them. Since no representation,\nexcept only the perception, refers immediately to the object, a\nconception is never referred immediately to an object, but to some\nother representation thereof, be that a perception or itself a\nconception. A judgement, therefore, is the mediate knowledge of an\nobject, consequently the representation of a representation of it. In\nevery judgement there is a conception which is valid for many\nrepresentations, and among these also comprehends a given\nrepresentation, this last being then immediately referred to the\nobject. For example, in the judgement 'All bodies are divisible', our\nconception of the divisible refers to various other conceptions; among\nthese, however, it is herein particularly referred to the conception\nof body, and this conception of body is referred to certain phenomena\nwhich present themselves to us. These objects, therefore, are\nmediately represented by the conception of divisibility. Accordingly,\nall judgements are functions of unity in our representations, since,\ninstead of an immediate, a higher representation, which comprehends\nthis and several others, is used for the knowledge of the object, and\nthereby many possible items of knowledge are collected into one. But\nwe can reduce all acts of the understanding to judgements, so that the\n_understanding_ in general can be represented as a _faculty of\njudging_.\"[11]\n\n    [11] B. 92-4, M. 56-7.\n\nIt is not worth while to go into all the difficulties of this confused\nand artificial passage. Three points are clear upon the surface. In\nthe first place, the account of the understanding now given differs\nfrom that given earlier in the _Critique_[12] in that, instead of\nmerely distinguishing, it separates the sensibility and the\nunderstanding, and treats them as contributing, not two inseparable\nfactors involved in all knowledge, but two kinds of knowledge. In the\nsecond place, the guise of argument is very thin, and while Kant\nostensibly _proves_, he really only _asserts_ that the understanding\nis the faculty of judgement. In the third place, in describing\njudgement Kant is hampered by trying to oppose it as the mediate\nknowledge of an object to perception as the immediate knowledge of an\nobject. A perception is said to relate immediately to an object; in\ncontrast with this, a conception is said to relate immediately only to\nanother conception or to a perception, and mediately to an object\nthrough relation to a perception, either directly or through another\nconception. Hence a judgement, as being the use of a conception, viz.\nthe predicate of the judgement, is said to be the mediate knowledge of\nan object. But if this distinction be examined, it will be found that\ntwo kinds of immediate relation are involved, and that the account of\nperception is not really compatible with that of judgement. When a\nperception is said to relate immediately to an object, the relation in\nquestion is that between a sensation or appearance produced by an\nobject acting upon or affecting the sensibility and the object which\nproduces it. But when a conception is said to relate immediately to\nanother conception or to a perception, the relation in question is\nthat of universal and particular, i. e. that of genus and species or\nof universal and individual. For the conception is said to be 'valid\nfor' (i. e. to 'apply to') and to 'comprehend' the conception or\nperception to which it is immediately related; and again, when a\nconception is said to relate mediately to an object, the relation\nmeant is its 'application' to the object, even though in this case\nthe application is indirect. Now if a perception to which a\nconception is related--either directly or indirectly through another\nconception--were an appearance produced by an object, the conception\ncould never be related to the object in the sense required, viz. that\nit applies to it; for an appearance does not _apply to_ but is\n_produced_ by the object. Consequently, when Kant is considering a\nconception, and therefore also when he is considering a judgement,\nwhich is the use of a conception, he is really thinking of the\nperception to which it is related as an _object of_ perception, i. e.\nas a perceived individual, and he has ceased to think of a perception\nas an appearance produced by an object.[13] Hence in considering\nKant's account of a conception and of judgement, we should ignore his\naccount of perception, and therefore also his statement that judgement\nis the mediate knowledge of an object.\n\n    [12] B. 74-6, M. 45-6.\n\n    [13] Kant, in _illustrating_ the nature of a judgement,\n    evades the difficulty occasioned by his account of\n    perception, by illustrating a 'perception' by the 'conception\n    of body', and 'objects' by 'certain phenomena'. He thereby\n    covertly substitutes the relation of universal and individual\n    for the relation of an appearance and the object which causes\n    it.\n\nIf we do so, we see that Kant's account of judgement simply amounts to\nthis: 'Judgement is the use of a conception or 'universal'; the use of\na conception or universal consists in bringing under it corresponding\nindividuals or species. Consequently, judgement is a function\nproducing unity. If, for instance, we judge 'All bodies are\ndivisible', we thereby unify 'bodies' with other kinds of divisible\nthings by bringing them under the conception of divisibility; and if\nwe judge 'This body is divisible' we thereby unify this divisible\nbody with others by bringing it and them under the conception of\ndivisibility.'[14] Again, since 'the understanding in general can be\nrepresented as a _faculty of judging_', it follows that the activity\nof the understanding consists in introducing unity into our\nrepresentations, by bringing individuals or species--both these being\nrepresentations--under the corresponding universal or conception.[15]\n\n    [14] It is not Kant's general account of judgement given in\n    this passage, but the account of perception incompatible with\n    it, which leads him to confine his illustrations to universal\n    judgements.\n\n    [15] We may note three minor points. (1) Kant's definition of\n    function as 'the unity of the act of arranging [i. e. the act\n    which produces unity by arranging] different representations\n    under a common representation' has no justification in its\n    immediate context, and is occasioned solely by the\n    forthcoming description of judgement. (2) Kant has no right\n    to distinguish the activity which _originates_ conceptions,\n    or upon which they depend, from the activity which _uses_\n    conceptions, viz. judgement. For the act of arranging diverse\n    representations under a common representation which\n    originates conceptions is the act of judgement as Kant\n    describes it. (3) It is wholly artificial to speak of\n    judgement as 'the representation of a representation of an\n    object'.\n\nHaving explained the nature of the understanding, Kant proceeds to\ntake the next step. His aim being to connect the understanding with\nthe categories, and the categories being a plurality, he has to show\nthat the activity of judgement can be differentiated into several\nkinds, each of which must subsequently be shown to involve a special\ncategory. Hence, solely in view of the desired conclusion, and in\nspite of the fact that he has described the activity of judgement as\nif it were always of the same kind, he passes in effect from the\nsingular to the plural and asserts that 'all the functions of the\nunderstanding can be discovered, when we can completely exhibit the\nfunctions of unity in judgements'. After this preliminary transition,\nhe proceeds to assert that, if we abstract in general from all content\nof a judgement and fix our attention upon the mere form of the\nunderstanding, we find that the function of thinking in a judgement\ncan be brought under four heads, each of which contains three\nsubdivisions. These, which are borrowed with slight modifications from\nFormal Logic, are expressed as follows.[16]\n\n    I. _Quantity._\n       Universal\n       Particular\n       Singular.\n\n   II. _Quality._\n       Affirmative\n       Negative\n       Infinite.\n\n  III. _Relation._\n       Categorical\n       Hypothetical\n       Disjunctive.\n\n   IV. _Modality._\n       Problematic\n       Assertoric\n       Apodeictic.\n\nThese distinctions, since they concern only the form of judgements,\nbelong, according to Kant, to the activity of judgement as such, and\nin fact constitute its essential differentiations.\n\n    [16] B. 95, M. 58.\n\nNow, before we consider whether this is really the case, we should\nask what answer Kant's account of judgement would lead us to expect\nto the question 'What are all the functions of unity in judgement?'\nThe question must mean 'What are the kinds of unity produced by\njudgement?' To this question three alternative answers are prima facie\npossible. (1) There is only one kind of unity, that of a group of\nparticulars unified through relation to the corresponding universal.\nThe special unity produced will differ for different judgements, since\nit will depend upon the special universal involved. The kind or form\nof unity, however, will always be the same, viz. that of particulars\nrelated through the corresponding universal. For instance, 'plants'\nand 'trees' are unified respectively by the judgements 'This body is a\nplant' and 'This body is a tree'; for 'this body' is in the one case\nrelated to other 'plants' and in the other case to other 'trees'. And\nthough the unity produced is different in each case, the kind of unity\nis the same; for plants and trees are, as members of a kind, unities\nof a special kind distinct from unities of another kind, such as the\nparts of a spatial or numerical whole. (2) There are as many kinds of\nunity as there are universals. Every group of particulars forms a\nunity of a special kind through relation to the corresponding\nuniversal. (3) There are as many kinds of unity as there are highest\nuniversals or _summa genera_. These _summa genera_ are the most\ngeneral sources of unity through which individuals are related in\ngroups, directly or indirectly. The kinds of unity are therefore in\nprinciple the Aristotelian categories, i. e. the highest forms of\nbeing under which all individuals fall.\n\nNevertheless, it is easy to see that the second and third answers\nshould be rejected in favour of the first. For though, according to\nKant, a judgement unifies particulars by bringing them under a\nuniversal, the special universal involved in a given judgement belongs\nnot to the judgement as such, but to the particulars unified. What\nbelongs to the judgement as such is simply the fact that the\nparticulars are brought under a universal. In other words, the\njudgement as such determines the kind of unity but not the particular\nunity. The judgements 'Gold is a metal' and 'Trees are green',\nconsidered merely as judgements and not as the particular judgements\nwhich they are, involve the same kind of unity, viz. that of\nparticulars as particulars of a universal; for the distinction between\n'metal' and 'green' is a distinction not of kinds of unity but of\nunities. Moreover, to anticipate the discussion of Kant's final\nconclusion, the moral is that Kant's account of judgement should have\nled him to recognize that judgement involves the reality, not of any\nspecial universals or--in Kant's language--conceptions, but of\nuniversality or conception as such. In other words, on his view of\njudgement the activity of the understanding implies simply that there\n_are_ universals or conceptions; it does not imply the existence of\nspecial conceptions which essentially belong to the understanding,\ne. g. that of 'cause' or 'plurality'.[17]\n\n    [17] To this failure in Kant's argument is due the difficulty\n    in following his transition from 'function' to 'functions' of\n    judgements. The judgement, as Kant describes it, always does\n    one and the same thing; it unifies particulars by bringing\n    them under a universal. This activity does not admit of\n    differentiation.\n\nIf we now turn to the list of the activities of thought in judgement,\nborrowed from Formal Logic, we shall see that it is not in any way\nconnected with Kant's account of judgement.[18] For if the kinds of\njudgement distinguished by Formal Logic are to be regarded as\ndifferent ways of unifying, the plurality unified must be allowed to\nbe not a special kind of group of particulars, but the two conceptions\nwhich constitute the terms of the judgement[19]; and the unity\nproduced must be allowed to be in no case a special form of the unity\nof particulars related through the corresponding universal. Thus the\nparticular judgement 'Some coroners are doctors' must be said to unify\nthe conceptions of 'coroner' and of 'doctor', and presumably by means\nof the conception of 'plurality'. Again, the hypothetical judgement\n'If it rains, the ground will be wet' must be said to unify the\njudgements 'It rains' and 'The ground will be wet', and presumably by\nmeans of the conception of 'reason and consequence'. In neither case\ncan the act of unification be considered a special form of the act of\nrecognizing particulars as particulars of the corresponding universal.\nThe fact is that the distinctions drawn by Formal Logic are based on a\nview of judgement which is different from, and even incompatible with,\nKant's, and they arise from the attempt to solve a different problem.\nThe problem before Kant in describing judgement is to distinguish the\nunderstanding from the sensibility, i. e. thought from perception.\nHence he regards judgement as the act of unifying a manifold given in\nperception, directly, or indirectly by means of a conception. But this\nis not the problem with which Formal Logic is occupied. Formal Logic\nassumes judgement to be an act which relates material given to it in\nthe shape of 'conceptions' or 'judgements' by analysis of this\nmaterial, and seeks to discover the various modes of relation thereby\neffected. The work of judgement, however, cannot consist _both_ in\nrelating particulars through a conception _and_ in relating two\nconceptions or judgements.\n\n    [18] Moreover, the forms of judgement clearly lack the\n    systematic character which Kant claims for them. Even if it\n    be allowed that the subdivisions within the four main heads\n    of quantity, quality, relation, and modality are based upon\n    single principles of division, it cannot be said that the\n    four heads themselves originate from a common principle.\n\n    [19] In the case of the third division, the plurality unified\n    will be two prior judgements.\n\nIt may be urged that this criticism only affects Kant's argument, but\nnot his conclusion. Possibly, it may be said, the list of types of\njudgement borrowed from Formal Logic really expresses the essential\ndifferentiations of judgement, and, in that case, Kant's only mistake\nis that he bases them upon a false or at least inappropriate account\nof judgement.[20] Moreover, since this list furnishes Kant with the\n'clue' to the categories, provided that it expresses the essential\ndifferentiations of judgement, the particular account of judgement\nupon which it is based is a matter of indifference.\n\n    [20] It may be noted that the account cannot be merely\n    inappropriate to the general problem, if it be _incompatible_\n    with that assumed by Formal Logic.\n\nThis contention leads us to consider the last stage of Kant's\nargument, in which he deduces the categories in detail from his list\nof the forms of judgement. For it is clear that unless the forms of\njudgement severally involve the categories, it will not matter whether\nthese forms are or are not the essential differentiations of\njudgement.\n\nKant's mode of connecting the categories in detail with the forms of\njudgement discovered by Formal Logic is at least as surprising as his\nmode of connecting the latter with the nature of judgement in general.\nSince the twelve distinctions within the form of judgement are to\nserve as a clue to the conceptions which belong to the understanding,\nwe naturally expect that each distinction will be found directly to\ninvolve a special conception or category, and that therefore, to\ndiscover the categories, we need only look for the special conception\ninvolved in each form of judgement.[21] Again, since the plurality\nunified in a judgement of each form is the two conceptions or\njudgements which form the matter of the judgement, we should expect\nthe conception involved in each form of judgement to be merely the\ntype of relationship established between these conceptions or\njudgements. This expectation is confirmed by a cursory glance at the\ntable of categories.[22]\n\n    I. _Of Quantity._\n       Unity\n       Plurality\n       Totality.\n\n   II. _Of Quality._\n       Reality\n       Negation\n       Limitation.\n\n  III. _Of Relation._\n       Inherence and Subsistence (_Substantia et Accidens_)\n       Causality and Dependence (_Cause and Effect_)\n       Community (_Reciprocity between the agent and patient._)\n\n   IV. _Of Modality._\n       Possibility--Impossibility\n       Existence--Non-existence\n       Necessity--Contingence.\n\nIf we compare the first division of these categories with the first\ndivision of judgements we naturally think that Kant conceived\nsingular, particular, and universal judgements to unify their terms\nby means of the conceptions of 'one', of 'some', and of 'all'\nrespectively; and we form corresponding, though less confident,\nexpectations in the case of the other divisions.\n\n    [21] This expectation is confirmed by Kant's view that\n    judgement introduces unity into a plurality by means of a\n    conception. This view leads us to expect that different forms\n    of judgement--if there be any--will be distinguished by the\n    different conceptions through which they unify the plurality;\n    for it will naturally be the different conceptions involved\n    which are responsible for the different kinds of unity\n    effected.\n\n    [22] B. 106, M. 64.\n\nKant, however, makes no attempt to show that each form of judgement\ndistinguished by Formal Logic involves a special conception. In fact,\nhis view is that the activities of thought studied by Formal Logic do\nnot originate or use any special conceptions at all. For his actual\ndeduction of the categories[23] is occupied in showing that although\nthought, when exercised under the conditions under which it is studied\nby Formal Logic, does not originate and use conceptions of its own, it\nis able under certain other conditions to originate and use such\nconceptions, i. e. categories.[24] Hence if we attend only to the\nprofessed procedure of the deduction, we are compelled to admit that\nthe deduction not only excludes any use of the 'clue' to the\ncategories, supposed to be furnished by Formal Logic, but even fails\nto deduce them at all. For it does not even nominally attempt to\ndiscover the categories in detail, but reverts to the prior task of\nshowing merely that there are categories. Doubtless Kant thinks that\nthe forms of judgement formulated by Formal Logic in some way\n_suggest_ the conceptions which become operative in thought under\nthese other conditions. Nevertheless, it is impossible to see how\nthese forms of judgement can suggest these conceptions, unless they\nactually presuppose them.\n\nIt is clear, however, that the professed link[25] between the forms of\njudgement and the categories does not represent the actual process by\nwhich Kant reached his list of categories; for he could never have\nreached any list of categories by an argument which was merely\ndirected to show that there are categories. Moreover, an inspection of\nthe list shows that he actually reached it partly by noticing the\nconceptions which the forms of judgement seemed to presuppose, and\npartly by bearing in mind the general conceptions underlying physics\nwhich it was his ultimate aim to vindicate. Since this is the case,\nand since the categories can only be connected with the forms of\njudgement by showing that they are presupposed in them, the proper\nquestion to be considered from the point of view of the metaphysical\ndeduction is simply whether the forms of judgement really presuppose\nthe categories.[26]\n\n    [23] B. 102-5, M. 62-3.\n\n    [24] Cf. p. 166.\n\n    [25] B. 102-5, M. 62-3.\n\n    [26] As we shall see later, the real importance of the\n    passage in which Kant professes to effect the transition from\n    the forms of judgement to the categories (B. 102-5, M. 62-3)\n    lies in its introduction of a new and important line of\n    thought, on which the transcendental deduction turns.\n    Consideration of it is therefore deferred to the next\n    chapter.\n\nIf, however, we examine the forms of judgement distinguished by Formal\nLogic, we find that they do not presuppose the categories. To see\nthis, it is only necessary to examine the four main divisions of\njudgement _seriatim_.\n\nThe first division of judgements is said to be a division in respect\nof quantity into singular, particular, and universal. So stated, the\ndivision is numerical. It is a division of judgements according as\nthey make an assertion about one, more than one, or all the members of\na kind. Each species may be said to presuppose (1) the conception of\nquantity, and (2) a conception peculiar to itself: the first\npresupposing the conception of one member of a kind, the second that\nof more than one but less than all members of a kind, the third that\nof all members of a kind. Moreover, a judgement of each kind may\nperhaps be said to relate the predicate conception to the subject\nconception by means of one of these three conceptions.\n\nThe fundamental division, however, into which universal and singular\njudgements enter is not numerical at all, and ignores particular\njudgements altogether. It is that between such judgements as\n'Three-sided figures, as such, are three-angled' and 'This man is\ntall'. The essential distinction is that in the universal judgement\nthe predicate term is apprehended to belong to the subject through\nour insight that it is necessitated by the nature of the subject term,\nwhile in the singular judgement our apprehension that the predicate\nterm belongs to the subject is based upon the perception or experience\nof the coexistence of predicate and subject terms in a common subject.\nIn other words, it is the distinction between an _a priori_ judgement\nand a judgement of perception.[27] The merely numerically universal\njudgement, and the merely numerically particular judgement[28] are\nsimply aggregates of singular judgements, and therefore are\nindistinguishable in principle from the singular judgement. If then we\nask what conceptions are really presupposed by the kinds of judgement\nwhich Kant seeks to distinguish in the first division, we can only\nreply that the universal judgement presupposes the conception of a\nconnected or systematic whole of attributes, and that the singular\njudgement presupposes the conception of the coexistence of two\nattributes in a common subject. Neither kind of judgement presupposes\nthe conception of quantity or the conceptions of unity, plurality, and\ntotality.\n\n    [27] I owe this view of the distinction to Professor Cook\n    Wilson's lectures on logic.\n\n    [28] 'Some coroners are doctors' of course in some contexts\n    means, 'it is possible for a coroner to be a doctor,' and is\n    therefore not numerical; but understood in this sense it is\n    merely a weakened form of the universal judgement in which\n    the connexion apprehended between subject and predicate terms\n    is incomplete.\n\nThe second division of judgements is said to be a division in respect\nof quality into affirmative, negative, and infinite, i. e. into\nspecies which may be illustrated by the judgements, 'A college is a\nplace of education,' 'A college is not a hotel,' and 'A college is a\nnot-hotel'. The conceptions involved are said to be those of reality,\nof negation, and of limitation respectively. The conception of\nlimitation may be ignored, since the infinite judgement said to\npresuppose it is a fiction. On the other hand, the conceptions of\nreality and negation, even if their existence be conceded, cannot be\nallowed to be the conceptions presupposed. For when we affirm or deny,\nwe affirm or deny of something not mere being, but being of a\nparticular kind. The conceptions presupposed are rather those of\nidentity and difference. It is only because differences fall within an\nidentity that we can affirm, and it is only because within an identity\nthere are differences that we can deny.\n\nThe third division of judgements is said to be in respect of relation\ninto categorical, hypothetical, and disjunctive judgements. Here,\nagain, the conclusion which Kant desires is clearly impossible. The\ncategorical judgement may be said to presuppose the conception of\nsubject and attribute, but not that of substance and accident. The\nhypothetical judgement may be conceded to presuppose the conception of\nreason and consequence, but it certainly does not presuppose the\nconception of cause and effect.[29] Lastly, while the disjunctive\njudgement may be said to presuppose the conception of mutually\nexclusive species of a genus, it certainly does not presuppose the\nconception of reciprocal action between physical things.\n\n    [29] No doubt, as the schematism of the categories shows,\n    Kant does not think that the hypothetical judgement\n    _directly_ involves the conception of cause and effect, i. e.\n    of the relation of necessary succession between the various\n    states of physical things. The point is, however, that the\n    hypothetical judgement does not involve it at all.\n\nThe fourth division of judgement is said to be in respect of modality\ninto assertoric, problematic, and apodeictic, the conceptions\ninvolved being respectively those of possibility and impossibility, of\nactuality and non-actuality, and of necessity and contingence. Now,\nfrom the point of view of Kant's argument, these conceptions, like\nthose which he holds to be involved in the other divisions of\njudgement, must be considered to relate to reality and not to our\nattitude towards it. Considered in this way, they resolve themselves\ninto the conceptions of--\n\n(1) the impossible (impossibility);\n(2) the possible but not actual (possibility, nonexistence);\n(3) the actual but not necessary (existence, contingence);\n(4) the necessary (necessity).\n\nBut since it must, in the end, be conceded that all fact is necessary,\nit is impossible to admit the reality of the conception of the\npossible but not actual, and of the actual but not necessary. There\nremain, therefore, only the conceptions of the necessary and of the\nimpossible. In fact, however, the distinctions between the assertoric,\nthe problematic, and the apodeictical judgement relate to our attitude\nto reality and not to reality, and therefore involve no different\nconceptions relating to reality. It must, therefore, be admitted\nthat the 'metaphysical' deduction of the categories breaks down\ndoubly. Judgement, as Kant describes it, does not involve the\nforms of judgement borrowed from Formal Logic as its essential\ndifferentiations; and these forms of judgement do not involve the\ncategories.\n\n\n\n\nCHAPTER VIII\n\nTHE TRANSCENDENTAL DEDUCTION OF THE CATEGORIES\n\n\nThe aim of the _Transcendental Deduction_ is to show that the\ncategories, though _a priori_ as originating in the understanding, are\nvalid, i. e. applicable to individual things. It is the part of the\n_Critique_ which has attracted most attention and which is the most\ndifficult to follow. The difficulty of interpretation is increased\nrather than diminished by the complete rewriting of this portion in\nthe second edition. For the second version, though it does not imply\na change of view, is undoubtedly even more obscure than the first.\nIt indeed makes one new contribution to the subject by adding an\nimportant link in the argument,[1] but the importance of the link\nis nullified by the fact that it is not really the link which it\nprofesses to be. The method of treatment adopted here will be to\nconsider only the minimum of passages necessary to elucidate Kant's\nmeaning and to make use primarily of the first edition.\n\n    [1] Cf. p. 206-10.\n\nIt is necessary, however, first to consider the passage in the\n_Metaphysical Deduction_ which nominally connects the list of\ncategories with the list of forms of judgement.[2] For its real\nfunction is to introduce a new and third account of knowledge, which\nforms the keynote of the _Transcendental Deduction_.[3]\n\n    [2] B. 102-5, M. 62-3. Cf. pp. 155-6.\n\n    [3] The first two accounts are (1) that of judgement given B.\n    92-4, M. 56-8, and (2) that of judgement implicit in the view\n    that the forms of judgement distinguished by Formal Logic are\n    functions of unity. In A. 126, Mah. 215, Kant seems to\n    imply--though untruly--that this new account coincides with\n    the other two, which he does not distinguish.\n\nIn this passage, the meaning of which it is difficult to state\nsatisfactorily, Kant's thought appears to be as follows: 'The activity\nof thought studied by Formal Logic relates by way of judgement\nconceptions previously obtained by an analysis of perceptions. For\ninstance, it relates the conceptions of body and of divisibility,\nobtained by analysis of perceptions of bodies, in the judgement\n'Bodies are divisible'. It effects this, however, merely by analysis\nof the conception 'body'. Consequently, the resulting knowledge or\njudgement, though _a priori_, is only analytic, and the conceptions\ninvolved originate not from thought but from the manifold previously\nanalysed. But besides the conceptions obtained by analysis of a given\nmanifold, there are others which belong to thought or the\nunderstanding as such, and in virtue of which thought originates\nsynthetic _a priori_ knowledge, this activity of thought being that\nstudied by Transcendental Logic. Two questions therefore arise.\nFirstly, how do these conceptions obtain a matter to which they can\napply and without which they would be without content or empty? And,\nsecondly, how does thought in virtue of these conceptions originate\nsynthetic _a priori_ knowledge? The first question is easily answered,\nfor the manifolds of space and time, i. e. individual spaces and\nindividual times, afford matter of the kind needed to give these\nconceptions content. As perceptions (i. e. as objects of perception),\nthey are that to which a conception can apply, and as pure or _a\npriori_ perceptions, they are that to which those conceptions can\napply which are pure or _a priori_, as belonging to the understanding.\nThe second question can be answered by considering the process by\nwhich this pure manifold of space and time enters into knowledge. All\nsynthetic knowledge, whether empirical or _a priori_, requires the\nrealization of three conditions. In the first place, there must be a\nmanifold given in perception. In the second place, this manifold must\nbe 'gone through, taken up, and combined'. In other words, if\nsynthesis be defined as 'the act of joining different representations\nto one another and of including their multiplicity in one knowledge',\nthe manifold must be subjected to an act of synthesis. This is\neffected by the imagination. In the third place, this synthesis\nproduced by the imagination must be brought to a conception, i. e.\nbrought under a conception which will constitute the synthesis a\nunity. This is the work of the understanding. The realization of _a\npriori_ knowledge, therefore, will require the realization of the\nthree conditions in a manner appropriate to its _a priori_ character.\nThere must be a pure or _a priori_ manifold; this is to be found in\nindividual spaces and individual times. There must be an act of pure\nsynthesis of this manifold; this is effected by the pure imagination.\nFinally, this pure synthesis must be brought under a conception. This\nis effected by the pure understanding by means of its pure or a priori\nconceptions, i. e. the categories. This, then, is the process by which\n_a priori_ knowledge is originated. The activity of thought or\nunderstanding, however, which unites two conceptions in a judgement by\nanalysis of them--this being the act studied by Formal Logic--is the\nsame as that which gives unity to the synthesis of the pure manifold\nof perception--this being the act studied by Transcendental Logic.\nConsequently, 'the same understanding, and indeed by the same\nactivities whereby in dealing with conceptions it unifies them in a\njudgement by an act of analysis, introduces by means of the\nsynthetical unity which it produces in the pure manifold of perception\na content into its own conceptions, in consequence of which these\nconceptions are called pure conceptions of the understanding,'[4] and\nwe are entitled to say _a priori_ that these conceptions apply to\nobjects because they are involved in the process by which we acquire\n_a priori_ knowledge of objects.'\n\n    [4] An interpretation of B. 105 init., M. 63 fin.\n\nA discussion of the various difficulties raised by the general drift\nof this passage, as well as by its details,[5] is unnecessary, and\nwould anticipate discussion of the _Transcendental Deduction_. But it\nis necessary to draw attention to three points.\n\n    [5] E. g. Kant's arbitrary assertion that the operation of\n    counting presupposes the conception of that number which\n    forms the scale of notation adopted as the source of the\n    unity of the synthesis. This is of course refuted among other\n    ways by the fact that a number of units less than the scale\n    of notation can be counted.\n\nIn the first place, as has been said, Kant here introduces--and\nintroduces without warning--a totally new account of knowledge. It has\nits origin in his theory of perception, according to which knowledge\nbegins with the production of sensations in us by things in\nthemselves. Since the spatial world which we come to know consists in\na multiplicity of related elements, it is clear that the isolated data\nof sensation have somehow to be combined and unified, if we are to\nhave this world before us or, in other words, to know it. Moreover,\nsince these empirical data are subject to space and time as the forms\nof perception, individual spaces and individual times, to which the\nempirical data will be related, have also to be combined and unified.\nOn this view, the process of knowledge consists in combining certain\ndata into an individual whole and in unifying them through a principle\nof combination.[6] If the data are empirical, the resulting knowledge\nwill be empirical; if the data are _a priori_, i. e. individual spaces\nand individual times, the resulting knowledge will be _a priori_.[7]\nThis account of knowledge is new, because, although it treats\nknowledge as a process or act of unifying a manifold, it describes a\ndifferent act of unification. As Kant first described the faculty of\njudgement,[8] it unifies a group of particulars through relation to\nthe corresponding universal. As Formal Logic, according to Kant,\ntreats the faculty of judgement, it unifies two conceptions or two\nprior judgements into a judgement. As Kant now describes the faculty\nof judgement or thought, it unifies an empirical or an _a priori_\nmanifold of perception combined into an individual whole, through a\nconception which constitutes a principle of unity. The difference\nbetween this last account and the others is also shown by the fact\nthat while the first two kinds of unification are held to be due to\nmere analysis of the material given to thought, the third kind of\nunification is held to be superinduced by thought, and to be in no way\ncapable of being extracted from the material by analysis. Further,\nthis new account of knowledge does not replace the others, but is\nplaced side by side with them. For, according to Kant, there exist\n_both_ the activity of thought which relates two conceptions in a\njudgement,[9] _and_ the activity by which it introduces a unity of its\nown into a manifold of perception. Nevertheless, this new account of\nknowledge, or rather this account of a new kind of knowledge, must be\nthe important one; for it is only the process now described for the\nfirst time which produces synthetic as opposed to analytic knowledge.\n\n    [6] Cf. A. 97, Mah. 193, 'Knowledge is a totality of compared\n    and connected representations.'\n\n    [7] No doubt Kant would allow that at least some categories,\n    e. g. the conception of cause and effect, are principles of\n    synthesis of a manifold which at any rate contains an\n    empirical element, but it _includes_ just one of the\n    difficulties of the passage that it implies that _a priori_\n    knowledge either is, or involves, a synthesis of pure or _a\n    priori_ elements.\n\n    [8] B. 92-4, M. 56-8.\n\n    [9] Kant, of course, thinks of this activity of thought,\n    as identical with that which brings particulars under a\n    conception.\n\nIn the second place, the passage incidentally explains why, according\nto Kant, the forms of judgement distinguished by Formal Logic do not\ninvolve the categories.[10] For its doctrine is that while thought, if\nexercised under the conditions under which it is studied by Formal\nLogic, can only analyse the manifold given to it, and so has, as it\nwere, to borrow from the manifold the unity through which it relates\nthe manifold,[11] yet if an _a priori_ manifold be given to it, it can\nby means of a conception introduce into the manifold a unity of its\nown which could not be discovered by analysis of the manifold. Thus\nthought as studied by Formal Logic merely analyses and consequently\ndoes not and cannot make use of conceptions of its own; it can use\nconceptions of its own only when an _a priori_ manifold is given to it\nto deal with.\n\n    [10] Cf. pp. 155-6.\n\n    [11] In bringing perceptions under a conception, thought,\n    according to Kant, finds the conception _in_ the perceptions\n    by analysis of them, and in relating two conceptions in\n    judgement, it determines the particular form of judgement by\n    analysis of the conceptions.\n\nIn the third place, there is great difficulty in following the part in\nknowledge assigned to the understanding. The synthesis of the manifold\nof perception is assigned to the imagination, a faculty which, like\nthe new kind of knowledge, is introduced without notice. The business\nof the understanding is to 'bring this synthesis to conceptions' and\nthereby to 'give unity to the synthesis'. Now the question arises\nwhether 'the activity of giving unity to the synthesis' really means\nwhat it says, i. e. an activity which _unifies_ or _introduces a unity\ninto_ the synthesis, or whether it only means an activity which\n_recognizes_ a unity already given to the synthesis by the\nimagination. Prima facie Kant is maintaining that the understanding\nreally unifies, or introduces the principle of unity. For the\ntwice-repeated phrase 'give unity to the synthesis' seems unmistakable\nin meaning, and the important role in knowledge is plainly meant to be\nassigned to the understanding. Kant's language, however, is not\ndecisive; for he speaks of the synthesis of the manifold as that which\n'first produces a knowledge which indeed at first may be crude and\nconfused and therefore needs _analysis_[12]', and he says of the\nconceptions which give unity to the synthesis that 'they consist\nsolely in the _representation_[13] of this necessary synthetical\nunity'.[14] Again, 'to bring the synthesis to a conception' may well\nbe understood to mean 'to recognize the synthesis as an instance of\nthe conception'; and, since Kant is speaking of knowledge, 'to give\nunity to the synthesis' may only mean 'to give unity to the synthesis\n_for us_', i. e. 'to make us aware of its unity'. Moreover,\nconsideration of what thought can possibly achieve with respect to a\nsynthesis presented to it by the imagination renders it necessary to\nhold that the understanding only recognizes the unity of the\nsynthesis. For if a synthesis has been effected, it must have been\neffected in accordance with a principle of construction or synthesis,\nand therefore it would seem that the only work left for the\nunderstanding is to discover the principle latent in the procedure of\nthe imagination. At any rate, if the synthesis does not involve a\nprinciple of synthesis, it is impossible to see how thought can\nsubsequently introduce a principle. The imagination, then, must be\nconsidered to have already introduced the principle of unity into the\nmanifold by combining it in accordance with a conception or principle\nof combination, and the work of the understanding must be considered\nto consist in recognizing that the manifold has been thereby combined\nand unified through the conception. We are therefore obliged to accept\none of two alternatives. _Either_ the understanding merely renders the\nmind conscious of the procedure of a faculty different from itself,\nviz. the imagination, in which case the important role in knowledge,\nviz. the effecting of the synthesis according to a principle, is\nplayed by a faculty different from the understanding; _or_ the\nimagination is the understanding working unreflectively, and the\nsubsequent process of bringing the synthesis to a conception is merely\na process by which the understanding becomes conscious of its own\nprocedure. Moreover, it is the latter alternative which we must accept\nas more in accordance with the general tenor of Kant's thought. For\nthe synthesis of the imagination is essentially the outcome of\nactivity or spontaneity, and, as such, it belongs to the understanding\nrather than to the sensibility; in fact we find Kant in one place\nactually saying that 'it is one and the same spontaneity which at one\ntime under the name of imagination, at another time under that of\nunderstanding, introduces connexion into the manifold of\nperception'.[15] Further, it should be noted that since the\nimagination must be the understanding working unreflectively, and\nsince it must be that which introduces unity into the manifold, there\nis some justification for his use of language which implies that the\nunderstanding is the source of the unity, though it will not be so in\nthe sense in which the passage under discussion might at first sight\nlead us to suppose.\n\n    [12] The italics are mine.\n\n    [13] The italics are mine.\n\n    [14] Cf. the description of the imagination as 'blind'.\n\n    [15] B. 162 note, M. 99 note. Cf. B. 152, M. 93. Similarly at\n    one point in the passage under discussion (B. 102 fin., M. 62\n    med.) the synthesis is expressly attributed to the\n    spontaneity of thought.\n\nWe can now turn to the argument of the _Transcendental Deduction_\nitself. Kant introduces it in effect by raising the question, 'How is\nit that, beginning with the isolated data of sense, we come to acquire\nknowledge?' His aim is to show (1) that knowledge requires the\nperformance of certain operations by the mind upon the manifold of\nsense; (2) that this process is a condition not merely of knowledge,\nbut also of self-consciousness; and (3) that, since the manifold is\ncapable of entering into knowledge, and since we are capable of being\nself-conscious, the categories, whose validity is implied by this\nprocess, are valid.\n\nKant begins by pointing out[16] that all knowledge, _a priori_ as well\nas empirical, requires the manifold, produced successively in the\nmind, to be subjected to three operations.\n\n    [16] A. 95-104, Mah. 194-8.\n\n1. Since the elements of the manifold are as given mere isolated\nunits, and since knowledge is the apprehension of a unity of connected\nelements, the mind must first run through the multiplicity of sense\nand then grasp it together into a whole, i. e. into an image.[17]\nThis act is an act of synthesis; it is called 'the synthesis of\napprehension' and is ascribed to the imagination. It must be carried\nout as much in respect of the pure or _a priori_ elements of space and\ntime as in respect of the manifold of sensation, for individual spaces\nand times contain a multiplicity which, to be apprehended, must be\ncombined.[18] The necessity of this act of synthesis is emphasized in\nthe second edition. \"We cannot represent anything as combined in the\nobject without having previously combined it ourselves. Of all\nrepresentations, _combination_ is the only one which cannot be given\nthrough objects,[19] but can be originated only by the subject itself\nbecause it is an act of its own activity.\"[20]\n\n    [17] Cf. A. 120, Mah. 211.\n\n    [18] 'Combine' is used as the verb corresponding to\n    'synthesis'.\n\n    [19] I. e. given to us through the operation of things in\n    themselves upon our sensibility.\n\n    [20] B. 130, M. 80.\n\n2. Since the data of perception are momentary, and pass away with\nperception, the act of grasping them together requires that the mind\nshall reproduce the past data in order to combine them with the\npresent datum. \"It is plain that if I draw a line in thought, or wish\nto think of the time from one midday to another, or even to represent\nto myself a certain number, I must first necessarily grasp in thought\nthese manifold representations one after another. But if I were\ncontinually to lose from my thoughts the preceding representations\n(the first parts of the line, the preceding parts of time or the units\nsuccessively represented), and were not to reproduce them, while I\nproceeded to the succeeding parts, there could never arise a complete\nrepresentation, nor any of the thoughts just named, not even the first\nand purest fundamental representations of space and time.\"[21] This\nact of reproduction is called 'the synthesis of reproduction in the\nimagination'.[22]\n\n    [21] A. 102, Mah. 197.\n\n    [22] The term 'synthesis' is undeserved, and is due to a\n    desire to find a verbal parallel to the 'synthesis of\n    apprehension in perception'. For the inappropriateness of\n    'reproduction' and of 'imagination' see pp. 239-41.\n\nFurther, the necessity of reproduction brings to light a\ncharacteristic of the synthesis of apprehension. \"It is indeed only an\nempirical law, according to which representations which have often\nfollowed or accompanied one another in the end become associated, and\nso form a connexion, according to which, even in the absence of the\nobject, one of these representations produces a transition of the mind\nto another by a fixed rule. But this law of reproduction presupposes\nthat phenomena themselves are actually subject to such a rule, and\nthat in the manifold of their representations there is a concomitance\nor sequence, according to a fixed rule; for, without this, our\nempirical imagination would never find anything to do suited to its\ncapacity, and would consequently remain hidden within the depths of\nthe mind as a dead faculty, unknown to ourselves. If cinnabar were now\nred, now black, now light, now heavy, if a man were changed now into\nthis, now into that animal shape, if our fields were covered on the\nlongest day, now with fruit, now with ice and snow, then my empirical\nfaculty of imagination could not even get an opportunity of thinking\nof the heavy cinnabar when there occurred the representation of red\ncolour; or if a certain name were given now to one thing, now to\nanother, or if the same thing were called now by one and now by\nanother name, without the control of some rule, to which the phenomena\nthemselves are already subject, no empirical synthesis of reproduction\ncould take place.\"\n\n\"There must then be something which makes this very reproduction of\nphenomena possible, by being the _a priori_ foundation of a necessary\nsynthetical unity of them. But we soon discover it, if we reflect that\nphenomena are not things in themselves, but the mere play of our\nrepresentations, which in the end resolve themselves into\ndeterminations of our internal sense. For if we can prove that even\nour purest _a priori_ perceptions afford us no knowledge, except so\nfar as they contain such a combination of the manifold as renders\npossible a thoroughgoing synthesis of reproduction, then this\nsynthesis of imagination is based, even before all experience, on _a\npriori_ principles, and we must assume a pure transcendental synthesis\nof the imagination which lies at the foundation of the very\npossibility of all experience (as that which necessarily presupposes\nthe reproducibility of phenomena).\"[23]\n\n    [23] A. 100-2, Mah. 195-7.\n\nIn other words, the faculty of reproduction, if it is to get to work,\npresupposes that the elements of the manifold are parts of a\nnecessarily related whole; or, as Kant expresses it later, it\npresupposes the _affinity_ of phenomena; and this affinity in turn\npresupposes that the synthesis of apprehension by combining the\nelements of the manifold on certain principles makes them parts of a\nnecessarily related whole.[24]\n\n    [24] Cf. A. 113, Mah. 205; A. 121-2, Mah. 211-12; and Caird,\n    i. 362-3. For a fuller account of these presuppositions, and\n    for a criticism of them, cf. Ch. IX, p. 219 and ff.\n\n3. Kant introduces the third operation, which he calls 'the synthesis\nof recognition in the conception',[25] as follows:\n\n    [25] This title also is a misnomer due to the desire to give\n    parallel titles to the three operations involved in\n    knowledge. There is really only one synthesis referred to,\n    and the title here should be 'the recognition of the\n    synthesis in the conception'.\n\n\"Without consciousness that what we are thinking is identical with\nwhat we thought a moment ago, all reproduction in the series of\nrepresentations would be in vain. For what we are thinking would be a\nnew representation at the present moment, which did not at all belong\nto the act by which it was bound to have been gradually produced, and\nthe manifold of the same would never constitute a whole, as lacking\nthe unity which only consciousness can give it. If in counting I\nforget that the units which now hover before my mind have been\ngradually added by me to one another, I should not know the generation\nof the group through this successive addition of one to one, and\nconsequently I should not know the number, for this conception\nconsists solely in the consciousness of this unity of the synthesis.\"\n\n\"The word 'conception'[26] might itself lead us to this remark. For\nit is this _one_ consciousness which unites the manifold gradually\nperceived and then also reproduced into one representation. This\nconsciousness may often be only weak, so that we connect it with the\nproduction of the representation only in the result but not in the act\nitself, i. e. immediately; but nevertheless there must always be one\nconsciousness, although it lacks striking clearness, and without it\nconceptions, and with them knowledge of objects, are wholly\nimpossible.\"[27]\n\n    [26] _Begriff._\n\n    [27] A. 103-4, Mah. 197-8.\n\nThough the passage is obscure and confused, its general drift is\nclear. Kant, having spoken hitherto only of the operation of the\nimagination in apprehension and reproduction, now wishes to introduce\nthe understanding. He naturally returns to the thought of it as that\nwhich recognizes a manifold as unified by a conception, the manifold,\nhowever, being not a group of particulars unified through the\ncorresponding universal or conception, but the parts of an individual\nimage, e. g. the parts of a line or the constituent units of a number,\nand the conception which unifies it being the principle on which these\nparts are combined.[28] His main point is that it is not enough for\nknowledge that we should combine the manifold of sense into a whole in\naccordance with a specific principle,[29] but we must also be in some\ndegree conscious of our continuously identical act of combination,[30]\nthis consciousness being at the same time a consciousness of the\nspecial unity of the manifold. For the conception which forms the\nprinciple of the combination has necessarily two sides; while from our\npoint of view it is the principle according to which we combine and\nwhich makes our combining activity one, from the point of view of the\nmanifold it is the special principle[31] by which the manifold is made\n_one_. If I am to count a group of five units, I must not only add\nthem, but also be conscious of my continuously identical act of\naddition, this consciousness consisting in the consciousness that I am\nsuccessively taking units up to, and only up to, five, and being at\nthe same time a consciousness that the units are acquiring the unity\nof being a group of five. It immediately follows, though Kant does not\nexplicitly say so, that all knowledge implies self-consciousness. For\nthe consciousness that we have been combining the manifold on a\ncertain definite principle is the consciousness of our identity\nthroughout the process, and, from the side of the manifold, it is just\nthat consciousness of the manifold as unified by being brought under a\nconception which constitutes knowledge. Even though it is Kant's view\nthat the self-consciousness need only be weak and need only arise\nafter the act of combination, when we are aware of its result, still,\nwithout it, there will be no consciousness of the manifold as unified\nthrough a conception and therefore no knowledge. Moreover, if the\nself-consciousness be weak, the knowledge will be weak also, so that\nif it be urged that knowledge in the strictest sense requires the full\nconsciousness that the manifold is unified through a conception, it\nmust be allowed that knowledge in this sense requires a full or clear\nself-consciousness.\n\n    [28] Cf. pp. 162-9.\n\n    [29] That the combination proceeds on a specific principle\n    only emerges in this account of the third operation.\n\n    [30] Kant's example shows that this consciousness is not the\n    mere consciousness of the act of combination as throughout\n    identical, but the consciousness of it as an identical act of\n    a particular kind.\n\n    [31] When Kant says 'this conception [i. e. the conception\n    of the number counted] consists in the consciousness of this\n    unity of the synthesis', he is momentarily and contrary to\n    his usual practice speaking of a conception in the sense\n    of the activity of conceiving a universal, and not in the\n    sense of the universal conceived. Similarly in appealing to\n    the meaning of _Begriff_ (conception) he is thinking of\n    'conceiving' as the activity of combining a manifold through\n    a conception.\n\nAs is to be expected, however, the passage involves a difficulty\nconcerning the respective functions of the imagination and the\nunderstanding. Is the understanding represented as only recognizing a\nprinciple of unity introduced into the manifold by the imagination, or\nas also for the first time introducing a principle of unity? At first\nsight the latter alternative may seem the right interpretation. For he\nsays that unless we were conscious that what we are thinking is\nidentical with what we thought a moment ago, 'what we are thinking\nwould _be_ a new representation which _did not at all belong_ to the\nact by which it was bound to have been gradually produced, and the\nmanifold of the same _would never_ constitute a whole, as lacking the\nunity which only _consciousness can give it_.'[32] Again, in speaking\nof a conception--which of course implies the understanding--he\nsays that 'it is this one consciousness which _unites_ the\nmanifold gradually perceived and then reproduced into _one_\nrepresentation'.[33] But these statements are not decisive, for he\nuses the term 'recognition' in his formula for the work of the\nunderstanding, and he illustrates its work by pointing out that in\ncounting we must _remember_ that we have added the units. Moreover,\nthere is a consideration which by itself makes it necessary to accept\nthe former interpretation. The passage certainly represents the\nunderstanding as recognizing the identical action of the mind in\ncombining the manifold on a principle, whether or not it also\nrepresents the understanding as the source of this activity. But if\nit were the understanding which combined the manifold, there would\nbe no synthesis which the imagination could be supposed to have\nperformed,[34] and therefore it could play no part in knowledge at\nall, a consequence which must be contrary to Kant's meaning. Further\nif, as the general tenor of the deduction shows, the imagination is\nreally only the understanding working unreflectively,[35] we are able\nto understand why Kant should for the moment cease to distinguish\nbetween the imagination and the understanding, and consequently should\nuse language which implies that the understanding both combines the\nmanifold on a principle and makes us conscious of our activity in so\ndoing. Hence we may say that the real meaning of the passage should\nbe stated thus: 'Knowledge requires one consciousness which,\nas imagination, combines the manifold on a definite principle\nconstituted by a conception,[36] and, as understanding, is to some\nextent conscious of its identical activity in so doing, this\nself-consciousness being, from the side of the whole produced by the\nsynthesis, the consciousness of the conception by which the manifold\nis unified.'\n\n    [32] The italics are mine. He does not say '_we should not be\n    conscious_ of what we are thinking as the same representation\n    and as belonging [Greek: ktl]., _and we should not be\n    conscious_ of the manifold as constituting a whole.\n\n    [33] The italics are mine.\n\n    [34] There could not, of course, be two syntheses, the one\n    being and the other not being upon a principle.\n\n    [35] Cf. pp. 168-9.\n\n    [36] In view of Kant's subsequent account of the function of\n    the categories it should be noticed that, according to the\n    present passage, the conception involved in an act of\n    knowledge is the conception not of an 'object in general',\n    but of 'an object of the particular kind which constitutes\n    the individual whole produced by the combination a whole of\n    the particular kind that it is of', and that, in accordance\n    with this, the self-consciousness involved is not the mere\n    consciousness that our combining activity is identical\n    throughout, but the consciousness that it is an identical\n    activity of a particular kind, e. g. that of counting five\n    units. Cf. pp. 184 fin.-186, 190-2, and 206-7.\n\nHitherto there has been no mention of an _object_ of knowledge, and\nsince knowledge is essentially knowledge of an object, Kant's next\ntask is to give such an account of an object of knowledge as will show\nthat the processes already described are precisely those which give\nour representations, i. e. the manifold of sense, relation to an\nobject, and consequently yield knowledge.\n\nHe begins by raising the question, 'What do we mean by the phrase 'an\nobject of representations'?'[37] He points out that a phenomenon, since\nit is a mere sensuous representation, and not a thing in itself\nexisting independently of the faculty of representations, is just not\nan object. To the question, therefore, 'What is meant by an object\ncorresponding to knowledge and therefore distinct from it?' we are\nbound to answer from the point of view of the distinction between\nphenomena and things in themselves, that the object is something in\ngeneral = _x_, i. e. the thing in itself of which we know only\n_that_ it is and not _what_ it is. There is, however, another point\nof view from which we can say something more about an object of\nrepresentations and the correspondence of our representations to it,\nviz. that from which we consider what is involved in the thought of\nthe relation of knowledge or of a representation to its object. \"We\nfind that our thought of the relation of all knowledge to its object\ncarries with it something of necessity, since its object is regarded\nas that which prevents our cognitions[38] being determined at random\nor capriciously, and causes them to be determined _a priori_ in a\ncertain way, because in that they are to relate to an object, they\nmust necessarily also, in relation to it, agree with one another, that\nis to say, they must have that unity which constitutes the conception\nof an object.\"[39]\n\n    [37] _Vorstellung_ in the present passage is perhaps better\n    rendered 'idea', but representation has been retained for the\n    sake of uniformity.\n\n    [38] _Erkenntnisse._\n\n    [39] A. 104, Mah. 199.\n\nKant's meaning seems to be this: 'If we think of certain\nrepresentations, e. g. certain lines[40] or the representations of\nextension, impenetrability, and shape,[41] as related to an object,\ne. g. to an individual triangle or an individual body, we think that\nthey must be mutually consistent or, in other words, that they must\nhave the unity of being parts of a necessarily related whole or\nsystem, this unity in fact constituting the conception of an object\nin general, in distinction from the conception of an object of a\nparticular kind. The latter thought in turn involves the thought of\nthe object of representations as that which prevents them being\nanything whatever and in fact makes them parts of a system. The\nthought therefore of representations as related to an object carries\nwith it the thought of a certain necessity, viz. the necessary or\nsystematic unity introduced into the representations by the object.\nHence by an object of representations we mean something which\nintroduces into the representations a systematic unity which\nconstitutes the nature of an object in general, and the relatedness of\nrepresentations to, or their correspondence with, an object involves\ntheir systematic unity.'[42]\n\n    [40] Cf. A. 105, Mah. 199.\n\n    [41] Cf. A. 106, Mah. 200.\n\n    [42] It may be noticed that possession of the unity of a\n    system does not really distinguish 'an object' from any other\n    whole of parts, nor in particular from 'a representation'.\n    Any whole of parts must be a systematic unity.\n\nCertain points, however, should be noticed. In the _first_ place, Kant\nis for the moment tacitly ignoring his own theory of knowledge, in\naccordance with which the object proper, i. e. the thing in itself, is\nunknowable, and is reverting to the ordinary conception of knowledge\nas really _knowledge_ of its object. For the elements which are said,\nin virtue of being related to an object, to agree and to have the\nunity which constitutes the conception of an object must be elements\nof an object which we know; for if the assertion that they agree is\nto be significant, they must be determinate parts or qualities of the\nobject, e. g. the sides of an individual triangle or the\nimpenetrability or shape of an individual body, and therefore it is\nimplied that we know that the object has these parts or qualities. In\nthe _second_ place, both the problem which Kant raises and the clue\nwhich he offers for its solution involve an impossible separation of\nknowledge or a representation from its object. Kant begins with the\nthought of a phenomenon as a mere representation which, as mental, and\nas the representation of an object, is just not an object, and asks,\n'What is meant by the object of it?' He finds the clue to the answer\nin the thought that though a representation or idea when considered in\nitself is a mere mental modification, yet, when considered as related\nto an object, it is subject to a certain necessity. In fact, however,\nan idea or knowledge is essentially an idea or knowledge of an object,\nand we are bound to think of it as such. There is no meaning whatever\nin saying that the thought of an idea as related to an object carries\nwith it something of necessity, for to say so implies that it is\npossible to think of it as unrelated to an object. Similarly there is\nreally no meaning in the question, 'What is meant by an object\ncorresponding to knowledge or to an idea?' for this in the same way\nimplies that we can first think of an idea as unrelated to an object\nand then ask, 'What can be meant by an object corresponding to\nit?'[43] In the _third_ place, Kant only escapes the absurdity\ninvolved in the thought of a mere idea or a mere representation\nby treating representations either as parts or as qualities of an\nobject. For although he speaks of our cognitions,[44] i. e. of our\nrepresentations, as being determined by the object, he says that they\nmust agree, i. e. they must have that unity which constitutes the\nconception of an object, and he illustrates representations by the\nsides of an individual triangle and the impenetrability and shape of\nan individual body, which are just as 'objective' as the objects to\nwhich they relate. The fact is that he really treats a representation\nnot as his problem requires that it should be treated, i. e. as a\nrepresentation of something, but as something represented,[45] i. e.\nas something of which we are aware, viz. a part or a quality of an\nobject. In the _fourth_ place, not only is that which Kant speaks of\nas related to an object really not a representation, but also--as we\nsee if we consider the fact which Kant has in mind--that to which he\nspeaks of it as related is really not _an_ object but _one and the\nsame object to which another so-called representation is related_. For\nwhat Kant says is that representations as related to an object must\nagree among themselves. But this statement, to be significant, implies\nthat the object to which various representations are related is _one\nand the same_. Otherwise why should the representations agree? In\nview, therefore, of these last two considerations we must admit that\nthe real thought underlying Kant's statement should be expressed thus:\n'We find that the thought that _two or more parts or qualities of an\nobject_ relate to _one and the same object_ carries with it a certain\nnecessity, since this object is considered to be that which _prevents\nthese parts or qualities which we know it to possess_ from being\ndetermined at random, because by being related to _one and the same\nobject_, they must agree among themselves.' The importance of the\ncorrection lies in the fact that what Kant is stating is not what he\nthinks he is stating. He is really stating the implication of the\nthought that two or more qualities or parts of some object or other,\nwhich, as such, already relate to an object, relate to one and the\nsame object. He thinks he is stating the implication of the thought\nthat a representation which in itself has no relation to an object,\nhas relation to an object. And since his problem is simply to\ndetermine what constitutes the relatedness to an object of that which\nin itself is a mere representation, the distinction is important; for\nit shows that he really elucidates it by an implication respecting\nsomething which already has relation to an object and is not a mental\nmodification at all, but a quality or a part of an object.\n\n    [43] Cf. pp. 230-3.\n\n    [44] _Erkenntnisse._\n\n    [45] _Vorgestellt._\n\nKant continues thus: \"But it is clear that, since we have to do only\nwith the manifold of our representations, and the _x_, which\ncorresponds to them (the object), since it is to be something distinct\nfrom all our representations, is for us nothing, the unity which the\nobject necessitates can be nothing else than the formal unity of\nconsciousness in the synthesis of the manifold of representations.\"\n[I. e. since the object which produces systematic unity in our\nrepresentations is after all only the unknown thing in itself, viz.\n_x_,[46] any of the parts or qualities of which it is impossible to\nknow, that to which it gives unity can be only our representations\nand not its own parts or qualities. For, since we do not know any\nof its parts or qualities, these representations cannot be its\nparts or qualities. Consequently, the unity produced by this _x_\ncan only be the formal unity of the combination of the manifold in\nconsciousness.[47]] \"Then and then only do we say that we know the\nobject,\" [i. e. we know that the manifold relates to an object[48]]\n\"if we have produced synthetical unity in the manifold of perception.\nBut this unity would be impossible, if the perception could not be\nproduced by means of such a function of synthesis according to a rule\nas renders the reproduction of the manifold a priori necessary, and a\nconception in which the manifold unifies itself possible. Thus we\nthink a triangle as an object, in that we are conscious of the\ncombination of three straight lines in accordance with a rule by which\nsuch a perception can at any time be presented. This _unity of the\nrule_ determines all the manifold and limits it to conditions which\nmake the unity of apperception possible, and the conception of this\nunity is the representation of the object=_x_, which I think through\nthe aforesaid predicates of a triangle.\" [I. e., apparently, 'to\nconceive this unity of the rule is to represent to myself the object\n_x_, i. e. the thing in itself,[49] of which I come to think by means\nof the rule of combination.']\n\n    [46] Cf. p. 183, note 2.\n\n    [47] 'The formal unity' means not the unity peculiar to any\n    particular synthesis, but the character shared by all\n    syntheses of being a systematic whole.\n\n    [48] The final sense is the same whether 'object' be here\n    understood to refer to the thing in itself or to a\n    phenomenon.\n\n    [49] A comparison of this passage (A. 104-5, Mah. 198-9) with\n    A. 108-9, Mah. 201-2 (which seems to reproduce A. 104-5, Mah.\n    198-9), B. 522-3, M. 309 and A. 250, Mah. 224, seems to\n    render it absolutely necessary to understand by _x_, and by\n    the transcendental object, the thing in itself. Cf. also B.\n    236, M. 143 ('so soon as I raise my conception of an object\n    to the transcendental meaning thereof, the house is not a\n    thing in itself but only a phenomenon, i. e. a representation\n    of which the transcendental object is unknown'), A. 372, Mah.\n    247 and A. 379, Mah. 253.\n\nIn this passage several points claim attention. In the _first_ place,\nit seems impossible to avoid the conclusion that in the second\nsentence the argument is exactly reversed. Up to this point, it is\nthe thing in itself which produces unity in our representations.\nHenceforward it is we who produce the unity by our activity of\ncombining the manifold. The discrepancy cannot be explained away, and\nits existence can only be accounted for by the exigencies of Kant's\nposition. When he is asking 'What is meant by the object (beyond the\nmind) corresponding to our representations?' he has to think of the\nunity of the representations as due to the object. But when he is\nasking 'How does the manifold of sense become unified?' his view that\nall synthesis is due to the mind compels him to hold that the unity is\nproduced by us. In the _second_ place, the passage introduces a second\nobject in addition to the thing in itself, viz. the phenomenal object,\ne. g. a triangle considered as a whole of parts unified on a definite\nprinciple.[50] It is this object which, as the object that we know, is\nhenceforward prominent in the first edition, and has exclusive\nattention in the second. The connexion between this object and the\nthing in itself appears to lie in the consideration that we are only\njustified in holding that the manifold of sense is related to a thing\nin itself when we have unified it and therefore know it to be a unity,\nand that to know it to be a unity is _ipso facto_ to be aware of it as\nrelated to a phenomenal object; in other words, the knowledge that the\nmanifold is related to an object beyond consciousness is acquired\nthrough our knowledge of its relatedness to an object within\nconsciousness. In the _third_ place, in view of Kant's forthcoming\nvindication of the categories, it is important to notice that the\nprocess by which the manifold is said to acquire relation to an\nobject is illustrated by a synthesis on a particular principle which\nconstitutes the phenomenal object an object of a particular kind. The\nsynthesis which enables us to recognize three lines as an object is\nnot a synthesis based on general principles constituted by the\ncategories, but a synthesis based on the particular principle that the\nthree lines must be so put together as to form an enclosed space.\nMoreover, it should be noticed that the need of a particular principle\nis really inconsistent with his view that relation to an object gives\nthe manifold the systematic unity which constitutes the conception of\nan object, or that at least a [Greek: hysteron proteron] is involved.\nFor if the knowledge that certain representations form a systematic\nunity justifies our holding that they relate to an object, it would\nseem that in order to know that they relate to an object we need not\nknow the special character of their unity. Yet, as Kant states the\nfacts, we really have to know the special character of their unity in\norder to know that they possess systematic unity in general.[51]\n_Lastly_, it is easy to see the connexion of this account of an object\nof representations with the preceding account of the synthesis\ninvolved in knowledge. Kant had said that knowledge requires a\nsynthesis of the imagination in accordance with a definite principle,\nand the recognition of the principle of the synthesis by the\nunderstanding. From this point of view it is clear that the aim of the\npresent passage is to show that this process yields knowledge of an\nobject; for it shows that this process yields knowledge of a\nphenomenal object of a particular kind, e. g. of a triangle or of a\nbody, and that this object as such refers to what after all is _the_\nobject, viz. the thing in itself.\n\n    [50] Compare 'The object of our perceptions is merely that\n    something of which the conception expresses such a necessity\n    of synthesis' (A. 106, Mah. 200), and 'An object is that in\n    the conception of which the manifold of a given perception is\n    united' (B. 137, M. 84). Cf. also A. 108, Mah. 201.\n\n    [51] Kant's position is no doubt explained by the fact that\n    since the object corresponding to our representations is the\n    thing in itself, and since we only know that this is of the\n    same kind in the case of every representation, it can only be\n    thought of as producing systematic unity, and not a unity of\n    a particular kind. The position is also in part due to the\n    fact that the principles of synthesis involved by the\n    phenomenal object are usually thought of by Kant as the\n    categories; these of course can only contribute a general\n    kind of unity, and not the special kind of unity belonging to\n    an individual object.\n\nThe position reached by Kant so far is this. Knowledge, as being\nknowledge of an object, consists in a process by which the manifold of\nperception acquires relation to an object. This process again is a\nprocess of combination of the manifold into a systematic whole upon a\ndefinite principle, accompanied by the consciousness in some degree of\nthe act of combination, and therefore also of the acquisition by the\nmanifold of the definite unity which forms the principle of\ncombination. In virtue of this process there is said to be 'unity of\nconsciousness in the synthesis of the manifold', a phrase which the\ncontext justifies us in understanding as a condensed expression for a\nsituation in which (1) the manifold of sense is a unity of necessarily\nrelated parts, (2) there is _consciousness_ of this unity, and (3) the\nconsciousness which combines and is conscious of combining the\nmanifold, as being necessarily one and the same throughout this\nprocess, is itself a unity.\n\nKant then proceeds to introduce what he evidently considers the\nkeystone of his system, viz. 'transcendental apperception.'\n\n\"There is always a transcendental condition at the basis of any\nnecessity. Hence we must be able to find a transcendental ground of\nthe unity of consciousness in the synthesis of the manifold of all our\nperceptions, and therefore also of the conceptions of objects in\ngeneral, consequently also of all objects of experience, a ground\nwithout which it would be impossible to think any object for our\nperceptions; for this object is no more than that something, the\nconception of which expresses such a necessity of synthesis.\"\n\n\"Now this original and transcendental condition is no other than\n_transcendental apperception_. The consciousness of self according to\nthe determinations of our state in internal sense-perception is merely\nempirical, always changeable; there can be no fixed or permanent self\nin this stream of internal phenomena, and this consciousness is\nusually called _internal sense_ or _empirical apperception_. That\nwhich is _necessarily_ to be represented as numerically identical\ncannot be thought as such by means of empirical data. The condition\nwhich is to make such a transcendental presupposition valid must be\none which precedes all experience, and makes experience itself\npossible.\"\n\n\"Now no cognitions[52] can occur in us, no combination and unity of\nthem with one another, without that unity of consciousness which\nprecedes all data of perception, and by relation to which alone all\nrepresentation of objects is possible. This pure original unchangeable\nconsciousness I shall call _transcendental apperception_. That it\ndeserves this name is clear from the fact that even the purest\nobjective unity, viz. that of _a priori_ conceptions (space and time)\nis only possible by relation of perceptions to it. The numerical unity\nof this apperception therefore forms the _a priori_ foundation of all\nconceptions, just as the multiplicity of space and time is the\nfoundation of the perceptions of the sensibility.\"[53]\n\n    [52] _Erkenntnisse._\n\n    [53] A. 106-7, Mah. 200-1.\n\nThe argument is clearly meant to be 'transcendental' in character; in\nother words, Kant continues to argue from the existence of knowledge\nto the existence of its presuppositions. We should therefore expect\nthe passage to do two things: firstly, to show what it is which is\npresupposed by the 'unity of consciousness in the synthesis of the\nmanifold'[54]; and secondly, to show that this presupposition deserves\nthe title 'transcendental apperception'. Unfortunately Kant introduces\n'transcendental apperception' after the manner in which he introduced\nthe 'sensibility', the 'imagination' and the 'understanding', as if it\nwere a term with which every one is familiar, and which therefore\nneeds little explanation. To interpret the passage, it seems necessary\nto take it in close connexion with the preceding account of the three\n'syntheses' involved in knowledge, and to bear in mind that, as a\ncomparison of passages will show, the term 'apperception', which Kant\nborrows from Leibniz, always has for Kant a reference to consciousness\nof self or self-consciousness. If this be done, the meaning of the\npassage seems to be as follows:\n\n    [54] We should have expected this to have been already\n    accomplished. For according to the account already\n    considered, it is we who by our imagination introduce\n    necessity into the synthesis of the manifold and by our\n    understanding become conscious of it. We shall therefore not\n    be surprised to find that 'transcendental apperception' is\n    really only ourselves as exercising imagination and\n    understanding in a new guise.\n\n'To vindicate the existence of a self which is necessarily one and the\nsame throughout its representations, and which is capable of being\naware of its own identity throughout, it is useless to appeal to that\nconsciousness of ourselves which we have when we reflect upon our\nsuccessive states. For, although in being conscious of our states we\nare conscious of ourselves we are not conscious of ourselves as\nunchanging. The self as going through successive states is changing,\nand even if in fact its states did not change, its identity would be\nonly contingent; it need not continue unchanged. Consequently, the\nonly course possible is to show that the self-consciousness in\nquestion is presupposed in any experience or knowledge. Now it is so\npresupposed. For, as we have already shown, the relation of\nrepresentations to an object presupposes one consciousness which\ncombines and unifies them, and is at the same time conscious of the\nidentity of its own action in unifying them. This consciousness is the\nground of the unity of consciousness in the synthesis of the manifold.\nIt may fairly be called transcendental, because even a conception\nwhich relates to space or time, and therefore is the most remote from\nsensation, presupposes one consciousness which combines and unifies\nthe manifold of space and time through the conception, and is\nconscious of the identity of its own action in so doing. It may,\ntherefore, be regarded as the presupposition of _all_ conceiving or\nbringing a manifold under a conception, and therefore of all\nknowledge. Consequently, since knowledge is possible, i. e. since the\nmanifold of representations can be related to an object, there must be\none self capable of being aware of its own identity throughout its\nrepresentations.'\n\nAt this point of Kant's argument, however, there seems to occur an\ninversion of the thought. Hitherto, Kant has been arguing from the\npossibility of knowledge to the possibility of the consciousness of\nour own identity. But in the next paragraph he appears to reverse this\nprocedure and to argue from the possibility of self-consciousness to\nthe possibility of knowledge.\n\n\"But it is just this transcendental unity of apperception[55] which\nforms, from all possible phenomena which can be together in one\nexperience, a connexion of them according to laws. For this unity of\nconsciousness would be impossible, if the mind in the knowledge of the\nmanifold could not become conscious of the identity of the function\nwhereby it unites the manifold synthetically in one knowledge.\nConsequently, the original and necessary consciousness of the identity\nof oneself is at the same time a consciousness of an equally necessary\nunity of the synthesis of all phenomena according to conceptions,\ni. e. according to rules which not only make them necessarily\nreproducible, but thereby determine an object for their perception,\ni. e. determine the conception of something in which they are\nnecessarily connected. For the mind could not possibly think the\nidentity of itself in the manifold of its representations, and this\nindeed _a priori_, if it had not before its eyes the identity of its\naction which subjects all synthesis of apprehension (which is\nempirical) to a transcendental unity, and first makes possible its\nconnexion according to rules.\"\n\n    [55] Kant seems here and elsewhere to use the phrase\n    'transcendental unity of apperception' as synonymous with\n    'transcendental apperception', the reason, presumably, being\n    that transcendental apperception is a unity.\n\nThe argument seems indisputably to be as follows: 'The mind is\nnecessarily able to be aware of its own identity throughout its\nmanifold representations. To be aware of this, it must be aware of the\nidentity of the activity by which it combines the manifold of\nrepresentations into a systematic whole. Therefore it must be capable\nof combining, and of being conscious of its activity in combining, all\nphenomena which can be its representations into such a whole. But\nthis process, from the point of view of the representations combined,\nis the process by which they become related to an object and so enter\ninto knowledge. Therefore, since we are capable of being conscious of\nour identity with respect to all phenomena which can be our\nrepresentations, the process of combination and consciousness of\ncombination which constitutes knowledge must be possible with respect\nto them.' Thus the thought of this and the preceding paragraph seems\nto involve a circle. First the possibility of self-consciousness is\ndeduced from the possibility of knowledge, and then the possibility of\nknowledge is deduced from the possibility of self-consciousness.\n\nAn issue therefore arises, the importance of which can be seen by\nreference to the final aim of the 'deduction', viz. the vindication of\nthe categories. The categories are 'fundamental conceptions which\nenable us to think objects in general[56] for phenomena'[57]; in other\nwords, they are the principles of the synthesis by which the manifold\nof sense becomes related to an object. Hence, if this be granted, the\nproof that the categories are applicable to objects consists in\nshowing that the manifold can be subjected to this synthesis. The\nquestion therefore arises whether Kant's real starting-point for\nestablishing the possibility of this synthesis and therefore the\napplicability of the categories, is to be found in the possibility\nof knowledge, or in the possibility of self-consciousness, or in\nboth. In other words, does Kant start from the position that all\nrepresentations must be capable of being related to an object, or\nfrom the position that we must be capable of being conscious of our\nidentity with respect to all of them, or from both?\n\n    [56] _Objecte ueberhaupt_, i. e. objects of any kind in\n    distinction not from objects of a particular kind but from no\n    objects at all.\n\n    [57] A. 111, Mah. 204\n\nPrima facie the second position is the more plausible basis for the\ndesired conclusion. On the one hand, it does not seem obvious that the\nmanifold _must_ be capable of being related to an object; for even if\nit be urged that otherwise we should have only 'a random play of\nrepresentations, less than a dream'[58], it may be replied, that this\nmight be or might come to be the case. On the other hand, the fact\nthat our representations are ours necessarily seems to presuppose that\nwe are identical subjects of these representations, and recognition of\nthis fact is the consciousness of our identity.\n\n    [58] A. 112, Mah. 204.\n\nIf we turn to the text for an answer to this question, we find that\nKant seems not only to use both starting-points, but even to regard\nthem as equivalents. Thus in introducing the categories[59] Kant\nbegins by appealing to the necessity for knowledge that\nrepresentations should relate to an object.\n\n    [59] A. 110-12, Mah. 203-4.\n\n\"Unity of synthesis according to empirical conceptions would be purely\ncontingent, and were these not based on a transcendental ground of\nunity, it would be possible for a confused crowd of phenomena to fill\nour soul, without the possibility of experience ever arising\ntherefrom. But then also all relation of knowledge to objects would\nfall away, because knowledge would lack connexion according to\nuniversal and necessary laws; it would be thoughtless perception but\nnever knowledge, and therefore for us as good as nothing.\"\n\n\"The _a priori_ conditions of any possible experience whatever are at\nthe same time conditions of the possibility of the objects of\nexperience. Now I assert that the above mentioned _categories_ are\nnothing but _the conditions of thinking in any possible experience_,\njust as _space and time_ are the _conditions of perception_ requisite\nfor the same. The former therefore are also fundamental conceptions by\nwhich we think objects in general for phenomena, and are therefore\nobjectively valid _a priori_--which is exactly what we wished to\nknow.\"\n\nThe next sentence, however, bases the necessity of the categories on\nthe possibility of self-consciousness, without giving any indication\nthat a change of standpoint is involved.\n\n\"But the possibility, nay, even the necessity, of these categories\nrests on the relation which the whole sensibility, and with it also\nall possible phenomena, have to original apperception, a relation\nwhich forces everything to conform to the conditions of the\nthoroughgoing unity of self-consciousness, i. e. to stand under\nuniversal functions of synthesis, i. e. of synthesis according to\nconceptions, as that wherein alone apperception can prove _a priori_\nits thorough-going and necessary identity.\"\n\nFinally, the conclusion of the paragraph seems definitely to treat\nboth starting-points as really the same.[60] \"Thus the conception of a\ncause is nothing but a synthesis (of the consequent in the time series\nwith other phenomena) _according to conceptions_; and without such a\nunity, which has its _a priori_ rule and subjects phenomena to itself,\nthorough-going and universal and therefore necessary unity of\nconsciousness in the manifold of sense-perceptions would not be met\nwith. But then also these perceptions would belong to no experience,\nconsequently they would have no object, and would be nothing but a\nblind play of representations, less than a dream.\"\n\n    [60] Cf. A. 113, Mah. 205-6 and A. 108-10, Mah. 202-3.\n\nThe fact is that since for Kant the synthesis of representations in\naccordance with the categories, accompanied by the consciousness of\nit, is at once the necessary and sufficient condition of the\nrelatedness of representations to an object and of the consciousness\nof our identity with respect to them, it seems to him to be one and\nthe same thing whether, in vindicating the synthesis, we appeal to the\npossibility of knowledge or to the possibility of self-consciousness,\nand it even seems possible to argue, _via_ the synthesis, from\nknowledge to self-consciousness and vice versa.\n\nNevertheless, it remains true that the vindication of the categories\nis different, according as it is based upon the possibility of\nrelating representations to an object or upon the possibility of\nbecoming self-conscious with respect to them. It also remains true\nthat Kant vindicates the categories in both ways. For while, in\nexpounding the three so-called syntheses involved in knowledge, he is\nvindicating the categories from the point of view of knowledge, when\nhe comes to speak of transcendental apperception, of which the central\ncharacteristic is the consciousness of self involved, there is a\nshifting of the centre of gravity. Instead of treating representations\nas something which can become related to an object, he now treats\nthem as something of which, as belonging to a self, the self must\nbe capable of being conscious as its own, and argues that a\nsynthesis in accordance with the categories is required for this\nself-consciousness. It must be admitted then--and the admission is\nonly to be made with reluctance--that when Kant reaches transcendental\napperception, he really adopts a new starting-point,[61] and that the\npassage which introduces transcendental apperception by showing it to\nbe implied in knowledge[62] only serves to conceal from Kant the fact\nthat, from the point of view of the deduction of the categories, he is\nreally assuming without proof the possibility of self-consciousness\nwith respect to all our representations, as a new basis for argument.\n\n    [61] The existence of this new starting-point is more\n    explicit, A. 116-7 (and note), Mah. 208 (and note), and A.\n    122, Mah. 212.\n\n    [62] A. 107, Mah. 200.\n\nThe approach to the categories from the side of self-consciousness is,\nhowever, more prominent in the second edition, and consequently we\nnaturally turn to it for more light on this side of Kant's position.\nThere Kant vindicates the necessity of the synthesis from the side of\nself-consciousness as follows:[63]\n\n    [63] The main clauses have been numbered for convenience of\n    reference.\n\n\"[1.] It must be possible that the 'I think' should accompany all my\nrepresentations; for otherwise something would be represented in me\nwhich could not be thought; in other words, the representation would\nbe either impossible or at least for me nothing. [2.] That\nrepresentation which can be given before all thought is called\n_perception_. All the manifold of perception has therefore a necessary\nrelation to the 'I think' in the same subject in which this manifold\nis found. [3.] But this representation[64] [i. e. the 'I think'] is an\nact of _spontaneity_, i. e. it cannot be regarded as belonging to\nsensibility. I call it _pure apperception_, to distinguish it from\n_empirical apperception_, or _original apperception_ also, because it\nis that self-consciousness which, while it gives birth to the\nrepresentation 'I think', which must be capable of accompanying all\nothers and is one and the same in all consciousness, cannot itself be\naccompanied by any other.[65] [4.] I also call the unity of it the\n_transcendental_ unity of self-consciousness, in order to indicate the\npossibility of _a priori_ knowledge arising from it. For the manifold\nrepresentations which are given in a perception would not all of them\nbe _my_ representations, if they did not all belong to one\nself-consciousness, that is, as my representations (even though I am\nnot conscious of them as such), they must necessarily conform to the\ncondition under which alone they _can_ stand together in a universal\nself-consciousness, because otherwise they would not all belong to me.\nFrom this original connexion much can be concluded.\"\n\n    [64] This is an indisputable case of the use of\n    representation in the sense of something represented or\n    presented.\n\n    [65] I. e. consciousness of our identity is final; we cannot,\n    for instance, go further back to a consciousness of the\n    consciousness of our identity.\n\n[5.] \"That is to say, this thorough-going identity of the apperception\nof a manifold given in perception contains a synthesis of\nrepresentations,[66] and is possible only through the consciousness of\nthis synthesis.[67] [6.] For the empirical consciousness which\naccompanies different representations is in itself fragmentary, and\nwithout relation to the identity of the subject. [7.] This relation,\ntherefore, takes place not by my merely accompanying every\nrepresentation with consciousness, but by my _adding_ one\nrepresentation to another, and being conscious of the synthesis of\nthem. [8.] Consequently, only because I can connect a manifold of\ngiven representations _in one consciousness_, is it possible for me to\nrepresent to myself the _identity of consciousness in these\nrepresentations_; i. e. the _analytical_ unity of apperception is\npossible only under the presupposition of a _synthetical_ unity. [9.]\nThe thought, 'These representations given in perception belong all of\nthem to me' is accordingly just the same as, 'I unite them in one\nself-consciousness, or at least can so unite them;' [10.] and although\nthis thought is not itself as yet the consciousness of the _synthesis_\nof representations, it nevertheless presupposes the possibility of\nthis synthesis; that is to say, it is only because I can comprehend\nthe manifold of representations in one consciousness, that I call them\nall _my_ representations; for otherwise I should have as many-coloured\nand varied a self as I have representations of which I am conscious.\n[11.] Synthetical unity of the manifold of perceptions, as given _a\npriori_, is therefore the ground of the identity of apperception\nitself, which precedes _a priori_ all _my_ determinate thinking. [12.]\nBut connexion does not lie in the objects, nor can it be borrowed from\nthem through perception and thereby first taken up into the\nunderstanding, but it is always an operation of the understanding\nwhich itself is nothing more than the faculty of connecting _a\npriori_, and of bringing the manifold of given representations under\nthe unity of apperception, which principle is the highest in all human\nknowledge.\"\n\n    [66] I understand this to mean 'This through and through\n    identical consciousness of myself as the identical subject of\n    a manifold given in perception involves a synthesis of\n    representations'.\n\n    [67] The drift of the passage as a whole (cf. especially\n    Sec. 16) seems to show that here 'the synthesis of\n    representations' means 'their connectedness' and not\n    'the act of connecting them'.\n\n[13.] \"Now this principle of the necessary unity of apperception is\nindeed an identical, and therefore an analytical, proposition, but\nnevertheless it declares a synthesis of the manifold given in a\nperception to be necessary, without which the thorough-going identity\nof self-consciousness cannot be thought. [14.] For through the Ego, as\na simple representation, is given no manifold content; in perception,\nwhich is different from it, a manifold can only be given, and through\n_connexion_ in one consciousness it can be thought. An understanding,\nthrough whose self-consciousness all the manifold would _eo ipso_ be\ngiven, would _perceive_; our understanding can only _think_ and must\nseek its perception in the senses. [15.] I am, therefore, conscious of\nthe identical self, in relation to the manifold of representations\ngiven to me in a perception, because I call all those representations\n_mine_, which constitute _one_. [16.] But this is the same as to say\nthat I am conscious _a priori_ of a necessary synthesis of them, which\nis called the original synthetic unity of apperception, under which\nall representations given to me stand, but also under which they must\nbe brought through a synthesis.\"[68]\n\n    [68] B. 131-5, M. 81-4.\n\nThough this passage involves many difficulties, the main drift of it\nis clear. Kant is anxious to establish the fact that the manifold of\nsense must be capable of being combined on principles, which\nafterwards turn out to be the categories, by showing this to be\ninvolved in the fact that we must be capable of being conscious of\nourselves as the identical subject of all our representations. To do\nthis, he seeks to prove in the first paragraph that self-consciousness\nin this sense must be possible, and in the second that this\nself-consciousness presupposes the synthesis of the manifold.\n\nExamination of the argument, however, shows that the view that\nself-consciousness must be possible is, so far as Kant is\nconcerned,[69] an assumption for which Kant succeeds in giving no\nreason at all, and that even if it be true, it cannot form a basis\nfrom which to deduce the possibility of the synthesis.\n\n    [69] Cf. p. 204, note 3.\n\nBefore, however, we attempt to prove this, it is necessary to draw\nattention to three features of the argument. In the _first_ place, it\nimplies a somewhat different account of self-consciousness to that\nimplied in the passages of the first edition which we have already\nconsidered. Self-consciousness, instead of being the consciousness of\nthe identity of our activity in combining the manifold, is now\nprimarily the consciousness of ourselves as identical subjects of all\nour representations, i. e. it is what Kant calls the analytical unity\nof apperception; and consequently it is somewhat differently related\nto the activity of synthesis involved in knowledge. Instead of being\nregarded as the consciousness of this activity, it is regarded as\npresupposing the consciousness of the product of this activity, i. e.\nof the connectedness[70] of the manifold produced by the activity,\nthis consciousness being what Kant calls the synthetical unity of\napperception.[71] In the _second_ place, it is plain that Kant's view\nis not that self-consciousness involves the consciousness of our\nrepresentations as a connected whole, but that it involves the\nconsciousness of them as capable of being connected by a synthesis.\nYet, if it is only because I can connect (and therefore apprehend as\nconnected) a manifold of representations in one consciousness, that I\ncan represent to myself the identity of consciousness in these\nrepresentations, self-consciousness really requires the consciousness\nof our representations as _already_ connected; the mere consciousness\nof our representations as _capable_ of being connected would not be\nenough. The explanation of the inconsistency seems to lie in the fact\nthat the synthetic unity of which Kant is thinking is the unity\nof nature. For, as Kant of course was aware, in our ordinary\nconsciousness we do not apprehend the interconnexion of the parts\nof nature in detail, but only believe that there is such an\ninterconnexion; consequently he naturally weakened the conclusion\nwhich he ought to have drawn, viz. that self-consciousness presupposes\nconsciousness of the synthesis, in order to make it conform to the\nfacts of our ordinary consciousness. Yet, if his _argument_ is\nto be defended, its conclusion must be taken in the form that\nself-consciousness presupposes consciousness of the actual synthesis\nor connexion and not merely of the possibility of it. In the _third_\nplace, Kant twice in this passage[72] definitely makes the act of\nsynthesis, which his argument maintains to be the condition of\n_consciousness of the identity_ of ourselves, the condition of the\n_identity_ of ourselves. The fact is that, on Kant's view, the act of\nsynthesis of the representations is really a condition of their\nbelonging to one self, the self being presupposed to be a self capable\nof self-consciousness.[73]\n\n    [70] More accurately, 'of the possibility of the\n    connectedness'.\n\n    [71] The same view seems implied A. 117-8, Mah. 208. Kant\n    apparently thinks of this consciousness as also a\n    self-consciousness (cf. Sec. 9), though it seems that he\n    should have considered it rather as a condition of\n    self-consciousness, cf. p. 204, note 2.\n\n    [72] Sections 6 and 10.\n\n    [73] Cf. pp. 202-3.\n\nWe may now turn to the first of the two main points to be considered,\nviz. the reason given by Kant for holding that self-consciousness must\nbe possible. In the first paragraph (Secs. 1-4) Kant appears twice to\nstate a reason, viz. in Secs. 1 and 4. What is meant by the first\nsentence, \"It must be possible that the 'I think' should accompany all\nmy representations; for otherwise something would be represented in me\nwhich could not be thought; in other words, the representation would\neither be impossible or at least for me nothing\"? It is difficult to\nhold that 'my representations' here means objects of which I am aware,\nand that the thesis to be established is that I must be capable of\nbeing conscious of my own identity throughout all awareness or thought\nof objects. For the next sentence refers to perceptions as\nrepresentations which can be given previously to all thought, and\ntherefore, presumably, as something of which I am not necessarily\naware. Again, the ground adduced for the thesis would be in part a\nmere restatement of it, and in part nonsense. It would be 'otherwise\nsomething would be apprehended with respect to which I could not be\naware that I was apprehending it; in other words, I could not\napprehend it [since otherwise I could be aware that I was apprehending\nit]', the last words being incapable of any interpretation. It is much\nmore probable that though Kant is leading up to self-consciousness,\nthe phrase 'I think' here refers not to 'consciousness that I am\nthinking', but to 'thinking'. He seems to mean 'It must be possible to\napprehend all my 'affections' (i. e. sensations or appearances in me),\nfor otherwise I should have an affection of which I could not be\naware; in other words, there could be no such affection, or at least\nit would be of no possible importance to me.'[74] And on this\ninterpretation self-consciousness is not introduced till Sec. 3, and\nthen only surreptitiously. On neither interpretation, however, does\nKant give the vestige of a _reason_ for the possibility of\nself-consciousness. Again, it seems clear that in Sec. 4 'my\nrepresentations', and 'representations which belong to me' mean\nobjects of which I am aware (i. e. something presented); for he says\nof my representations, not that I may not be conscious of them--which\nhe should have said if 'my representations' meant my mental affections\nof which I could become conscious--but that I may not be conscious of\nthem as my representations. Consequently in Sec. 4 he is merely\nasserting that I must be able to be conscious of my identity throughout\nmy awareness of objects. So far, then, we find merely the _assertion_\nthat self-consciousness must be possible.[75]\n\n    [74] A third alternative is to understand Kant to be thinking\n    of all thought as self-conscious, i. e. as thinking\n    accompanied by the consciousness of thinking. But since in\n    that case Kant would be arguing from thinking as _thinking_,\n    i. e. as apprehending objects, the possibility of\n    self-consciousness would only be glaringly assumed.\n\n    [75] The same is true of A. 116 and A. 117 note, Mah. 208,\n    where Kant also appears to be offering what he considers to\n    be an argument.\n\nIn the next paragraph[76]--which is clearly meant to be the important\none--Kant, though he can hardly be said to be aware of it, seems to\n_assume_ that it is the very nature of a knowing self, not only to be\nidentical throughout its thoughts or apprehendings, but to be capable\nof being conscious of its own identity. Sec. 6 runs: \"The empirical\nconsciousness which accompanies different representations is in itself\nfragmentary, and without relation to the identity of the subject.\"\nKant is saying that if there existed merely a consciousness of A which\nwas not at the same time a consciousness of B and a consciousness\nof B which was not at the same time a consciousness of A, these\nconsciousnesses would not be the consciousnesses belonging to one\nself. But this is only true, if the one self to which the\nconsciousness of A and the consciousness of B are to belong must be\ncapable of being aware of its own identity. Otherwise it might be one\nself which apprehended A and then, forgetting A, apprehended B. No\ndoubt in that case the self could not be aware of its own identity in\napprehending A and in apprehending B, but none the less it would _be_\nidentical in so doing. We reach the same conclusion if we consider the\nconcluding sentence of Sec. 10. \"It is only because I can comprehend the\nmanifold of representations in one consciousness, that I call them all\nmy representations; for otherwise I should have as many-coloured and\nvaried a self as I have representations of which I am conscious.\"\nDoubtless if I am to _be aware of_ myself as the same in apprehending\nA and B, then, in coming to apprehend B, I must continue to apprehend\nA, and therefore must apprehend A and B as related; and such a\nconsciousness on Kant's view involves a synthesis. But if I am merely\nto _be_ the same subject which apprehends A and B, or rather if the\napprehension of A and that of B are merely to _be_ apprehensions on\nthe part of one and the same subject, no such consciousness of A and B\nas related and, therefore, no synthesis is involved.\n\n    [76] Secs. 5-11.\n\nAgain, the third paragraph assumes the possibility of\nself-consciousness as the starting-point for argument. The thought[77]\nseems to be this: 'For a self to be aware of its own identity, there\nmust be a manifold in relation to which it can apprehend itself as one\nand the same throughout. An understanding which was perceptive, i. e.\nwhich originated objects by its own act of thinking, would necessarily\nby its own thinking originate a manifold in relation to which it could\nbe aware of its own identity in thinking, and therefore its\nself-consciousness would need no synthesis. But our understanding,\nwhich is not perceptive, requires a manifold to be given to it, in\nrelation to which it can be aware of its own identity by means of a\nsynthesis of the manifold.' If this be the thought, it is clearly\npresupposed that _any_ understanding must be capable of being\nconscious of its own identity.[78]\n\n    [77] Cf. B. 138 fin.-139 init., M. 85 fin.\n\n    [78] B. 139 init., M. 85 fin. also assumes that it is\n    impossible for a mind to be a unity without being able to be\n    conscious of its unity.\n\nFurther, it is easy to see how Kant came to take for granted the\npossibility of self-consciousness, in the sense of the consciousness\nof ourselves as the identical subject of all our representations. He\napproaches self-consciousness with the presupposition derived from his\nanalysis of knowledge that our apprehension of a manifold does not\nconsist in separate apprehensions of its elements, but is one\napprehension or consciousness of the elements as related.[79] He\nthinks of this as a general presupposition of all apprehension of a\nmanifold, and, of course, to discover this presupposition is to be\nself-conscious. To recognize the oneness of our apprehension is to be\nconscious of our own identity.[80]\n\n    [79] It is in consequence of this that the statement that 'a\n    manifold of representations belongs to me' means, with the\n    probable exception of Sec. 1, not, 'I am aware of A, I am\n    aware of B, I am aware of C,' but, 'I am aware, in one act of\n    awareness, of A B C as related' (= ABC are 'connected in' or\n    'belong to' one consciousness). Cf. Secs. 4, 8 ('in one\n    consciousness'), 9, 10 ('in one consciousness'), and A. 116,\n    Mah. 208 ('These representations only represent anything in\n    me by belonging with all the rest to one consciousness\n    [accepting Erdmann's emendation _mit allen anderen_], in\n    which at any rate they can be connected').\n\n    [80] The above criticism of Kant's thought has not implied\n    that it may not be true that a knowing mind is, as such,\n    capable of being aware of its own unity; the argument has\n    only been that Kant's proof is unsuccessful.\n\nAgain, to pass to the second main point to be considered,[81]\nKant has no justification for arguing from the possibility of\nself-consciousness to that of the synthesis. This can be seen from the\nmere form of his argument. Kant, as has been said, seems first to\nestablish the possibility of self-consciousness, and thence to\nconclude that a synthesis must be possible. But if, as it is his point\nto urge, consciousness of our identity only takes place through\nconsciousness of the synthesis, this method of argument must be\ninvalid. It would clearly be necessary to know that the synthesis is\npossible, _before_ and _in order that_ we could know that\nself-consciousness is possible. An objector has only to urge that the\nmanifold might be such that it could not be combined into a systematic\nwhole, in order to secure the admission that in that case\nself-consciousness would not be possible.\n\n    [81] Cf. p. 198.\n\nNevertheless, the passage under consideration may be said to lay bare\nan important presupposition of self-consciousness. It is true that\nself-consciousness would be impossible, if we merely apprehended the\nparts of the world in isolation. To be conscious that I who am\nperceiving C perceived B and A, I must be conscious at once of A, B,\nand C, in one act of consciousness or apprehension. To be conscious\nseparately of A and B and C is not to be conscious of A and B and C.\nAnd, to be conscious of A and B and C in one act of consciousness, I\nmust apprehend A, B, and C as related, i. e. as forming parts of a\nwhole or system. Hence it is only because our consciousness of A, B,\nand C is never the consciousness of a mere A, a mere B, and a mere C,\nbut is always the consciousness of A B C as elements in one world that\nwe can be conscious of our identity in apprehending A, B, and C. If\n_per impossibile_ our apprehension be supposed to cease to be an\napprehension of a plurality of objects in relation, self-consciousness\nmust be supposed to cease also. At the same time, it is impossible to\nargue from the consciousness of our identity in apprehending to the\nconsciousness of what is apprehended as a unity, and thence to the\nexistence of that unity. For, apart from the consideration that in\nfact all thinking presupposes the relatedness or--what is the same\nthing--the necessary relatedness of objects to one another, and that\ntherefore any assertion to the contrary is meaningless, the\nconsciousness of objects as a unity is a condition of the\nconsciousness of our identity, and therefore any doubt that can be\nraised in regard to the former can be raised equally with regard to\nthe latter.\n\nWe may now pass to the concluding portion of the deduction. For the\npurpose of considering it, we may sum up the results of the preceding\ndiscussion by saying that Kant establishes the synthesis of the\nmanifold on certain principles by what are really two independent\nlines of thought. The manifold may be regarded either as something\nwhich, in order to enter into knowledge, must be given relation to an\nobject, or as something with respect to which self-consciousness must\nbe possible. Regarded in either way, the manifold, according to Kant,\ninvolves a process of synthesis on certain principles, which makes it\na systematic unity. Now Kant introduces the categories by maintaining\nthat they are the principles of synthesis in question. \"I assert that\nthe above mentioned _categories_ are nothing but the _conditions of\nthinking in a possible experience_.... They are fundamental\nconceptions by which we think objects in general for phenomena.\"[82]\nA synthesis according to the categories is 'that wherein alone\napperception can prove _a priori_ its thorough-going and necessary\nidentity'.[83] In the first edition this identification is simply\nasserted, but in the second Kant offers a proof.[84]\n\n    [82] A. 111, Mah. 204. Cf. A. 119, Mah. 210.\n\n    [83] A. 112, Mah. 204.\n\n    [84] Cf. p. 161.\n\nBefore, however, we consider the proof, it is necessary to refer\nto a difficulty which seems to have escaped Kant altogether. The\npreceding account of the synthesis involved in knowledge and in\nself-consciousness implies, as his illustrations conclusively show,\nthat the synthesis requires a particular principle which constitutes\nthe individual manifold a whole of a particular kind.[85] But, if this\nbe the case, it is clear that the categories, which are merely\nconceptions of an object in general, and are consequently quite\ngeneral, cannot possibly be sufficient for the purpose. And since the\nmanifold in itself includes no synthesis and therefore no principle of\nsynthesis, Kant fails to give any account of the source of the\nparticular principles of synthesis required for particular acts of\nknowledge.[86] This difficulty--which admits of no solution--is\nconcealed from Kant in two ways. In the first place, when he describes\nwhat really must be stated as the process by which parts or qualities\nof an object become related to an object of a particular kind, he\nthinks that he is describing a process by which representations become\nrelated to an object in general.[87] Secondly, he thinks of the\nunderstanding as the source of general principles of synthesis,\nindividual syntheses and the particular principles involved being\nattributed to the imagination; and so, when he comes to consider the\npart played in knowledge by the understanding, he is apt to ignore the\nneed of particular principles.[88] Hence, Kant's proof that the\ncategories are the principles of synthesis can at best be taken only\nas a proof that the categories, though not sufficient for the\nsynthesis, are involved in it.\n\n    [85] Cf. p. 177, note 2, and p. 185.\n\n    [86] Cf. pp. 215-17.\n\n    [87] Cf. pp. 181-2.\n\n    [88] Cf. p. 217.\n\nThe proof runs thus:\n\n\"I could never satisfy myself with the definition which logicians\ngive of a judgement in general. It is, according to them, the\nrepresentation of a relation between two conceptions....\"\n\n\"But if I examine more closely the relation of given\nrepresentations[89] in every judgement, and distinguish it, as\nbelonging to the understanding, from their relation according to the\nlaws of the reproductive imagination (which has only subjective\nvalidity), I find that a judgement is nothing but the mode of bringing\ngiven representations under the _objective_ unity of apperception.\nThis is what is intended by the term of relation 'is' in judgements,\nwhich is meant to distinguish the objective unity of given\nrepresentations from the subjective. For this term indicates the\nrelation of these representations to the original apperception, and\nalso their _necessary unity_, even though the judgement itself is\nempirical, and therefore contingent, e. g. 'Bodies are heavy.' By this\nI do not mean that these representations _necessarily_ belong _to each\nother_ in empirical perception, but that they belong to each other _by\nmeans of the necessary unity_ of apperception in the synthesis of\nperceptions, that is, according to principles of the objective\ndetermination of all our representations, in so far as knowledge can\narise from them, these principles being all derived from the principle\nof the transcendental unity of apperception. In this way alone can\nthere arise from this relation _a judgement_, that is, a relation\nwhich is _objectively valid_, and is adequately distinguished from the\nrelation of the very same representations which would be only\nsubjectively valid, e. g. according to laws of association. According\nto these laws, I could only say, 'If I carry a body, I feel an\nimpression of weight', but not 'It, the body, _is_ heavy'; for this is\ntantamount to saying, 'These two representations are connected in the\nobject, that is, without distinction as to the condition of the\nsubject, and are not merely connected together in the perception,\nhowever often it may be repeated.'\"[90]\n\n    [89] _Erkenntnisse_ here is clearly used as a synonym for\n    representations. Cf. A. 104, Mah. 199.\n\n    [90] B. 140-2, M. 86-8; cf. _Prol._, Secs. 18-20.\n\nThis ground for the identification of the categories with the\nprinciples of synthesis involved in knowledge may be ignored, as on\nthe face of it unsuccessful. For the argument is that since the\nactivity by which the synthesis is affected is that of judgement, the\nconceptions shown by the _Metaphysical Deduction_ to be involved in\njudgement must constitute the principles of synthesis. But it is\nessential to this argument that the present account of judgement and\nthat which forms the basis of the _Metaphysical Deduction_ should be\nthe same; and this is plainly not the case.[91] Judgement is now\nrepresented as an act by which we relate the manifold of sense in\ncertain necessary ways as parts of the physical world,[92] whereas in\nthe _Metaphysical_ _Deduction_ it was treated as an act by which we\nrelate conceptions; and Kant now actually says that this latter\naccount is faulty. Hence even if the metaphysical deduction had\nsuccessfully derived the categories from the account of judgement\nwhich it presupposed, the present argument would not justify the\nidentification of the categories so deduced with the principles of\nsynthesis. The fact is that Kant's vindication of the categories is in\nsubstance independent of the _Metaphysical Deduction_. Kant's real\nthought, as opposed to his formal presentation of it, is simply that\nwhen we come to consider what are the principles of synthesis involved\nin the reference of the manifold to an object, we find that they are\nthe categories.[93] The success, then, of this step in Kant's\nvindication of the categories is independent of that of the\nmetaphysical deduction, and depends solely upon the question whether\nthe principles of synthesis involved in knowledge are in fact the\ncategories.\n\n    [91] Cf. Caird, i. 348-9 note.\n\n    [92] We may notice in passing that this passage renders\n    explicit the extreme difficulty of Kant's view that 'the\n    objective unity of apperception' is the unity of the parts of\n    nature or of the physical world. How can the 'very same\n    representations' stand at once in the subjective relation of\n    association and in the objective relation which consists in\n    their being related as parts of nature? There is plainly\n    involved a transition from representation, in the sense of\n    the apprehension of something, to representation, in the\n    sense of something apprehended. It is objects apprehended\n    which are objectively related; it is our apprehensions of\n    objects which are associated, cf. pp. 233 and 281-2. Current\n    psychology seems to share Kant's mistake in its doctrine of\n    association of ideas, by treating the elements associated,\n    which are really apprehensions of objects, as if they were\n    objects apprehended.\n\n    [93] Cf. A. 112, Mah. 204; B. 162, M. 99.\n\nThe substance of Kant's vindication of the categories may therefore be\nepitomized thus: 'We may take either of two starting-points. On the\none hand, we may start from the fact that our experience is no mere\ndream, but an intelligent experience in which we are aware of a world\nof individual objects. This fact is conceded even by those who, like\nHume, deny that we are aware of any necessity of relation between\nthese objects. We may then go on to ask how it comes about that,\nbeginning as we do with a manifold of sense given in succession, we\ncome to apprehend this world of individual objects. If we do so, we\nfind that there is presupposed a synthesis on our part of the manifold\nupon principles constituted by the categories.\n\nTo deny, therefore, that the manifold is so connected is implicitly to\ndeny that we have an apprehension of objects at all. But the existence\nof this apprehension is plainly a fact which even Hume did not\ndispute. On the other hand, we may start with the equally obvious fact\nthat we must be capable of apprehending our own identity throughout\nour apprehension of the manifold of sense, and look for the\npresupposition of this fact. If we do this, we again find that there\nis involved a combination of the manifold according to the\ncategories.'\n\nIn conclusion, attention may be drawn to two points. In the first\nplace, Kant completes his account by at once emphasizing and\nexplaining the paradoxical character of his conclusion. \"Accordingly,\nthe order and conformity to law in the phenomena which we call\n_nature_ we ourselves introduce, and we could never find it there,\nif we, or the nature of our mind, had not originally placed it\nthere.\"[94] \"However exaggerated or absurd then it may sound to say\nthat the understanding itself is the source of the laws of nature and\nconsequently of the formal unity of nature, such an assertion is\nnevertheless correct and in accordance with the object, i. e. with\nexperience.\"[95] The explanation of the paradox is found in the fact\nthat objects of nature are phenomena. \"But if we reflect that this\nnature is in itself nothing else than a totality[96] of phenomena and\nconsequently no thing in itself but merely a number of representations\nof the mind, we shall not be surprised that only in the radical\nfaculty of all our knowledge, viz. transcendental apperception, do we\nsee it in that unity through which alone it can be called object of\nall possible experience, i. e. nature.\"[97] \"It is no more surprising\nthat the laws of the phenomena in nature must agree with the\nunderstanding and with its _a priori_ form, that is, its faculty of\nconnecting the manifold in general, than that the phenomena themselves\nmust agree with the _a priori_ form of our sensuous perception. For\nlaws exist in the phenomena as little as phenomena exist in\nthemselves; on the contrary, laws exist only relatively to the subject\nin which the phenomena inhere, so far as it has understanding, just as\nphenomena exist only relatively to the subject, so far as it has\nsenses. To things in themselves their conformity to law would\nnecessarily also belong independently of an understanding which knows\nthem. But phenomena are only representations of things which exist\nunknown in respect of what they may be in themselves. But, as mere\nrepresentations, they stand under no law of connexion except that\nwhich the connecting faculty prescribes.\"[98]\n\n    [94] A. 125, Mah. 214.\n\n    [95] A. 127, Mah. 216.\n\n    [96] _Inbegriff._\n\n    [97] A. 114, Mah. 206.\n\n    [98] B. 164, M. 100.\n\nIn the second place, this last paragraph contains the real reason from\nthe point of view of the deduction[99] of the categories for what may\nbe called the negative side of his doctrine, viz. that the categories\nonly apply to objects of experience and not to things in themselves.\nAccording to Kant, we can only say that certain principles of\nconnexion apply to a reality into which we introduce the connexion.\nThings in themselves, if connected, are connected in themselves and\napart from us. Hence there can be no guarantee that any principles of\nconnexion which we might assert them to possess are those which they\ndo possess.\n\n    [99] The main passage (B. 146-9, M. 90-2), in which he argues\n    that the categories do not apply to things in themselves,\n    ignores the account of a conception as a principle of\n    synthesis, upon which the deduction turns, and returns to the\n    earlier account of a conception as something opposed to a\n    perception, i. e. as that by which an object is thought as\n    opposed to a perception by which an object is given.\n    Consequently, it argues merely that the categories, as\n    conceptions, are empty or without an object, unless an object\n    is given in perception, and that, since things in themselves\n    are not objects of perception, the categories are no more\n    applicable to things in themselves than are any other\n    conceptions.\n\n\n\n\nCHAPTER IX\n\nGENERAL CRITICISM OF THE TRANSCENDENTAL DEDUCTION OF THE CATEGORIES\n\n\nThe preceding account of Kant's vindication of the categories has\nincluded much criticism. But the criticism has been as far as possible\nrestricted to details, and has dealt with matters of principle only so\nfar as has been necessary in order to follow Kant's thought. We must\nnow consider the position as a whole, even though this may involve\nsome repetition.[1] The general difficulties of the position may be\ndivided into two kinds, (1) difficulties involved in the working out\nof the theory, even if its main principles are not questioned, and (2)\ndifficulties involved in accepting its main principles at all.\n\n    [1] Difficulties connected with Kant's view of\n    self-consciousness will be ignored, as having been\n    sufficiently considered.\n\nThe initial difficulty of the first kind, which naturally strikes the\nreader, concerns the possibility of performing the synthesis. The mind\nhas certain general ways of combining the manifold, viz. the\ncategories. But on general grounds we should expect the mind to\npossess only one mode of combining the manifold. For the character of\nthe manifold to be combined cannot affect the mind's power of\ncombination, and, if the power of the mind consists in combining, the\ncombining should always be of the same kind. Thus, suppose the\nmanifold given to the mind to be combined consisted of musical notes,\nwe could think of the mind's power of combination as exercised in\ncombining the notes by way of succession, _provided that_ this be\nregarded as the only mode of combination. But if the mind were thought\nalso capable of combining notes by way of simultaneity, we should at\nonce be confronted with the insoluble problem of determining why the\none mode of combination was exercised in any given case rather than\nthe other. If, several kinds of synthesis being allowed, this\ndifficulty be avoided by the supposition that, not being incompatible,\nthey are all exercised together, we have the alternative task of\nexplaining how the same manifold can be combined in each of these\nways. As a matter of fact, Kant thinks of manifolds of different kinds\nas combined or related in different ways; thus events are related\ncausally and quantities quantitatively. But since, on Kant's view, the\nmanifold as given is unrelated and all combination comes from the\nmind, the mind should not be held capable of combining manifolds of\ndifferent kinds differently. Otherwise the manifold would in its own\nnature imply the need of a particular kind of synthesis, and would\ntherefore not be unrelated.\n\nSuppose, however, we waive the difficulty involved in the plurality of\nthe categories. There remains the equally fundamental difficulty that\nany single principle of synthesis contains in itself no ground for the\ndifferent ways of its application.[2] Suppose it to be conceded that\nin the apprehension of definite shapes we combine the manifold in\naccordance with the conception of figure, and, for the purpose of the\nargument, that the conception of figure can be treated as equivalent\nto the category of quantity. It is plain that we apprehend different\nshapes, e. g. lines[3] and triangles[4], of which, if we take into\naccount differences of relative length of sides, there is an infinite\nvariety, and houses,[5] which may also have an infinite variety of\nshape. But there is nothing in the mind's capacity of relating the\nmanifold by way of figure to determine it to combine a given manifold\ninto a figure of one kind rather than into a figure of any other kind;\nfor to combine the manifold into a particular shape, there is needed\nnot merely the thought of a figure in general, but the thought of a\ndefinite figure. No 'cue' can be furnished by the manifold itself, for\nany such cue would involve the conception of a particular figure, and\nwould therefore imply that the particular synthesis was implicit in\nthe manifold itself, in which case it would not be true that all\nsynthesis comes from the mind.\n\n    [2] Cf. p. 207.\n\n    [3] B. 137, M. 85.\n\n    [4] A. 105, Mah. 199.\n\n    [5] B. 162, M. 99.\n\nThis difficulty takes a somewhat different form in the case of the\ncategories of relation. To take the case of cause and effect,\nthe conception of which, according to Kant, is involved in our\napprehension of a succession, Kant's view seems to be that we become\naware of two elements of the manifold A B as a succession of events in\nthe world of nature by combining them as necessarily successive in a\ncausal order, in which the state of affairs which precedes B and which\ncontains A contains something upon which B must follow (i. e. a cause\nof B), which therefore makes it necessary that B must follow A.[6] But\nif we are to do this, we must in some way succeed in selecting or\npicking out from among the elements of the manifold that element A\nwhich is to be thus combined with B. We therefore need something more\nthan the category. It is not enough that we should think that B has a\ncause; we must think of something in particular as the cause of B,\nand we must think of it either as coexistent with, or as identical\nwith, A.\n\n    [6] Cf. pp. 291-3.\n\nKant fails to notice this second difficulty,[7] and up to a certain\npoint avoids it owing to his distinction between the imagination and\nthe understanding. For he thinks of the understanding as the source of\ngeneral principles of synthesis, viz. the categories, and attributes\nindividual syntheses to the imagination. Hence the individual\nsyntheses, which involve particular principles, are already effected\nbefore the understanding comes into play. But to throw the work of\neffecting individual syntheses upon the imagination is only to evade\nthe difficulty. For in the end, as has been pointed out,[8] the\nimagination must be the understanding working unreflectively, and,\nwhether this is so or not, some account must be given of the way in\nwhich the imagination furnishes the particular principles of synthesis\nrequired.\n\n    [7] We should have expected Kant to have noticed this\n    difficulty in A. 105, Mah. 199, where he describes what is\n    involved in the relation of representations to an object, for\n    his instance of representations becoming so related is the\n    process of combining elements into a triangle, which plainly\n    requires a synthesis of a very definite kind. For the reasons\n    of his failure to notice the difficulty cf. p. 207.\n\n    [8] Pp. 168-9.\n\nThe third and last main difficulty of the first kind concerns the\nrelation of the elements of the manifold and the kinds of synthesis\nby which they are combined. This involves the distinction between\nrelating in general and terms to be related. For to perform a\nsynthesis is in general to relate, and the elements to be combined\nare the terms to be related.[9] Now it is only necessary to take\ninstances to realize that the possibility of relating terms in certain\nways involves two presuppositions, which concern respectively the\ngeneral and the special nature of the terms to be related.\n\n    [9] 'To relate' is used rather than 'to recognize as\n    related', in order to conform to Kant's view of knowledge.\n    But if it be desired to take the argument which follows in\n    connexion with knowledge proper (cf. p. 242), it is only\n    necessary to substitute throughout 'to recognize as related'\n    for 'to relate' and to make the other changes consequent\n    thereon.\n\nIn the first place, it is clear that the general nature of the terms\nmust correspond with or be adapted to the general nature of the\nrelationship to be effected. Thus if two terms are to be related as\nmore or less loud, they must be sounds, since the relation in question\nis one in respect of sound and not, e. g., of time or colour or space.\nSimilarly, terms to be related as right and left must be bodies in\nspace, right and left being a spatial relation. Again, only human\nbeings can be related as parent and child. Kant's doctrine, however,\ndoes not conform to this presupposition. For the manifold to be\nrelated consists solely of sensations, and of individual spaces, and\nperhaps individual times, as elements of pure perception; and such a\nmanifold is not of the kind required. Possibly individual spaces may\nbe regarded as adequate terms to be related or combined into\ngeometrical figures, e. g. into lines or triangles. But a house as a\nsynthesis of a manifold cannot be a synthesis of spaces, or of times,\nor of sensations. Its parts are bodies, which, whatever they may be,\nare neither sensations nor spaces nor times, nor combinations of them.\nIn reality they are substances of a special kind. Again, the relation\nof cause and effect is not a relation of sensations or spaces or\ntimes, but of successive states of physical things or substances, the\nrelation consisting in the necessity of their succession.\n\nIn the second place, it is clear that the special nature of the\nrelation to be effected presupposes a special nature on the part of\nthe terms to be related. If one sound is to be related to another by\nway of the octave, that other must be its octave. If one quantity is\nto be related to another as the double of it, that quantity must be\ntwice as large as the other. In the same way, proceeding to Kant's\ninstances, we see that if we are to combine or relate a manifold into\na triangle, and therefore into a triangle of a particular size and\nshape, the elements of the manifold must be lines, and lines of a\nparticular size. If we are to combine a manifold into a house, and\ntherefore into a house of a certain shape and size, the manifold must\nconsist of bodies of a suitable shape and size. If we are to relate a\nmanifold by way of necessary succession, the manifold must be such\nthat it can be so related; in other words, if we are to relate an\nelement X of the manifold with some other Y as the necessary\nantecedent of X, there must be some definite element Y which is\nconnected with, and always occurs along with, X. To put the matter\ngenerally, we may say that the manifold must be adapted to or 'fit'\nthe categories not only, as has been pointed out, in the sense that it\nmust be of the right kind, but also in the sense that its individual\nelements must have that orderly character which enables them to be\nrelated according to the categories.\n\nNow it is plain from Kant's vindication of what he calls the affinity\nof phenomena,[10] that he recognizes the existence of this\npresupposition. But the question arises whether this vindication can\nbe successful. For since the manifold is originated by the thing in\nitself, it seems prima facie impossible to prove that the elements of\nthe manifold must have affinity, and so be capable of being related\naccording to the categories. Before, however, we consider the chief\npassage in which Kant tries to make good his position, we may notice a\ndefence which might naturally be offered on his behalf. It might be\nsaid that he establishes the conformity of the manifold to the\ncategories at least hypothetically, i. e. upon the supposition that\nthe manifold is capable of entering into knowledge, and also upon the\nsupposition that we are capable of being conscious of our identity\nwith respect to it; for upon either supposition any element of the\nmanifold must be capable of being combined with all the rest into one\nworld of nature. Moreover, it might be added that these suppositions\nare justified, for our experience is not a mere dream, but is\nthroughout the consciousness of a world, and we are self-conscious\nthroughout our experience; and therefore it is clear that the manifold\ndoes in fact 'fit' the categories. But the retort is obvious. Any\nactual conformity of the manifold to the categories would upon this\nview be at best but an empirical fact, and, although, if the\nconformity ceased, we should cease to be aware of a world and of\nourselves, no reason has been or can be given why the conformity\nshould not cease.\n\n    [10] Cf. A. 100-2, Mah. 195-7 (quoted pp. 171-2); A. 113,\n    Mah. 205; A. 121-2, Mah. 211-2.\n\nThe passage in which Kant vindicates the affinity of phenomena in the\ngreatest detail is the following:\n\n\"We will now try to exhibit the necessary connexion of the\nunderstanding with phenomena by means of the categories, by beginning\nfrom below, i. e. from the empirical end. The first that is given us\nis a phenomenon, which if connected with consciousness is called\nperception[11].... But because every phenomenon contains a manifold,\nand consequently different perceptions are found in the mind scattered\nand single, a connexion of them is necessary, which they cannot have\nin mere sense. There is, therefore, in us an active power of synthesis\nof this manifold, which we call imagination, and the action of which,\nwhen exercised immediately upon perceptions, I call apprehension. The\nbusiness of the imagination, that is to say, is to bring the manifold\nof intuition[12] into an _image_; it must, therefore, first receive\nthe impressions into its activity, i. e. apprehend them.\"\n\n    [11] _Wahrnehmung._\n\n    [12] _Anschauung._\n\n\"But it is clear that even this apprehension of the manifold would not\nby itself produce an image and a connexion of the impressions, unless\nthere were a subjective ground in virtue of which one perception, from\nwhich the mind has passed to another, is summoned to join that which\nfollows, and thus whole series of perceptions are presented, i. e. a\nreproductive power of imagination, which power, however, is also only\nempirical.\"\n\n\"But if representations reproduced one another at haphazard just as\nthey happened to meet together, once more no determinate connexion\nwould arise, but merely chaotic heaps of them, and consequently no\nknowledge would arise; therefore the reproduction of them must have a\nrule, according to which a representation enters into connexion with\nthis rather than with another in the imagination. This subjective and\n_empirical_ ground of reproduction according to rules is called the\n_association_ of representations.\"\n\n\"But now, if this unity of association had not also an objective\nground, so that it was impossible that phenomena should be apprehended\nby the imagination otherwise than under the condition of a possible\nsynthetic unity of this apprehension, it would also be a pure accident\nthat phenomena were adapted to a connected system of human knowledge.\nFor although we should have the power of associating perceptions, it\nwould still remain wholly undetermined and accidental whether they\nwere associable; and in the event of their not being so, a multitude\nof perceptions and even perhaps a whole sensibility would be possible,\nin which much empirical consciousness would be met with in my mind,\nbut divided and without belonging to _one_ consciousness of myself,\nwhich however is impossible. For only in that I ascribe all\nperceptions to one consciousness (the original apperception) can I say\nof all of them that I am conscious of them. There must therefore be an\nobjective ground, i. e. a ground to be recognized _a priori_ before\nall empirical laws of the imagination, on which rests the possibility,\nnay even the necessity, of a law which extends throughout all\nphenomena, according to which we regard them without exception as such\ndata of the senses, as are in themselves associable and subjected to\nuniversal rules of a thorough-going connexion in reproduction. This\nobjective ground of all association of phenomena I call the _affinity_\nof phenomena. But we can meet this nowhere else than in the principle\nof the unity of apperception as regards all cognitions which are to\nbelong to me. According to it, all phenomena without exception must so\nenter into the mind or be apprehended as to agree with the unity of\napperception, which agreement would be impossible without synthetical\nunity in their connexion, which therefore is also objectively\nnecessary.\"\n\n\"The objective unity of all (empirical) consciousness in one\nconsciousness (the original apperception) is therefore the necessary\ncondition even of all possible perception, and the affinity of all\nphenomena (near or remote) is a necessary consequence of a synthesis\nin the imagination, which is _a priori_ founded upon rules.\"\n\n\"The imagination is therefore also a power of _a priori_ synthesis,\nfor which reason we give it the name of the productive imagination;\nand so far as it, in relation to all the manifold of the phenomenon,\nhas no further aim than the necessary unity in the synthesis of the\nphenomenon, it can be called the transcendental function of the\nimagination. It is therefore strange indeed, but nevertheless clear\nfrom the preceding, that only by means of this transcendental function\nof the imagination does even the affinity of phenomena, and with it\ntheir association and, through this, lastly their reproduction\naccording to laws, and consequently experience itself become possible,\nbecause without it no conceptions of objects would ever come together\ninto one experience.\"[13]\n\n    [13] A. 119-23, Mah. 210-3.\n\nIf it were not for the last two paragraphs[14], we should understand\nthis difficult passage to be substantially identical in meaning with\nthe defence of the affinity of phenomena just given.[15] We should\nunderstand Kant to be saying (1) that the synthesis which knowledge\nrequires presupposes not merely a faculty of association on our part\nby which we reproduce elements of the manifold according to rules, but\nalso an affinity on the part of the manifold to be apprehended, which\nenables our faculty of association to get to work, and (2) that this\naffinity can be vindicated as a presupposition at once of knowledge\nand of self-consciousness.\n\n    [14] And also the first and last sentence of the fourth\n    paragraph, where Kant speaks not of 'phenomena which are to\n    be apprehended', but of the 'apprehension of phenomena' as\n    necessarily agreeing with the unity of apperception.\n\n    [15] p. 220.\n\nIn view, however, of the fact that, according to the last two\nparagraphs, the affinity is due to the imagination,[16] it seems\nnecessary to interpret the passage thus:\n\n    [16] It should be noted that in the last paragraph but one\n    Kant does not say '_our knowledge_ that phenomena must have\n    affinity is a consequence of _our knowledge_ that there must\n    be a synthesis of the imagination', but 'the affinity of all\n    phenomena is a consequence of a synthesis in the\n    imagination'. And the last paragraph precludes the view that\n    in making the latter statement he meant the former. Cf. also\n    A. 101, Mah. 196.\n\n'Since the given manifold of sense consists of isolated elements, this\nmanifold, in order to enter into knowledge, must be combined into an\nimage. This combination is effected by the imagination, which however\nmust first apprehend the elements one by one.'\n\n'But this apprehension of the manifold by the imagination could\nproduce no image, unless the imagination also possessed the power of\nreproducing past elements of the manifold, and, if knowledge is to\narise, of reproducing them according to rules. This faculty of\nreproduction by which, on perceiving the element A, we are led to\nthink of or reproduce a past element B--B being reproduced according\nto some rule--rather than C or D is called the faculty of association;\nand since the rules according to which it works depend on empirical\nconditions, and therefore cannot be anticipated _a priori_, it may be\ncalled the subjective ground of reproduction.'\n\n'But if the image produced by association is to play a part in\nknowledge, the empirical faculty of reproduction is not a sufficient\ncondition or ground of it. A further condition is implied, which may\nbe called objective in the sense that it is _a priori_ and prior to\nall empirical laws of imagination. This condition is that the act by\nwhich the data of sense enter the mind or are apprehended, i. e. the\nact by which the imagination _apprehends and combines_ the data of\nsense into a sensuous image, must _make_ the elements such that they\nhave affinity, and therefore such that they can subsequently be\nrecognized as parts of a necessarily related whole.[17] Unless this\ncondition is satisfied, even if we possessed the faculty of\nassociation, our experience would be a chaos of disconnected elements,\nand we could not be self-conscious, which is impossible. Starting,\ntherefore, with the principle that we must be capable of being\nself-conscious with respect to all the elements of the manifold, we\ncan lay down _a priori_ that this condition is a fact.'\n\n    [17] On this interpretation 'entering the mind' or 'being\n    apprehended' in the fourth paragraph does not refer merely to\n    the apprehension of elements one by one, which is preliminary\n    to the act of combining them, but includes the act by which\n    they are combined. If so, Kant's argument formally involves a\n    circle. For in the second and third paragraphs he argues that\n    the synthesis of perceptions involves reproduction according\n    to rules, and then, in the fourth paragraph, he argues that\n    this reproduction presupposes a synthesis of perceptions. We\n    may, however, perhaps regard his argument as being in\n    substance that knowledge involves _re_production by the\n    imagination of elements capable of connexion, and that this\n    reproduction involves _pro_duction by the imagination of the\n    data of sense, which are to be reproduced, into an image.\n\n'It follows, then, that the affinity or connectedness of the data of\nsense presupposed by the _re_production which is presupposed in\nknowledge, is actually produced by the _pro_ductive faculty of\nimagination, which, in combining the data into a sensuous image,\ngives them the unity required.'\n\nIf, as it seems necessary to believe, this be the correct\ninterpretation of the passage,[18] Kant is here trying to carry out\nto the full his doctrine that _all_ unity or connectedness comes from\nthe mind's activity. He is maintaining that the imagination, acting\n_pro_ductively on the data of sense and thereby combining them into an\nimage, gives the data a connectedness which the understanding can\nsubsequently recognize. But to maintain this is, of course, only to\nthrow the problem one stage further back. If reproduction, in order to\nenter into knowledge, implies a manifold which has such connexion that\nit is capable of being reproduced according to rules, so the\nproduction of sense-elements into a coherent image in turn implies\nsense-elements capable of being so combined. The act of combination\ncannot confer upon them or introduce into them a unity which they do\nnot already possess.\n\n    [18] If the preceding interpretation (pp. 223-4) be thought\n    the correct one, it must be admitted that Kant's vindication\n    of the affinity breaks down for the reason given, p. 220.\n\nThe fact is that this step in Kant's argument exhibits the final\nbreakdown of his view that all unity or connectedness or relatedness\nis conferred upon the data of sense by the activity of the mind.\nConsequently, this forms a convenient point at which to consider what\nseems to be the fundamental mistake of this view. The mistake stated\nin its most general form appears to be that, misled by his theory of\nperception, he regards 'terms' as given by things in themselves\nacting on the sensibility, and 'relations' as introduced by the\nunderstanding,[19] whereas the fact is that in the sense in which terms\ncan be said to be given, relations can and must also be said to be\ngiven.\n\n    [19] The understanding being taken to include the\n    imagination, as being the faculty of _spontaneity_ in\n    distinction from the _passive_ sensibility.\n\nTo realize that this is the case, we need only consider Kant's\nfavourite instance of knowledge, the apprehension of a straight line.\nAccording to him, this presupposes that there is given to us a\nmanifold, which--whether he admits it or not--must really be parts of\nthe line, and that we combine this manifold on a principle involved in\nthe nature of straightness. Now suppose that the manifold given is the\nparts AB, BC, CD, DE of the line AE. It is clearly only possible to\nrecognize AB and BC as contiguous parts of a straight line, if we\nimmediately apprehend that AB and BC form one line of which these\nparts are identical in direction. Otherwise, we might just as well\njoin AB and BC at a right angle, and in fact at any angle; we need not\neven make AB and BC contiguous.[20] Similarly, the relation of BC to\nCD and of CD to DE must be just as immediately apprehended as the\nparts themselves. Is there, however, any relation of which it could be\nsaid that it is not given, and to which therefore Kant's doctrine\nmight seem to apply? There is. Suppose AB, BC, CD to be of such a size\nthat, though we can see AB and BC, or BC and CD, together, we cannot\nsee AB and CD together. It is clear that in this case we can only\nlearn that AB and CD are parts of the same straight line through an\ninference. We have to infer that, because each is in the same straight\nline with BC, the one is in the same straight line with the other.\nHere the fact that AB and CD are in the same straight line is not\nimmediately apprehended. This relation, therefore, may be said not to\nbe given; and, from Kant's point of view, we could say that we\nintroduce this relation into the manifold through our activity of\nthinking, which combines AB and CD together in accordance with the\nprinciple that two straight lines which are in the same line with a\nthird are in line with one another. Nevertheless, this case is no\nexception to the general principle that relations must be given\nequally with terms; for we only become aware of the relation between\nAB and CD, which is not given, because we are already aware of other\nrelations, viz. those between AB and BC, and BC and CD, which are\ngiven. Relations then, or, in Kant's language, particular syntheses\nmust be said to be given, in the sense in which the elements to be\ncombined can be said to be given.\n\n    [20] In order to meet a possible objection, it may be pointed\n    out that if AB and BC be given in isolation, the contiguity\n    implied in referring to them as A_B_ and _B_C will not be\n    known.\n\nFurther, we can better see the nature of Kant's mistake in this\nrespect, if we bear in mind that Kant originally and rightly\nintroduced the distinction between the sensibility and the\nunderstanding as that between the passive faculty by which an\nindividual is given or presented to us and the active faculty by which\nwe bring an individual under, or recognize it as an instance of a\nuniversal.[21] For we then see that Kant in the _Transcendental\nDeduction_, by treating what is given by the sensibility as terms and\nwhat is contributed by the understanding as relations, is really\nconfusing the distinction between a relation and its terms with that\nbetween universal and individual; in other words, he says of terms\nwhat ought to be said of individuals, and of relations what ought to\nbe said of universals. That the confusion is a confusion, and not a\nlegitimate identification, it is easy to see. For, on the one hand, a\nrelation between terms is as much an individual as either of the\nterms. That a body A is to the right of a body B is as much an\nindividual fact as either A or B.[22] And if terms, as being\nindividuals, belong to perception and are given, in the sense that\nthey are in an immediate relation to us, relations, as being\nindividuals, equally belong to perception and are given. On the other\nhand, individual terms just as much as individual relations imply\ncorresponding universals. An individual body implies 'bodiness', just\nas much as the fact that a body A is to the right of a body B implies\nthe relationship of 'being to the right of something'. And if, as is\nthe case, thinking or conceiving in distinction from perceiving, is\nthat activity by which we recognize an individual, given in\nperception, as one of a kind, conceiving is involved as much in the\napprehension of a term as in the apprehension of a relation. The\napprehension of 'this red body' as much involves the recognition of an\nindividual as an instance of a kind, i. e. as much involves an act of\nthe understanding, as does the apprehension of the fact that it is\nbrighter than some other body.\n\n    [21] Cf. pp. 27-9.\n\n    [22] I can attach no meaning to Mr. Bertrand Russell's\n    assertion that relations have no instances. See _The\n    Principles of Mathematics_, Sec. 55.\n\nKant has failed to notice this confusion for two reasons. In the first\nplace, beginning in the _Analytic_ with the thought that the thing in\nitself, by acting on our sensibility, produces isolated sense data, he\nis led to adopt a different view of the understanding from that which\nhe originally gave, and to conceive its business as consisting in\nrelating these data. In the second place, by distinguishing the\nimagination from the understanding, he is able to confine the\nunderstanding to being the source of universals or principles of\nrelation in distinction from individual relations.[23] Since, however,\nas has been pointed out, and as Kant himself sees at times, the\nimagination is the understanding working unreflectively, this\nlimitation cannot be successful.\n\n    [23] Cf. p. 217.\n\nThere remain for consideration the difficulties of the second kind,\ni. e. the difficulties involved in accepting its main principles at\nall. These are of course the most important. Throughout the deduction\nKant is attempting to formulate the nature of knowledge. According to\nhim, it consists in an activity of the mind by which it combines the\nmanifold of sense on certain principles and is to some extent aware\nthat it does so, and by which it thereby gives the manifold relation\nto an object. Now the fundamental and final objection to this account\nis that what it describes is not knowledge at all. The justice of this\nobjection may be seen by considering the two leading thoughts\nunderlying the view, which, though closely connected, may be treated\nseparately. These are the thought of knowledge as a process by which\nrepresentations acquire relation to an object, and the thought of\nknowledge as a process of synthesis.\n\nIt is in reality meaningless to speak of 'a process by which\nrepresentations or ideas acquire relation to an object'.[24] The\nphrase must mean a process by which a mere apprehension, which, as\nsuch, is not the apprehension of an object, becomes the apprehension\nof an object. Apprehension, however, is essentially and from the very\nbeginning the apprehension of an object, i. e. of a reality\napprehended. If there is no object which the apprehension is 'of',\nthere is no apprehension. It is therefore wholly meaningless to speak\nof a process by which an apprehension _becomes_ the apprehension of an\nobject. If when we reflected we were not aware of an object, i. e. a\nreality apprehended, we could not be aware of our apprehension; for\nour apprehension is the apprehension of it, and is itself only\napprehended in relation to, though in distinction from, it. It is\ntherefore impossible to suppose a condition of mind in which, knowing\nwhat 'apprehension' means, we proceed to ask, 'What is meant by an\nobject of it?' and 'How does an apprehension become related to an\nobject?'; for both questions involve the thought of a mere\nrepresentation, i. e. of an apprehension which as yet is not the\napprehension of anything.\n\n    [24] Cf. p. 180, and pp. 280-3.\n\nThese questions, when their real nature is exhibited, are plainly\nabsurd. Kant's special theory, however, enables him to evade the real\nabsurdity involved. For, according to his view, a representation is\nthe representation or apprehension of something only from the point of\nview of the thing in itself. As an appearance or perhaps more strictly\nspeaking as a sensation, it has also a being of its own which is not\nrelative[25]; and from this point of view it _is_ possible to speak of\n'mere' representations and to raise questions which presuppose their\nreality.[26]\n\n    [25] Cf. p. 137 init.\n\n    [26] The absurdity of the problem really propounded is also\n    concealed from Kant in the way indicated, pp. 180 fin.-181\n    init.\n\nBut this remedy, if remedy it can be called, is at least as bad as the\ndisease. For, in the first place, the change of standpoint is\nnecessarily illegitimate. An appearance or sensation is not from any\npoint of view a representation in the proper sense, i. e. a\nrepresentation or apprehension of something. It is simply a reality to\nbe apprehended, of the special kind called mental. If it be called a\nrepresentation, the word must have a new meaning; it must mean\nsomething represented, or presented,[27] i. e. object of apprehension,\nwith the implication that what is presented, or is object of\napprehension, is mental or a modification of the mind. Kant therefore\nonly avoids the original absurdity by an illegitimate change of\nstandpoint, the change being concealed by a tacit transition in the\nmeaning of representation. In the second place, the change of\nstandpoint only saves the main problem from being absurd by rendering\nit insoluble. For if a representation be taken to be an appearance or\na sensation, the main problem becomes that of explaining how it is\nthat, beginning with the apprehension of mere appearances or\nsensations, we come to apprehend an object, in the sense of an object\nin nature, which, as such, is not an appearance or sensation but a\npart of the physical world. But if the immediate object of\napprehension were in this way confined to appearances, which are, to\nuse Kant's phrase, determinations of our mind, our apprehension would\nbe limited to these appearances, and any apprehension of an object in\nnature would be impossible.[28] In fact, it is just the view that the\nimmediate object of apprehension consists in a determination of the\nmind which forms the basis of the solipsist position. Kant's own\nsolution involves an absurdity at least as great as that involved in\nthe thought of a mere representation, in the proper sense of\nrepresentation. For the solution is that appearances or sensations\nbecome related to an object, in the sense of an object in nature, by\nbeing combined on certain principles. Yet it is plainly impossible to\ncombine appearances or sensations into an object in nature. If a\ntriangle, or a house, or 'a freezing of water'[29] is the result of\nany process of combination, the elements combined must be respectively\nlines, and bricks, and physical events; these are objects in the sense\nin which the whole produced by the combination is an object, and are\ncertainly not appearances or sensations. Kant conceals the difficulty\nfrom himself by the use of language to which he is not entitled. For\nwhile his instances of objects are always of the kind indicated, he\npersists in calling the manifold combined 'representations', i. e.\npresented mental modifications. This procedure is of course\nfacilitated for him by his view that nature is a phenomenon or\nappearance, but the difficulty which it presents to the reader\nculminates when he speaks of the very same representations as having\nboth a subjective and an objective relation, i. e. as being both\nmodifications of the mind and parts of nature.[30]\n\n    [27] _Vorgestellt._\n\n    [28] Cf. p. 123.\n\n    [29] B. 162, M. 99.\n\n    [30] B. 139-42, M. 87-8. Cf. 209, note 3, and pp. 281-2.\n\nWe may now turn to Kant's thought of knowledge as a process of\nsynthesis. When Kant speaks of synthesis, the kind of synthesis of\nwhich he usually is thinking is that of spatial elements into a\nspatial whole; and although he refers to other kinds, e. g. of units\ninto numbers, and of events into a temporal series, nevertheless it is\nthe thought of spatial synthesis which guides his view. Now we must in\nthe end admit that the spatial synthesis of which he is thinking is\nreally the _construction_ or _making_ of spatial objects in the\nliteral sense. It would be rightly illustrated by making figures out\nof matches or spelicans, or by drawing a circle with compasses, or by\nbuilding a house out of bricks. Further, if we extend this view of the\nprocess of which Kant is thinking, we have to allow that the process\nof synthesis in which, according to Kant, knowledge consists is that\nof making or constructing parts of the physical world, and in fact the\nphysical world itself, out of elements given in perception.[31] The\ndeduction throughout presupposes that the synthesis is really\n_manufacture_, and Kant is at pains to emphasize the fact. \"The order\nand conformity to law in the phenomena which we call _nature_ we\nourselves introduce, and we could not find it there, if we or the\nnature of our mind had not originally placed it there.\"[32] He\nnaturally rejoices in the manufacture, because it is just this which\nmakes the categories valid. If knowing is really making, the\nprinciples of synthesis must apply to the reality known, because it is\nby these very principles that the reality is made. Moreover,\nrecognition of this fact enables us to understand certain features of\nhis view which would otherwise be inexplicable. For if the synthesis\nconsists in literal construction, we are able to understand why Kant\nshould think (1) that in the process of knowledge the mind\n_introduces_ order into the manifold, (2) that the mind is limited in\nits activity of synthesis by having to conform to certain principles\nof construction which constitute the nature of the understanding, and\n(3) that the manifold of phenomena must possess affinity. If, for\nexample, we build a house, it can be said (1) that we introduce into\nthe materials a plan or principle of arrangement which they do not\npossess in themselves, (2) that the particular plan is limited by, and\nmust conform to, the laws of spatial relation and to the general\npresuppositions of physics, such as the uniformity of nature, and (3)\nthat only such materials are capable of the particular combination as\npossess a nature suitable to it. Moreover, if, for Kant, knowing is\nreally making, we are able to understand two other prominent features\nof his view. We can understand why Kant should lay so much stress upon\nthe 'recognition' of the synthesis, and upon the self-consciousness\ninvolved in knowledge. For if the synthesis of the manifold is really\nthe making of an object, it results merely in the existence of the\nobject; knowledge of it is still to be effected. Consequently,\nknowledge of the object only finds a place in Kant's view by the\n_recognition_ (on the necessity of which he insists) of the manifold\nas combined on a principle. This recognition, which Kant considers\nonly an element in knowledge, is really the knowledge itself. Again,\nsince the reality to be known is a whole of parts which we construct\non a principle, we know that it is such a whole, and therefore that\n'the manifold is related to one object', because, and only because, we\nknow that we have combined the elements on a principle.\nSelf-consciousness therefore _must_ be inseparable from consciousness\nof an object.\n\n    [31] It is for this reason that the mathematical\n    illustrations of the synthesis are the most plausible for his\n    theory. While we can be said to construct geometrical\n    figures, and while the construction of geometrical figures\n    can easily be mistaken for the apprehension of them, we\n    cannot with any plausibility be said to construct the\n    physical world.\n\n    [32] A. 125, Mah. 214. Cf. the other passages quoted pp.\n    211-12.\n\nThe fundamental objection to this account of knowledge seems so\nobvious as to be hardly worth stating; it is of course that knowing\nand making are not the same. The very nature of knowing presupposes\nthat the thing known is already made, or, to speak more accurately,\nalready exists.[33] In other words, knowing is essentially the\ndiscovery of what already is. Even if the reality known happens to be\nsomething which we make, e. g. a house, the knowing it is distinct\nfrom the making it, and, so far from being identical with the making,\npresupposes that the reality in question is already made. Music and\npoetry are, no doubt, realities which in some sense are 'made' or\n'composed', but the apprehension of them is distinct from and\npresupposes the process by which they are composed.\n\n    [33] Cf. Ch. VI.\n\nHow difficult it is to resolve knowing into making may be seen by\nconsideration of a difficulty in the interpretation of Kant's phrase\n'relation of the manifold to an object', to which no allusion has yet\nbeen made. When it is said that a certain manifold is related to, or\nstands[34] in relation to, an object, does the relatedness referred to\nconsist in the fact that the manifold is combined into a whole, or in\nthe fact that we are conscious of the combination, or in both? If we\naccept the first alternative we must allow that, while relatedness to\nan object implies a process of synthesis, yet the relatedness, and\ntherefore the synthesis, have nothing to do with knowledge. For the\nrelatedness of the manifold to an object will be the combination of\nthe elements of the manifold as parts of an object constructed, and\nthe process of synthesis involved will be that by which the object is\nconstructed. This process of synthesis will have nothing to do with\nknowledge; for since it is merely the process by which the object is\nconstructed, knowledge so far is not effected at all, and no clue is\ngiven to the way in which it comes about. If, however, we accept the\nsecond alternative, we have to allow that while relatedness to an\nobject has to do with knowledge, yet it in no way implies a process of\nsynthesis. For since in that case it consists in the fact that we are\nconscious of the manifold as together forming an object, it in no way\nimplies that the object has been produced by a process of synthesis.\nKant, of course, would accept the third alternative. For, firstly,\nsince it is knowledge which he is describing, the phrase 'relatedness\nto an object' cannot refer simply to the _existence_ of a combination\nof the manifold, and of a process by which it has been produced; its\nmeaning must include _consciousness_ of the combination. In the second\nplace, it is definitely his view that we cannot represent anything as\ncombined in the object without having previously combined it\nourselves.[35] Moreover, it is just with respect to this connexion\nbetween the synthesis and the consciousness of the synthesis that his\nreduction of knowing to making helps him; for to make an object, e. g.\na house, is to make it consciously, i. e. to combine materials on a\nprinciple of which we are aware. Since, then, the combining of which\nhe speaks is really making, it seems to him impossible to combine a\nmanifold without being aware of the nature of the act of combination,\nand therefore of the nature of the whole thereby produced.[36] But\nthough this is clearly Kant's view, it is not justified. In the first\nplace, 'relatedness of the manifold to an object' ought not to refer\n_both_ to its combination in a whole _and_ to our consciousness of the\ncombination; and in strictness it should refer to the former only. For\nas referring to the former it indicates a relation of the manifold _to\nthe object_, as being the parts of the object, and as referring to the\nlatter it indicates a relation of the manifold _to us_, as being\napprehended by us as the parts of the object. But two relations which,\nthough they are of one and the same thing, are nevertheless relations\nof it to two different things, should not be referred to by the same\nphrase. Moreover, since the relatedness is referred to as relatedness\nto an object, the phrase properly indicates the relation of the\nmanifold to an object, and not to us as apprehending it. Again, in the\nsecond place, Kant cannot successfully maintain that the phrase is\nprimarily a loose expression for our consciousness of the manifold as\nrelated to an object, and that since this implies a process of\nsynthesis, the phrase may fairly include in its meaning the thought of\nthe combination of the manifold by us into a whole. For although Kant\nasserts--and with some plausibility--that we can only apprehend as\ncombined what we have ourselves combined, yet when we consider this\nassertion seriously we see it to be in no sense true.\n\n    [34] A. 109, Mah. 202.\n\n    [35] B. 130, M. 80.\n\n    [36] To say that 'combining', in the sense of making,\n    _really_ presupposes consciousness of the nature of the whole\n    produced, would be inconsistent with the previous assertion\n    that even where the reality known is something made, the\n    knowledge of it presupposes that the reality is already made.\n    Strictly speaking, the activity of combining presupposes\n    consciousness not of the whole which we _succeed_ in\n    producing, but of the whole which we _want_ to produce.\n\n    It may be noted that, from the point of view of the above\n    argument, the activity of combining presupposes actual\n    consciousness of the act of combination and of its principle,\n    and does not imply merely the possibility of it. Kant, of\n    course, does not hold this.\n\nThe general conclusion, therefore, to be drawn is that the process of\nsynthesis by which the manifold is said to become related to an object\nis a process not of knowledge but of construction in the literal\nsense, and that it leaves knowledge of the thing constructed still to\nbe effected. But if knowing is obviously different from making, why\nshould Kant have apparently felt no difficulty in resolving knowing\ninto making? Three reasons may be given.\n\nIn the first place, the very question, 'What does the process of\nknowing consist in?' at least suggests that knowing can be resolved\ninto and stated in terms of something else. In this respect it\nresembles the modern phrase '_theory_ of knowledge'. Moreover, since\nit is plain that in knowing we are active, the question is apt to\nassume the form, 'What do we _do_ when we know or think?' and since\none of the commonest forms of doing something is to perform a physical\noperation on physical things, whereby we effect a recombination of\nthem on some plan, it is natural to try to resolve knowing into this\nkind of doing, i. e. into making in a wide sense of the word.\n\nIn the second place, Kant never relaxed his hold upon the thing in\nitself. Consequently, there always remained for him a reality which\nexisted in itself and was not made by us. This was to him the\nfundamental reality, and the proper object of knowledge, although\nunfortunately inaccessible to _our_ faculties of knowing. Hence to\nKant it did not seriously matter that an inferior reality, viz. the\nphenomenal world, was made by us in the process of knowing.\n\nIn the third place, it is difficult, if not impossible, to read the\n_Deduction_ without realizing that Kant failed to distinguish knowing\nfrom that formation of mental imagery which accompanies knowing. The\nprocess of synthesis, if it is even to seem to constitute knowledge\nand to involve the validity of the categories, must really be a\nprocess by which we construct, and recognize our construction of,\nan individual reality in nature out of certain physical data.\nNevertheless, it is plain that what Kant normally describes as the\nprocess of synthesis is really the process by which we construct an\nimaginary picture of a reality in nature not present to perception,\ni. e. by which we imagine to ourselves what it would look like if we\nwere present to perceive it. This is implied by his continued use of\nthe terms 'reproduction' and 'imagination' in describing the\nsynthesis. To be aware of an object of past perception, it is\nnecessary, according to him, that the object should be _re_produced.\nIt is thereby implied that the object of our present awareness is not\nthe object of past perception, but a mental image which copies or\nreproduces it. The same implication is conveyed by his use of the term\n'imagination' to describe the faculty by which the synthesis is\neffected; for 'imagination' normally means the power of making a\nmental image of something not present to perception, and this\ninterpretation is confirmed by Kant's own description of the\nimagination as 'the faculty of representing an object even without\nits presence in perception'.[37] Further, that Kant really fails\nto distinguish the construction of mental imagery from literal\nconstruction is shown by the fact that, although he insists that\nthe formation of an image and reproduction are both necessary for\nknowledge, he does not consistently adhere to this. For his general\nview is that the elements combined and recognized as combined are the\noriginal data of sense, and not reproductions of them which together\nform an image, and his instances imply that the elements retained in\nthought, i. e. the elements of which we are aware subsequently to\nperception, are the elements originally perceived, e. g. the parts of\na line or the units counted.[38] Moreover, in one passage Kant\ndefinitely describes certain _objects_ of _perception_ taken together\nas an _image_ of that 'kind' of which, when taken together, they are\nan instance. \"If I place five points one after another, . . . . . this\nis an image of the number five.\"[39] Now, if it be granted that Kant\nhas in mind normally the process of imagining, we can see why he found\nno difficulty in the thought of knowledge as construction. For while\nwe cannot reasonably speak of making _an object of knowledge_, we can\nreasonably speak of making _a mental image_ through our own activity,\nand also of making it in accordance with the categories and the\nempirical laws which presuppose them. Moreover, the ease with which it\nis possible to take the imagining which accompanies knowing for\nknowing[40]--the image formed being taken to be the object known and\nthe forming it being taken to be the knowing it--renders it easy to\ntransfer the thought of construction to the knowledge itself. The only\ndefect, however, under which the view labours is the important one\nthat, whatever be the extent to which imagination must accompany\nknowledge, it is distinct from knowledge. To realize the difference we\nhave only to notice that the process by which we present to ourselves\nin imagination realities not present to perception presupposes, and is\nthroughout guided by, the knowledge of them. It should be noted,\nhowever, that, although the process of which Kant is normally thinking\nis doubtless that of constructing mental imagery, his real view must\nbe that knowledge consists in constructing a world out of the data of\nsense, or, more accurately, as his instances show, out of the objects\nof isolated perceptions, e. g. parts of a line or units to be counted.\nOtherwise the final act of recognition would be an apprehension not of\nthe world of nature, but of an image of it.\n\n    [37] B. 152, M. 93; cf. also Mah. 211, A. 120.\n\n    [38] Cf. A. 102-3, Mah. 197-8. The fact is that the appeal to\n    reproduction is a useless device intended by Kant--and by\n    'empirical psychologists'--to get round the difficulty of\n    allowing that in the apprehension (in memory or otherwise) of\n    a reality not present to perception, we are really aware of\n    the reality. The difficulty is in reality due to a\n    sensationalistic standpoint, avowed or unavowed, and the\n    device is useless, because the assumption has in the end to\n    be made, covertly or otherwise, that we are really aware of\n    the reality in question.\n\n    [39] B. 179, M. 109. Cf. the whole passage B. 176-81, M.\n    107-10 (part quoted pp. 249-51), and p. 251.\n\n    [40] Cf. Locke and Hume.\n\n'This criticism,' it may be said, 'is too sweeping. It may be true\nthat the process which Kant describes is really making in the literal\nsense and not knowing, but Kant's mistake may have been merely that of\nthinking of the wrong kind of synthesis. For both ordinary language\nand that of philosophical discussion imply that synthesis plays some\npart in knowledge. Thus we find in ordinary language the phrases\n'_putting_ 2 and 2 _together_' and '2 and 2 _make_ 4'. Even in\nphilosophical discussions we find it said that a complex conception,\ne. g. gold, is a _synthesis_ of simple conceptions, e. g. yellowness,\nweight, &c.; that in judgement we _relate_ or _refer_ the predicate to\nthe subject; and that in inference we _construct_ reality, though only\nmentally or ideally. Further, in any case it is by thinking or knowing\nthat the world comes to be _for us_; the more we think, the more of\nreality there is for us. Hence at least the world _for us_ or _our_\nworld is due to our activity of knowing, and so is in some sense made\nby us, i. e. by our relating activity.'\n\nThis position, however, seems in reality to be based on a simple but\nillegitimate transition, viz. the transition to the assertion that in\nknowing we relate, or combine, or construct from the assertion that in\nknowing we recognize as related, or combined, or constructed--the last\ntwo terms being retained to preserve the parallelism.[41] While the\nlatter assertion may be said to be true, although the terms 'combined'\nand 'constructed' should be rejected as misleading, the former\nassertion must be admitted to be wholly false, i. e. true in no sense\nwhatever. Moreover, the considerations adduced in favour of the\nposition should, it seems, be met by a flat denial of their truth or,\nif not, of their relevance. For when it is said that _our_ world, or\nthe world _for us_, is due to our activity of thinking, and so is in\nsome sense _made_ by us, all that should be meant is that our\n_apprehending_ the world as whatever we apprehend it to be\n_presupposes_ activity on our part. But since the activity is after\nall only the activity itself of apprehending or knowing, this\nassertion is only a way of saying that apprehending or knowing is not\na condition of mind which can be produced in us _ab extra_, but is\nsomething which we have to do for ourselves. Nothing is implied to be\nmade. If anything is to be said to be made, it must be not our world\nbut our activity of apprehending the world; but even we and our\nactivity of apprehending the world are not related as maker and thing\nmade. Again, to speak of a complex conception, e. g. gold, and to say\nthat it involves a synthesis of simple conceptions by the mind is mere\n'conceptualism'. If, as we ought to do, we replace the term\n'conception' by 'universal', and speak of gold as a synthesis of\nuniversals, any suggestion that the mind performs the synthesis will\nvanish, for a 'synthesis of universals' will mean simply a connexion\nof universals. All that is mental is our apprehension of their\nconnexion. Again, in judgement we cannot be said to _relate_ predicate\nto subject. Such an assertion would mean either that we relate a\nconception to a conception, or a conception to a reality[42], or a\nreality to a reality; and, on any of these interpretations, it is\nplainly false. To retain the language of 'relation' or of\n'combination' at all, we must say that in judgement we recognize real\nelements as related or combined. Again, when we infer, we do not\nconstruct, ideally or otherwise. 'Ideal construction'[43] is a\ncontradiction in terms, unless it refers solely to mental imagining,\nin which case it is not inference. Construction which is not 'ideal',\ni. e. literal construction, plainly cannot constitute the nature of\ninference; for inference would cease to be inference, if by it we\nmade, and did not apprehend, a necessity of connexion. Again, the\nphrase '2 and 2 _make_ 4' does not justify the view that in some sense\nwe 'make' reality. It of course suggests that 2 and 2 are not 4 until\nthey are added, i. e. that the addition makes them 4.[44] But the\nlanguage is only appropriate when we are literally making a group of 4\nby physically placing 2 pairs of bodies in one group. Where we are\ncounting, we should say merely that 2 and 2 _are_ 4. Lastly, it must\nbe allowed that the use of the phrase 'putting two and two together',\nto describe an inference from facts not quite obviously connected, is\nloose and inexact. If we meet a dog with a blood-stained mouth and\nshortly afterwards see a dead fowl, we may be said to put two and two\ntogether and to conclude thereby that the dog killed the fowl. But,\nstrictly speaking, in drawing the inference we do not put anything\ntogether. We certainly do not put together the facts that the mouth of\nthe dog is blood-stained and that the fowl has just been killed. We do\nnot even put the premises together, i. e. our apprehensions of these\nfacts. What takes place should be described by saying simply that\nseeing that the fowl is killed, we also remember that the dog's mouth\nwas stained, and then apprehend a connexion between these facts.\n\n    [41] Cf. Caird, i. 394, where Dr. Caird speaks of 'the\n    distinction of the activity of thought from the matter which\n    it _combines or recognizes as combined_ in the idea of an\n    object'. (The italics are mine.) The context seems to\n    indicate that the phrase is meant to express the truth, and\n    not merely Kant's view.\n\n    [42] Cf. the account of judgement in Mr. Bradley's _Logic_.\n\n\n    [43] Cf. the account of inference in Mr. Bradley's _Logic_.\n\n    [44] Cf. Bradley, _Logic_, pp. 370 and 506.\n\nThe fact seems to be that the thought of synthesis in no way helps to\nelucidate the nature of knowing, and that the mistake in principle\nwhich underlies Kant's view lies in the implicit supposition that it\nis possible to elucidate the nature of knowledge by means of something\nother than itself. Knowledge is _sui generis_ and therefore a 'theory'\nof it is impossible. Knowledge is simply knowledge, and any attempt to\nstate it in terms of something else must end in describing something\nwhich is not knowledge.[45]\n\n    [45] Cf. p. 124.\n\n\n\n\nCHAPTER X\n\nTHE SCHEMATISM OF THE CATEGORIES\n\n\nAs has already been pointed out,[1] the _Analytic_ is divided into\ntwo parts, the _Analytic of Conceptions_, of which the aim is to\ndiscover and vindicate the validity of the categories, and the\n_Analytic of Principles_, of which the aim is to determine the use\nof the categories in judgement. The latter part, which has now to be\nconsidered, is subdivided into two. It has, according to Kant, firstly\nto determine the sensuous conditions under which the categories are\nused, and secondly to discover the _a priori_ principles involved in\nthe categories, as exercised under these sensuous conditions, such,\nfor instance, as the law that all changes take place according to\nthe law of cause and effect. The first problem is dealt with in\nthe chapter on the 'schematism of the pure conceptions of the\nunderstanding', the second in the chapter on the 'system of all\nprinciples of the pure understanding'.\n\n    [1] p. 141.\n\nWe naturally feel a preliminary difficulty with respect to the\nexistence of this second part of the _Analytic_ at all. It seems clear\nthat if the first part is successful, the second must be unnecessary.\nFor if Kant is in a position to lay down that the categories must\napply to objects, no special conditions of their application need be\nsubsequently determined. If, for instance, it can be laid down that\nthe category of quantity must apply to objects, it is implied either\nthat there are no special conditions of its application, or that they\nhave already been discovered and shown to exist. Again, to assert the\napplicability of the categories is really to assert the existence of\nprinciples, and in fact of just those principles which it is the aim\nof the _System of Principles_ to prove. Thus to assert the\napplicability of the categories of quantity and of cause and effect is\nto assert respectively the principles that all objects of perception\nare extensive quantities, and that all changes take place according\nto the law of cause and effect. The _Deduction of the Categories_\ntherefore, if successful, must have already proved the principles\nnow to be vindicated; and it is a matter for legitimate surprise\nthat we find Kant in the _System of Principles_ giving proofs of\nthese principles which make no appeal to the _Deduction of the\nCategories_.[2] On the other hand, for the existence of the account of\nthe schematism of the categories Kant has a better show of reason. For\nthe conceptions derived in the _Metaphysical Deduction_ from the\nnature of formal judgement are in themselves too abstract to be the\nconceptions which are to be shown applicable to the sensible world,\nsince all the latter involve the thought of time. Thus, the conception\nof cause and effect derived from the nature of the hypothetical\njudgement includes no thought of time, while the conception of which\nhe wishes to show the validity is that of necessary succession in\ntime. Hence the conceptions discovered by analysis of formal judgement\nhave in some way to be rendered more concrete in respect of time. The\naccount of the schematism, therefore, is an attempt to get out of the\nfalse position reached by appealing to Formal Logic for the list of\ncategories. Nevertheless, the mention of a sensuous condition under\nwhich alone the categories can be employed[3] should have suggested to\nKant that the transcendental deduction was defective, and, in fact, in\nthe second version of the transcendental deduction two paragraphs[4]\nare inserted which take account of this sensuous condition.\n\n    [2] The cause of Kant's procedure is, of course, to be found\n    in the unreal way in which he isolates conception from\n    judgement.\n\n    [3] B. 175, M. 106.\n\n    [4] B. Secs. 24 and 26, M. Secs. 20 and 22.\n\nThe beginning of Kant's account of schematism may be summarized thus:\n'Whenever we subsume an individual object of a certain kind, e. g.\na plate, under a conception, e. g. a circle, the object and the\nconception must be homogeneous, that is to say, the individual must\npossess the characteristic which constitutes the conception, or, in\nother words, must be an instance of it. Pure conceptions, however, and\nempirical perceptions, i. e. objects of empirical perception, are\nquite heterogeneous. We do not, for instance, perceive cases of cause\nand effect. Hence the problem arises, 'How is it possible to subsume\nobjects of empirical perception under pure conceptions?' The\npossibility of this subsumption presupposes a _tertium quid_, which is\nhomogeneous both with the object of empirical perception and with the\nconception, and so makes the subsumption mediately possible. This\n_tertium quid_ must be, on the one side, intellectual and, on the\nother side, sensuous. It is to be found in a 'transcendental\ndetermination of time', i. e. a conception involving time and involved\nin experience. For in the first place this is on the one side\nintellectual and on the other sensuous, and in the second place it\nis so far homogeneous with the category which constitutes its unity\nthat it is universal and rests on an _a priori_ rule, and so far\nhomogeneous with the phenomenon that all phenomena are in time[5].\nSuch transcendental determinations of time are the schemata of the\npure conceptions of the understanding.' Kant continues as follows:\n\n    [5] It may be noted that the argument here really fails. For\n    though phenomena as involving temporal relations, might\n    possibly be said to be instances of a transcendental\n    determination of time, the fact that the latter agrees with\n    the corresponding category by being universal and _a priori_\n    does not constitute it homogeneous with the category, in the\n    sense required for subsumption, viz. that it is an instance\n    of or a species of the category.\n\n\"The schema is in itself always a mere product of the imagination.\nBut since the synthesis of the imagination has for its aim no single\nperception, but merely unity in the determination of the sensibility,\nthe schema should be distinguished from the image. Thus, if I place\nfive points one after another, . . . . . this is an image of the\nnumber five. On the other hand, if I only just think a number in\ngeneral--no matter what it may be, five or a hundred--this thinking\nis rather the representation of a method of representing in an image\na group (e. g. a thousand), in conformity with a certain conception,\nthan the image itself, an image which, in the instance given, I should\nfind difficulty in surveying and comparing with the conception. Now\nthis representation of a general procedure of the imagination to\nsupply its image to a conception, I call the schema of this\nconception.\"\n\n\"The fact is that it is not images of objects, but schemata, which lie\nat the foundation of our pure sensuous conceptions. No image could\never be adequate to our conception of a triangle in general. For it\nwould not attain the generality of the conception which makes it valid\nfor all triangles, whether right-angled, acute-angled, &c., but would\nalways be limited to one part only of this sphere. The schema of the\ntriangle can exist nowhere else than in thought, and signifies a rule\nof the synthesis of the imagination in regard to pure figures in\nspace. An object of experience or an image of it always falls short of\nthe empirical conception to a far greater degree than does the schema;\nthe empirical conception always relates immediately to the schema of\nthe imagination as a rule for the determination of our perception in\nconformity with a certain general conception. The conception of 'dog'\nsignifies a rule according to which my imagination can draw the\ngeneral outline of the figure of a four-footed animal, without being\nlimited to any particular single form which experience presents to me,\nor indeed to any possible image that I can represent to myself _in\nconcreto_. This schematism of our understanding in regard to phenomena\nand their mere form is an art hidden in the depths of the human soul,\nwhose true modes of action we are not likely ever to discover from\nNature and unveil. Thus much only can we say: the _image_ is a product\nof the empirical faculty of the productive imagination, while the\n_schema_ of sensuous conceptions (such as of figures in space) is a\nproduct and, as it were, a monogram of the pure _a priori_\nimagination, through which, and according to which, images first\nbecome possible, though the images must be connected with the\nconception only by means of the schema which they express, and are in\nthemselves not fully adequate to it. On the other hand, the schema of\na pure conception of the understanding is something which cannot be\nbrought to an image; on the contrary, it is only the pure synthesis in\naccordance with a rule of unity according to conceptions in general, a\nrule of unity which the category expresses, and it is a transcendental\nproduct of the imagination which concerns the determination of the\ninner sense in general according to conditions of its form (time)\nwith reference to all representations, so far as these are to be\nconnected _a priori_ in one conception according to the unity of\napperception.\"[6]\n\n    [6] B. 179-81, M. 109-10.\n\nNow, in order to determine whether schemata can constitute the desired\nlink between the pure conceptions or categories and the manifold of\nsense, it is necessary to follow closely this account of a schema.\nKant unquestionably in this passage treats as a mental image related\nto a conception what really is, and what on his own theory ought to\nhave been, an individual object related to a conception, i. e. an\ninstance of it. In other words, he takes a mental image of an\nindividual for the individual itself.[7] On the one hand, he treats a\nschema of a conception throughout as the thought of a procedure of the\nimagination to present to the conception its _image_, and he opposes\nschemata not to objects but to _images_; on the other hand, his\nproblem concerns subsumption under a conception, and what is subsumed\nmust be an instance of the conception, i. e. an individual object of\nthe kind in question.[8] Again, in asserting that if I place five\npoints one after another, . . . . . this is an image of the number\nfive, he is actually saying that an individual group of five points is\nan image of a group of five in general.[9] Further, if the process of\nschematizing is to enter--as it must--into knowledge of the phenomenal\nworld, what Kant here speaks of as the images related to a conception\nmust be taken to be individual instances of the conception, whatever\nhis language may be. For, in order to enter into knowledge, the\nprocess referred to must be that by which _objects of experience_ are\nconstructed. Hence the passage should be interpreted as if throughout\nthere had been written for 'image' 'individual instance' or more\nsimply 'instance'. Again, the process of schematizing, although\n_introduced_ simply as a process by which an individual is to be\nsubsumed indirectly under a conception, is assumed in the passage\nquoted to be a process of _synthesis_. Hence we may say that the\nprocess of schematizing is a process by which we combine the manifold\nof perception into an individual whole in accordance with a\nconception, and that the schema of a conception is the thought of the\nrule of procedure on our part by which we combine the manifold in\naccordance with the conception, and so bring the manifold under the\nconception. Thus the schema of the conception of 100 is the thought of\na process of synthesis by which we combine say 10 groups of 10 units\ninto 100, and the schematizing of the conception of 100 is the process\nby which we do so. Here it is essential to notice three points. In the\nfirst place, the schema is a conception which relates not to the\nreality apprehended but to us. It is the thought of a rule of\nprocedure on our part by which an instance of a conception is\nconstructed, and not the thought of a characteristic of the reality\nconstructed. For instance, the thought of a rule by which we can\ncombine points to make 100 is a thought which concerns us and not the\npoints; it is only the conception corresponding to this schema, viz.\nthe thought of 100, which concerns the points. In the second place,\nalthough the thought of time is involved in the schema, the succession\nin question lies not in the object, but in our act of construction\nor apprehension. In the third place, the schema presupposes the\ncorresponding conception and the process of schematizing directly\nbrings the manifold of perception under the conception. Thus the\nthought of combining 10 groups of 10 units to make 100 presupposes the\nthought of 100, and the process of combination brings the units under\nthe conception of 100.\n\n    [7] Cf. pp. 240-1. The mistake is, of course, facilitated by\n    the fact that 'objects in nature', being for Kant only\n    'appearances', resemble mental images more closely than they\n    do as usually conceived.\n\n    [8] Cf. B. 176, M. 107. That individuals are really referred\n    to is also implied in the assertion that 'the synthesis of\n    imagination has for its aim no single _perception_, but\n    merely unity in the determination of sensibility'. (The\n    italics are mine.)\n\n    [9] Two sentences treat individual objects and images as\n    if they might be mentioned indifferently. \"An object of\n    experience or an image of it always falls short of the\n    empirical conception to a far greater degree than does\n    the schema.\" \"The conception of a 'dog' signifies a rule\n    according to which my imagination can draw the general\n    outline of the figure of a four-footed animal without being\n    limited to any single particular form which experience\n    presents to me, or indeed to any possible image that I can\n    represent to myself _in concreto_.\"\n\nIf, however, we go on to ask what is required of schemata and of the\nprocess of schematizing, if they are to enable the manifold to be\nsubsumed under the categories, we see that each of these three\ncharacteristics makes it impossible for them to fulfil this purpose.\nFor firstly, an individual manifold A has to be brought under a\ncategory B. Since _ex hypothesi_ this cannot be effected directly,\nthere is needed a mediating conception C. C, therefore, it would seem,\nmust be at once a species of B and a conception of which A is an\ninstance. In any case C must be a conception relating to the reality\nto be known, and not to any process of knowing on our part, and,\nagain, it must be more concrete than B. This is borne out by the list\nof the schemata of the categories. But, although a schema may be said\nto be more concrete than the corresponding conception, in that it\npresupposes the conception, it neither is nor involves a more\nconcrete conception of an _object_ and in fact, as has been pointed\nout, relates not to the reality to be known but to the process on our\npart by which we construct or apprehend it.[10] In the second place,\nthe time in respect of which the category B has to be made more\nconcrete must relate to the object, and not to the successive process\nby which we apprehend it, whereas the time involved in a schema\nconcerns the latter and not the former. In the third place, from the\npoint of view of the categories, the process of schematizing should be\na process whereby we combine the manifold into a whole A in accordance\nwith the conception C, and thereby render _possible_ the subsumption\nof A under the category B. If it be a process which actually subsumes\nthe manifold under B, it will _actually_ perform that, the very\nimpossibility of which has made it necessary to postulate such a\nprocess at all. For, according to Kant, it is just the fact that the\nmanifold cannot be subsumed directly under the categories that renders\nschematism necessary. Yet, on Kant's general account of a schema, the\nschematizing must actually bring a manifold under the corresponding\nconception. If we present to ourselves an individual triangle by\nsuccessively joining three lines according to the conception of a\ntriangle, i. e. so that they enclose a space, we are directly bringing\nthe manifold, i. e. the lines, under the conception of a triangle.\nAgain, if we present to ourselves an instance of a group of 100 by\ncombining 10 groups of 10 units of any kind, we are directly bringing\nthe units under the conception of 100. If this consideration be\napplied to the schematism of a category, we see that the process said\nto be necessary because a certain other process is impossible is the\nvery process said to be impossible.\n\n    [10] It may be objected that, from Kant's point of view, the\n    thought of a rule of construction, and the thought of the\n    principle of the whole to be constructed, are the same thing\n    from different points of view. But if this be insisted on,\n    the schema and its corresponding conception become the same\n    thing regarded from different points of view; consequently\n    the schema will not be a more concrete conception of an\n    object than the corresponding conception, but it will be the\n    conception itself.\n\nIf, therefore, Kant succeeds in finding schemata of the categories in\ndetail in the sense in which they are required for the solution of his\nproblem, i. e. in the sense of more concrete conceptions involving the\nthought of time and relating to objects, we should expect either that\nhe ignores his general account of a schema, or that if he appeals to\nit, the appeal is irrelevant. This we find to be the case. His account\nof the first two transcendental schemata makes a wholly irrelevant\nappeal to the temporal process of synthesis on our part, while his\naccount of the remaining schemata makes no attempt to appeal to it at\nall.\n\n\"The pure _schema_ of _quantity_, as a conception of the\nunderstanding, is _number_, a representation which comprises the\nsuccessive addition of one to one (homogeneous elements). Accordingly,\nnumber is nothing else than the unity of the synthesis of the manifold\nof a homogeneous perception in general, in that I generate time itself\nin the apprehension of the perception.\"[11]\n\n    [11] B. 182, M. 110.\n\nIt is clear that this passage, whatever its precise interpretation may\nbe,[12] involves a confusion between the thought of counting and that\nof number. The thought of number relates to objects of apprehension\nand does not involve the thought of time. The thought of counting,\nwhich presupposes the thought of number, relates to our apprehension\nof objects and involves the thought of time; it is the thought of a\nsuccessive process on our part by which we count the number of units\ncontained in what we already know to consist of units.[13] Now we must\nassume that the schema of quantity is really what Kant says it is,\nviz. number, or to express it more accurately, the thought of number,\nand not the thought of counting, with which he wrongly identifies it.\nFor his main problem is to find conceptions which at once are more\nconcrete than the categories and, at the same time, like the\ncategories, relate to objects, and the thought of counting, though\nmore concrete than that of number, does not relate to objects. Three\nconsequences follow. In the first place, although the schema of\nquantity, i. e. the thought of number, is more concrete than the\nthought of quantity,[14] it is not, as it should be, more concrete in\nrespect of time; for the thought of number does not include the\nthought of time. Secondly, the thought of time is only introduced into\nthe schema of quantity irrelevantly by reference to the temporal\nprocess of _counting_, by which we come to apprehend the number of a\ngiven group of units. Thirdly, the schema of quantity is only in\nappearance connected with the nature of a schema in general, as Kant\ndescribes it, by a false identification of the thought of number with\nthe thought of the process on our part by which we count groups of\nunits, i. e. numbers.\n\n    [12] The drift of the passage would seem to be this: 'If we\n    are to present to ourselves an instance of a quantity, we\n    must successively combine similar units until they form a\n    quantity. This process involves the thought of a successive\n    process by which we add units according to the conception of\n    a quantity. This thought is the thought of number, and since\n    by it we present to ourselves an instance of a quantity, it\n    is the schema of quantity.' But if this be its drift,\n    considerations of sense demand that it should be rewritten,\n    at least to the following extent: 'If we are to present to\n    ourselves an instance of a _particular_ quantity [which will\n    really be a particular number, for it must be regarded as\n    discrete, (cf. B. 212, M. 128 fin., 129 init.)] e. g. three,\n    we must successively combine units until they form _that_\n    quantity. This process involves the thought of a successive\n    process, by which we add units according to the conception of\n    _that_ quantity. This thought is the thought of a particular\n    number, and since by it we present to ourselves an instance\n    of _that_ quantity, this thought is the schema of _that_\n    quantity.' If this rewriting be admitted to be necessary, it\n    must be allowed that Kant has confused (_a_) the thoughts of\n    particular quantities and of particular numbers with those of\n    quantity and of number in general respectively, (_b_) the\n    thought of a particular quantity with that of a particular\n    number (for the process referred to presupposes that the\n    particular quantity taken is known to consist of a number of\n    equal units) and (_c_) the thought of counting with that of\n    number.\n\n    [13] This statement is, of course, not meant as a definition\n    of counting, but as a means of bringing out the distinction\n    between a process of counting and a number.\n\n    [14] For the thought of a number is the thought of a quantity\n    of a special kind, viz. of a quantity made up of a number of\n    similar units without remainder.\n\nThe account of the schema of reality, the second category, runs as\nfollows: \"Reality is in the pure conception of the understanding that\nwhich corresponds to a sensation in general, that therefore of which\nthe conception in itself indicates a being (in time), while negation\nis that of which the conception indicates a not being (in time). Their\nopposition, therefore, arises in the distinction between one and the\nsame time as filled or empty. Since time is only the form of\nperception, consequently of objects as phenomena, that which in\nobjects corresponds to sensation is the transcendental matter of all\nobjects as things in themselves (thinghood, reality).[15] Now every\nsensation has a degree or magnitude by which it can fill the same\ntime, i. e. the internal sense, in respect of the same representation\nof an object, more or less, until it vanishes into nothing ( = 0 =\n_negatio_). There is, therefore, a relation and connexion between\nreality and negation, or rather a transition from the former to the\nlatter, which makes every reality representable as a _quantum_; and\nthe schema of a reality, as the quantity of something so far as it\nfills time, is just this continuous and uniform generation of the\nreality in time, as we descend in time from the sensation which has a\ncertain degree, down to the vanishing thereof, or gradually ascend\nfrom negation to the magnitude thereof.\"[16]\n\n    [15] It is difficult to see how Kant could meet the criticism\n    that here, contrary to his intention, he is treating physical\n    objects as things in themselves. Cf. p. 265.\n\n    [16] B. 182-3, M. 110-11.\n\nThis passage, if it be taken in connexion with the account of the\nanticipations of perception,[17] seems to have the following meaning:\n'In thinking of something as a reality, we think of it as that which\ncorresponds to, i. e. produces, a sensation, and therefore as\nsomething which, like the sensation, is in time; and just as every\nsensation, which, as such, occupies time, has a certain degree of\nintensity, so has the reality which produces it. Now to produce for\nourselves an instance of a reality in this sense, we must add units of\nreality till a reality of the required degree is produced, and the\nthought of this method on our part of constructing an individual\nreality is the schema of reality.' But if this represents Kant's\nmeaning, the schema of reality relates only to our process of\napprehension, and therefore is not a conception which relates to\nobjects and is more concrete than the corresponding category in\nrespect of time. Moreover, it is matter for surprise that in the case\nof this category Kant should have thought schematism necessary, for\ntime is actually included in his own statement of the category.\n\n    [17] B. 207-18, M. 125-32.\n\nThe account of the schemata of the remaining categories need not be\nconsidered. It merely _asserts_ that certain conceptions relating\nto objects and involving the thought of time are the schemata\ncorresponding to the remaining categories, without any attempt to\nconnect them with the nature of a schema. Thus, the schema of\nsubstance is asserted to be the _permanence_ of the real _in time_,\nthat of cause the _succession_ of the manifold, in so far as that\nsuccession is subjected to a rule, that of interaction the\n_coexistence_ of the determinations or accidents of one substance\nwith those of another according to a universal rule.[18] Again, the\nschemata of possibility, of actuality and of necessity are said to be\nrespectively the accordance of the synthesis of representations with\nthe conditions of time in general, existence in a determined time, and\nexistence of an object in all time.\n\n    [18] The italics are mine.\n\nThe main confusion pervading the chapter is of course that between\ntemporal relations which concern the process of apprehension and\ntemporal relations which concern the realities apprehended. Kant is\ncontinually referring to the former as if they were the latter. The\ncause of this confusion lies in Kant's reduction of physical realities\nto representations. Since, according to him, these realities are only\nour representations, all temporal relations are really relations of\nour representations, and these relations have to be treated at one\ntime as relations of our apprehensions, and at another as relations of\nthe realities apprehended, as the context requires.\n\n\n\n\nCHAPTER XI\n\nTHE MATHEMATICAL PRINCIPLES\n\n\nAs has been pointed out,[1] the aim of the second part of the\n_Analytic of Principles_ is to determine the _a priori_ principles\ninvolved in the use of the categories under the necessary sensuous\nconditions. These principles Kant divides into four classes,\ncorresponding to the four groups of categories, and he calls\nthem respectively 'axioms of perception', 'anticipations of\nsense-perception', 'analogies of experience', and 'postulates of\nempirical thought'. The first two and the last two classes are grouped\ntogether as 'mathematical' and 'dynamical' respectively, on the ground\nthat the former group concerns the perception of objects, i. e. their\nnature apprehended in perception, while the latter group concerns\ntheir existence, and that consequently, since assertions concerning\nthe existence of objects presuppose the realization of empirical\nconditions which assertions concerning their nature do not, only the\nformer possesses an absolute necessity and an immediate evidence such\nas is found in mathematics.[2]\n\n    [1] p. 246.\n\n    [2] The assertion that all perceptions (i. e. all objects of\n    perception) are extensive quantities relates, according to\n    Kant, to the nature of objects, while the assertion that an\n    event must have a necessary antecedent affirms that such an\n    antecedent must exist, but gives no clue to its specific\n    nature. Compare \"But the existence of phenomena cannot be\n    known _a priori_, and although we could be led in this way to\n    infer the fact of some existence, we should not know this\n    existence determinately, i. e. we could not anticipate the\n    respect in which the empirical perception of it differed from\n    that of other existences\". (B. 221, M. 134). Kant seems to\n    think that the fact that the dynamical principles relate to\n    the existence of objects is a sufficient justification of\n    their name.\n\n    It needs but little reflection to see that the distinctions\n    which Kant draws between the mathematical and the dynamical\n    principles must break down.\n\nThese two groups of principles are not, as their names might suggest,\nprinciples within mathematics and physics, but presuppositions of\nmathematics and physics respectively. Kant also claims appropriateness\nfor the special terms used of each minor group to indicate the kind of\nprinciples in question, viz. 'axioms', 'anticipations', 'analogies',\n'postulates'. But it may be noted as an indication of the\nartificiality of the scheme that each of the first two groups contains\nonly one principle, although Kant refers to them in the plural as\naxioms and anticipations respectively, and although the existence of\nthree categories corresponding to each group would suggest the\nexistence of three principles.\n\nThe axiom of perception is that 'All perceptions are extensive\nquantities'. The proof of it runs thus:\n\n\"An extensive quantity I call that in which the representation of the\nparts renders possible the representation of the whole (and therefore\nnecessarily precedes it). I cannot represent to myself any line,\nhowever small it may be, without drawing it in thought, that is,\nwithout generating from a point all its parts one after another, and\nthereby first drawing this perception. Precisely the same is the case\nwith every, even the smallest, time.... Since the pure perception in\nall phenomena is either time or space, every phenomenon as a\nperception is an extensive quantity, because it can be known in\napprehension only by a successive synthesis (of part with part). All\nphenomena, therefore, are already perceived as aggregates (groups of\npreviously given parts), which is not the case with quantities of\nevery kind, but only with those which are represented and apprehended\nby us as _extensive_.\"[3]\n\n    [3] B. 203-4, M. 123.\n\nKant opposes an extensive quantity to an intensive quantity or a\nquantity which has a degree. \"That quantity which is apprehended\nonly as unity and in which plurality can be represented only by\napproximation to negation = 0, I call _intensive quantity_.\"[4] The\naspect of this ultimate distinction which underlies Kant's mode of\nstating it is that only an extensive quantity is a whole, i. e.\nsomething made up of parts. Thus a mile can be said to be made up\nof two half-miles, but a velocity of one foot per second, though\ncomparable with a velocity of half a foot per second, cannot be said\nto be made up of two such velocities; it is essentially one and\nindivisible. Hence, from Kant's point of view, it follows that it is\nonly an extensive magnitude which can, and indeed must, be apprehended\nthrough a successive synthesis of the parts. The proof of the axiom\nseems to be simply this: 'All phenomena as objects of perception are\nsubject to the forms of perception, space and time. Space and time are\n[homogeneous manifolds, and therefore] extensive quantities, only\nto be apprehended by a successive synthesis of the parts. Hence\nphenomena, or objects of experience, must also be extensive\nquantities, to be similarly apprehended.' And Kant goes on to add that\nit is for this reason that geometry and pure mathematics generally\napply to objects of experience.\n\n    [4] B. 210, M. 127.\n\nWe need only draw attention to three points. Firstly, no justification\nis given of the term 'axiom'. Secondly, the argument does not really\nappeal to the doctrine of the categories, but only to the character\nof space and time as forms of perception. Thirdly, it need not appeal\nto space and time as forms of perception in the proper sense of ways\nin which we apprehend objects, but only in the sense of ways in which\nobjects are related[5]; in other words, it need not appeal to Kant's\ntheory of knowledge. The conclusion follows simply from the nature of\nobjects as spatially and temporally related, whether they are\nphenomena or not. It may be objected that Kant's thesis is that _all_\nobjects of perception are extensive quantities, and that unless space\nand time are allowed to be ways in which _we must perceive_ objects,\nwe cannot say that all objects will be spatially and temporally\nrelated, and so extensive quantities. But to this it may be replied\nthat it is only true that all objects of perception are extensive\nquantities if the term 'object of perception' be restricted to parts\nof the physical world, i. e. to just those realities which Kant is\nthinking of as spatially and temporally related,[6] and that this\nrestriction is not justified, since a sensation or a pain which has\nonly intensive quantity is just as much entitled to be called an\nobject of perception.\n\n    [5] Cf. pp. 37-9.\n\n    [6] The context shows that Kant is thinking only of such\n    temporal relations as belong to the physical world, and not\n    of those which belong to us as apprehending it. Cf. p. 139.\n\nThe anticipation of sense-perception consists in the principle that\n'In all phenomena, the real, which is an object of sensation, has\nintensive magnitude, i. e. a degree'. The proof is stated thus:\n\n\"Apprehension merely by means of sensation fills only one moment\n(that is, if I do not take into consideration the succession of many\nsensations). Sensation, therefore, as that in the phenomenon the\napprehension of which is not a successive synthesis advancing from\nparts to a complete representation, has no extensive quantity; the\nlack of sensation in one and the same moment would represent it as\nempty, consequently = 0. Now that which in the empirical perception\ncorresponds to sensation is reality (_realitas phaenomenon_); that\nwhich corresponds to the lack of it is negation = 0. But every\nsensation is capable of a diminution, so that it can decrease and thus\ngradually vanish. Therefore, between reality in the phenomenon and\nnegation there exists a continuous connexion of many possible\nintermediate sensations, the difference of which from each other is\nalways smaller than that between the given sensation and zero, or\ncomplete negation. That is to say, the real in the phenomenon has\nalways a quantity, which, however, is not found in apprehension, since\napprehension takes place by means of mere sensation in one moment and\nnot by a successive synthesis of many sensations, and therefore does\nnot proceed from parts to the whole. Consequently, it has a quantity,\nbut not an extensive quantity.\"\n\n\"Now that quantity which is apprehended only as unity, and in which\nplurality can be represented only by approximation to negation = 0,\nI call an _intensive quantity_. Every reality, therefore, in a\nphenomenon has intensive quantity, that is, a degree.\"[7]\n\n    [7] B. 209-10, M. 127.\n\nIn other words, 'We can lay down _a priori_ that all sensations have a\ncertain degree of intensity, and that between a sensation of a given\nintensity and the total absence of sensation there is possible an\ninfinite number of sensations varying in intensity from nothing to\nthat degree of intensity. Therefore the real, which corresponds to\nsensation, can also be said _a priori_ to admit of an infinite variety\nof degree.'\n\nThough the principle established is of little intrinsic importance,\nthe account of it is noticeable for two reasons. In the first place,\nalthough Kant clearly means by the 'real corresponding to sensation' a\nbody in space, and regards it as a phenomenon, it is impossible to see\nhow he can avoid the charge that he in fact treats it as a thing in\nitself.[8] For the correspondence must consist in the fact that the\nreal causes or excites sensation in us, and therefore the real, i. e.\na body in space, is implied to be a thing in itself. In fact, Kant\nhimself speaks of considering the real in the phenomenon as the cause\nof sensation,[9] and, in a passage added in the second edition, after\nproving that sensation must have an intensive quantity, he says that,\ncorresponding to the intensive quantity of sensation, an intensive\nquantity, i. e. _a degree of influence on sense_, must be attributed\nto all objects of sense-perception.[10] The difficulty of consistently\nmaintaining that the real, which corresponds to sensation, is a\nphenomenon is, of course, due to the impossibility of distinguishing\nbetween reality and appearance within phenomena.[11]\n\n    [8] Cf. p. 257 note.\n\n    [9] B. 210, M. 128.\n\n    [10] B. 208, M. 126. The italics are mine. Cf. from the same\n    passage, \"Phenomena contain, over and above perception, the\n    materials for some object (through which is represented\n    something existing in space and time), i. e. they contain the\n    real of sensation as a merely subjective representation of\n    which we can only become conscious that _the subject is\n    affected_, and which we relate _to an object in general_.\"\n    (The italics are mine.)\n\n    [11] Cf. pp. 94-100.\n\nIn the second place, Kant expressly allows that in this anticipation\nwe succeed in discovering _a priori_ a characteristic of sensation,\nalthough sensation constitutes that empirical element in phenomena,\nwhich on Kant's general view cannot be apprehended _a priori_.\n\n\"Nevertheless, this anticipation of sense-perception must always be\nsomewhat surprising to an inquirer who is used to transcendental\nreflection, and is thereby rendered cautious. It leads us to feel some\nmisgiving as to whether the understanding can anticipate such a\nsynthetic proposition as that respecting the degree of all that is\nreal in phenomena, and consequently respecting the possibility of the\ninternal distinction of sensation itself, if we abstract from its\nempirical quality. There remains, therefore, a problem not unworthy of\nsolution, viz. 'How can the understanding pronounce synthetically and\n_a priori_ upon phenomena in this respect, and thus anticipate\nphenomena even in that which is specially and merely empirical, viz.\nthat which concerns sensations?'\"[12] But although Kant recognizes\nthat the anticipation is surprising, he is not led to revise his\ngeneral theory, as being inconsistent with the existence of the\nanticipation. He indeed makes an attempt[13] to deal with the\ndifficulty; but his solution consists not in showing that the\nanticipation is consistent with his general theory--as he should have\ndone, if the theory was to be retained--but in showing that, in the\ncase of the degree of sensation, we do apprehend the nature of\nsensation _a priori_.\n\n    [12] B. 217, M. 131; cf. B. 209, M. 127.\n\n    [13] B. 217-18, M. 132.\n\nStrangely enough, Hume finds himself face to face with what is in\nprinciple the same difficulty, and treats it in a not dissimilar way.\n\"There is, however, one contradictory phenomenon, which may prove,\nthat 'tis not absolutely impossible for ideas to go before their\ncorrespondent impressions. I believe it will readily be allow'd, that\nthe several distinct ideas of colours, which enter by the eyes, or\nthose of sounds, which are convey'd by the hearing, are really\ndifferent from each other, tho' at the same time resembling. Now if\nthis be true of different colours, it must be no less so of the\ndifferent shades of the same colour, that each of them produces a\ndistinct idea, independent of the rest. For if this shou'd be deny'd,\n'tis possible, by the continual gradation of shades, to run a colour\ninsensibly into what is most remote from it; and if you will not allow\nany of the means to be different, you cannot without absurdity deny\nthe extremes to be the same. Suppose therefore a person to have\nenjoyed his sight for thirty years, and to have become perfectly well\nacquainted with colours of all kinds, excepting one particular shade\nof blue, for instance, which it never has been his fortune to meet\nwith. Let all the different shades of that colour, except that single\none, be plac'd before him, descending gradually from the deepest to\nthe lightest; 'tis plain that he will perceive a blank, where that\nshade is wanting, and will be sensible, that there is a greater\ndistance in that place betwixt the contiguous colours, than in any\nother. Now I ask, whether 'tis possible for him, from his own\nimagination, to supply this deficiency, and raise up to himself the\nidea of that particular shade, tho' it had never been conveyed to him\nby his senses? I believe there are few but will be of opinion that he\ncan; and this may serve as a proof, that the simple ideas are not\nalways derived from the correspondent impressions; tho' the instance\nis so particular and singular, that 'tis scarce worth our observing,\nand does not merit that for it alone we should alter our general\nmaxim.\"[14]\n\n    [14] Hume, _Treatise_, Bk. I, Part 1, Sec. 1.\n\n\n\n\nCHAPTER XII\n\nTHE ANALOGIES OF EXPERIENCE\n\n\nEach of the three categories of relation, i. e. those of substance and\naccident, of cause and effect, and of interaction between agent and\npatient involves, according to Kant, a special principle, and these\nspecial principles he calls 'analogies of experience'. They are stated\nthus:[1] (1) In all changes of phenomena the substance is permanent,\nand its quantity in nature is neither increased nor diminished. (2)\nAll changes take place according to the law of the connexion of cause\nand effect. (3) All substances, so far as they can be perceived in\nspace as coexistent, are in complete interaction. The justification of\nthe term _analogy_ of experience is as follows. In mathematics an\nanalogy is a formula which asserts the equality of two _quantitative_\nrelations, and is such that, if three of the terms are given, we can\ndiscover the fourth, e. g. if we know that _a_ : _b_ = _c_ : _d_, and\nthat _a_ = 2, _b_ = 4, _c_ = 6 we can discover that _d_ = 12. But in\nphilosophy an analogy is the assertion of the equality of two\n_qualitative_ relations and is such that, if three of the terms are\ngiven, we can discover, not the fourth, but only the relation of the\nthird to the fourth, though at the same time we are furnished with a\nclue whereby to search for the fourth in experience. In this\nphilosophical sense, the principles involved in the categories of\nrelation are analogies. For instance, the principles of causality can\nbe stated in the form 'Any known event _X_ is to _some other_ event\n_Y_, whatever it be, as effect to cause'; so stated, it clearly\ninforms us not of the character of _Y_ but only of the fact that there\nmust be a _Y_, i. e. a necessary antecedent, though at the same time\nthis knowledge enables us to search in experience for the special\ncharacter of _Y_.\n\n    [1] The formulation of them in the first edition is slightly\n    different.\n\nThe principles to be established relate to the two kinds of temporal\nrelation apprehended in the world of nature, viz. coexistence and\nsuccession. The _method_ of proof, which is to be gathered from the\nproofs themselves rather than from Kant's general remarks[2] on the\nsubject, is the same in each case. Kant expressly rejects any proof\nwhich is 'dogmatical' or 'from conceptions', e. g. any attempt to show\nthat the very conception of change presupposes the thought of an\nidentical subject of change.[3] The proof is transcendental in\ncharacter, i. e. it argues that the principle to be established is a\ncondition of the possibility of _apprehending_ the temporal relation\nin question, e. g. that the existence of a permanent subject of change\nis presupposed in any _apprehension_ of change. It assumes that we\nbecome aware of sequences and coexistences in the world of nature by a\nprocess which begins with a succession of mere perceptions, i. e.\nperceptions which are so far not the perceptions of a sequence or of a\ncoexistence or indeed of anything;[4] and it seeks to show that this\nprocess involves an appeal to one of the principles in question--the\nparticular principle involved depending on the temporal relation\napprehended--and consequently, that since we do apprehend this\ntemporal relation, which, as belonging to the world of nature, must be\ndistinct from any temporal relation of our perceptions, the principle\nappealed to is valid.\n\n    [2] B. 218-24, M. 132-6; and B. 262-5, M. 159-61.\n\n    [3] B. 263-4, M. 160-1; B. 289, M. 174-5.\n\n    [4] This assumption is of course analogous to the assumption\n    which underlies the _Transcendental Deduction of the\n    Categories_, that knowledge begins with the successive\n    origination in us of isolated data of sense.\n\nThe proof of the first analogy is given somewhat differently in the\nfirst edition, and in a passage added in the second. The earlier\nversion, which is a better expression of the attitude underlying\nKant's general remarks on the analogy, is as follows:\n\n\"Our _apprehension_ of the manifold of a phenomenon is always\nsuccessive, and is therefore always changing. By it alone, therefore,\nwe can never determine whether this manifold, as an object of\nexperience, is coexistent or successive, unless there lies at the base\nof it something that exists _always_, that is, something _enduring_\nand _permanent_, of which all succession and coexistence are nothing\nbut so many ways (_modi_ of time) in which the permanent exists. Only\nin the permanent, then, are time relations possible (for simultaneity\nand succession are the only relations in time); i. e. the permanent is\nthe _substratum_ of the empirical representation of time itself, in\nwhich alone all time-determination is possible. Permanence expresses\nin general time, as the persisting correlate of all existence of\nphenomena, of all change, and of all concomitance.... Only through the\npermanent does _existence_ in different parts of the successive series\nof time gain a _quantity_ which we call _duration_. For, in mere\nsuccession, existence is always vanishing and beginning, and never\nhas the least quantity. Without this permanent, then, no time\nrelation is possible. Now, time in itself cannot be perceived[5];\nconsequently this permanent in phenomena is the substratum of all\ntime-determination, and therefore also the condition of the\npossibility of all synthetic unity of sense-perceptions, that is, of\nexperience, and in this permanent all existence and all change in time\ncan only be regarded as a mode of the existence of that which endures\nand is permanent. Therefore in all phenomena the permanent is the\nobject itself, i. e. the substance (_phenomenon_); but all that\nchanges or can change belongs only to the way in which this substance\nor substances exist, consequently to their determinations.\"[6]\n\"Accordingly since substance cannot change in existence, its quantity\nin nature can neither be increased nor diminished.\"[7] The argument\nbecomes plainer if it be realized that in the interval between the two\neditions, Kant came to think that the permanent in question was matter\nor bodies in space.[8] \"We find that in order to give something\n_permanent_ in perception corresponding to the conception of\n_substance_ (and thereby to exhibit the objective reality of this\nconception), we need a perception _in space_ (of matter), because\nspace alone has permanent determinations, while time, and consequently\neverything which is in the internal sense, is continually flowing.\"[9]\n\n    [5] _Wahrgenommen._\n\n    [6] A. 182-4 and B. 225-7, M. 137-8. This formulation of the\n    conclusion is adapted only to the form in which the first\n    analogy is stated in the first edition, viz. \"All phenomena\n    contain the permanent (_substance_) as the object itself and\n    the changeable as its mere determination, i. e. as a way\n    in which the object exists.\" Hence a sentence from the\n    conclusion of the proof added in the second edition is quoted\n    to elucidate Kant's meaning; its doctrine is as legitimate a\n    conclusion of the argument given in the first edition as of\n    that peculiar to the second.\n\n    [7] B. 225, M. 137.\n\n    [8] Cf. Caird, i. 541-2.\n\n    [9] B. 291, M. 176 (in 2nd ed. only). Cf. B. 277 fin.-278\n    init., M. 168 (in 2nd ed. only).\n\nKant's thought appears to be as follows: 'Our apprehension of the\nmanifold consists of a series of successive acts in which we apprehend\nits elements one by one and in isolation. This apprehension,\ntherefore, does not enable us to determine that its elements are\ntemporally related either as successive or as coexistent.[10] In order\nto determine this, we must apprehend the elements of the manifold as\nrelated to something permanent. For a succession proper, i. e. a\nchange, is a succession of states or determinations of something\npermanent or unchanging. A mere succession which is not a succession\nof states of something which remains identical is an unconnected\nseries of endings and beginnings, and with respect to it, 'duration',\nwhich has meaning with regard to changes, i. e. successions proper,\nhas no meaning at all. Similarly, coexistence is a coexistence of\nstates of two permanents. Hence, to apprehend elements of the manifold\nas successive or coexistent, we must apprehend them in relation to a\npermanent or permanents. Therefore, to apprehend a coexistence or a\nsuccession, we must perceive something permanent. But this permanent\nsomething cannot be time, for time cannot be perceived. It must\ntherefore be a permanent in phenomena; and this must be the object\nitself or the substance of a phenomenon, i. e. the substratum of the\nchanges which it undergoes, or that of which the elements of the\nmanifold are states or modifications.[11] Consequently, there must be\na permanent substance of a phenomenon, and the quantity of substances\ntaken together must be constant.'\n\n    [10] The account of the first analogy as a whole makes it\n    necessary to think that Kant in the first two sentences of\n    the proof quoted does not mean exactly what he says, what he\n    says being due to a desire to secure conformity with his\n    treatment of the second and third analogies. What he _says_\n    suggests (1) that he is about to discuss the implications,\n    not of the process by which we come to apprehend the manifold\n    as temporally related in one of the two ways possible, i. e.\n    either as successive or as coexistent, but of the process by\n    which we decide whether the relation of the manifold which we\n    already know to be temporal is that of succession or that of\n    coexistence, and (2) that the necessity for this process is\n    due to the fact that our _apprehension_ of the manifold is\n    always successive. The context, however, refutes both\n    suggestions, and in any case it is the special function of\n    the processes which involve the second and third analogies to\n    determine the relations of the manifold as that of succession\n    and that of coexistence respectively.\n\n    [11] Cf. B. 225, M. 137 (first half).\n\nNow, if Kant's thought has been here represented fairly, it is open to\nthe following comments. In the first place, even if his position be\nright in the main, Kant should not introduce the thought of the\n_quantity_ of substance, and speak of the quantity as constant. For he\nthereby implies that in a plurality of substances--if such a plurality\ncan in the end be admitted--there may be total extinction of, or\npartial loss in, some, if only there be a corresponding compensation\nin others; whereas such extinction and creation would be inconsistent\nwith the nature of a substance.[12] Even Kant himself speaks of having\nestablished the impossibility of the origin and extinction of\nsubstance.[13]\n\n    [12] I owe this comment to Professor Cook Wilson.\n\n    [13] B. 232-3, M. 141 fin.\n\nIn the second place, it is impossible to see how it can be legitimate\nfor Kant to speak of a permanent substratum of change at all.[14] For\nphenomena or appearances neither are nor imply the substratum of which\nKant is thinking. They might be held to imply ourselves as the\nidentical substratum of which they are successive states, but this\nview would be irrelevant to, if not inconsistent with, Kant's\ndoctrine. It is all very well to _say_ that the substratum is to be\nfound in matter, i. e. in bodies in space,[15] but the assertion is\nincompatible with the phenomenal character of the world; for the\nsensations or appearances produced in us by the thing in itself cannot\nbe successive states of bodies in space. In the third place, in spite\nof Kant's protests against any proof which is 'dogmatical' or 'from\nconceptions', such a proof really forms the basis of his thought. For\nif the argument is to proceed not from the nature of change as such\nbut from the possibility of perceiving change, it must not take into\naccount any implications of the possibility of perceiving change which\nrest upon implications of the nature of change as such. Yet this is\nwhat the argument does. For the reason really given for the view that\nthe apprehension of change involves the apprehension of the manifold\nas related to a permanent substratum is that a change, as such,\nimplies a permanent substratum. It is only because change is held to\nimply a substratum that we are said to be able to apprehend a change\nonly in relation to a substratum. Moreover, shortly afterwards, Kant,\napparently without realizing what he is doing, actually uses what is,\non the very face of it, the dogmatic method, and in accordance with it\ndevelops the implications of the perception of change. \"Upon this\npermanence is based the justification of the conception of _change_.\nComing into being and perishing are not changes of that which comes to\nbe or perishes. Change is but a mode of existence, which follows on\nanother mode of existence of the same object. Hence everything which\nchanges _endures_ and only its _condition changes_.... Change,\ntherefore, can be perceived only in substances, and absolute coming to\nbe or perishing, which does not concern merely a determination of the\npermanent, cannot be a possible perception.\"[16] Surely the fact that\nKant is constrained in spite of himself to use the dogmatic method is\nsome indication that it is the right method. It is in reality\nimpossible to make any discoveries about change, or indeed about\nanything, except by consideration of the nature of the thing itself;\nno study of the conditions under which it can be apprehended can throw\nany light upon its nature.[17] Lastly, although the supposition is not\nso explicit as the corresponding supposition made in the case of the\nother analogies, Kant's argument really assumes, and assumes wrongly,\nthe existence of a process by which, starting with the successive\napprehension of elements of the manifold in isolation, we come to\napprehend them as temporally related.\n\n    [14] The term 'permanent' is retained to conform to Kant's\n    language. Strictly speaking, only a state of that which\n    changes can be said to persist or to be permanent; for the\n    substratum of change is not susceptible of any temporal\n    predicates. Cf. p. 306.\n\n    [15] B. 291, M. 176.\n\n    [16] B. 230-1, M. 176.\n\n    [17] Cf. pp. 300-1.\n\nThe deduction of the second and third analogies argues that the\nprinciples of causality and reciprocal action are involved\nrespectively in the processes by which we become aware of successions\nand of coexistences in the world of nature. From this point of view it\nwould seem that the first analogy is a presupposition of the others,\nand that the process which involves the first is presupposed by the\nprocess which involves the others. It would seem that it is only upon\nthe conclusion of a process by which, beginning with the successive\napprehension of elements of the manifold in isolation, we come to\napprehend them as _either_ successive or coexistent elements in the\nworld of nature, that there can arise a process by which we come to\ndecide _whether_ the specific relation is that of succession or of\ncoexistence. For if the latter process can take place independently of\nthe former, i. e. if it can start from the successive apprehension of\nthe manifold, the former process will be unnecessary, and in that\ncase the vindication of the first analogy will be invalid. It is\nnecessary, however, to distinguish between Kant's nominal and his\nactual procedure. Though he nominally regards the first analogy as the\npresupposition of the others,[18] he really does not. For he does not\nin fact treat the process which involves the validity of the first\nanalogy as an antecedent condition of the processes which involve the\nvalidity of the others. On the contrary, the latter processes begin\n_ab initio_ with the mere successive apprehension of the manifold,\ni. e. they begin at a stage where we are not aware of any relation in\nthe physical world at all; and Kant, in his account of them, nowhere\nurges that they involve the first analogy.[19]\n\n    [18] Cf. B. 229, M. 140; B. 232-3, M. 141-2; and Caird, i.\n    545 and ff.\n\n    [19] This is not disproved by B. 247-51, M. 150-2, which\n    involves a different conception of cause and effect.\n\nMoreover, just because Kant does not face the difficulties involved in\nthe thought of a process which begins in this way until he comes to\nvindicate causality, it is only when we come to this vindication that\nwe realize the real nature of his deduction of the analogies, and, in\nparticular, of that of the first.\n\nKant, prompted no doubt by his desire to answer Hume, treats the\nprinciple of causality very fully. The length of the discussion,\nhowever, is due not so much to the complication of the argument as to\nKant's desire to make his meaning unmistakable; his account consists\nmainly in a repetition of what is substantially the same argument no\nless than five times. Hence it will suffice to consider those passages\nwhich best express Kant's meaning. At the same time, the prominence of\nthe principle of causality in Kant's theory, and in the history of\nphilosophy generally, and also the way in which Kant's treatment of\nit reveals the true nature of his general position, makes it necessary\nto consider these passages in some detail.\n\nHume had denied that we are justified in asserting any causal\nconnexion, i. e. any necessity of succession in the various events\nwhich we perceive, but even this denial presupposed that we do\napprehend particular sequences in the world of nature, and therefore\nthat we succeed in distinguishing between a sequence of events in\nnature and a mere sequence of perceptions, such as is also to be found\nwhen we apprehend a coexistence of bodies in space. Kant urges, in\neffect, that this denial renders it impossible to explain, as we\nshould be able to do, the possibility of making the distinction in\nquestion, which even the denial itself presupposes that we make.\nHolding, with Hume, that in all cases of perception what we are\ndirectly aware of is a succession of perceptions, he contends that it\nis necessary to explain how in certain cases we succeed in passing\nfrom the knowledge of our successive perceptions to the knowledge of a\nsuccession in what we perceive. How is it that we know, when, as we\nsay, we see a boat going down stream, that there is a succession in\nwhat we perceive, and not merely a succession in our perception of it,\nas is the case when, as we say, we see the parts of a house? Hume,\naccording to Kant, cannot answer this question; he has only the right\nto say that in all cases we have a succession of perceptions; for in\nreality an answer to the question will show that the acquisition of\nthis knowledge involves an appeal to the principle of causality.\nSince, then, we do in fact, as even Hume implicitly allowed, succeed\nin distinguishing between a succession in objects in nature and a\nsuccession in our apprehension of them, the law of causality must be\ntrue. \"It is only under this presupposition (i. e. of causality) that\neven the experience of an event is possible.\"[20]\n\n    [20] B. 240, M. 146. For the general view, cf. Caird, i.\n    556-61.\n\nKant begins[21] his proof as follows: \"Our apprehension of the\nmanifold of a phenomenon is always successive. The representations of\nthe parts succeed one another. Whether they succeed one another in the\nobject also is a second point for reflection which is not contained in\nthe first.\"[22] But, before he can continue, the very nature of these\nopening sentences compels him to consider a general problem which they\nraise. The distinction referred to between a succession in our\napprehensions or representations and a succession in the object\nimplies an object distinct from the apprehensions or representations.\nWhat, then, can be meant by such an object? For prima facie, if we\nignore the thing in itself as unknowable, there is no object; there\nare only representations. But, in that case, what can be meant by a\nsuccession in the object? Kant is therefore once more[23] forced to\nconsider the question 'What is meant by object of representations?'\nalthough on this occasion with special reference to the meaning of a\nsuccession in the object; and the vindication of causality is bound up\nwith the answer. The answer is stated thus:\n\n    [21] The preceding paragraph is an addition of the second\n    edition.\n\n    [22] B. 234, M. 142.\n\n    [23] Cf. A. 104-5, Mah. 198-9, and pp. 178-86 and 230-3.\n\n\"Now we may certainly give the name of object to everything, and even\nto every representation, so far as we are conscious thereof; but what\nthis word may mean in the case of phenomena, not in so far as they (as\nrepresentations) are objects, but in so far as they only indicate an\nobject, is a question requiring deeper consideration. So far as they,\nas representations only, are at the same time objects of\nconsciousness, they are not to be distinguished from apprehension,\ni. e. reception into the synthesis of imagination, and we must\ntherefore say, 'The manifold of phenomena is always produced\nsuccessively in the mind'. If phenomena were things in themselves, no\nman would be able to infer from the succession of the representations\nof their manifold how this manifold is connected in the object. For\nafter all we have to do only with our representations; how things may\nbe in themselves, without regard to the representations through which\nthey affect us, is wholly outside the sphere of our knowledge. Now,\nalthough phenomena are not things in themselves, and are nevertheless\nthe only thing which can be given to us as data for knowledge, it is\nmy business to show what kind of connexion in time belongs to the\nmanifold in phenomena themselves, while the representation of this\nmanifold in apprehension is always successive. Thus, for example, the\napprehension of the manifold in the phenomenon of a house which stands\nbefore me is successive. Now arises the question, whether the manifold\nof this house itself is in itself also successive, which of course no\none will grant. But, so soon as I raise my conceptions of an object to\nthe transcendental meaning thereof, the house is not a thing in\nitself, but only a phenomenon, i. e. a representation, the\ntranscendental object of which is unknown. What, then, am I to\nunderstand by the question, 'How may the manifold be connected in the\nphenomenon itself (which is nevertheless nothing in itself)?' Here\nthat which lies in the successive apprehension is regarded as\nrepresentation, while the phenomenon which is given me, although it\nis nothing more than a complex of these representations, is regarded\nas the object thereof, with which my conception, drawn from the\nrepresentations of apprehension, is to agree. It is soon seen that,\nsince agreement of knowledge with the object is truth, we can ask here\nonly for the formal conditions of empirical truth, and that the\nphenomenon, in opposition to the representations of apprehension, can\nonly be represented as the object of the same, distinct therefrom, if\nit stands under a rule, which distinguishes it from every other\napprehension, and which renders necessary a mode of conjunction of the\nmanifold. That in the phenomenon which contains the condition of this\nnecessary rule of apprehension is the object.\"[24]\n\n    [24] B. 234-6, M. 143-4. Cf. B. 242, M. 147.\n\nThis passage is only intelligible if we realize the _impasse_ into\nwhich Kant has been led by his doctrine that objects, i. e. realities\nin the physical world, are only representations or ideas. As has\nalready been pointed out,[25] an apprehension is essentially\ninseparable from a reality of which it is the apprehension. In other\nwords, an apprehension is always the apprehension of a reality, and a\nreality apprehended, i. e. an object of apprehension, cannot be stated\nin terms of the apprehension of it. We never confuse an apprehension\nand its object; nor do we take the temporal relations which belong to\nthe one for the temporal relations which belong to the other, for\nthese relations involve different terms which are never confused, viz.\napprehensions and the objects apprehended. Now Kant, by his doctrine\nof the unknowability of the thing in itself, has really deprived\nhimself of an object of apprehension or, in his language, of an\nobject of representations. For it is the thing in itself which is,\nproperly speaking, the object of the representations of which he is\nthinking, i. e. representations of a reality in nature; and yet the\nthing in itself, being on his view inapprehensible, can never be for\nhim an object in the proper sense, i. e. a reality apprehended. Hence\nhe is only able to state the fact of knowledge in terms of mere\napprehensions, or ideas, or representations--the particular name is a\nmatter of indifference--and consequently his efforts to recover an\nobject of apprehension are fruitless. As a matter of fact, these\nefforts only result in the assertion that the object of\nrepresentations consists in the representations themselves related in\na certain necessary way. But this view is open to two fatal\nobjections. In the first place, a complex of representations is just\nnot an object in the proper sense, i. e. a reality apprehended. It\nessentially falls on the subject side of the distinction between an\napprehension and the reality apprehended. The _complexity_ of a\ncomplex of representations in no way divests it of the character which\nit has as a complex of _representations_. In the second place, on this\nview the same terms have to enter at once into two incompatible\nrelations. Representations have to be related successively as our\nrepresentations or apprehensions--as in fact they are related--and, at\nthe same time, successively or otherwise, as the case may be, as parts\nof the object apprehended, viz. a reality in nature. In other words,\nthe same terms have to enter into both a subjective and an objective\nrelation, i. e. both a relation concerning us, the knowing subjects,\nand a relation concerning the object which we know.[26] \"A phenomenon\nin opposition to the representations of apprehension can only be\nrepresented as the object of the same, distinct therefrom, if it\nstands under a rule which distinguishes it from _every other_\napprehension, and renders necessary a mode of conjunction of the\nmanifold.\"[27] A representation, however, cannot be so related by a\nrule to another representation, for the rule meant relates to\nrealities in nature, and, however much Kant may try to maintain the\ncontrary, two representations, not being realities in nature, cannot\nbe so related. Kant is in fact only driven to treat rules of nature as\nrelating to representations, because there is nothing else to which he\ncan regard them as relating. The result is that he is unable to\njustify the very distinction, the implications of which it is his aim\nto discover, and he is unable to do so for the very reason which would\nhave rendered Hume unable to justify it. Like Hume, he is committed to\na philosophical vocabulary which makes it meaningless to speak of\nrelations of objects at all in distinction from relations of\napprehensions. It has been said that for Kant the road to objectivity\nlay through necessity.[28] But whatever Kant may have thought, in\npoint of fact there is no road to objectivity, and, in particular, no\nroad through necessity. No necessity in the relation between two\nrepresentations can render the relation objective, i. e. a relation\nbetween objects. No doubt the successive acts in which we come to\napprehend the world are necessarily related; we certainly do not\nsuppose their order to be fortuitous. Nevertheless, their relations\nare not in consequence a relation of realities apprehended.\n\n    [25] pp. 133-4; cf. pp. 180 and 230-1.\n\n    [26] Cf. p. 209, note 3, and p. 233.\n\n    [27] The italics are mine.\n\n    [28] Caird, i. 557.\n\nKant only renders his own view plausible by treating an apprehension\nor representation as if it consisted in a sensation or an appearance.\nA sensation or an appearance, so far from being the apprehension of\nanything, is in fact a reality which can be apprehended, of the kind\ncalled mental. Hence it can be treated as an object, i. e. something\napprehended or presented, though not really as an object in nature. On\nthe other hand, from the point of view of the thing in itself it can\nbe treated as only an apprehension, even though it is an unsuccessful\napprehension. Thus, for Kant, there is something which can with some\nplausibility be treated as an object as well as an apprehension, and\ntherefore as capable of standing in both a subjective and an objective\nrelation to other realities of the same kind.[29]\n\n    [29] Cf. pp. 137 and 231.\n\nIf we now turn to the passage under discussion, we find it easy to\nvindicate the justice of the criticism that Kant, inconsistently with\nthe distinction which he desires to elucidate, treats the same thing\nas at once the representation of an object and the object represented.\nHe is trying to give such an account of 'object of representations' as\nwill explain what is meant by a succession in an object in nature,\ni. e. a phenomenon, in distinction from the succession in our\napprehension of it. In order to state this distinction at all, he has\nto speak of what enters into the two successions as different. \"It is\nmy business to show what sort of connexion in time belongs to the\n_manifold_ in phenomena themselves, while the _representation_ of this\nmanifold in apprehension is always successive.\"[30] Here an element of\nthe manifold is distinguished from the representation of it. Yet Kant,\nthough he thus distinguishes them, repeatedly identifies them; in\nother words, he identifies a representation with that of which it is a\nrepresentation, viz. an element in or part of the object itself. \"_Our\napprehension_ of the manifold of the phenomenon is always successive.\n_The representations_ of the parts succeed one another. Whether _they_\n[i. e. _the representations_[31]] succeed one another _in the object_\nalso, is a second point for reflection.... So far as they [i. e.\nphenomena], as representations only, are at the same time objects of\nconsciousness, they are not to be distinguished from apprehension,\ni. e. reception into the synthesis of imagination, and we must\ntherefore say, _'The manifold of phenomena_ is always produced\nsuccessively in the mind'. If phenomena were things in themselves, no\nman would be able to infer from the succession of the representations\nhow _this manifold_ is connected _in the object_.... The phenomenon,\nin opposition to the representations of apprehension, can only be\nrepresented as the object of the same, distinct therefrom, if it\nstands under a rule, which distinguishes _it_ from every _other_\nrepresentation and which renders necessary a mode of conjunction of\nthe manifold.\"[32]\n\n    [30] The italics are mine.\n\n    [31] This is implied both by the use of 'also' and by the\n    context.\n\n    [32] The italics are mine.\n\nSince Kant in introducing his vindication of causality thus identifies\nelements in the object apprehended (i. e. the manifold of phenomena)\nwith the apprehensions of them, we approach the vindication itself\nwith the expectation that he will identify a causal rule, which\nconsists in a necessity in the succession of objects, viz. of events\nin nature, with the necessity in the succession of our apprehensions\nof them. This expectation turns out justified. The following passage\nadequately expresses the vindication:\n\n\"Let us now proceed to our task. That something happens, i. e. that\nsomething or some state comes to be which before was not, cannot be\nempirically perceived, unless a phenomenon precedes, which does not\ncontain in itself this state; for a reality which follows upon an\nempty time, and therefore a coming into existence preceded by no state\nof things, can just as little be apprehended as empty time itself.\nEvery apprehension of an event is therefore a perception which follows\nupon another perception. But because this is the case with all\nsynthesis of apprehension, as I have shown above[33] in the phenomenon\nof a house, the apprehension of an event is thereby not yet\ndistinguished from other apprehensions. But I notice also, that if in\na phenomenon which contains an event, I call the preceding state of my\nperception A, and the following state B, B can only follow A in\napprehension, while the perception A cannot follow B but can only\nprecede it. For example, I see a ship float down a stream. My\nperception of its place lower down follows upon my perception of its\nplace higher up the course of the river, and it is impossible that in\nthe apprehension of this phenomenon the vessel should be perceived\nfirst below and afterwards higher up the stream. Here, therefore, the\norder in the sequence of perceptions in apprehension is determined,\nand apprehension is bound to this order. In the former example of a\nhouse, my perceptions in apprehension could begin at the roof and end\nat the foundation, or begin below and end above; in the same way they\ncould apprehend the manifold of the empirical perception from left to\nright, or from right to left. Accordingly, in the series of these\nperceptions, there was no determined order, which necessitated my\nbeginning at a certain point, in order to combine the manifold\nempirically. But this rule is always to be found in the perception of\nthat which happens, and it makes the order of the successive\nperceptions (in the apprehension of this phenomenon) _necessary_.\"\n\n    [33] B. 235-6, M. 143 (quoted p. 279).\n\n\"In the present case, therefore, I shall have to derive the\n_subjective sequence_ of apprehension from the _objective sequence_ of\nphenomena, for otherwise the former is wholly undetermined, and does\nnot distinguish one phenomenon from another. The former alone proves\nnothing as to the connexion of the manifold in the object, for it is\nwholly arbitrary. The latter, therefore [i. e. the objective sequence\nof phenomena[34]], will consist in that order of the manifold of the\nphenomenon, according to which the apprehension of the one (that which\nhappens) follows that of the other (that which precedes) _according to\na rule_. In this way alone can I be justified in saying of the\nphenomenon itself, and not merely of my apprehension, that a sequence\nis to be found therein, which is the same as to say that I cannot\narrange my apprehension otherwise than in just this sequence.\"\n\n    [34] The sense is not affected if 'the latter' be understood\n    to refer to the connexion of the manifold in the object.\n\n\"In conformity with such a rule, therefore, there must exist in that\nwhich in general precedes an event the condition of a rule, according\nto which this event follows always and necessarily, but I cannot\nconversely go back from the event, and determine (by apprehension)\nthat which precedes it. For no phenomenon goes back from the\nsucceeding point of time to the preceding point, although it does\ncertainly relate to _some preceding point of time_; on the other hand,\nthe advance from a given time to the determinate succeeding time is\nnecessary. Therefore, because there certainly is something which\nfollows, I must relate it necessarily to something else in general,\nwhich precedes, and upon which it follows in conformity with a rule,\nthat is necessarily, so that the event, as the conditioned, affords\ncertain indication of some condition, while this condition determines\nthe event.\"\n\n\"If we suppose that nothing precedes an event, upon which this event\nmust follow in conformity with a rule, all sequence of perception\nwould exist only in apprehension, i. e. would be merely subjective,\nbut it would not thereby be objectively determined which of the\nperceptions must in fact be the preceding and which the succeeding\none. We should in this manner have only a play of representations,\nwhich would not be related to any object, i. e. no phenomenon would be\ndistinguished through our perception in respect of time relations from\nany other, because the succession in apprehension is always of the\nsame kind, and so there is nothing in the phenomenon to determine the\nsuccession, so as to render a certain sequence objectively necessary.\nI could therefore not say that in the phenomenon two states follow\neach other, but only that one apprehension follows on another, a fact\nwhich is merely _subjective_ and does not determine any object, and\ncannot therefore be considered as knowledge of an object (not even in\nthe phenomenon).\"\n\n\"If therefore we experience that something happens, we always thereby\npresuppose that something precedes, on which it follows according to a\nrule. For otherwise, I should not say of the object, that it follows,\nbecause the mere sequence in my apprehension, if it is not determined\nby a rule in relation to something preceding, does not justify the\nassumption of a sequence in the object. It is therefore always in\nreference to a rule, according to which phenomena are determined in\ntheir sequence (i. e. as they happen) by the preceding state, that I\nmake my subjective synthesis (of apprehension) objective, and it is\nsolely upon this presupposition that even the experience of something\nwhich happens is possible.\"[35]\n\n    [35] B. 236-41, M. 144-6.\n\nThe meaning of the first paragraph is plain. Kant is saying that when\nwe reflect upon the process by which we come to apprehend the world of\nnature, we can lay down two propositions. The first is that the\nprocess is equally successive whether the object apprehended be a\nsuccession in nature or a coexistence of bodies in space, so that the\nknowledge that we have a succession of apprehensions would not by\nitself enable us to decide whether the object of the apprehensions is\na sequence or not. The second proposition is that, nevertheless, there\nis this difference between the succession of our apprehensions where\nwe apprehend a succession and where we apprehend a coexistence, that\nin the former case, and in that only, the succession of our\napprehensions is irreversible or, in other words, is the expression of\na rule of order which makes it a necessary succession. So far we find\nno mention of causality, i. e. of a necessity of succession in\nobjects, but only a necessity of succession in our apprehension of\nthem. So far, again, we find no contribution to the problem of\nexplaining how we distinguish between successive perceptions which are\nthe perceptions of an event and those which are not. For it is\nreasonable to object that it is only possible to say that the order of\nour perceptions is irreversible, if and because we already know that\nwhat we have been perceiving is an event, and that therefore any\nattempt to argue from the irreversibility of our perceptions to the\nexistence of a sequence in the object must involve a [Greek: hysteron\nproteron]. And it is clear that, if irreversibility in our perceptions\nwere the only irreversibility to which appeal could be made, even Kant\nwould not have supposed that the apprehension of a succession was\nreached through belief in an irreversibility.\n\nThe next paragraph, of which the interpretation is difficult, appears\nto introduce a causal rule, i. e. an irreversibility in objects, by\nidentifying it with the irreversibility in our perceptions of which\nKant has been speaking. The first step to this identification is taken\nby the assertion: \"In the present case, therefore, I shall have to\nderive the subjective sequence of perceptions from the objective\nsequence of phenomena.... The latter will consist in the order of the\n_manifold of the phenomenon_, according to which _the apprehension_ of\nthe one (that which happens) follows that of the other (that which\nprecedes) according to a rule.\"[36] Here Kant definitely implies that\nan objective sequence, i. e. an order or sequence of the _manifold_ of\na phenomenon, consists in a sequence of _perceptions or apprehensions_\nof which the order is necessary or according to a rule; in other\nwords, that a succession of perceptions in the special case where the\nsuccession is necessary is a succession of events perceived.[37] This\nimplication enables us to understand the meaning of the assertion that\n'we must therefore derive the subjective sequence of perceptions from\nthe objective sequence of phenomena', and to see its connexion with\nthe preceding paragraph. It means, 'in view of the fact that in all\napprehensions of a succession, and in them alone, the sequence of\nperceptions is irreversible, we are justified in saying that a given\nsequence of perceptions is the apprehension of a succession, if we\nknow that the sequence is irreversible; in that case we must be\napprehending a real succession, for an irreversible sequence of\nperceptions _is_ a sequence of events perceived.' Having thus implied\nthat irreversibility of perceptions constitutes them events perceived,\nhe is naturally enough able to go on to speak of the irreversibility\nof perceptions as if it were the same thing as an irreversibility of\nevents perceived, and thus to bring in a causal rule. \"In this way\nalone [i. e. only by deriving the subjective from the objective\nsequence] can I be justified in saying of the phenomenon itself, and\nnot merely of my apprehension, that a sequence is to be found therein,\n_which is the same as to say_ that I cannot _arrange_ my apprehension\notherwise than in just this sequence. In conformity with _such a\nrule_, therefore, there must exist in that which in general precedes\n_an event_ the condition of a rule, according to which _this event\nfollows always and necessarily_.\"[38] Here the use of the word\n'arrange'[39] and the statement about the rule in the next sentence\nimply that Kant has now come to think of the rule of succession as a\ncausal rule relating to the objective succession. Moreover, if any\ndoubt remains as to whether Kant really confuses the two\nirreversibilities or necessities of succession, it is removed by the\nlast paragraph of the passage quoted. \"If therefore we experience that\nsomething happens, we always thereby presuppose that something\nprecedes on which _it_ follows according to a rule. For otherwise I\nshould not say of the object that _it_ follows; because the mere\nsuccession of my apprehension, if _it_ is not determined by a rule in\nrelation to something preceding, does not justify the assumption of a\nsuccession in the object. It is therefore always in reference to a\nrule, according to which _phenomena_ are determined in their sequence\n(i. e. as they happen) by the preceding state, that I make my\nsubjective sequence (of apprehension) objective.\"[40] The fact is\nsimply that Kant _must_ identify the two irreversibilities, because,\nas has been pointed out, he has only one set of terms to be related as\nirreversible, viz. the elements of the manifold, which have to be,\nfrom one point of view, elements of an object and, from another,\nrepresentations or apprehensions of it.\n\n    [36] The italics are mine. 'According to which' does not\n    appear to indicate that the two orders referred to are\n    different.\n\n    [37] Cf. B. 242 fin., M. 147 fin.\n\n    [38] The italics are mine\n\n    [39] _Anstellen._\n\n    [40] The italics are mine.\n\nAs soon, therefore, as the real nature of Kant's vindication of\ncausality has been laid bare, it is difficult to describe it as an\nargument at all. He is anxious to show that in apprehending A B as a\nreal or objective succession we presuppose that they are elements in a\ncausal order of succession. Yet in support of his contention he points\nonly to the quite different fact that where we apprehend a succession\nA B, we think of the _perception_ of A and the _perception_ of B as\nelements in a necessary but subjective succession.\n\nBefore we attempt to consider the facts with which Kant is dealing, we\nmust refer to a feature in Kant's account to which no allusion has\nbeen made. We should on the whole expect from the passage quoted that,\nin the case where we regard two perceptions A B as necessarily\nsuccessive and therefore as constituting an objective succession, the\nnecessity of succession consists in the fact that A is the cause of B.\nThis, however, is apparently not Kant's view; on the contrary, he\nseems to hold that, in thinking of A B as an objective succession, we\npresuppose not that A causes B, but only that the state of affairs\nwhich precedes B, and which therefore includes A, contains a cause of\nB, the coexistence or identity of this cause with A rendering the\nparticular succession A B necessary. \"Thus [if I perceive that\nsomething happens] it arises that there comes to be an order among our\nrepresentations in which the present (so far as it has taken place)\npoints to some preceding state as a correlate, _though a still\nundetermined correlate_,[41] of this event which is given, and this\ncorrelate relates to the event by determining the event as its\nconsequence, and connects the event with itself necessarily in the\nseries of time.\"[42]\n\n    [41] The italics are mine.\n\n    [42] B. 244, M. 148. Cf. B. 243, M. 148 (first half) and B.\n    239, M. 145 (second paragraph). The same implication is to\n    be found in his formulation of the rule involved in the\n    perception of an event, e. g. \"In conformity with such a\n    rule, there must exist in that which in general precedes an\n    event, the condition of a rule, according to which this event\n    follows always and necessarily.\" Here the condition of a rule\n    is the necessary antecedent of the event, whatever it may be.\n\nThe fact is that Kant is in a difficulty which he feels obscurely\nhimself. He seems driven to this view for two reasons. If he were to\nmaintain that A was necessarily the cause of B, he would be\nmaintaining that all observed sequences are causal, i. e. that in them\nthe antecedent and consequent are always cause and effect, which is\npalpably contrary to fact. Again, his aim is to show that we become\naware of a succession by presupposing the law of causality. This law,\nhowever, is quite general, and only asserts that _something_ must\nprecede an event upon which it follows always and necessarily. Hence\nby itself it palpably gives no means of determining whether this\nsomething is A rather than anything else.[43] Therefore if he were\nto maintain that the antecedent member of an apprehended objective\nsuccession must be thought of as its cause, the analogy would\nobviously provide no means of determining the antecedent member, and\ntherefore the succession itself, for the succession must be the\nsequence of B upon some definite antecedent. On the other hand, the\nview that the cause of B need not be A only incurs the same difficulty\nin a rather less obvious form. For, even on this view, the argument\nimplies that in order to apprehend two individual perceptions A B as\nan objective succession, we must know that A _must_ precede B, and the\npresupposition that B implies a cause in the state of affairs\npreceding B in no way enables us to say either that A coexists with\nthe cause, or that it is identical with it, and therefore that it must\nprecede B.\n\n    [43] Cf. B. 165, M. 101, where Kant points out that the\n    determination of particular laws of nature requires\n    experience.\n\nNevertheless, it cannot be regarded as certain that Kant did not think\nof A, the apprehended antecedent of B, as necessarily the cause of B,\nfor his language is both ambiguous and inconsistent. When he considers\nthe apprehension of a succession from the side of the successive\nperceptions, he at least tends to think of A B as cause and\neffect;[44] and it may well be that in discussing the problem from\nthe side of the law of causality, he means the cause of B to be A,\nalthough the generality of the law compels him to refer to it as\n_something_ upon which B follows according to a rule.\n\n    [44] He definitely implies this, B. 234, M. 142.\n\nFurther, it should be noticed that to allow as Kant, in effect, does\nelsewhere[45], that experience is needed to determine the cause of B is\nreally to concede that the apprehension of objective successions is\n_prior to_, and _presupposed by_, any process which appeals to the\nprinciple of causality; for if the principle of causality does not by\nitself enable us to determine the cause of B, it cannot do more than\nenable us to pick out the cause of B among events known to precede B\nindependently of the principle. Hence, from this point of view, there\ncan be no process such as Kant is trying to describe, and therefore\nits precise nature is a matter of indifference.\n\n    [45] Cf. B. 165, M. 101, where Kant points out that the\n    determination of particular laws of nature requires\n    experience.\n\nWe may now turn to the facts. There is, it seems, no such thing as a\nprocess by which, beginning with the knowledge of successive\napprehensions or representations, of the object of which we are\nunaware, we come to be aware of their object. Still less is there a\nprocess--and it is really this which Kant is trying to describe--by\nwhich, so beginning, we come to apprehend these successive\nrepresentations as objects, i. e. as parts of the physical world,\nthrough the thought of them as necessarily related. We may take Kant's\ninstance of our apprehension of a boat going down stream. We do not\nfirst apprehend two perceptions of which the object is undetermined\nand then decide that their object is a succession rather than a\ncoexistence. Still less do we first apprehend two perceptions or\nrepresentations and then decide that they are related as successive\nevents in the physical world. From the beginning we apprehend a real\nsequence, viz. the fact that the boat having left one place is\narriving at another; there is no process _to_ this apprehension. In\nother words, from the beginning we are aware of real elements, viz. of\nevents in nature, and we are aware of them as really related, viz. as\nsuccessive in nature. This must be so. For if we begin with the\nawareness of two mere perceptions, we could never thence reach the\nknowledge that their object was a succession, or even the knowledge\nthat they had an object; nor, so beginning, could we become aware of\nthe perceptions themselves as successive events in the physical world.\nFor suppose, _per impossibile_, the existence of a process by which we\ncome to be aware of two elements A and B as standing in a relation of\nsequence in the physical world. In the first place, A and B, with the\nawareness of which we begin, must be, and be known to be, real or\nobjective, and not perceptions or apprehensions; otherwise we could\nnever come to apprehend them as related in the physical world. In the\nsecond place, A and B must be, and be known to be, real with the\nreality of a physical event, otherwise we could never come to\napprehend them as related by way of succession in the physical world.\nIf A and B were bodies, as they are when we apprehend the parts of a\nhouse, they could never be apprehended as successive. In other words,\nthe process by which, on Kant's view, A and B become, and become known\nto be, events presupposes that they already are, and are known to be,\nevents. Again, even if it be granted that A and B are real events, it\nis clear that there can be no process by which we come to apprehend\nthem as successive. For if we apprehended events A and B separately,\nwe could never thence advance to the apprehension of their relation,\nor, in other words, we could never discover which came first. Kant\nhimself saw clearly that the perception of A followed by the\nperception of B does not by itself yield the perception that B follows\nA. In fact it was this insight which formed the starting-point of his\ndiscussion.[46] Unfortunately, instead of concluding that the\napprehension of a succession is ultimate and underivable from a more\nprimitive apprehension, he tried to formulate the nature of the\nprocess by which, starting from such a succession of perceptions, we\nreach the apprehension of a succession. The truth is simply that there\nis and can be no _process to_ the apprehension of a succession; in\nother words, that we do and must apprehend a real succession\nimmediately or not at all. The same considerations can of course be\nsupplied _mutatis mutandis_ to the apprehension of the coexistence of\nbodies in space, e. g. of the parts of a house.\n\n    [46] Cf. B. 237, M. 144.\n\nIt may be objected that this denial of the existence of the process\nwhich Kant is trying to describe must at least be an overstatement.\nFor the assertion that the apprehension of a succession or of a\ncoexistence is immediate may seem to imply that the apprehension of\nthe course of a boat or of the shape of a house involves no process at\nall; yet either apprehension clearly takes time and so must involve a\nprocess. But though a process is obviously involved, it is not a\nprocess from the apprehension of what is not a succession to the\napprehension of a succession, but a process from the apprehension of\none succession to that of another. It is the process by which we pass\nfrom the apprehension of one part of a succession which may have, and\nwhich it is known may have, other parts to the apprehension of what\nis, and what is known to be, another part of the same succession.\nMoreover, the assertion that the apprehension of a succession must be\nimmediate does not imply that it may not be reached by a process. It\nis not inconsistent with the obvious fact that to apprehend that the\nboat is now turning a corner is really to apprehend that what before\nwas going straight is now changing its course, and therefore\npresupposes a previous apprehension of the boat's course as straight.\nIt only implies that the apprehension of a succession, if reached by a\nprocess at all, is not reached by a process of which the\nstarting-point is not itself the apprehension of a succession.\n\nNevertheless, a plausible defence of Kant's treatment of causality can\nbe found, which may be formulated thus: 'Time, just as much as space,\nis a sphere within which we have to distinguish between appearance and\nreality. For instance, when moving in a lift, we see, as we say, the\nwalls moving, while the lift remains stationary. When sitting in a\ntrain which is beginning to move out of a station, we see, as we say,\nanother train beginning to move, although it is in fact standing\nstill. When looking at distant trees from a fast train, we see, as we\nsay, the buildings in the intermediate space moving backwards. In\nthese cases the events seen are not real, and we only succeed in\ndetermining what is really happening, by a process which presupposes\nthe law of causality. Thus, in the last case we only believe that the\nintermediate buildings do not move, by realizing that, given the\nuniformity of nature, belief in their motion is incompatible with what\nwe believe on the strength of experience of these buildings on other\noccasions and of the rest of the world. These cases prove the\nexistence of a process which enables us, and is required to enable us,\nto decide whether a given change is objective or subjective, i. e.\nwhether it lies in the reality apprehended or in our apprehension of\nit; and this process involves an appeal to causality. Kant's mistake\nlay in his choice of illustrations. His illustrations implied that the\nprocess which involves causality is one by which we distinguish a\nsuccession in the object apprehended from another relation in the\nobject, viz. a coexistence of bodies. But he ought to have taken\nillustrations which implied that the process is one by which we\ndistinguish a succession in the object from a succession in our\nperception of it. In other words, the illustrations should, like those\njust given, have illustrated the process by which we distinguish an\nobjective from a subjective change, and not a process by which we\ndistinguish an objective change from something else also objective.\nConsequently, Kant's conclusion and his _general_ method of treatment\nare right, even if, misled by his instances, he supports his position\nby arguments which are wrong.'\n\nThis defence is, however, open to the following reply: 'At first sight\nthe cases taken undoubtedly seem to illustrate a process in which we\nseek to discover whether a certain change belongs to objects or only\nto our apprehension of them, and in which we appeal to causality in\narriving at a decision. But this is only because we ignore the\nrelativity of motion. To take the third case: our first statement of\nthe facts is that we saw the intermediate buildings moving, but that\nsubsequent reflection on the results of other experience forced us to\nconclude that the change perceived was after all only in our\napprehension and not in the things apprehended. The statement,\nhowever, that we saw the buildings moving really assumes that we, the\nobservers, were stationary; and it states too much. What we really\nperceived was a relative changing of position between us, the near\nbuildings, and the distant trees. This is a fact, and the apprehension\nof it, therefore, does not afterwards prove mistaken. It is equally\ncompatible with motion on the part of the trees, or of the buildings,\nor of the observers, or of a combination of them; and that for which\nan appeal to causality is needed is the problem of deciding which of\nthese alternatives is correct. Moreover, the perceived relative change\nof position is objective; it concerns the things apprehended. Hence,\nin this case too, it can be said that we perceive an objective\nsuccession from the beginning, and that the appeal to causality is\nonly needed to determine something further about it. It is useless to\nurge that to be aware of an event is to be aware of it in all its\ndefiniteness, and that this awareness admittedly involves an appeal to\ncausality; for it is easy to see that unless our awareness of the\nrelative motion formed the starting-point of any subsequent process in\nwhich we appealed to the law of causality, we could never use the law\nto determine which body really moved.'\n\nTwo remarks may be made in conclusion. In the first place, the basis\nof Kant's account, viz. the view that in our apprehension of the world\nwe advance from the apprehension of a succession of perceptions to\nthe apprehension of objects perceived, involves a [Greek: hysteron\nproteron]. As Kant himself in effect urges in the _Refutation of\nIdealism_,[47] self-consciousness, in the sense of the consciousness\nof the successive process in which we apprehend the world, is plainly\nonly attained by reflecting upon our apprehension of the world. We\nfirst apprehend the world and only by subsequent reflection become\naware of our activity in apprehending it. Even if consciousness of\nthe world must lead to, and so is in a sense inseparable from,\nself-consciousness, it is none the less its presupposition.\n\n    [47] Cf. p. 320.\n\nIn the second place, it seems that the true vindication of causality,\nlike that of the first analogy, lies in the dogmatic method which Kant\nrejects. It consists in insight into the fact that it is of the very\nnature of a physical event to be an element in a process of change\nundergone by a system of substances in space, this process being\nthrough and through necessary in the sense that any event (i. e. the\nattainment of any state by a substance) is the outcome of certain\npreceding events (i. e. the previous attainment of certain states by\nit and other substances), and is similarly the condition of certain\nsubsequent events.[48] To attain this insight, we have only to reflect\nupon what we really mean by a 'physical event'. The vindication can\nalso be expressed in the form that the very _thought_ of a physical\nevent presupposes the _thought_ of it as an element in a necessary\nprocess of change--provided, however, that no distinction is implied\nbetween the nature of a thing and what we think its nature to be. But\nto vindicate causality in this way is to pursue the dogmatic method;\nit is to argue from the nature, or, to use Kant's phrase, from the\nconception, of a physical event. On the other hand, it seems that the\nmethod of arguing transcendentally, or from the possibility of\nperceiving events, must be doomed to failure in principle. For if, as\nhas been argued to be the case,[49] apprehension is essentially the\napprehension of a reality as it exists independently of the\napprehension of it, only those characteristics can be attributed to\nit, as characteristics which it must have if it is to be apprehended,\nwhich belong to it in its own nature or in virtue of its being what it\nis. It can only be because we think that a thing has some\ncharacteristic in virtue of its own nature, and so think\n'dogmatically', that we can think that in apprehending it we must\napprehend it as having that characteristic.[50]\n\n    [48] This statement of course includes the third analogy.\n\n    [49] Cf. Chh. IV and VI.\n\n    [50] Cf. p. 275.\n\nThere remains to be considered Kant's proof of the third analogy,\ni. e. the principle that all substances, so far as they can be\nperceived in space as coexistent, are in thorough-going interaction.\nThe account is extremely confused, and it is difficult to extract from\nit a consistent view. We shall consider here the version added in the\nsecond edition, as being the fuller and the less unintelligible.\n\n\"Things are _coexistent_, when in empirical intuition[51] the\nperception[52] of the one can follow upon the perception of the other,\nand vice versa (which cannot occur in the temporal succession of\nphenomena, as we have shown in the second principle). Thus I can\ndirect my perception first to the moon and afterwards to the earth, or\nconversely, first to the earth and then to the moon, and because the\nperceptions of these objects can reciprocally follow each other, I say\nthat they coexist. Now coexistence is the existence of the manifold in\nthe same time. But we cannot perceive time itself, so as to conclude\nfrom the fact that things are placed in the same time that the\nperceptions of them can follow each other reciprocally. The synthesis\nof the imagination in apprehension, therefore, would only give us each\nof these perceptions as existing in the subject when the other is\nabsent and vice versa; but it would not give us that the objects are\ncoexistent, i. e. that, if the one exists, the other also exists in\nthe same time, and that this is necessary in order that the\nperceptions can follow each other reciprocally. Hence there is needed\na conception-of-the-understanding[53] of the reciprocal sequence of\nthe determinations of these things coexisting externally to one\nanother, in order to say that the reciprocal succession of perceptions\nis grounded in the object, and thereby to represent the coexistence as\nobjective. But the relation of substances in which the one contains\ndeterminations the ground of which is contained in the other is the\nrelation of influence, and if, reciprocally, the former contains the\nground of the determinations in the latter, it is the relation of\ncommunity or interaction. Consequently, the coexistence of substances\nin space cannot be known in experience otherwise than under the\npresupposition of their interaction; this is therefore also the\ncondition of the possibility of things themselves as objects of\nexperience.\"[54]\n\n    [51] _Anschauung._\n\n    [52] _Wahrnehmung._\n\n    [53] _Verstandesbegriff._\n\n    [54] B. 257-8, M. 156-7.\n\nThe proof begins, as we should expect, in a way parallel to that of\ncausality. Just as Kant had apparently argued that we learn that a\nsuccession of perceptions is the perception of a sequence when we find\nthe order of the perceptions to be irreversible, so he now definitely\nasserts that we learn that certain perceptions are the perceptions of\na coexistence of bodies in space when we find that the order of the\nperceptions is reversible, or, to use Kant's language, that there can\nbe a reciprocal sequence of the perceptions. This beginning, if read\nby itself, seems as though it should also be the end. There seems\nnothing more which need be said. Just as we should have expected Kant\nto have completed his account of the apprehension of a succession when\nhe pointed out that it is distinguished by the irreversibility of the\nperceptions, so here we should expect him to have said enough when he\npoints out that the earth and the moon are said to be coexistent\nbecause our perceptions of them can follow one another reciprocally.\n\nThe analogy, however, has in some way to be brought in, and to this\nthe rest of the proof is devoted. In order to consider how this is\ndone, we must first consider the nature of the analogy itself. Kant\nspeaks of 'a conception-of-the-understanding of the reciprocal\nsequence of the determinations of things which coexist externally to\none another'; and he says that 'that relation of substances in which\nthe one contains determinations, the ground of which is contained in\nthe other substance, is the relation of influence'. His meaning can be\nillustrated thus. Suppose two bodies, A, a lump of ice, and B, a fire,\nclose together, yet at such a distance that they can be observed in\nsuccession. Suppose that A passes through changes of temperature a_{1}\na_{2} a_{3} ... in certain times, the changes ending in states\n[alpha]_{1} [alpha]_{2} [alpha]_{3} ..., and that B passes through\nchanges of temperature b_{1} b_{2} b_{3} ... in the same times, the\nchanges ending in states [beta]_{1} [beta]_{2} [beta]_{3}. Suppose\nalso, as we must, that A and B interact, i. e. that A in passing\nthrough its changes conditions the changes through which B passes, and\ntherefore also the states in which B ends, and vice versa, so that\na_{2} and [alpha]_{2} will be the outcome not of a_{1} and [alpha]_{1}\nalone, but of a_{1} and [alpha]_{1}, and b_{1} and [beta]_{1} jointly.\nThen we can say (1) that A and B are in the relation of influence, and\nalso of interaction or reciprocal influence, in the sense that they\n_mutually_ (not alternately) determine one another's states. Again, if\nwe first perceive A in the state [alpha]_1 by a perception A_{1}, then\nB in the state [beta]_{2} by a perception B_{2}, then A in the state\n[alpha]_{3} by a perception A_{3} and so on, we can speak (2) of a\nreciprocal sequence of perceptions, in the sense of a sequence of\nperceptions in which alternately a perception of B follows a\nperception of A and a perception of A follows a perception of B; for\nfirst a perception of B, viz. B_{2}, follows a perception of A, viz.\nA_{1}, and then a perception of A, viz. A_{3}, follows a perception of\nB, viz. B_{2}. We can also speak (3) of a reciprocal sequence of the\ndeterminations of two things in the sense of a necessary succession of\nstates which _alternately_ are states of A and of B; for [alpha]_{1},\nwhich is perceived first, can be said to contribute to determine\n[beta]_{2}, which is perceived next, and [beta]_{2} can be said to\ncontribute to determine [alpha]_{3}, which is perceived next, and so\non; and this reciprocal sequence can be said to be involved in the\nvery nature of interaction. Further, it can be said (4) that if we\nperceive A and B alternately, and so only in the states [alpha]_{1}\n[alpha]_{3} ... [beta]_{2} [beta]_{4} ... respectively, we can\nonly fill in the blanks, i. e. discover the states [alpha]_{2}\n[alpha]_{4} ... [beta]_{1} [beta]_{3} ... _coexistent_ with [beta]_{2}\n[beta]_{4} ... and [alpha]_{1} [alpha]_{3} ... respectively, if we\npresuppose the thought of interaction. For it is only possible to use\nthe observed states as a clue to the unobserved states, if we\npresuppose that the observed states are members of a necessary\nsuccession of which the unobserved states are also members and\ntherefore have partially determined and been determined by the\nobserved states. Hence it may be said that the determination of the\nunobserved states coexistent with the observed states presupposes the\nthought of interaction.\n\nHow then does Kant advance from the assertion that the apprehension of\na coexistence requires the knowledge that our _perceptions_ can be\nreciprocally sequent to the assertion that it presupposes the thought\nthat the _determinations of phenomena_ are reciprocally sequent? The\npassage in which the transition is effected is obscure and confused,\nbut it is capable of interpretation as soon as we see that it is\nintended to run parallel to the proof of the second analogy which is\nadded in the second edition.[55] Kant apparently puts to himself the\nquestion, 'How are we to know when we have a reciprocal sequence of\nperceptions from which we can infer a coexistence in what we\nperceived?' and apparently answers it thus: 'Since we cannot perceive\ntime, and therefore cannot perceive objects as dated in time with\nrespect to one another, we cannot begin with the apprehension of the\ncoexistence of two objects, and thence infer the possibility of\nreciprocal sequence in our perceptions. This being so, the synthesis\nof imagination in apprehension can indeed combine these perceptions\n[these now being really considered as determinations or states of an\nobject perceived] in a reciprocal sequence, but there is so far no\nguarantee that the sequence produced by the synthesis is not an\narbitrary product of the imagination, and therefore we cannot think of\nit as a reciprocal sequence in objects. In order to think of such a\nreciprocal sequence as not arbitrary but as constituting a real\nsequence in objects [ = 'as grounded in the object'], we must think of\nthe states reciprocally sequent [as necessarily related and therefore]\nas successive states of two coexisting substances which interact or\nmutually determine one another's successive states. Only then shall we\nbe able to think of the coexistence of objects involved in the\nreciprocal sequence as an objective fact, and not merely as an\narbitrary product of the imagination.' But, if this fairly expresses\nKant's meaning, his argument is clearly vitiated by two confusions. In\nthe first place, it confuses a subjective sequence of perceptions\nwhich are alternately perceptions of A and of B, two bodies in space,\nwith an objective sequence of perceived states of bodies, [alpha]_{1}\n[beta]_{2} [alpha]_{3} [beta]_{4}, which are alternately states of two\nbodies A and B, the same thing being regarded at once as a perception\nand as a state of a physical object. In the second place, mainly in\nconsequence of the first confusion, it confuses the necessity that the\nperceptions of A and of B can follow one another alternately with the\nnecessity of succession in the alternately perceived states of A and B\nas interacting. Moreover, there is really a change in the cases under\nconsideration. The case with which he begins, i. e. when he is\nconsidering merely the reciprocal sequence of perceptions, is the\nsuccessive perceptions of two _bodies in space_ alternately, e. g. of\nthe moon and the earth, the nature of their states at the time of\nperception not being in question. But the case with which he ends is\nthe successive perception of the _states of two bodies_ alternately,\ne. g. of the states of the fire and of the lump of ice. Moreover, it\nis only in the latter case that the objective relation apprehended is\nthat of coexistence in the proper sense, and in the sense which Kant\nintends throughout, viz. that of being contemporaneous in distinction\nfrom being successive. For when we say that two bodies, e. g. the moon\nand the earth, coexist, we should only mean that both exist, and not,\nas Kant means, that they are contemporaneous. For to a substance,\nbeing as it is the substratum of changes, we can ascribe no temporal\npredicates. That which changes cannot be said either to begin, or to\nend, or to exist at a certain moment of time, or, therefore, to exist\ncontemporaneously with, or after, or before anything else; it cannot\neven be said to persist through a portion of time or, to use the\nphrase of the first analogy, to be permanent. It will be objected\nthat, though the cases are different, yet the transition from the one\nto the other is justified, for it is precisely Kant's point that the\nexistence together of two substances in space can only be discovered\nby consideration of their successive states under the presupposition\nthat they mutually determine one another's states. \"Besides the mere\nfact of existence there must be something by which A determines the\nplace in time for B, and conversely B the place for A, because only\nunder this condition can these substances be empirically represented\nas coexistent.\"[56] The objection, however, should be met by two\nconsiderations, each of which is of some intrinsic importance. In the\nfirst place, the apprehension of a body in space in itself involves\nthe apprehension that it exists together with all other bodies in\nspace, for the apprehension of something as spatial involves the\napprehension of it as spatially related to, and therefore as existing\ntogether with, everything else which is spatial. No process,\ntherefore, such as Kant describes is required in order that we may\nlearn that it exists along with some other body. In the second place,\nthat for which the principle of interaction is really required is not,\nas Kant supposes, the determination of the coexistence of an\nunperceived body with a perceived body, but the determination of that\nunperceived state of a body already known to exist which is coexistent\nwith a perceived state of a perceived body. As has been pointed out,\nif we perceive A and B alternately in the states [alpha]_{1}\n[beta]_{2} [alpha]_{3} [beta]_{4} ... we need the thought of\ninteraction to determine the nature of [beta]_{1} [alpha]_{2}\n[beta]_{3} [alpha]_{4} ... Thus it appears that Kant in his\nvindication of the third analogy omits altogether to notice the one\nprocess which really presupposes it.\n\n    [55] B. 233-4, M. 142.\n\n    [56] B. 259, M. 157.\n\n\n\n\nCHAPTER XIII\n\nTHE POSTULATES OF EMPIRICAL THOUGHT\n\n\nThe postulates of empirical thought, which correspond to the\ncategories of modality, are stated as follows:\n\n\"1. That which agrees with the formal conditions of experience\n(according to perception and conceptions) is _possible_.\n\n2. That which is connected with the material conditions of experience\n(sensation) is _actual_.\n\n3. That of which the connexion with the actual is determined according\nto universal conditions of experience is _necessary_ (exists\nnecessarily).\"[1]\n\n    [1] B. 265-6, M. 161.\n\nThese principles, described as only 'explanations of the conceptions\nof possibility, actuality, and necessity as employed in experience',\nare really treated as principles by which we decide what is possible,\nwhat is actual, and what is necessary. The three conceptions involved\ndo not, according to Kant, enlarge our knowledge of the nature of\nobjects, but only 'express their relation to the faculty of\nknowledge'[2]; i. e. they only concern our ability to apprehend an\nobject whose nature is already determined for us otherwise as at least\npossible, or as real, or as even necessary. Moreover, it is because\nthese principles do not enlarge our knowledge of the nature of objects\nthat they are called postulates; for a postulate in geometry, from\nwhich science the term is borrowed (e. g. that it is possible with a\ngiven line to describe a circle from a given point), does not augment\nthe conception of the figure to which it relates, but only asserts the\npossibility of the conception itself.[3] The discussion of these\nprinciples is described, contrary to the terminology adopted in the\ncase of the preceding principles, as 'explanation' and not as 'proof'.\nThe discussion, however, certainly includes a proof of them, for it is\nKant's main object to _prove_ that these principles constitute the\ngeneral character of what can be asserted to be possible, actual, or\nnecessary respectively. Again, as before, the basis of proof lies in a\ntheory of knowledge, and in particular in Kant's theory of knowledge;\nfor it consists in the principle that everything knowable must conform\nto the conditions involved in its being an object of possible\nexperience.\n\n    [2] B. 266, M. 161. Cf. B. 286-7, M. 173-4.\n\n    [3] B. 286-7, M. 173-4.\n\nTo understand these principles and the proof of them, we must notice\ncertain preliminary considerations. In the _first_ place, the very\nproblem of distinguishing the possible, the actual, and the necessary\npresupposes the existence of distinctions which may prove open to\nquestion. It presupposes that something may be possible without being\nactual, and again that something may be actual without being\nnecessary. In the _second_ place, Kant's mode of approaching the\nproblem assumes that we can begin with a conception of an object,\ne. g. of a man with six toes, and then ask whether the object of it is\npossible, whether, if possible, it is also actual, and whether, if\nactual, it is also necessary. In other words, it assumes the\npossibility of separating what is conceived from what is possible, and\ntherefore _a fortiori_ from what is actual,[4] and from what is\nnecessary. _Thirdly_, in this context, as in most others, Kant in\nspeaking of a conception is thinking, to use Locke's phraseology, not\nof a 'simple' conception, such as that of equality or of redness, but\nof a 'complex' conception, such as that of a centaur, or of a triangle\nin the sense of a three-sided three-angled figure. It is the\napprehension of a 'complex' of elements.[5] _Fourthly_, what is said\nto be possible, real, or necessary is not the conception but the\ncorresponding object. The question is not, for instance, whether the\nconception of a triangle or of a centaur is possible, actual, or\nnecessary, but whether a triangle or a centaur is possible, actual, or\nnecessary. Kant sometimes speaks loosely of conceptions as\npossible,[6] but the terms which he normally and, from the point of\nview of his theory, rightly applies to conceptions are 'objectively\nreal' and 'fictitious'.[7] _Lastly_, Kant distinguishes 'objectively\nreal' and 'fictitious' conceptions in two ways. He speaks of\nestablishing the objective reality of a conception as consisting in\nestablishing the possibility of a corresponding object,[8] implying\ntherefore that a fictitious conception is a conception of which the\ncorresponding object is not known to be possible. Again, he describes\nas fictitious new conceptions of substances, powers, and interactions,\nwhich we might form from the material offered to us by perception\nwithout borrowing from experience itself the example of their\nconnexions, e. g. the conception of a power of the mind to perceive\nthe future; and he says that the possibility of these conceptions\n(i. e. the possibility of corresponding objects) cannot, like that of\nthe categories, be acquired _a priori_ through their being conditions\non which all experience depends, but must be discovered empirically or\nnot at all. Of such conceptions he says that, without being based upon\nexperience and its known laws, they are arbitrary syntheses which,\nalthough they contain no contradiction, have no claim to objective\nreality, and therefore to the possibility of corresponding objects.[9]\nHe implies, therefore, that the object of a conception can be said to\nbe possible only when the conception is the apprehension of a complex\nof elements together with the apprehension--which, if not _a priori_,\nmust be based upon experience--that they are connected. Hence a\nconception may be regarded as 'objectively real', or as 'fictitious'\naccording as it is the apprehension of a complex of elements\naccompanied by the apprehension that they are connected, or the\napprehension of a complex of elements not so accompanied.\n\n    [4] The view that 'in the mere conception of a thing no sign\n    of its existence is to be found' (B. 272, M. 165) forms, of\n    course, the basis of Kant's criticism of the ontological\n    argument for the existence of God. Cf. _Dialectic_, Bk. II,\n    Ch. III, Sec. 4.\n\n    [5] Cf. 'a conception which includes in itself a synthesis'\n    (B. 267 med., M. 162 med.).\n\n    [6] E. g. B. 269 fin., M. 163 fin.; B. 270 med., M. 164 init.\n    The formulation which really expresses Kant's thought is to\n    be found B. 266 med., M. 161 fin.; B. 268 init., M. 162 fin.;\n    B. 268 med., M. 163 init.; and B. 270 med., M. 164 init.\n\n    [7] _Gedichtete._\n\n    [8] B. 268 init., M. 162 fin.\n\n    [9] B. 269-70, M. 163-4.\n\nIt is now possible to state Kant's problem more precisely. With regard\nto a given complex conception he wishes to determine the way in which\nwe can answer the questions (1) 'Has the conception a possible object\nto correspond to it', or, in other words, 'Is the conception\n'objectively real' or 'fictitious'?' (2) 'Given that a corresponding\nobject is possible, is it also real?' (3) 'Given that it is real, is\nit also necessary?'\n\nThe substance of Kant's answer to this problem may be stated thus:\n'The most obvious guarantee of the objective reality of a conception,\ni. e. of the possibility of a corresponding object, is the experience\nof such an object. For instance, our experience of water guarantees\nthe objective reality of the conception of a liquid which expands as\nit solidifies. This appeal to experience, however, takes us beyond the\npossibility of the object to its reality, for the experience\nvindicates the possibility of the object only through its reality.\nMoreover, here the basis of our assertion of possibility is only\nempirical, whereas our aim is to discover the conceptions of which the\nobjects can be determined _a priori_ to be possible. What then is the\nanswer to this, the real problem? To take the case of cause and\neffect, we cannot reach any conclusion by the mere study of the\nconception of cause and effect. For although the conception of a\nnecessary succession contains no contradiction, the necessary\nsuccession of events is a mere arbitrary synthesis as far as our\nthought of it is concerned; we have no direct insight into the\nnecessity. Therefore we cannot argue from this conception to the\npossibility of a corresponding object, viz. a necessarily successive\nseries of events in nature. We can, however, say that that synthesis\nis not arbitrary but necessary to which any object must conform, if it\nis to be an object of experience. From this point of view we can say\nthat there must be a possible object corresponding to the conception\nof cause and effect, because only as subjected to this synthesis are\nthere objects of experience at all. Hence, if we take this point of\nview, we can say generally that all spatial and temporal conceptions,\nas constituting the conditions of perceiving in experience, and all\nthe categories, as constituting the conditions of conceiving in\nexperience, must have possible objects. In other words, 'that which\nagrees with the formal conditions of experience (according to\nperception and conceptions) is _possible_'. Again, if we know that\nthe object of a conception is possible, how are we to determine\nwhether it is also actual? It is clear that, since we cannot advance\nfrom the mere conception, objectively real though it may be, to the\nreality of the corresponding object, we need perception. The case,\nhowever, where the corresponding object is directly perceived may be\nignored, for it involves no inference or process of thought; the\nappeal is to experience alone. Therefore the question to be considered\nis, 'How do we determine the actuality of the object of a conception\ncomparatively _a priori_, i. e. without direct experience of it[10]?'\nThe answer must be that we do so by finding it to be 'connected\nwith an actual perception in accordance with the analogies of\nexperience'[11]. For instance, we must establish the actuality of an\nobject corresponding to the conception of a volcanic eruption by\nshowing it to be involved, in accordance with the analogies (and with\nparticular empirical laws), in the state of a place which we are now\nperceiving. In other words, we can say that 'that which is connected\nwith the material conditions of existence (sensation) is _actual_'.\nFinally, since we cannot learn the existence of any object of\nexperience wholly _a priori_, but only relatively to another existence\nalready given, the necessity of the existence of an object can never\nbe known from conceptions, but only from its connexion with what is\nperceived; this necessity, however, is not the necessity of the\nexistence of a substance, but only the necessity of connexion of an\nunobserved state of a substance with some observed state of a\nsubstance. Therefore we can (and indeed must) say of an unobserved\nobject corresponding to a conception, not only that it is real, but\nalso that it is necessary, when we know it to be connected with a\nperceived reality 'according to universal conditions of experience';\nbut the necessity can be attributed only to states of substances and\nnot to substances themselves.'\n\n    [10] Cf. B. 279, M. 169 and p. 4, note 1.\n\n    [11] B. 273, M. 165.\n\nThroughout this account there runs one fatal mistake, that of\nsupposing that we can separate our knowledge of things as possible, as\nactual, and as necessary. Even if this supposition be tenable in\ncertain cases,[12] it is not tenable in respect of the objects of a\ncomplex conception, with which Kant is dealing. If we know the object\nof a complex conception to be possible, we already know it to be\nactual, and if we know it to be actual, we already know it to be\nnecessary. A complex conception in the proper sense is the\napprehension of a complex of elements together with the apprehension\nof, or insight into, their connexion.[13] Thus, in the case of the\nconception of a triangle we see that the possession of three sides\nnecessitates the possession of three angles. From such a conception\nmust be distinguished Kant's 'fictitious' conception, i. e. the\napprehension of a complex of elements without the apprehension of\nconnexion between them. Thus, in the case of the conception of a man\nwith six toes, there is no apprehension of connexion between the\npossession of the characteristics indicated by the term 'man' and the\npossession of six toes. In such a case, since we do not apprehend any\nconnexion between the elements, we do not really 'conceive' or 'think'\nthe object in question, e. g. a man with six toes. Now in the case of\na complex conception proper, it is impossible to think of a\ncorresponding individual as only possible. The question 'Is a\ntriangle, in the sense of a figure with three sides and three angles,\npossible?' really means 'Is it possible for a three-sided figure to\nhave three angles?' To this question we can only answer that we see\nthat a three-sided figure can have three angles, because we see that\nit must have, and therefore has, and can have, three angles; in other\nwords, that we see a triangle in the sense in question to be possible,\nbecause we see it to be necessary, and, therefore, actual, and\npossible. It cannot be argued that our insight is limited to the fact\nthat if there are three-sided figures they must be three-angled, and\nthat therefore we only know a triangle in the sense in question to be\npossible. Our apprehension of the fact that the possession of three\nsides necessitates the possession of three angles presupposes\nknowledge of the existence of three-sided figures, for it is only in\nan actual three-sided figure that we can apprehend the necessity. It\nmay, however, be objected that the question ought to mean simply 'Is a\nthree-sided figure possible?' and that, understood in this sense, it\ncannot be answered in a similar way. Nevertheless, a similar answer is\nthe right answer. For the question 'Is a three-sided figure possible?'\nreally means 'Is it possible for three straight lines to form a\nfigure, i. e. to enclose a space?' and we can only answer it for\nourselves by seeing that a group of three straight lines or\ndirections, no two of which are parallel, must, as such, enclose a\nspace, this insight presupposing the apprehension of an actual group\nof three straight lines. It may be said, therefore, that we can only\ndetermine the possibility of the object of a complex conception in\nthe proper sense, through an act in which we apprehend its necessity\nand its actuality at once. It is only where conceptions are\n'fictitious', and so not properly conceptions, that appeal to\nexperience is necessary. The question 'Is an object corresponding to\nthe conception of a man with six toes possible?' presupposes the\nreality of man and asks whether any man can have six toes. If we\nunderstood the nature of man and could thereby apprehend either that\nthe possession of six toes was, or that it was not, involved in one of\nthe possible differentiations of man, we could decide the question of\npossibility _a priori_, i. e. through our conceiving alone without an\nappeal to experience; but we could do so only because we apprehended\neither that a certain kind of man with six toes was necessary and\nactual, or that such a man was impossible and not actual. If, however,\nas is the case, we do not understand the nature of man, we can only\ndecide the question of possibility by an appeal to experience, i. e.\nto the experience of a corresponding object, or of an object from\nwhich the existence of such an object could be inferred. Here,\ntherefore--assuming the required experience to be forthcoming--we\ncan appeal to Kant's formula and say that we know that such a man,\ni. e. an object corresponding to the conception, is actual, as\nbeing connected with the material conditions of experience. But the\nperception which constitutes the material conditions of experience\nin the case in question is only of use because it carries us beyond\npossibility to actuality, and appeal to it is only necessary because\nthe object is not really conceived or, in other words, because the\nso-called conception is not really a conception.\n\n    [12] For instance, it might at least be _argued_ that we know\n    space to be actual without knowing it to be necessary.\n\n    [13] _Not_ 'together with the apprehension _that_ the\n    elements are connected'. Cf. p. 311.\n\nKant really treats his 'objectively real' conceptions as if they were\n'fictitious', even though he speaks of them as complete. Consequently,\nhis conceptions not being conceptions proper, he is necessarily led to\nhold that an appeal to experience is needed in order to establish the\nreality of a corresponding object. Yet, this being so, he should have\nasked himself whether, without an appeal to perception, we could even\nsay that a corresponding object was possible. That he did not ask this\nquestion is partly due to the fact that he attributes the form and the\nmatter of knowledge to different sources, viz. to the mind and to\nthings in themselves. While the conceptions involved in the forms of\nperception, space, and time, and also the categories are the\nmanifestations of the mind's own nature, sensations, which form the\nmatter of knowledge, are due to the action of things in themselves on\nour sensibility, and of this activity we can say nothing. Hence, from\nthe point of view of our mind--and since we do not know things in\nthemselves, this is the only point of view we can take--the existence\nof sensations, and therefore of objects, which must be given in\nperception, is wholly contingent and only to be discovered through\nexperience. On the other hand, since the forms of perception and\nconception necessarily determine in certain ways the nature of\nobjects, _if_ there prove to be any objects, the conceptions involved\nmay be thought to determine what objects are possible, even though the\nvery existence of the objects is uncertain. Nevertheless, on his own\nprinciples, Kant should have allowed that, apart from perception, we\ncould discover _a priori_ at least the reality, even if not the\nnecessity, of the objects of these conceptions. For his general view\nis that the forms of perception and the categories are only actualized\non the occasion of the stimulus afforded by the action of things in\nthemselves on the sensibility. Hence the fact that the categories and\nforms of perception are actualized--a fact implied in the very\nexistence of the _Critique_--involves the existence of objects\ncorresponding to the categories and to the conceptions involved in the\nforms of perception. On Kant's own principles, therefore, we could say\n_a priori_ that there must be objects corresponding to these\nconceptions, even though their nature in detail could only be filled\nin by experience.[14]\n\n    [14] Cf. Caird, i. 604-5.\n\n\n\n\nNOTE ON THE REFUTATION OF IDEALISM\n\n\nThis well-known passage[1] practically replaces a long section,[2]\ncontained only in the first edition, on the fourth paralogism of\npure reason. Its aim is to vindicate against 'idealism' the reality\nof objects in space, and it is for this reason inserted after the\ndiscussion of the second postulate. The interest which it has excited\nis due to Kant's use of language which at least seems to imply that\nbodies in space are things in themselves, and therefore that here he\nreally abandons his main thesis.\n\n    [1] B. 274-9, M. 167-9. Cf. B. xxxix (note), M. xl (note).\n\n    [2] A. 367-80, Mah. 241-53.\n\nIdealism is the general name which Kant gives to any view which\nquestions or denies the reality of the physical world; and, as has\nbeen pointed out before,[3] he repeatedly tries to defend himself\nagainst the charge of being an idealist in this general sense. This\npassage is the expression of his final attempt. Kant begins by\ndistinguishing two forms which idealism can take according as it\nregards the existence of objects in space as false and _impossible_,\nor as doubtful and _indemonstrable_. His own view, which regards their\nexistence as certain and demonstrable, and which he elsewhere[4]\ncalls transcendental idealism, constitutes a third form. The first\nform is the dogmatic idealism of Berkeley. This view, Kant says,\nis unavoidable, if space be regarded as a property of things in\nthemselves, and the basis of it has been destroyed in the _Aesthetic_.\nThe second form is the problematic idealism of Descartes, according to\nwhich we are immediately aware only of our own existence, and belief\nin the existence of bodies in space can be only an inference, and an\nuncertain inference, from the immediate apprehension of our own\nexistence. This view, according to Kant, is the outcome of a\nphilosophical attitude of mind, in that it demands that a belief\nshould be proved, and apparently--to judge from what Kant says of\nBerkeley--it does not commit Descartes to the view that bodies in\nspace, if their reality can be vindicated, are things in themselves.\n\n    [3] Cf. p. 76.\n\n    [4] A. 369, Mah. 243; cf. B. 44, M. 27.\n\nThe assertion that the _Aesthetic_ has destroyed the basis of\nBerkeley's view, taken together with the drift of the _Refutation_\nas a whole, and especially of Remark I, renders it clear that the\n_Refutation_ is directed against Descartes and not Berkeley. Kant\nregards himself as having already refuted Berkeley's view, as he\nhere states it, viz. that the existence of objects in space is\n_impossible_, on the ground that it arose from the mistake of\nsupposing that space, if real at all, must be a property of things in\nthemselves, whereas the _Aesthetic_ has as he thinks, shown that space\ncan be, and in point of fact is, a property of phenomena. He now wants\nto prove--compatibly with their character as phenomena--that the\nexistence of bodies in space is not even, as Descartes contends,\n_doubtful_. To prove this he seeks to show that Descartes is wrong in\nsupposing that we have no immediate experience of these objects. His\nmethod is to argue that reflection shows that internal experience\npresupposes external experience, i. e. that unless we were directly\naware of spatial objects, we could not be aware of the succession of\nour own states, and consequently that it is an inversion to hold that\nwe must reach the knowledge of objects in space, if at all, by an\ninference from the immediate apprehension of our own states.\n\nAn examination of the proof itself, however, forces us to allow that\nKant, without realizing what he is doing, really abandons the view\nthat objects in space are phenomena, and uses an argument the very\nnature of which implies that these objects are things in themselves.\nThe proof runs thus:\n\n_Theorem._ \"The mere but empirically determined consciousness of my\nown existence proves the existence of objects in space external to\nme.\"\n\n\"_Proof._ I am conscious of my own existence as determined in\ntime. All time-determination presupposes something permanent in\nperception.[5] This permanent, however, cannot be an intuition[6]\nin me. For all grounds of determination of my own existence, which\ncan be found in me, are representations, and as such themselves need\na permanent different from them, in relation to which their change\nand consequently my existence in the time in which they change can\nbe determined.[7] The perception of this permanent, therefore, is\npossible only through a _thing_ external to me, and not through the\nmere _representation_ of a thing external to me. Consequently, the\ndetermination of my existence in time is possible only through the\nexistence of actual things, which I perceive external to me. Now\nconsciousness in time is necessarily connected with the consciousness\nof the possibility of this time-determination; hence it is necessarily\nconnected also with the existence of things external to me, as the\ncondition of time-determination, i. e. the consciousness of my own\nexistence, is at the same time an immediate consciousness of the\nexistence of other things external to me.\"[8]\n\n    [5] _Wahrnehmung._\n\n    [6] _Anschauung._\n\n    [7] The text has been corrected in accordance with Kant's\n    note in the preface to the second edition, B. xxxix, M. xl.\n\n    [8] B. 275-6, M. 167.\n\nThe nature of the argument is clear. 'In order to be conscious, as\nI am, of a determinate succession of my states, I must perceive\nsomething permanent as that in relation to which alone I can perceive\nmy states as having a definite order.[9] But this permanent cannot be\na perception in me, for in that case it would only be a representation\nof mine, which, as such, could only be apprehended in relation to\nanother permanent. Consequently, this permanent must be a thing\nexternal to me and not a representation of a thing external to me.\nConsequently, the consciousness of my own existence, which is\nnecessarily a consciousness of my successive states, involves the\nimmediate consciousness of things external to me.'\n\n    [9] Cf. Kant's proof of the first analogy.\n\nHere there is no way of avoiding the conclusion that Kant is deceived\nby the ambiguity of the phrase 'a thing external to me' into thinking\nthat he has given a proof of the existence of bodies in space which is\ncompatible with the view that they are only phenomena, although in\nreality the proof presupposes that they are things in themselves. In\nthe 'proof', the phrase 'a thing external to me' must have a double\nmeaning. It must mean a thing external to my body, i. e. any body\nwhich is not my body; in other words, it must be a loose expression\nfor a body in space. For, though the 'proof' makes us appeal to the\nspatial character of things external to me, the _Refutation_ as a\nwhole, and especially Remark II, shows that it is of bodies in space\nthat he is thinking throughout. The phrase must also, and primarily,\nmean a thing external to, in the sense of independent of, my mind,\ni. e. a thing in itself. For the nerve of the argument consists in the\ncontention that the permanent the perception of which is required for\nthe consciousness of my successive states must be a _thing_ external\nto me in opposition to the representation of a thing external to me,\nand a thing external to me in opposition to a thing external to me can\nonly be a thing in itself. On the other hand, in Kant's conclusion,\n'a thing external to me' can only mean a body in space, this being\nsupposed to be a phenomenon; for his aim is to establish the reality\nof bodies in space compatibly with his general view that they are only\nphenomena. The proof therefore requires that things external to me, in\norder that they may render possible the consciousness of my successive\nstates, should have the very character which is withheld from them in\nthe conclusion, viz. that of existing independently of me; in other\nwords, if Kant establishes the existence of bodies in space at all,\nhe does so only at the cost of allowing that they are things in\nthemselves.[10]\n\n    [10] The ambiguity of the phrase 'external to me' is pointed\n    out in the suppressed account of the fourth paralogism,\n    where it is expressly declared that objects in space are\n    only representations. (A. 372-3, Mah. 247). Possibly the\n    introduction of an argument which turns on the view that they\n    are not representations may have had something to do with the\n    suppression.\n\nNevertheless, the _Refutation_ may be considered to suggest the proper\nrefutation of Descartes. It is possible to ignore Kant's demand for a\npermanent as a condition of the apprehension of our successive states,\nand to confine attention to his remark that he has shown that external\nexperience is really immediate, and that only by means of it is the\nconsciousness of our existence as determined in time possible.[11] If\nwe do so, we may consider the _Refutation_ as suggesting the view that\nDescartes' position is precisely an inversion of the truth; in other\nwords, that our consciousness of the world, so far from being an\nuncertain inference from the consciousness of our successive states,\nis in reality a presupposition of the latter consciousness, in that\nthis latter consciousness only arises through reflection upon the\nformer, and that therefore Descartes' admission of the validity of\nself-consciousness implicitly involves the admission _a fortiori_ of\nthe validity of our consciousness of the world.[12]\n\n    [11] B. 277, M. 167 fin.\n\n    [12] Cf. Caird, i. 632 and ff.\n\n\n      Oxford: Printed at the Clarendon Press by HORACE HART, M.A.","title":"Kant's Theory of Knowledge","tradition":"philosophy-ethics","translated_text":null,"word_count":98995}