Kant's theory of knowledge;: An outline of one central argument in the Critique of pure reason

Citation preview

International Libray of

KANT'S THEORY OF KNOWLEDGE An Outline of One Central Argument in the

EDITOR: A. J. AYER ASSISTANT EDITOR: BE.RNARD WILLIAMS

lN3UCTIVE PROBABILITY by John Patrick Day. SENSATION AND PERCEPTION: A History of the Philosophy of Perception by D. W. Harnlyn. '"'RACTATUS LOGICO-PHILOSOPHICUS: Ludwig Wittgenstein's Logiscbph~loropbi~che Abhandlug with a new Translation by D. F. Pears and B. F. McGuimess and with the Introduction by Bertrand Russell. PERCEPTION AND T H E PHYSICAL WORLD by D. M. Armstrong. I3UME'S PHILOSOPHY O F BELIEF: A Study of His First Ingriir_yby Antony Flew. KANT'S THEORY O F KNOWLEDGE: An Outline of One Central Argument in the Critiqtte of Pure Reason by Graham Bird. CONDITIONS FOR DESCRIPTION by Peter Zinkernagel, translated from the Danish by Olaf Lindum. AN EXAMINATION O F PLATO'S DOCTRINES by I. M. Crombie. Two volumes. I: Plato on Man and Society. 11: Plato on Knowledge and Reality. PFENOMENOLOGY O F PERCEPTION by M. Merleau-Ponty, translated from the French by Colin Smith. THE IDEA O F JUSTICE AND T H E PROBLEM O F ARGUMENT by Ch. Perelman, translated from the French by John Petrie. 1,EC'I'URES O N PSYCHICAL RESEARCH by C. D. Broad. Incorporating the Perrott Lectures given in Cambridge University in 1959and 1960. T H E VARIETIES O F GOODNESS by Georg Heniik von Wright. METHOD I N T H E PHYSICAL SCIENCES by G. Schlesinger. METHOD IN ETHICAL THEORY by Abraham Edel. SCIENCE, PERCEPTION AND REALITY by Wilfrid Sellars. NORM AND ACTION: A Logical Enquiry by Georg Henrik von Wright. PHILOSOPHY AND SCIENTIFIC REALISM by J, J. C. Smart. STUDIES I N METAPIIILOSOPHY by Morris Lazerowitz. REASON, ACTION AND MORALITY by J. Kemp. T H E HIDDEN GOD: A Study of Tragic Vision in the PensCesof Pascal and the Tragedies of Racine by Lucien Goldmann, translated from the French by Philip Thody. THE EXPLANATION OF BEHAVIOUR by Charles Taylor.

Critiqtle of Ptlre Reason

by

Graham Bird Lecturer in Logic UItiverJity of Aber&en

LONDON

ROUTLEDGE & KEGAN PAUL NEW YORK

: THE H U M A N I T I E S P R E S S

CONTENTS page i x

PREFACE

I PHENOhiENA AND PHENOMENALIShl

(i) Prichard's theory (ii) Literal construction (iii) The first edition Deduction and phenomenalism 2 NOUMENA A N D NOUMENALIShf

3

TRANSCENDENTAL AND ESIPIRICAL

(i) The distinction in general (ii) The Fourth Paralogism (A) (iii) The Aesthetic 4

SENSIBILITY AND UNDERSTANDING

(i) Data of sense and phenomenal objects (ii) The 'representation' problem for categories I

I N T E L L I G I B L E OBJECTS

(i) The occasions of misuse (ii) Concepts and objects (iii) The 'object of representations'

7

CATEGORIES AND JUDGMENTS

(i) Analysis and synthesis (ii) Concepts and judgments 8

THE TRANSCENDENTAL DEDUCTION

(i) Introduction (ii) Apperception (iii) Apperception, categories, and objective validity

vii

CONTENTS

9

D E D U C T I O N (CONTINUED) page 126 (i) Language and experience 126 (ii) Categories and personality 136 (iii) Transcendental distinctions and empirical illustrations 140

THE TRANSCENDENTAL

PREFACE

I 0 THE FIRST AND SECOND ANALOGIES

(i) The Analogies I1 P E R S O N A L I T Y

(i) Inner sense and apperception (ii) Categories and persons (iii) Kant and Strawson I2 THE TRANSITION TO MORAL PHILOSOPHY

(i) Causality and freedom (ii) Kant's solution of the Third Antinomy LIST O F BOOKS

2of

INDEX

207

viii

THIS book is intended as a limited account of one central argument in the Critique of Pure Reason. I t is not an elementary or introductory account of the Critique, and presupposes some acquaintance with the main features of Kant's theory of knowledge. I t is not simply a commentary on Kant's text, but an attempt to pursue one continuous argument through the Critique. That this argument extends from the Prefaces and Introduction, through the Aesthetic and Analytic, to the Paralogisms and the Third Antinomy suggests that it forms a central part of Kant's doctrines. But there is much else in the Critique which has been left out here. In particular nothing has been said specifically of what we should call Kant's philosophy of science. The Critiqge of Pure Reason has had some influence on recent British philosophy. Significant references to Kant have been made in several recent books (for example, Hampshire: Tholrght and Action, Strawson : Individt/dIs, Stenius : Wittgenstein's Tractatr-), in which Kant's views are either re-stated in modern terms, or compared with the views of modern philosophers, like Wittgenstein (Cf. Strawson: op. cit., Ch. 3, Stenius: op. cit., Ch. XI). It is true that Kant, perhaps more than other traditional philosophers, invites such interpretation and comparison, and there are good historical reasons for this. In this book an attempt is made to provide a basis for such interpretations by discussing Kant's arguments in terms of a modern philosophical idiom. But since much of it is concerned to explain what Kant said, the philosophic issues cannot be exhaustively discussed, and are only introduced through Kant's attitudes towards them. No reasonably comprehensive account of a traditional philosopher could do more without assuming from the start what his views actually are; and in Kant7scase no such assumption can sensibly be made. This book will, therefore, appear defective for two opposed reasons. On the one hand it will not satisfy those who look for the ix

PREFACE

thoroughness of a commentary. On the other it will disappoint those who hope for exhaustive discussion of purely philosophical issues. It is very likely that any short account of the Critique, which does more than merely paraphrase the text, will fall somewhere between the two stools of history and philosophy. I have tried to fall nearer to that of philosophical interest. This recipe seems to me to be recommended by another book on Kant with a similar aim, namely Prichard's Kant's Theory of IGzowledge. Prichard provided in his book more genuine philosophical interest and argument than most conlmentaries on Kant; but he almost always misunderstood Kant's views. I have tried to give an account of the Critique which is as sympathetic to Kant as Prichard's was unsympdthetic to him. I would like to thank the non-philosophers who have helped and encouraged me. I owe also very much to the late T. D. Weldon, and to the kindness and help of Professor R. C. Cross and Dr. W. Bednarowski.

I

PHENOMENA A N D PHENOMENALISM '. . . which constrains us to regard all appearances (Erscheinungen) as data of the senses

. . .' (A 122).

THERE is no one passage in the Critique of Pgre Reason where Kant develops his account of perception. Such an account is nonetheless implicit in Kant's use of perceptual terms, and especially of his key term 'Erscheinung', which it is natural to translate either as 'appearance' or as 'phenomenon'. This lack of systematic development reflects important differences of aim and emphasis between Kant and modern philosophers of perception, which make it hard to attach any modem label to his account without adding some qualification. Yet both natural translations of the term 'Erscheinung' in English, as well as some of the ways in which Kant uses the term, strongly encourage rhe assumption that Kant must be a phenomenalist of some kind. Many philosophers, including Kemp Smith, Paton, Vleeschauwer, and Ayer, have indeed ascribed such a theory to Kant, although there is no guarantee that they have all understood this attribution in precisely the same way. For these reasons it is as well to be careful before classing Kant as an adherent of this, or any other, theory, however firmly such labels have stuck to him. In this chapter I want to indicate some difficulties in the way of at least some attributions of such a theory to Kant. It would not be sensible, in the space of one short chapter, to try to show that all such attributions are wholly wrong. It may nevertheless be useful to indicate defects in one such attribution, so that permission may be provisionally I

PRICHARD'S

PHENOMENA AND PHENOMENALISM

granted to produce some alternative interpretation. Prichard's account of Kant has been chosen as the most vigorous and clearest exposition of the view to be disputed.

I (i) PRICHARD'S THEORY Prichard's account of Kant's perceptual terms can be found throughout his book Kdnt's Theory of Knowledge, and is neatly summarised in a passage from pp. 2 3 1-23 j. In this passage Prichard argues (p. 231) that the central question of the Transcendental Deduction, which he identifies as 'How does an apprehension become related to an object?', contains an absurdity. He claims that Kant manages to conceal this absurdity from himself by supposing that representations, or appearances, or sensations, both 'have a being of their own' and also represent something, namely the thing in itself. This latter point is reinforced by Prichard elsewhere (p. 137) through the claim, specifically about appearances, that 'though from the point of view of the thing in itself an appearance is an appearance or perception of it, yet regarded from the standpoint of what it is in itself, an appearance is a reality perceived of the kind called mental'. Prichard's picture of Kant is thus one in which appearances, or phenomena, connect the real world of things in themselves with perceiving subjects, by representing the former and also belonging to the latter. Clearly on such a view something needs to be said about the relations of 'representing' and 'belonging to', and Prichard now reintroduces Kant's problem in this way. 'If a representation is taken to be an appearance, or sensation, the main problem becomes that of explaining how it is that, beginning with the apprehension of mere appearances or sensations, we come to apprehend an object, in the sense of an object in nature, which, as such, is not an appearance or sensation, but a part of the physical world' (p. 232). According to Prichard Kant's solution to this problem is simply to say that '. . appearances, or sensations, become related to an object, in the sense of an object in nature, by being combined on certain principles' (p. 232); or, as he later puts it, by the construction or making of 'parts of the physical world, and in fact the physical world itself, out of elements given in perception7 (pp. 23 3-234). Prichard explains, in a footnote, that the mathematical illustrations of this supposed 'construction' are

.

2

THEORY

the most plausible, for 'while we can be said to construct geometrical figures, and while the construction of geometrical figures can easily l e mistaken for the apprehension of them, we cannot with any plausibility be said to construct the physical world' (p. 234). This constitutes Prichard's basic objection to Kant's principal claim, that on such an account the construction envisaged has to be of the physical world itself (literalb (p. 2 3 3)), and, impossibly, out of materials of a mental kind, namely appearances,or sensations. It would be easy to say that Prichard's is a very unsubtle account of Kant, and to leave the way open for a more sympathetic interpretation of 'construction' along more sophisticated phenomenalist lines. Prichard does not say that Kant is a phenomenalist, but it is not hard to see in Prichard's exposition the outline of a crude phenomenalist theory. Prichard's main point is that Kant's term 'synthesis' refers to a construction of physical objects out of appearances, or sensations. A more up-to-date phenomenalist view might be expressed by saying that physical objects can be 'reduced' to sense-data. Where Kant, on Prichard's view, speaks of the construction of the external world out of our sensations, phenomenalists speak of the 'reduction' of things in the external world to sense-data. One obvious difference is that, on Prichard's view, Kant's construction is literally of objects from sensations, while contemporary phenomenalists prefer to speak instead of the reduction of physical object statements to sense-datum statements. These similarities and differences would allow one to say either that Kant was a phenomenalist of a primitive type, as Prichard implies, or else that he was simply a phenomenalist, on the ground that Prichard exaggerates the extent to which Kant was misled by talking of the construction of objects instead of the analysis of statements. Certainly it is true that Prichard exaggerates the psychological flavour of Kant's claims, though it is also true that some things which Kant says appear quite strongly to support Prichard's interpretation. Indeed if this were not so it would be hard to understand the tenacity with which such an interpretation survives. Nevertheless, examination of Prichard's central theme shows that his account is defective at almost every point. Prichard evidently believed that for Kant the physical world is one of things in themselves, as well as that this world is represented to us by appearances, or sensations, of these things. It is not worth spending much time on the first of these interpretations, 3

PHENOMENA A N D PHENOMENALISM

since it is indisputable that Kant persistently denies it throughout the Critique. One of his most important aims is to show that our knowledge of the physical world is not of things in themselves, but is restricted to what he variously calls objects of experience, sensible objects, or appearances. Even the knowledge which we have from Newtonian mechanics, to which Kant ascribes a status superior to that of our ordinary experience, is said to be restricted in this way (I3 3 I 3). A characteristic passage where Kant makes it clear that on his view all our theoretical knowledge is so restricted occurs in sections 2 2 and 23 of the second edition Deduction (B 146-1 jo), but the same point is made frequently elsewhere. Even in the first edition Deduction, where, as we shall see, support for Prichard's interpretation is at its strongest, Kant shows that his argument would be pointless if our knowledge was of things in themselves. At A 128 he says: 'If the objects with which our knowledge has to deal were things in themselves, we could have no a priori concepts of them.' The point of this assertion is to summarise Kant's argument that we can employ a priori concepts only because our knowledge is limited to phenomena, or appearances. The second claim, that appearances represent things in themselves, loses some of its plausibility as an interpretation of Kant once Prichard's first claim has been rejected in this way. It is, nevertheless, a more disputable claim than the first, for Kant certainly seems on occasion to say things which support it. Yet it is also true that some such passages which naturally seem to support this view do not in the end do so. Two such typical passages occur at B I 64 and A I 09. In the former Kant says : 'But appearances are only representations of things which are unknown as regards what they may be in themselves.' And in the latter: 'But these appearances are not things in themselves; they are only representations which in turn have their o b j e c t a n object which (therefore) cannot be itself intuited by us, and which may hence be called the non-empirical, i.e transcendental object = X.' In the first type of case it must be noted that Kant does not say that appearances are representations of things in themselves, which are nevertheless unknown to us. He says rather that appearances represent things, which are unknown to us as regards what they may be in themselves. These are clearly quite different claims, for the second is compatible with the statement that appearances do

not represent things in themselves, while the first is not. There are other passages where the same clistinction is easy to overlook (e.g. Prol., Sect. I 3, note 11; Prol., Sect. 36. Ali., Vol. 4, p. 289 and p. j I a). The first construction of this passaoe which is that required b.' for Prichard's interpretation, is paradoxical if not inconsistent. For it is hard to understand how we could be entitled to say that appearances represent something, and also that that thing is unknown to us. It would not be surprising that Kant could be so easily dismissed by Prichard if this were what he meant. It is certainly not what he here says. For what he says implies that while it is proper to speak of appearances' representing things in some way, it is not proper to speak of knowing things as they are in themselves. And this suggests that the things which it is proper to say that appearances represent are not things in themselves. Such a view is confirmed by the second passage. For there Kant claims that appearances might be said to represent not things in themselves, but the transcendental object. Nothing would have been easier for I