This page intentionally left blank
G. E. MOORE Early Philosophical Writings
G. E. Moore’s fame as a philosopher rest...
61 downloads
1231 Views
1MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
This page intentionally left blank
G. E. MOORE Early Philosophical Writings
G. E. Moore’s fame as a philosopher rests on his ethics of love and beauty, which inspired Bloomsbury, and on his ‘common sense’ certainties, which challenge abstract philosophical theory. Behind these themes lies his critical engagement with Kant’s idealist philosophy, which is published here for the first time. These early writings, Moore’s fellowship dissertations of 1897 and 1898, show how he initiated his influential break with idealism. In 1897 his main target was Kant’s ethics; but by 1898 it was the whole Kantian project of transcendental philosophy that he rejected, and the theory which he developed to replace it gave rise to the new project of philosophy as logical analysis. This edition includes comments by Moore’s examiners, Henry Sidgwick, Edward Caird and Bernard Bosanquet, and in a substantial introduction the editors explore the crucial importance of the dissertations to the history of twentieth-century philosophical thought. t h o m a s b a l d w i n is Professor of Philosophy at the University of York. His previous publications include G. E. Moore (1990), Contemporary Philosophy: Philosophy in English since 1945 (2001) and The Cambridge History of Philosophy 1870–1945 (Cambridge, 2003). He has also edited the revised edition of Principia Ethica (Cambridge, 1993) and a selection of Moore’s papers, Selected Writings (1993). c o n s u e l o pr e t i is Associate Professor of Philosophy at the College of New Jersey. She is the author of On Kripke (2003) and On Fodor (2000).
Frontispiece: First page of the introduction to Moore’s 1897 dissertation. Reproduced with permission from the Master and Fellows of Trinity College Cambridge.
G. E. MOORE Early Philosophical Writings
edited and with an introduction by THOMAS BALDWIN AND CONSUELO PRETI
cambridge university press Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, S˜ao Paulo, Delhi, Tokyo, Mexico City Cambridge University Press The Edinburgh Building, Cambridge cb2 8ru, UK Published in the United States of America by Cambridge University Press, New York www.cambridge.org Information on this title: www.cambridge.org/9780521190145 c Consuelo Preti and Thomas Baldwin 2011
This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2011 Printed in the United Kingdom at the University Press, Cambridge A catalogue record for this publication is available from the British Library Library of Congress Cataloguing in Publication data G. E. Moore : Early Philosophical Writings / edited and with an introduction by Thomas Baldwin and Consuelo Preti. p. cm. Includes index. isbn 978-0-521-19014-5 (hardback) 1. Moore, G. E. (George Edward), 1873–1958. I. Baldwin, Thomas, 1947– editor. II. Preti, Consuelo, editor. b1647.m74g22 2011 192 – dc22 2010048129 isbn 978-0-521-19014-5 Hardback
Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.
Contents
Preface Abbreviations and notes Editors’ introduction i ii iii iv v vi vii viii
Moore on his dissertations Early life The Trinity College Prize Fellowship competition Six years as a Prize Fellow: 1898–1904 Intellectual background Moore’s debate with Kant Moore’s examiners The text of the dissertations
page vii x xii xii xiv xvii xx xxiii xlvi lxvii lxxi
the 1897 dissertation: the metaphysical basis of ethics Preface
3
Introduction
6
Chapter I Freedom
20
Appendix: Professor Sidgwick’s Hedonism
87
examiners’ reports on the 1897 dissertation H. S. on G. E. Moore, Dissertation on the Metaphysical Basis of Ethics Henry Sidgwick
97 99
Report on Mr. Moore’s Essay Edward Caird
v
vi
Contents
the 1898 dissertation: the metaphysical basis of ethics Preface
117
Introduction
120
Chapter I On the meaning of ‘Reason’ in Kant
133
Chapter II Reason
161
Chapter III The meaning of ‘Freedom’ in Kant
181
Chapter IV Freedom
211
Chapter V Ethical Conclusions
232
Appendix on the chronology of Kant’s ethical writings
238
examiner’s report on the 1898 dissertation Report on a Dissertation entitled ‘The Metaphysical Basis of Ethics’ by Mr. G. E. Moore Bernard Bosanquet
Index
245 250
Preface
About twenty-five years ago, when I was working on my book on G. E. Moore (G. E. Moore, Routledge, 1990), I came across the manuscripts of Moore’s Fellowship dissertations which had recently been deposited in the Wren Library at Trinity College Cambridge. Realising the importance and significance of these dissertations, I formed a plan to prepare an edition of them for publication. But before this could proceed the permission of Moore’s son Timothy Moore was needed, and Timothy was unwilling for this to go ahead, on the grounds that his father’s reputation might be harmed by the publication of these juvenilia which his father had regarded as confused and unsatisfactory. The matter rested there until Consuelo Preti, while working recently on Moore’s early papers, persuaded me to revive the plan to publish his dissertations. By this time Timothy Moore was dead and control of his literary estate had passed to Moore’s grandson Peregrine Moore, who was happy to agree to the publication of his grandfather’s Fellowship dissertations. We are grateful to him for agreeing to this, and we are confident that, far from being harmed, his grandfather’s reputation will only be enhanced by this edition of his early philosophical writings. In preparing this edition we have both spent a good deal of time in the Wren Library at Trinity College, and it is a pleasure to record our thanks to the Librarian, David McKitterick, and to his staff for their unfailing help and patience as we have returned again and again to check our transcripts of Moore’s manuscripts. We have also made considerable use of the Moore archive at Cambridge University Library, and we are grateful for the help we have received from the staff there. At Cambridge University Press Joanna Garbutt has made valuable suggestions while helping us to prepare this book for publication and we are much indebted to her. I am grateful to my Department at the University of York and the UK Arts and Humanities Council for the research leave in 2009 during which I worked on this volume. vii
viii
Preface
We have worked together on this volume and as well as sharing the work of preparing the edition of Moore’s dissertations, we have composed our editors’ introduction together. We are therefore grateful to each other for correcting our mistakes, but also happy to blame one another for the mistakes which remain. thomas baldwin
I must acknowledge the help of many people for making the research for this volume possible, most of which took place between 2005 and 2009. I am especially grateful to the American Philosophical Society for the award of a Franklin Grant which enabled me to spend four weeks in Cambridge in 2007. Dean Susan Albertine and Dean Deborah Compte at the College of New Jersey supported my research with Dean’s Mini-Grants between 2005 and 2009, for which I am very grateful. My particular thanks are due to Godfrey Waller and the staff at the Cambridge University Library Manuscripts Reading Room for their help with the Moore archives, and also to Jonathan Smith and the staff at the Wren Library, Trinity College, for their help with the Moore papers there. In Cambridge, Bruno and Polly Kenway took me in good-naturedly for weeks at a time every year, during (among other things) Royal College of Surgeons exams, Ph.D. writing, new babies and some quite lurid degrees of illness. I could not have managed this project without their generosity, for which I am immeasurably grateful. Also in Cambridge, Francesca Stubbins provided much-needed distraction and company after the library every day (even though I was always there during her exams). At Woodhouse in Yorkshire, Liz and Jeremy Stubbins, along with Christina and Hugh, provided my weekends away from the library with a peaceful agricultural antidote, for which I am very thankful. In London, Twyla Howse (with help from Nathan and Danny) hosted me and ferried me around when I was able to get away from the library; I am very grateful to her. I am likewise grateful to Simon Blackburn, Kenneth Blackwell, Malcolm Budd, Nick Griffin and Gary Ostertag for all their support of my work (and to Gary for German translation). My greatest and most heartfelt debt, of course, is to Tom Baldwin, who welcomed my interest in Moore with great kindness, support and enthusiasm. His knowledge of G. E. Moore’s work is unparalleled, and he has patiently answered my many questions. Tom is also indefatigably affable – even when he disagrees with me on some point of Moore
Preface
ix
interpretation – and I have learned much from talking with him (but not enough). I am exceedingly grateful for his generosity and his keenness to take on this project. His friendship has meant a great deal to me. My deepest thanks finally are to Michael Esposito, for his inexhaustible supply of devotion, steadiness and encouragement; he makes everything possible. consuelo preti
Abbreviations and notes
ABBREVIATIONS
In our edition of Moore’s dissertations we have retained Moore’s abbreviations for the works to which he refers frequently; Kant’s main works and Sidgwick’s The Methods of Ethics, whose title he abbreviates as ME. Moore’s references to ME are to the fifth edition of 1893. But the only edition readily available to present-day readers is the seventh edition of 1907 and since this turns out to differ from the fifth edition in some significant details these are noted in the footnotes; we have also noted where the seventh edition has different page numbers. In the case of Kant’s main works, Moore’s abbreviations are: R.V. for Kritik der reinen Vernunft (Critique of Pure Reason) P.V. for Kritik der praktischen Vernunft (Critique of Practical Reason) G. for Grundlegung zur Metaphysik der Sitten (Groundwork of the Metaphysic of Morals) M. for Metaphysik der Sitten (Metaphysic of Morals) Moore’s references are to the page numbers in Hartenstein’s 1867 edition of Kant’s works. This edition is unlikely to be available to contemporary readers and we have supplemented them with references to the standard Prussian Academy edition of Kant’s works. In the case of the Critique of Pure Reason we usually just give the page number from the second ‘B’ edition, though where the reference is to a passage from the ‘A’ edition we identify it as such; in the case of the other works by Kant we give the volume and page number from the academy edition (e.g. ‘AK 5: 63’). We also give a page reference to the English translations whose titles are abbreviated as follows: GW for the translation of Critique of Pure Reason by Paul Guyer and Allen Wood (Cambridge University Press, 1997); MG for the translation of Kant’s Practical Philosophy by Mary Gregor, which includes translations of P.V., G. and M. (Cambridge University Press, 1996). x
Abbreviations and notes
xi
In our editors’ introduction we also employ these abbreviations; we also follow the common practice of using PE to abbreviate the title of Moore’s Principia Ethica; page references are to the revised edition (ed. T. Baldwin, Cambridge University Press, 1993). When providing cross-references within his own theses Moore left the page numbers blank, as he did not have the numbers of his final typescript. Here these have been completed with the page references of the present edition. NOTES
In our edition of Moore’s dissertations we use a dual system of footnotes in order to distinguish Moore’s notes from our own. Moore’s notes are identified by lower case roman letters on a page-by-page basis, with editorial comments in square brackets; our own notes are identified by arabic numerals which run sequentially throughout each dissertation.
Editors’ introduction
I MOORE ON HIS DISSERTATIONS
In this volume we publish, for the first time, G. E. Moore’s 1897 and 1898 Trinity College Prize Fellowship dissertations. The most noteworthy of his early philosophical writings, the dissertations represent a significant stage in the history and development of Moore’s early thought, a stage that culminated with the publication of his 1903 Principia Ethica. In the autobiographical introduction to The Philosophy of G. E. Moore Moore gives a brief account of his ‘Two years working for a Fellowship: 1896–1898’ (Moore 1942, 20–2): I did well enough in the Moral Sciences Tripos to make my advisers think it worth while that I should compete in Philosophy in the annual Fellowship examination at Trinity. In order to compete, it was necessary to submit a dissertation; and, after consulting Ward, I decided to try to write one on Kant’s Ethics. Accordingly for the next two years, 1896–1898, I was engaged in trying to do this; and, of course, a great deal of my time was spent in puzzling over Kant’s three Critiques, his Prolegomena, and his Grundlegung zur Metaphysik der Sitten. During the first of these two years the part of Kant’s ethical doctrines with which I was chiefly concerned was some of the things he said about freedom. He seemed to me to say or imply that each of us had two ‘selves’ or ‘Egos,’ one of which he called a ‘no¨umenal’ self, the other an ‘empirical’ self, and he seemed also to say or imply that the ‘no¨umenal’ self was free, whereas the ‘empirical’ self was not; and what I wrote during this first year can, I think, be described as, in the main, an attempt to make sense of these extremely mysterious assertions. I found something which seemed to me at the time to give them an intelligible meaning, but I have no doubt that the meaning I found was as far as or further from anything which Kant actually meant, as was McTaggart’s interpretation of Hegel’s ‘absolute Idea’ from anything which Hegel meant. The substance of what I wrote on this topic was published shortly afterwards as an article in Mind, entitled ‘Freedom;’ and, though I have not looked at that article for a long time, I have no doubt that it was absolutely worthless. I expect that Sidgwick, who was the representative of Philosophy among the Fellowship electors in that year (1897), xii
Editors’ introduction
xiii
must have felt about my dissertation much the same as he is said to have felt about the dissertation on Hegel by which McTaggart won his fellowship a few years earlier. Sidgwick is reported to have said about McTaggart’s dissertation (and I believe this is authentic): ‘I can see that this is nonsense, but what I want to know is whether it is the right kind of nonsense.’ I think he must have decided about my nonsense, as he had decided about McTaggart’s, that it was the right kind; for, though I was not elected that year, Ward told me that in the next year (1898), when he had taken Sidgwick’s place on the Board of Electors, Sidgwick spoke to him just before the final meeting of the Electors and warned him that he must be careful not ruin my chances of election by failing to speak sufficiently favourably of my work. In the second year’s work (1897–98) I got on to what I think was a much more profitable line of inquiry, though one which had a much less direct connection with Kant’s Ethics – had, indeed, a more direct connection with the Critique of Pure Reason than with the Critique of Practical Reason. It seemed to me that it was extremely difficult to see clearly what Kant meant by ‘Reason.’ This was a term which occurred not only in the title of both these works, but also frequently in the text, and, as it seemed to me, in a very mystifying manner. What on earth did Kant mean by it? He must be referring, more or less directly, to something which was to be found in the world, and which could be described in other terms. But to what exactly? This was what I set myself to think about; and it led me to think first about the notion of ‘truth,’ since it seemed to me that, in some of its uses at all events, Kant’s term ‘Reason’ involved a reference to the notion of ‘truth;’ and, in thinking about truth, I was led to take as my text a passage from the beginning of Bradley’s Logic, in which after saying that ‘Truth and falsehood depend on the relation of our ideas to reality,’ he goes on to say that the ‘meaning’ of an idea consists in a part of its content ‘cut off, fixed by the mind, and considered apart from the existence’ of the idea in question. It seemed to me, if I remember right, that the meaning of an idea was not anything ‘cut off’ from it, but something wholly independent of mind. I tried to argue for this position, and this was the beginning, I think, of certain tendencies in me which have led some people to call me a ‘Realist,’ and was also the beginning of a breakaway from belief in Bradley’s philosophy, of which, up till about then, both Russell and I had, following McTaggart, been enthusiastic admirers. I remember McTaggart once saying of an occasion when he met Bradley at Oxford that, when Bradley came in, he felt ‘as if a Platonic Idea had walked into the room.’ I added what I had written this year about ‘reason’ and ‘ideas’ as a concluding chapter to what I had written the year before, and submitted the whole at the Fellowship Examination in 1898. This time I was elected. The substance of the new chapter was published soon afterwards in Mind under the title of ‘The Nature of Judgment;’ and though I am sure that article must have been full of confusions, I think there was probably some good in it. Ward, although he had secured my election to a Fellowship, was not very happy about me. When I went to see him after the election, he told me he thought I was too sceptical, and that I seemed to take a pride and pleasure in picking
xiv
Editors’ introduction
holes in accepted views; this he did not like, and he compared me in that respect to Hume. II EARLY LIFE
George Edward Moore was born in November 1873 into a comfortably prosperous middle-class family. He was the fourth of seven children of Dr Daniel Moore and his wife Henrietta,1 and named for his paternal grandfather.2 Moore’s father had retired from his medical practice two years before Moore was born to live in a suburb of London not far from Dulwich College, in order to send his boys to school there. The family was close – the great majority of the letters in the Moore archive comprise letters between Moore and his parents, sisters and brothers. Moore was sent to school at Dulwich College from 1882 until 1892.3 Moore’s studies at Dulwich, particularly in the last six of his schoolboy years, mainly centred on translation of English prose and verse into Greek and Latin, with ‘very few hours per week being given to French and German and to some mathematics . . . ’ (Moore 1942, 5–6). In spite of this Moore was not prepared to deny that he had received a good education at Dulwich, averring that the translation of so much English literature exposed him to qualities of the language that he would not otherwise have had the opportunity to appreciate (Moore 1942, 6). When Moore was ten years old he and his brother Tom, along with two of their friends (also brothers), formed what they called the ‘Boomerang Club’, which they kept up for a few years thereafter. The future editor of Mind and his friends produced a club journal, whose issues each contained a contents page, editorial, articles, poems, reports of sporting events, jokes, letters and reviews. As a young man, Moore continued the practice of keeping accounts of his activities;4 and the close companionship of a trusted group of friends remained a key feature of his days as an undergraduate and Fellow. 1 2 3
4
Dr Moore had had a daughter by his first wife, so there were eight children in the Moore family. Moore much disliked his given name, and called himself ‘Bill’ in the company of his own wife and sons. Former scholars of Dulwich are known as Old Alleynians, after the founder of the school, Edward Alleyn. Dulwich College today describes Moore as one of its ‘Eminent Old Alleynians’, and (curiously) as the author of ‘Ethics (1912) . . . a key work which pointed out that moral philosophers, and particularly utilitarians, were logically confused’. These mainly take the form of lists that record events from his arrival in Cambridge in 1892 and continue up to 1913. Moore did keep diaries but seems to have destroyed most of them. Those that survive record events in 1908, 1909–16 and 1924. There are some supplementary diaries/notebooks from 1909 to 1928 and extracts from diaries from 1929 to 1939. The content of the surviving diaries makes it difficult to avoid the supposition that Moore preserved (just) the ones that mentioned Wittgenstein.
Editors’ introduction
xv
In October 1892, following his brother Henry, Moore arrived at Trinity College Cambridge to study classics. Moore described his first two years’ work for part i of the Classics Tripos as consisting of nothing in which he was not already proficient, given the intensity of the schooling in classics he had received at Dulwich (Moore 1942, 5). But in his first and second years at Trinity, Moore began to get to know the young men who would have the greatest impact on him, personally and intellectually. One of Moore’s classics tutors at Trinity, A. W. Verrall, had himself been at Dulwich College. Verrall had been tipped off by the masters at Dulwich of Moore’s promise and, as a member of the selective and covert Cambridge Conversazione Society – the Apostles – he was among those who recommended Moore for membership.5 The society was the defining intellectual experience of Moore’s undergraduate life. Never more than a handful of active members at any time (usually twelve, mostly drawn from Trinity and King’s College), the group met every Saturday evening in the rooms of a member. Someone would present a paper on a topic; debate would ensue; and a vote would be taken at the end on a question that was only sometimes directly related to the subject of the paper. The members of the society in 1894 included Russell, McTaggart, the Davies brothers, Theodore and Crompton, Robert Trevelyan, Ralph Wedgwood, Edward Marsh and Charles Sanger; a few years later, the society elected George Trevelyan, Desmond MacCarthy and Alfred Ainsworth. These were, and mostly remained, Moore’s closest friends.6 Moore’s aptitude for argument, which surfaced early on, was what caught the attention of notable Apostles (like Russell) – he was considered as quite ferocious in philosophical discussions, at least in the company of close friends. In letters to his parents in 1893 Moore describes meetings with McTaggart, ‘a fellow at Trinity for metaphysics, who is very interesting’ who ‘spent nearly an hour in trying to explain to me the metaphysical aspect of “time”’. Arguing forcefully with McTaggart about the nature of time was later described by Moore as an early experience of the appeal of 5
6
Such was Moore’s regard for the clandestine nature of the society that he only obliquely mentions the influence of ‘friends’ (whose conversation was nevertheless characterised as brilliant) in his autobiography (Moore 1942, 12). In late 1898 he seems to have gone so far as to scold Russell for a lapse in discretion about the society; Russell replies ‘It is not very unwise to write about the Society on a p[ost] c[card]; none but the porters would read it and none but the initiates would understand it. However, it is perhaps better to avoid it’ (Griffin 1991, 186). It is not clear that the society was as secret as its members thought it was. See Levy (1979) for details on the Apostles’ Society and its members during Moore’s time at Cambridge. Moore was particularly close to MacCarthy and Ainsworth (who married one of Moore’s sisters).
xvi
Editors’ introduction
philosophy.7 McTaggart himself was an Apostle, and Moore was elected to the society on 10 February 1894. His first contribution at a meeting, a week later, was an objection to a point of Russell’s on the question posed for the evening, ‘What ought Cambridge to give?’ Russell had argued that the end result of a Cambridge education was young men unfit for practical life, as a result of the profound scepticism produced in them. Moore instead countered that ‘we should . . . spread scepticism until at last everybody knows that we can know absolutely nothing’. Moore’s manner, perhaps more than his inaugural remarks, prompted Russell to write to his wife Alys that ‘the scene was so perfectly wonderful and unprecedented that I would give anything to be able to describe it adequately’, managing none the less to paint a picture of Moore as having ‘electrified’ the assembly, who had ‘never realized what fearless intellect pure and unadulterated really means’.8 Moore went on to deliver his first paper to the society later that year, entitled ‘What end?’9 This, Moore’s earliest licensed incursion into philosophy proper, is a discussion of hedonism, a subject Moore was to revisit in his Trinity Prize Fellowship dissertations and develop further in Principia Ethica. In 1894 Russell had successfully achieved a first-class result in part ii of the Moral Sciences Tripos after only a year’s study, having already achieved a first-class result in part ii of the Mathematics Tripos in 1893. Russell encouraged Moore to do something similar after he had completed part i of the Classics Tripos in 1894, as did Moore’s teachers Jackson, Ward and Verrall. So instead of just sticking with the Classics Tripos and qualifying himself thereby for the career as a classics teacher in a public school he had initially envisaged (Moore 1942, 13), Moore took their advice;10 and 7
8
9 10
Add. MS 8330 2/1/16 (18 March 1893). Moore later accords to McTaggart almost as much credit for philosophical influence as he does to Russell (Moore 1942, 18) – in particular for what has become the hallmark of Moore’s own style, ‘trying to give a precise meaning to philosophical expressions, on asking the question “What does this mean?”’ An interesting point about Moore’s philosophical development arises, however, in his retelling of his first meeting with McTaggart (Moore 1942, 14). Moore there claims that upon hearing McTaggart express ‘his well-known view that Time is unreal’ he was sufficiently moved to argue against it, given that ‘this must have seemed to me then (as it still does) a perfectly monstrous proposition’. But Moore’s first publication was part of a symposium on the nature of time (Moore 1897), in the course of which he seems to find the doctrine of the unreality of time less ‘monstrous’ than he later claims he did. Bertrand Russell to Alys Pearsall Smith, 18 February, 1894. Moore’s own recollection of these discussions, characteristically, takes a more diffident line: ‘I felt (and was) extremely crude compared to them; and did not feel able to make any contributions to the discussion which would bear comparison with those which they were making. I felt greatly flattered, and rather surprised, that they seemed to think me worthy of associating with them’ (Moore 1942, 12–13). Add. MS 8775 12/1/1. Moore wrote to his parents to tell them of his decision on 11 June 1894 (Add. MS 8330 2/1/33).
Editors’ introduction
xvii
from 1894 until 1896 he combined his studies for part ii of the Classics Tripos with further study for part ii of the Moral Sciences Tripos. In preparation for the Moral Sciences Tripos, Moore attended lectures by Stout (history of philosophy, 1894) and Sidgwick (ethical systems, 1894), as well as McTaggart’s lectures on Hegel (1895–6),11 and he successfully passed the Moral Sciences part ii Tripos exams in May, 1896 (first class, with distinction, though he only achieved a second-class result in part ii of the Classics Tripos). Committed now to philosophy, Moore’s thoughts turned towards submitting a dissertation in the hope of winning a Trinity Prize Fellowship. III THE TRINITY COLLEGE PRIZE FELLOWSHIP COMPETITION
Moore submitted his dissertations in the hope that he would be elected to a Prize Fellowship at Trinity College Cambridge, and he was successful in this attempt in 1898, having been unsuccessful the previous year. These Prize Fellowships were immensely valuable positions for those who aspired to enter an academic career. At this time there were no postgraduate programmes or degrees in Britain even though they were already well established in Germany and had been introduced into the USA (the Ph.D. degree was first introduced as a postgraduate research degree, as opposed to an honorary award, in Britain in 1917). Since those British students who aspired to an academic career still needed a way of becoming qualified for such a career after the completion of an undergraduate degree, the Prize Fellowships offered by Oxford and Cambridge colleges provided one of the main ways of achieving this, even though these Fellowships were typically available only to students of the colleges providing them. Among these Fellowships those at Trinity were among the most valuable and prestigious: a Trinity Prize Fellowship lasted for six years, provided free board and lodging in college, and included an annual ‘dividend’ of £200 (which would be worth about £16,000 today if one inflates by relative prices of goods).12 In aspiring to become a Prize Fellow, Moore was attempting to follow in the footsteps of his friends McTaggart and Russell. McTaggart had achieved first-class honours in part ii of the Moral Sciences Tripos in 1888, and in 1891 he was elected to a Prize Fellowship at Trinity; his dissertation formed the basis of his book Studies in the Hegelian Dialectic (McTaggart 1896). 11 12
Moore’s lecture notes survive (Add. MS 8775, 10/1/1; 10/2/1; 10/3/1). £200 then would be worth about £85,000 today if one inflates by average earnings. Perhaps this tells us something about the relative decline in academic salaries.
xviii
Editors’ introduction
So McTaggart was already a Prize Fellow when Moore met him in 1893. In 1897, when his Prize Fellowship ended and Moore first competed for a Prize Fellowship, McTaggart was appointed a lecturer in moral sciences at Trinity College; so for him the Prize Fellowship precisely enabled him to make the transition from undergraduate studies to an academic career. Russell’s progress was initially similar: having achieved first-class honours in part ii of the Moral Sciences Tripos in 1894 he was awarded a Prize Fellowship at Trinity in 1895; his dissertation formed the basis of his book An Essay on the Foundations of Geometry (Russell 1897). So Russell was an undergraduate when Moore met him during his first year at Trinity, but had become a Prize Fellow by the time Moore was starting his third year. Unlike McTaggart Russell did not become a college lecturer when his Prize Fellowship ended in 1901. He had a substantial private income and following his discovery of his ‘paradox’ in 1901, he was in a position to undertake his ground-breaking research into the foundations of mathematics without any teaching duties. But in 1910, when his private income was becoming depleted, he did accept an invitation to become a college lecturer in mathematics at Trinity College; he was deprived of this position in 1916 because of his opposition to the war. Whereas McTaggart took three years between his graduation in 1888 and his election to a Fellowship in 1891, Russell achieved the same result in one year. Moore attempted to match Russell’s achievement: he graduated in 1896 and entered the competition in 1897. But, as Moore himself acknowledged, his 1897 dissertation is incomplete. Hence although these points are not mentioned explicitly in the examiners’ reports by Sidgwick and Caird, it is easy to imagine the 1897 Trinity electors thinking that the award of one of their valuable Prize Fellowships to Moore just one year after his graduation would be premature. It may also be that preference was given to some candidates who had been turned down the year before and for whom the 1897 competition was their last chance of election to Fellowships.13 One year later, however, Moore’s new dissertation was more polished and more original and, despite Bosanquet’s negative report on it (see vii below), this time Moore was successful. As Moore’s account indicates (see i above), for this success Moore was primarily dependent on Ward’s support, even though, as Moore also indicates, Ward did not in fact much like the work. 13
Russell wrote to Moore from Venice on 19 October 1897 to say: ‘I am sorry to see from “Nature” – the only English paper I get – that you did not, apparently, get your Fellowship at this shot. However, as Fletcher and Lawrence were both at their last chance, I suppose they could not be passed by’ (Add. MS 8330 8R/33/3).
Editors’ introduction
xix
Moore’s election to a Prize Fellowship did not in fact depend wholly on the assessment of his dissertation. The 1896 edition of the Trinity College ‘Ordinances concerning Fellowships and the Fellowship Examination’ includes the following clauses:14 7. Candidates may be examined in the subjects of their dissertations and in matters connected with them as well as in the branches of study to which they refer, but the questions set in the examination will not be confined necessarily to the subjects indicated by the candidates. 8. Every Candidate is expected to take the questions on Modern Philosophy in both the Philosophy papers, and also the paper on English Essays. Any Candidate may send to the Secretary of the Council not later than July 31 preceding the examination a list of works on Modern Philosophy upon which he wishes to be examined, but the questions will not be confined necessarily to these works. ... 10. Every candidate in the Moral Sciences is expected to take the questions on Ancient Philosophy.
In accordance with clause 8 here Moore did indeed send a list of works on which he wished to be examined, namely parts of Kant’s major works:15 Trinity College July 31st, 1897 Dear Sir In accordance with Regulation 8 of the Fellowship Regulations issued to candidates, I wish to give notice that I should prefer to be examined in the following works on Modern Philosophy. Kant’s works on the Principles of Morals: namely all those translated by T. R. Abbott (Longmans 1889), including the ‘Grundlegung zur Metaphysik der Sitten’; the ‘Kritik der Praktischen vernunft’; out of the ‘Metaphysik der sitten’ Part i, the ‘Einleitung in die Metaphysik der Sitten’, and out of Part ii the ‘Vorrede’ and ‘Einleitung zur Origenlehre’; and the ‘Erstes Stuck’, of the ‘Religion innerhalb der Grenzen der blosen Vernunft’; and also out of the ‘Kritik der Reinen Vernunft’, the section in the ‘Dialectik’ on the ‘Aufl¨osung der kosmologischen Idee von der Totalit¨at der Ableitung der Weltbegebenheiten aus ihren Ursachen’, and in the ‘Methodenlehre’ on the ‘Kanon der reinen Vernunft’. yours sincerely G. E. Moore
No similar letter survives with a request concerning the works on which he is to be examined in 1898. But in a letter to Russell of September 14 15
There do not appear to be new editions for 1897 and 1898. So it is reasonable to assume that Moore’s attempts to gain a Fellowship were governed by the rules specified in the 1896 edition. The letter is bundled in with the dissertation manuscripts at Trinity College.
xx
Editors’ introduction
1898 Moore writes ‘I must begin to stir again now, doing Classics for the exam’, where the ‘exam’ in question is presumably that for the Prize Fellowship competition and the ‘Classics’ questions arise from clause 10 of the ordinances. Moore’s period as a Prize Fellow of Trinity College lasted from 1898 until 1904. Moore records that soon after his election Sidgwick, slightly surprisingly, advised him ‘to spend a year or two at some German university’ (Moore 1942, 17). Moore did not take this advice, although he had earlier followed Ward’s advice and spent the summer of 1895 in T¨ubingen. One can speculate whether Moore’s philosophical development would have been different had he taken Sidgwick’s advice; Moore himself comments ‘I still feel very doubtful whether I should have got as much benefit by studying in Germany as I did from staying at home’. And there is no question but that Moore did use the six years of his Prize Fellowship in a very productive way. IV SIX YEARS AS A PRIZE FELLOW: 1898–1904
Moore’s list called ‘Chronological Table of My Life’ contains an underlined entry for October 1898: ‘Become a Fellow’. In one way the life of the scholar suited Moore down to the ground – he had, he recalls, ‘a set of Fellows’ rooms on the north side of Nevile’s Court – a very pleasant place and a very pleasant life’ (Moore 1942, 23). But Moore’s letters to friends and family during this time indicate that he worried all the time about the difficulty of getting on with his work. He even later claimed that had it not been for the stimulus of requests from the Aristotelian Society for papers for symposia, which he felt he could not refuse, he would have published much less than he did (Moore 1942, 24). In fact Moore was already engaged on writing a series of entries for two volumes of J. M. Baldwin’s Dictionary of Philosophy and Psychology (Baldwin 1902).16 More importantly he had also already accepted a commitment to provide some lectures for the London School of Ethics and Social Philosophy, an institution founded by Bosanquet and Sidgwick to provide the working men of London with an opportunity for some university-level instruction.17 Moore provided two courses of lectures, one in the autumn of 1898 on the ‘Elements of Ethics’ and the other in spring 1899 on ‘Kant’s 16
17
He contributed the entries in volume i on ‘Cause and effect’ and ‘Change’; and in volume ii the entries on ‘Nativism’, ‘Quality’, ‘Real’, ‘Reason’, ‘Relativity of Knowledge’, ‘Relative and Absolute’, ‘Substance and Essence’, ‘Spirit’, ‘Teleology’, and ‘Truth and error’. See Regan (1991) for a description of the London School of Ethics and Social Philosophy and for the text of the 1898 lectures.
Editors’ introduction
xxi
Moral Philosophy’. Each course was ten lectures long and Moore went to London every Thursday to deliver his lecture. Moore had the text of the autumn lectures typed up with a view to preparing them for publication;18 significantly perhaps, Moore did not take the same attitude to the text of his 1899 lectures.19 Complaints to his parents and others notwithstanding, during the period of his Fellowship Moore in fact produced nine papers, including the influential paper in which he set out his critical response to idealism, ‘The Refutation of Idealism’ (Moore 1903b). Crucially, too, the Fellowship enabled him to complete his single most important book, Principia Ethica (Moore 1903a). This evolved out of the 1898 London lectures on the ‘Elements of Ethics’ which Moore had had typed up. In 1902 Moore submitted a revised version of these lectures to Cambridge University Press for publication. In reply Moore received a letter from Sorley, who told Moore that ‘he was responsible for the Syndics’ decision’ to accept the lectures for publication, but then provided a list of criticisms to which Moore was expected to respond. Moore struggled at times over the revisions Sorley had requested, but eventually produced his masterpiece which was published in October 1903.20 In retrospect Moore was characteristically doubtful about the value of his work during this period. In the preface that he wrote in 1922 for the second edition of Principia Ethica, but which he did not in the end publish, he wrote, concerning Principia Ethica, ‘I now see that this book, as it stands, is full of mistakes and confusions’ (Moore 1993a, 2); and the only paper from this period which he included in his 1922 collection of his papers Philosophical Studies was ‘The Refutation of Idealism’, concerning which he commented ‘This paper now appears to me to be very confused, as well as to embody a good many down-right mistakes’ (Moore 1922, viii). Moore himself was not, however, best placed to appreciate the value of his own work during this period. For by 1922 he had changed his general approach to philosophy, increasingly emphasising the importance of philosophical analysis and the ‘Common Sense view of the world’. That perspective is not, however, the appropriate one for an appreciation of the 18
19 20
Two typescripts of the 1898 lectures survive in the Moore archive in Cambridge University Library (Add. MS 8875 14/1/1 and 14/1/2). Typescript 14/1/1 contains over 300 marginal comments, some in Moore’s hand, some in Russell’s and some in what we believe is Sorley’s, who read the typescript for Cambridge University Press and corresponded with Moore about it. Moore kept his notes for these lectures and they are now in the Moore archive in Cambridge University Library (Add. MS 8875 14/2/1). See Baldwin (1993) for details of the relationship between ‘The Elements of Ethics’ and Principia Ethica.
xxii
Editors’ introduction
value of his early work and the two dissertations which initiate it. Instead, the value of this work lies in the ways in which, starting from the idealist standpoint of the 1897 dissertation, he thinks his way, via the 1898 dissertation, to the analytical realism of Principia Ethica and ‘The Refutation of Idealism’. Moore himself, once immersed in his distinctive analysis of common sense, did not recognise his own achievement; but in the context of contemporary inquiries into the origins of analytic philosophy, the contribution made by Moore’s early work is obvious. In 1904 Moore’s Prize Fellowship came to an end. Given that McTaggart was well established as a college lecturer in moral sciences at Trinity, there was no opportunity for him to make a similar transition to a college lectureship, since the college did not need two lecturers in moral sciences. Instead Moore applied to have his Fellowship continued as a Research Fellowship. This application was not successful; in his autobiography Moore comments ‘election to a Research Fellowship was, and still is a very rare and exceptional thing at Trinity, and I was not surprised that my application was refused’ (Moore 1942, 25).21 It appears that the college had again consulted Bosanquet about this application and that his advice was again negative;22 but it is likely that Ward’s opinion will have been crucial since Russell had by now left Trinity and even though Moore may have had McTaggart’s support, Ward was the senior philosopher at Trinity. At this juncture, therefore, Moore might have had to fall back after all on the option of taking a position as a classics teacher in a public school. But he was saved from this fate by the fact that, following the recent death of his parents, he had obtained an inheritance which enabled him ‘to live in moderate comfort without needing to earn anything’, so that he could ‘go on working at philosophy, which was what I wanted to do, without a Fellowship and without needing to try to obtain any paid employment’ (Moore 1942, 26). Moore did, however, leave Cambridge at this time, moving first to Edinburgh and then in 1908 to London where he shared a house with his sister Hettie. During these years he continued to publish reviews and articles, delivered papers to the Aristotelian Society and gave an important series of lectures on ‘Some Main Problems of Philosophy’ at Morley College, London, in 1910–11.23 In 1911 he was offered, and accepted, a lectureship 21 22
23
One person who was later awarded a Research Fellowship at Trinity was Wittgenstein, partly through Moore’s strong support for his application. A diary entry dated 27 February 1914 reads: ‘Feel depressed. Fletcher tells me that why the older members of Council voted against my Research Fellowship [added in margin ‘1904’] was because of unfavourable reports from English philosophers (Bosanquet)’ (Add. MS 8330 1/3/4). These lectures were published in 1953 – see Moore (1953).
Editors’ introduction
xxiii
at Cambridge after the resignation of J. N. Keynes. For the next twentyeight years until his retirement in 1939 he lectured at Cambridge, first on psychology, and subsequently on metaphysics (but never on ethics).24 He quickly re-established himself there, and in 1925 he succeeded Ward as Professor of Philosophy. His tenure of this professorship spans the golden period of Cambridge philosophy when a succession of visitors from abroad came to study with him and Wittgenstein, who returned to Cambridge in 1929. Moore did not publish much during these years, but he exercised his influence through his friendships, lectures, the meetings of the Moral Sciences Club and his role as editor of Mind (which he edited from 1921 until 1945).25 V INTELLECTUAL BACKGROUND
The story of the rise of analytic philosophy has long emphasised the dominance of idealist philosophy in nineteenth-century Oxford and Cambridge, and the ‘rebellion’ of Moore and Russell from its central doctrines.26 A key element in that story is Russell’s credit to Moore for having rescued him from its stifling effects: On fundamental questions of philosophy, my position, in all its chief features, is derived from Mr G. E. Moore. I have accepted from him the non-existential nature of propositions (except such as happen to assert existence) and their independence of any knowing mind; also the pluralism which regards the world, both that of existents and that of entities, as composed of an infinite number of mutually independent entities, with relations which are ultimate, and not reducible to adjectives of their terms or of the whole which those compose. Before learning these views from him, I found myself completely unable to construct any philosophy of arithmetic . . . (Russell 1903, xviii) But it was not only these rather dry, logical, doctrines that made me rejoice in the new philosophy. I felt it, in fact, as a great liberation, as if I had escaped from a hot-house on to a wind-swept headland. (Russell 1959, 61)
However, partly due to the uncritical acceptance of Russell’s own descriptions, some of the details of this story remain unexplored. In his early writings Russell certainly accepted the framework assumptions of the neoHegelian metaphysics he had learnt from McTaggart. And Moore cites 24 25 26
Moore kept his lecture notes and they are now preserved in the Moore archive in Cambridge University Library. Moore’s undogmatic openness in discussion is affectionately described by Gilbert Ryle in his memoir of Moore: ‘G. E. Moore’ in Ryle (1971). See Hylton (1990) and Griffin (1991) for detail.
xxiv
Editors’ introduction
Bradley – and no one else – as an influence in the preface to his 1897 dissertation: ‘It is to Mr. Bradley’s “Principles of Logic” and “Appearance and Reality” (2nd. edn. 1897) that I chiefly owe my conception of the fundamental problem of Metaphysics’ (1897 dissertation, p. 4). But the story here is not simply one of the dominant influence of idealist philosophy on Russell and Moore. Idealist philosophy itself was not a unified doctrine, and an important aspect of Moore’s acknowledgement to Bradley is that he credits Bradley’s metaphysics with preventing him from accepting Caird’s neo-Kantian idealism. Thus at this stage Bradley’s Absolute, which transcends not only our own consciousness, but consciousness generally, provided Moore with a way of avoiding the subjective idealism that he held had infected Kant’s ethics and metaphysics. Furthermore a fuller picture of the rise of analytic philosophy will link Moore’s early philosophical development to other intellectual developments in the late nineteenth century, in particular to the rise of ‘mental science’, i.e. scientific psychology. For an important aspect of the way in which Moore became equipped to distance himself from his early embrace of Bradley’s idealism in 1897 to the realism at the core of his 1898 dissertation is to be found in his reaction to this newly emerging field of scientific psychology. Finally, it is also important to take account of the influence on Moore’s still developing ethical theory of his teacher, Henry Sidgwick, who remained a powerful critic of idealist ethics.27 v.1 british idealism
In order to understand the connections between Moore’s early philosophy and the idealist philosophy of the period it is sensible to start with a brief account of T. H. Green’s neo-Kantian metaphysics since this sets the scene for Bradley’s metaphysics which, as we have observed, initially provided Moore with his philosophical framework.28 T. H. Green (1836–82) died at the age of forty-five when he was at the height of his powers and influence. His most important work, Prolegomena to Ethics (Green 1883), was nearly complete at his death; among his earlier works the most significant was the long introduction he wrote for the 1874 edition of Hume’s Treatise of Human Nature which he prepared with T. H. Grose (Green 1874) and which is reprinted in volume i of his 27
28
There were other important British critics of idealist philosophy, most notably ‘Oxford Realists’ such as J. Cook Wilson. But little of their work was published at this time and it does not appear to have had any significant influence outside Oxford until later. Hylton (1990) argues that Bradley’s distinctive idealism is not comprehensible except in the context of Green’s more canonical idealist and anti-empiricist views.
Editors’ introduction
xxv
works (Green 1885–9). Green’s views were formulated in reaction against the empiricist account of knowledge to be found in Hume, though in fact he concentrates mainly on Locke’s position. Green’s reaction was Kantian in spirit: he argued that knowledge, even that based on the simplest experiences, involved an active consciousness which relates and organises disparate experiences as perceptions of the world. The Lockean, empiricist, position that perceptual knowledge is based upon the passive imprint of ideas of sensation upon the mind is, according to Green, untenable. Green accepts that it is not Locke’s view that the mere occurrence of sensation is by itself knowledge of anything; instead, as Green notes, Locke holds that knowledge requires the ‘perception’ of relations between one’s ideas; but, Green asks, how does the mind come by these relational perceptions? For Locke this is ultimately a matter of reflection on one’s ideas of sensation, and it is his account of this which provides Green with his target. Perceptual knowledge of the presence of a red cup here and now, for Locke, will employ the general ideas of redness and cuphood, which relate this perception to others; but what account does Locke provide of these general ideas? Locke claims that they are reflective abstractions of what there is in common among a series of similar ideas, but Green objects that in order to be able to abstract a general idea from a series of ideas of sensation, most of which are now past and gone, I must already have an idea of what they have in common. Green puts his criticism in the following way: 17. It is not difficult to show that to have a simple idea, according to Locke’s account of it, means to have already the conception of substance and relation, which are yet according to him ‘complex and derived ideas’, ‘the workmanship of the mind’ in opposition to its original material, the result of its action in opposition to what is given it as passive. (Green 1885, 12–13) 24. The fact is that the ‘simple idea’ with Locke, as the beginning of knowledge, is already, at its minimum, the judgment, ‘I have an idea different from other ideas, which I did not make for myself.’ (Green 1885, 19)
How do Green’s criticisms of Locke introduce idealism? For Green nothing can be thought or meant or experienced without being in some way related to other things, and relating is an activity of mind (or ‘spirit’, Green’s preferred term). Green infers from this that whatever is an object of a mental state is dependent upon ‘spiritual’ activity. Kant of course hypothesised the existence of things-in-themselves (the Ding-an-Sich) to provide an independent source of what is given to sensation. But Green rejects the possibility that we can make sense of the existence of any such realm of things inaccessible to the mind and, once this is dropped, what
xxvi
Editors’ introduction
is left is the claim that all reality is dependent upon the activity of ‘the spiritual principle in nature’ (Green 1884, 59), a conclusion which is clearly idealist. Green’s conception of reality’s dependence upon the spiritual principle in nature brings him to the brink of monism (Green 1884, 104). But it is Bradley who explicitly affirms this monistic implication in Appearance and Reality (Bradley 1893). Although things appear to be several, distinct but related, reality cannot be fundamentally relational; and if there are no relations, then there is just one all-encompassing whole. Green took the unifying role of mind in constituting a relational reality as foundational, but Bradley went further into speculative metaphysics by denying the reality of relations. Bradley’s main argument against relations is that they involve a vicious regress (Bradley 1893, ch. III): where two wholly separate things are thought of as related by a relation, the unified character of the resulting relationship requires a second-order relation to relate the first-order relation to the things to which it is related; and that relationship requires in turn another relation which relates that relation to its relata; and so on. So relational being is not self-sufficient and makes sense only as abstracted from something more comprehensive whose unity is non-relational. It follows that relations are not ‘real’, since for Bradley ‘real’ implies ‘unconditional’. Metaphysics is the inquiry into unconditional reality of this kind, and it leads to the conception of the Absolute as the one independent reality, which encompasses everything in a form sufficiently ‘transmuted’ to become harmonious with everything else (Bradley 1893, ch. XIV). It is then Bradley’s Absolute which Moore implicitly invokes in his 1897 dissertation to provide a conception of ‘Reality’ which can serve as a ground for empirical appearances, including human life. As Moore recognises, ‘Reality’ so conceived is not a Kantian Ding-an-Sich, and his description of it is recognisably Bradleian: On this view, therefore, it is unnecessary to deny that the Real World appears to Intuition – our own experience. We can only deny that it appears as it is . . . There is no longer any need for conceiving Reality as external to all particular Appearances in the same way in which one Appearance is external to another – a false conception, which seems to have led Kant to call the Reality a cause. It is, indeed, their ‘ground’; but that relation is to be conceived not merely like that of formal logic nor like that of cause and effect, but as something between the two. The Reality is not an Individual separated from particulars as they are from one another, not yet a mere universal from which they might be deduced; it is an Individual both implied and existing in them. (1897 dissertation, p. 35)
Editors’ introduction
xxvii
Moore goes on to suggest that this absolute Reality provides the best way of making sense of Kant’s conception of ‘Transcendental Freedom’, though he acknowledges that Kant himself would not accept this account of his position (see vii.2 for discussion of this issue): The answer, then, to the question what Kant means by Transcendental Freedom is this. Transcendental Freedom is the relation in which the world as it really is stands to events as we know them. It is the relation of Reality to Appearance. This relation necessarily appears to us as the logical relation of reason to consequent. The reason is free cause of its consequence. (1897 dissertation, p. 35)
Bradley himself does not connect the Absolute with Freedom in this way. But he does connect it with goodness – ‘In a sense, therefore, the Absolute is actually good’ (Bradley 1893, 412) – though he has also to allow that, because the Absolute includes all phenomena, including bad actions, ‘it manifests itself throughout in various degrees of goodness and badness’ (Bradley 1893, 411). Moore postulates much the same connection in his 1897 dissertation, with the same baffling complications: Appearance may partake more or less of Reality, and thus gives rise to the differences in the categories by which its relation to Reality is expressed. It is thus that evil, though, like good, it is only possible through the union of Reality with Appearance, yet expresses less, than good, of the nature of Reality and more of mere Appearance. The object which is said to be good and that which is said to be real, are identical, if that object is taken to be Reality, as such, and it is not bad, only because ‘good’ is more adequate to that object than bad. (1897 dissertation, p. 83)
A year later Moore remarks in the Preface to his 1898 dissertation that he no longer shares Bradley’s general philosophical attitude: ‘I have come to disagree with him on so many points, and those points of importance, that I doubt if I can name any special obligations’ (p. 117). It is indicative of this changed attitude that Moore eliminates all talk of a ‘Reality’ which embraces all appearances from his dissertation. All the passages quoted above are dropped, and although Moore continues to discuss the conception of a ‘Transcendental Freedom’ which grounds all empirical action, his attitude now is sceptical: ‘Transcendental Freedom, upon which the main stress of the theory lies, is possible, as a chimaera is possible, but in no other sense’ (p. 209). In particular, therefore, all discussion of the relation between goodness and ‘Reality’ disappears. But not only does Moore in this way seek to purge his philosophy of Bradley’s Absolute idealism, he also introduces in chapter ii of the 1898 dissertation a critical discussion of Bradley’s account of judgement. Although in his account of his dissertations Moore highlights this discussion as the
xxviii
Editors’ introduction
central point of his new dissertation (see i above), it is in some respects a somewhat unsatisfactory episode. Moore focuses on the opening pages of Bradley’s The Principles of Logic (Bradley 1883) where Bradley is concerned to reject empiricist conceptions of ideas and judgement. To this end he distinguishes between the conception of an idea as a particular ‘psychical state’, or ‘sign’, with its own properties, and the logical conception of an idea as what is signified, or meant, by such a sign, and he takes it to have been a fundamental error of empiricism to confuse these two, by treating universal meanings as if they were just very thin abstract particular psychical states (Bradley 1883, 6). Thus far Bradley is in fact largely following in Green’s footsteps; Bradley, however, does not continue by introducing Green’s Kantian conception of the mind’s synthetic activity to explain how signified meanings are derived from signifying psychical states. Instead, he just says: A sign is any fact that has a meaning, and meaning just consists of a part of the content (original or acquired), cut off, fixed by the mind, and considered apart from the existence of the sign. (Bradley 1883, 4)
It is this passage which attracts Moore’s attention; he argues (1898 dissertation, p. 164) that so far from avoiding the errors of the old empiricist theory of ideas, Bradley has just repeated them. For, Moore argues, I cannot ‘cut off’ part of the content of a signifying psychical state unless I can identify that part, which presupposes that I already grasp the signified meaning which is supposedly constituted by ‘cutting it off’ from the total content of the original psychical state. In effect, Moore argues, Bradley has reverted to the empiricist doctrine of ‘abstraction’ according to which the signified universal meaning is just an abstraction, ‘cut off’ from some particular signifying states (1898 dissertation, p. 163). If Moore’s interpretation of Bradley were correct, the objection would be appropriate; and it is fair to note that Bradley does write of the signified meaning as ‘an adjective divorced, a parasite cut loose, a spirit without the body seeking rest in another, an abstraction from the concrete’ (Bradley 1883, 8). None the less this interpretation is so implausible in the context of Bradley’s assault on the empiricist conception of ideas that it must be wrong, despite Bradley’s suggestive language. And in this case we have Bradley’s word for it: for once Moore had published this part of his 1898 dissertation as ‘The Nature of Judgment’ in Mind (Moore 1899; see viii.2 below), Bradley wrote to Moore to explain that Moore had misunderstood him:
Editors’ introduction
xxix
The first [argument] seems to be that the separation of meaning from existence required for judgment presupposes a previous judgment. Well certainly it may do so – a psychological judgment, that is, but then again it may not and often does not . . . I suppose that my phrase ‘cut off’ etc. has been taken to imply a going about to cut off and therefore a previous idea. I never meant this . . . But I admit my language was loose.29
Things would indeed have been clearer had Bradley explained what he meant by ‘cutting off’, and it is remarkable that neither here, nor elsewhere, does he provide an extended account of the meaning of signs. Perhaps he wanted to avoid tangling with Green’s complex neo-Kantian theory of the synthetic unity of consciousness, and had no simple alternative to offer. Moore, of course, did not see things this way: his view was that once one had understood what is wrong with Bradley’s account, one should infer that meanings, or ‘concepts’ as he calls them, are irreducible non-mental entities which: may come into relation with a thinker; and in order that they may do anything, they must already be something. It is indifferent to their nature whether anybody thinks them or not. They are incapable of change; and the relation into which they enter with the knowing subject implies no action or reaction. (1898 dissertation, p. 165)
Moore’s concepts are the constituents of ‘propositions’, and we now turn to the background to the thesis he introduces at the start of chapter ii of the 1898 dissertation, that ‘that which we call a proposition is something independent of consciousness, and something of fundamental importance for philosophy’ (1898 dissertation, p. 162). v.2 mental science
Moore was strongly attached to robustly objectivist intuitions about the nature of ethics (ethics is a science, not an art, as he claims in the preface of the 1897 dissertation). His intellectual development at this period can be seen as a path toward the discovery, formulation and application of a metaphysics that would help to support those intuitions. Moore was ready and willing enough to adopt aspects of the views dominant in his orbit at Cambridge so long as he thought that these would contribute substantively to his steadily sharpening convictions. Thus, as we have seen, Bradley’s 29
Bradley’s letter is preserved in the Moore archive at Cambridge University Library: Add. MS 8830 8B/21/1. For further discussion of Moore’s criticisms of Bradley’s account of ideas, see Baldwin 1990, 14–15.
xxx
Editors’ introduction
Absolute idealism could and did play a brief (if seemingly paradoxical) role in providing a metaphysical foundation for an early version of Moore’s ethical objectivism, by providing a non-subjective formulation of Reality as the ground of appearances. But it made demands that were increasingly unattractive to Moore; so it was discarded once Moore had found his way to a preferable alternative. In this search Moore was assisted by the quickly evolving shift in views about the nature of judgement in the discipline then known as mental science. Moore’s 1942 acknowledgement to his teachers at Cambridge emphasises the influence of McTaggart (Moore 1942, 18), though Moore mainly stresses the influence of McTaggart’s method, crediting McTaggart’s Hegelianism principally as an improvement on the incoherence of Hegel himself. Apart from Ward, the person Moore credits most is Stout, and there is significant archival evidence of this influence. Moore, like Russell before him, attended Stout’s history of philosophy lectures in 1894, where the emphasis was on Kant and Lotze,30 and Stout thought well enough of the young Moore to encourage him to speak at meetings of the Moral Sciences Club, even when Moore was the only student present.31 Stout was a figure of some intellectual authority both at Cambridge and in British philosophy at this period, not least as editor of Mind, a position he held from 1892 (just before Moore arrived at Cambridge) until 1921. Although Russell called Stout a Hegelian (Russell 1959, 30), this is misleading: Stout’s work is a useful signpost to the progressive emergence of twentieth-century philosophy of mind from nineteenth-century mental science.32 The metaphysical demarcations in this discipline were not always consistent, as the psychologically minded philosophers fought to account for a study of mind that would avoid collapse either into metaphysical mentalism or physical reductionism. Central to this intellectual debate was the Psychologismus-Streit, which took hold on the continent, mainly in Germany, Austria and Poland.33 The dispute over the nature of logic and reasoning inevitably developed into a dispute over the nature of thought, which drew the discipline of psychology, by now in a fully extended advance toward the status of empirical science, into the fray. The distinction between the act of thought and its objects became the focus of the literature. Stout (and Ward) had both studied in 30 31 32 33
The notebooks survive (Add. MS 8875 10/1/1). Moore to parents, 4 February 1894 (Add. MS 8330 2/1/27). Stout was not, however, a member of the Apostles Society. See Griffin (1991); Preti (2008); Van der Schaar (1996). See Kusch (1995).
Editors’ introduction
xxxi
Germany, and were thoroughly familiar with the literature in Germanlanguage philosophy and psychology. Stout further developed his own views, largely influenced by that literature, in his Analytic Psychology (Stout 1896).34 Analytic Psychology is an extended discussion of what Stout calls the ‘Scope and Method of Psychology’. It includes a clear commitment to an empirical science of psychology which distinguishes it from both metaphysics and physiology. This conception of psychology is due to Brentano, whose Psychology from an Empirical Standpoint (Brentano 1874) was deeply influential on the discipline.35 Brentano had by this time introduced the distinction between ‘descriptive’ and ‘genetic’ approaches in psychology, defending the former, which Stout approvingly cites as properly focusing on the nature of consciousness, and not its origin (1896, vol. i: 30, 40).36 Stout is sometimes held to have espoused an undiluted Brentanianism in Analytic Psychology, but he does not rely on Brentano alone to supply the criteria for the study of consciousness that he takes to be central to psychology: among others, he mentions Ward (who fought in vain for years for a psychological laboratory at Cambridge) and Herbart (who along with Ward had earlier defended an act/object distinction for judgement). Brentano had proposed that the object of judgement was a form of representation (an ‘intentional inexistence’), a formulation that appeared to his critics to render the object of judgement too psychological in nature.37 Following Twardowski (1977), Stout modified this view by distinguishing the content of judgement from the object of judgement, on grounds that (i) the object may have properties that the content of my judgement about it will not have and (ii) my judgement always has content (is about something), even if what it is about does not exist. On this view, although it is the content of judgement that determines, in a given context, its object, 34
35
36
37
Russell read Analytic Psychology when it was published (Griffin 1991, 34). Moore also appears to have known it well (Preti 2008); his psychology lectures after 1911 involved close examination of Stout’s work (Moore 1942; Preti 2008). See Schumann (in Jacquette 2004: 277–97); Jacquette (2004: 1–19). Stout’s title is taken from Brentano’s conception of a ‘descriptive psychology’. But see Van der Schaar (1996) for critical differences between Stout and Brentano. The distinction between descriptive and genetic psychology is mirrored in Moore’s comment in the 1898 dissertation concerning propositions: ‘We are not concerned with its origin, but with its nature’ (p. 167). Stout (1896, i: 439–41). One point that bears repeating is that Brentano’s noted connection to contemporary formulations of propositional content as intentional can obscure the point that he himself did not argue either for a propositional or a realist formulation of objects of judgement.
xxxii
Editors’ introduction
the full specification of a judgement requires a specification of both content and object.38 This, then, is the position that Moore will have encountered in Stout’s writings and in other discussions of judgement when he was preparing his dissertations. Moore does not refer to Stout in his dissertations; but the position he takes in chapter II of his 1898 dissertation can be read as a reaction against Stout’s tripartite act/content/object distinction. For what is most remarkable about the position Moore presents here is not the quasi-Platonist conception of concepts which he proposes as an alternative to Bradley’s conception of meaning. By itself that move would fit readily within Stout’s tripartite schema, with Moore’s concepts providing a nonpsychologistic account of the ‘content’ of judgement, without implying anything about the ‘object’ of judgement. Yet what is astonishing is the move which Moore then makes, which is, in effect, to collapse Stout’s content/object distinction by taking it that ‘All that exists is thus composed of concepts necessarily related to one another in specific manners, and likewise to the concept of existence’ (1898 dissertation, p. 167). On this new view, truth is inherent in true propositions, existents are true existential propositions and the concepts which constitute propositions are ‘the ultimate elements of everything that is’.39 Moore explicitly rejects the alternative option of affirming a correspondence theory of truth according to which a propositional ‘content’ is true where it corresponds to a state of the world composed of non-conceptual ‘objects’. Instead he embraces a view of the world ‘as formed of concepts’. Moore does not explain clearly why he makes this move. He remarks concerning this new theory, ‘It will be apparent how much this theory has in common with Kant’s theory of perception’, though he goes on to add ‘It differs chiefly in substituting for sensations, as the data of knowledge, concepts; and in refusing to regard the relations in which they stand as, in some obscure sense, the work of the mind’ (1898 dissertation, p. 169). This suggests that Moore’s route to his position has a Kantian starting point, in Kant’s theory of perception (at least as this theory was then understood by Moore), although he then distances himself from the Transcendental Idealism that Kant himself would have regarded as fundamental to his 38
39
Stout (1896). For Twardowski’s role, see Van der Schaar (1996, 300). There are obvious similarities between Twardowski’s content/object distinction and Frege’s sense/reference distinction, but the distinctions should not be assimilated since different considerations are used to justify and apply them. As Moore put it in his letter to Russell of 11 September 1898 which is reproduced below p. xxxv.
Editors’ introduction
xxxiii
theory of perception. But this explanation does not really provide much of a justification of the new theory. The most likely suggestion is that it was Moore’s determination to purge the nature of judgement of any hint of psychologism that is responsible for his rejection of the tripartite schema. Certainly in ‘The Refutation of Idealism’ (1993b, 38–9) he argues that the basic error of idealism has been that it treats the ‘object’ of sensation as if it were part of its ‘content’, where, following Bradley’s account of the content of an idea, the ‘content’ of a sensation is understood as comprising its intrinsic qualities. Thus a similar account of the content of judgement would imply that this content is a purely qualitative aspect of the judgement, the psychological act; and if it were held that this content determines the judgement’s ‘object’, which for Moore is the proposition judged, the resulting position would be strongly psychologistic. In fact it need be no part of those who uphold a content/object distinction that the content of judgement is in this way a psychological quality of the judgement.40 But we suspect that Moore thought otherwise, at least at this time, and that this belief lies behind his adoption of a simple act/object conception of both judgement and sensation.41 However that may be, the results of Moore’s new position are startling. In effect a theory of judgement turns into a metaphysics, thanks to the role of propositions as (nearly) ‘ultimate’ entities, both objects of thought and elements of existence: Even the description of an existent as a proposition (a true existential proposition) seems to lose its strangeness, when it is remembered that a proposition is here to be understood, not as anything subjective – an assertion or affirmation of something – but as the combination of concepts which is affirmed. (1898 dissertation, p. 168) . . . thus, in the end, the concept turns out to be the only substantive or subject, and no one concept either more or less an adjective than any other. From our description of a proposition, there must, then, disappear all reference either to our mind or to the world. Neither of these can furnish ‘ground’ for anything, save in so far as they are complex propositions. The nature of the proposition is more ultimate than either, and less ultimate only than the nature of its constituents – the nature of the concept or logical idea. (1898 dissertation, pp. 173–4)
40 41
As the comparison with Frege’s sense/reference distinction indicates: Frege strongly rejected psychologistic conceptions of sense. For a more speculative explanation which starts from Bradley’s conception of truth as identity, see Baldwin (1990), 43–4.
xxxiv
Editors’ introduction
Despite Moore’s comment here, his position is ‘strange’, but also exciting and potentially liberating. As the passages quoted at the start of this section from Russell indicate, that is how Russell reacted to it, though there may have been an element of exaggeration in Russell’s account (Griffin 1991 306–8). None the less Russell always felt that Moore’s line of thought here, which formed the core of the paper ‘The Nature of Judgment’ which Moore extracted from the 1898 dissertation (see viii.2 below), changed the world. Here is a passage from Russell’s obituary notice for Moore: It was towards the end of 1898 that Moore and I rebelled against both Kant and Hegel. Moore led the way, but I followed closely in his footsteps. I think that the first published account of the new philosophy was Moore’s article in Mind on ‘The Nature of Judgement’ . . . I still think that this article gave conclusive proof of philosophical genius.42 Addendum: Moore’s letter to Russell about the 1898 Dissertation43 Cheriton, Cockington, Torquay. Sept. 11th , ’98 Dear Russell, Almost all the addition to my dissertation was metaphysical, so that the whole does not hang together at all well. I have found a great difficulty too, in fitting in my last year’s work, since that was directed, with a tolerable unity of purpose, to shewing that Transcendental Freedom might be taken as an attempt to describe the relation between Reality and Appearance; which I cannot claim for it now, since I see no proof that there is any existent reality, beside Appearance. However, I see no reason for thinking that the quality of the new matter is not as good as that of the old; and there is quite as large a quantity as I had hoped. Of course I think it is more correct; but I am much less proud of it than I was last year, having begun to see how a really thorough account of Kant could be written, and the sort of exhaustiveness that might be put into original writing. Moreover, I fear the present dissertation is much more paradoxical than the last, and will deprive me completely of Caird’s sympathy. My chief discovery, which shocked me a good deal when I made it, is expressed in the form that an existent is a proposition. I see now that I might have put this more mildly. Of course by an existent must be understood an existent existent – not what exists, but that + its existence. I carefully state that a proposition is not to be understood as any thought or words, but the concepts + their relation of which we think. It is only propositions in this sense, which can be true, and from which 42 43
The Times, 28 October 1958. This letter is held in the Russell archive at McMaster University.
Editors’ introduction
xxxv
inference can be made. Truth therefore does not depend on any relation between ideas and reality, but is an inherent property of the whole formed by certain concepts and their relations; falsehood similarly. True existential propositions are those in which certain concepts stand in a specific relation to the concept existence; and I see no way of distinguishing such from what are commonly called ‘existents’, i.e. what exists + its existence. This explains how it should commonly be thought that a proposition can be inferred from an existent. Existents are in reality only one kind of proposition. The ultimate elements of everything that is are concepts, and a part of these, when compounded in a special way, form the existent world. With regard to the special method of composition I said nothing. There would need, I think, to be several kinds of ultimate relation between concepts – each, of course, necessary (I shall read on this to the Aristotelian [Society] on Dec 9th ). I haven’t looked at your work yet, having had a cold & been doing nothing since I came down. I must begin to stir again now, doing Classics for the exam, and having to begin lecturing for the School of Ethics on Oct. 13th . Yours fraternally G. E. Moore v.3 sidgwick
In the preface to the 1898 dissertation Moore wrote: For my ethical views it will be obvious how much I owe to Prof. Sidgwick. I could not have devoted so large a space to criticizing his views on certain points, had he not supplied me with much of the common ground on which I take my stand. (1898 dissertation, p. 117)
What Moore says here is right; he owes the basic orientation of his ethical theory to Sidgwick – in particular the thesis presented in the introduction that ethics requires its own metaphysics to elucidate its fundamental concept and, further, that this fundamental concept is not, as Kant held, the moral law which prescribes what ought to be done, but the good, which defines what ‘ought to be’ (though this point needs some qualification – see below). As Moore also implies here, there remained many points on which he differed from Sidgwick, though in the 1898 dissertation Moore had moved closer to Sidgwick than he had been a year earlier in the 1897 dissertation, where his idealist sympathies had led him to conclusions altogether alien to Sidgwick. This difference is reflected in the preface: instead of the deferential passage quoted above, in 1897 Moore had adopted a more critical stance: Finally, with regard to the frequency with which I have introduced Professor Sidgwick’s views, and that almost exclusively for criticism, I think it fair to state that I
xxxvi
Editors’ introduction
have criticised him precisely because he seems to me to have discussed the difficulties of ethics with such extraordinary care and acuteness, and because it is to his clear solution of many problems, which would almost certainly have embarrassed me, that I owe the possibility of having attempted, however unsuccessfully, to contest the finality of his conclusions. (1897 dissertation, p. 4)
Sidgwick, who was at this time a Fellow of Trinity College as well as Professor of Philosophy at the University of Cambridge, had been one of Moore’s teachers during his undergraduate studies. Moore records that he found Sidgwick’s lectures ‘rather dull’ (Moore 1942, 16), and he does not seem to have gained much intellectual excitement from his personal contacts with Sidgwick at this time. Moore’s papers include several undergraduate essays which he wrote for Sidgwick on topics such as ‘Egoism and altruism’ in which the young Moore takes issue with Sidgwick’s account of practical reason. Moore’s Hegelian enthusiasm at this time (no doubt inspired by McTaggart), and Sidgwick’s reaction to it, are visible in an essay on the ‘Relation of reason and moral action’.44 Moore writes: Thought ‘determines’ itself into various forms of thought; it is at least given to us as a unity in difference: as we cannot say that its aspect as unity, its definition, has any truth at all, unless it can also be shewn that its different forms are necessarily connected with that unity, i.e. unless thought, as abstract, ‘determines’ and is determined by particular forms of thought – unless they involve one another.
Alongside this passage Sidgwick has commented in the margin ‘metaphysics into which I cannot enter’! This intellectual distance between Moore and Sidgwick was probably accentuated by the difference in age between them: Moore was twenty-two whereas Sidgwick was by then in his late fifties and not in good health. In fact he resigned on grounds of ill health in 1900 and died that year, aged sixty-two. Sidgwick’s major philosophical work was The Methods of Ethics, first published in 1874 (Sidgwick 1874). Sidgwick attempted here to combine the insights of the two main traditions of English moral philosophy, utilitarianism and intuitionism, which were usually regarded as opposed to each other. Although he also says a little about Kant’s moral philosophy, he does not attempt to address it fully; and he says even less about Kant’s successors. Sidgwick attempted to keep his book up to date in relation to contemporary ethical debate in Britain by bringing out a new edition every five years or so (the last significantly changed edition is the sixth edition of 44
The essay is in Cambridge University Library: Add. MS 8875, 11/2/5.
Editors’ introduction
xxxvii
1901).45 But in one respect ethical debate in fact moved beyond Sidgwick’s book. The most important contribution to ethics by the British idealists was T. H. Green’s Prolegomena to Ethics (Green 1884) which we have already discussed briefly (see v.1). Green here propounded a perfectionist synthesis of the ethical theories of Aristotle and Kant while also importing themes from his idealist metaphysics concerning ‘the self-realising subject’ and this book became for a time the most influential ethical work of the period. Although Sidgwick, who had known and admired Green personally, alludes briefly to Green’s Prolegomena in later editions of The Methods of Ethics, he did not deal with it properly there. But he discussed it later in lectures on Green (subsequently published)46 which are very critical of Green’s Prolegomena. Moore attended these lectures as an undergraduate, and even if he found these lectures dull, it may none the less be that Moore had been sufficiently impressed by Sidgwick’s criticisms of Green never to be attracted by Green’s project even when he was himself both sympathetic to idealism and discussing Kant’s ethics. It is certainly striking, and in retrospect disappointing, that Moore never took the trouble to discuss Green’s work at any length. There is a brief discussion in Principia Ethica §84, but it does not do justice to Green’s position. It would have been interesting to see what Moore had to say in the projected, but unwritten, second chapter of the 1897 dissertation in which he had planned to discuss Green’s ethics. Sidgwick was of course one of Moore’s examiners in 1897. In his brief report Sidgwick chooses not to say much about Moore’s discussion of his own work and concentrates instead on Moore’s discussion of Kant ‘as I am able to judge this more impartially’, though he goes on to add ‘I may say generally that the part relating to me has certainly the same merits and perhaps the same defects as the rest of his initial work’ (p. 97). The main ‘defects’ Sidgwick identifies in Moore’s dissertation concern Moore’s failure to grasp the ways in which Kant modifies his position in the writings he discusses and the obscurity of his own positive response to Kant. By contrast the main ‘merits’ of Moore’s dissertation are that it is ‘first-rate, in respect of critical acumen and dialectical vigour, and independence of thought’ (p. 97). Hence, Sidgwick concludes, the dissertation’s chief merit ‘seems to me to lie in promise rather than performance: but I judge it to be very promising’ (p. 98). Presumably, therefore, Sidgwick felt that 45
46
There is a seventh edition of 1907, which is indeed the edition most readily available today. This was published long after Sidgwick’s death and differs from the sixth edition only by correcting some minor errors. Sidgwick’s lectures were edited by Constance Jones, who was Sidgwick’s closest Cambridge disciple, and published posthumously (Sidgwick 1902).
xxxviii
Editors’ introduction
in some respects Moore had misunderstood his own position; but, equally, he was prepared to acknowledge the ‘vigour’ and ‘promise’ inherent in Moore’s critical discussion of his views. One point on which Moore misunderstood Sidgwick concerns the thesis that he was to make famous in Principia Ethica, that the fundamental ethical concept ‘good’ is ‘simple’ and ‘unanalysable’: in the 1898 dissertation Moore both affirms this thesis and attributes it to Sidgwick (1898 dissertation, p. 178; the 1897 dissertation does not contain a similarly explicit affirmation of this thesis, but it is implicit in the introduction). Yet if one goes back to Sidgwick’s discussion of ethical judgements in The Methods of Ethics (book 1, ch. III) where Sidgwick is addressing the question of whether ethical judgement involves a distinctive ‘fundamental notion’ of its own, one finds that although Sidgwick does indeed affirm that there is a fundamental ethical notion which is ‘ultimate and unanalysable’ (ME 34),47 he never uses the word ‘good’ in this connection; instead he generally writes of what is ‘right’ (ME 33). Indeed, not only does Sidgwick not affirm that ‘good’ is unanalysable, he actually proposes a definition of it in terms of desire: Similarly, ‘ultimate good’ without qualification must be taken to mean what rational beings as such would desire, assuming them to have an impartial concern for all existence. (ME 112)48
Admittedly (as Sidgwick makes clear in the sixth edition of ME) the reference here to ‘what rational beings as such would desire’ is to be understood as including an implicit reference to Sidgwick’s fundamental ethical notion; none the less it remains the case that Sidgwick does analyse ‘good’, and thus that Moore’s attribution to him of the thesis that ‘good’ is unanalysable is mistaken. Moore devotes most attention to Sidgwick in the context of his discussion of free will. Indeed in the 1897 dissertation (p. 20; cf. 1898 dissertation, ch. iii, p. 181) Moore starts straight off from Sidgwick’s account of the free will debate in order to set the terms for his discussion of Kant; and he then returns at length to Sidgwick’s attitude to this debate when discussing the libertarian conception of the ‘liberty of indifference’ of the radically free will (1897 dissertation, pp. 39ff.; cf. 1898 dissertation, ch. iv, pp. 212ff.). In The Methods of Ethics Sidgwick had not of course attempted to defend a libertarian position; he makes it clear that his 47
48
This phrase comes from the fifth edition of ME; it is omitted from the comparable passage in the seventh edition, where Sidgwick just says, concerning ‘right’ and ‘ought’, that ‘the notion which these terms have in common is too elementary to admit of any formal definition’ (p. 32). Sidgwick advances essentially the same definition in the seventh edition, p. 112.
Editors’ introduction
xxxix
sympathies lie instead with the determinist position (see ME 62–5). But his main aim had been to demonstrate the ‘practical unimportance’ of the Free Will controversy (ME 75), and thus demonstrate that for the purpose of ethical theory, as opposed to metaphysics, it is not necessary to take a position on it. In his dissertations Moore takes issue with Sidgwick on this very point. For Moore claims that the libertarian thesis threatens the possibility of a serious ethics by making it dependent on the ‘metaphysical monstrosity’ of the conception of the human will as an uncaused cause (1897 dissertation, p. 39; 1898 dissertation, p. 212). Moore’s justification for this claim comes directly from Kant: Moore takes it that in the ‘Transcendental Analytic’ of his Critique of Pure Reason Kant has demonstrated ‘the a priori certainty of the Deterministic view’ and, he remarks, ‘it seems inevitable to agree with him’ (1897 dissertation, p. 41; 1898 dissertation, p. 213). But, Moore acknowledges, it is not enough to leave the matter there; it is also necessary to undermine the familiar consideration in favour of the libertarian position whose strength even Sidgwick had acknowledged (while not endorsing the position) – namely the ‘affirmation of consciousness’ that, whatever one’s inclinations and past bad habits, one can always step back from one’s past and choose to do what one conceives to be ‘right or reasonable’ (ME 65). Moore begins his critical appraisal of this line of thought in a rather oblique way by complaining that Sidgwick’s formulation of the libertarian position places too much emphasis on the capacity to choose that which is conceived to be ‘right or reasonable’. Moore objects that the libertarian conception of free will must also be one that can be exercised in choices of what one recognises as ‘wrong or unreasonable’, and thus that the key question is whether any weight is to be placed on an affirmation of consciousness ‘when alternatives are presented, that I can choose any of them that I think either good or bad’ (1897 dissertation, p. 49; 1898 dissertation, p. 222). When reading these criticisms of his position, Sidgwick might well have thought that Moore was misrepresenting him; after all, as Moore also notes, he himself says that the Free Will controversy revolves around the question whether ‘my voluntary action, for good or for evil, is at any moment completely caused’ (ME 62). None the less it is perhaps significant that Sidgwick did alter several of the passages Moore criticises when preparing the sixth edition of The Methods of Ethics (see 1897 dissertation, p. 41 notes d and 30). Still, what is much more interesting is the move that Moore then makes. Where Sidgwick deliberately leaves the question of the merits of the
xl
Editors’ introduction
libertarian’s ‘affirmation of consciousness’ unresolved, Moore argues that the positive affirmation that ‘I can choose’ does not have the significance for the libertarian position which it has often been supposed to have. For to say ‘I can choose so-and-so’, he says, is to say that one does not know that one will not choose so-and-so and, he argues, even if in some cases on determinist grounds one can know that an event will not happen, one can never have grounds for knowledge of this kind concerning a choice of one’s own since it is always possible for it to be ‘conditioned by a conscious forecast’ (1897 dissertation, p. 50; 1898 dissertation, p. 222). This claim rests on the thesis that because one’s knowledge of a prediction of one’s choice is always liable to make a difference to one’s actual choice, ‘it [is] possible to pronounce a priori that complete prediction of the results of mental process must always be impossible’ (1897 dissertation, p. 40; 1898 dissertation, p. 213). It follows from this that, even if determinism is true, one cannot oneself predict one’s own choices; hence the ‘affirmation of consciousness’ that choice of an alternative course of action always remains possible does not refute determinism. As Moore puts it: ‘the “affirmation of consciousness” as against Determinism disappears on the attempt to make it precise’ (1897 dissertation, p. 50; 1898 dissertation, p. 223). In this argument Moore relies on the thesis that the ‘can’ of ‘I can choose’ is epistemic – I do not know that I will not choose. Despite Caird’s suggestion in his examiner’s report that this affirmation of the unpredictability of choice confirms the validity of the libertarian’s conception of the free will (p. 106), Moore’s argument is that the affirmation is warranted only in so far as its content is epistemic and not metaphysical; hence it does not support the libertarian’s ‘metaphysical monstrosity’. Moore later returned to this point at the end of chapter vi (‘Free Will’) of Ethics.49 Significantly, however, he does not introduce it when discussing here the significance of the ‘could’ of ‘I could have done otherwise’, where the possibility is, on the face of it, metaphysical. It is only when he has argued that this possibility is compatible with determinism because it is the conditional possibility that one would have done otherwise if one had so chosen that he introduces epistemic possibility in order to show how the further requirement that we could have so chosen is also compatible with determinism.50 In his discussion of freedom Moore is not, of course, primarily concerned to assess Sidgwick’s treatment of this matter; his main target is 49 50
Moore (1912), esp. 220–1; reprinted in Moore (1993a), 310. Much later Moore was to discuss this conception of epistemic possibility in his 1941 Howison lecture ‘Certainty’, first published in Moore (1959) and reprinted in Moore (1993b); see esp. 176–7.
Editors’ introduction
xli
Kant, and we will therefore return to Moore’s discussion of freedom when discussing Moore’s approach to Kant in these dissertations. But there is one topic where Moore’s target is Sidgwick, and Sidgwick alone – hedonism. Moore’s main discussion of this topic occurs in the appendix on Professor Sidgwick’s hedonism which he attached to the 1897 dissertation and, so far as we can tell, repeated unchanged as an appendix to the 1898 dissertation. Before introducing Moore’s appendix, however, it is necessary to say a little about Sidgwick’s treatment of this topic. Sidgwick’s commitment to hedonism was theoretical; his thesis in The Methods of Ethics was that the best way to organise our moral convictions is to take it that pleasure, or happiness (Sidgwick makes no significant distinction here), is the only thing of ultimate value. As Sidgwick was well aware, that thesis had often been given an egoist slant, to the effect that the only thing of ultimate value for oneself is one’s own happiness; but Sidgwick was very clear in distinguishing between the egoistic hedonism which enjoins each of us to maximise our own happiness and ‘universalistic hedonism’, which requires us to promote to the greatest degree possible ‘universal happiness’, which Sidgwick describes as ‘desirable consciousness or feeling for the innumerable multitude of sentient beings, present and to come’ (ME 404). Having made this distinction, however, Sidgwick notoriously found it difficult to identify any decisive a priori arguments in favour of universalistic as opposed to egoistic hedonism (ME 420–2). But he also held that, unlike egoistic hedonism, universalistic hedonism does at least provide a theoretical framework within which our ordinary moral convictions and practices can be accommodated, and thus that our deep-seated commitment to these convictions and practices provides us with a reason for endorsing universalistic, rather than egoistic, hedonism (ME 496–500). In his dissertations Moore accepts Sidgwick’s universalistic conceptions of value and practical reason: right at the start of his appendix on Professor Sidgwick’s hedonism he comments that: In so far as the Intuitional method is held to establish a Universalistic doctrine, I fully agree; in fact, I should go further than Professor Sidgwick, in as much as I cannot see that Egoistic Hedonism has any claim to an independent validity. (1897 dissertation, p. 87)
But he goes on to add: I disagree only in so far as it is a Hedonism that the Intuitional Method is said to yield as its result. (1897 dissertation, p. 87)
xlii
Editors’ introduction
Moore then seeks to substantiate this disagreement by a series of arguments. The first concerns a presupposition of Sidgwick’s position, that when we seek to identify the things which are ultimately good, apart from human qualities, ‘we can find nothing that, on reflection, appears to possess this quality of goodness out of relation to human existence, or at least to some consciousness or feeling’ (ME 113). Thus, Sidgwick continues, beauty has no value out of relation to human existence: ‘no one would consider it rational to aim at the production of beauty in external nature, apart from any possible contemplation of it by human beings’ (ME 114). In his appendix Moore directly contradicts this claim: Now I must confess I do consider this rational. Even if it were absolutely impossible that any part of nature should ever be an object of contemplation either to myself or to anyone else, I should think it reasonable to prefer that it should be beautiful rather than that it should be ugly. (1897 dissertation, p. 88)
As Moore acknowledges (1897 dissertation, p. 89) he does not give any ‘strict argument’ for this point; but, he maintains, Sidgwick equally gives none on his side. Sidgwick might object that in the claim which Moore disputes he writes of the ‘production’ of beauty, and that it was primarily the rationality of producing beauty which no one could contemplate that he was rejecting, whereas Moore writes of its being reasonable to prefer the simple existence of a beautiful object to that of an ugly one, without reference to the rationality of producing the former where it is not possible that either should be contemplated. Yet although there is a distinction here, for the purposes of his overall argument, Sidgwick does need the stronger claim, that the bare existence of beauty ‘out of relation to human existence’ is of no value in itself; and Moore’s objection to this is intuitively potent, though the difficult issue here is whether the ethical value of the appreciation of beauty (which Sidgwick acknowledges) implies that the objects whose beauty is appreciated are of intrinsic ethical value in addition to their aesthetic value. Sidgwick inferred from his thesis that anything ultimately good is somehow related to human existence that ‘if there be any Good other than Happiness to be sought by man, as an ultimate practical end, it can only be the Goodness, Perfection, or Excellence of Human Existence’ (ME 115). This position is the starting point for his later discussion whose conclusion is that ‘the Intuitional method rigorously applied yields as its final result the doctrine of pure Universalistic Hedonism’ (ME 406–7). This discussion has two parts: first, Sidgwick suggests that if we align our thoughts to ‘the sober judgment of reflective persons’ (ME 400), we will intuitively
Editors’ introduction
xliii
recognise that putative goods such as ‘cognition of Truth, contemplation of Beauty, Free or Virtuous action’ (ME 400) are good only in so far as they lead to ‘the happiness of sentient beings’ (ME 401). Second, he argues that if we are to provide a ‘systematic answer’ (ME 406) to ethical questions, we must have a ‘common standard’ for comparing different values, which hedonism is uniquely placed to provide. In his appendix Moore disputes both these points: first, he rejects Sidgwick’s intuitive judgement: as far as his own sober judgement is concerned, ‘knowledge of the truth, so far from being good merely as a means to happiness, however great, is in itself better than happiness’ (1897 dissertation, p. 89). Second, he argues (1897 dissertation, p. 90), it is just not necessary that ethics be organised into the kind of theoretical system which Sidgwick assumes. It is better to rely on the ‘Intuitional method’ to assess the importance of different fundamental values than to reduce them all to a single system. In his first draft for the appendix Moore then added a third point (1897 dissertation, p. 94 note 52), namely that Sidgwick’s claim that his utilitarian system captured the morality of common sense was not right. Moore attempted to substantiate this criticism by the following example: our common sense morality implies that a man who is in serious danger of drowning has a duty to help a child unrelated to him who is in the same situation even if the chance of success is slight and the man thereby puts himself at greater risk. Moore argues that where it is assumed that the chance of either being saved is slight, it might be that if the man were to refuse to help the child, he would not thereby significantly diminish the sum total of happiness; and thus that a utilitarian would not be justified in condemning the man who behaves in this way. But, Moore asserts, in this situation ‘Common Sense . . . would seem to reprobate him utterly’. Despite Moore’s initial confidence about this case, however, on reflection it is not as decisive as it needs to be – it is not clear how far the duty to help others extends to a duty of potential self-sacrifice, and ‘common sense’ surely refrains from reprobation in situations of this kind. So it was sensible of Moore to omit this section of his draft, though the underlying point, that utilitarianism conflicts with common sense morality, seems right. Moore adds to these criticisms of Sidgwick’s ‘universalistic hedonism’ further criticisms of his account of pleasure which start from Sidgwick’s definition of pleasure as ‘a feeling which . . . is at least implicitly apprehended as desirable or – in cases of comparison – as preferable’ (ME 127). This definition implies that where pleasures are being compared that which
xliv
Editors’ introduction
is more pleasant is that which is ‘apprehended as preferable’. Moore objects to this: My intuitive judgment is, in fact, that I apprehend some feelings as pleasanter than others, and yet that I sometimes apprehend those very same feelings as less desirable, and that, considered in themselves, entirely without regard either to consequences or concomitants. (1897 dissertation, p. 93)
Moore certainly seems right about this – Sidgwick’s attempt to defend the view that the ‘higher’ pleasures of the intellect are more pleasant than simpler ones because of their ‘conditions and concomitants’ is not persuasive. Indeed, as Moore observes, if one combines Sidgwick’s definition of pleasure with his definition of good, his thesis that ‘Pleasure is the ultimate good’ means just that that which is apprehended as desirable is the only thing which is desirable for itself which, so far from being a fundamental intuitive truth, ‘seems, on the face of it, absurd; for it will hardly be maintained that an erroneous judgment is impossible’ (1897 dissertation, p. 93). Furthermore, since pleasure defined as that which is apprehended as desirable or preferable is plainly a subjective matter, Sidgwick’s universalistic hedonism implies that morality is basically a matter of finding principles which maximise the satisfaction of preferences. Moore takes this to be a singularly unattractive conclusion; and although thanks to rational-choice theory we are today familiar with sophisticated versions of this position, it remains hard to view it as a basis for morality in general, as opposed to acknowledging its value as an important tool for determining public policy. In the 1898 dissertation Moore adds a further criticism of hedonism. In his new chapter v (‘Ethical conclusions’) he endorses Plato’s argument in the Philebus that because pleasure by itself does not necessarily bring with it awareness of pleasure, Sidgwick’s thesis that pleasure is the ultimate good implies that the best life need include no awareness, memory or anticipation of pleasure (1898 dissertation, p. 236). Moore repeats this argument at greater length in Principia Ethica (§§52–3, 139–42); but he also acknowledges here that the hedonist can easily modify his position to avoid this objection by taking the ultimate good to be consciousness of pleasure rather than pleasure itself. Moore’s discussion of this point occurs in chapter iii of Principia Ethica which is devoted to a critical discussion of hedonism. The first part of this chapter includes Moore’s notorious discussion of Mill’s ‘proof’ of his principle of utility; but once he has dealt with Mill Moore turns to Sidgwick and develops some of the points of his early dissertations. He refines his objection to Sidgwick’s thesis that
Editors’ introduction
xlv
all ultimate goods must be somehow related to human existence into the famous two worlds argument (PE 135): where one world is ‘exceedingly beautiful’ and another is ‘simply one heap of filth’ (PE 135) ‘is it irrational to hold that it is better that the beautiful world should exist, than the one which is ugly?’; and he also criticises the way in which Sidgwick argues that, of the valuable things which are related to human existence, such as contemplation of beauty, only pleasure in the end should be counted as of ultimate value (PE 144–5). Despite these similarities between Principia Ethica and the earlier dissertations, however, the differences are equally striking. The most significant is that in Principia Ethica Moore is generally much more sympathetic to hedonism than before: he accepts that ‘Pleasure does seem to be a necessary constituent of most valuable wholes’ (PE 145); indeed, when discussing the value of beauty in chapter vi of Principia Ethica, he now writes: I have myself urged in Chap. iii (§50) that the mere existence of what is beautiful does appear to have some intrinsic value; but I regard it as indubitable that Prof. Sidgwick was so far right, in the view there discussed, that such mere existence of what is beautiful has value, so small as to be negligible, in comparison with that which attaches to the consciousness of beauty. (PE 237–8)
It is not clear what motivates this change. One consideration is perhaps the conception of intrinsic value employed in Principia Ethica as that which is ‘worth having purely for its own sake’ (PE 237), since this applies much more readily to states of consciousness than to worlds, however beautiful. A second reason is the introduction of the principle of organic unities, for Moore now relies on this principle a great deal both in his critical discussion of Sidgwick’s account of pleasure (PE 144–7) and in his positive account of the way in which pleasure contributes to the ‘most valuable wholes’ even though pleasure by itself is not of great intrinsic value. This point is a little surprising, for Moore’s principle of organic unities is clearly a residue within Principia Ethica of his earlier enthusiasm for idealism and one might have expected to find this principle employed in his dissertations. But in fact, although Moore does discuss the metaphysics of organic unities in his dissertations (1897 dissertation, p. 54; 1898 dissertation, p. 226) in the 1897 dissertation he does not give them any ethical significance. In the 1898 dissertation he does add a brief discussion in which he in effect states a restricted version of his principle – ‘two things, which would separately be comparatively worthless, may by a special conjunction become highly valuable as parts of that whole which they combine to form’ (1898 dissertation, p. 228). But even here he does not put this new principle to work in his
xlvi
Editors’ introduction
ethical discussion. It does begin to play a role in the account of the ‘Ideal’ which Moore propounds in his lectures on ‘The Elements of Ethics’ which he composed later in autumn 1898 after he had been awarded his Prize Fellowship.51 But it is not until Moore rethinks his ethical theory while revising these lectures in 1902–3 that the principle of organic unities comes to assume the central place in his thought that it occupies in Principia Ethica. VI MOORE’S DEBATE WITH KANT
Moore took Ward’s advice that he should try to write his Fellowship dissertation on Kant’s Ethics. In the end he gave his 1897 dissertation the title ‘The Metaphysical Basis of Ethics’, though it is primarily concerned with questions about the place of freedom in Kant’s metaphysics of ethics. As his preface indicates he had originally intended to deal with other aspects of Kant’s Ethics (his ‘attitude towards Hedonism and Practical Ethics’ – 1897 dissertation, p. 4); and he starts his 1898 paper ‘Freedom’ which is lifted from the 1897 dissertation (see viii, p. lxxix below) by announcing an even more ambitious intention: The present paper is selected from a much longer essay on Kant’s notion of Freedom, which I hope in future to rearrange and enlarge into a treatise on the whole of his Ethical Philosophy. (Moore 1898, 179)
During 1898, however, Moore’s interests changed, and although he retains the title ‘The Metaphysical Basis of Ethics’ for the 1898 dissertation his focus of attention shifts in the new chapters i and ii to a critique of Kant’s Pure Reason and the development of his new theory of propositions and concepts. None the less Kant’s philosophy remains central to this dissertation, and he also adds a brief chapter (chapter v ‘Ethical conclusions’) to provide a critical discussion of Kant’s ‘Practical Ethics’, i.e. Kant’s account of the moral law, which had been largely missing from the 1897 dissertation. Moore had been introduced to Kant’s philosophy when he was an undergraduate and attended Stout’s lectures on the history of modern philosophy (Moore 1942, 18). On Ward’s advice he spent his summer vacation in 1895 in T¨ubingen where he attended lectures on Kant by Sigwart (which he did not enjoy), but he did improve his grasp of German sufficiently to be able to read Kant’s writings in German. At this time the neo-Kantian movement was dominant in German philosophy, and in the prefaces to his dissertations Moore mentions that he has consulted the writings of Cohen, 51
See Moore (1991), lecture 10, 186–91.
Editors’ introduction
xlvii
Fischer, Vaihinger and others, though he does not appear to have taken much notice of them. Kant’s philosophy also commanded increasing attention in Britain at this time, having been especially brought to attention by T. H. Green’s use of Kantian themes in his Prolegomena to Ethics (see v.1). The most notable current British neo-Kantian was Moore’s 1897 examiner, Edward Caird, whose works included his massive two-volume study The Critical Philosophy of Immanuel Kant (Caird 1889) which Moore mentions in his 1897 preface and alludes to occasionally in both dissertations (see especially 1898 dissertation, pp. 148–50). Moore’s assessment of Kant’s ‘metaphysics of ethics’, and especially his account of freedom, led him in the end to a firmly negative conclusion: at the end of chapter iii of the 1898 dissertation he writes ‘[w]ith this the whole construction of Kantian Freedom seems finally to fall to the ground’ (1898 dissertation, p. 209). But his criticisms are initially combined with enthusiasm for Kant’s approach to philosophy; in the 1897 dissertation he expresses his attitude in the following terms: In this Kant betrays the too psychological standpoint above which he seems never to have completely risen in treating epistemological questions, in spite of the enormous services which he did to epistemology, as well in the Metaphysics of Ethics as elsewhere. He supplies, as it seems to me, more materials for a true view than any one else, and those, too in a wonderfully forward state of preparation, but nevertheless they are still for him encumbered and confused with the irrelevant matter, from which it was his merit to set them free for others. (1897 dissertation, pp. 62–3)
It is no doubt indicative of Moore’s revised attitude that this passage (apart from the initial reference to Kant’s ‘psychological standpoint’) is omitted from the 1898 dissertation (p. 192). Yet even while he is here developing his new account of propositions and concepts he remarks ‘how much this theory has in common with Kant’s theory of perception’ and that ‘it retains his doctrine of Transcendentalism’ (1898 dissertation, p. 169), although, as we observed above (p. xxxii) it is not clear what Moore means by these remarks. The main topic of Moore’s critical discussion of Kant in the 1897 dissertation is Kant’s account of freedom; in the 1898 dissertation Moore adds to this his discussion of Kant’s account of pure reason. The main theme throughout is indeed Kant’s ‘too psychological standpoint’ which, according to Moore, leads him to treat metaphysical grounds or logical presuppositions as quasi-psychological causes. ‘Transcendental freedom’, according to Moore (at least in 1897), makes some sense as a unified reality
xlviii
Editors’ introduction
which provides a complete ground of appearances; but Kant’s attempt to use it as a basis for ‘practical’ freedom involves a misconceived attempt to give it a psychological basis in the will. Similarly, Moore argues, although there are a priori truths whose truth is presupposed by the truth of common sense judgements about objects and their properties, Kant’s ‘Copernican’ suggestion that these truths have their basis in the understanding rests on a ‘metaphorical personification of Reason, of which I think Kant never cleared his thought’ (1898 dissertation, p. 134). Apart from the new topics that come into the 1898 dissertation, however, the main difference between them when it comes to the assessment of Kant is Moore’s increasing scepticism concerning the existence of a reality which grounds appearances. In 1897, although Moore is critical of Kant’s conception of things-inthemselves, he does not in the end reject the thesis that appearances are grounded in an all-embracing timeless reality (‘the world as it really is’ – 1897 dissertation, p. 35). As noted above (see v, p. xxvi) Moore uses Bradley’s absolute idealism to correct what he takes to be mistakes in Kant’s position in this respect. By 1898, however, although Moore still affirms the unreality of time, he now questions the existence of any such reality – which is why ‘the whole construction of Kantian Freedom seems finally to fall to the ground’. Before dealing with Moore’s debates with Kant on the topics of freedom and reason, however, it is worth considering briefly the two introductions to his dissertations, where there is already a complex engagement with Kantian themes. vi.1 moore’s introductions
Having taken as his title ‘The Metaphysical Basis of Ethics’ Moore recognises in 1897 that he needs to show there is a ‘metaphysical’ subject matter for his study. His argument to this effect is that there is a ‘fallacy involved in all empirical definitions of the good’ (1897 dissertation, p. 10). Moore’s argument here is essentially that which he was to propound in Principia Ethica as demonstrating the existence of the ‘naturalistic fallacy’ (it is in his 1898 lectures ‘The Elements of Ethics’ that he first uses the phrase ‘naturalistic fallacy’, though in the 1898 dissertation he writes of ‘naturalistic ethics’ – 1898 dissertation, p. 232). It is difficult to pin down exactly what the ‘fallacy’ is supposed to be; Moore says here, as if this explained the fallacy, that there is a ‘tautology’ in any supposed definition (1897 dissertation, p. 11), by which he seems to mean that there is a problematic circularity in it. But the underlying thought seems to be that which recurs in Principia
Editors’ introduction
xlix
Ethica, that the existence of persisting deep ethical disagreements which cannot be resolved by identifying differences in the definitions of ethical terms shows us that ethical terms do not admit of definition in empirical terms (1897 dissertation, pp. 10–11). Moore here expresses the conclusion of his argument in Kantian language, as the thesis that the meaning of good is fundamentally ‘transcendental’ (1897 dissertation, p. 12). Indeed he goes on attribute his conclusion to Kant: It is Kant who seems most clearly to have recognised this distinction and to have done most towards a systematic exposition of the nature of the concept ‘good’; but since he states the problem in a different way some preliminary statement of my grounds of dissent may not be out of place. (1897 dissertation, p. 12)
As Moore’s talk here of ‘dissent’ indicates, there are in fact significant disagreements between himself and Kant at this point. Kant’s most ‘systematic exposition of the nature of the concept “good”’ occurs in his Critique of Practical Reason (AK 5: 58–64, MG 186ff.) where he defines the concept in terms of the objects of ‘the faculty of desire . . . in accordance with a principle of reason’ (AK 5: 58, MG 186). Because of the reference here to ‘reason’, this is not an empirical definition of good; but none the less it indicates that for Kant good is not the fundamental moral concept; instead this is the concept of pure practical reason which, Kant holds, identifies by itself the principles which determine what rational beings such as ourselves ought to do. Moore indicates briefly in the introduction (1897 dissertation, p. 15) that he disagrees with Kant on this central point, and he argues the point in much more detail later in the dissertation (1897 dissertation, pp. 75ff.; cf. p. lvi below). For Moore it is the good, and not the rational will, that is the foundation of ethics. One implication of this aspect of Moore’s position is that fundamental inquiry into ethics aims to provide knowledge of the good, and, Moore thinks, there is no good reason for taking this kind of knowledge to be radically different from other types of knowledge of fundamental truth. So ethics should be regarded as a branch of ‘Theoretical Philosophy’, and not treated as a separate branch of ‘Practical Philosophy’ (1897 dissertation, p. 7). A position of this latter kind would, Moore allows in the 1897 dissertation (p. 13), be legitimate if Kant had been correct in holding that ethics is the study of pure practical reason; but, as we have seen, Moore rejects this position. Moore indeed recognises that his
l
Editors’ introduction
position is much closer to that of Plato than Kant: ‘So far therefore as general philosophical scheme goes, the standpoint here taken up seems to agree most with that of Plato’ (1897 dissertation, p. 14). But, having abandoned the study of classics for philosophy, Moore does not pursue this connection with Plato; instead he returns directly to Kant as a ‘startingpoint for a Metaphysic of Ethics’ (1897 dissertation, p. 14), despite this fundamental disagreement between them. He gives several reasons for doing so. Partly because his exposition is so much fuller and more systematic; partly because he was able to make use of so much more speculation; and partly because his direct influence upon modern thought seems so much greater. And much of what he said seems to lead directly to the view above expressed. (1897 dissertation, p. 14)
It looks as though the main reason here is the connection with ‘modern thought’ – i.e. the idealist philosophy of Kant and his successors, most notably Bradley to whom Moore has just deferred in his preface (1897 dissertation, p. 4). So, one might say, Moore’s project in 1897 was to draw on the resources of Bradley’s Absolute idealism in order to adapt the idealism of Kant’s ‘Metaphysic of Morals’ to serve the goals of his own neo-Platonist theory of the good, and thereby construct a new ‘Metaphysic of Ethics’. A year later Moore’s 1897 introduction is incorporated almost entirely unchanged into the introduction to his new 1898 dissertation. Moore still uses Kantian language to express the conclusion of his demonstration that there is a fallacy in empiricist definitions of the good – namely that good has a ‘transcendental’ meaning (1898 dissertation, p. 126). And, having alluded as before to his agreement with Plato concerning the priority of good in ethics, he again reverts straight back to Kant as his ‘starting-point’ (1898 dissertation, p. 128). Does that mean that his project is the same? It looks as though it must be very much the same; after all, he retains the title ‘The Metaphysical Basis of Ethics’. But there is one significant difference: in 1897 Moore’s reference to ‘modern Idealism’ (1897 dissertation, p. 14) suggested that he intended to use Bradley’s Absolute idealism as a resource for his own metaphysical project. One year later the reference is toned down, as was the earlier reference to Bradley in the new preface (1898 dissertation, p. 127). So although Moore still presents himself here as aiming to adapt Kant’s idealist metaphysics for his own Platonic ends, Bradley’s Absolute idealism (‘modern Idealism’) is no longer presented as the obvious resource for achieving this, though Moore does not suggest
Editors’ introduction
li
the alternative he is in fact going to offer – the radically new non-Kantian theory of the a priori he presents in chapters i and ii (perhaps he ran out of time to rewrite his old introduction). But there is none the less one significant difference between the introductions, in that Moore adds to the 1898 dissertation a discussion of his approach to writing about the history of philosophy (1898 dissertation, pp. 129–32). This is an interesting passage and unique in Moore’s writings. Moore seems to have been prompted to write it by Caird’s report on his 1897 dissertation, in which Caird complained that in several respects Moore’s criticisms of Kant arose from misunderstandings, and also that ‘[s]ometimes it is not easy to see whether Mr. Moore is interpreting Kant, or expressing his own views’ (Caird, p. 99). Moore’s response to Caird is anticipated by an intriguing paragraph found on the verso side of the title page of the 1898 dissertation, in Moore’s hand: You should not merely find your own views in Kant, but unless you carefully compare him with what you yourself can really understand and think to be true, you are in great danger of never finding what he meant at all, but only collecting different formulae which he used. You can never hope to understand what an author means, unless you come to him with an independent judgment of your own on the subject which he is discussing. The purely historical treatment may be useful as a handbook, but it is impossible that it should contribute anything to philosophy or to the knowledge of your author as a philosopher. The man who is not an independent (if not original) thinker himself, has no chance of discovering new meaning in an author’s views – even the author’s own meaning; he can only make it a little easier for others to do so.
In his introduction this line of thought is developed in more detail. Despite the remark in his preface, that ‘I have followed the suggestions of my examiners in attempting to distinguish more clearly between my own views and those of Kant’, he argues that this distinction is not a deep one, for ‘beyond somewhat narrow limits, there seems no ground for determining what a philosopher actually thought except a judgment of what it is right to think’ (1898 dissertation, p. 130). Moore here propounds an extreme version of the principle of charity which, arguably, does not allow sufficiently for the possibility of learning from the past, or for the kind of dialogue with the past which leads to a ‘fusion of horizons’ (Gadamer 1979, 268–73). None the less Moore is certainly right in his thesis that the interpretation of a past philosopher is inseparable from the achievement of a satisfactory understanding of the topics the philosopher is addressing.
lii
Editors’ introduction vi.2 freedom
The 1897 dissertation consists of one very long chapter entitled ‘Freedom’, together with a much shorter appendix on Sidgwick’s hedonism (see v, p. xli). Much of the discussion concerns Kant’s account of freedom, and a warning of what to expect, which exemplifies the approach to the history of philosophy outlined here, comes from the opening paragraph of Moore’s paper ‘Freedom’: ‘It is not my main object to expound Kant, but to arrive at the truth on the subjects which he discusses’ (Moore 1898, 179). Moore’s discussion of Kant’s account of freedom addresses the familiar question of how Kant can combine his unqualified determinism with respect to the empirical world with his insistence that morality presupposes a capacity for self-determination which cannot be accommodated within a deterministic order of events. In his Groundwork of the Metaphysics of Morals Kant famously suggests that the way to resolve this issue is by recognising that we can regard ourselves ‘from two standpoints’, as belonging to the empirical world, but equally as belonging to an ‘intelligible world’ (4:452, MG 99) in so far as we have the rational capacity both to make sense of our experience as experience of an empirical world and to project ourselves beyond this world by deploying ‘ideas’ such as the idea of freedom. Moore approaches the question of the validity of this position by starting from Kant’s discussion of it in the Transcendental Dialectic of Critique of Pure Reason (book ii, ch. ii, section ix §3 ‘Resolution of the cosmological idea of the totality of the derivation of occurrences in the world from their causes’ – B560, GW 532), where Kant introduces the idea of freedom in the context of a discussion of causality. Kant here states that the ‘practical’ concept of the freedom of a rational agent is grounded on a ‘transcendental idea of freedom’ which is the idea of a kind of causality which is ‘the faculty of beginning a state from itself ’ (B561, GW 533); and it is then on this fundamental ‘Transcendental Freedom’, as Moore calls it (1897 dissertation, p. 35), that Moore concentrates in his 1897 dissertation. Moore’s discussion of transcendental freedom can appear at first somewhat misconceived, since he tends to treat it as if for Kant it were a type of freedom that can be considered apart from practical freedom, whereas Kant’s claim is basically that practical freedom has a transcendental condition, namely the possibility of there being an uncaused ground of human action. In fact Kant does himself distinguish between practical and transcendental freedom as such in the ‘Canon of Pure Reason’ (B831, GW 676), but, setting aside the issue of Kant’s own use of this terminology, if one simply interprets Moore’s discussion of ‘Transcendental Freedom’
Editors’ introduction
liii
as a discussion of the transcendental condition of practical freedom, one can make reasonable sense of much of Moore’s argument. Thus, given determinism with respect to empirical phenomena, it follows that transcendental freedom can consist only in the existence of a non-empirical uncaused ground, or ‘ultimate ground’ (as Moore prefers to think of it – 1897 dissertation, p. 26), of human action and Moore therefore takes it that such a cause must be a non-empirical thing-in-itself. This leads Moore to consider the merits of Kant’s discussions of things-in-themselves and other non-empirical objects (1897 dissertation, pp. 26–35), from which he concludes that the only thesis of this type which can be defended is the conception of ‘Reality’ as an unconditioned totality which provides an ungrounded ‘ground’ for all empirical appearances and therefore meets the transcendental condition for being a ‘free cause’ of empirical human action (1897 dissertation, p. 35). This position is recognisable as a version of Bradley’s Absolute idealism. The question which now arises is whether this transcendental condition, Moore’s ‘Transcendental Freedom’, has any close connection with human agency and thus with the kind of ‘practical freedom’ which, by giving a special metaphysical status to the human will, legitimates the moral law, at least as Kant conceives it. In the central part of his 1897 dissertation Moore argues that there is no such connection and thereby reaches the conclusion that there is no special practical freedom of the kind Kant supposed: My conclusion, then, is this: That ‘will’ is only a special form of natural causality, or rather, a natural causal process, where the cause is of one definite sort. It is a special form of natural causality, just as explosion of gunpowder by a match is one special form of natural causality, and explosion of gunpowder by percussion is another. And, that on which I wish to insist, is that voluntary action, of whatever sort, whether autonomous or heteronomous, exhibits ‘freedom’, in the sense which I have hitherto explained as essential to Kant’s notion, no more and no less than gunpowder explosions or any other natural process whatever . . . ‘Freedom’, then, for Kant means only ‘transcendental freedom’, and ‘transcendental freedom’ is not ‘practical’, in the sense that it is inseparably connected with ‘action’ alone. It is true that actions are dependent on ‘transcendental freedom’, but that is only because it is the relation which holds between the empirical causes of those actions and the transcendental ground of such causes; whether sensible objects produced effects, and so indicated their right to be considered practical (as they always must), or not, they would equally be results of ‘transcendental freedom’. (1897 dissertation, pp. 72–3)
Moore’s route to this conclusion involves a detailed discussion of the supposed freedom of the free will. Having set aside Sidgwick’s intuitive
liv
Editors’ introduction
‘affirmation of consciousness’ in favour of free will (see v.3, p. xl above), Moore discusses whether the freedom of a free will can be thought of as an extrapolation from the kind of ‘self-caused’ relative autonomy of organisms (as Kant had suggested in his Critique of Judgment). Moore’s conclusion is negative: ‘It is, no doubt, convenient to treat certain groups as wholes, and to ascribe freedom to them in the sense that their parts and not anything external to them are immediate cause of their changes; but this grouping must be admitted to be wholly arbitrary’ (1897 dissertation, pp. 56–7). So there is no special metaphysical freedom here, and ‘the case of human volition would seem to be precisely the same’ (1897 dissertation, p. 57). Moore recognises, however, that he needs to address the central Kantian consideration that the status of the will is special because it is capable of being ‘determined by reason’ alone and therefore ‘free’ in so far as this determination by reason does not depend on further antecedent causes. In response to this, Moore argues that the claim that ‘reason determines the will’ is crucially ambiguous (1897 dissertation, p. 63): if it is to warrant the attribution of a special status to the will, then ‘practical reason’ itself must be construed as an uncaused cause, or ultimate ground, of the will’s determinations. But, Moore objects, the only sense in which it is legitimate to accept that ‘reason determines the will’ involves the ordinary psychological process whereby someone’s more or less rational beliefs determine their will, and here no special type of causation is involved; instead the causation remains entirely ‘natural’ and thus does not give the will any special metaphysical status (1897 dissertation, p. 62). Moore’s argument rests on the assumption that there is an ‘impassable gulf’ which separates an ‘idea’ conceived ‘as a psychical existent’ from ‘the content of which it is an idea’ (1897 dissertation, p. 62). This distinction is reminiscent of Bradley’s distinction between ideas as ‘psychical states’ and ideas as ‘meanings’ which Moore discusses in chapter ii of his 1898 dissertation (see v.1 on British idealism, p. xxviii above), though Moore does not here refer to Bradley. Where ‘reason’ is conceived as a ‘psychical existent’ there is no reason to deny that human action is often ‘determined by reason’; but no special form of causality is involved. Whereas if ‘reason’ is conceived as the ‘content’ of these ideas, and thus perhaps as the moral law itself, some very special grounding would be involved if this were indeed to determine human action; but, Moore holds, the moral law itself, as distinct from moral beliefs, has no such power to determine the will. Kant’s conception of pure practical reason presents a challenge to Moore’s assumption that there is an impassable gulf between ideas and their
Editors’ introduction
lv
content: for he took the view that our capacity for practical reason provides by itself the basis for the moral law. Thus, in Moore’s terminology, the ‘content’, i.e. the moral law, is grounded upon a human capacity for practical reason, even if, because it is a priori, this capacity is not happily described as just a ‘psychical existent’. Thus Kant’s position implies that, in so far as human beings exercise their capacity for practical reason by acting on principles which express this capacity, their actions have a rational basis which legitimates the description of them as ‘free’, since they are thereby conceived as grounded in a fundamental a priori capacity. Notoriously, it remains unclear how this rational determination of the will is to be combined with Kant’s unqualified assertion of the will’s causal determination by antecedent empirical causes, and in this respect Moore’s distinction between an idea and its content resembles Kant’s own distinction between the empirical and the intelligible worlds. But it is clear, in fact, that this latter distinction is not, for Kant, an ‘impassable gulf’, and that, on the contrary, the will (Willk¨ur) is supposed to be a bridge between the two52 (Kant is clear on this point in his discussion of the will in Religion within the limits of Reason Alone). In effect Kant allows that the empirical order includes the exercise of a priori rational capacities in such a way that the norms inherent in these capacities play a fundamental causal role in the determination of action. To set out this Kantian response to Moore’s challenge is certainly not to vindicate it, but it is not necessary or appropriate to take a stand on this question here.53 Moore’s achievement was primarily to articulate in a precise way a fundamental challenge to this aspect of Kant’s position. Not surprisingly Moore also rejects Kant’s conception of pure practical reason as the foundation for the moral law. What he meant by his conception of a ‘pure’ will was a will which should command consistency in your view of what ought to be done – a consistency which is necessary to ‘reason’ as such. But all that a ‘pure’ will could really mean would be a will which was consistent with itself in the fact that it always commanded, not in respect of that which it commanded. We may admit that the former notion can be obtained by an analysis of will; but it would be by no means sufficient to give a moral principle. (1897 dissertation, pp. 75–6; 1898 dissertation, p. 203)
Kant will of course object that there is more to pure practical reason than this simple test of consistency, but Moore takes it that what is missing 52 53
This point is discussed at length in Caird (1889): see vol. ii, ch. iii, ‘The Idea of Freedom’. For an exposition and defence of Kant’s position, see Korsgaard (1996).
lvi
Editors’ introduction
from Kant’s ethics is the primacy of considerations concerning the good. Skipping a couple of lines, the passage above continues: Thus, what really distinguishes a moral law from laws of nature, is that it expresses the connection of natural objects, not with one another, but with the notion of ‘end’ or ‘goodness’. Its primary form is ‘This is good’ . . . (1897 dissertation, p. 76; 1898 dissertation, p. 204)
Moore acknowledges in a footnote here that he is inverting Kant’s own account of the relative priority of the good and the right, since Kant takes it that the good is that which is rightly desired (see Critique of Practical Reason, AK 5: 58ff., MG 186ff.). In truth, as noted above (p. xlix) Moore’s treatment of Kant on this topic is rather cavalier, since when endorsing Kant’s rejection of empirical theories of ethics, he represents Kant as affirming the fundamental a priori status, not of pure practical reason, but of the good, but then not seeing that he (Kant) needed some further truths concerning what is good in order to provide a moral theory. This leads to odd passages, such as: While therefore we accept Kant’s metaphysics of ethics as pointing out the fundamental principle that ‘Good is a concept’, with its consequence that ‘What is good once, is good always’, it seems impossible to accept his theory that any practical principle can be obtained, without a second premiss, not to be deduced from the first, of the form ‘This is good’. (1898 dissertation, p. 235)
Plainly Kant would not have been impressed by this comment which largely ignores his own conception of pure practical reason. Moore rejects Kant’s rationalist conception of the practical a priori as incompatible with the demands of the causal connectedness of our empiricial psychology. In its place he affirms a ‘Platonist’ account of fundamental ethical truths as abstract a priori truths that are just ‘intuitively known’ or ‘given’, as in the following passage: To call an ethical principle ‘rational’ will therefore mean, in the first place, that it is a proposition. The question of its origin, whether it be regarded as a precept of Reason, as innate, or derived from experience, must in any case be wholly irrelevant. For such explanations can, at best, only give a history of how we come to apprehend it, of our knowledge of it: itself can not properly be said to have any origin at all. Nevertheless it may be well to state that the description of it as intuitively known seems nearest to the truth; for, whether it be inferred or immediately apprehended, it is, in either case, itself merely ‘given’, as much as any so-called sense-datum. (1898 dissertation, pp. 176–7)
Editors’ introduction
lvii
The contrast here, between Kantian rationalism and Platonist intuitionism, need not be confined to ethics: it can be applied to ‘pure reason’, i.e. a priori truth, generally. Moore recognised this already in the 1897 dissertation where the critical discussion of practical freedom which had led him to his criticisms of Kant’s conception of practical reason is briefly extended to criticisms of theoretical reason (see e.g. p. 76), but this generalised line of thought is first worked out in detail in the new chapter i of the 1898 dissertation.54 vi.3 reason
Moore begins chapter i (‘On the meaning of “Reason” in Kant’) with an odd discussion of the concept of a priori truth. He starts from Kant’s claim that the marks of a priori truth are necessity and strict universality, but then argues that these marks do not pick out the judgements which we intuitively regard as a priori: many apparently empirical judgements are universal and, more importantly, most, or even all, such judgements are necessary. Moore’s initial reason for this position is that any judgement (or proposition) which ‘involves necessity’ in the sense that its truth presupposes that of a manifestly necessary proposition, is itself to be counted as a priori by Kant’s criterion because of this involvement; thus because the proposition ‘This body is heavy’ presupposes ‘Heaviness is an attribute’, and the latter is a necessary proposition, the former should count as a priori (1898 dissertation, p. 136). Moore takes it that this argument still allows one to regard simple existential propositions such as ‘Heaviness exists here and now’ as empirical (1898 dissertation, p. 139), but even this result is set aside in the next chapter where Moore argues that absolutely all propositions are, if true, necessarily true on the grounds that the truth of any proposition is internal to it (1898 dissertation, p. 173). The truth of a proposition, he maintains, cannot depend on the contingency of there being a state of affairs external to the proposition to which it corresponds, for the existence of a state of affairs just comprises the truth of an existential proposition. Instead, he holds, the truth of a proposition arises from the way in which its constituent concepts are combined in the proposition (1898 dissertation, pp. 167, 173), and since a different combination of concepts would constitute a different proposition, it does seem to follow that the truth of a proposition is internal to it. As Moore acknowledges, 54
As we explain in viii.2 of this introduction, the text of this chapter and of the following chapter of the 1898 dissertation has had to be to some extent reconstructed. In this discussion we assume that this reconstruction is reliable.
lviii
Editors’ introduction
this position is ‘paradoxical’ (1898 dissertation, p. 167) and it is not easy to make good sense of it; but it is not necessary to pursue the matter here. The result of this initial discussion of a priori truth is that an alternative to Kant’s account of the a priori is required, and Moore suggests therefore that what is distinctive of empirical propositions is that they involve empirical concepts whose definitive feature is that they ‘can exist in parts of time’ (1898 dissertation, p. 137); a priori propositions, therefore, involve only a priori concepts which ‘can only claim a precarious sort of existence’ (1898 dissertation, p. 136) in virtue of the fact that they cannot exist in time. Hence Moore’s a priori/empirical distinction is essentially that between non-temporal and temporal truths. One complication here is the fact that Moore carries forward into the 1898 dissertation the thesis that time is unreal (p. 127); hence he ends up accepting that all empirical propositions (presumably all positive ones, at any rate) are false (1898 dissertation, p. 175). But that point can be set aside for the moment; what is more important is the recognition that Moore has replaced Kant’s essentially epistemological a priori/empirical distinction with a metaphysical non-temporal/temporal one. As we shall see, this substitution is characteristic of Moore’s critical discussion of Kant’s conception of pure reason. For Moore’s fundamental thesis is that epistemological conditions are either disguised psychological causal connections or logical connections intended to establish the validity of metaphysical truths. Before saying more about this, however, it is worth noting briefly Moore’s dismissive discussion of analytic truths, which he interprets as definitions which state empirical truths concerning the use of words (1898 dissertation, p. 140). So although analysis does play an important role in Moore’s philosophical method in his dissertations, for example in his metaphysical theory of concepts and propositions where he writes that ‘a thing becomes intelligible first when it is analysed into its constituent concepts’ (1898 dissertation, p. 168), it is essential to realise that Moore does not think of analysis of this kind as issuing in ‘analytic’ truths, at least in Kant’s sense. Thus in so far as these early writings of Moore are taken to constitute a starting point for ‘analytic philosophy’, it is important to recognise that Moore does not think that philosophical analysis brings with it a commitment to regarding a priori philosophical truth as ‘analytic’. The core of Moore’s critical discussion of Kant’s transcendental idealism is his complete rejection of Kant’s ‘Copernican Revolution’, concerning which he writes:
Editors’ introduction
lix
it appears to me an exceedingly unfortunate suggestion; and in discussing this connection which Kant finds between the a priori and the Reason, I shall hope to shew that there is no reason for asserting it; that it does not serve the purpose for which it was devised; and that it is inconsistent with other contentions of the Critical Philosophy, which appear to me to be true and of great importance. (1898 dissertation, p. 143)
Moore recognises that in his Critique of Pure Reason Kant deploys three different lines of thought to substantiate his idealist thesis: the first, in the Transcendental Aesthetic, concerns the status of space and time; the second, in the Transcendental Analytic, concerns the categories of the understanding; and the third, in the Transcendental Dialectic, arises from the antinomies which are implicit in the cosmological ideas of totality. In each case, Moore argues, Kant’s discussion does reveal important truths; but in no case is an idealist conclusion warranted. It is convenient to start from Moore’s treatment of Kant’s arguments in the Dialectic. Moore takes it that Kant’s arguments in the first two antinomies concerning time and space do demonstrate that once one applies the idea of unconditioned totality to space and time by asking whether they are finite or infinite one is led to contradiction: space and time are both finite and infinite. But he rejects Kant’s solution to these antinomies, according to which the mistake has been to apply ‘the idea of absolute totality, which is valid only as a condition of things in themselves, to appearances that exist only in representation’ (B534, GW 519). For, Moore argues, even if it is a mistake to apply the idea of totality to time and space, it does not follow that they ‘are appearances which exist only in representation’. Plainly Moore is not here engaging with the details of Kant’s argument, which has the merit of both suggesting an explanation for the apparent contradictions in space and time and dissolving them – whereas Moore just leaves the conclusion of the antinomies standing: ‘the Antinomies prove that our conception of the world, as in space and time, is not a true one’ (1898 dissertation, p. 152). And since he proceeds to infer that ‘all our propositions with regard to time and space are absolutely false’ (1898 dissertation, p. 174), he ends up embracing a form of metaphysical scepticism which Kant’s transcendental idealism was intended to avoid. In his dissertations Moore does not appear especially unnerved by this result; but in the brief discussion of the antinomies of space and time in the lectures on ‘Kant’s Moral Philosophy’ which he gave early in 1899 he does not affirm this sceptical result although he still rejects Kant’s idealist
lx
Editors’ introduction
solution.55 Moore did later provide a much better treatment of these issues in his 1910–11 lectures Some Main Problems of Philosophy (see Moore 1953, lectures 9 and 10); and what is notable here is that Moore argues that there are errors in the arguments concerning infinity by which Kant aims to establish the antinomies of time and space, in that they involve an inadequate grasp of reasoning concerning infinite totalities. Moore here (Moore 1953, 198) alludes to Russell’s discussion of infinity in The Principles of Mathematics (Russell 1903; especially chapter lii). Moore knew this work well because in 1905 he had attempted to write a review of this book (see Schilpp 1942, 27). Because he remained dissatisfied with his review, he did not in the end publish it, but his manuscript survives56 and shows that Moore here concentrated on Russell’s arguments against Kant’s treatment of infinity in the antinomies.57 In the case of the argument of the Aesthetic that space and time are a priori and, for this reason, transcendentally ideal, Moore objects that this idealist conclusion does not follow (1898 dissertation, p. 144). However, here too Moore does not go into the detail of Kant’s treatment of space and time, so his rejection of Kant’s conclusion is rather abrupt. One thing worth noting about Moore’s discussion here, however, is that he himself does not notice that his own acceptance that space and time are a priori does not sit comfortably alongside his revised conception of the a priori; for one would expect space to be counted as empirical in so far as ‘it can exist in parts of time’. Indeed, when discussing Kant’s argument that space is a priori, Moore says: To this argument, in a modified form, as shewing that space is logically prior to the objects of experience, no exception need be taken: but it only shews that space is logically prior or a priori, not its required subjectivity. (1898 dissertation, p. 143)
The key concept here is that of ‘logical priority’, which is not explained but seems to involve what Moore elsewhere calls ‘logical presupposition’. This suggests that in fact Moore’s conception of the a priori is one of propositions which are logically presupposed by all propositions concerning experience in the sense that the truth of the former are necessary conditions of the truth of the latter. This has no direct connection with his official account 55
56 57
Moore’s manuscript notes for these lectures are in the Moore archive in Cambridge University Library (Add. MS 8875 14/2/5). We quote a short passage from Moore’s exposition of Kant’s first antinomy on p. 174 n. 89 below. Cambridge University Library (Add. MS 8875 15/2). Moore’s conclusion is: ‘We may conclude, then, that such arguments are one and all fallacious; and considering how many philosophical theories have been supported by the supposition that infinity is impossible, such a conclusion is not without importance.’
Editors’ introduction
lxi
of the a priori in terms of the lack of relation to time, and is not far from the traditional conception of the a priori, except that presupposition is a logical, and not an epistemological, relation. Whereas Moore’s critical discussions of Kant’s Aesthetic and Dialectic do not really engage with the considerations Kant advances for transcendental idealism, Moore’s discussion of Kant’s Analytic is much more revealing. Moore accepts that Kant does show that categories such as substance and causation are ‘logically presupposed in empirical judgments’ (1898 dissertation, p. 147); but, again, he asks what this has to do with transcendental idealism. Moore then acknowledges that the answer to this question is bound up with the way in which Kant holds that the validity of the categories rests upon their role as a priori conditions for the possibility of empirical knowledge, and at this point he tries to set up a dilemma comparable to that discussed above in connection with Kant’s conception of the role of pure practical reason. He maintains, first, that knowledge is just true consciousness and, second, that the truth of a proposition is independent of its being an object of consciousness (1898 dissertation, p. 148). Once these points are granted, Moore can return to the Kantian conception of the categories as conditions for the possibility of empirical knowledge and insist that there are here two questions, separated in effect by an ‘impassable gulf’: first, whether the categories are conditions for the possibility of truth, and second, whether they are conditions for the possibility of conscious thought. The first question involves the validity of the categories as logical presuppositions, whereas the second is a causal question concerning the psychological origins of conscious thought. Moore, as we have seen, gives an affirmative answer to the first question; but he insists that because this answer is not dependent on an affirmative answer to the second, psychological, question, it does not bring with it any idealist implications. It is, he alleges, only by confusing psychological questions about the causal conditions for conscious thought with non-psychological questions about the logical conditions for the truth of what is consciously thought that Kant and his followers, such as Caird (whose discussion of this point Moore criticises on pp. 149–50), have been to led to think that the validity of the categories requires the truth of transcendental idealism. As before, the issue is whether there really is an impassable gulf between these questions. Moore’s account of knowledge omits all reference to justification by reason and once this is included, it opens the way for Kant to argue that pure reason has a legitimate constructivist role in so far as the theoretical reasoning which justifies empirical knowledge brings with it a commitment to the validity of the categories and thereby shows how a
lxii
Editors’ introduction
priori truths concerning these categories can be grounded on the exercise of pure theoretical reason. Whether a ‘Transcendental Deduction’ of this kind can be made to work is, of course, much disputed;58 but because Moore’s critique of Kant fails to address it he fails to dispose of the core thesis of the Transcendental Analytic. None the less just as applies to the case of Moore’s criticism of pure practical reason (see, p. lv above), Moore’s challenge shows clearly here what the Kantian project requires if it is to succeed. Moore sums up his challenge in the following way: Finally the subjective explanations of Aesthetic and Analytic alike are inconsistent with the very necessity which they attempt to establish. For they presume to deduce necessity from a mere fact, namely that our mind is so and so constituted, and this, on Kant’s own principles effectually excludes the propositions deduced from any claim to be absolutely necessary. So far as this explanation is concerned, the Critique must lose its title to a priori demonstration, and thus the very hypothesis by which Kant hoped to establish once for all a scientific philosophy, would, if it were true, render such philosophy impossible. (1898 dissertation, p. 151)
Moore draws two not altogether coherent conclusions from this critical discussion of Kant’s ‘Copernican Revolution’. The first is that epistemology is not a legitimate philosophical discipline; instead it is just a mixture of psychology and logic which tempts philosophers to confuse logical demonstrations of validity with causal explanations of origin (1898 dissertation, p. 150). The second, which appears in fact to be an epistemological thesis rather than a logical or psychological one, is that a priori truths are apprehended in much the way that empirical truths are: the fact that 2 + 2 = 4 ‘presents itself to me as a fact forced upon me from outside just as much as any sensation’ (1898 dissertation, p. 153). Moore suggests that what lies behind Kant’s failure to recognise this point is his adoption of the ‘vicious theory’ (1898 dissertation, p. 155) of ‘the English philosophers’ (such as Locke) who held that sensations are ‘given’ as raw experiences. For once this myth of the given is accepted, it seems that the ‘mind’ is needed to synthesise and conceptualise these sensations in such a way that empirical knowledge is possible (see v, p. xxv for Green’s version of this position); and if the mind is needed to do this work, then it is natural to suppose that the fundamental ‘a priori’ principles which guide this work are themselves the work of the mind. According to Moore, however, this is all a mistake. No doubt there are psychological processes at work in perception and thought, but these deal only in contingent causal connections and can provide no basis for genuine a priori truth. As far as 58
For a selection of recent debates, see Stern 1999.
Editors’ introduction
lxiii
the acquisition of knowledge, either empirical or a priori, is concerned, no contribution by the subject is required: In the relation of subject and object, the object always appears to me as something merely ‘presented’, merely there, not as something produced by the subject, which contemplates it, and this equally whether it be a sensation, a thought, or a feeling. (1898 dissertation, p. 155)
This passage looks forward to a central conception of chapter ii, that the ‘object’ which is thus presented in sensation, thought and feeling is just a proposition; for we are familiar with the idea of affirming or ‘positing’ an existent, of knowing objects as well as propositions; and the difficulty hitherto has been to discover wherein the two processes were akin. It now appears that perception is to be regarded philosophically as the cognition of an existential proposition; and it is thus apparent how it can furnish a basis for inference, which uniformly exhibits the connexion between propositions. (1898 dissertation, pp. 168–9)
Hence as in ethics, so more generally (see pp. lvi–lvii above): Kantian rationalism is replaced by an intuitionist account of a priori knowledge which does not need to engage with the implications of the contribution to knowledge made by the mind’s conceptualising activities. On the contrary, concepts are not the work of the mind at all: indeed ‘in the end, the concept turns out to be the only substantive or subject’ (1898 dissertation, p. 173). In his subsequent writings on Kant Moore is sufficiently confident of these criticisms of Kant’s Transcendental idealism to repeat most of them without significant change;59 the only important change is that mentioned above which occurs in his 1910–11 lectures and concerns his attitude to Kant’s antinomies of time and space. After these lectures Moore in fact writes very little about Kant; he has a new agenda, the vindication of common sense, and a new method, philosophical analysis, and neither of these leads him back to Kant’s critical philosophy. And yet there is one paper where he does just this – his 1939 paper ‘Proof of an External World’ (Moore 1939). For Moore here challenges Kant’s conception of what it would be to prove the existence of an external world by displaying his hands to his audience and affirming that in doing this he does all that is necessary to prove the existence of his hands and thus of an external world. Whether Moore is right about this is much disputed; but one thing that is absolutely clear here is that Moore still rejects the transcendental idealism 59
See, for example, Moore (1903a), ch. iv, Moore (1904), Moore (1953), lecture 8.
lxiv
Editors’ introduction
which would imply that his hands are ‘merely appearances, i.e. mere modes of representation’ (A372, GW 428). vi.4 ethical conclusions
After the discussions in 1898 of ‘Reason’ (chs. i, ii) and ‘Freedom’ (chs. iii, iv), Moore adds a short new chapter in which he discusses Kant’s moral theory. The way Moore opens the chapter suggests that he felt that if he was to do justice to Kant’s ethical theory he needed to discuss Kant’s ‘Practical Ethics’ in addition to his ‘Metaphysics of Ethics’ which he had already discussed at length (though since the title of the dissertation was ‘The Metaphysical Basis of Ethics’ it is not obvious that this extra task was necessary). Kant’s ethics, he says, is a fusion of two ‘wholly distinct’ doctrines: an investigation into the concept ‘ought’, and the ‘peculiar and elusive’ doctrine of Practical Ethics. Moore regards the former favourably as what may be called ‘The Metaphysics of Ethics’. Of the latter, he is less approving. Moore describes Kant’s practical doctrine as the claim that no action is good unless it springs from a disposition on the part of the agent to act on the belief that the action is good. This disposition is the good will – a will which acts in conformity to and also ‘from’ duty. Moore’s objection to this is not so much that Kant tries to elevate the practical doctrine into the sole principle of practical ethics, but that Kant believes he can deduce it from his metaphysical doctrine. This, Moore says, is put forward as two claims: the attempt to deduce which actions conform to duty from the Categorical Imperative, and the attempt to prove that the only good actions are the ones whose motive is the desire to act in conformity to duty. Moore’s objection to the first point is that the test of the Categorical Imperative is not sufficient to detect that a maxim, a suggested principle for action, is ‘undutiful’, i.e. morally wrong (1898 dissertation, p. 234). Moore’s argument is the familiar complaint that the formal test of universalisability rules out nothing by itself; according to Moore, the only way in which Kant’s test implies that a maxim is undutiful is where it permits acts which are separately known to be undutiful. Even less convincing, according to Moore, is Kant’s contention that good actions are only those whose maxim is in conformity to the moral law. For, Moore argues, since actions in conformity to the law ‘have some special quality which it is difficult to call by any other name than “good”’ (1898 dissertation, p. 234), we get the ‘curious result’ that action is not good unless it is done because it is good. Not only does Kant’s moral theory lead in this way to a ‘vicious
Editors’ introduction
lxv
circle’ (1898 dissertation, p. 235), it also has implications which are morally unacceptable: For, if the value of the willing does not at all depend upon the value of the object willed, but solely on the fact that conformity to the moral law is the motive; then the object willed need not necessarily be in conformity to the moral law, and we shall be acting perfectly well, when we intend to lie and murder, provided only we are mistaken in thinking those actions to be in conformity with the law, and are moved by that goodness which we attribute to them. (1898 dissertation, p. 235)
Kantians will of course reject these criticisms on the grounds that they reveal an inadequate understanding of the objective implications of pure practical reason. Once it is properly understood, it will be argued, the test of the Categorical Imperative is not merely formal, since it universalises the requirements of rational agency (see O’Neill 1975). None the less it remains much disputed whether a satisfactory moral theory can be constructed on this basis alone, or whether it is essential to introduce separate valuations of actions, dispositions and outcomes.60 Moore’s view is that this separate component is essential: It appears, then, that before any practical ethical precepts can be laid down, the principle involved in Kant’s Moral Law, must be supplemented by some other principle or principles declaring that ‘This or that is good’. (1898 dissertation, p. 236)
For Moore, however, these principles identifying the good are not just essential preconditions of normative moral theory, as in contemporary constructivist ethical theory. Instead he takes it that moral theory is fundamentally based on principles of this kind and thus comprises a theory of the good which specifies ‘an ethical principle determining our ideal’ (1898 dissertation, p. 237) and thereby implies that once it is decided what course of action will produce this ideal, it follows that ‘It is good to do that’ or ‘You ought to do that’ (1898 dissertation, p. 237). Moore thus ends up affirming here the ideal utilitarian position he was to make famous in Principia Ethica. He continues in a vein which is also reminiscent of Principia Ethica, by emphasising the difficulty that now arises for Practical Ethics from our ignorance about the results of action, and thus concerning which course of action will produce the ideal outcome. His response to this quandary is to suggest that, for the purposes of ‘rational choice’, one should rely on considerations of probability: 60
For an influential hybrid position of this kind, see Scanlon (1998), esp. 150–1.
lxvi
Editors’ introduction
A rational precept will be one which bids you do what will probably have the best results. In attempting to make a rational choice you will, therefore, have to consider (1) What is good (2) What will most probably bring it about. (1898 dissertation, p. 237)
Moore’s distinction here between rational choice and right action is interesting, but not developed. Instead, as in Principia Ethica, he observes that even considerations of probability do not provide certainty as to what one ought to do: The question, therefore, arises whether a greater good that is less probable is to be preferred to a lesser good that is more probable. And this seems to be a question susceptible of no certain solution. (1898 dissertation, p. 237)
So even in the context of Moore’s own, non-Kantian, moral theory Practical Ethics is at best a flawed attempt to reconcile irreconcilable principles: ‘choose what is best’ and ‘choose what will happen’ – neither of which is itself sufficient to provide a rational guide to practice. (1898 dissertation p. 237) vi.5 the chronology of kant’s ethical writings
Moore added a brief discussion titled ‘Appendix on the Chronology of Kant’s Ethical Writings’ to the end of his 1898 dissertation. As Moore states in his preface, he prepared this appendix as a response to the comments of the examiners of his 1897 dissertation (especially Sidgwick) who had complained that he had not done justice to the changes in Kant’s position over the period of the main ethical works under discussion. Moore’s reply was that there is no reason to think that Kant did make any substantive changes to his position. The first part of the discussion covers relatively uncontentious points concerning different formulations of the Categorical Imperative where Moore argues that there are no significant changes. He then addresses the more substantive issue of whether Kant’s position in the Groundwork allows for the conception of agents who freely will what is bad, which involves an exercise of freedom that is only explicitly identified in the second Critique and later writings. Moore argues persuasively that in fact this possibility is allowed for in the Groundwork. Right at the end of the appendix, however, he acknowledges that in his late works Kant also introduces the distinction between Wille and Willk¨uhr to clarify his treatment of the will as a source both of moral principles and of actions. According to Moore this is only a ‘verbal’ change; but whether that is correct is disputed and Moore does not say enough to settle the issue (for a
Editors’ introduction
lxvii
different view, see Silber 1960). None the less Moore certainly displays an impressive familiarity with Kant’s texts in this appendix. VII MOORE’S EXAMINERS
Rule 13 of the rules for election to a Prize Fellowship (see iii) state that: Dissertations are referred by the Council, if they think fit, to special referees, so that the electors may have before them the opinion of two authorities with regard to each Candidate recommended for election.
In 1897 Moore’s dissertation was referred to Sidgwick and Caird; in 1898 to Bosanquet and Ward. Unfortunately Ward’s report, if indeed he ever wrote one, has been lost, but Moore preserved the others and they are reproduced immediately after the dissertation to which they apply. We have already discussed Sidgwick’s report in an earlier section of this introduction (v.3), so here we comment on the reports by Caird and Bosanquet, two of the leading British idealists of the period. vii.1 caird
Edward Caird (1835–1908) is an unjustly neglected philosopher. His twovolume study of Kant, The Critical Philosophy of Immanuel Kant, was a major contribution to the critical understanding of Kant and is still worth reading. Caird was born in Greenock but studied philosophy at Balliol College Oxford where he was taught by T. H. Green. In 1866 he was appointed Professor of Philosophy at Glasgow and held this post until 1893, when he returned to Oxford as Master of Balliol College following Benjamin Jowett’s death. Caird retired in 1907 and died a year later. Caird’s long report on Moore’s dissertation begins with considerable praise: I wish therefore to say that I think very highly of Mr. Moore’s philosophical powers, especially of his power of following his ideas to their ultimate results. What one has most to fear in philosophy is that this should not be done, and therefore the weakness and strength of ideas should never be distinctly seen, but hidden under some plausible compromise. It is therefore very high praise to say that a writer on philosophy can, or does realise fully the consequences of his principles. (Caird, p. 99)
Indeed Caird nowhere says in this report that he would not support Moore’s election to a Fellowship. But he does devote most of his report to explaining his disagreements with Moore and attempting to make a better case for
lxviii
Editors’ introduction
Kant on freedom. In doing so, however, Caird does not directly address Moore’s arguments in their own terms. Indeed he remarks at the start (p. 99) that he has found Moore’s dissertation ‘extremely difficult to understand’, because, he says, ‘Mr. Moore has not sufficiently studied how to be clear to those who are not looking at things at his precise angle’. Instead he confines himself to reiterating an interpretation of Kant that acknowledges the force of some of Moore’s criticisms, but nevertheless dismisses them as insufficiently attentive to what Caird, following Green (see v, p. xxv) defends as the central element in Kantian metaphysics: the role of the unifying nature of consciousness. Caird agrees with Moore that it is a serious weakness in Kant that he often adopts a psychological standpoint: His main weakness is, I agree, the introduction of the psychological point of view into epistemology or rather as I should prefer to put it, his stopping at epistemology, and not going on to Metaphysics; for it seems to me that epistemology is a mere compromise between psychology and metaphysics. (Caird, p. 103)
Indeed, Caird allows, it is understandable that the defects of Kant’s psychologistic language should have led Moore to reject his account of the role of things-in-themselves in perception and thought and to adopt instead the Bradleian account of the Reality as that which is fundamentally ‘free’ (p. 103). But, according to Caird, this position is absurd, and the way of reading Kant that has led Moore to adopt it is inadequate because it fails to take account of Kant’s key insight, ‘viz. the unity implied in the connecting process, a unity of which we become conscious when we become conscious of the self’ (p. 105). The proper Kantian view is that all experience, including moral experience, must be interpreted in relation to the unity of self-consciousness, even though, Caird acknowledges, Kant himself does not always express this insight. Thus Caird acknowledges that Kant sometimes presents his position in such a way as to invite Moore’s dilemma with respect to practical reason, that either it involves the natural causality of psychological processes (which Caird calls a ‘synthetic relation’) or it is just a matter of the logical connection of truths (which Caird calls ‘an analytic development of conceptions’): Either, he [i.e. Moore] thinks, we have a synthetic relation of causality between presentations or an analytic development of conceptions, and there is no third possibility. This is a dilemma for Kant, just so far as he conceives the judgment of self consciousness with the ideas of reason and the moral law, as all expressing an analytic unity. (Caird, p. 109)
Editors’ introduction
lxix
But, Caird holds, once Kant’s position is understood properly, it will be seen that he has a way out of Moore’s dilemma. The passage quoted above continues as follows: But just so far as he points to a better idea, and suggests that the view of experience which we get by looking only to the connections of objects with each other, is essentially abstract, and that we have to correct this view by the idea that the object is essentially an existence-for-a-self, we can reject both horns of the dilemma. (Caird, pp. 109–10)
What underpins this ‘better idea’ is Caird’s view that for Kant the transcendental self is active both in the activity of understanding of the empirical world and in the rational agency by which we seek to change it. Caird’s reference here to the conception of ‘the object’ as ‘essentially an existencefor-a-self ’ arises in the context of epistemology; but Caird earlier makes much the same point with respect to moral experience in his defence of Kant’s account of practical freedom: Kant defines practical freedom as the power of acting upon motives not derived from desire, and says that we have empirical evidence of it in our consciousness of the moral law as a motive. This however, might be explained away, if we were able to conceive the ego as a mere phenomenon, as merely one object among the other objects of experience, or if its consciousness were limited to its own states as such an object. But both of these assertions Kant attempts by his transcendental Regress to shew to be untrue. The attempt therefore to prove that Kant is inconsistent from his own point of view, rests upon a misunderstanding. Kant’s point is just that this is an experience which cannot be explained on the principle on which we usually explain objects of experience, i.e. without taking account of their relation to the subject. (Caird, p. 100)
It has to be admitted that this is not as unequivocal and helpful as one might like; and it is disappointing that Caird’s defence of Kant in this report is not as clear on this matter as he is in chapter iii (‘The Idea of Freedom’) of volume ii of The Critical Philosophy of Kant. None the less, it is clear enough that Caird holds that the ‘impassable gulf’ between pure practical reason and the will which lies at the core of Moore’s critique of Kant’s conception of practical freedom can be bridged over once the role of the will is conceived on the model of the role of the subject of knowledge in the ‘connecting process’ through which knowledge is possible. It has to be recognised, however, that Caird’s way of writing about this issue, like that of his teacher Green, readily invites the charge that all we are being offered here is a ‘Transcendental psychology’ which remains vulnerable to Moore’s critical dialectic (cf. 1898 dissertation, p. 156).
lxx
Editors’ introduction vii.2 bosanquet
Bernard Bosanquet (1848–1923) was an unusual philosopher. Like Caird, he studied under T. H. Green at Balliol College; he then held a post at University College Oxford. But he left in 1881 to pursue an independent career and took up an active role in public affairs. One aspect of this was his role, along with Sidgwick, in founding the short-lived London School of Ethics and Social Philosophy where, at their invitation, Moore lectured on ethics in 1898 and 1899.61 Bosanquet is now viewed primarily as an associate of Bradley, though less original; but this view fails to take account of his important writings in aesthetics and political philosophy. Bosanquet begins his report by noting that the dissertation presents ‘an exceedingly difficult problem’. The criteria for a Fellowship dissertation must include more than that the work shows merely promise – the work should show a ‘brilliant’ capacity for study and also that the candidate has taken up a ‘fruitful’ line of work. Bosanquet does not stint in his tribute to Moore’s ‘great earnestness in the pursuit of truth’ and ‘knowledge, ingenuity, and power of continuous persistence’ (p. 245). But he cannot, he says, ‘appreciate the theoretical point of view which the author has adopted. It appears to me to lie beyond the limits of paradox which is permissible in philosophy’ (p. 246). And there is no way in which this point of view can be isolated from ‘independent judgment of the criticism of Kant, of which, as explained and supported by the author, the Dissertation consists’ (p. 246). For, in accordance with ‘his just conception of the duty of the historian of philosophy’ Moore has linked ‘his positive views so closely to his critical standpoint, that it is wholly impossible to estimate them separately’ (p. 246). What then is it in Moore’s theoretical point of view that Bosanquet takes to be ‘beyond the limits of paradox’? It is the fact that Moore’s aim is ‘to dissociate Truth from the nature of Knowledge, and Good from the nature 61
Shortly after Moore’s election to his Prize Fellowship, but at a time when Moore must have seen Bosanquet’s negative report on his dissertation, Bosanquet wrote to invite Moore to meet him in London: It would be pleasant to me to improve your acquaintance, as you are good enough to lecture for the School of Ethics, even apart from the question of the Dissertation. I am very glad indeed that you have got such a good audience. I should be uncomfortable if I did not say that I hope we meet as fellow students of philosophy. I formed a clear opinion about the Dissertation, and will tell you what it is as well as I can. But I do not feel any mission to speak as an authority, or to ‘give advice’ in any sense of that kind. We will talk of our own views, and we ought to be able to understand each other. I thought the Dissertation full of interest, though as you have heard I took a strong view against its main contentions. (30 October 1898) (Add. MS 8330 8B/16/1)
Editors’ introduction
lxxi
of Will, so as to free Metaphysics from all confusion with Psychology’ (p. 247). But, Bosanquet protests, to abandon the idealist conception of consciousness as an ‘endeavour towards unity’, and decree a ‘divorce between the “good” and the nature of the will, as also between Reason with its law and this same nature’ is a ‘hopeless surrender of the most important connections’ (p. 247). It is clear that this protest is motivated by the Bradleian philosophy examined earlier (v, p. xxvi). Moore’s dissociation of truth and good from knowledge and will leaves them separate, and thus merely potentially related. Moore’s position is one in which ‘relational truth has been hypostasised as a self-subsistent form of Reality’: thus ‘the child [i.e. the Absolute] has been thrown away in emptying the bath’ (i.e. in getting rid of mere psychology). And matters are, if anything, made worse by Moore’s theory of propositions and concepts: ‘I confess that I feel a difficulty in regarding it as serious’ (p. 247). It cannot be said that Bosanquet provides any arguments for the assumptions which motivate these comments. The situation might be regarded as exemplifying a radical incommensurability between British idealism and Moore’s incipient analytical realism, though we have tried in our earlier comments (see v) to identify and clarify some of the areas for genuine debate that connect these philosophical projects. It was perhaps a pity that Bosanquet was not invited to report on Moore’s 1897 dissertation where Moore tries to combine his critical stance towards Kant’s practical philosophy with a version of Bradleian idealism. As it is, Bosanquet’s failure here to do much more than express his horror at Moore’s suggestions without showing why Moore is mistaken may have undermined his implicit advice that Moore did not merit election to a Prize Fellowship – advice which Bosanquet himself rather undermines by his correct prediction that if the electors were to elect Moore ‘it is quite possible that they may secure in him a considerable philosopher’ (p. 246).62 VIII THE TEXT OF THE DISSERTATIONS viii.1 the manuscripts
The draft manuscripts of Moore’s Trinity Prize Fellowship dissertations were the first item among the ‘Philosophical papers of G. E. Moore’ that were then offered for sale at Sotheby’s in December 1979. Fortunately for scholars, Moore’s papers were purchased by Cambridge 62
Bosanquet later reviewed Principia Ethica unfavourably for Mind 13 (1904), 254–61. He also advised Trinity College not to award Moore a Research Fellowship after the expiry of his Prize Fellowship in 1904 (see IV, p. xxii).
lxxii
Editors’ introduction
University and most of them are now to be found in the Moore archive in Cambridge University Library, along with other papers by Moore that have been subsequently donated to the library. The manuscripts of Moore’s Fellowship dissertations are, however, kept on loan in the Wren Library of Trinity College, as a token of gratitude for the financial assistance that Trinity College provided towards the purchase of Moore’s papers. One might well wonder why the successful 1898 Fellowship dissertation is not preserved in the Wren Library of Trinity College anyway, given the usual practice for the library to keep a copy of successful Prize Fellowship dissertations. The answer is that although this is the current practice, it was not the practice in Moore’s time. So if the college ever possessed a copy, it has been lost (but perhaps it will turn up one day in some old college cupboard or attic). Characteristically, however, Moore himself preserved the draft manuscripts which he had used to prepare the typescripts he submitted as dissertations, and it is these manuscripts which we have used to prepare this edition. Not surprisingly, the manuscripts have many corrections and, in the case of that for the 1898 dissertation, are in important respects incomplete. None the less we have been able to prepare an edition which, we believe, is reliable. We say more about the way in which we have reconstructed the missing parts of the 1898 dissertation below. Here we just describe the manuscripts as they are. 1897 dissertation
Most of this is written by Moore himself on folded half demy sheets of paper. The writing is such that the fold is on the left side (as with ordinary folded note cards) so that for each sheet there are in effect four sides of paper: a single side (1) on the front, two sides (2, 3) inside, and one (4) on the back. Moore initially wrote only on sides 1 and 3; he then used sides 2 and 4 for changes. Each folded sheet has a single number on side 1, but the numbers are not always continuous. Sometimes Moore has cut a folded sheet in half along the fold, and he then uses the top side of these single sheets, with changes on the reverse. These single sheets occur, numbered, in among the folded sheets. It is clear from the contents page that Moore used this draft manuscript to produce a clean manuscript copy of his dissertation on foolscap, presumably for his typist. Most of this clean manuscript does not survive, but in the case of the appendix on Professor Sidgwick’s hedonism we have both Moore’s initial draft and his clear copy.
Editors’ introduction
lxxiii
Because the page numbers of the draft as we have it are not continuous, and do not distinguish between the folded double half demy sheets and the single sheets which occur in among them, we have not included these numbers in our edition. But the details of the structure of the manuscript are as follows: Cover sheet: single unnumbered sheet. Preface: three unnumbered single sheets. Contents: single unnumbered half sheet. Introduction: eight folded double sheets, numbered 1–8. Chapter i (i) This begins with thirteen folded double sheets numbered 1–13. (ii) Three single sheets, numbered 1–3, though the text is continuous at this point (the change occurs on p. 35 of this edition at ‘The answer, then, to the question . . . ’). (iii) Thirty-seven numbered sheets, numbered 4–40; most of these are folded half demy sheets, but the pages numbered 31 and 35 are single sheets, though the text is continuous on both occasions. Appendix (i) The original version comprises eight folded double sheets, numbered 1–8. (ii) The tidy manuscript comprises nine foolscap sheets with writing on one side only, numbered 98–106, and with an abbreviated ending (see p. 94). 1898 dissertation
The draft manuscript is in this case made up from a variety of materials. Much of it started off as foolscap pages of a typescript of the 1897 dissertation, with changes added in handwriting by Moore. When making more radical changes, Moore sometimes pastes parts of old pages on to fresh sheets of paper and adds further new material; in other cases he inserts new sheets of paper, usually single sheets obtained as before by cutting folded half demy sheets in half. Substantial parts of chapters i and ii are missing. There are fresh page numbers for the introduction, for each chapter and the appendix. These page numbers are generally in sequence within each chapter, but not always, and where Moore has re-used pages of his previous typescript the pages have often been renumbered two or three times. We have omitted these page numbers from our edition since they are not informative. In this case there is no evidence that Moore systematically
lxxiv
Editors’ introduction
prepared a clean manuscript for his typist, but the existing manuscript of chapter ii is not in Moore’s hand (though it has a correction in Moore’s hand), so it must have been produced from a manuscript by Moore which has, alas, been lost. The details of the manuscript are as follows: Cover Single foolscap sheet in typescript, with handwriting by Moore on the verso. Preface Two unnumbered single half demy sheets, in Moore’s handwriting. Contents One unnumbered single half demy sheet, in Moore’s handwriting. Introduction (i) Fourteen foolscap sheets, numbered (i)–(xiv). These are a typescript of the introduction to the 1897 dissertation, with several changes in Moore’s handwriting. (ii) Six single demy quarto sheets in Moore’s handwriting. These are numbered (xv), 2–5. This material is new; see pp. 129–32 of this edition. Chapter i (i) Three sides of handwriting by Moore, on single demy quarto sheets numbered 1–3. (ii) Pages numbered 4–11 are missing. (iii) Thirty-three sides of single demy quarto sheets, numbered 12–44. Chapter ii (i) Two single demy quarto sheets, numbered 1–2 not in Moore’s handwriting (as applies throughout this chapter). (ii) Pages numbered 3–5 are missing. (iii) One single demy quarto sheet, numbered 6. (iv) One part sheet, with no number. For discussion of its location within this chapter see viii.2 below. (v) Fourteen single demy quarto sheets, numbered 25–38. Chapter iii This chapter is mainly built up from the foolscap pages of the typescript of the 1897 dissertation with many changes introduced by Moore in handwriting. The pages are numbered from 2 to 51, with number 6 twice. These are new handwritten numbers, the old page numbers of the typescript having been crossed out, and sometimes other handwritten numbers too; e.g. page 58 of the typescript was first renumbered 27 and then as 16. The final eight pages, numbered 44–51, are single demy
Editors’ introduction
lxxv
quarto sheets of Moore’s handwriting, which indicate that the end of this chapter is new material (see viii.3 below for more details). Chapter iv As with the previous chapter, this chapter is largely constructed from the foolscap pages of the typescript of the 1897 dissertation with changes introduced by Moore in handwriting. The numbering of these pages is rather unsystematic: the first page has no number, the next four are numbered 12–15, the next nine are numbered 33–41, the next group 16–24, and the chapter then ends with pages numbered 51–56 which include two new pages, 52a and 52b, inserted into page 52. Chapter v This is a manuscript by Moore with eleven single demy quarto sheets, numbered 1–11. Appendix on the Chronology of Kant’s Ethical Writings This is a manuscript by Moore with single demy quarto sheets, numbered 1–7; page 7 is much damaged. Appendix on Sidgwick’s Hedonism There is no manuscript for this, although it is listed in the table of contents. viii.2 the 1898 dissertation and ‘the nature of judgment’
Moore’s paper ‘The Nature of Judgment’ (Moore 1899) was published in Mind in April 1899. It is based on Moore’s 1898 dissertation, in particular on the ‘metaphysical’ parts of this dissertation, as Moore described chapters i and ii in his letter to Russell of September 1898 which was reproduced above (at the end of v). At the end of this letter Moore mentions that he is going to present a paper based on these parts of his dissertation to the Aristotelian Society on 9 December; and as the published paper indicates, that paper was ‘The Nature of Judgment’. Moore had used a similar procedure when composing his 1898 Mind paper ‘Freedom’ (Moore 1898). This paper is based on the 1897 dissertation, and by comparing the published paper with the dissertation one can see how Moore stitched together parts of the 1897 dissertation to construct his paper, with relatively minor alterations (we have provided details at the end of this section). The hypothesis that Moore used a similar procedure to construct ‘The Nature of Judgment’ is confirmed by the surviving manuscript of the 1898 dissertation, but in a way which is also disappointing: the manuscript lacks large parts of chapters i and ii, and the obvious explanation for these gaps is precisely that Moore extracted these pages in order to compose ‘The Nature of Judgment’. The
lxxvi
Editors’ introduction
result in this case is that the manuscript does not provide a text which can be compared with that of ‘The Nature of Judgment’. All is not lost, however. For at the start and at the end of most of the gaps in the manuscript there are passages which overlap with parts of ‘The Nature of Judgment’ and as a result it is possible to use ‘The Nature of Judgment’ in order to reconstruct the chapters of the 1898 dissertation on which it is based. The general shape of this reconstruction is supported by Moore’s letter to Russell to which we alluded above and also by Bosanquet’s examination report on the dissertation. In his letter Moore describes to Russell the ‘chief discovery’ he has attempted to vindicate in the new parts of the dissertation, namely that ‘an existent is a proposition’, and he combines this with the further claims that truth is an inherent property of propositions, that propositions are just relationships among concepts, and that concepts are ‘the ultimate elements of everything that is’ – all of which are readily recognisable as central themes of ‘The Nature of Judgment’. In his report Bosanquet broadly confirms this description of chapter ii of the dissertation; for it is precisely his complaint against Moore that in this chapter of the dissertation ‘the child has been thrown away in emptying the bath’ (p. 247) thanks to Moore’s account of the world as an immutable network of true propositions, whose truth depends on the relationship among the concepts which constitute them. We are sufficiently confident of our reconstruction of chapters i and ii to have included the chapters as reconstructed in our edition of the 1898 dissertation (albeit with clear indications as to how the text has been reconstructed). But it is appropriate to add here a few further comments to explain our procedure and to discuss briefly the results. Moore starts chapter i by announcing that he will here provide a critical assessment of Kant’s conception of Practical Reason, but most of the chapter is devoted to a critical discussion of Kant’s general conception of Reason, especially as developed in his Critique of Pure Reason (hence the chapter’s title – ‘On the meaning of “Reason” in Kant’). At the start of this discussion Moore observes that Kant connects the exercise of Pure Reason with the possession of a priori knowledge, and he then appears to embark upon a critical discussion of Kant’s account of the distinction between a priori and empirical knowledge. But it is at this point that the gap in the manuscript of chapter i opens up, after page 3 of the manuscript (which is complete) to the top of page 12 of the manuscript, where Moore begins a discussion of Kant’s distinction between analytic and synthetic judgements. Since page 3 of the manuscript ends with a paragraph which matches exactly a paragraph
Editors’ introduction
lxxvii
that occurs on page 184 of ‘The Nature of Judgment’, it is reasonable to suppose that the material which follows that paragraph in ‘The Nature of Judgment’ is that which had originally occurred in chapter i of the 1898 dissertation. This hypothesis is strongly confirmed by a fragment from the manuscript of the 1898 dissertation which certainly belongs in chapter ii (because of the handwriting involved) and which we have inserted at p. 170. For this fragment closely resembles a paragraph which occurs on page 189 of ‘The Nature of Judgment’ except that at the start of the fragment Moore remarks that it was in the previous chapter that the basis of the distinction between a priori and empirical propositions was discussed, and that in that discussion ‘an empirical concept was defined as one which could exist in an actual part of time’. It is precisely in pages 184–9 of ‘The Nature of Judgment’ that Moore discusses the a priori/empirical distinction and includes the definition of empirical concepts which is mentioned in this fragment as having been given in chapter i. So the fact that there is a close overlap between this fragment from chapter ii and a paragraph on page 189 of ‘The Nature of Judgment’ strongly suggests that the material taken from chapter i ends just before this paragraph. These considerations do not prove that the text of the missing pages 4–11 of the manuscript of chapter i of the 1898 dissertation comprises pages 184–9 of ‘The Nature of Judgment’. But Moore’s practice in constructing ‘Freedom’ from separate parts of the 1897 dissertation with very few changes to each part makes it reasonable to suppose that he followed a similar practice when constructing ‘The Nature of Judgment’ from the 1898 dissertation. This hypothesis is supported by the word counts: each page of Moore’s manuscript has about 250 words, so eight pages should comprise about 2,000 words. In fact the total for the text taken from ‘The Nature of Judgment’ is 2,180 words. This difference of 180 words (less than a page) falls within a reasonable margin of error and certainly does not suggest that Moore made major changes while constructing this part of ‘The Nature of Judgment’ on the basis of material from chapters i and ii of the 1898 dissertation. Moore begins chapter ii (‘Reason’) by announcing his intention ‘to expound and support the ultimate philosophical position, which was presupposed in the last chapter’ in the course of his critical discussion there of Kant’s conception of Reason. Since this was also his aim in ‘The Nature of Judgment’, it is no surprise to find that the manuscript for chapter ii is very much more fragmentary than that for chapter i. Moore clearly used a good deal of it in the composition of ‘The Nature of Judgment’.
lxxviii
Editors’ introduction
After his brief methodological introduction to the chapter, Moore quotes a passage from F. H. Bradley’s The Principles of Logic (Bradley 1883) and has just begun to discuss Bradley’s position when the manuscript of chapter ii stops short at the bottom of page 2. Since exactly the same passage from Bradley is quoted right at the start of ‘The Nature of Judgment’ (p. 176) and the discussion of it starts just as in chapter ii of the 1898 dissertation, it is reasonable to suppose that the critical discussion of Bradley’s position in ‘The Nature of Judgment’ (pp. 176–9) is taken from these pages. This is confirmed by the next surviving page of the manuscript of chapter ii, part of page 6 of the manuscript, which closely resembles a passage on pages 177–8 of ‘The Nature of Judgment’. There is then a large gap in the manuscript, from this part of page 6 until page 25 of the manuscript; however, as mentioned earlier in connection with chapter i, there is also an unnumbered fragment of a page which belongs somewhere in this gap. This fragment resembles a paragraph which occurs on page 189 of ‘The Nature of Judgment’ but also alludes to a discussion of the a priori/empirical distinction which has occurred in chapter i of the dissertation. We have therefore inferred that this paragraph indicates that the material taken from chapter i and incorporated into ‘The Nature of Judgment’ ends just before the paragraph on page 189 of ‘The Nature of Judgment’ which resembles this fragment; and hence that the material which follows this paragraph in ‘The Nature of Judgment’ came from chapter ii of the 1898 dissertation. It seems reasonable to assume that, in fact, the remaining part of ‘The Nature of Judgment’ comes from chapter ii of the 1898 dissertation. The only place where a significant change is likely occurs in the final paragraph of ‘The Nature of Judgment’ where Moore suddenly switches from talking about ‘propositions’, which is the idiom he has been employing up to that point and which fits well with his stated aim in chapter ii of the dissertation of showing the ‘fundamental importance for philosophy’ of propositions, to talking about ‘judgments’, which does not fit well with that aim and whose use he in fact criticises at the start of the chapter (p. 161). This change of idiom is no doubt to be explained by the need to accommodate the paper’s title; but it raises the question as to whether the whole paragraph was not in fact specially written for the paper as a summary of its conclusions. On balance, we think not – if one simply replaces Moore’s talk here of ‘judgments’ by ‘propositions’, the paragraph fits smoothly onto the preceding text which we take to belong to the dissertation, and we have therefore included this paragraph thus modified in our reconstruction of chapter ii (see pp. 173–4).
Editors’ introduction
lxxix
There is no way of demonstrating that this is correct. But an important test again is whether this remaining part of ‘The Nature of Judgment’ provides the right amount of material to fill the gap in the manuscript of chapter ii. The main gap in the manuscript of chapter ii (pages 7– 24) comprises 18 pages, which, at 250 words per page, should add up to about 4,500 words; the material identified in ‘The Nature of Judgment’ as coming from this part of chapter ii (including the final paragraph) adds up to 4,300 words. So there is a shortfall of 200 words – less than a page. This is what one would expect since the discussion of Kant’s antinomies on page 25 of the manuscript (p. 174 here) will have been preceded by an introduction to this topic which also connects it to the preceding discussion of Kant; but this introduction is unlikely to have filled a whole page of the manuscript, since the reason it is missing must be that the same page also included material Moore wanted to use for the composition of ‘The Nature of Judgment’. Hence, overall, adding these parts of ‘The Nature of Judgment’ to the existing manuscript provides a plausible reconstruction of most of this part of chapter ii of the 1898 dissertation. We provide below a brief summary of this account of the relationship between ‘The Nature of Judgment’ and the 1898 dissertation, alongside that for ‘Freedom’ and the 1897 dissertation: ‘Freedom’ (Mind, n.s. 7 (1898) 179–204)
1897 dissertation (page numbers in this volume)
179 two new paragraphs 179–83 183–8 188–94 194–201 201–4 (end)
20–3, 35–6 36–41 48–54 57–64 70–4
‘The Nature of Judgment’ (Mind, n.s. 8 (1899) 176–93)
1898 dissertation (page numbers in this volume)
176–7 178 178–83 184 (top) new paragraph 184–9 189–93 (end)
Chapter ii, 162–3 Chapter ii, 163–4 Chapter ii, 164–9 Chapter i, 135–9 Chapter ii, 170–4
lxxx
Editors’ introduction
viii.3 from 1897 to 1898: how moore changed his dissertation
Moore begins the preface to his second, 1898, dissertation with the following paragraph: The greater part of the Dissertation, which I submitted for examination last year, has been included in the present work. Some omissions and alterations, involving an important change of view have been made; and nearly as much again of new matter has been added. I have followed the suggestions of my examiners in attempting to distinguish more clearly between my own views and those of Kant; and, in deference to the same suggestions, I have added an appendix on the chronology of Kant’s ethical writings.
The structure of the dissertations differs a good deal. The 1897 dissertation is mainly constituted by one very long chapter (Chapter i ‘Freedom’), together with a short preface, a substantial introduction and an appendix on Professor Sidgwick’s hedonism. As Moore explains in his preface, he had originally intended to add two further chapters, one on ‘Kant’s attitude towards Hedonism and Practical Ethics’63 and one on his own ‘positive theory of practical applications’, together with appendices with ‘special criticisms on Green and Bradley’. In his preface he says that lack of time explains his failure to fulfil these intentions; but it is notable that, except for the brief discussion of ‘practical ethics’ in chapter v of the 1898 dissertation, he makes no great effort in his second dissertation to address these issues. The structure of the 1898 dissertation is very different. After a similar preface and introduction, there are four substantial chapters, the first two of which, chapter i (‘On the meaning of “Reason” in Kant’) and chapter ii (‘Reason’), are entirely new. The next two, chapter iii (‘The meaning of “Freedom” in Kant’) and chapter iv (‘Freedom’), are largely derived from the single long chapter i of the first dissertation, though its contents are very substantially reorganised (see below). Chapter v, ‘Ethical Conclusions’, is then a brief venture into practical ethics, followed by two appendices, a new one on the chronology of Kant’s ethical writings and the previous one on Sidgwick’s hedonism. Moore does not suggest that in the case of this second dissertation lack of time has prevented him from completing it; so we can take it as basically a finished piece of work. 63
In Moore’s note (b) for this dissertation (p. 63) he remarks: ‘This is the ground of Kant’s distinction between the Categorical Imperative, or objective Law, and the mere Maxim or subjective principle which will receive fuller treatment in my second chapter.’ Presumably the ‘fuller treatment’ of this distinction would have formed part of his discussion of Kant’s ‘Practical Ethics’.
Editors’ introduction
lxxxi
The ways in which Moore modifies and reorganises his long 1897 chapter on ‘Freedom’ into two shorter chapters for the 1898 dissertation reveal a good deal about the development of his philosophical stance at this time. Almost all the ‘Hegelian’ themes of the 1897 dissertation do not recur in the 1898 dissertation; and whereas the tone of the 1897 dissertation, though not uncritical of Kant, implies that Kant’s treatment of freedom remains of great importance, by 1898 Moore is much more critical. It is easy to set out the way in which Moore, in effect, used a cut and paste technique to extract chapters iii and iv of the 1898 dissertation from the 1897 dissertation: 1898 chapter iii: ‘The meaning of “Freedom” in Kant’
1897 ‘Freedom’
181–5 186–8 188–9 189 189–206 206–10
20–5 New 37–8 New 59–78 New
1898 chapter iv: ‘Freedom’ 211–27 227–8 228–30 230–1 (final paragraph)
38–55 New 55–7 New
What is immediately clear from this is, first, that chapter iv has been extracted from the middle of the 1897 chapter; and, second, that a substantial chunk of the 1897 chapter has been omitted altogether, namely pages 25–37. There is one other important change which is not apparent, namely that Moore has omitted the end of the 1897 chapter (pages 78–86) from the 1898 dissertation; it is replaced by pages 206–10 of chapter iii. In the material moved into chapter iv of the 1898 dissertation Moore is primarily concerned to discuss the libertarian thesis that the human will is a metaphysically special uncaused cause, and this discussion proceeds without any significant reference to Kant. So, since Moore clearly wanted to replicate with respect to Freedom the division he had made in chapters i and ii of the 1898 dissertation between a critical discussion of Kant’s
lxxxii
Editors’ introduction
account of Reason and the presentation of his own account of Reason, it made sense for him to move this discussion into a separate chapter in which he, in effect, set out his own essentially compatibilist account of the freedom of the will. The material omitted from the middle of the 1897 dissertation, by contrast, is closely focused on Kant; it deals primarily with Kant’s speculative metaphysics, and includes his sympathetic Bradleian interpretation of ‘Transcendental Freedom’. A year later Moore has lost faith in this conception of Transcendental Freedom; so it is omitted. Finally, the material omitted from the end of the 1897 chapter is Moore’s tentative attempt at a Hegelian argument to show ‘that what is real or free must be good, and that what is absolutely good must be real’ (p. 84). It is no surprise that this is dropped; the surprise was that it was attempted in the first place. What replaces it at the end of chapter iii of the 1898 dissertation is a new argument, designed to show that Kant provides no good reason for thinking that there is a special connection between goodness and freedom (pp. 206–10). Moore’s reputation in twentieth-century philosophy of course rests largely on Principia Ethica and the work he published after 1903. His exacting standards led to a reluctance to think highly of his early work and moreover a reluctance to publish or reprint it. But Moore’s dissertations engage fiercely with many of the main threads of discussion in philosophy at the turn of the century, both in Cambridge and on the continent, and provide a fresh perspective on the theoretical developments that led to Principia Ethica as well as a salient perspective on the origins of analytic philosophy.
References
Moore’s Fellowship dissertations and examiners’ reports are held in the Wren Library of Trinity College Cambridge, where they are catalogued as Add. MS a 247. Most of Moore’s other papers, including diaries and correspondence, are held in the Moore archive in Cambridge University Library, where they are catalogued as Add. MSS 8330 and 8875. Baldwin, J. M. (1902). Dictionary of Philosophy and Psychology, London: Macmillan. Baldwin, T. (1990). G. E. Moore, London: Routledge. (1993). ‘Appendix: “Principia Ethica” and “The Elements of Ethics”’, in Moore (1993a), 312–13. Bradley, F. H. (1883). The Principles of Logic, 2 vols., Oxford: Clarendon Press. (1893). Appearance and Reality, London: George, Allen and Unwin. Brentano, F. (1874). Psychology from an Empirical Standpoint, Leipzig: Dunker and Humblot. Caird, E. (1889). The Critical Philosophy of Immanuel Kant, 2 vols., Glasgow: James Maclehose and Sons. Gadamer, H.-G. (1979). Truth and Method, trans. W. Glen-Doepel, London: Sheed and Ward. Green, T. H. (1874). ‘Introductions to Hume’s “Treatise of Human Nature”’, in D. Hume, Treatise of Human Nature, eds. T. H. Green and T. H. Grose, London: Longman. (1883). Prolegomena to Ethics, ed. A. C. Bradley, Oxford: Clarendon Press. (1885). The Works of Thomas Hill Green, vol. i, ed. R. Nettleship, London: Longmans, Green & Co. Griffin, N. (1991). Russell’s Idealist Apprenticeship, Oxford: Clarendon Press. Hylton, P. (1990). Russell, Idealism and the Emergence of Analytic Philosophy, Oxford: Clarendon Press. Jacquette, D. (2004). Cambridge Companion to Brentano, Cambridge University Press. Korsgaard, C. (1996). Creating the Kingdom of Ends, Cambridge University Press. Kusch, M. (1995). Psychologism, London: Routledge. Levy, P. (1979). G. E. Moore and the Cambridge Apostles, London: Weidenfeld and Nicolson. lxxxiii
lxxxiv
References
McTaggart, J. M. E. (1896). Studies in the Hegelian Dialectic, Cambridge University Press. Moore, G. E. (1897). ‘In what sense, if any, do past and future time exist?’, Mind, n.s. 6, 235–40. (1898). ‘Freedom’, Mind, n.s. 7, 179–204. (1899). ‘The Nature of Judgment’, Mind, n.s. 8, 176–93. (1903a). Principia Ethica, Cambridge University Press. (1903b). ‘The Refutation of Idealism’, Mind, n.s. 12, 433–53. Reprinted in Moore (1993b), 23–44. (1904). ‘Kant’s Idealism’, Proceedings of the Aristotelian Society, n.s. 4, 127– 40. (1912). Ethics, London: Williams and Norgate. (1922). Philosophical Studies, London: George, Allen and Unwin. (1939). ‘Proof of an External World’, Proceedings of the British Academy 25, 273–300. Reprinted in Moore (1993b), 147–70. (1942). ‘An Autobiography’, in Schilpp (1942), 3–39. (1953). Some Main Problems of Philosophy, London: Allen and Unwin. (1959). Philosophical Papers, London: Allen and Unwin. (1991). The Elements of Ethics, ed. T. Regan, Philadelphia: Temple University Press. (1993a). Principia Ethica, rev. edn, ed. T. Baldwin, Cambridge University Press. (1993b). Selected Writings, ed. T. Baldwin, London: Routledge. O’Neill, O. (1975). Acting on Principle, New York: Columbia University Press. Passmore, J. (1957). A Hundred Years of Philosophy, London: Duckworth. Preti, C. (2008). ‘On the origins of the contemporary notion of propositional content: anti-psychologism in nineteenth-century psychology and G. E. Moore’s early theory of judgment’, Studies in History and Philosophy of Science, 39, 176–85. Regan, T. (1991). Editor’s introduction in Moore (1991), xiii–xxxviii. Russell, B. (1897). An Essay on the Foundations of Geometry, Cambridge University Press. (1903). The Principles of Mathematics, London: Routledge. (1959). My Philosophical Development, London: Routledge. Ryle, G. (1971). Collected Papers, vol. i, London: Hutchinson. Scanlon, T. (1998). What We Owe to Each Other, Cambridge, MA: Harvard University Press. Schilpp, P. A. (1942). The Philosophy of G. E. Moore, Evanston, IL: Open Court. Sidgwick, H. (1874). The Methods of Ethics, London: Macmillan. (1902). Lectures on the Ethics of Green, Spencer and Martineau, ed. E. E. C. Jones, London: Macmillan. Silber, J. (1960). ‘The Ethical Significance of Kant’s Religion’, in Religion within the Limits of Reason Alone, trans. T. M. Greene and H. H. Hudson, New York: Harper. Stern, R. (ed.) (1999). Transcendental Arguments, Oxford University Press.
References
lxxxv
Stout, G. F. (1896). Analytic Psychology, 2 vols., London: Routledge and Kegan Paul. Twardowski, K. (1977). On the Content and Object of Presentations, trans. R. Grossmann, The Hague: Nijhoff. Van der Schaar, M. (1996). ‘From Analytic Psychology to Analytic Philosophy: the Reception of Twardowski’s Ideas in Cambridge’, Axiomathes, 7, 295–324.
The 1897 dissertation The Metaphysical Basis of Ethics by George Edward Moore
Preface
I believe that almost the whole of this dissertation is original, so far, at least, that in some cases, the views expressed in it, and, in almost all, the particular exposition of those views and the arguments adduced in their support, are not consciously derived from any other author. My obligation to Kant will be sufficiently evident from the Dissertation itself. The criticism of his ethical doctrines contained in the first two chapters1 is based upon those of his works translated by T. R. Abbott (4th ed., Longman’s, 1889), and upon the two sections in the ‘Critique of Pure Reason’ dealing with ‘Freedom’, namely that in the ‘Dialectic’ entitled ‘Aufl¨osung der kosmologischen Idee von der Totalit¨at der Ableitung der Weltbegebenheiten aus ihren Ursachen’,2 and that in the ‘Methodenlehre’ entitled ‘Der Kanon der Reinen Vernunft’. Of those commentators on Kant, whom I have read, the one with whom I imagine myself to be most in agreement is Dr Hermann Cohen in his ‘Kant’s Begr¨undung der Ethik’ (Berlin, 1877);3 but Dr. Cohen’s excessive use of technical terms, nowhere clearly explained, and his dogmatic, rather than closely argumentative, style of exposition, have prevented me from availing myself of his work in detail. To Dr. Caird’s ‘[The] Critical Philosophy of Immanuel Kant’ (2 vols. Glasgow, 1889) I am, no doubt, largely indebted for my general conception of Kant, especially in my discussion of the ‘Ding an Sich’. But with Dr. Caird’s consistent use of ‘the unity of consciousness’, as a principle of explanation and criticism, I am prevented from sympathising very much by my far greater agreement with Mr.
1 2 3
As Moore explains below, there is in fact only one, very long, chapter. ‘Resolution of the cosmological idea of the totality in the derivation of occurrences in the world from their causes’. Hermann Cohen (1842–1918) was one of the founders of the Marburg school of neo-Kantian philosophy.
3
4
The 1897 dissertation
F. H. Bradley’s general philosophical attitude. It is to Mr. Bradley’s ‘Principles of Logic’ and ‘Appearance and Reality’ (2nd. edn. 1897) that I chiefly owe my conception of the fundamental problem of Metaphysics, though I would be far from claiming to have understood him rightly; especially, as I have used what I imagined to be his own principles, in order to attack ethical doctrines which he himself appears to hold. Finally, with regard to the frequency with which I have introduced Professor Sidgwick’s views, and that almost exclusively for criticism, I think it fair to state that I have criticised him precisely because he seems to me to have discussed the difficulties of ethics with such extraordinary care and acuteness, and because it is to his clear solution of many problems, which would almost certainly have embarrassed me, that I owe the possibility of having attempted, however unsuccessfully, to contest the finality of his conclusions. Had it not been for him, this essay would certainly have been far more rash and hasty, than it is. Lack of time has prevented me from completing more than a small part of what I had intended. I have been obliged to omit entirely (1) the discussion of Kant’s attitude towards Hedonism and Practical Ethics, of which I meant to make a second chapter, (2) the special criticisms of Green and Bradley, which were to form appendices, and (3) my positive theory of practical applications, which I had intended to form into a third chapter. The dissertation, therefore, consists solely of a general Introduction, a chapter on Freedom, with special reference to Kant, and a short Appendix on some of Professor Sidgwick’s fundamental positions, in his ‘[The] Methods of Ethics’.4 The chapter on Freedom could, with more time, have been profitably enlarged in several places, especially the last part, in which the brevity both of exposition and argument is greatly disproportionate to the importance of the topic discussed. My references to Kant are in every case to the pages of Hartenstein’s edition (8 vols., Leipzig, 1867). I hoped that that edition was sufficiently widely known, to justify this method of reference, which is a great convenience. ‘R.V.’ stands for the ‘Kritik der reinen Vernunft’ (vol. 3); ‘P.V.’ for the ‘Kritik der praktischen Vernunft’ (vol. 5); ‘G.’ for the ‘Grundlegung 4
As mentioned earlier (p. x) we follow Moore’s practice of using ‘ME ’ to abbreviate The Methods of Ethics. When citing passages from this book Moore often gives a page number without specifying that it is from ME, and we have not altered his main text to make this explicit, though where the page numbers are different in the seventh edition of ME, as opposed to the fifth edition which Moore uses, we have noted this in our footnotes.
Preface
5
zur Metaphysik der Sitten’ (vol. 4); and ‘M.’ for the ‘Metaphysik der Sitten’ (vol. 7).5 5
As explained earlier (p. x) Moore’s references and abbreviations are retained here, but we have added cross-references to the standard Prussian Academy edition of Kant’s works. In the case of the Critique of Pure Reason we usually just give the page number from the second ‘B’ edition; in the case of the other works by Kant we give the volume number of the Academy edition (AK). We also add references to the translation of the Critique of Pure Reason by Paul Guyer and Allen Wood (GW) and to the translation of Kant’s writings on Practical Philosophy by Mary Gregor (MG). So far as we have been able to ascertain, the translations Moore gives of passages by Kant are his own. Moore omits here to mention Kant’s ‘Kritik der Urtheilskraft’ (Critique of Judgment) which was collected into volume 5 of Hartenstein’s edition and to which he in fact refers twice, with the correct page numbers from Hartenstein’s edition but with no further details (see p. 55 notes a, b); he also refers once to Kant’s Prolegomena to any Future Metaphysics (see p. 28 note a), though in this case without the page numbers from volume 4 of Hartenstein’s edition.
Introduction
The scope of ‘Ethics’ has been very variously defined. Without prejudging any of the questions which it will be necessary to discuss hereafter, it may be stated summarily that the subject of the present essay is an enquiry into the nature of that which we denote by the terms ‘good’ or ‘what ought to be’. It may, perhaps, be well to confine the term ‘Ethics’, as does Professor Sidgwick, for example, to ‘the science or study of what is right or what ought to be, so far as this depends upon the voluntary action of individuals’.a Such a view is often roughly expressed by defining Ethics as ‘The Science’ or ‘Art of Conduct’;b and such is the scope of Aristotle’s Ethics, the book from which the term has been derived. ‘Ethics’ would thus take its place beside ‘Politics’, in the sense in which the latter is distinguished by Professor Sidgwick from Political Philosophy on the one hand and Political Science on the other. It is, perhaps, best on this view of its scope, to adopt Professor Mackenzie’s term and call it a ‘Normative Science’6 ; for the term ‘Art’ would seem to be most properly confined to the actual pursuit of some end or group of ends, in so far as such pursuit involves a systematic use of certain definite means, and not to include any statement of or enquiry into the rules by which such end or ends may be attained. If, for example, a book be entitled ‘The Art of Music,’ that title seems to denote the subject-matter described, just as does the title ‘Origin of Species’7 or ‘The a b
ME p. 4. Professor Sidgwick’s own view (ME p. 4) that his definition, as including Intuitionist conceptions, can not be brought under the notion of Art, seems inconclusive, since any intuitionist who pretends to have a system, and therefore an Ethics at all, must admit the subordination of means to end in some cases, and the assumption of several ultimate ends, not intimately connected, seems in no way contrary to the notion of an Art.
6
Cf. J. S. Mackenzie, Manual of Ethics, London: W. B. Clive, 1897, p. 4. Moore slightly misstates the title of Darwin’s famous book, On the Origin of Species, London: Murray, 1859.
7
6
Introduction
7
Cray-fish’,8 not the description or inquiry itself. Such an inquiry, in so far as it described the rules by which music of certain definite sorts, assumed to be the ends pursued in the Art, is actually to be produced, would rather itself be termed ‘Rules of the Art of Music’ or, in so far as it tried to give a reasoned connection between the rules by subordinating the various assumed ends to one main end, assumed to be the most general definition of the object pursued in the Art, ‘Theory of the Art of Music’. In this sense, the inquiry or the knowledge resulting might be termed a Normative Science, and such a ‘Normative Science’ Ethics as defined by Professor Sidgwick and as expounded by Aristotle, seems to be. The distinction between Art and Science may seem almost too obvious to need mention; but it seems to be almost uniformly overlooked, and to lead to a misapprehension of the relation of Theoretical to so-called Practical Philosophy. E.g., if Art be defined as a ‘scientific discipline’ whose aim is ‘the moulding of things by man’s activity’ while the aim of Science is said to be knowledge,a it becomes plausible to treat Practical [Philosophy] as a study really coordinate with Theoretical Philosophy. But though Art, as the actual doing of things, in which sense alone the moulding of things can be its direct aim, may really be treated as so coordinate with knowing – the distinction being that between volition and cognition in psychology; when Art is treated as a ‘scientific discipline’ its direct object becomes ‘knowing’ just as much as that of science. The object of Ethics, ‘what ought to be’, is certainly different from that of any science, but in as much as the direct aim of Ethics is to know this and not to do it, it becomes pure theory and is subordinate to the general conditions of knowledge. This common confusion between doing and knowing what to do has not a little connection with Kant’s distinction between Theoretical and Practical Reason. It is from this confusion, too, that Professor Sidgwick was able to sayb that Practical Philosophy including ‘the study of the fundamental principles of Ethics and Politics’ . . . ‘seems to hold a position in reference to Arts in general,.. similar to that which Theoretical Philosophy holds with reference to Sciences in general;’ although he subsequently defines a b
8
F. Paulsen System der Ethik (Hertz, Berlin 1889) p. 1. [Friedrich Paulsen (1846–1908) was a professor of moral philosophy at Berlin.] Printed lecture on ‘Scope of Philosophy’: 1897, p. 11. [This lecture is reprinted in H. Sidgwick, Philosophy: Its Scope and Relations, London: Macmillan, 1902; the passages Moore cites come from p. 28.] Moore here alludes to Thomas Huxley’s classic introduction to zoology, The Crayfish, London: Kegan Paul, 1879.
8
The 1897 dissertation
Arts (rightly, as I think) as ‘all departments of human activity, carried out systematically with reasoned adaptation of means to ends, for the attainment of some particular end, other than the knowledge applied a in the Art’. But it is to be noted that in ‘Normative Sciences’ or ‘Theories of Art’, the end, for the attainment of which rules are given, is itself merely assumed to be good. All the imperatives laid down are merely hypothetical. Aristotle, for instance, assumes that eÉdaimon©a is the end which both the individual and the state ought to pursue; and that Ethics and Politics are therefore only concerned to define what is meant by eÉdaimon©a and to lay down the rules by which it may be attained. Professor Sidgwick makes it an axiom that pleasure alone can be an end in itself. But if the ultimate end to the attainment of which all Arts are only means (the Art of Conduct like the rest) be itself made the object of enquiry; if it is attempted to discover not even what it is we ought to pursue, but why we ought to pursue anything at all; or what is meant by saying that a thing is good or ‘ought to be’ – this enquiry seems to be different in kind from that of a Normative Science, and, since it is the subject-matter of the present essay, I have preferred to entitle it ‘The Metaphysical basis of Ethics’ or ‘Metaphysics of Ethics’, although authority might have been found for calling it ‘Ethics’ simply. The ground to be traversed will thus correspond to that covered by Kant in his Critique of Practical Reason, and to the province which Professor Sidgwick would assign to Practical Philosophy; but it will appear later why I prefer not to use the word ‘Practical’ in the title of my subject. It may perhaps be doubted whether such a subject-matter exists at all; but, pending proof, the meaning of the distinction drawn between Metaphysics of Ethics and Ethics may be elucidated by the analogy of the following distinction between Science and Metaphysics. Science is wholly occupied in ascertaining the laws which obtain among data immediately given as existent. In so far as it answers the question ‘What exists?’, it only points to what may be called the ‘matter of knowledge’, and ascribes to it the predicate of existence, without inquiring what that predicate means in itself. For Metaphysics, on the other hand, approaching its problem through epistemology, the stress must lie on the meaning of the predicate:b a b
Italics are mine. All so-called Metaphysics does not perhaps approach its problem through epistemology; but in so far as it can claim to be rational knowledge, it would seem that it must not assume any type of existent as the ultimately real; and, if it proceeds to justification of its assumption, no other method seems open to it.
Introduction
9
its ultimate strength against Scepticism lies in the impossibility of making any proposition that shall not involve an assertion of ‘being’ and the validity of certain logical laws. For it therefore the question ‘What exists?’ must be answered by a reference to the only possible tests of the validity of knowledge – tests which can never guarantee the necessary existence of the ‘matter of knowledge.’ Hence Kant, with his conception of a Ding an Sich, and the later Idealisms, have spoken of an ultimate reality, however little they have succeeded in fully defining its content, in comparison of which all the matter dealt with by Science can be condemned as mere Appearance. A similar distinction seems possible in the region of ‘the good’; we can not only ask ‘what is good’ with the meaning ‘Tell us what are the data of experience to which we may apply the term “good”’; but it is at least possible that just as the question ‘what is being?’ cannot be answered by pointing to anything which is, or even to the whole world which seems to be, so the question ‘what is good?’ may involve a metaphysical enquiry to which no identification of the good with any one empirical datum, such as pleasure, or with that which under certain conditions leads to it, can ever by the nature of the case furnish an adequate answer. It is possible pleasure may be good, and that a maximum of pleasure may be what we ought to aim at; and yet nevertheless to identify ‘the good’ with pleasure or ‘what ought to be’ with a maximum of pleasure may involve a fallacy like that of identifying a chair or a table, because they are, with ‘being’, or matter, when it is held to be real, with reality. Indeed such definitions must become tautologous, unless (1) either it be maintained that by ‘A maximum of pleasure is the good’ is meant ‘A maximum of pleasure, or that which is productive of a maximum of pleasure, is what is meant whenever the word good is used’ – a statement which is flagrant contradiction with the facts. For it is impossible to take refuge in the hypothesis that people have erred as to what they meant, since error implies that some definite meaning is attached to each of the terms falsely conjoined. Hence it must be admitted people meant something definite by ‘virtue’, when they said it was the good; and as the error, if error there be, can only lie in the identification of it with good – not in mistaking it for maximum pleasure.9 (2) Or10 else it must be allowed that good has a meaning of its own, not exhausted by any empirical concept or any definition involving such. The same dilemma 9 10
Moore’s MS has a detached comment which appears to apply here: ‘Is pleasure “what I ought to pursue”. What does “ought to” mean ?’ Russell: ‘The disjunction is not obviously exhaustive.’ (Russell added a few marginal comments to Moore’s manuscript; we have included these comments among our notes with the prefix ‘Russell’.)
10
The 1897 dissertation
may be put with reference to Aristotle’s definition of the good – nerge©a katé retn11 etc. Is this a significant proposition? If so, then something other than nerge©a katé retn must have been meant by ‘good’. But it is impossible that that something other can have been any other empirical definition of good; since the significance of Aristotle’s doctrine rests on his refusal to identify nerge©a etc. with any such ordinary conception, e.g. pleasure. Good, in short, must have some meaning of its own, apart from any reference to empirical concepts, unless we are to mean e.g. by ‘pleasure is the good’ only that we shall limit the word good to pleasure; and in that case the question ‘Is pleasure good?’ must be meaningless. Perhaps the fallacy involved in all empirical definitions of the good may be quite plainly exhibited as follows. One party, let us suppose, holds that the good is pleasure, another party holds that it is self-realisation, it being admitted that self is to be defined as in psychology by reference to the results of empirical enquiry. Our contention is that if each party is to hold that the other is wrong, each must presuppose ‘the good’ to have some other meaning than that which by their definition they assign to it. Either the controversy must be of the same nature as if one party maintained that light was ether-waves and the other that it was air-waves, in which case it is plain that each must have some direct knowledge of what they mean by light as distinct either from ether-waves or from air-waves; or else it must be sought to maintain that the controversy is merely ‘about words’ – about the definition of the term ‘good’. But in this latter case, either each party must be content to allow that the other may also be right – in which case both propositions become insignificant; or else there must be some reference to a standard by which the right meaning of words can be determined. And this latter contention involves the term that is to be defined: the argument is circular. Nor can this conclusion be escaped by the contention that we should have said the true meaning of words. For mere words have no true meaning, in the only sense in which truth can be established without reference to an ethical standard. Truth or error consist only in the consistency or inconsistency of the relations thought to hold between real objects, with those which actually obtain. The true meaning of a word can only be the sense which it ought to convey either absolutely or because it is usually used to convey it. But (1) it is impossible for either party to contend in the present instance that ‘pleasure’ or ‘self-realisation’ are what the term good is usually employed to convey. For in the first instance their propositions would be palpable contradiction with the facts; 11
‘Activity of soul in accordance with virtue’, Nicomachean Ethics 1098a16.
11
Introduction
and in the second place they would not be ethical but purely lexicographical, and in order to bring them into the former sphere they would be forced to consider whether the vulgar notion of good was a true one or not; for which purpose the contention that good is a mere term must be wholly given up. If on the other hand (2) it is maintained that the word ‘good’ ought absolutely to be used in this or that sense; the moral standard implied, if it is not the lexicographical one of ordinary usage, must be that of truth, and truth, as said above, implies the assertion of a relation between real things, in which case ‘good’ must have a meaning other than that assigned to it. In other words, if the discussion what a word ought to be used to be12 mean or what it really does mean is to be of philosophical importance, it must always be implied that there is some real object or relation, perhaps not sufficiently disengaged by former analysis, for which a fixed term is required for the purposes of scientific discussion, and to which the term in question has always had reference, although perhaps confusedly. There can be no true meaning of words in a strict sense; but, if we are to discuss the truth about reality, it may be necessary to confine a word to one meaning; and when the word is already in common use, it must be shewn that the meaning proposed has been usually attached to it. It is in this sense that the present essay proposes to discuss the meaning of ‘good’ and ‘ought to be’: and it is maintained that there is a real object or relation corresponding to them, and to which they have always had more or less definite reference, and the identification of which with anything empirical always involves a tautology. We may, then, I think, fairly consider it proved that ‘good’ must have a meaning other than an empirical one; though freely admitting that the discovery of the empirical objects to which the term may be truly applied is highly important. All that we contend is that the proposition ‘This or this directly given object, or that which stands in such and such a uniform relation to this object, is the good’ is not capable of empirical13 verification. It is not of the nature of empirical or scientific propositions, which uniformly include some particular phenomenon in a larger class – but a class of phenomena only: e.g. when we say man is an animal, where the two notions conjoined are both of them defined by directly given empirical characteristics. To take a final instance: if it be said that the good is only that which pleases, that proposition is certainly significant, for we know what it is to be pleased, and it tells us that all things which cause us to have that feeling of pleasure are good. But if the proposition is to remain 12
Sic; this second occurrence of ‘be’ is clearly a slip.
13
Russell: ‘Purely empirical?’
12
The 1897 dissertation
significant and not to become tautologous, it must not be maintained that good is a merely empirical conception. It may be attempted by pointing e.g. to a virtuous man: it may be said ‘See, he is good; and he does produce a balance of pleasure over pain.’ But this tells us no more than that the virtuous man is to be included in the class of things that produce more pleasure than pain: it tells us nothing whatever about the good except that where we apply that predicate it can be shewn that a balance of pleasure is produced. But our whole question was of whether we were justified in applying that predicate and the present answer is a mere petitio principi. The justice of the above criticism seems to be tacitly implied in any ethical system which coordinates ‘what ought to be’ with ‘what is’ as if they were two fundamentally opposed conceptions; but its importance for speculative purposes seems much too generally neglected.14 It is only if the object of Ethics is conceived to be purely practical, to tell you ‘what you ought to do’, as is definitely announced by Aristotle, that there can be any excuse for assuming an Ethical axiom. If the object of the enquiry is knowledge for its own sake no axiom the denial of which would not involve an affirmation can be admitted; and though the results so obtained may be very inadequate to tell us how to live, quˆa knowledge, their value depends upon their truth and not upon their use. The main object of the present essay, then, is a purely speculative one; though it will be attempted later to shew how the theory here to be developed is connected with the Normative Science of conduct. So far I have been engaged in defining what I mean by Metaphysics of Ethics or the attempt to give a ‘transcendental’ meaning to good. This serves to distinguish the present enquiry from all such as start from an empirical definition of the good, and are only further concerned with the attempt to shew how their principle may be consistently applied and what are the means of attaining their end. It is Kant who seems most clearly to have recognised this distinction and to have done most towards a systematic exposition of the nature of the concept ‘good’;a,15 but since he states the problem in a different way some preliminary statement of my grounds of dissent may not be out of place. a 14 15
Pref to P. V. Russell: ‘This point should be explained’. It is not obvious which passage Moore is here referring to; presumably it is Kant’s discussion of the concept of the good at AK 5: 9ff. (MG 143ff.) despite the fact that Kant’s account of this concept is very different from Moore’s.
Introduction
13
Kant considers the discussion to fall within the province of Practical Reason, just as Professor Sidgwick would assign it to Practical Philosophy: and this nomenclature has been so universally adopted, that a discussion of it seems necessary, to avoid misunderstanding, at the outset of the present essay. In its more restricted sense Practice seems to have an essential reference to human nature; when some might even think that this exhausted its significance. On this understanding, Practical Philosophy should be confined to the determination of what ‘we ought to do’ in Ethics and Politics. But most would admit that a complete answer to this question implied primˆa facie a discussion of ‘what ought to be’, whether we can do it or not.a And when this further problem is taken into account, as by Kant it certainly is, the meaning of ‘Practical Philosophy’ must necessarily be extended beyond the sphere of mere human action, and embrace the discussion of what Nature ‘ought to’ produce, or the ‘end’ of the whole process of things in time. With this widened conception of ‘Practice’, the discussion would however still assume that what ought to be necessarily was not; that it was only something which might be in the future. And this contrast between ‘what is’ and ‘what ought to be’, as involving a distinction in time, is generally assumed to be fundamental. But from the point of view of the present essay no distinction involving Time can be assumed as fundamental. Time seems to be a merely empirical datum which, as such, must be discarded, when a logical test of reality is set up.16 It seems therefore necessary that a discussion which professes to deal with the most fundamental conceptions of Ethics, should start without prejudging the question whether ‘what ought to be’ really is or not. In other words, the conception of ‘good’ seems logically more fundamental than that of what ought to be if the latter is understood to imply anything empirical – in short, than any conception in which ‘practice’, by the largest extension of its meaning, can be involved. On the other hand it will be attempted to shew that what ‘ought to be’ may rightly be taken as identical with ‘good,’ the limitation of the former term being only due to a confusion between ‘being’ as existence in time and ‘being’ in a transcendental sense; for what only ‘ought’ to exist in time certainly may primˆa facie, and (as will appear if a
Cf. Professor Sidgwick p. 33 [ME 7th edn, p. 35]. I do not however accept his restriction, that ‘ought’ in the wider sense ‘merely implies an ideal or pattern which I ‘ought’ – in the stricter sense – to seek to imitate as far as possible (see later fn.). [There is no later footnote relevant to this issue; but Moore does discuss it on pp. 45–6.]
16
Russell: ‘At this point you might refer to your remarks at the Arist’ i.e. Moore’s paper ‘In What Sense, if any, do Past and Future Time exist?’ which was read to the Aristotelian Society and then published in Mind, n.s. 6 (1897) 228–40.
14
The 1897 dissertation
the present discussion is successful) must, ‘be’ in the transcendental sense – or, to adopt the language of modern Idealism, must be real. It will follow indeed that what ought to be includes also what we ‘ought to do’ or what ought to be done; but that because it is ‘what ought to be’, and not vice versˆa. That this rejection of ‘practice’ is justified seems proved by the mere conceivability of a ‘good’ which is not merely to be: the wider notion is always logically the more fundamental. On our view, then, the distinction between ‘what ought to be’ and ‘what is’ is not such as to justify a fundamental division of Philosophy or of the Reason into the two parts ‘Theoretical’ and ‘Practical’. The persistence of this division is probably to be ascribed to Aristotle, whose Ethics, however, are professedly not metaphysical. In fact the distinction between ‘what ought to be’ and ‘what is’ is not as great as the distinction between what is and what exists; and the ‘practical’ is so far from representing ‘what ought to be’ that it is inextricably bound up with mere existence. It does indeed imply the transcendental concept; our very first point is that it necessarily leads thereto: but as soon as it is recognised that this concept to which it leads is of a wider nature than itself, it appears as much a misnomer to call the inquiry into that concept ‘practical’, as to call any branch of science ‘metaphysics’. So far therefore as general philosophical scheme goes, the standpoint here taken up seems to agree most with that of Plato. The ‘good’ is to be considered as an Idea pkeina ts oÉs©as17 perhaps – (the meaning that can be attached to such a phrase will be subsequently considered):18 but the really essential distinction is between Ànta and gign»mena19 and the ‘practical’, by its very notion, belongs to the latter class. Kant, however, is a much more convenient starting-point for a Metaphysic of Ethics than Plato: partly because his exposition is so much fuller and more systematic; partly because he was able to make use of so much more speculation; and partly because his direct influence upon modern thought seems so much greater. And much of what he said seems to lead directly to the view above expressed. He himself says that ‘freedom’ of the will rests upon ‘transcendental freedom’, which latter is a speculative notion, though, according to him, not demonstrable by speculative means; and the fault in his statement seems to proceed largely from his not sufficiently distinguishing between the relation of reason to consequent and 17 18 19
The phrase comes from Republic 509b, and is often translated as ‘beyond being’. Moore returns to this theme on p. 83. This distinction, usually interpreted as a distinction between ‘being’ and ‘becoming’, is drawn in various ways in Republic e.g. 521d, 525b, 525c, 526e (there is no canonical context in which Plato directly contrasts Ànta and gign»mena).
Introduction
15
that of cause to effect. Our will, as he defines it, is a power of being cause;20 but what he calls a ‘pure will’ cannot be a ‘will’ at all, since by his own definition of causality as applying only to relations of phenomena it must be put out of court by the mere fact of its ‘purity’. This is the old objection raised against his statement of the relation between phenomena and things in themselves – an objection which seems fatal to the basis of Schopenhauer’s professedly Kantian system. He himself, in answer to a critic of the ‘Grundlegung’, gives his reason for not taking ‘the good’ as the fundamental conception in his Metaphysic of Ethics. This however seems to amount to no more than that ‘the good’ necessarily implies an ‘object’ to which it may be applied.a Now an ‘object’ must, according to the ‘Pure Reason’,21 be given in intuition. The ‘good’ must therefore either contain some empirical notion or it must refer to the object of an ‘Idea’, the existence of which we cannot ratify. But in the ‘Practical Reason’, he proceeds to infer the existence of ‘objects’ of Ideas from the moral law. His insistence, then, on making the ‘law’ and not the ‘object’ the fundamental notion of his ethical philosophy seems to be due to the fallacy of mistaking the datum from which you infer, for the logical ground of the inference.b There is nothing in what he says to shew that the notion ‘good’ is not logically more fundamental than the moral law: he only insists that if it were not for the moral law we should know nothing of that notion. In short he seems to confuse the ratio cognoscendi with the ratio essendi, a confusion against which he himself gives a warning elsewhere.c It is the latter which in philosophy must always form the basis of distinctions; and therefore it would be more appropriate to distinguish the ‘Critique of Practical Reason’ as ‘philosophy of the good’ from ‘philosophy of that which is’, than by means of the ‘practical law’, which is only the datum from which it sets to work. Enough has perhaps been said in preliminary justification of our subject matter. As Kant himself says, philosophy differs from Mathematics in that it cannot start with definitions because it has no intuition whereon to base them: a philosophical treatise must justify and explain itself by its whole course. I propose, therefore, in this essay, to criticise Kant’s Metaphysics of a b c 20
21
P. V. p. 67 [AK 5: 63, MG 191]. He seems himself to recognise this, P. V. p. 31 [AK 5: 29, MG 163]. P. V. p. 4 note [AK 5: 5, MG 140]. Sic – presumably ‘being a cause’ is intended. Moore provides no reference for this imputed definition, but the opening of section iii of The Groundwork of the Metaphysics of Morals is characteristic: ‘Will is a kind of causality of living beings’ (AK 4: 446, MG 94). I.e. Kant’s Critique of Pure Reason.
16
The 1897 dissertation
Ethics in some detail, as a means of making plainer my own view. I shall try to convict him of inconsistencies, which seem not unimportant, since they seem to have been adopted from him by later writers such as Green, Bradley, and Mackenzie. Finally I shall state my objection to Hedonism, as represented by Professor Sidgwick, who puts it in the only form in which it appears to me at all defensible.
chapter i
Freedom
17
contents of chapter i Moore’s MS22
Heading Kant a declared Determinist. Freedom not ‘naturally’ possible Meaning and ambiguity of Causality. Ding an Sich as Free Cause Ding an Sich implies positive knowledge of real Free Cause, which is no cause Reality thus has Transcendental Freedom. But Kant applies it to ‘Will’ ‘Liberty of Indifference’ an attempt to find in the world of Experience an absolutely self-caused cause. Impossibility of completely predicting human action Libertarianism means uncaused choosing. No defence of it. Sense in which empirical objects have ‘self’ and are free Notion of ‘organism’, as possessed of freedom, untenable Kant’s Freedom takes account of freedom, in only sense, in which we have found it to belong to natural objects, but his Freedom does not belong to ‘Will’ either Reasons why he thought it did. Confusions exposed ‘Pure Will’ impossible as implying time. No Practical Freedom
22
23
1–5
This edition23 pages 20–5
5–10
25–9
10–23
29–35
24–8
35–9
28–31
39–41
31–45
41–51
45–50
51–4
50–5
54–7
55–64
58–64
65–74
64–70
74–9
70–73
These are the page numbers of the clean manuscript Moore prepared for his typist. This manuscript does not survive, except for the pages of the appendix entitled ‘Professor Sidgwick’s Hedonism’ which is numbered ‘98–106’. We have inserted Moore’s section headings into the text of Chapter i at what appear to be the appropriate places, although we cannot be certain that these places are correct.
18
19
Freedom Kant’s connection of Freedom with ‘End’, vitiated by being referred to ‘Will’ Notion ‘ good’ commonly connected with freedom, and really capable of connection with Transcendental Freedom
79–85
73–8
85–97
78–86
chapter i
Freedom
kant a declared determinist. freedom not ‘naturally’ possible In beginning a discussion of Kant’s notion of ‘Freedom’, which he himself considers to be essentially connected with his Ethical system, it seems most important to emphasize the fact that, so far as his express statements are concerned, he accepts unconditionally the view of Determinism and rejects that of ‘Freedom’, in the only sense in which the two have been generally discussed by English thinkers. In ordinary controversies on the subject, no such absolute distinction is drawn between two kinds of ‘causality’, two kinds of ‘determination’ (Bestimmung – the sense which is implied in ‘Determinism’), two kinds of ‘possibility’, or finally an ‘intelligible’ and an ‘empirical’ character, as is drawn by Kant. Professor Sidgwick, indeed, puts the question in such a form that Kant’s answer would probably have to be on the Libertarian side; but this result seems only to be obtained at the cost of the above-mentioned ambiguity. ‘Is the self ’ he says ‘to which I refer my deliberate volitions a self of strictly determinate moral qualities, a definite character partly inherited, partly formed by my past actions and feelings, and by any physical influences that it may have unconsciously received; so that my voluntary action, for good or for evil, is at any moment completely caused by the determinate qualities of this character together with my circumstances, or the external influences acting on me at the moment – including under the latter term my present bodily conditions – or is there always a possibility of my choosing to act in the manner that I now judge to be reasonable and right, whatever my previous actions and experiences may have been?’a Now to the first half of the first alternative, ‘Is the self to which I refer my deliberate volitions a definite character (etc.)’ Kant would be compelled to give what Professor Sidgwick considers to be the Libertarian answer a
ME pp. 61–2.
20
Freedom
21
of ‘No’, because there seems to be implied in it the alternative of what he would call an ‘intelligible character’; though even here, he would be in some doubt, because it seems implied that the ‘intelligible character’ cannot be ‘of strictly determinate moral qualities’ or ‘definite’. And with this presumption that Professor Sidgwick accepted his distinction, he would probably answer ‘No’ to the second half ‘Is my voluntary action, . . . at any moment completely caused by the determinate qualities of this character’, although, had that question come by itself, his answer would probably have been ‘Yes’, since the sequel shews that when Professor Sidgwick says ‘completely caused’ he is only thinking of what Kant calls ‘natural causality’ (Natur-causalit¨at). So, too, in answer to the second question, he would only say ‘Yes’, on the presumption that Professor Sidgwick might mean by ‘possibility’, intelligible as well as empirical possibility. But when Professor Sidgwick goes on to exemplify the determinate view by reference to the principle of causality as employed in the Natural Sciences; when he says (p. 62) that ‘the substantial dispute relates to the completeness of the causal dependence of the volition upon the state of things at the preceding instant’, there could no longer be any doubt that only that causality was meant of which Kant had been at such pains to prove the universal validity in the Critique of Pure Reason; and only some reason for surprise that reference should have been made to the possibility of a self with any other than a psychological character. Professor Clifforda gives a statement of the doctrine of Free Will as commonly understood, which seems so clear as to be worth quoting. ‘Whenever a man exercises his will, and makes a voluntary choice of one out of various possible courses, an event occurs whose relation to contiguous events cannot be included in a general statement applicable to all similar cases. There is something wholly capricious and arbitrary, belonging to that moment only; and we have no right to conclude that if circumstances were exactly repeated, and the man himself absolutely unaltered he would choose the same course.’ Now this doctrine Kant would absolutely condemn. In fact if Determinism only means that all men’s actions conform to the laws of nature, and so with the progress of psychology, could ultimately be predicted as certainly as the motions of the planets (and this is what Professor Sidgwick seems obviously to mean, and what is usually meant by it) Kant would have no hesitation in calling himself a Determinist. ‘All actions of man in Appearance’ says he ‘are “determined” (bestimmt) by a
Essay on ‘Right and Wrong’ in ‘Lectures and Essays’ (1886) p. 318. [W. K. Clifford, eds. L. Stephen and F. Pollock, Lectures and Essays, London: Macmillan, 1886, repr. 1901, vol. ii, pp. 126–7. The emphasis in the quoted passage is Moore’s.]
22
The 1897 dissertation
his empirical character and the other contributory causes according to the order of nature, and if we could investigate all Appearances of his choice (Willk¨uhra ) to the bottom, there would be no single human action, which we could not foretell with certainty and recognise as following necessarily from its preceding conditions’.b Freedom, according to him, is absolutely impossible, if reality is ascribed to events in space and time. ‘Since the thorough-going connection of all Appearances in a context of Nature is a law that admits of no exception, this must necessarily upset all Freedom, if one were determined to cling obstinately to the Reality of Appearances. Hence also those, who in this latter respect follow the common opinion, have never been able to succeed in uniting Nature and Freedom with one another.’c Now the dispute between Libertarians and Determinists is undoubtedly conducted in general by those who do ‘follow the common opinion’ of ascribing reality to what Kant calls Appearances, e.g. matter as treated in Physics and mind as treated in Psychology. In so far as Determinism is regarded as bringing the phenomenon of Will into harmony with the results established by experimental investigation of Nature, it can only be a doctrine concerned with what Kant calls Appearances, and as such the above quotations seem to prove his unqualified adherence to it. It would, in fact, appear absurd to the ordinary champion of Free Will, to declare that ‘actions . . . which never have happened and perhaps will not happen’d are yet ‘necessary’; and yet it is only on this basis that Kant is prepared to defend Free Will. If this be absurd, there is no choice but Determinism. Kant, in fact, uses ‘necessity’ here in a totally different sense from that in which common sense usually understands it. ‘“Ought” expresses a kind of necessity and connection with reasons, which is found nowhere else in the whole of Nature. . . . It is impossible that anything else ought to happen in Nature, than what in all these temporal relations actually is; indeed “ought”, if we only look at the course of Nature has absolutely no meaning.’e If you declare future action to be ‘necessary’ the ordinary man would suppose you must mean ‘it will happen’ – that you are predicting something according to the Laws of Nature; if you do mean that ‘perhaps it won’t happen’, he would say that you are using terms inaccurately: you ought to have said it was only probable or possible. But meanwhile it is sufficient to point out that Kant does say this absurd thing; a
b c
‘Willk¨uhr’ seems to correspond to the ordinary notion of ‘will’, in the sense in which its freedom is disputed – that is as a subject matter of psychology. ‘willk¨uhrlich’ may perhaps be translated ‘arbitrary’. The translators of Hegel use ‘caprice’ and ‘capricious’; but these appear to have a disparaging sense, which primˆa facie is not justified. R.V. pp. 380–1 [B577–8, GW 541; the emphasis here is Moore’s]. d R.V. p. 380 [B576, GW 541]. e R.V. p. 379 [B575, GW 540]. R.V. p. 373 [B565, GW 535].
Freedom
23
and that from this second meaning of ‘necessity’ there follows a second meaning of possibility also. Since that which according to the Laws of Nature is only possible can be called necessary, that which according to the Laws of Nature is absolutely impossible, may, from Kant’s point of view be regarded as ‘possible’.a It is only on this supposition of the possibility of the impossible, that Kant could have answered ‘Yes to Professor Sidgwick’s second question. What, then, if Kant is a Determinist, does he mean by that Freedom, the reality of which he asserts? It is noteworthy that he first approaches the problem in connection with the Critique of Speculative Reason, discussing the application of that notion of causality, to defend the objectivity of which against Hume was one of the chief motives to the formation of his Critical Philosophy. The principle of causality is a principle by which the existence (Dasein) of one object is necessarily connected with the existence of another. Hume had not denied the necessity of some kinds of relations between objects; for instance he seems never to have suspected any insufficiency in the grounds we have for saying that five apples and seven apples are equal to twelve apples. This, he would have said, is a relation which, while the objects remain the same, must also remain the same, because, whether there ever were twelve apples or not, the mere idea of them would always stand in the same relation to the ideas of seven and of five. He recognises, in fact, that the certainty of a relation of this sort, is not, like the certainty of the existence of any object, to be tested by our experience of it. Though we could never learn that 12 = 7 + 5, unless we had seen twelve objects, yet our ground for saying that 12 = 7 + 5, is not only that we have seen cases in which they are, as, according to him, our only ground for saying that there are apples, is the perception of them. He would entitle us here to the further statement, that if there are twelve objects of any sort, it is impossible but that they should be equal to seven and five of the same sort. Our statement that 12 = 7 + 5, does not only mean, that in all cases which we have observed, this is so; as would (according to him) such a statement, as that, if there are apples, there were apple-blossoms before them. In the former instance, he would say a priori that there never could be a case in which 12 was not equal to 7 + 5; whereas in the latter, he would utterly deny our right to make a generalisation a
The words of R.V. p. 379 [B576, GW 540] ‘Now the action must undoubtedly be possible under natural conditions if it is conformed to the “ought”’ must be understood to mean, that any actual action, which was in accordance with ‘ought’, must also have natural possibility i.e. have been capable of prediction according to natural laws: not that for any conceivable action to be moral, it must also be naturally possible. So in M. p. 18 [AK 6: 221, MG 376], it is obvious that the ‘morally possible’, the ‘permitted’, may be something which you cannot actually do.
24
The 1897 dissertation
universal. According to him apples are preceded by apple-blossoms, when they are; but that they must be, or that anything must ever be so connected with anything else, is what, he would say, we have no ground to be certain of. In fact Hume treats propositions about relations of quantity, as if they were purely analytic. Kant saw that they were not so. Only insignificant propositions can be purely analytic. Mathematical propositions are therefore synthetic; and at the same time, as Hume admitted, they are a priori. Kant’s object, therefore, in the Critique of Pure Reason, is to consider the validity of a priori synthetic propositions, as well those mathematical ones, which Hume thought to be analytic, as those expressing a causal relation to which Hume, because he saw that they were not analytic, had denied any a priori necessity. The precise manner of Kant’s Deduction is not here to the point. What the above brief account of the question at issue is designed to bring out, is his distinction between the category of causality and the mathematical categories, which makes him call the former ‘real’ or ‘dynamical’; for it is upon this distinction that he conceives to rest the possibility of Freedom. It is solely in virtue of the dynamical categories that we are able to say that if one thing exists, another must exist. The mathematical categories only tell us that between things appearing in space and time, certain relations must hold; they do not allow us to conclude from the existence of any one thing to the existence of any other. Hence, when it appears that, in trying to obtain a complete notion of the world as ‘determined’ by the various categories, we are involved in antinomies, the contradictory propositions resulting from such an application of the mathematical categories must both be false, since both relate solely to the temporal and spatial properties of things. But with the dynamical categories the case is different. These profess to deal with the existence of things. Now Kant always asserts the existence not only of appearances, but also of something of which these are appearances. It is therefore possible that any proposition which concerns the existence and not merely the relations of things, may be true with regard to these Things-in-themselves. And hence, though the contradictory propositions resulting from the attempt to obtain a complete view of the world under the dynamical categories, must both be false, if it is presupposed that we are talking only of things existing in time, it is possible that one of them may be true, if we are entitled also to take into account the existence of a Thing-in-Itself. It is with regard to the proposition that things have an unconditional or first cause, i.e. a cause which is not itself the effect of something else, that Kant points out this possibility; and the causality of
Freedom
25
such a first cause would, he says, be Freedom. If, therefore, it can be shewn that this proposition is true, not merely that it may be, then the existence of Freedom is proved. meaning and ambiguity of causality. ding an sich as free cause However, before proceeding to discuss Kant’s proof that there is free causality, it seems desirable to explain at greater length precisely what he means by this conception – why he calls it causality and why he calls it free. And, in the first place, it seems a pity that he should have called it causality at all; since there must evidently be a most important difference between the temporal relation of successive phenomena, usually denoted by the term, and the relation of real to phenomenal things which is here in question. It is true that Kant is not so defenceless as might appear against the charge of inconsistency sometimes brought against him, when in the Critique of Pure Reason he speaks of the Ding an Sich as cause of a phenomenon; since, according to him, causality is a pure category, and therefore, though only applicable by us when schematised through the form of Time, in which alone objects can appear to us, it may conceivably apply to other than temporal relations of other objects, to us unknown. He is therefore only exceeding the limits within which he has for the present restricted human knowledge, when he says that the Ding an Sich actually is a cause: he is not applying to the relation of Reality to Appearance a notion which he has affirmed to be from its very nature only applicable to the mutual relations of Appearances. But nevertheless causality is commonly understood of a temporal relation, and Kant’s employment of it also to denote the relation of objects under the unschematised category seems liable to cause grave misunderstandings, especially in the interpretation of his notion of Freedom. Under the ‘pure’ category he would seem to include every sort of necessary dependence, whether analytic or synthetic. Thus (to take his example of the Hypothetical Judgment, from which he derives the category of cause) ‘If perfect justice exists, the man who perseveres in wickedness is punished’ asserts a connection between the universals ‘perfect justice’ and ‘the punishment of the wicked’ of such a sort that from the existence of perfect justice you would be entitled to infer the punishment of the wicked. Now such a connection might, according to Kant, be of two sorts: Either the punishment of the wicked is included in the notion of perfect justice, in which case the connection is that of logical implication, i.e. analytic; or the
26
The 1897 dissertation
appearance of perfect justice (if it could appear) might be the invariable temporal antecedent of the punishment of the wicked, in which case the connection is causal in the ordinary sense (that of the schematised category), and synthetic. It was a relation which should combine the absolute certainty of the former with the synthetic nature of the latter which Kant regarded as the ideal of rational knowledge. Both, as they stand, are obviously imperfect: logical implication because it is purely formal and can never enlarge our knowledge by assuring us of the existence of any object; causality, because it deals only with appearances in the subjective form of time. The former gives us only the necessary connection of part of a notion with the whole, but does not in any way warrant the existence of an object corresponding to our notion; the latter only enables us to pass continuously from one object to another which has always in its turn the same contingency. What Kant wanted for complete knowledge was some object which could be regarded as ultimate ground of all others, in the same way in which a universal law is the ultimate ground of its exemplifications. Such an object, it has been generally assumed, was the Ding an Sich, of which Kant does assert the existence in the Critique of Pure Reason. It is neither cause in the sense in which a particular appearance is cause, nor reason in the sense in which a logical universal is reason. It certainly has Freedom, and therefore it has justly been taken as a starting point in discussions of this fundamental notion of Kant’s Ethics. And, since there is as yet no agreement as to its significance in this connection; since on one side Cohen asseverates fiercely that it is a Law; and on the other Kuno Fischer, following Schopenhauer, identifies it no less positively with Will, there would seem sufficient reason for attempting to sift the matter thoroughly here. It will, perhaps, be a convenient method of discussing this problem, if we first put briefly side by side the various doctrines of Kant which seem to have the most intimate bearing on his conception of the Ding an Sich, and then try to draw a consistent conclusion from them. Too many commentators seem to have failed in estimating the conception justly, from not taking a sufficiently wide view of its bearings. I shall therefore try to enumerate completely the main points, without, however, entering into any detail upon each separately – labour which would be irrelevant to the purpose of this essay. First, then, as has been already said, Kant does in the Critique of Pure Reason assert the reality of the Ding an Sich. We do know that it exists. We know, also, that it is the ground of Appearances; or, as Kant also puts it, causally related to them. But we do not know what it is. This defect in
Freedom
27
our knowledge is denoted by its name – the thing-in-itself. All our a priori or necessary knowledge, Kant maintains, is only of the relations of things; we can never get a knowledge of what they are in their inner nature.a This ignorance he maintained against the school of Leibniz, whose doctrine was that our spatiotemporal experience gave us as direct a notion of the real world, though this notion was more confused, as did the principle of contradiction and other laws of Formal Logic, to which everything real must conform. This objection to Leibniz’s metaphysics would seem to be connected with that objection to his epistemology, which led to the transcendental deduction of the categories in the Analytic, and which is also clearly put in the answer to Eberhard.24 Here the point turns on the possibility of proving a priori that every thing must have a ground or cause. Kant holds that we can only be certain of this in the region of Appearances, because we are there speaking of the way in which things affect us through the forms of space and time, which forms, as lying in our own reason, are within the reach of our investigation. To this view he was led by the study of Hume, with whom he agreed that the matter of knowledge, that about which we can make propositions, was given in sensation. It, therefore, is on this view essentially a posteriori;b it is what it is, and no more: no necessity can be shewn why it should be such as it is. On the other hand Reason requires for complete knowledge, that there should be necessity in it; that the matter of knowledge, no less than its form should be known a priori; that the nature of things should be deducible from the nature of Reason itself. It is this last point of view which is elaborated in the Dialectic. Our understanding is merely discursive; it only gives us universal forms for the necessary connection of a material always presented in endless series. What Reason demands is, as it were, to get behind these series. It is not content with what the Understanding can perform – the connection of each member with all the rest. It wants to bring in a third party – a complete reason of all these connections, some self-subsistent reality, which, by a b
24
R.V. p. 76 [B67, GW 189]. Kuno Fischer seems very perverse on this point, when he tries to draw a distinction between the statement that the matter is a posteriori and that it is given a posteriori (Critique of Kant p. 7). What his argument seems to bring out is that since the matter is a necessary presupposition of experience, the knowledge that there is a matter cannot be a posteriori; but he does not distinguish this from the knowledge of the matter. The materiality of knowledge is not given empirically, but the matter is. He does, indeed, refute himself later by quoting from Kant the very phrase against which he is here contending. (pp. 24–5; R.V. p. 56 [B34, GW 156]) Kant gave his ‘answer to Eberhard’ in On a Discovery (1790). This text is translated and discussed in H. E. Allison’s The Kant–Eberhard Controversy, Baltimore: Johns Hopkins University Press, 1973.
28
The 1897 dissertation
necessarily connecting itself with each of them, should shew why they are connected with one another. This our Reason cannot get, because all the matter presented to it necessarily appears in the Forms of Space and Time, which condition its receptivity. Hence when it seeks for the unconditioned, it only obtains it in the form of abstract laws – the Laws of Nature. It is these from which, in its formal use, it deduces the particular connections, whose reality is only vouchsafed by the action of the Understanding upon the material presented in the sense-forms. They have the highest form of unconditionality to which we can attain; but they have a twofold defect. They are not things, and therefore cannot be real, as they are; the cause to which the Understanding refers each particular event is superior to them in this respect. And secondly the material to which they refer is still merely this given Appearance – something a posteriori; this alone gives them actual significance. Lastly we must take account of the knowing faculty which Kant ascribes to God, and which he says differs absolutely in kind from ours. This he calls an ‘intellectual intuition’ (Intellectuelle Anschauung) or ‘creative understanding’ (Intellectus originalich). That is to say, the matter of knowledge, presented in intuition, would to such a faculty have the a priori character which now belongs only to our forms of intuition and to our understanding. A posteriori knowledge would be entirely excluded; but at the same time that which we can only conceive as known a posterioria,25 ‘things, as they are in themselves’ would still be an object of knowledge. This is only possible on the assumption that reason should create the matter of its knowledge as well as give it form, unless we have recourse to pre-established harmony, which is a mere hypothesis, quite without justification. The intellect would not be limited, as it is with us, to merely arranging given data, but would rise to the full dignity of a synthetic reason, since it would itself create an object of such a nature as to be capable of that absolute totality or unconditionality which reason demands. This conception combines the different conceptions of truth in rationalism and empiricism. The necessity expressed in the universal judgment it takes from the first; and the assertion of existence expressed in the categorical judgment from the second. Were we possessed of such a knowing faculty philosophy would not a 25
Prol. §9. Moore provides this reference; but it is puzzling that the position he attributes to Kant appears to be the opposite of that which Kant here affirms, which is that for an intellectual intuition of things as they are in themselves ‘absolutely no intuition a priori would take place, but it would always be empirical’.
Freedom
29
oscillate between the two extremes of saying that our sensations are all that we can know, while so-called laws by which we attempt to group them are mere abstractions and represent nothing in the world itself; and of asserting that nothing is truly known except universal connections or predicates. With our present knowing faculty we can only attain a union of unity with plurality, or identity with difference, in two ways, each imperfect. Either the unity or identity takes the form of a mere abstract notion or logical universal, which exists only in each given particular; or it is the given particular which persists through change and unites in itself several universals. Kant imagines that for God these two forms of unity would be the same. Each object would be identical with every other in the sense in which the universal is identical with itself in each particular; but at the same time each would be substantial and so unite in itself all the others in the same way as now the particular unites universals. How this should be, he rightly says, is to us incomprehensible. It is so because we cannot reduce the ‘individuation’ given by the forms of space and time to the individuation given to concepts by thought, or vice versˆa. The conception resembles that of Leibniz in that each object is complete ground of all the others; but differs in that Kant, by definitely rejecting the forms of space and time as mere forms of Appearance, does not need after all to represent the objects as having grades of perfection, which is contradictory to their unity from the analytic point of view. But for the same reason, since we cannot escape our limitation to space and time, Kant thinks we cannot vouch for the reality of this conception. Now Kant himself never clearly connects these various doctrines. They follow singly from the application of his new epistemology to the various questions which had been raised in the previous progress of modern philosophy. He was naturally chiefly concerned to put his doctrine in relation to the past; but in doing so he failed to give to itself such a systematic unity, as he had achieved between previous Rationalism and Empiricism. He thus leaves us with several ultimate doctrines which he nowhere combines with one another in such a way as to shew that Reason is really one in all its functions, and how the various truths at which it arrives are related to the one ultimate truth which the unity of knowledge presupposes. ding an sich implies positive knowledge of real free cause, which is no cause The crucial question, which comes to light in a review of Kant’s system, appears to be this. Why is our knowledge to be condemned as merely
30
The 1897 dissertation
knowledge of Appearance? Why are we to be merely Transcendental Idealists and merely Empirical Realists? Kant’s doctrine of Appearance was designed to recognise the truth, on the one hand, of the English empirical position that we can only know what is given us; and, on the other hand, of the German Rationalists that we have a priori knowledge – that some propositions are true of everything that can be given. These two points of view, he thought, could only be united by examining the nature of our knowing faculty, because, if definite characteristics could be found in it, these must necessarily qualify in a definite way everything which it could know, and thus give a priori propositions; whereas the ‘manifold of sense’ – that which seemed to be left, after these universal predicates imposed by mind had been abstracted, was obviously, as Berkeley had pointed out, subjective. But by this too psychological statement of the nature of knowledge, Kant did in reality lay himself open to the charge of Berkleian Idealism, which he indignantly repudiates by asserting the existence of the Ding an Sich. For he has no answer to the question: How do we know that these conditions imposed by our knowing faculty are universal? To this question he would have been bound to find an answer if he had ever tried definitely to determine whether the defect in our knowledge lay in the fact that the sense-manifold was only given, or in the nature of the a priori forms, space and time, in which it is given. The first view seems to be implied by his conception of the perfect understanding as necessarily creative. But, then, he cannot shew that our reason is not creative in this sense: since the sense-manifold is purely subjective, we cannot know that its nature is not wholly determined by the subject: the assumption that it is the resultant of subject and object, ego and Ding an Sich, is just one of those dogmatic assumptions, which, because we cannot know it to be true should not be made in an epistemology which aims at being ‘scientific’. The other view, that space and time are defects, seems to be implied in the solution of the antinomies; since it is only the endless relativity of space and time which is there assigned as the obstacle to our knowing the world as a whole. His argument here is that we are involved in Antinomies only because we try to apply to the world of Appearance a notion which can only hold of Things-in-themselves – the notion of totality, and unconditionality. But this implies a definite notion of the Ding an Sich – that it is unconditioned, the inapplicability of which to things in space and time would be sufficient to condemn them as mere appearance. Kant, however, does not see that this is the real consequence of his argument because he has started with the assumption that our world is a world of Appearance, merely because it appears to us; and because he identifies the fact that it appears
Freedom
31
to us, with the fact that its matter is given. But the mere fact of appearance to an intelligence could not be an argument against the ultimate reality of what appears, since the real itself must appear in some form to God’s ideal intelligence. The argument is only valid on one of two conditions: (1) either it must be assumed that the appearance is given, that is to say, that there is some utterly external object which gives it to us and so far makes our knowledge merely passive or receptive. But this assumption, as was pointed out above, cannot be justified epistemologically. Kant seems only to have been led to use it by accepting the English description of sensation as passive, coupled with the further doctrine, which really contradicts the point, that all that we can know must be in a sense mental phenomena or, in Locke’s phrase, ‘ideas’ (Vorstellungen). Berkeley saw this contradiction and accepted the latter alternative. But Kant’s common sense cannot brook the reduction of things to mere illusion (Schein), and he therefore takes refuge in the Ding an Sich; offering, however, little justification for so doing, except that ‘Appearance’ must be appearance of something. Or (2) the contrast of Appearance with Reality, may be defended on the ground that the manner in which things must appear to us is inconsistent with their real Nature – ‘in themselves’; and this is the argument implied in the Antinomies. This latter argument, however, is inconsistent with the notion of a Ding an Sich, in the form in which that notion first appears, as a necessary corollary of the passive theory of sensation. We are now, then, in a position to sum up with regard to the Ding an Sich. The Ding an Sich, as such, was a notion connected primarily with Kant’s theory of the sense-manifold as something given. On this theory there must be something to give that which we know as given; and this, which gives what we know, not being itself given, cannot be known. This is the Ding an Sich. Or else the notion may be said to have arisen in this way. We can only know that which we have received. But what we receive is necessarily subjective, since we have received it; and so is merely Appearance. Appearance, however, must be Appearance of something; and this something, not itself known, because it is not Appearance, is the Ding an Sich. Now both these arguments Kant would, I suppose, have regarded as merely analytic. It follows from the mere notion of something given, or of something appearing, that there must also be something to give and something, as it is in itself. But Kant does not seem to have recognised that his statement that what we know is merely given or appearance, was itself synthetic and therefore required justification. His argument seems no better than the one which he, after Hume, derides, that every effect must
32
The 1897 dissertation
have a cause, because the notion of cause is implied in that of an effect. This, he rightly urges, is a petitio principi. The significant proposition would be that every event must have a cause; and in this form the plausibility of the argument disappears. So in his own case, if he had substituted for ‘what appears to us’ quite simply ‘what we know’, he would have seen that the inference to a Ding an Sich required elaborate justification. Moreover, as was pointed out above, the unjustifiable nature of his fundamental assumption is shewn by the contradictory forms in which he expresses it. It cannot be maintained at the same time that the defect in our knowledge is (1) that we only know ‘what is given’, and (2) that we only know ‘what we receive’. For the first expression implies that what we know is not sufficiently dependent upon our reason, not deducible therefrom; whereas the complaint of the second is that the object is too dependent on our reason, so dependent that its own nature, apart from us, cannot be known. Both these defects may be traced from the fundamental form in which Kant puts the problem of the Critique. He does not sufficiently distinguish it from Locke’s psychological problem. When he sets out to examine the limits of our Reason, he presupposes that the distinction between subject and object is fundamental for epistemology, that we have knowing faculties (Verm¨ogen) which we can examine by themselves, and that on the other hand there is also a world, which is what it is, whether we can know it or not. Within the limits which he had marked out for possible knowledge, the region of appearance, he does indeed give to a priori a meaning that is not psychological. He shews that what he calls the a priori Forms of Intuition as well as the Principles of the Understanding are implied in any particular object which we can experience; they are necessary to the possibility of our experience; we could not have any of those perceptions which the English school supposed to be ultimate and unanalysable, unless these universal relations were involved in them. In this method of determining the a priori, as what is implied in all our particular judgments lay the great advance which Kant made upon previous philosophy. He thus cut away the whole ground of dispute between the supporters of innate ideas and their opponents, by shewing that neither hypothesis would account for the knowledge we actually possess and by supplying a third which would. But, when he limited the validity of the Principles of Understanding to what merely appears to be the world, he was again introducing on a higher level the old method. He thought it necessary to justify his claim for the possibility of a priori knowledge, which he had already really justified by shewing its implication in experience, by explaining that it depended on
Freedom
33
the nature of our knowing faculties and thus re-establishing a different sort of innate idea. The investigation of knowledge, upon such a presupposition of its opposition to reality, can obviously lead to nothing but to its confinement to Appearance. When our knowledge is from the beginning investigated as belonging to us as opposed to the world, it can never be brought into relation with the world, except by some such doctrine as Preestablished Harmony. The world is from the beginning for Kant a Ding an Sich; and, in this sense, which is its primary explicit one, the conception is totally invalid. At the same time, in this form, it allows neither of those descriptions upon which Cohen and Kuno Fischer (after Schopenhauer) appear to base their theories of Kant’s Ethics. It is not a ‘law’ because it is a ‘thing’; and nobody is more careful than Kant in distinguishing the ‘thing’, as a substantial existence, a centre of relations, a possible object of intuition (Gegenstand), from the merely relational forms of reason, to which all ‘laws’ belong. On the other hand it is not ‘will’ (in the ordinary sense), because it is totally indeterminate – the mere logical ‘reason’ of appearances; which is the very aspect of it which made it seem like a ‘law’. But when we reach the ‘Dialectic’, we get a conception, which takes the place of the Ding an Sich as opposed to phenomena, and which is justified and has important bearings. We here seem to have it laid down that the true object must be capable of being conceived as a whole and unconditioned. This gives us the basis of the Hegelian doctrine that the rational and the real are one. ‘Thought’ or ‘Reason’ here presumes to say something about what the world must be. Kant’s solution of the Antinomies is as follows. There could be no solution, if the objects which we know in space and time were the world as it really is. In the world, as it really is, the sum of conditions must be given along with the conditioned; otherwise it would be in contradiction with the formal laws of thought. The Antinomies result from the application of this character of reality to things which are wholly dependent on our sensibility; and, Kant maintains, there is no justification for such an application. Now it would seem to be a valid inference from this argument that the reason why we know that objects in space and time are not the world as it really is, is that to treat them as such would lead to irreconcilable contradictions. But this, it must be confessed, is not the way Kant puts it. He still retains the assumption, on the totally invalid grounds discussed above, that things in space and time are mere phenomena and that therefore there are also Dinge an Sich; and his conclusion is that we can never prove by theoretical investigation, that the notion of unconditionality applies to any existent object; not even to the Dinge an Sich, because ex hypothesi we cannot know
34
The 1897 dissertation
them. What he does not see is that he has already assumed the objective validity of the notion of unconditionality in asserting the existence of the Ding an Sich; and that it is only this assumption which enables him to deny that we know the unconditioned. For he says that, if things in space and time were really things as they really are, we should be bound to admit that the sum of conditions necessary to each was given along with each. It is only because they are not real things that we need not be disconcerted by the Antinomies. It is, then, only because Kant has already assumed the world as it really is is to be ‘free’ in his sense, that he is able to deny at the end of the Critique the possibility of knowing that such freedom actually exists.26 So that we are justified in saying, on the results of the Pure Reason alone, that the Ding an Sich is free, but not in so far as it is a mere Ding an Sich. This result, and with it the meaning of freedom, is obtained by a reconsideration of Kant’s premises. Kant has in fact no title to talk of the Ding an Sich, or of experience as phenomenal in contrast thereto, if the Ding an Sich is really inaccessible to knowledge. What he has really done, in making this presupposition, is to apply objectively the very Ideas of Reason, of which he denies the objective or constitutive applicability. He has really, as Professor Caird puts it,27 assumed that the pure thought is never purely analytic but also synthetic, since in some way it must always combine phenomena with the real world. He has himself made use of the principle of sufficient ground unschematised, while he denied the validity of such uses because as a matter of fact it had been misused as if it applied to things in space and time. One object stands in the relation of free cause to another, when from the existence of that other you can necessarily infer its existence, whereas it does not itself need to be referred to any other object as its ground. Now starting from any object in space and time you can infer the existence of such an unconditioned ground for it, but that ground can never appear as an object (Gegenstand) in contrast to the first, because it would then at once be limited, by exclusion from the first, and something further would be needed to hold the two together. It is therefore true that we can never know Reality, in the way in which we know Appearances, as separate existences merely related to one another in a necessary manner. We cannot directly experience it as Reality, which would alone be complete knowledge. But Kant is wrong in supposing that the direct experience which we have 26 27
Russell: ‘This seems to me not quite clear, and susceptible of explanation’. Moore provides no reference, but Caird makes a claim of this kind on p. 220 of The Critical Philosophy of Immanuel Kant, 2 vols., Glasgow: James Maclehose and Sons, vol. i.
Freedom
35
of phenomena is any more worthy to be called knowledge than the merely formal certainty that their ground exists. The one is as deficient in its way as the other. Indeed, it seems absurd, on reflection, to call that knowledge which is merely of phenomena, since knowledge is not knowledge unless it is true, and so in some way must be information concerning things as they really are, i.e. Dinge an Sich. On this view, therefore, it is unnecessary to deny that the Real World appears to Intuition – our own experience. We can only deny that it appears as it is. The character of objectivity, that which makes a Gegenstand a Gegenstand, can be separated from particular objects which possess it, and can be directly predicated of the real, no less than can such abstract characters as ‘being’, and ‘unconditionality’, which do not involve space and time. There is no longer any need for conceiving Reality as external to all particular Appearances in the same way in which one Appearance is external to another – a false conception, which seems to have led Kant to call the Reality a cause. It is, indeed, their ‘ground’; but that relation is to be conceived not merely like that of formal logic nor like that of cause and effect, but as something between the two. The Reality is not an Individual separated from particulars as they are from one another, not yet a mere universal from which they might be deduced; it is an Individual both implied and existing in them. It is ‘transcendental’, in the sense in which the categories are said to be so, but no longer ‘transcendent’ since nothing can be so. reality thus has transcendental freedom. but kant applies it to ‘will’ The answer, then, to the question what Kant means by Transcendental Freedom is this. Transcendental Freedom is the relation in which the world as it really is stands to events as we know them. It is the relation of Reality to Appearance. This relation necessarily appears to us as the logical relation of reason to consequent. The reason is free cause of its consequence. But though the relation is of this kind, and Transcendental Freedom is, by this aspect of its nature, absolutely distinguished from empirical causality and from human volition, as it appears in psychology, which is merely one form of such causality; nevertheless it differs from the logical relation of reason and consequent, in that neither reason nor consequent is here an abstract notion, but must be considered as having existence. A mere logical reason can never, as such, be considered as actual. If we seek for an actual existence as the ground for another, we get a mere cause.
36
The 1897 dissertation
But that which has Transcendental Freedom is not a mere cause because it is no part of the temporal series of events; and it is not a mere logical reason because it has all the self-subsistence which appears to belong to the given temporal series. Our result is that the Intelligible Character, which is free, is (1) not the Ding an Sich; because the Ding an Sich, as such, is a purely negative conception, which must be superseded, since our positive experience, in opposition to which it is defined, could not be known to be in opposition to it, unless it were itself known to be possessed of a positive character. (2) It is not a ‘law’, though its relation to events, is that of a law to its instances; because a law, as such is abstract, not self-subsistent: and a merely hypostatised law, like the Platonic Ideas, is a mere union of two characters felt to be equally necessary, but of which a direct union is strictly inconceivable. (3) It is not ‘Will’, because will, as we know it, is merely a cause in the order of nature, itself an effect of other preceding causes. Such a description either gives us no information, if ‘intelligible’ will be meant; or it is false, if it be meant to illustrate an absolute reality by the nature of an empirical process. But on this last head, it may be maintained that Schopenhauer did mean ‘intelligible’ will;a and the proof that will, if it becomes ‘intelligible’, ceases to be will, can only follow from a much wider discussion. We shall reach it, by answering our second question. Why Kant should have called his free causality ‘free’? We have already explained that this ‘free causality’ is not causality in the ordinary sense; and there may well seem a good case for the contention that it is not free either, on the ground that freedom has an essential reference to human volition. [We have already tried to shew that the objective validity of the Idea of ‘Freedom’ in Kant’s sense can be found by theoretical arguments alone, without reference to the validity of the Categorical Imperative, which Kant regards as the necessary datum for his proof.]b,28 Kant’s conclusion at the end of the Critique of Pure Reason should have been that Transcendental Freedom was not merely possible but actual. But this independence of the proof of ‘Freedom’ from the Categorical Imperative, would seem to justify a suspicion that this ‘Freedom’ is not freedom, since its connection with human action is by that independence certainly lessened. And, indeed, it must be admitted that there is no longer a b 28
Kant also uses this conception; he says that the object of Metaphysics of Morals is ‘a possible pure Will.’ G. p. 238 [AK 4: 390, MG 46]; M. p. 18 [AK 6: 221, MG 376] passim. M. pp. 18 [AK 6: 221, MG 376], 22 [AK 6: 225, MG 379]. This passage has been marked for omission, but only very lightly, in pencil.
Freedom
37
any reason for connecting the ‘Intelligible Character’ with the psychological character, which distinguishes one individual from another. The ‘Intelligible Character’ is the one sufficient reason of all phenomena, whether processes of inanimate nature, or human actions. It is not proved that it is individualised in a multiplicity of souls; and it is certain that in any case it is the same in each. Our conclusions will not enable us to decide between a Monadism and a Monism; but they shew, that, if there be Monads, they will be identical in so far as each exemplifies the Intelligible Character. The ‘Intelligible Character’ cannot be used to explain why one man is different from another, so that you could say A is soand-so because his ‘Intelligible Character’ is of this sort; and B is different because his ‘Intelligible Character’ is of that sort. All differences can only be explained by referring them to different causes. But the Intelligible Character is the one reason of the whole world with all its differences, and so not more the reason of one part than of another. This is another of the differences between the Intelligible Character and the Platonic Ideas. Kant however does not admit that everything may be regarded as a result of intelligible or free causality, just as everything is a result of natural causality. All he claims in his discussion of Transcendental Freedom at the end of the Critique of Pure Reason, is that ‘among natural causes there may also be some, which have a faculty that is only intelligible’.a And he goes on to explain that by these he means only mankind. ‘In inanimate or merely animal nature we find no reason to suppose any faculty conditioned otherwise than through sense. But man, who knows the rest of nature only through senses, gets knowledge of himself also through mere apperception . . . and is to himself partly, we must admit, a phenomenon, but partly also, namely in view of certain faculties, a merely intelligible object’.b Now setting aside the statement that man knows himself through mere apperception, – a kind of knowledge of which Kant has not elsewhere explained the possibility, and which seems here temporarily to take the place of the Categorical Imperative as affording a ratio cognoscendi for the applicability of freedom to him – it is plain that he here regards man as on an absolutely different level from other things, in respect of freedom. Man has freedom, and nothing else shares it in any degree. And throughout his ethical works this attitude is maintained. Free causality is attributed to man alone, among the objects of experience.c So that, whereas natural causality a c
b R.V. p. 379 [B574, GW 540]. R.V. p. 378 [B573, GW 539]. Though also it would belong to any other ‘reasonable beings’, if such there be. G. p. 237 [AK 4: 389, MG 45].
38
The 1897 dissertation
applies with absolute universality – to him as well as to all other objects, freedom appears as a sort of miraculous power, whose influence may be traced in some events, but not in others. In the Critique of Judgment he is led partly to correct this view and to see that if Freedom is to be brought in to explain anything at all, it must be brought in to explain everything. But, meanwhile, his restricted view of freedom makes it easier for him to establish a connection with the vulgar notion. In the vulgar notion, too, some actions are free and some are not; and though it would not be admitted, as for Kant is necessary, that those which were free, might also, in another aspect, be seen to be completely determined by natural causality, yet the mere fact that the application of freedom is so partial, and also its especial connection with man, assimilate the view more to that of Kant, than is possible with that here advocated, according to which freedom is universal. In the vulgar notion of freedom the most universal characteristic seems to be the absence of external constraint, whether asserted to impel or to prevent. Where the immediate cause of a motion or change seems to lie in the thing which moves or changes and not in anything outside it, there, in a sense at all events, freedom is predicable. But this is a notion which is obviously not limited to human actions. Many of the movements and changes of animals and plants have their proximate causes in the things themselves; and the same might probably be said of any body in so far as it moved in accordance with Newton’s second Law of Motion. It is thus we seem to talk of ‘free as air’, or of the wheels of a watch moving ‘freely’. But there is an obvious defect in this wide notion, in that the limits, whether spatial or temporal, of any group we may take for our unit or thing, are always more or less arbitrary. A watch may be moving freely when its spring is driving it; but the movement of any one of its wheels is not free, because the wheel is driven by the spring or by another wheel. And, again, there seems no reason why we should single out the proximate or immediate cause for such pre-eminence, nor anything to determine how far back in the past a cause ceases to be proximate. It is difficulties of this sort which seem to have gradually tended to restrict the notion of freedom to men; because in man the notion of self is far more striking than elsewhere and the distinction between the internally and externally caused, therefore, primˆa facie more satisfactory. The difference between himself and anything else whatever is more constantly forced on a man’s notice and more practically important to him than any other difference; and it is therefore not unnatural that the notion of free, in the sense of
Freedom
39
self-caused, action, whether or not it is originally derived from his experience and transferred anthropomorphically to other things, should at all events be more widely applied and less easy to dispense with in his own case and that of other beings like him, than elsewhere. ‘liberty of indifference’ an attempt to find in the world of experience an absolutely self-caused cause. impossibility of completely predicting human action Now the vulgar doctrine of Free Will, as ‘Liberty of Indifference’, seems to be in the main an attempt to raise this distinction between self and the world entirely above the level of an arbitrary distinction. It was seen that this could not be done, if the self were regarded as a part in the causal chain of events, since it must then be subject to the infinite divisibility inherent in time, and the ultimate causal unit remains as arbitrary as any unity of time. It was therefore maintained that man’s soul was an agent undetermined by previous events in time; it was the absolutely simple unity of Rational Psychology, and as such distinguished from all natural objects, which were always both themselves divisible into parts and also incapable of certain discrimination from an ever wider whole. Such a notion of a finite uncaused cause inevitably follows from the attempt to distinguish within the world of experience cases of purely internal and purely external, of immanent and transeunt, causation. And there are good reasons why the human will should have been taken as the final instance of a cause which is not also an effect. The progress made in the analysis of mental processes has been very slight in comparison with that made in physical science (1) because of their greater complexity (2) because experiment in psychology must be either indirect or encumbered by the fact that the observed is also the observer and (3) because subconsciousness must be taken into account. And the region of the incompletely known is the favourite abode of a metaphysical monstrosity. In plain language, where facts are not completely understood, some shortsighted metaphysical theory is generally introduced as affording an easy road past the difficulties which stand in the way of thorough investigation. And, secondly, apart from the general difficulty of establishing exhaustive causal laws, which applies in a less degree to physical science also and prevents certain prediction even there, there seems to be a real reason, which from the nature of the case can never disappear, why human volition should produce the illusion of so-called Freedom. It is this, that, in virtue of the deterministic hypothesis itself, the knowledge that a certain course
40
The 1897 dissertation
of action was about to be pursueda must always exert some influence upon the course actually pursued, and so make the result different from what was foreseen after a consideration of all the other elements that would contribute to it. And even if the fact of this knowledge were taken to account in the calculation, and the prediction modified accordingly, the knowledge of this modification would again introduce a new element, which would require a fresh calculation, and so on ad infinitum. This seems to be a difficulty inherent in the double nature of the mind as subject and object – a difficulty which makes it possible to pronounce a priori that complete prediction of the results of mental process must always be impossible. It is a difficulty which does not apply to prediction in the physical world of space, considered, as seems necessary at present, in abstraction from the world of mind. It could only modify our view of that, if the real connection of body and mind were fully discovered. As it is, mental processes, though obviously corresponding to physiological [ones], and useful for their investigation, have only too much the appearance of a totally independent world from the point of view of causality and reciprocity. So that the distinction is justifiable, when we say that the results of human volition, alone among causes, must of necessity remain incapable of prediction. And this fact, along with the greater empirical difficulties of prediction in the case of mind, seems sufficient to account for the illusive belief that the will, at any rate, is free, though it be admitted that nothing else is. The failure to discover a cause in any particular instance of itself encourages a belief in the uncaused; and when to mere failure is added an absolute impossibility of discovery, the case is naturally strengthened. That the belief in uncaused volition is illusory, the progress in scientific method, with the resultant growth of empirical psychology, has rendered it more and more difficult to doubt. Nor is it only, as Professor Sidgwick says, that ‘as29 the Determinist side there is a cumulative argument of great force’.b For, in however many instances causation were proved, though that would, on Hume’s principle, be a cause of our expecting it in others, yet it would not be, by itself, any reason for that expectation. An inductive argument always needs, as empiricists put it, to be supplemented by the assumption of the uniformity of nature. And that this assumption is not a
b 29
Unless, indeed, we are to carry out logically Professor Huxley’s doctrine (Hume p. 86) that ‘there is only a verbal difference between having a sensation and knowing one has it.’ [The passage Moore refers to occurs on p. 73 in T. Huxley, Hume, London: Macmillan, 1879.] ME p. 62. Sic – but it should be ‘On’.
Freedom
41
in this case an assumption but an a priori necessity, may, I think, be considered to have been sufficiently proved by Kant’s argument in the Analytic. He there shews that every event must have a cause if there is to be an objective succession in time; and such an objective succession is certainly presupposed by all actual experience. Accordingly Kant himself fully recognises the a priori certainty of the Deterministic view, as was shewn at the beginning of this chapter;a and it seems inevitable to agree with him. As for ‘the immediate affirmation of consciousness in the moment of deliberate action’b which is asserted to stand against Determinism, great care is needed in deciding what it is that consciousness then affirms. Professor Sidgwick expresses it thus. ‘Certainly when I have a distinct consciousness of choosing between alternatives of conduct, one of which I conceive as right or reasonable, I find it impossible not to think that I can now choose to do what I so conceive – supposing that there is no obstacle to my doing it except absence of adequate motive – however strong may be my inclination to act unreasonably, and however uniformly I may have yielded to such inclinations in the past’.c,30 libertarianism means uncaused choosing. no defence of it. Now with regard to this statement, it is to be noted first that what consciousness is said to affirm as on the Libertarian side in the Free Will controversy is only the possibility of choosing the reasonable course; the note appended seems to be intended to defend this limitation.31 This note runs: ‘It is not the possibility of merely indeterminate choice, of an “arbitrary freak of unmotived willing”, with which we are concerned from an ethical point of view; but the possibility of action in conformity with practical reason.’d The same limitation is also introduced in the passage quoted a b d
30
31
Cf., on this subject, P.V. Pref. p. 12-end [AK 5: 12–13, MG 146–7]. c ME p. 65. ME p. 65. Note that ‘action in conformity with practical reason’ seems here to be identical with the ‘doing what I conceive to be reasonable’ of the text. [This note is altered in the seventh edition: instead of the phrase ‘the possibility of action in conformity with practical reason’, Sidgwick now writes ‘the The phrase ‘supposing that there is no obstacle to my doing it except absence of adequate motive’ which Moore cites does not occur in the seventh edition of ME; Sidgwick here writes instead ‘supposing that there is no obstacle to my doing it other than the condition of my desires and voluntary habits’, presumably in order to emphasise the presence of other motives. It is conceivable that Sidgwick was motivated to make this change by Moore’s discussion here of the earlier version which he will have read in his capacity as one of Moore’s examiners. This is note 1 in ME p. 65.
42
The 1897 dissertation
above: ‘Is there always a possibility of my choosing to act in the manner that I now judge to be reasonable and right?’a Now, what precisely is meant by ‘the possibility of merely indeterminate choice’ mentioned in the note, I am quite sure that I do not understand; especially as, in the text, the choice, of which the possibility is under consideration, is pointedly distinguished as one to which the presence of adequate motive cannot be assured so that the question presumably is of the possibility of unmotived choice, in so far as there ceases to be a question, if the motive is alone sufficient to produce the choice. We are considering the possibility of a choice in conformity ‘with practical reason’, where something other than the motive supplied by that conformity is necessary to produce the choice; and where, therefore, the choice if possible must be assumed to be to some extent unmotived. But, however this may be, I do not see how it can be denied that ‘we are concerned from an ethical point of view’ with the possibility of a wrong or unreasonable choice – a choice not ‘in conformity with’, but contrary to ‘practical reason’; and, whether such a choice be an ethical concern or not, Professor Sidgwick has certainly elsewhere stated that the question of its possibility is essential to the Free Will controversy. For instance, in the passage on p. 6132 just quoted, he says the question is whether ‘my voluntary action, for good or for evil, is at any moment completely caused’; though, as we have seen, in the following sentence, he drops the ‘for evil’, and inquires only into the possibility of my choosing to act rightly. So too, on p. 58, he says ‘It is obvious that the Freedom thus connected with Responsibility is not the Freedom that is only manifested in rational action, but the Freedom to choose between right and wrong which is manifested equally in either choice’. In fact, when on p. 65 Professor Sidgwick limits the possibility of Freedom to the possibility of choosing rational actions, he seems to have fallen partially into the error of which he accuses Kant and his followers – the error of ‘identifying Rational and Free action’ (p. 57). For he seems to suppose that only rational action can be free, though he does not complete the identification by also implying that only free action can be rational. It is noteworthy, however, that it is only against the first of these two statements, the one which he himself subsequently seems to adopt on the immediate affirmation of consciousness, that he directs his
a 32
possibility of choosing between rational and irrational motives’. Again (cf. note 30) it is possible that this change was motivated by Moore’s critical discussion here of the earlier phrase.] ME p. 62. Moore is of course referring to the fifth edition of ME, both here and in the subsequent page references; in the seventh edition this passage occurs on p. 62.
Freedom
43
attack, as the supposed view of Kant; whereas it is only the latter, which, as will be shewn, Kant is really concerned to maintain. In fact even the statement which Professor Sidgwick quotes, as of Kant’s disciple, that a man ‘is a free agent in so far as he acts under the guidance of reason’,33 does not imply, as Professor Sidgwick thinks it does, that irrational action may not be also free – but only that all rational action is free, which is equivalent to ‘only free action is rational’, but not to ‘only rational action is free’. We may therefore use the weapons supplied by Professor Sidgwick himself on pp. 57–9, but there misdirected against a Kantian view, to defend ourselves in considering the possibility not only of our choosing the right course but also of our choosing the wrong one, whatever our previous character may have been. But it seems worth while, in passing, to dwell a little longer on the apparent inconsistency with which after maintaining that wrong must be just as free as right actions he himself seems immediately to adopt the one-sided view which he has been attacking. Thus he concludes the anti-Kantian paragraph on p. 59 with the words ‘Can we say, then, of the wilful wrongdoer, that his wrong choice was ‘free’, in the sense that he might have chosen rightly, not merely if the antecedents of his volition, external and internal, had been different, but supposing these antecedents unchanged? This, I conceive, is the substantial issue raised in the Free Will controversy’ – and does not seem to see that the whole tendency of his foregoing argument is to shew that it would be just as pertinent to ask: ‘Can we say of the wilful rightdoer, that his choice was free, in the sense that he might have chosen wrongly etc . . . ’? Now it is curious that in Aristotle’s discussion of the voluntary, wherever he touches the point, he always emphasizes the remark that vice must be considered to be as voluntary as virtue, not vice versˆa; as if it were a wide-spread notion in his day that virtue was voluntary, but vice necessitated. This, indeed, was probably due to the Socratic doctrine, that all vicious choice is merely due to our ignorance, and since to choose ignorance would itself be a vicious choice, to something beyond a man’s control. In Christianity, on the other hand, just the opposite doctrine prevails, that sin is due to our free will, and virtue solely to the merit of Christ; though, there, indeed, the other one-sidedness also appears, hopelessly entangled with it. But what is to be noted is the general tendency not to accept both doctrines at once, but each separately. For this, I think, whichever be the one taken, is evidence of an almost inevitable feeling that good and evil are somehow not on a par 33
ME p. 58.
44
The 1897 dissertation
with respect to reality; that good is somehow more natural and therefore not in so much need of notice or explanation as evil. It is the Origin of Evil which has always appeared as a problem; as if the Origin of Good were not primˆa facie equally inexplicable. With the Greek, who has no personal God to start with, as the author of nature, the explanation is found in blind natural necessity: man being the apparent best, it seems less obvious to suppose him cause of the worst. Whereas the Christian’s notion of God forces him to find an explanation of what is contrary to God, and therefore cannot be regarded as directly created by him, in a power approaching nearer to God’s than that of blind nature, and so in free Will. The onesidedness of our judgments is also illustrated by the far greater provision made in almost every state of society for punishment than for reward, and the general feeling that we need be much more careful when we single out for blame than when we praise; and here again the one-sidedness may be connected with either of the two opposite theories, according as punishment is regarded as corrective and deterrent, which harmonises with the Greek view, or as retributive, as suits the Christian. Nevertheless, it is plain, that this one-sidedness cannot be defended on such ethical principles as Professor Sidgwick’s, any more than it could on Aristotle’s. For those who take the world of experience to be real, there can (as Kant saysa ) be no freedom; and similarly, if freedom be nevertheless maintained there can be no grounds for asserting that good is more real than evil, even if there be a greater quantity of things which are good than of things which are evil. We shall hereafter34 try to explain how, upon Kant’s principles, this one-sidedness points to a truth: not, indeed, in the vulgar form of either ‘only rational action is free’ or ‘only irrational action is free’; but in what we said above to be the true Kantian formula ‘only free action is rational’, where however neither free nor rational can be understood in a popular sense. But, let us now consider, how the admission that wrong choice must be considered by the ordinary Libertarian to be as much the work of Free Will as right choice – an admission which, as we have seen, Professor Sidgwick is urgent to justify against Kant – affects Professor Sidgwick’s statement of what he holds to be ‘immediately affirmed by consciousness in the moment of choice’ on the Libertarian side. This statement, as we have seen, runs ‘that I can now choose to do what I so conceive’ (i.e. conceive a 34
R.V. p. 373 [B564, GW 535]. Moore returns to this issue on pp. 78ff.
Freedom
45
as right and reasonable) – ‘supposing that there is no obstacle to my doing it except absence of adequate motive.’35 Now the condition here asserted in parenthesis implies that if the doing of a thing is impossible for other reasons than the absence of motive, there can be no question of choosing to do it. In other words the only Free Will, which is at all affirmed by consciousness, is only free to choose actions, to the performance of which there is no obstacle but absence of motive. And this restriction to what we may call physically possible actions, seems plausible when the question is also restricted to reasonable actions; for to the ordinary reader it would seem that an action to which there were insuperable physical obstacles could not be called a reasonable one to choose. But we have seen that the restriction of the question to reasonable actions cannot be defended; and, if we recognise this and accordingly put the question as to the possibility of choosing wrong and unreasonable actions, it is no longer so plain that an action, which is physically impossible, is not an unreasonable one to choose. Part therefore of the plausibility of Professor Sidgwick’s statement as to the affirmation of consciousness, seems to be due to his erroneous limitation of the Free Will question to reasonable actions. But this objection is not fundamental. It only serves to lead up to another, which I believe to be of great importance, since it is based on a distinction without recognising which it seems impossible to arrive at a clear decision in the Free Will controversy. It is obvious that my last argument depended largely on the meaning to be given to ‘reasonable’ and ‘unreasonable’. Now Professor Sidgwick has discussed at length in chapter iii the meaning to be given to ‘reasonable’. He distinguishes two senses of ‘ought’, an ‘ethical’ and a wider sense, and notes (ME p. 3636 n. 4) that it is in the former sense that he will henceforward use the word. This ‘ethical’ sense, when it is said that an action ‘ought to be’ done, implies not only that it is reasonable in the sense in which Reason may apply the unanalysable notion of ‘ought’ to ‘a pattern to which no more than an approximation is practically possible’ (ME p. 3537 n.), but also that the action ‘is thought capable of being brought about by the volition of any individual, in the circumstances to which the judgment applies’. (p. 35) Now it may, I think, be assumed that when an action is termed reasonable, Professor Sidgwick generally means by it that it ‘ought to be’ done in this narrower ethical sense; since his definition of reasonable conduct seems to be that which ‘ought to be done’. There is therefore included in the 35 37
36 P. 34 in the seventh edition. ME p. 65; but see note 30 above. P. 33 in the seventh edition, as are all Moore’s subsequent references to p. 35 of the fifth edition.
46
The 1897 dissertation
notion ‘reasonable’ as applied to conduct both that it partakes of the ‘ideal’ and also that it is ‘practically possible’. But unfortunately Prof Sidgwick does not tell us with regard to ‘unreasonable’ action, whether the negative is to be taken as denying both these elements or only one of them. If it is to be taken as denying both, then our previous argument not only destroys the plausibility of Professor Sidgwick’s statement of the Libertarian contention, but renders it manifestly unfair. But I imagine that Professor Sidgwick would adopt the other alternative, that the negative only denies that the action partakes of the ‘ideal’, not that it is thought to be ‘practically possible’; for this certainly seems to be in accordance with common usage. In that case, however, the notions of ‘reasonable’ and ‘unreasonable’, as applied to conduct have for a common property that both are reasonable:38 a result which seems to give reason for reflection. For what is it, then, that distinguishes the ‘reasonable’ from the ‘unreasonable’? Only that one is ‘ideal’ and the other the opposite of that. In short, we cannot get on with an ‘unanalysable notion’39 which, after all turns out to be analysable into two distinct and separable elements. Our inclusion then of ‘unreasonable’ actions among the possible objects of free choice, if free choice there be, serves to make it plainer that the practical possibility of the action contemplated has nothing to do with the possibility of the choice; though this indeed would seem to be plain enough of itself. In other words, what it may seem unreasonable to choose to do, it is not therefore unreasonable to choose. The appearance of irrationality in the first case may be due to the fact that, when I say ‘choose to do’, I do imply that, if the choice is made, the action may reasonably be expected to follow from it. But this reasonable expectation cannot be admitted as an ultimate element in the definition of a reasonable action as that ‘which ought to be’. The ‘wider sense’ in which Prof Sidgwick admits ‘I sometimes judge that I “ought” to know what a wiser man would know’ is not only unable to be ‘conveniently discarded in ordinary discourse’ (p. 35), but must be regarded as the fundamental one, if we are to get a clear notion of the issue in the Free Will controversy. It would indeed be a ‘futile’ conception, ‘if it 38 39
Sic – but surely Moore intended ‘practically possible’ (as indeed occurs in the corresponding passage in the 1898 dissertation, p. 218). In the fifth edition Sidgwick writes, concerning ‘the notion of “ought” or “moral obligation” as used in our common moral judgments’ (ME p. 33): ‘I find the notion that we have been examining elementary and unanalysable. As it now exists in our thought, it cannot be resolved into any more simple notions’ (ME pp. 34–5). By the seventh edition all reference to ‘ought’ as ‘unanalysable’ has been removed, and all he says here is that ‘ought’ is a fundamental notion which ‘cannot be resolved into any more simple notions’ (ME 7th edn, pp. 32–3). As with the other changes noted above (see note 30), one wonders whether reading Moore’s dissertation motivated this change.
Freedom
47
had no relation to practice’ (p. 35 n); but we cannot, without begging the question with regard to Free Will, assume that the relation consists in the fact that what is so judged is also thereby judged to be always practically possible. And this I think is also plain from what Prof Sidgwick himself subsequently says in his chapter on Free Will (p. 67). He here admits that, after we have made up our minds on the question whether I can do a thing, if I choose, ‘still the question remains “Can I choose what – if I can choose – I judge to be right to do?”.’ Here the ambiguous form ‘Can I choose to do?’ is at last discarded; and Professor Sidgwick here explains that what he thinks his consciousness affirms is ‘that I can choose’. He adds, it is true, ‘within the limits above explained’. But my point is that by putting the question in this form he has actually removed ‘the limits above explained’. They appear in the sentence just quoted in the proviso ‘what – if I can choose – I judge to be right to do’. But does not that proviso make the question into nonsense? For note that the question is can I choose what I conceive to be right (p. 65). But if, before I can decide whether it is right, I have to decide that I can choose it, this question must be answered before it can be asked: which is absurd. Professor Sidgwick seems here expressly to state a relation of complete mutual dependence between the questions ‘Is this right?’ and ‘Is this a possible object of free choice?’, a dependence which he maintained in a qualified form on p. 35; and, if this holds, it is certainly impossible to answer the Free Will question in the form ‘can I choose what I conceive to be right or reasonable?’. Surely the only escape from this dilemma is the view expressed above that from the definition of ‘reasonable’ as ‘what is in conformity with practical reason’ we must absolutely exclude all consideration of physical possibility: and this, I suspect, does make the Free Will question one of ‘an arbitrary freak of unmotived willing’,40 and yet a question which must be answered, before what Professor Sidgwick takes to be the question ‘from an ethical point of view’, can be asked. These considerations, I think, sufficiently prove that we cannot separate the Free Will question, in a supposed ethical sense, from the wider question; nor suppose that consciousness affirms the Libertarian doctrine in the former sense, without prejudging the latter. It may be that this makes the question one even more unimportant for Ethics than Professor Sidgwick thinks it; but that will be afterwards discussed. There still remains the question[;] does consciousness affirm in the moment of choice that I can choose what I judge to be good or bad? – if I may state the issue 40
ME p. 65 n. 1.
48
The 1897 dissertation
in that way, to get rid of the question-begging ‘choose to do’ and doubly question-begging ‘reasonable’. Now it is first to be noted with regard to this question, that it only concerns the possibility of choosing; that the free will controversy seems to have narrowed itself to the question of free choice. For it is only choice which distinguishes voluntary from non-voluntary action, and the ordinary Libertarian would hardly maintain that non-voluntary actions could be free. But, on the other hand, the question is also seen to be a wider one than that which is ordinarily discussed – than is stated, for example, by Professor Sidgwick. For since, as has been shewn, the physical possibility of the action, which is the possible object of choice, cannot be considered to be a necessary element in constituting it good or reasonable, in the sense which is fundamental for practical reason, it seems hardly possible to exclude mere choices, such as that I should have the genius of Shakespeare, though that I should have it might be reasonably considered physically impossible. Even such a case as a choice to prevent the sun rising tomorrow can hardly be excluded from the class which Professor Sidgwick recognises as ‘a species of volition’ – ‘the adoption of an object of desire as an end to be aimed at’ (p. 6041 ). For though, perhaps, no one but a madman would make such a choice, yet his choice would prove that it can be made: and we ourselves do often choose through ignorance what is impossible in this sense; the only reason why we do not choose what we also think impossible, seems to be not that we cannot, (either in the deterministic or libertarian sense) but that it does not seem worth while. The question, whether a choice will produce in any degree the effect chosen, seems to be merely one for experience to decide, and we judge of it just as we judge of the probabilities and possibilities of events in the physical world. It does not seem to be concerned in the Free Will controversy, if the issue of that controversy be clearly stated. Locke and Humea , indeed, agree marvellously in their treatment of Liberty, both asserting that it means simply ‘a power to act, as we choose’. But it would seem to be for this very reason, that they are able to treat the Free Will controversy so cavalierly as they do. If the question were merely as to whether we did not sometimes do what we chose, it would, as they
a
41
Locke: Essay ii 21, 14 foll. Hume: Treatise iv p. 110 foll. [Moore’s reference to Hume’s Treatise is unintelligible – it has no book ‘iv’ and we have not found an edition of it available to Moore where there is a relevant discussion on pages 110ff. In fact the clearest statement by Hume of the position Moore discusses occurs in An Enquiry concerning Human Understanding section viii (‘Of Liberty and Necessity’.] ME 7th edn, p. 61.
Freedom
49
say, be obvious what answer we should give; but it would not, as they also say, be obvious that liberty in this sense was not contrary to ‘necessity’ since the question ‘Can we choose?’ would still remain unanswered. They both, it is true, leave an ambiguity even on the first point by not sufficiently considering what is implied in their notion of ‘power’; but, nevertheless, when they speak of a power to act, as we choose, they would appear to mean only as Locke saysa that the existence or not of the action is dependent upon our choice. In this part of free action, then, it may be admitted that they leave no room for anything contrary to ‘necessity’; since their notion is that the action is necessitated by the choice. Locke, however, sees that the point in dispute occurs not here but in the question: Are we free to choose?.b And this he dismisses as absurd, on the ground that it means: Can a man will, what he wills? But it does not mean this, unless his definition of freedom, as power to do what I choose, be already accepted. It would indeed be absurd to ask ‘Can I choose to choose?’, in the sense ‘Am I free to choose which of two alternatives I will choose?’ But Locke has no right to assume that this is meant by the question ‘Am I free to choose?’ That question may mean ‘Am I the original cause of my choice?’; and this he leaves undiscussed. Both Locke and Hume, therefore, neglect the point of the controversy by their definition of freedom. They have, however, done some service to the question, in as much as their treatment of it is a protest against that confusion of freedom ‘to do, if I choose’, and freedom ‘to choose’, which I have just pointed out. Their defect is that they assume that it was an answer to the first only which was really wanted; and hence their contempt of the dispute. As a matter of fact, I am free, in the ordinary political sense, when ‘I do what I choose, because I choose it’, since there the immediate cause of my action lies in myself, i.e. in my choice. But that is not the freedom demanded by Libertarians. What they wish to maintain is that the choice itself is caused only by a self, which is an uncaused entity; and this implies that where alternatives are presented, the choice between them is wholly independent of their previous habits, dispositions etc. The question then is: ‘Does consciousness affirm, when alternatives are presented, that I can choose any of them that I think either good or bad’, which would seem to be equivalent to ‘any conceivable presented alternative’? And with this we come to the last ambiguity of statement which seems to me to stand in the way of our giving a clear answer to the question. a
§27 [i.e. – Essay book ii ch. 21 §27].
b
§22 [i.e. – Essay book ii ch. 21 §22].
50
The 1897 dissertation
Professor Clifforda rather ingeniously urges that, if the deliverance of my consciousness is to be ‘of any use in the controversy’, it must be ‘competent to assure me of the non-existence of something which by hypothesis is not in my consciousness’ i.e. the subconscious mental elements which the Determinist must suppose to determine the choice. But it seems possible to surmount this objection by maintaining that it is enough, if consciousness can make a positive affirmation as to what is cause of the choice, without requiring it to prove exhaustively that nothing else in the world can be. If the man of science, before he enunciates a law, is always bound to prove, that no other elements besides those, whose constant connection with the effect has been observed by him, really contributed to it, no scientific laws have been discovered yet. If therefore consciousness does affirm that ‘I’ am the cause of the choice, that should be sufficient. But then the question arises what can it mean by ‘I’? Is it quite certain that when consciousness seems to affirm that ‘I can choose so and so’, it means more than ‘it is possible that such and such a choice will take place in my mind’? If it does not mean more than this, its affirmation is not against Determinism; since, as we have tried to shew above (p. 40), even on the Deterministic hypothesis, it must always be entitled to affirm the latter proposition, even if it does not always exercise its right. For by saying that such and such a choice is possible I imagine we can mean no more than that we do not know but that it will happen; and even if ‘the uniformity of nature’ can be proved in such a sense as to justify an assertion with regard to any event whatever that it certainly will not happen, this can never be the case with regard to an event conditioned by a conscious forecast. The cases in which it is ‘on Determinist principles legitimate to conclude it to be certain – and not merely highly probable – that I shall deliberately choose to do what I judge to be unwise’ seem to be not merely ‘rare’ as Professor Sidgwick says (p. 67)42 but absolutely non-existent. It seems, therefore, that the ‘affirmation of consciousness’ as against Determinism disappears on the attempt to make it precise. The attempt to find a more exact meaning for the vulgar notion of freedom has thrown us back upon the exception with which we started. Instead of free action being the action of an ‘uncaused self’, we have to be content with it as selfcaused action: anything may be said to act freely in so far as the immediate a
42
Lectures on Ethics p. 327. [Moore here refers to the same essay ‘Right and Wrong: the Scientific Ground of their Distinction’ as he had referred to earlier on p. 21 note a; the passage Moore refers to occurs on pp. 143–4 of volume ii of Clifford’s Lectures and Essays, London: Macmillan, 1901.] ME 7th edn, p. 65.
Freedom
51
cause of its change lies in itself. We have now to see how this notion is connected with that explained above as derived from Kant; and to examine whether there is any justification for applying it in an exclusive sense to Will – a restriction which Kant seems to adopt in the application of his notion also. If the restriction turns out to be unjustifiable in both cases, we shall have disposed of Schopenhauer’s view of the ultimate reality – a view which according to Kuno Fischer is also that of Kant. It will then remain to discuss the connection of Freedom with Ethics. sense in which empirical objects have ‘self’ and are free It was one of Kant’s great merits in the Critique of Pure Reason to have pointed out that there is nothing absolutely ‘inner’a in the objects of experience, either of the outer or inner sense, either in nature or in mind. He gave the final blow to the doctrine of ‘essences’ and ‘faculties’, as principles of explanation, by shewing that advance in scientific knowledge presupposed the complete interdependence of things; that all we can know for certain about them is their relation to one another. This indeed was one of his motives to his distinction of Phenomena and Dinge an Sich; for he could not avoid the conviction, though he could not justify it, that there must be something self-subsistent somewhere. But his main point was that if you treated natural objects as if they were self-subsistent, you could not escape the most unbearable contradictions. This was the ‘natural dialectic of Reason’. In the Critique of Judgment, however, he began to see that he had overemphasised the doctrine that all we can know is mere relations. He here recognises that a philosophy of nature must take into account the ‘matter of knowledge’ as such, since it too must have some element of necessity. Thus it is not only the categories and the pure forms of Intuition which have an a priori certainty, but the sense-manifold must also be of such a nature that the categories and forms of intuition will apply to it. It must be of such a nature as to supply terms to these relations. And, though the amount of the nature of objects of experience, which is thus determined a priori, is far from giving them a claim to be considered completely rational, it yet gives them a certain amount of inwardness and self-subsistence. Thus, in considering the course of Nature, it becomes obvious that, though we try to explain what happens, by referring it in each case to something prior in time, and so on ad infinitum, there is also presented a
e.g. in the ‘Amphibole’ R.V. p. 235 [B334, GW 375].
52
The 1897 dissertation
another element, left out of account by this method (the only one allowed by Kant in the Pure Reason), which also helps to explain what happens. This element is the actual qualitative nature of the events we are trying to explain. So far as merely causal connection is concerned, there is no reason why there should be any change in the world whatever, except that which is involved in the lapse of time. Each moment of time is different from the one before, and, if the world were quite without other differences, there would yet be a necessary connection between its state at one moment and its state at the next, exactly fulfilling the type which Kant sought to prove against Hume. For the state of the world at one moment would be a different thing from its state at the next, in the sense in which Hume denied that you were really entitled to infer from the existence of one thing the subsequent existence of another. But, even if this were so, causation would obviously not afford a complete explanation of the course of nature. The world which did thus persist unchanged through time, would still itself be part of the reason of the course of Nature. We could not exhaust our knowledge of each successive state by saying it was such as to have been the effect of the one before and the cause of the one which followed it. It would still remain true that each state was what it was, besides being related to those before and after it; each would have a content – the content in virtue of which each was identical with that of every other; and the nature of this content would require to be taken into account in explaining each state. We can assert a priori not only that each state of the world must be necessarily connected with those that precede and follow it, but also that it must have some definite qualitative nature. It is not only what it is because the previous state was what it is, but because it is what it is. This consideration seems obvious enough, but yet it is one which is very apt to be neglected. It was recognized in the Aristotelian doctrine of ‘formal’ causes; but has been put out of sight by the procedure of modern science, which seeks always for efficient causes, without sufficiently considering that there could be no efficiency unless there were also ‘form’. It is no doubt of more practical importance to establish the relations between things than just to recognise those things; and Kant in his desire to justify the methods of natural science, seems to have been misled by the prominence given in it to the discovery of relations, into an unjustifiable neglect of the qualitative aspect of things. There was also, as has been pointed out above, another reason for the emphasis which he lays on relation – namely his desire to protest against the assumption that the objects of experience were real, or absolutely self-subsistent. And finally, quality, so far as it is necessary, is only
Freedom
53
one: there are not, as in the case of relation, a number of different forms to justify. But still, from a philosophical point of view, it seems to be of equal importance, and is always presupposed by science in discriminating the things between which relations are to be discovered. Things, then, in so far as they must be terms of relations, may be said to have a self. But the degree of selfhood would not suffice to define the notion of freedom. For we are as yet not entitled a priori to infer in the world any difference of quality. And if there were none, as in the case above supposed, there would be no reason to suppose that the causal connection between the successive world-states was in any way dependent on their qualities. The quality would necessarily be taken into account in explaining the series as a whole; but the causal connection might be considered to hold between them purely as existing, in so far as they had matter in the Aristotelian sense, not in virtue of their form. And this, it is to be noted, is all that Kant proves for causality in the Critique of Pure Reason; the necessary connection is between the existence (Dasein) of things. But as a matter of fact there are differences of quality in the world of experience; and whatever be the justification for it, there is connected with this difference of quality a most important addition to the notion of causality. Causality in Kant’s sense would not justify any Law of Nature, and yet without these science would be impossible. There is implied in any law, that ‘Like cause has like effect’ and vice-versˆa; and in this conception we have at once the causal relation between things conceived as depending on their qualitative nature. It is no longer the thing, considered as individuated merely in time, which is necessarily connected with those preceding and following, but the thing, as distinguished by a particular quality, is considered to have a necessary connection with other things so distinguished. It is not assumed that all the qualities in the world might not be different from what they are; but it is assumed that, given any one quality, it has a unique causal relation with some other one, in the sense that only the thing of which it is a quality can be cause of the thing of which that other is a quality, and only that other thing can be the effect of the first thing. With this we seem to have arrived at the notion of a thing with a distinguishable self, having a distinct efficiency in virtue of that self. And in this conception of the course of nature there is contained the union of Determination with Freedom, in its simplest form. Each thing, marked by a simple qualitative nature, is no doubt determined in that it is the effect of some other thing, and given that other thing, it was forced to appear. But also it is itself similarly the cause of something else, and free so far as its effect depends upon its own nature. It is nothing against this that its
54
The 1897 dissertation
own nature depends in its turn upon something else; for that something else could not by itself have produced the effect which it produces. It is an essential link in the chain, and though the effect is not solely due to it, some part of the effect is due to it and to it only. notion of ‘organism’, as possessed of freedom, untenable In this notion, then, we seem to have a quite precise distinction between free and necessary action. Every action is both free and necessary; each is an aspect under which every event must be necessarily regarded. And thus we have an ultimate and valid meaning for freedom, as appearing in the world of experience; a meaning, which in consequence of the Kantian criticism, substitutes within that world the notion of the self-caused for that of the uncaused or original, which had been maintained by the pre-Kantian metaphysics. But neither common sense nor Kant himself are satisfied with this. Both wish to drag into the world of experience a notion of freedom which would be inconsistent with complete mechanical determination, as applying only to some of the objects of experience, and not merely as an aspect from which every mechanically determined event alike may be regarded. This is attempted no longer by maintaining the freedom of conscious will, in the sense of ‘Liberty of Indifference’, – the old notion, discarded by Kant, against which I have hitherto directed my attack – but by treating certain groups of qualities, discoverable in the world of experience, as bound together in a so-called ‘organic’ unity, which confers on them a ‘self’ in a special sense, and makes not their parts, but them, ultimate factors in natural processes. The human mind may then be regarded as a special instance of this sort of unity,43 and so capable of free action as a whole, differing in degree perhaps, but of the same kind as that exercised by any organism. Kant, it is true, in treating this notion of ‘organism’ in the Critique of Judgment, still asserts that it is ‘regulative’ or ‘reflective’ only, not ‘constitutive’ or ‘determinant’. By this he means that our experience would be possible without it. We shall still be able to have knowledge of an objective world, if all things in it were only mechanically determined. But nevertheless he does now maintain that the explanation of certain things presented in experience, namely organisms, is absolutely impossible according to mechanical laws. By this he appears to mean that their ‘form’ is wholly 43
Moore’s marginal comment: ‘Kant?’
Freedom
55
‘contingent’ from a mechanical point of view;a and this contingent form is such that all the parts of the thing may be regarded as acting for the sake of each other’s ‘form’, i.e. the form of each is end as well as means to all the rest.b Thus, in order to define an organism, Kant is obliged to use, in describing the relation of the parts to one another, the notion of ‘end’, which he admits to be only given by the Practical Reason. He says that the form of the whole organism can only be regarded as the ‘end’ of the parts (and that only ‘regulatively’) on condition that the parts are ends to one another. But apparently his only reason for thinking that they are so is that their ‘form’ is inexplicable by mere mechanical causation; and this ‘form’ is no longer the systematic unity which it was in the case of the whole, but the presupposition of that. So that it would seem to be undistinguished from ‘form’ in the sense above explained, in which every natural object equally may be said to have a ‘form’. There is therefore no reason primˆa facie, why an organism should be distinguished as having a systematic unity, which would not apply to any group of interacting parts whatever; since in every case, as was shewn above, the action must be conceived as affecting the ‘form’, and likewise in every case the mere principles of the understanding will give no reason why the form should be such as it is. And as for the inference from inexplicability of form to the necessity of ‘end’, that requires a long discussion, which will occupy us later, in seeking to connect ‘freedom’ with Ethics. But, apart from this it seems to be a fact that Kant’s only reason for saying that in organisms we have presented a whole or systematic unity, such as cannot be explained by natural laws, is the assumption that its parts can be seen to [be] related in a non-mechanical way. And his ground for assuming this, again, is one that applies equally to all causally related elements. So that he fails, on the whole, to justify even as a regulative principle the special position which he attributes to organic things. On the other hand, if we confine our attention to the relation of the parts, there is, without employing the notion of end, something in this relation which cannot be explained by the mere ‘principles of the understanding’. Our knowledge of nature would be impossible, unless things did act according to rules, i.e. unless the form of a thing could be taken to indicate what its effect would be. This is a principle which Kant seems a b
[Kritik der Urteilskraft, ed. Hartenstein, vol. 5] p. 372 [AK 5: 360; Critique of Judgment trans. Meredith, ed. Walker, Oxford University Press, Oxford 2007, p. 188]. [Kritik der Urteilskraft, ed. Hartenstein, vol. 5] p. 385 [AK 5: 373; Critique of Judgment trans. Meredith, ed. Walker, Oxford University Press, Oxford 2007, p. 201].
56
The 1897 dissertation
merely to assumea , but it would seem to be a properly constitutive one. In the Critique of Judgment, it is, however, confused with another principle which Kant does declare to be ‘regulative’. It is here asserted that, if we are to have knowledge of nature, the variety of qualities presented in the sense-manifold must not wholly exclude recurrence; and this is taken to be equivalent to the principles of ‘homogeneity, specification and continuity’. These are principles according to which we should always aim at a systematic classification of the Laws of Nature themselves, hoping to find a continuous chain from one, general enough to embrace them all, down through regulative subdivisions, to infinite particularity. Kant’s view is that we cannot know a priori that Nature is capable of such continuous classification, and there do actually appear to be gaps which spoil its symmetry; but that the ideal of science is to reach it, and that its progress consists in continual approximation to it. It is therefore a ‘regulative’ but not a ‘constitutive’ Idea. Now, whether this be so or not, it seems extremely important to distinguish these principles, from the notion with which Kant connects them. It is true that there must be some recurrence of identical elements in experience, in order that we may find any Laws of Nature at all; but it does not follow from this that these Laws will themselves be capable of a completely systematic classification. The two principles seem to be on an entirely different level, whereas Kant takes them both indifferently as evidence that Nature as a whole must be regarded as if it had been designedly adapted to our intelligence. The truth is that the one is presupposed by the principle ‘Like cause, like effect’, and so is necessary to a knowledge of nature as mechanically determined; whereas the other may indeed be regarded as a mere ideal. This ideal of perfect classification would indeed presuppose the other; but the inverse proposition does not hold. Kant’s reasons, therefore, for holding the Idea of complete system to be purely regulative, do not invalidate the objectivity of the principle that all the elements in nature reciprocally determine one another’s form, and that, therefore, though the form of Nature as a whole must be regarded as contingent, the form of the parts in relation to one another is necessary. On the other hand, the notion of a systematic whole does not apply objectively either to Nature as a whole, or to any group of parts in nature. We must therefore be content, so far as the doctrine of experience is concerned, with regarding every element in Nature as equally free and determined; but this much we may regard as fact. It is, no doubt, convenient to treat a
R.V. p. 380 [B577, GW 541]: ‘ . . . jede Ursache eine Regel voraussetzt . . . und jede Regel eine Gleichf¨ormingkeit den Wirkungen erfordet . . . ’ [‘every cause presupposes a rule . . . and every rule requires a uniformity in its effects’].
Freedom
57
certain groups as wholes, and to ascribe freedom to them in the sense that their parts and not anything external to them are immediate cause of their changes; but this grouping must be admitted to be wholly arbitrary. It is not merely, as Kant thinks, that a mechanical determination of the parts is the only one that can be presupposed as actual; but also that there is nothing to explain in the whole, which is not also in each of the parts, taken by themselves, and that presupposed by their mechanical connection. Now the case of human volition would seem to be precisely the same. This, too, is to be explained mechanically; but is free, in virtue of that very mechanical explanation, in so far as the form of each element in the series may be regarded as determining the form of the next. And there is no more ground for treating the mind as a whole as free, than there is for treating an organism as free in nature. If a precise and ultimate meaning is to be given to such freedom, it must be explained as that which properly belongs to the ultimate mental elements, not as anything which resides in the whole, as such. For there is here again no special form to fix the limits of a unity anywhere between the smallest distinct element and the whole mental world. But, on the other hand, it may be admitted that the activity of mental elements is sometimes directly perceived, in a sense in which that of physical elements never can be. For consciousness itself is an element in mental processes, so that here the form which determines the change knows itself. However, the important point to emphasize is that, from the point of view of explanation of experience, this makes no difference. In the context of inner experience this form plays just the same part as any other form in outer experience. The difference is only to itself; it is not of an objective significance. From the common point of view, then, which takes the world of experience as ultimately real, this, in which every part of that world is alike free and alike determined, is the only sense of freedom, which can withstand criticism, as in no way based on arbitrary distinctions. It is a sense, which would to most seem to be the same as that of determination. But it can, I think, be seen to underlie all common uses of freedom: and it is only to the difficulty of distinguishing it as an irreducible aspect in mechanical causation, that there is to be attributed the mistaken attempt to shew that the notion of freedom is irreducible, by maintaining the existence of uncaused choice. We have now to shew the connection between this and Kant’s sense of Freedom; and to consider whether he can give any justification for speaking as if his, any more than the popular, sense were to be found exhibited in special cases in the world of experience, i.e. especially in human volition, and not everywhere alike.
58
The 1897 dissertation
kant’s freedom takes account of freedom, in only sense, in which we have found it to belong to natural objects, but his freedom does not belong to ‘will’ either Kant’s use of the term Freedom does seem justified in that it coincides with the popular one in opposing the view that the aspect of the world as causally determined is alone sufficient for its explanation. Both alike recognise that to define a thing’s relation to other things is not the most that can be done in knowing the world. But we have seen that Kant does not, in the Critique of Pure Reason, seem to allow that freedom, in the sense just explained, where it consists in recognising the part played in nature by each thing’s form, is an objective notion. His failure to do so has been explained as due to his confusion of this notion with the wider one of systematic unity in Nature, which involves it. It is in the same way that his notion of Freedom involves the common one, but also goes much further. According to him Freedom means not only that each part of Nature is necessarily connected with all the other parts in respect of its form as well as in respect of its existence; but also that all these different forms, considered in themselves, together with their differences and the laws of their connection, must be taken into account in explaining the world as a whole: and since the world as a whole is an impossible conception, if the objects of experience be taken to be its ultimate constituents, since they are necessarily conceived as in the infinite forms of space and time, the complete reason of all that appears must be placed in a supersensible reality. This supersensible reality is the world as a whole, and is the reason of everything that appears; and, as such, it has Freedom. As such, too, Kant will not allow it to be known as more than a mere Idea; but we have seen reason to think that this was only due to his failure to reconcile two different criteria of reality: so that he generally considers the being given in the context of experience essential, and since the context of experience can never offer the required completeness, such completeness must be condemned as merely regulative. Now our question is, whether a relation, really analogous to this of the real world to the world of experience, is presented in the relation of the human will to its actions, and in that alone among the objects of experience. Kant himself distinguishes between freedom ‘in the cosmological sense’ (which is the one we have hitherto discussed as his) and freedom ‘in the practical sense’; but he asserts that the latter is possible, only if the former be also possible.a In the Critique of Practical Reason, he proves that ‘practical a
R.V. p. 371 [B561, GW 533].
Freedom
59
Freedom’ is actual, and from that infers that ‘cosmological Freedom’ is also actual.a What then is his account of ‘practical Freedom’? ‘Practical Freedom’ is something which must belong to all ‘reasonable’ beings, as such.b It is defined negatively as ‘the independence of our choice from compulsion through impulses of sense’;c and positively as ‘a power’ or ‘causality’ of ‘reason’, ‘to begin a series of events entirely of itself’.d ‘Pure practical reason’ is identified with ‘pure will’; and ‘will’ again, up to the end of the Critique of Practical Reason, seems to be identical with ‘choice’ (Willk¨uhr), though in the Preface to the Metaphysic of Morals,e they are distinguished in a very important manner; for it is there declared that only ‘choice’ can be called ‘free’, ‘will’ being concerned not with ‘actions’, but only, like practical Reason, with the giving of Moral Laws. Kant’s account of the way in which we must conceive ‘practical freedom’ in relation to experience is as follows. Every ‘cause’ (Ursache) has a ‘power’ (Verm¨ogen), which may also be called its ‘causality’ (Causalit¨at) which ‘power’ is necessarily connected with the subsequent appearance of a definite effect (Wirkung); and the law of this connection is called the ‘character’ of the cause. The transition from the ‘causality’ to the ‘effect’, however it be conceived, is called the ‘action’ of the cause (Handlung). Now in ‘natural causation’, the ‘causality’ of every cause is also an effect of some previous cause and so on ad infinitum; and the ‘action’ therefore is merely a transition in time. But for every natural object, we must also suppose there to be an intelligible ground; and there is no contradiction in thinking of this intelligible ground as cause (in another sense) of the ‘causality’ of the natural object. The ‘causality’ of the natural object would thus be effect both of some preceding natural object and also of its intelligible ground. But the intelligible ground is, as such, in no way subject to ‘time-conditions’, and therefore its ‘action’ in producing the ‘causality’ which is its appearance, is not a time-transition. It cannot therefore be said to ‘begin to act’ at any time; although its effect, i.e. the ‘causality’ of the natural object had a beginning. It is thus original cause of an appearance, which is on another side also effect of a conditional cause and in its turn cause of other appearances. It begins ‘of itself’ a series of events in time, without itself beginning to act. a c d e
b G. p. 296 [AK 4: 449, MG 96]. P.V. Pref. pp. 3–4 [AK 5: 4–5, MG 139–40]. R.V. p. 371 [B562, GW 533]; cf. M. p. 11 [AK 6: 213–14, MG 375]. R.V. p. 372 [B562, GW 534]; cf. M. p. 11 [AK 6: 213–14, MG 375]. M. p. 23 [AK 6: 226, MG 380].
60
The 1897 dissertation
Now, so far, except for the ambiguity of the word ‘cause’ as applied to an intelligible object, which was discussed above, and except for a lack of fixity about almost all his terms, many of which are at one time distinguished, and at another used as synonymous (e.g. Causalit¨at = Charakter = Handlung), there seems no reason to object to Kant’s account. But it is an account which would apply to any natural object whatever, and we have now to consider whether it will apply in a special sense to human volition. I quoted above (p. 37) a passage of Kant,a in which he says that ‘man knows himself ’ not only through his senses but ‘also through mere apperception, and that too in actions and inner determinations, which he cannot ascribe to the impression of the senses. He is to himself, it must be admitted, partly a phenomenon, but partly also, namely in view of certain faculties, a merely intelligible object, because his action cannot be ascribed to the receptivity of sensibility. We call these faculties Understanding and Reason.’b And he goes on to say that Reason appears pre-eminently as the faculty of supersensible being. We are next told that the Imperatives expressed by ‘ought’ make it plain that ‘this Reason has a Causality, or at least that we represent it as having one.’ And finally we have the following sentence: ‘Now this “ought” expresses a possible action, of which the ground is nothing more than a mere conception; whereas, on the contrary, the ground of a merely natural action must always be an appearance.’c In this passage I think we have presented the full extent to which Kant’s error of restricting practical freedom to reasonable beings goes, together with the confusions on which that restriction was based. One ambiguity occurs in the last sentence; and it is a very important one, since it seems to have given rise to many false notions of what Kant meant by freedom. This sentence expresses in an antithetical form the difference between ‘free’ and ‘natural’ causality – which he frequently says are the only two kinds of causation possible. The first is distinguished by this that its ground is a mere conception; whereas the second is always a phenomenon. Now from the account given above of Transcendental Freedom it will appear in what sense I accept this description of free causality. A free cause must necessarily appear to us as a logical reason, and so far as ‘a mere conception’; because it is not, as such, presented to us as an object of intuition. It is always a universal, and though we can know that it must be an individual; we cannot experience it as uniting both characters. But from what Kant says in the preceding context, as a c
R.V. p. 379 [B574, GW 540]. R.V. p. 379 [B575, GW 540].
b
R.V. p. 379 [B574–5, GW 540].
61
Freedom
well as from his general account of will elsewhere, I think it is plain that he is not thinking of ‘a mere conception’ in this sense. When our will is singled out as having a special kind of causality, inasmuch as it can be ‘determined to action by the presentation (Vorstellung) of certain laws’,a Kant shews what it is he is thinking of. The ‘mere conception’, in the only justifiable sense for freedom, would be the laws themselves, and not the ‘presentation’ of the laws. Every ‘conception’ may be regarded from two points of view, either as a psychical existent, or from the point of view of its content; and it is this very important (and obvious) distinction which Kant appears to have neglected. If the causation exercised by the presentation of a conception were enough to justify freedom, freedom would be no more than that aspect of every mechanical process, which was distinguished above as the only precise sense assignable to freedom on the common view which regards the objects of experience as real; and thus there would not even be an appearance of conflict between it and natural causality. For it is precisely ‘presentations’ to which Kant repeatedly asserts that the objects of experience are reduced, when they are viewed, as he holds they must be, as appearances. An appearance is a ‘mere presentation’, and it is only between such that the causal laws will hold. There would therefore be no difference between ‘an action of which the ground was no more than’ the presentation of ‘a conception’, and an action of which the ground ‘must always be an appearance’: for the presentation of anything whatever is, as such, an appearance. Kant himself would seem to recognise this in a passage of the ‘Canon of Pure Reason’, in which for that very reason he is driven to an almost direct contradiction of what he says in the context quoted above. In this passage he says: ‘Practical Freedom can be proved through experience. For not only that which charms, i.e. affects the senses directly, determines human choice, but we have a power to overcome impressions upon our sensual desiderative faculty (Begehrungsverm¨ogen), through presentationsb of what, even in a somewhat remote way, is useful or harmful; and these considerations of that which, in view of our whole state is desirable, i.e. good and useful, are based upon Reason. Hence also Reason gives laws, which are Imperatives, i.e. objective Laws of Freedom, and which tell us, what ought to happen, even though perhaps it never does happen, and are distinguished in that respect from Natural Laws, which deal only with that, which happens.’c He then goes on to suggest that on a wider view, what here appears as freedom, might be seen to be nature (which would, indeed, with regard to part of his a
G. p. 275 [AK 4: 427, MG 78].
b
My italics.
c
R.V. p. 530 [B830, GW 675].
62
The 1897 dissertation
statement, be certainly the case); but this, he says, is a speculative question, irrelevant just here. Finally he comes to this: ‘Accordingly we know practical freedom through experience as one among natural causes, namely a causality of the reason in determination of the will; whereas Transcendental Freedom demands an independence of this reason itself (in view of its causal power to begin a series of appearances) from all determining causes of the senseworld, and so far appears to be contrary to the Law of Nature, and hence to all possible experience; it therefore remains problematic. But for reason in its practical use this problem is irrelevant . . . . . . .The question with regard to Transcendental Freedom concerns solely speculative knowledge. We can set it aside as wholly irrelevant, when we have to do with the practical.’a Now in this passage Kant states very well what is characteristic of human volition; and his definitions of ‘will’ are constantly expressed in the same fashion. Will differs from other instances of natural causation, in as much as in it the idea (to use the common English word for ‘Vorstellung’) of something, which is not yet real, tends to bring about the realisation of that thing; and he may be justified in saying that the process ‘is based upon Reason’, since to have an idea of anything either real or imaginary presupposes that faculty of cognition which distinguishes man from beasts, and still more from inanimate nature. Nay, more than this, in the special instance, which Kant takes to be the only truly ‘moral’ willing, where the idea which acts as cause, is the idea of conformity to a universal law, the content of the idea is so abstract, that it may be confidently asserted that only reasonable beings are capable of having such an idea. But nevertheless the idea is even here still ‘an appearance’, and, as such separated by an impassable gulf from the content, of which it is an idea. And, inasmuch as it is in its character of idea, i.e. as a psychical existent, that it produces an effect, the causation is still merely ‘natural’. This, as we have said, Kant in the present passage fully recognises. But it is only the more remarkable that he should speak of Reason in the same context, as ‘giving laws of Freedom’, as if it were Reason in the same sense, which is the source on the one hand of objectivity, and on the other hand of abstract ideas, whether true or false. In this Kant betrays the too psychological standpoint above which he seems never to have completely risen in treating epistemological questions, in spite of the enormous services which he did to epistemology, as well in the Metaphysics of Ethics as elsewhere. He supplies, as it seems to me, more materials for a true view than any one else, and those, too in a wonderfully forward state of preparation, but nevertheless they are still a
R.V. p. 531 [A831, GW 676; Moore’s emphasis].
Freedom
63
for him encumbered and confused with the irrelevant matter, from which it was his merit to set them free for others. It is perhaps impossible to dispense with the term ’rational’, for what is true or objective, especially after its full adoption by Hegel; but it is extremely important to avoid confusing the ‘rational’ in this sense, which is the fundamental one for Kant’s system, with the ‘rational’ in the sense of that which implies the psychological faculty of making judgments and inferences. The distinction between what is true and what is only believed (although only a ‘rational’ being can believe) is one which cannot be either done away or bridged over,a however small be the amount of what we may be thought to really know in comparison of what we must be content to believe; and it is this distinction which is here in question. Knowing, the function of Reason, is on one side a natural function, and, as such, it is indistinguishable from believing; but, in so far as knowing is distinct from believing, i.e. in so far as that which is known is true, there are no two words which express a difference more profound. When Kant talks of the only true morality as based upon the laws which Reason gives itself, the whole course of his work shews that he means laws which tell us truly what ought to be done; it is, indeed, only on this condition that he could claim universality for them.b In this sense ‘Reason determines the Will’ whenever the idea which is cause of our action, is an idea of what is truly good. But it is only in an utterly different sense that ‘Reason’ can be said to ‘determine the Will’, whenever the idea, which causes our actions, implies the power of abstraction. And it is only in this second sense that such determination of the will can be called a ‘practical freedom’ which is independent of ‘Transcendental Freedom’. Accordingly Kant himself, as we have said, recognises elsewhere that ‘the practical conception of freedom is based upon’ the ‘transcendental Idea of Freedom’;c and again, speaking of freedom, ‘as one of the faculties, which contain the cause of the appearances of our sense-world’, i.e. as practical, in distinction from transcendental, freedom, he declares that we cannot a
b c
Prof Sidgwick seems to recognise this distinction in the passage discussed above (p. 41), by his use of the term ‘what I conceive to be rational’; but this does not defend him from the charge of assuming ‘only rational action is free’, since he does not point out the consequences of the distinction, and indeed (as was noted) uses ‘rational’ and ‘conceived to be rational’ in different parts of the same discussion, as convertible terms. [In the passage to which Moore alludes Sidgwick does not use the phrase ‘what I conceive to be rational’; instead Sidgwick writes of that ‘which I conceive as right or reasonable’ (ME p. 65); but Sidgwick seems not to distinguish between the rational and reasonable, as when he writes that ‘I spoke of actions that we judge to be right and what ought to be done as “reasonable” or “rational”’ (ME p. 25, 7th edn p. 23).] This is the ground of Kant’s distinction between the Categorical Imperative, or objective Law, and the mere Maxim or subjective principle which will receive fuller treatment in my second chapter. R.V. p. 371 [B561, GW 533].
64
The 1897 dissertation
hope to establish its actuality in experience, ‘inasmuch as we can never infer from experience to anything, which must not be conceived at all according to laws of experience’.a And this seems sufficient to contradict his statement in the ‘Canon’ that ‘we know practical freedom through experience as one among natural causes’ and that ‘for reason in its practical use’ the problem of transcendental freedom ‘is irrelevant’. reasons why he thought it did. confusions exposed 44 Kant has therefore confused the purely natural process of human volition, with the transcendental aspect of it, which alone entitles us to ascribe to man ‘practical freedom’; and it is solely on this confusion that the special place he assigns to man as a ‘free’ agent seems to be based. It is true that the content of the idea, which acts as cause in volition, is different from the content of any other natural cause; but that content is merely the form of the cause, and difference of form is something which in no way renders one natural cause more or less of a natural cause than any other. The question is, however, complicated by the fact that we are dealing, in the case of volition, with an ‘appearance of the inner sense’; and as this point also touches the legitimacy of Kant’s assertion that ‘appearances’ are mere ‘presentations’, an assertion which I used in my argument to prove that for him human volition was really a causation by ‘appearances’, something should be said of it. It is true, as Berkeley saw, that everything which we directly experience, may be regarded from one point of view as a state of our own mind; but it is not true that everything is directly experienced, as such a state only. Kant pointed out that Berkeley’s hypothesis would leave us no means of making a rational distinction between dreams and perceptions – between mere illusion (lauter Schein) and appearance (Erscheinung). He shews that we necessarily conceive things extended in space (the form of outer sense) as forming a connected series, of which the succession is objective, i.e. not existent merely as and when we perceive them. In fact we could not know a thing to be a state of ours, except by contrasting it with something not so dependent on us; since we can only know ourselves in contrast to other things or persons, and our knowledge of the existence of other persons depends upon inference from their manifestation in a world of things. a 44
R.V. p. 385 [B586, GW 546]. I.e. ‘Reasons why Kant thought Freedom belongs to “Will”. Confusions exposed’.
Freedom
65
Our ‘ideas’, therefore, in so far as they are states of ourselves, must be conceived as forming a causal series of their own, different in the order of its connections from that of the content of those ideas, considered as extended in space. The knowledge of the mental series, as such, involves the existence of the series of things in space; but the latter must be conceived as so far independent of the former that it might exist, without being represented in any mental series. Nevertheless it remains true that the spatial series can only be conceived as having identical elements with those which occur in the mental series. And thus it is true, in a sense, that the spatial world is composed of mental elements; only these mental elements, in so far as they are conceived as existing in one space, and belonging to one series, of which the parts are in a relation of complete causal interdependence, must not be confused with the elements, whose content is partially the same, but which form part of the mental series. Kant’s denomination of the elements of the spatial world as ‘mere presentations’ is justifiable, if it be clearly understood that these ‘presentations’ are the same in content only, with those which belong to the mental world, but not the same in respect of their causal relations with one another. And this, on the whole, he makes plain enough by his distinction of the ‘outer’ from the ‘inner’ sense. But the ambiguity of the name nevertheless lends a handle to confusion, into which he sometimes seems to fall, as, for instance, when he speaks of spatial objects as being actually ‘in our mind’ (in unserem Gem¨uthe).45 A concrete instance will make the distinction plain. The ‘presentation’ of a cow munching upon grass, may in the mental series be cause of a feeling of pleasure, or of the recollection of the words ‘cud’ or ‘ruminate’; but the same munching will probably be the cause, in the material series, of digestion, and, perhaps, of such movements in the cow’s brain as correspond to her feeling of pleasure. Yet the latter munching is certainly composed of elements which can only be known as ‘presentational’. The difference is not between the ‘states’ between which, as Kant insists, the causal connection must always be conceived to hold, but between the substances, of which, as Kant also insists, they must be conceived as states. In the case of the material cow, the substance is conceived as extended in a unique space; but in the case of the presentation of her, the substance is mental – not extended in space, 45
We have not been able to locate a passage in the Critique of Pure Reason in which Kant makes quite this claim; but Kant’s transcendental idealism is exemplified by the following passage: ‘Space itself, however, along with time, and, with both, all appearances, are not things, but rather nothing but representations, and they cannot exist at all outside our mind’ (ausser unserem Gem¨uths): B520 (GW 511).
66
The 1897 dissertation
but also conceived as persisting through time. This difference of substance adds an element of content to the states of each series, which distinguishes a state conceived as belonging to the one from the same state conceived as belonging to the other. Part of the content, then, of some mental states is necessarily conceived as forming a world by itself. But there are other mental states, as, when we are dreaming imagining, feeling pleasure and pain, or engaged in abstract thinking, of which the content is not so conceived. In these cases the content, in so far as it is conceived as existing, exists only as the form of mental states. The difference between such contents and those which are conceived as constituting the spatial world, is what is marked by the fact that Kant calls the latter contents themselves ‘presentations’. For these latter are not mere content, i.e. merely what distinguishes one mental element from another, but appear to be given as existent and to be bound together in one context of experience, exactly in the same way as the presentations of them, along with the other mental elements, appear to be given as existing and to be bound together in another context of experience. There is exactly the same reason for calling these contents themselves ‘presentations’, as for giving that name to the mental states of which they are contents and to all other mental states. The first group are ‘presentations’ of ‘the outer sense’, the second of ‘the inner’. When a psychologist examines his mind, he has, to use Hume’s terms, ‘impressions’ both of ‘impressions’ and of ‘ideas’; and ‘impression’ is used in both these cases in the same sense, that of something both immediately given, as an object of intuition (Kant’s Anschauung), and as forming part of a unique series. But when his ‘impression’ is of an ‘impression’, the latter impression necessarily appears as part of his mental series, and it was only as such that Berkeley erroneously wished to consider it; whereas when his impression is of an ‘idea’, the ‘idea’ is really a part of his mental series and nothing more. When Kant, therefore, speaks of the ‘appearances’ to which he confines our knowledge of the real world, as ‘mere presentations’, he means what might perhaps be more properly termed ‘contents of presentation’, since they need not always have actually been ‘presented’; but still he means those contents, considered as having the character conferred by presentation, i.e. as absolutely particularised or individuated by the unique time-series, not in their abstract character. Anything whatever, which ‘determines the will’, i.e. which causes action, is a ‘presentation’ in this sense, since it must be an element in the mental series and an ‘object’ (Gegenstand) of the inner sense, either possible (if it be subconscious) or actual. But confusion is apt to arise because some of these ‘presentations’ of the inner sense, are also ‘presentations’ of the outer.
Freedom
67
And a further confusion also occurs, because the ‘presentations’ of the inner sense, may be ‘ideas’ either of other ‘presentations’ of the inner sense, e.g. the ‘idea’ of my future pleasure, or of presentations of the outer, e.g. of some very good wine, not actually existent. And all these cases are different in respect of the content of the ‘presentation’, from those in which it is something abstract, like the moral law. Nevertheless, what it is important to notice and what is too often left out of sight, is that in all cases alike what causes action is some ‘presentation’ of the inner sense, whether its content be the moral law or the ‘idea’ of another presentation; and that thus in all cases the ‘will’ is subject to the ‘natural’ law of causality, and ‘determined’ by an ‘appearance’. Now it is characteristic of ‘voluntary’ action, i.e. of human volition, as opposed to conation in general and to some forms even of desire, not only that the presentation of the inner sense which is its cause, must always be an actual presentation, as opposed to one that is subconscious, but also that this presentation must have for content a mere ‘idea’ of another presentation of the outer or inner sense, and not an actual presentation either of the one or the other. Thus it is possible to desire a pear, which you actually see, and this desire may be the cause of your eating it; but, if this is all, the action will not be voluntary. To make it voluntary, the desire must not be of the pear, but of the taking and eating it, actions which are not yet actually prevented. It is this which would seem to distinguish actions, which are automatic, as a result of habit, from those which are merely instinctive. Both alike are caused by ‘presentations’ of the inner sense, which are not actually presented but subconscious; only in the first the subconscious ‘presentation’ must be of something itself not actual, whereas in the case of instinct, it may be of an actual presentation of the outer sense. For instance, there seems little doubt, that a dog may be directly moved by the sight of a man, whom he hates, to attack him. In which case the presentation of the man, is a subconscious presentation from the point of view of the inner sense (for the dog probably does not perceive that he sees the man); but it is not a subconscious presentation of the attack, as it might be in a man who had consciously formed the habit of caning a certain person, whenever he met him. These distinctions are all of them important in considering Kant’s account of volition, as with regard to all of them he uses at times ambiguous language; and we shall have to take account of them later in considering his treatment of Hedonism. What we are concerned with now, however, is their bearing on Kant’s view of practical freedom, as belonging only to man. And I think they serve to explain another confusion in the passage quoted
68
The 1897 dissertation
above.a,46 It is not only that ‘conception’ is there used as equivalent to ‘presentation of conception’, or conception as psychical existent; but there seems also to be an ambiguity in the sense in which he speaks of man’s action as characterised by the fact that it ‘cannot be ascribed to the receptivity of sensibility’. This description is opposed to that of ‘lifeless or merely animal’ nature, in which we have no reason to ‘think of any power, as other than sensuously conditioned (sinnlich bedingtb ). Now by this latter expression it seems plain that what Kant means, is that the actions of all other parts of nature have sensible objects for their causes. It is in this sense that it forms a true antithesis to what immediately follows: ‘Man, who knows the whole of nature besides solely through senses, gets knowledge of himself also through mere apperception’; i.e. other animals and inanimate nature are mere objects of the senses, and therefore their actions can only be caused by objects of the senses, which is the type of natural causality; whereas man is something more. But, when Kant goes on to speak of man as recognising actions of his own ‘which he cannot ascribe to the impressions of the senses’, it is hard to believe that those ‘actions which can be ascribed to sense impressions’ are not actions of man’s will, under the aspect in which he calls it ‘sensual’ (sinnlich), a term which he explains as meaning that it is ‘affected pathologically (through motives (Bewegursachen) of sense)’.c Now, if this be so, Kant would seem to be treating three totally distinct notions as all equally relevant for his purpose. Man is sensual (1) in so far as his actions may be regarded as sensible objects, belonging to a context of experience (2) in so far as he receives impressions of the senses, i.e. in so far as he is capable of perceiving sensible objects (3) in so far as the ‘idea’ of sensible objects is the cause of his actions. And again, in all these divisions, an ambiguity remains because the ‘sensible objects’ may be objects either of the outer or of the inner sense. Now, when Kant speaks of ‘heteronomy’ of the will, that is a distinction based upon the third sense in which man is ‘sensual’. The will is heteronomous, when that which causes the action is either the ‘idea’ of some sensible object, either of outer or inner sense, or the actual presentation of some object of outer sense. Even here, however, Kant is ambiguous, since he sometimes speaks as if the actual presentation of an object of outer sense, were necessary a c 46
R.V. p. 379 [B575, GW 540]. R.V. p. 371 [B562, GW 533].
b
R.V. p. 379 [B574, GW 540].
The passage Moore is here alluding to as ‘quoted above’ is, we think, ‘Now this “ought” expresses a possible action, of which the ground is nothing more than a mere conception’ (see p. 60 note c above). Moore here uses the word ‘conception’ to translate Kant’s Begriff, which is more often translated ‘concept’.
Freedom
69
to constitute heteronomy – such a case as that given above of the desire of a seen pear; and, where he does not appear to limit the conception so narrowly as this, he seems almost always to be thinking of the ‘idea’ of some object of the outer sense, e.g. the taking of the pear, and hardly ever of the ‘idea’ of some object of the inner sense, e.g. the sensations of touch and taste that accompany the eating of the pear, or the pleasure they excite. Yet it is plain that where the cause of action is the ‘idea’ of some future state of the agent’s mind, the action is fully heteronomous. But the confusion which bears on ‘practical freedom’ is that Kant’s language encourages the notion that where action is not heteronomous, i.e. does not fall under (3), it is therefore not naturally caused, i.e. does not fall under (1); which is by no means the case, although all action which is heteronomous must fall under (1). Man’s sensuousness under (2) is irrelevant both to heteronomy and to practical freedom. Kant is here committing the error of regarding the relation of subject and object, in experience, as the relation of ordinary causality, which holds only between objects. He traces the possibility of heteronomy quite rightly to the fact that our experience is sensuous – that we do not know objects a priori, in virtue of an intellectual intuition. But he seems to regard the effect of objects upon us, whereby we know them, as the same thing as their effect in determining the will; whereas the knowledge of objects cannot be rightly regarded as an effect upon our sensibility at all, and, if it were, it would not be the same thing with the effect which the actual presentation of objects or the ‘idea’ of them, regarded as psychical existents, produces in subsequent states of mind. Yet Kant seems to regard the undoubted fact of sense-perception as a confirmation of the reality of heteronomy, and, through that, of his misleading description of ‘practical freedom’ as something empirically knowable. It is in man’s sensuousness under (1) that we reach what is really relevant to freedom. When this aspect of his nature is clearly distinguished from (2) and (3) it becomes plain that all his actions fall simply under ‘natural laws’. They are all mere ‘appearances’ of the inner sense, and are completely caused by previous appearances of the inner sense. It is true, that, where the cause of an action is the presentation of a law, the law which is presented is not itself a sensuous object, and herein there lies a real distinction between the actions of men and beasts. But what is important is that the law must be presented to the inner sense, before it can be conceived as causing the action; and as a presentation of the inner sense, it is as truly a sensuous object as any other, though only of the inner sense. Thus we might truly say of man, what Kant says of beasts, that we have no reason to think of any of his
70
The 1897 dissertation
faculties as ‘other than sensuously conditioned’. The difference between men and beasts is only that beasts are apparently incapable of forming abstractions, i.e. the content of the presentation which causes their action must always be itself a presentation: in other words, because they know nothing but sensible objects, their action is always heteronomous. But this difference between man and beasts only proves him to be a different sort of natural object from them; it is by no means sufficient to entitle us to ascribe ‘practical freedom’ to him. Man does not, when he acts morally, exhibit any ‘pure activity’.a This ‘pure activity’, in so far as it can be ascribed to him at all, consists only in his being able to conceive the Ideas of Reason;b that, when they have been conceived, they should influence his action, cannot be regarded as an instance of the same activity. And this power of conception itself cannot, without danger of misleading, be called a ‘pure activity’. That is an expression which suggests that the Ego to which Kant attributes it, may from the cognitive point of view, at all events, be regarded as itself an uncaused cause of its pure notions, e.g. the Ideas of Reason. But, as has been pointed out at length above, Kant could not maintain that the Ding an Sich or the Transcendental Ego really stood in a causal relation to experience: their relation to it is rather that of reason and consequence. To maintain the opposite is to identify the ‘synthesis’ of the Transcendental Ego, with the actual process of judging, and to make the Transcendental Ego itself the psychological subject. If the relation of the premises of a syllogism to the conclusion could be called ‘activity’, there might be nothing misleading in the notion of a ‘pure activity’. But, as a matter of fact, it is rather the relation of the recognition of the premises to the recognition of the conclusion as conditioned by them, that seems to correspond to our notion of cognitive activity; and such a process Kant enables us to set aside, as not what he means by ‘pure activity’, since it involves time. ‘pure will’ impossible as implying time. no practical freedom Indeed, Kant himself sufficiently prevents misapprehension, by the rigour with which he rejects the attempt to conceive as prior in time, that which, if it determine the will, shews that will to be ‘practically free’. In this rejection he is quite consistent. ‘The action’ he says ‘so far as it is to be imputed to thought’ (Denkungsart, identified just before with ‘intelligible a
R.V. p. 371 [B562, GW 533].
b
G. p. 300 [AK 4: 452, MG 99].
Freedom
71
character’) ‘as its cause, nevertheless does not follow from it at all according to empirical laws, that is, so that the conditions of pure reason, but only so that the effects of pure reason in the appearance of the inner sense, precede’.a In other words that which is to be regarded as the condition, or, as Kant calls it, cause, of the action, in so far as that action exhibits practical freedom, does not precede the action in time. The action is only preceded by the consequence, or, as Kant calls it, effect, of this ‘intelligible’ condition; and hence the action itself may be said to ‘follow from’ the condition as a conclusion follows from premises, but not to follow it in the time-order (cf. above, p. 59). Now, in the case of moral action, this ‘effect’, which produces the action, is just the presentation of the moral law; and the intelligible condition of that effect is the moral law itself. Kant himself allows that this effect or presentation, must always be present in human volition; and, what I wish to maintain, is that this is all that the analysis of human volition as such can ever shew to be present. I have examined the confusions, upon which Kant’s contrary view that the law itself is somehow to be obtained by analysis of volition, that it is given by a ‘pure Will’ or ‘practical Reason’, seems to be based; and those confusions seem sufficient to explain the view and to shew that, for Kant at any rate, it was baseless. It only remains to give a positive summary of the reasons against the legitimacy of any such view; and then to shew what is the real meaning of ‘practical freedom’ and how it is universal. The point at issue is this: Whether ‘will’ can be understood at all as other than a form of ‘activity’; and whether, if it be an activity, it must not be conceived as essentially conditioned by time, and therefore, in Kant’s language a mere ‘appearance’. If it be a mere ‘appearance’, the conception of a ‘pure Will’ is nonsense; and ‘will’ cannot be ascribed as an attribute to anything real – either to God or to the Transcendent Ego. That ‘will’ is a form of ‘activity’ has, I suppose, never been disputed. Kant himself, as we have seen, refers us, for our notion of pure Will, to the pure activity of the Ego. What is disputed is whether psychical activity, at least, may not be considered as fundamentally real. Our contention is that it cannot be so, because it is inconceivable except as taking place in time. That time itself cannot be conceived to be fundamentally real is always admitted by Kant himself, and indeed he has attempted a proof of it. How far his proof is satisfactory, and whether, if unsatisfactory, any other proof is forthcoming, is too large a question to be fully discussed here. I have tried to justify a notion of reality, which excludes time, and to point out that a
R.V. p. 381 [B579, GW 542].
72
The 1897 dissertation
defect in our temporal experience, which really justified Kant in reducing it to mere ‘appearance’. The question which remains, then, is whether we cannot conceive a timeless activity; for it is to such that Kant must be referring us, for justification of the notion of ‘pure Will’. That such a conception is very difficult to maintain appears plainly enough from the attempt of Lotze,a who assumes psychical activity to be the fundamental reality, and then finds himself forced, in consequence, much against his will, to accept the ultimate reality of time.b And we have seen that Kant has nothing valid to say for it. It is a notion which would seem to rest on a combination of the notion of causal dependence between empirical things in time, with that of logical dependence. Both are necessary connections, but in the one case between things, in the other between concepts. That the relation of reality to appearance, or the interrelation of realities, must be conceived as that of logical necessity, changed (in a way which we cannot understand, because we have no intelligible intuition) by the fact that it holds between things, has been maintained above. And if this relation be all that is meant by ‘pure activity’, there seems no objection to the notion; only ‘activity’ seems a misnomer for it, as great as Kant’s ‘causality’ for the same notion, apt only to create confusion. If anything else be meant, it is to be wished it were brought forward; as it would then be possible to discuss it. Meanwhile I must be content to let the matter rest in this result: That, if the logical relation of reason to consequent, regarded as synthetic i.e. holding between real objects,c is to be considered as established by ‘Will’ and the type of its ‘activity’, then pure ‘Will’ may indeed be the fundamental reality; but, I would still protest, that it would be better to keep that word for the distinct notion which it ordinarily conveys, instead of transferring it to another notion, which has long had a sufficiently distinctive name of its own. My conclusion, then, is this: That ‘will’ is only a special form of natural causality, or rather, a natural causal process, where the cause is of one definite sort. It is a special form of natural causality, just as explosion of gunpowder by a match is one special form of natural causality, and explosion of gunpowder by percussion is another. And, that on which I wish to insist, is that voluntary action, of whatever sort, whether autonomous or heteronomous, exhibits ‘freedom’, in the sense which I have hitherto a b c
R. H. Lotze Metaphysics [trans. T. H. Green, ed. B. Bosanquet, 2nd ed. Clarendon Press, Oxford 1887] §156 [§155 actually contains the argument; see pp. 348–50]. See my article on Time: Mind N.S. 22, p. 240 [Moore here argues against the reality of time; see ‘In What Sense, if any, do Past and Future Time exist?’, Mind n.s. 6 (1897) 228–40]. see P. V. p. 52 [AK 5: 49, MG 179].
Freedom
73
explained as essential to Kant’s notion, no more and no less than gunpowder explosions or any other natural process whatever. It seems, indeed, strange that this conclusion from his doctrine should have escaped the notice, both of himself and others, to the extent to which it has. For he repeatedly asserts that for every ‘appearance’ we must suppose an intelligible ground (the Ding an Sich) and it is just this dependence of the cause of his actions on an intelligible ground (the Transcendental Ego), which he describes as constituting man’s practical freedom. Moreover even the identity of the Ding an Sich and the Transcendental Ego has been suggested by him and accepted by others; though this would not be necessary to justify the inference, since the dependence on an intelligible ground is by itself sufficient for practical freedom. When this is acknowledged, ‘practical’ freedom disappears altogether, as something intermediate between natural causality and transcendental freedom. For, as Kant himself says, nothing intermediate is possible; only two sorts of causality can be conceived at all. ‘Freedom’, then, for Kant means only ‘transcendental freedom’, and ‘transcendental freedom’ is not ‘practical’, in the sense that it is inseparably connected with ‘action’ alone. It is true that actions are dependent on ‘transcendental freedom’, but that is only because it is the relation which holds between the empirical causes of those actions and the transcendental ground of such causes; whether sensible objects produced effects, and so indicated their right to be considered practical (as they always must), or not, they would equally be results of ‘transcendental freedom’. kant’s connection of freedom with ‘end’, vitiated by being referred to ‘will’ The degree to which Kant himself was forced to recognise the unpractical nature of his conception of Freedom, is singularly illustrated by a passage in the ‘Metaphysics of Morals’ to which I have referred above (p. 59).a He here declares that ‘Will’, which he has hitherto regarded as identical with ‘pure practical Reason’, and so as that which is alone endowed with ‘Freedom’ in his special sense, cannot be called either ‘free’ or the reverse, because it is not ‘susceptible of compulsion’. This ‘susceptibility of compulsion’ implies subjection to natural law, and, as so subject, he declares that human ‘choice’ (Willk¨uhr) may be called ‘free’. He would seem, therefore here to recognise that ‘action’ can only be conceived as a time-process; indeed he says that ‘Will’ does not refer to ‘actions’ (Handlungen); and a
M. p. 23 [AK 6: 227, MG 380].
74
The 1897 dissertation
it is only because he sees that he would be departing too far from the ordinary use of ‘freedom’, if he disconnected it from action, that he now denies freedom to ‘Will’. The fact is that his previous doctrine has already departed from the ordinary usage, further than he himself was fully aware; and hence the inconsistency, with which he now tries to patch up the discrepancy. The true way of meeting the difficulty would have been, as has been pointed out, to insist on his meaning of Freedom as the true one, and to give up the special connection which he had hitherto asserted between it and human volition: to recognise that ‘Willk¨uhr’ was a mere ‘appearance’, and therefore not ‘free’, and that, that which was free, had not even so much connection with volition, as to deserve the name of ‘Will’. There would, then, have stood out clearly the problem, which must be next discussed – the justification of his use of ‘Freedom’ and of the importance which he assigns to it in relation to Ethics. One justification has already been pointed out – namely that Kant’s view recognises that other element in the causal process, neglected by a Determinism of pure relativity, which I have called ‘form’ and have tried to exhibit as the only basis of ‘Freedom’ for those who take the causal process as the ultimate reality. Kant’s notion includes this element, but it also includes much more, since he sees that finite things, though things and not mere relations, are yet, as finite, partly defined by their relations, and therefore not self-subsistent or capable of being regarded as complete real grounds. But there is, I think, another justification, in so far as ‘Transcendental Freedom’ can be shewn to have an essential connection with Ethics. It is, in Aristotelian language, if Transcendental Freedom can be shewn to supplement the notion of a mere efficient cause, not only by that of a formal, but also of a final cause. Now for Kant ‘Freedom’ undoubtedly has this significance. He deduces the reality of Freedom from the reality of the moral law; and the moral law is what determines the only objective ‘end’ (Zweck). Our reason tells us that we ‘ought’ to do so and so; and it is merely, because we know that we ‘ought’, that we are able to conclude that we ‘can’, i.e. that we are free. ‘If I ought, I can’ does not mean for him, as for Prof Sidgwick (p. 66),47 that only actions, which will follow upon my choice of them, can be reasonably included in any notion of duty. Kant fully recognises, what I have tried to shew above (p. 46), that the fundamental sense of ‘ought’ is that in which it prescribes an ideal, which need not be physically or psychologically possible.a The connection of ‘ought’ with ‘can’, is one a 47
cf. P. V. p. 101 [AK 5: 97, MG 217]. ME 7th edn, p. 67.
Freedom
75
of complete dependence of the latter on the former. What I ought to do, or what ought to be, is ‘morally possible’,a even if, according to natural laws, and so far as experience enables me to predict, it never will happen. This is what Kant means by the ‘Primacy of the Practical Reason’. It is not that ‘Will’, as equivalent to Practical Reason, is to be regarded as the fundamental reality; but that the Practical Reason, as prescribing the moral law, enables us to enlarge our knowledge of reality in a way that the Speculative Reason alone could not do. From the fact that a ‘kingdom of ends’ ‘ought to be’, we can conclude that a ‘kingdom of ends’ is not only possible but necessary, although, if our knowledge were confined to experience alone, we should be bound to conclude that it was neither actual nor possible. Now when Kant takes the moral Imperative, as such, for the primary ‘fact’ in the metaphysics of ethics, he seems to have been influenced by his theory that will is more than a psychological faculty. Will is no doubt the source of the imperative mood as distinguished from the indicative. But, if the above analysis of will has been correct, it can not possibly be the source of the universality, which distinguishes the moral imperative from every other. The will is capable of commanding, as of doing, what is wrong as well as what is right; and so far as it, and therefore the imperative form, is concerned, an injunction to wash your hands is precisely the same thing as an injunction to pursue what is good. The question what ought to be willed, cannot be answered by an analysis of will itself; for such analysis can never tell you more than what is willed, which differs with different times and different persons. Kant rejects the theory that ‘the good’ is the foundation of ethics, precisely because he thinks it can only be directly determined by an appeal to the feelings, which vary subjectively with individuals; and it is the similar subjectivity of the will, which seems to me to condemn his own ‘imperative’. He was perhaps misled by his desire to explain the ambiguity of the term ‘law’ into thinking that the only possible way of distinguishing ‘moral’ from ‘natural’ laws, was the implication of command in the former. What he meant by his conception of a ‘pure’ will, was a will which should command consistency in your view of what ought to be done – a consistency which is necessary to ‘reason’ as such. But all that a ‘pure’ will could really mean, would be, a will which was consistent with itself in the fact that it always commanded, not in respect of that which it commanded. We may admit that the former notion can be obtained by an analysis of will; but it would be by no means sufficient to give a moral a
P. V. p. 61 [AK 5: 58, MG 186].
76
The 1897 dissertation
principle. It is true that will cannot but be consistent with itself, in that it always commands; but what Kant did not see, was that it was a consistency in the object commanded a that he was after all requiring. It is true that this consistency is directly given as no more than a formal condition of good willing; but it is a formal condition of good willing, and not of willing in general. Thus, what really distinguishes the moral law from laws of nature, is that it expresses the connection of natural objects, not with one another, but with the notion of ‘end’ or ‘goodness’. Its primary form is ‘This is good’ or ‘this is an end-in-itself’ or ‘this ought to be’; the command ‘do this’ is no more than a corollary from such a judgment. It is a ‘law’, because such a judgment, if true at all, expresses a universal truth; whereas, subjective maxims, cannot be interpreted into more than a particular statement of fact – ‘I mean to do this’ or at most ‘I think you ought to do it’. To try to obtain this objective validity out of ‘Will’ – ‘this ought to be willed’ out of ‘I or you or all the world, do, or can, or must will this’ – is a procedure similar to that of trying to prove that this is true from the fact that all the world believe it. The assertion of ‘goodness’ claims rationality in the sense explained above (pp. 63–4), precisely as does the assertion of truth: the first has no more connection with volition, than the latter with cognition; both rest on the same ‘Theoretic Reason’ – if we are to adhere to the custom of ascribing them to ‘reason’ at all. If, therefore, we examine Kant’s first expression of the moral law ‘Act only on that maxim whereby thou canst at the same time will that it should become a universal law’, we get the following results. The significance of the expression rests on the words ‘thou canst will’. But there seem no limits to the empirical possibilities of willing. It is possible to will one thing one day, and the opposite the next. It is only, if the question has been begged, by a previous assumption that will cannot be inconsistent with itself, with regard to the objects willed, (in respect of always being will, it must certainly be consistent) that this expression will serve for what Kant means. Kant assumes that I cannot will a thing to be a universal law, to which I might afterwards wish an exception. Perhaps, I cannot rationally; but what is to compel me to will rationally? It is admitted that most of us very seldom do. To set up this as the fundamental principle of morals is as if one were to find the basis of epistemology in the principle ‘Think that only to be true, which you can at the same time think to be universally true’; whereby it would be assumed that no one can think otherwise than rationally. The a
cf. P. V. p. 114 [AK 5: 108–9, MG 227]. He does not see that the mere notion of ‘ought’ is an object or matter of a special sort, though, indeed, not an object of intuition.
Freedom
77
insertion of ‘At the same time’ only makes matters worse, by involving the theory (which has sometimes been held) that the principle of contradiction means no more or less than that two contradictory propositions cannot both be true at the same time. Moreover our epistemological principle would have just as much right to the imperative form, as the moral one. That form is in each case totally irrelevant to the substance of the principle. The reason why Kant uses it is precisely the same fallacy, which he betrays in ‘thou canst will’. He imagines that the objective validity of the principle proceeds from the nature of the will itself, and hence represents it as so proceeding. In fact, however, this reference to will merely conceals the fact that he is dealing with the notion ‘ought’; and the only essential element in the expression is that of universality. Its whole significance can, therefore, be rendered in the proposition ‘You ought to do that only which ought universally to be done.’ Now, if in this way we reject the theory that the principle of Ethics has essential reference to will, we thereby give up the precise connection, which Kant found, between ethics and freedom. Their connection, for him, was that the moral law, necessarily implying will, yet obviously not identical with the law which governs the natural course of human volitions, could have no meaning unless we suppose it to be the guiding principle of another or pure will, which is hence possessed of that ‘intelligible causality’ which he calls freedom. Unless therefore such a free being were theoretically possible, the moral law would actually be contradictory to the course of nature, and we could not say it was reasonable to obey it. The ‘Critique of Pure Reason’, however, shewed it to be possible; and, there being, therefore, nothing against the conclusion to which the moral law would lead us, we can accept its validity unconditionally and along with it the reality of that freedom which could alone make it valid. It is, thus, only because there is a free will, that there is a moral imperative. But if, as I have tried to shew, the principle of ethics is not an imperative, there might seem no need for ‘free will’ or ‘freedom’ to ground it. That principle amounts merely to this, that there is an objective ‘end’, something which ought unconditionally to be, an absolute good. And Kant, though he does express the moral law in terms of these notions, yet thinks them to be merely derivative from that primary imperative, and connected with ‘Freedom’ only through it. There is an unconditional prescription: there must therefore be something unconditionally prescribed, and again an unconditional prescriber. The former is that which is absolutely good, or the sole objective ‘end’, and the latter is pure will. But, says Kant, there is none among the objects of experience which will answer to the former notion, since none is fit to
78
The 1897 dissertation
be a universal object of pursuit. The only objective end must therefore be the only object, which has such universality, and whose reality at the same time is vouched by the moral law itself – namely the prescriber or pure will. Hence pure will and its object are the same: as is expressed in the famous saying with which Kant opens his ‘Foundations for the Metaphysics of Morals’; ‘There is nothing anywhere in the world, nay, nothing at all that is conceivable even outside it, that could be considered good without limitation, except only a good Will’. Thus ‘good’ or ‘end’ is for Kant nothing but the necessary object of a necessary ‘will’, which object can be none other than that ‘will’ itself. But, if that which is free is not ‘will’, there is no necessary object connected with it. I have tried to shew that there is something ‘free’, from theoretic considerations alone. And since it thus has no necessary connection with any specific sort of object, the notion of ‘good’, and Ethics with it, seems to be hopelessly sundered from the doctrine of Transcendental Freedom. Yet if there is any truth in Kant’s ‘Metaphysics of Ethics’, and if that ‘Freedom’ is fully to deserve its name, they must be somehow connected. Only the connection must be sought directly, discarding the irrelevancies which made the transition seem easier to Kant. We have a supersensible reality, which Kant calls ‘free’, and we have a notion ‘good’, which bases Ethics, and the problem is, taking them as they stand, reduced to their simplest terms, to justify their connection. The connection which it is desired to establish is this: That, that which is ultimately real, appearing to us as the reason of whatever happens, is also necessarily ‘good’, and alone absolutely good. This is the same as to say that the complete ‘formal’ cause of the world, that which alone has the theoretically perfect form of self-subsistence, is necessarily also its ‘final’ cause; whence it would follow, that since that formal cause has been vindicated as not only possible but actual, the final cause must also be actual, and we should be entitled to declare that the world is ‘good’. This view may be made plainer to anyone acquainted with the Hegelian dialectic, by the explanation that, according to it, ‘good’ is simply a category – a universal notion and one connected with reality by a logical necessity. notion ‘good’ commonly connected with freedom, and really capable of connection with transcendental freedom Now I have hitherto, in discussing the meaning of freedom, confined myself to conceptions which did not seem to involve the notion of ‘good’
Freedom
79
or ‘end’. But, as a matter of fact, that notion does seem to be involved in the vulgar conception of ‘freedom’ as applied to man. Responsibility would lose its meaning, unless we distinguished objects and actions as good or bad. There would be no reason to think of a man as responsible for what he had done, unless what he had done could be condemned or praised. And when responsibility is referred to freedom for its meaning, it is not only for that part of its meaning, which requires that man, who is responsible, should have been the cause of that for which he is responsible. That the notion of ‘good’ infects that of freedom, too, is, I think, evidenced by the meaning quoted above as given by Locke and Hume to freedom. A man, they say, is free, when he can do what he wills; that is, he must will or choose something, if he is to be considered free: he must not only be the cause of what he so wills or chooses. But this willing or choosing is, strictly speaking, impossible except for a man who has the conception of ‘good’ and ‘bad’. There could be no reason, why a man should choose this or that, unless he could judge the one to be better than the other. But it is characteristic of a voluntary agent that he can give a reason for what he does, and this giving a reason always implies that his action may be regarded by him either as means to something, which he judges to be good, or as itself good. It is, therefore, I think not unfair to say that, in common usage, a responsible agent is conceived not only as uniting the capability of distinguishing good and evil (whether what he thinks good, is really good or not) with freedom, but also as free, only in so far as he makes this distinction. And, this restriction of freedom, to agents capable of conceiving good and evil, already involves the wider proposition that, unless there were such notions as good and evil, there could be no such notion as freedom. The common notion of free action, if completely analysed, involves not only that the action should be self-caused, in the sense explained above, but also that it should be capable of being regarded as good or evil. In fact the reason, why a man feels guilty for having done wrong, is always that he believes something else ought to have been done, even though he has reason to think that owing to his character and habits be could not have done it. Kant has, therefore, a further justification for calling his Transcendental Freedom by that name, if it can be shewn to have a necessary connection with good and evil; since common sense seems to recognise some necessary connection between freedom on the one hand and these ethical notions on the other. But the importance of freedom for Ethics will depend upon what the connection is. If it were only true that, for an action to be considered free, it must be good or evil; these latter notions would thereby have an
80
The 1897 dissertation
independence given to them: we should not need to know what freedom was in order to know what was good and evil, but only to know these, if we wished to know what freedom was. On the other hand, common sense does seem to recognise the converse proposition, that an action must be free, if it is to be good or evil in a special sense, the sense of ‘moral’ good or evil; a distinction corresponding to that between Prof Sidgwick’s ‘ethical’ and wider ‘ought’, and to the Stoic distinction between ‘kaqkonta’ and ‘difora’.48 So that, in this case, we should have to know whether an action was free, before we could tell whether it was good or evil; whereas, according to the former connection, we have to know whether it is good or evil before we know whether it is free. We thus seem to reach that dilemma to which Prof Sidgwick’s statement of the free will problem was formerly reduced (see p. 47). Our escape was effected in that case by rejecting the doctrine that the good or evil of an action depended upon our freedom to do it or to leave it undone. And this, I think, is true, if freedom be taken (as there it was) only to imply empirical possibility – possibility according to natural laws. But when we have to do with Transcendental Freedom, this way of escape is closed, since nothing can be at all, unless it depend on Transcendental Freedom. Hence in this case it must be true that ‘only free actions can be good or bad’. But, though we thus have freedom and Ethics connected in the form, ‘Only free actions can be good or bad’, this is not yet sufficient to make the knowledge of what actions are free of primary importance for Ethics; since that science must find some way of distinguishing between what is good and what is bad. This lack would be supplemented by the further formula, with which it was proposed to replace Kant’s theory of the dependence of the moral law upon pure will – the formula that the source of freedom is necessarily good. But with this, the source of evil seems to become, as Kant himself calls it, incomprehensible.a For that which is nothing but good does not seem to supply a reason for the existence of evil. We have here the result that since the intelligible character is necessarily good, and also the sole reason of everything that happens, everything that happens is necessarily good. And this seems to conflict with the actual fact
a 48
M. p. 24 [AK 6: 226, MG 381]. Moore alludes here to the Stoic distinction between acts which manifest man’s ‘proper function’ (kaqkonta) and ‘indifferent’ states of man (difora), though this seems a different distinction from Sidgwick’s distinction between the ‘ethical’ and the wider ‘ought’. For discussion of Stoic ethics, see A. Long and D. Sedley The Hellenistic Philosophers, vol. I, Cambridge University Press, 1987, esp. pp. 354–68.
Freedom
81
that some things are evil. They can be evil only in so far as Freedom is their reason; and yet freedom is unable to supply a reason for evil. This is, I think, what more properly deserves the name of the ‘Antinomy of Practical Reason’, than that which Kant has discussed under that head. And it is noteworthy that its solution is effected by the same method; for the difficulty arises with respect to the object which can be said to be good or evil. Kant recognises two sorts of good. That which is a cause of pleasure is, he thinks, a good in itself, and that which is a cause of pain, an evil. But these causes can only be objects of experience, belonging to the world of sense; and hence perfect virtue, which is moral good, can never guarantee the existence of perfect happiness, since the existence of that does not, by natural laws, depend upon it. Kant solves this Antinomy simply by saying that though virtue, as a natural cause, does not necessarily produce happiness, yet as the property of an intelligible being, such as the moral law shews that we are, it may do so.a But, according to what has been said above, virtue is not the property of an intelligible being; nor, on the other hand, does Kant give any reason for thinking that happiness is necessarily a part of the complete good. What seems to have led him to this view is his perception that the objects of experience must in some way be capable of representation as good; and the false assumption, consequent on his restriction of the application of freedom to a peculiar type of human volition, that they could only be so represented in so far as they affected the feelings and were subjective ends. According, however, to the doctrine of Freedom explained above, by which the free Cause is represented as a universal reason of things, coupled with the doctrine, which he himself admits, that the Free Cause is necessarily good, all the objects of experience are good, as dependent upon it, and there is no need for the reference to feeling. The Antinomy then appears in the fact that the objects of experience are not all good, and that the Free Cause, which is the reason of their goodness, is also necessarily responsible for their badness. The solution lies in the fact that the objects of experience are not Things-in-Themselves; and hence that, though the Thing-in-Itself, i.e. the Free Cause, cannot be both good and bad, yet when the character of goodness, which properly applies to it only, is transferred, as it must be, to those things of which it is the reason, it is found equally necessary both to affirm and to deny it of them. Evil thus appears to be absolute, only when appearances are taken to be real. And just as it must be asserted that appearances are real, if they are to be appearances, but on the other hand that they must be unreal, else a
P. V. p.121 [AK 5: 114–15, MG 232].
82
The 1897 dissertation
they would not be appearances; so evil only is, in so far as it is good, and yet it cannot be good, else it would not be evil. This theory, which, whether true or not, is, I think, a fair deduction from Kant’s connection of Freedom with Ethics, accounts for the distinction between moral and natural good and evil in the following way. The relation between them is that natural good and evil though actually dependent on Freedom, are presented as such directly, and in so far as they are good, enable us to conclude their free causation; but that the knowledge of good and evil depends on the previous knowledge of their free causation. It is in so far as some of the objects of experience (particularly mankind) appear to approximate more to the conception of a complete formal cause, i.e. to reality as such, particularly in respect of their inclusiveness,a that they can be said, according as they are judged good and bad, to be morally good or bad. Thus, though it is not true, as Kant says, that a man’s ‘good will’ can be called unconditionally good, yet it is true that, in virtue of its formal completeness, it may be called the best of things, and similarly a bad will, the worst. This is because they unite in themselves more of natural goods, than is otherwise possible, to make of them respectively a good or a bad whole. Or to put the relation in terms of ‘end’, more of the subordinate objective ends contribute in the good will, as means to a result which is also an end-in-itself, and in the bad will are emptied of the goodness which they possess in their own right, by being united into a whole which is the most complete negation of an end-in-itself. But this statement, it must be noted, is only possible when the notion of end has already been determined as that which is naturally good. (see above p. 76) The result of this theory for Ethics is mainly negative, since it prevents the assumption that any one natural datum, such as pleasure, is to be considered as sole end-in-itself, and other things as good merely as means to it. Everything, according to it, must be regarded as an end-in-itself, and the superiority of one over another consists only in the degree in which it combines such natural goods into a whole, that can itself be judged a good. The doctrine of Freedom, with its definition of reality, thus helps out ethical judgments by the assurance that formal completeness is necessary to goodness; but at the same time the Primacy of Practical Reason is still maintained in a sense, inasmuch as we require to make a direct judgment in any particular case, whether the combination in question and the things a
That such a criterion of reality is capable of precise application, I do not mean to imply. It seems to me that the direct judgment that one thing is better than another, is far more certain. I am here only offering an explanation of that appearance of difference in kind, e.g. between man and animals, which is really due to the lack of continuity in degree, which characterises our actual experience.
Freedom
83
combined are good or bad. We cannot, as Spinoza proposed, take the reality of a thing as a direct guide whether a thing is good or bad; since it is only the completely real which must be wholly good, and nothing in experience can be completely real. But we may use this criterion in distinguishing the degree of goodness of two different goods. Thus though the fundamental proposition, in such a Metaphysics of Ethics, is ‘Only the free is good’, it is yet true that this must be supplemented in considering any natural object which make a claim to goodness by the proposition that ‘Only the good is free’; and goodness, though pkeina ts oÉs©as,49 with regard to any natural object, since the best, we know, is therefore, the most completely real, is inferior to reality, with regard to the absolutely real, since the goodness of that proceeds only from its reality. Such a relation between the concepts ‘good’ and ‘free’, a relation in which each may be necessarily inferred from the other, although they are not identical, is only possible with truly universal or a priori concepts, such as the Hegelian categories. Of these, wherever one is asserted, all the others must be asserted too; whereas, with empirical concepts, in proportion as the extension increases, the intension seems to decrease. But, while each of the categories embraces the whole extension of Appearance, they have a difference of intension in respect of the degree, in which they express Reality. They all characterise Reality, but only in relation to Appearance, and hence though there is no difference between them in respect of analytic priority, in respect of synthetic priority there is; since, though the concept of Reality, as such, is only analytically related to itself, and the concept of Appearance, as such, likewise, Appearance may partake more or less of Reality, and thus gives rise to the differences in the categories by which its relation to Reality is expressed. It is thus that evil, though, like good, it is only possible through the union of Reality with Appearance, yet expresses less, than good, of the nature of Reality and more of mere Appearance. The object which is said to be good and that which is said to be real, are identical, if that object is taken to be Reality, as such, and it is not bad, only because ‘good’ is more adequate to that object than bad; whereas the objects denoted by good and bad and real are all identical, if those objects be treated as mere Appearance. But in fact good, like mere ‘being’, is no more than an abstract notion, and hence, though always identical with itself and different from every other predicate, the actual objects denoted by it admit of difference in degree and also of the simultaneous predication of evil. 49
See note 18.
84
The 1897 dissertation
Thus the only ‘end’, properly so-called, is what is good, whether it be actually pursued by any will or not. It is so because it is the only thing which ought to be pursued; and this is the only meaning which can be given to ‘objective ends’. ‘Subjective ends’, objects actually pursued, such as pleasure, are not ‘ends’ at all, merely in so far as they are pursued, but only if they also happen to be good, which is quite contingent from the point of view of will. ‘Willk¨uhr’, as such, is mere ‘caprice’. The goodness of Kant’s ‘good will’ does not consist in the mere will to do good; for that may be accompanied by want of insight into what is really good. It is, no doubt, a necessary part of a completely good will, and, perhaps, the most important part, when that will is to be judged as a good in itself; but nevertheless it is only a part: for completeness, the ‘moral disposition’ to adopt the objective as your subjective end, must be accompanied by a knowledge of what the objective end is. The connection between freedom and goodness is not that ‘to choose the good for its own sake’ is alone truly good; but that to choose the good for its own sake is the best form of choice, and that choice, by being a very good thing in itself, is shewn to be ‘free’. But it has not yet been shewn that what is real or free must be good, and that what is absolutely good must be real; and this, it may as well be confessed at once, I do not see any way of completely proving. There should, no doubt, be some chain of argument similar to that of Hegel’s Dialectic; but, in the absence of that, I have only to offer the following considerations, which seem to render the view, though not certain, probable. The question may be put in various forms. The form in which Kant gives it an affirmative answer, is this: Can we infer, from the fact that we ought, that we can? And with regard to this form I have tried to shew that ‘ought’ cannot be limited to what we regard as possible according to natural laws (cf. p. 46). The question therefore means: From the fact that we ought to be perfect, may we infer that we can be perfect? And with regard to this I have tried to shew (Introd. pp. 9–12) that the conception of perfection is one that must have meaning, apart from any description we may give of what is perfect. ‘Can’ too, since it does not apply to empirical possibility, must mean: Is there actually an object, which is a sufficient reason of perfection, so that from its reality the reality of perfection can be inferred? The question therefore becomes identical with that which Descartes answered in the affirmative, when he said that, because we have the idea of perfection, there must be a perfect being to give it us. Or the question may be put in a form more like that, in which Leibniz, Plato and Aristotle answered ‘Yes’ to it: Is the reason why anything is at all, the same
Freedom
85
as the only reason we can give for doing a thing – that it is good? Or, finally, from the fact that what is good is good, can we infer that what is good, is? because good is an objective notion, must it have an object? Now when we say that ‘good is good’ we certainly do not mean that our idea of good is our idea of good. And yet that proposition has some meaning and must give us some information about the world. This meaning cannot be that the notion of goodness exists in our minds. That is not the only form in which goodness is necessarily connected with reality. For we are able to assert of it that it is itself, and that it is something different from everything else. And this assertion is not equivalent to such a one as ‘A chimaera is a chimaera’. For though a chimaera also does not mean our idea of a chimaera, it is not something which we conceive it necessary to predicate of the world. But this ‘good’ is. Just as we cannot conceive of a world except as ‘being’, we cannot I think conceive of a world of which it would not be possible to assert that it was either good or bad. If we actually said it was indifferent; the case would remain the same for we should still be regarding it under what I may call the ‘moral category’. But, with regard to a chimaera, we can perfectly well conceive a world, in which there should be no such creature; a world, too, of which a complete description might be given, without a necessity of stating that it contained no chimaeras. A world, indeed, must either contain a chimaera, or not contain one; but with regard to good, bad, and indifferent, there is in them some common, positive, unique notion which must be predicated of any world. By denying good you imply bad or indifferent, but by denying chimaera you imply nothing but mere existence. The difference consists in this that a chimaera is made up of actually presented empirical qualities, i.e. is merely imagined; whereas the ethical notions are neither made up of any such, nor are themselves any one such simple datum. They have a positive content which neither is a sense-datum, nor presupposes, for its definition, any actual sense-datum. In fact, it seems to me certain that we cannot apply to them the criticism, which Hegel makesa,50 upon Kant’s use of a hundred thalers, to refute the ontological proof. These too are ‘objects of another kind than any hundred sovereigns, and unlike any one particular notion, representation, or however else it may be styled’. a 50
Logic, Wallace’s trans. 2nd edn. pp. 107–8. Hegel argues that Kant’s objection to the ontological argument (B627, GW 567) fails to take proper account of the difference between the notion of a hundred thalers and that of God, which expressly involves God’s actual being; Moore suggests that the moral categories, goodness, badness and indifference should be understood in a way which is similar to this way of thinking about God.
86
The 1897 dissertation
But though it is true that the world must have the ‘moral category’ predicated of it, it is a harder thing to shew that this category is ‘goodness’ rather than badness or indifference. However, from the formal notion which we got of the world as ‘Free Cause’, we can infer that it must be one of these three alternatives, not all three at once, as are objects in the world of experience. Indifference, too, may, I think, be excluded by this test, as only possible through that mixture of good and bad, inseparable from the finitude of the objects which constitute Appearance. We have therefore the choice left between its being good or bad, and the only means of deciding between the two that I can suggest, is that good seems to have the preeminence as a more positive notion than the other. Even if reality were wholly evil, the only reason for choosing one thing rather than another, would still be that it was better, not that it was worse; and the problem is whether this latter truth cannot be shewn to conflict, with the hypothesis of the reality of evil; whether the Reason which tells us that only the good is rational, does not, after all, mean ‘rational’ in the same sense as when it tells us that only the rational is real.
Appendix: Professor Sidgwick’s Hedonism
The following criticism only touches directly a very small part of the ‘Methods of Ethics’. But the points discussed seem to be of fundamental importance for any Hedonistic Ethics. I
Professor Sidgwick says: ‘I am finally led to the conclusion . . . that the Intuitional method rigorously applied yields as its final result the doctrine of pure Universalistic Hedonism – which it is convenient to denote by the single word, Utilitarianism’a . In this statement I find the cardinal point of dispute. But it is necessary to distinguish. In so far as the Intuitional method is held to establish a Universalistic doctrine, I fully agree; in fact, I should go further than Professor Sidgwick, in as much as I cannot see that Egoistic Hedonism has any claim to an independent validity. I disagree only in so far as it is a Hedonism that the Intuitional Method is said to yield as its result. The argument by which Professor Sidgwick attempts to support his conclusion on this point has two main steps, in neither of which can I see any necessity. (1) He has already in Book i, chap. ix, ‘confidently laid down’, ‘that, if there be any good other than Happiness to be sought by man, as an ultimate practical end, it can only be the Goodness, Perfection, or Excellence of Human Existence’.b Those of his reasons for this view, which I dispute, are wholly contained in the short fourth section of this chapter. ‘I think’, he says, ‘that if we consider carefully such permanent results as are commonly judged to be good, other than qualities of human beings, we can find nothing that, on reflection, appears to possess this quality of goodness out of relation to human existence, or at least to some consciousness or feeling’.c This sentence, in itself, is not sufficiently definite for me to take exception to it; for I do not know of anything whatever, which is strictly ‘out of relation to human existence’, and which, therefore, could have the quality of goodness out of that relation. Yet the sentence is obviously meant a
ME pp. 406–7.
b
ME p. 115.
c
ME p. 113.
87
88
The 1897 dissertation
to exclude certain objects, considered in a certain way, from the claim to possess goodness; for which purpose it is necessary to define the precise relation to consciousness or feeling, on which a valid claim of goodness is held to depend. This relation is partially limited in the note. ‘For practical purposes, we require to conceive some parts of the universe as at least less good than they might be. And we do not seem to have any ground for drawing such a distinction between different portions of the non-sentient universe, considered in themselves and out of relation to conscious or sentient beings’.a We are here, I suppose, entitled to infer that ‘portions of the non-sentient universe’ can be ‘considered in themselves’, and therefore that the relation to consciousness which gives them a claim to goodness, is not of that universal kind, in which I should hold that everything whatever must be considered as related to the cognitive consciousness. Understood in this sense, then, which, however, I must admit, is still deplorably vague, I can only say of Professor Sidgwick’s proposition, that, after the most careful consideration and reflection of which I am capable, it seems to me to be untrue. I cannot avoid thinking that we have as much ground for distinguishing degrees of goodness in different portions of the non-sentient universe, considered in themselves, as when they are considered in relation to sentient beings. And I do not think my objection is in any degree based on a quibble, as might be suspected from the vagueness of the expression which I am trying to combat. To the best of my belief, I am denying the proposition in precisely the same sense, in which Professor Sidgwick means to affirm it. To avoid all mistake, let me take the example, given on p. 114. ‘No one’ it is stated ‘would consider it rational to aim at the production of beauty in external nature, apart from any possible contemplation of it by human beings.’b Now I must confess I do consider this rational. Even if it were absolutely impossible that any part of nature should ever be an object of contemplation either to myself or to anyone else, I should think it reasonable to prefer that it should be beautiful rather than that it should be ugly. And I think this statement is not nonsensical, in any sense in which it is not nonsensical to talk of things being ‘considered in themselves’ at all. For, though I might accept Professor Sidgwick’s statement that ‘when beauty is maintained to be objective, it is not commonly meant that it exists as beauty out of relation to any mind whatsoever’;c nay, though I hold myself, that beauty cannot exist as beauty out of relation to mind, I only hold it in the sense in which I hold that nature cannot exist out of relation to mind; and, since Professor Sidgwick appears to think that there is a sense in which nature may be considered to exist out of relation to a
ME p. 113 fn. 2.
b
ME p. 114.
c
ME p. 114.
Professor Sidgwick’s Hedonism
89
mind, as to which I perfectly agree with him, and whereas he seems to give no reason why beauty should not, in the same sense, be considered to exist, as beauty, out of relation to mind, I can content myself with saying that I for my part see no reason why it should not. The question is, I believe, whether a thing can be considered as good or beautiful, except as a means to a possible contemplation of it; this relation of means to end, seems to be the one in which Professor Sidgwick conceives it must stand to sentience if it is to have a claim to the quality of goodness. On the issue, thus defined, my most careful reflection only leads me to the judgments: (1) That a thing may be beautiful, whether it is possible any one should contemplate it or not (2) That, in so far as it is beautiful, it is better than that which is uglier, quite without regard to its possible effect on an observer. I have not given any strict argument on these points, because Professor Sidgwick gives none. He seems to ‘lay down’ his propositions as matter of intuitive judgment; and though I should not like to assert, for my part, that such cannot be disproved, yet, in face of a carefully considered judgment to the contrary, the onus probandi seems at least to rest as much on one side as on the other. (2) The case is the same with regard to the second important step, by which Professor Sidgwick proceeds from the conclusion ‘that Ultimate Good can only be conceived as Desirable Consciousness’a to the conclusion that ‘when . . . we “sit down in a cool hour”, we can only justify to ourselves the importance that we attach to any of ’ certain ‘objective relations of the conscious subject’, ‘by considering its conduciveness, in one way or another, to the happiness of sentient beings’.b In this case, Professor Sidgwick, explicitly appeals to his reader’s ‘intuitive judgment after due consideration of the question when fairly placed before it’c . Now, I have fulfilled these conditions to the best of my ability; but yet I have not been led to this conclusion. On the contrary. I think that ‘these objective relations of the conscious subject, when distinguished from the consciousness accompanying and resulting from them,’ are ‘ultimately and intrinsically desirable’; just as much as ‘material or other objects are’ even ‘when considered apart from any relation to conscious existence’.d To take an example: I do consider that it would be good and desirable to know the truth, whether or not any pleasure either to myself or anybody else were to accompany, or even could result from that knowledge; nay I think that knowledge of the truth, so far from being good merely as a means to happiness, however great, is in itself better than happiness, understood as a balance of pleasure. I therefore think myself, on the ground of intuitive judgment, as fully a d
b ME p. 401. c ME p. 400. ME p. 398. ME p. 401 (7th edn, pp. 400–1) [for Sidgwick ‘these objective relations of the conscious subject (etc.) . . . are not ultimately and intrinsically desirable’].
90
The 1897 dissertation
justified in rejecting Professor Sidgwick’s Hedonism, as he can be in accepting it. With regard to Professor Sidgwick’s appeal to a ‘comprehensive comparison of the ordinary judgments of mankind’a in support of his contention, he admits that it ‘cannot be made completely cogent’; and I shall, later on, have several points to make against what he urges in its favour. I think he would agree with me that the main issue of Universalistic Hedonism, quˆa Hedonism, lies in the direct answer to the question ‘Can Pleasure alone be considered as good for its own sake ?’, – a question which he seems to consider as answerable in the end solely by intuition; and to this question, as matter of intuition, I feel myself compelled to give a negative answer. But there is one other consideration, appearing in the form of an argument, which I think sufficiently important to deserve notice. Professor Sidgwick asksb : ‘If we are not to systematize human activities by taking Universal Happiness as their common end, on what other principles are we to systematize them ?’ And, lower down, he says ‘I have failed to find – and am unable to construct – any systematic answer to this question that appears to me deserving of serious consideration: and hence c I am finally led’ – to the conclusion, which was quoted at the beginning of this paragraph. Now, this language would seem to suggest that the absence of any other system was in itself a reason for adopting that of Utilitarianism. But this is only so, if it be assumed that some system must be possible. But are there any grounds for this assumption? It seems to me important to make a distinction. I should have thought that the mere reference of human activities to the notion of ‘goodness’ (a notion which Professor Sidgwick admits to have a unique meaning) as their proper end, was deserving of the title ‘system’; since we are agreed that what is once truly judged to be good, cannot, in so far as it is really the same, be ever truly judged not to be so; that is to say, the objectivity of ethical judgments introduces some system into them. And this, I think, is enough to satisfy the demands of the Intuitional method; and is certainly important, for it is sometimes denied. However, I quite admit that Ethics would, in our present state, be made more systematic, if there were some universal criterion of the proper end of human activities, such as ‘happiness’ supplies. Only I see no reason, except wishes, for supposing that such a criterion is even possible. Moreover I do not know that it is even to be wished. The science of counting seems to have been very well established, without any criterion of what objects are two and what are four. It is even, just because there is no universal empirical criterion of numbers, that the science seems capable of such a
ME p. 401 [7th edn p. 400].
b
ME p. 406.
c
My italics.
Professor Sidgwick’s Hedonism
91
complete system as it has attained. There is no system so complete as that of a priori notions. Though, therefore, I must admit that at present, we cannot systematise experience in relation to this notion,51 so well as if we took Happiness as its criterion; yet I see no reason to despair that we may ultimately be able to apply it with as much certainty as number, and to be as confident in our measurements of degrees of goodness as we now are in our measurement of degrees of quantity. Meanwhile the adoption of Pleasure as our standard, merely for the sake of having some definite system, would, if it happened not to be a true one (and we have seen no reason to think that it is), lead us only into systematic error. We should only be the more certain to be wrong, the more certainly we could determine the relations of actions to Greatest Happiness: we should thus be in a worse state, than if we trusted to the judgments of common sense, systematised only in relation to the undefined notion of goodness. I do not, indeed, think that the danger from Utilitarianism is very great, partly because I accept the common sense judgment that pleasure is, in itself, better than pain, and to be sought so far as it does not hinder the pursuit of still better ends; but chiefly because the application of the hedonistic calculus seems, in all the most important matters, to yield whatever conclusion would have been maintained without it, owing to the uncertainty both of measurements of pleasure and of calculation of results. But I shall presently give reasons for thinking that, in proportion as it is made more capable of just application, it will give results more and more divergent from those of Common Sense and those which I myself intuitively judge to be true. II
The second topic which I think it important to discuss is Professor Sidgwick’s definition of Pleasure. ‘I propose’, he says ‘to define it as feeling which, when experienced by intelligent beings, is at least implicitly apprehended as desirable or – in cases of comparison – preferable’.a Now, in the first place, it is remarkable that Professor Sidgwick seems here to have fallen into the very same confusion, with regard to ‘preferable’, against which he has himself so carefully warned us with regard to ‘desirable’. We should certainly expect that just as ‘desirable’ means not ‘desired’, but what ‘ought to be desired’,b so ‘preferable’ would mean, not what is, but what ought to be preferred. But Professor Sidgwick goes on to defend himself against a supposed self-contradiction, as if, after all, a
ME p. 128 [7th edn p. 127)
51
I.e. good. The previous sentence was originally: ‘There is no system so complete as that of a priori notions; and it has been attempted to shew that good is such a notion.’ Moore has then deleted the second part of this sentence.
b
ME pp. 110–13.
92
The 1897 dissertation
‘apprehended as preferable’ meant the same thing as ‘preferred’: indeed in order to substantiate his charge against himself, he actually converts the proposition that ‘preference of pleasures on grounds of quality as opposed to quantity . . . is actually possible’, into the proposition that ‘the less pleasant feeling can . . . be thought preferable to the more pleasant’.a I shall assume, however, that this does not express his meaning in the definition, but that he does really intend us to understand by ‘preferable’, ‘what ought to be preferred’ or ‘what it is reasonable to prefer’, not ‘what it is actually possible to prefer’. If he does not mean this, the objections to his definition would be still more serious than those which I am about to make. My first objection is that, on Professor Sidgwick’s own showing, he is giving us a definition not of ‘Pleasure’ but of ‘a pleasure’. He has just said that the only common quality he can find ‘in the feelings so designated’ (i.e. by the title ‘pleasure’) ‘seems to be that expressed by the general term “good” or “desirable”’. But surely ‘feelings’ cannot be ‘Pleasure’, but only ‘pleasures’; and thus ‘Pleasure’ would be their common quality, i.e. that which is apprehended as desirable in feeling, not ‘feeling apprehended as desirable’. And this conclusion that ‘Pleasure’ denotes a ‘quality’ of feeling, not feeling itself, seems supported by a later passage where we are told that, when Professor Sidgwick contemplates any feeling ‘merely as the transient feeling of a single subject; it seems to’ him ‘impossible to find in it any other preferable quality than that which we call its pleasantness’ (p. 129).b It is here, therefore, declared expressly that it is a quality of feeling, called pleasantness, that is recognised as ‘preferable’. Now a pleasant feeling, i.e. a feeling which has the quality of ‘pleasantness’, may be ‘a pleasure’, but it is certainly not ‘Pleasure’. I think, therefore, that the definition must be corrected into the form: ‘Pleasure is that quality of feeling, which . . . is . . . apprehended as desirable or . . . preferable’. To this, my first objection, is that other qualities of feeling, e.g. ‘sweet’ (to use Professor Sidgwick’s own examplec ) may be directly apprehended as desirable. I think I agree with Professor Sidgwick’s definition of ‘desirable’ or ‘good’; but, as was explained above, I think it intuitively certain, that it can be applied not only to other elements of consciousness, beside one specific quality of feeling, but also to material objects. I must, therefore, give another meaning to Pleasure; and my view is just that which Professor Sidgwick rejects, on the evidence of ‘reflection’, namely that it is a directly cognisable quality of feeling, ‘strictly undefinable from its simplicity’.d I am aware that a psychologist of such high repute a c
b ME p. 129 [7th edn p. 128]. ME p. 128. d ME p. 127. ME p. 127–8 [7th edn p. 127].
Professor Sidgwick’s Hedonism
93
as Dr. Ward, refuses to allow that it is directly cognisable. But he would allow, at all events, that degrees of pleasure are measurable by other tests than that of degrees of desirability; and this is all that I am concerned to maintain against Professor Sidgwick. My intuitive judgment is, in fact, that I apprehend some feelings as pleasanter than others, and yet that I sometimes apprehend those very same feelings as less desirable, and that, considered in themselves, entirely without regard either to consequences or concomitants. Here again, then, we have reached the cul de sac of contrary intuitions. My intuition will, therefore, justify me in condemning this definition as very ‘vague and indefinite’; not because I do not think ‘desirable’ adequately defined, but because ‘desirable feelings’ seem to me to cover a much wider field, than what I recognise as pleasure. It seems to me, also, that it is just because of this vagueness in his notion of pleasure, that Professor Sidgwick is able to exhibit Utilitarian judgments as so little divergent from those of Common Sense. If my view as to the true application of ‘Pleasure’ and ‘desirable’ is correct, he would be in danger of supposing that he had judged to be ‘pleasanter’, what, in my view, he had only judged to be ‘preferable’; and this would furnish a reason why he finds the ‘most refined and subtle intellectual gratifications’ to compare so favourably, in point of pleasurable intensity, with the ‘coarser and more definite sensual enjoyments’.a I cannot help thinking that Common Sense is more in the right, when the term Pleasure ‘suggests’ to it ‘pre-eminently the coarser and commoner kinds of feelings’.b Lastly, this definition of Pleasure, as feeling merely apprehended to be desirable seems to come into serious conflict with the principle of ethical Hedonism, that Pleasure is desirable per se. To say that which is merely judged to be desirable, is desirable, nay the only desirable thing, seems, on the face of it, absurd; for it will hardly be maintained that an erroneous judgment is impossible. But there is another argument to be drawn from the subjectivity of this definition which would seem to touch Utilitarianism more nearly; since it affects its claim to give systematic guidance. A man may judge that one kind of feeling is preferable to another, because it is so to him: but it can hardly be denied that different persons find different feelings pleasanter; and whether your most intense pleasure or my most intense pleasure is the more intense, and therefore whether it is more desirable I should train my child, or should advocate public measures tending to train my fellow-citizens, to take after me or to take after you, in respect of the feelings from which they are to expect most pleasure – this, with Professor Sidgwick’s definition, there is absolutely no a
ME p. 128 [7th edn p. 127].
b
ME p. 402.
94
The 1897 dissertation
means of deciding. Universalistic Hedonism may, therefore, be a guide so far as to tell us to give men the most possible opportunities of enjoying, what they actually do enjoy most; but what it is desirable that they should enjoy most, because it is most enjoyable, this definition of Pleasure has debarred it from telling us. It offers no warrant that, by carefully following it, we should not end in Plato’s Ëwn poliv, because the swinish pleasures there enjoyed happened to be those which were most intense for the greater number, and we could not know they were not the most intense possible. It is surely not impossible that such an ideal, by the adoption of Utilitarianism, might become gradually practicable; but it is not a little divergent from that of Common Sense.52 52
In the first manuscript Moore concludes his criticisms of Sidgwick with the following passage, which is unfortunately very messy in parts; he omitted it from the clean manuscript which provided the source for the dissertation itself. iii I think, however, that Universalistic Hedonism may be shewn to be in very strong opposition to Common sense, even with regard to the judgment of a particular action in the present state of society without taking into account the direction of general social development. Take the instance of a man and a boy capsized in the sea, when the probability that they will be saved is very small. The man is much stronger than the boy, and might, if he chose, slightly increase the chance that the boy would be saved, by holding him up and thus decreasing his own chance. He absolutely refuses to help the boy, in order to keep his own chance as great as possible. Both eventually are drowned. Common Sense would, I think, judge that the man had behaved just as wrongly, as if he had refused to help the boy, in a case where there was a very strong probability that, with that help, both might escape, where his refusal was witnessed, and where he himself escaped easily. Yet could the Utilitarian calculus justify this judgment? The infelicific effects which it might take into account seem to be chiefly these. (1) The loss of the boy’s life. (2) The general good effect of example, if the action became known. (3) The effect on the man’s character, in weakening the disposition to act morally, and, perhaps, in making his general influence, on his associates, less conducive to good. Now if there was even a possibility that both might be saved, such a mere possibility of preventing these bad effects, and of producing the corresponding good ones, would, I admit, be sufficient to justify a Utilitarian in condemning the man’s action as wrong. But suppose, if the wrongness is to be measured it would necessarily be conceived to vary in proportion to the probability of these consequences ensuing; and where, as in the case supposed, the probability of being saved is practically nothing, the man’s guilt would be judged very slight indeed. Common Sense, however, if his motive was merely to give himself the best chance, however poor a one, would seem to reprobate him utterly. In conclusion, then, the principle of Universalistic Hedonism, as defined by Professor Sidgwick, seems to me to lack three at least out of the characteristics which he himself declares to be requisite for an Axiom: and yet I suppose it is meant to be axiomatic, since it does not pretend to have a more rational basis than intuition. (1) It is not ‘stated in precise terms’, since the definition of Pleasure, which it involves, is vague, and vague too . . . . . . ., as was maintained in (ii). (2) It is not really self-evident, since it can be disputed, as was maintained in (i). (3) It is not ‘supported by an adequate “consensus of experts”’, since, as was maintained in (ii), it does not seem in harmony with Common Sense, nor has it been by any means generally held by Ethical philosophers. It would also seem to lack the fourth necessary characteristic for an axiom, by conflicting with another truth, if the necessary universality of ‘good’ could be considered as fully proved.
Examiners’ reports on the 1897 dissertation
H. S. on G. E. Moore, Dissertation on the Metaphysical Basis of Ethics Henry Sidgwick
Mr. Moore’s Dissertation is on the ‘Metaphysical Basis of Ethics’. It only covers, as he says, a small part of the ground which he had marked out for it. It’s [sic] main part (97 pages) consists of ‘a chapter on Freedom with special reference to Kant’; but about 1/5 of it, including 10 pages of this chapter, consists of criticism, on my views. I propose to confine my remarks to that part of the Dissertation in which Mr. Moore criticizes Kant or expounds his own views as I am able to judge this more impartially: but I may say generally that the part relating to me has certainly the same merits and perhaps the same defects as the rest of his initial work. The ability shown in the dissertation seems to me of a high order; quite first-rate, in respect of critical acumen and dialectical vigour, and independence of thought; and at least very promising in respect of philosophical grasp and penetration. Mr. Moore, however, is in my opinion more successful in making points against particular statements of Kant, than in understanding his views as a whole, where that understanding is rendered difficult and complicated by [the] discrepancy between the old ideas that Kant retains and the new ones that he is introducing. At the same time he has a thorough grasp of Kant’s main conception and his criticism is almost always close and penetrating and of value in the way of suggestion even when he seems to miss Kant’s real view. The criticism of the special notion of Freedom, fundamentally important in Kant’s system appears to me of less value than it might have been, because his examination of the passages in Kant’s sections in which this notion is expounded and applied, is somewhat defective in completeness and in method. That is, he does not seem to me to have carefully considered all the passages, nor to have adequately borne in mind – though he has not Sidgwick’s report is in the Wren Library at Trinity College, catalogued as Add. MS A 247 3. It comprises two large foolscap sheets in Sidgwick’s own autograph.
97
98
1897 examiners’ reports
entirely overlooked – the changes of view that are likely to be found in works written with different aims during a period of 16 years. His exposition of his own views is rather difficult to judge, as in the earlier part of his dissertation he does not quite sufficiently distinguish these from the views he attributes to Kant, and in the closing portion where his exposition becomes more independent it becomes obscure from brevity – owing, I gather, to pressure of time. He impresses me as having a strong, independent, and definite grasp of the fundamental problems of philosophy: and if his conclusions appear to me somewhat crudely paradoxical, this is at any rate evidence of fearlessness in following out his trains of thought to the logical results. On the whole, the merit of this part of his work seems to me to lie in promise rather than performance: but I judge it to be very promising. P. S. Mr Moore also sends in a short paper in Mind on the question ‘In what sense, if any, do Past and Future Time exist?’ This paper comes third in a discussion and has therefore naturally a preponderantly critical character: it shows, like the Dissertation critical acumen and a vigorous grasp of metaphysical reasoning.
Report on Mr. Moore’s Essay Edward Caird
I have carefully gone over Mr. Moore’s Essay and his article in Mind, and intend to criticize it so far as it is possible to do so without going into too great detail; but I should first wish to express my appreciation of his philosophical ability. This is the more necessary, as it is difficult to criticize without unduly emphasizing what one thinks to be defects. I wish therefore to say that I think very highly of Mr. Moore’s philosophical powers, especially of his power of following his ideas to their ultimate results. What one has most to fear in philosophy is that this should not be done, and therefore the weakness and strength of ideas should never be distinctly seen, but hidden under some plausible compromise. It is therefore very high praise to say that a writer on philosophy can, or does realise fully the consequences of his principles. By such thinking he may make a great contribution to philosophy, even if the result is that the principles break down; for they cannot do so without throwing light upon the relations of the element of truth which they exaggerate to the other elements of it. Partly in consequence of this, but partly I think because Mr. Moore has not sufficiently studied how to be clear to those who are not looking at things at his precise angle; he is extremely difficult to understand. It has cost me much trouble to do so, and I do not think the fault in [sic] entirely mine. This fault however is very natural in a philosopher who has not had much experience of teaching, and should not be counted as taking much from the merit of the essay. The difficulty, however, I should add, partly arises from the fact that Kant is read so much through the eyes of Bradley and Lotze, which leads I think, to an imperfect realization of the best points in Kant’s work, and an exaggeration of his inconsistencies. Sometimes it is not quite easy to see whether Mr. Moore is interpreting Kant, or expressing his own views. Caird’s report is in the Wren Library at Trinity College where it is catalogued as Add. MS A 247 2(2). It comprises twenty-five folded small demy sheets in Caird’s autograph.
99
100
1897 examiners’ reports
At the same time, I think that, in spite of these defects, Mr. Moore shows himself to be a thinker of no ordinary power, and that he has established his claim to any reward that is given for such work. Of the Introduction, I need not say much. Mr. Moore rightly contends I think, that such predicates as ‘true’ ‘real’ ‘good’, cannot be explained without metaphysical enquiry. And he objects to the division of ‘Philosophy’ or ‘Reason’ as theoretical and practical. Kant’s full phrase is ‘criticism of Pure and Practical Reason’. That is, he arrives in his Transcendental Regress, at the Theory of Knowledge and the Theory of Morality or Practice, these being the two great facts of human life. Mr. Moore, for reasons afterwards given, holds that the reason that knows cannot be practical, i.e. cannot create motives for itself. But this is no legitimate criticism upon Kant’s division from Kant’s point of view. Perhaps, however, I may take occasion from what he says p. xii [p. 14] (‘He himself says that freedom of the will rests upon transcendental freedom, which latter is a speculative notion’,) to correct an error as to Kant’s meaning, which frequently appears in the Essay (see p. 64 seq. [pp. 61ff.]). Kant defines practical freedom as the power of acting upon motives not derived from desire, and says that we have empirical evidence of it in our consciousness of the moral law as a motive. This however, might be explained away, if we were able to conceive the ego as a mere phenomenon, as merely one object among the other objects of experience, or if its consciousness were limited to its own states as such an object. But both of these assertions Kant attempts by his transcendental Regress1 to shew to be untrue. The attempt therefore to prove that Kant is inconsistent from his own point of view, rests upon a misunderstanding. Kant’s point is just that this is an experience which cannot be explained on the principle on which we usually explain objects of experience, i.e. without taking account of their relation to the subject. There is no doubt a difficulty in speaking of the moral consciousness as experience at all, considering Kant’s narrow definition of experience; and Kant shews that he is quite aware of this, when he speaks of the moral law as only a quasi factum. (See e.g. K.P.R. p. 50)2 But the point is that we have here also an immediate unreflected consciousness, of which we have transcendentally to explain the possibility. 1
2
Caird introduces his conception of Kant’s method of ‘transcendental Regress’ in The Critical Philosophy of Immanuel Kant, Glasgow: James Maclehose and Sons, vol. i, pp. 22–4, 1889. The basic thought is that a ‘critical regress’ is required in order to resolve apparent inconsistency among appearances. This critical regress becomes a ‘transcendental regress’ when it draws on considerations arising from the possibility of knowledge of the relevant kind – see vol. i, pp. 483–7, 595. Caird here refers to p. 50 in Hartenstein’s edition of Kant’s Kritik der practischen Vernunft (AK 5: 47, MG 177).
Edward Caird
101
(I may refer to my Kant vol. ii p. 123).3 There is a sense in which it is true that no object of experience can be explained without taking into account its relation to the subject; and Mr. Moore, (p. 50. seq [pp. 53ff.]) maintains that this is true of all objects equally; but this is not Kant’s view. This point we shall have to discuss hereafter. The first chapter of the Essay begins with a comparison of Kant’s statement of the question of Freedom with that of Mr. Sidgwick, which is at least misleading. Kant holds that there is a necessity according to the law of Nature, but that that necessity pre-supposes a principle, which at once explains and limits it, and which also we must take into account in order to explain the moral consciousness. So far as I see, the words quoted from Mr. Sidgwick, do not bring in this distinction at all, and to make Kant answer the question in a form in which he has not asked it, cannot tend to clearness. Again, to call Kant a ‘determinist’, because he says that moral actions cannot be understood by taking them as merely a succession of states in an object, is surely misleading. The contrast that what must be possible because it ought to be done, need not be capable of appearing and indeed cannot appear, is necessary to Kant; but the action viewed objectively as an event and viewed in reference to the subject as its action, is not strictly the same thing. The contrast holds in much the same way as if I should say that a line without breadth is possible in Mathematics, but not in physical reality. If one has to take an abstract view of things, and then a point of view that does away with the abstraction, it is impossible to put the abstract on a level with the concrete view. The only defect in this analogy is that Kant, for reasons to be mentioned hereafter, does not think that we can ever understand how the real includes the appearance. There is nothing that directly contradicts the above in Mr. Moore’s statement, but I think that the way he puts the distinction is somewhat misleading. The section that follows p. 5 seq. [pp. 20ff.] is perhaps the most important in the Essay, as it shows how Mr. Moore understands Kant’s view of reality and freedom, and what view he opposes to it. There are two particular mistakes, which I may first get rid of. 1. (p. 8 [p. 24]) What Kant says of the Third Antinomy is that both, not merely one, of the alternatives may be true; one horn of the dilemma referring to the world of appearances, the other to the world of reality. This 3
This is an odd reference, not obviously relevant to the point at issue. Caird does provide a much fuller discussion in vol. ii, pp. 243–8.
102
1897 examiners’ reports
is not much more than a verbal mistake, due to Mr. Moore’s not noticing the way in which Kant puts the antinomy. 2. Mr. Moore states (p. 26 seq [pp. 37ff.]) that Kant only applies the idea of freedom or of an unconditioned cause to some things or beings within the system of appearance (i.e. to men) and not to the system as a whole. This is incorrect; for in stating the antinomy, Kant deals in the first instance with the necessity of a free cause for the whole system; although he goes on to speak more fully of free causes which are partial existences within the system, because of the special interest of this question in relation to morality (See K.R.V p. 320 ‘Nun haben wir diese Nothwendigkeit’ etc. – and following sentences).4 Passing now to Mr. Moore’s general argument, let us first see what is his view of Kant. Kant, according to Mr. Moore, confuses his epistemological view of things, after which he is striving, with a psychological view of them as determined by their relation to man’s peculiar kind of subjectivity. He therefore (p. 16 [p. 30]) is unable to answer the question: ‘How do we know that these conditions imposed by our knowing faculty are universal?’ All we know is, that our peculiar subjectivity imposes these conditions on any reality that is presented to us; for it is only as the given matter is brought under them, that it becomes an object for us. Kant, indeed, assumes a Ding an sich as a cause for the affections of our subjectivity which constitute this matter. But therein he makes illegitimate use of the Category of Cause; for in this application, the Category of Cause must lose its Schema, and be reduced to the merely analytic relation of Reason and Consequent (p. 10 [p. 25]). Now, as Kant holds, (quite correctly, according to Mr. Moore) that all the a priori forms that we apply to the given, are merely relational, and as he has surrendered, or never adopted, the notion that in the given we are in contact with reality, he is reduced to a merely subjective view of experience. His only escape from this is through the ideas of reason: ideas which represent, on the one hand, the mind’s dissatisfaction with a merely relational experience (such as alone we can have through the Categories as applied to a matter under the conditions of space and time); and on the other hand, its demand for a res completa, in which the universality of thought should be really united with the particularity of sense, and not externally applied to it as is the case in our experience. And this, for Kant, remains at least so far as theoretical reason is concerned, a mere demand, a problematical conception. For the idea of the totality thus introduced 4
‘K.R.V’ is Kritik der Reinen Vernunft, and p. 320 in Hartenstein’s edition is B476, GW 486: ‘We have really established this necessity’.
Edward Caird
103
cannot be legitimately combined with the idea of the Ding an Sich, whose existence was originally assumed, or was justified only by the illegitimate application of the Categories of Causality. We are, therefore, left with, on the one hand, a demand for totality which cannot be satisfied, because of the essential universality and relational character of thought; and with, on the other hand, an experience which as it is constructed according to such relational forms out of the matter of sense, is purely subjective. Now Mr. Moore holds that this view, (which he attributes to Kant) is quite accurate, in so far as it asserts the abstractly universal and relational character of our thought; and therefore of the experience or knowledge which we construct by means of it. But he maintains that it is quite untrue as regards the matter to which our thought is applied. He holds, p. 23 [p. 35], that reality directly appears in intuition, though it does not appear as it is; for the moment we go beyond the fact that it appears, we connect it as in space and time with other things according to the relational focus of our thought, and construct an appearance which is necessarily subjective, as is shewn by the antinomies to which it leads. The thing that appears, on the other hand, must be taken as an individual and unconditional totality, a self-determined whole, and in that sense as free. Thus for Mr. Moore its negative unconditionedness as a mere ‘that’, converts itself at a stroke into complete self-determination. This seems like saying that ‘0 = the absolute’, but Mr. Moore, I think, would answer, that in the intuition, the something perceived is given not as a mere ‘that’, but as qualitative – holding, as he does, that qualities can be sharply distinguished from relations (cf. p. 24 [p. 36]).5 This view seems to me untenable, and the criticism of Kant by which it is introduced, unfair, or at least reached by merely emphasizing Kant’s weaknesses. His main weakness is, I agree, the introduction of the psychological point of view into epistemology or rather as I should prefer to put it, his stopping at epistemology, and not going on to Metaphysics; for it seems to me that epistemology is a mere compromise between psychology and metaphysics. Kant tells us that his criticism has not to do with ‘the origin of experience’ but aims at discovering ‘what lies in experience’. (Proleg. p. 52).6 But as Mr. Moore sees, in working this out, Kant often falls back – and still oftener I think appears to fall back – on the process of constructing experience out of elements previously conceived in separation, as 5 6
The page to which Caird here refers does not include an explicit discussion of the distinction between qualities and relations; but Moore does maintain this distinction on pp. 52ff. Caird here refers to p. 52 of Hartenstein’s edition of Kant’s Prolegomena, in particular to §21a (AK 4: 304; trans. Hatfield, 55).
104
1897 examiners’ reports
our understanding and our sensations. Mr. Moore, as I understand him, accepts this view, only putting intuition for sensation, but maintaining the essential disparateness, and therefore coming to the same conclusion that what we know is merely appearances; though, as it is in intuition, and not in sensation that the object appears, what appears is objective reality, the thing in its unconditioned individuality. Now this seems to me to be exactly Kant’s point of view in the Aesthetic, where he goes on the supposition that individual things are given in perception, though the forms through which we apprehend them are subjective, and therefore we do not know them as they are. (Cf. the Dissertation in which the ideas of the Aesthetic were first expressed). In the Analytic, however, Kant bring in the further criticism that the forms also through which we determine them as objects are a priori, and further that this determination is not possible unless there is also a synthesis of imagination (in conformity with the pure synthesis of understanding) whereby the manifold – which is now all that is attributed to sense – is brought together in an image, to which the conception can be attached as a predicate. (K.R.V p. 99).7 The result towards which Kant is now working, therefore, is that the image is nothing without the conception, and the conception nothing without the image. We are thus brought to the conclusion that perception involves a judgment in which the subject and the predicate are necessarily distinguished, but at the same time necessarily related; a judgment which therefore is objective (a fact which Kant considers to be expressed by the use of the verb of existence as the copula. (K.R.V. p. 121).8 Kant thus rejects as inadequate the view which Mr. Moore adopts, and in the Analytic he is developing a quite different view, especially in the Chapters on the Transcendental Deduction and in the Schematism, though with some imperfection, owing to his never completely giving up the point of view of the Aesthetic. I have tried in my treatise on Kant (see especially i. p. 594 seq) to show how he fails to bring the two sides together. So far, however, as this new view is worked out by Kant, he is not constructing experience out of pre-existing factors, but simply taking our perception of things as in time and space, and showing that such perception could not exist for us without conceptions, which determine the images of perception as objective, and connect the objects so determined in one experience. In other words, we cannot even image the world in space and time, except by the use of the categories of quantity and quality and we cannot 7
B103–4, GW 211.
8
B141–2, GW 251.
Edward Caird
105
determine it as an objective world, unless all its elements are necessarily combined into one experience by the principles of Substance, Causality, and reciprocity. But again, Kant cannot stop here; for, in speaking of the synthesis by means of the categories, he has brought to light another element of experience; viz. the unity implied in the connecting process, a unity of which we become conscious when we become conscious of the self. Hence Kant says that all our experience is subject to the condition that we must be able to combine it with the ‘Ich denke’. It must, in other words, be capable of being combined with self-consciousness, else its elements could not be combined with each other in one experience. Instead, however, of taking in this new element and asking how it modifies the results so far gained, – which he must have done if he had been faithful to his programme of simply enquiring what is in experience, – Kant treats consciousness which the Ego has of itself, as utterly disparate from that which it has of objects, and that, in spite of his confession that the consciousness of self is mediated by the consciousness of objects. The judgment of self consciousness is, he holds, analytic, though it presupposes a synthesis (a view which is equivalent to saying that it is, and it is not, a synthesis). Hence it follows that the ideal demand of a self conscious being, for an object corresponding to itself, as it is a demand which springs out of the bare analytic unity of self consciousness, can never be satisfied in experience, which has only a synthetic unity. In other words, it is impossible to bring together abstract unity with abstract difference. Hence the demand of reason takes the paradoxical form on which Mr. Moore has insisted on p. 14 [p. 28]. The error, however, lies not in the demand of reason for an intelligible world, but in its being put in a way which could make its satisfaction an absurdity. It is not merely that we do not see how it is possible, but that we see that it is impossible. Now it seems to me that the whole subsequent difficulty of Kant’s philosophy, and especially of his view of freedom and morality, first arises from the fact that he is obliged to treat the idea of reason and the moral law (which is the form the ideal takes in Ethics) as at once synthetic and not synthetic, as a fruitful idea and yet a mere tautology, as a new principle for the explanation of experience, especially moral experience, and yet as a mere ‘A is A’, a formal principle, which would not authorize us to draw from premises any conclusion which was not explicitly contained in them. Mr. Moore, however, accepts this deadlock as the truth about human reason, whenever it attempts to go beyond our immediate contact with reality in intuition. Hence he is obliged, so to speak, to read Kant backwards, and to empty Kant’s use of the ideas of reason, in relation to knowledge and
106
1897 examiners’ reports
practice, of all its meaning. But it seems to me that the only legitimate way of correcting the deficiency in Kant’s procedure, is to reconsider his analytic view of reason, a correction which Kant himself suggests when he speaks of the analytic unity as implying a synthesis. What however I am more concerned to point out in this place is that, if the problem how we were to reach knowledge of reality, meant that the abstract particular had to be identified with the abstract universal, without either of them ceasing to be what they are, it would not only be insoluble in the sense that we cannot see how it is to be done, but in the sense that we see clearly that it cannot be done. On the other hand, I think Mr. Moore is right in holding that the very beginning of knowledge implies the highest ideal of it, though he gives no reason for this supposition, such as is given by Kant when he points out that all experience implies, and must be interpreted in relation to, the unity of self consciousness. Mr. Moore goes on in p. 28 [p. 39] to the discussion of the application of the idea of freedom to the will of man in the two forms 1) of liberty of indifference, as maintained on the ground of the immediate affirmation of consciousness; and 2) of self determination as maintained by Kant. Of the former, Mr. Moore tries to show, – and, I think, shows – that when we endeavour to state the nature of freedom to which consciousness testifies, we can get no distinct deliverance in favour of a motiveless decision between motives. And he points out the ambiguity of the question when we merely ask whether man can act rationally, and not whether he can either rationally or irrationally, since the former necessarily brings in the other view of freedom as self determination.a I have not Mr. Sidgwick’s book with me, and am not able to say whether Mr. Moore’s criticism upon his statements on this point are valid. But I should like in passing, to advert to Mr. Moore’s statement that the consciousness of an end may change the character of that end, and that in a processus ad infinitum; and that therefore, ‘the results of human volition alone among causes, must remain incapable of prediction’ [p. 40], though they are really determined like other effects. If consciousness thus changes our motives, it must be because it brings with it some new idea which continually reconstitutes them; and if this be admitted it will be difficult to maintain that the conscious self is not free, so far as it then determines its own motives. Such a process would be impossible, if we looked at ourselves and what was a
See, however, on this point what I say on p. 20 of his paper which has no such necessity of application. [Caird is here alluding to Moore’s reference to his book on p. 34 of the 1897 dissertation; see footnote 27.]
Edward Caird
107
happening in ourselves, as, to use Kant’s simile, a conscious Bratenwender (turnspit)9 might look at its own movements. Mr. Moore then goes onto deal with the Kantian notion of self determination – he thinks that Kant’s great error was that he did not see that every thing is unconditioned or free from our point of view, and determined by its relations from another; but that he conceived some things or beings in the phenomenal world as organic wholes, and so free, and others as not free (p. 50 [p. 54]). Freedom and necessity indeed are combined in the phenomenal world; for we can not only say that the existence of one thing is necessarily connected with the existence of another, but also that the quality belonging to the one is necessarily connected with the quality of the other, although as a quality it cannot be explained by this relation, and is therefore, free. Hence Kant was right in maintaining that the phenomena in their particularity must be such that the principles of the understanding can be applied to them. It is quite different, however, with the ideas of reason which have no such necessity of application, and Kant confused the two points when he said that the world of experience as a whole must be taken (even regulatively) to be such that we can realise the idea of totality-in-unity or organic unity in it. A fortiori he was wrong when he supposed that that idea could be used even as a regulative conception to determine any particular existence within the world. The only point according to Mr. Moore at which thought and reality meet, is the immediate intuition in which reality is given to us, yet not as it is, but as it appears. Whenever we go beyond the something intuited, the idea of the unconditioned can only be applied to the whole, and as so applied, it remains a mere idea; for it is impossible to make an experience conditioned in time and space, into a whole. If we take anything less than the whole, e.g. an animal, or a man, it is impossible to treat it as whole at all: for the line of division between it and other things is quite ‘arbitrary’. (p. 54 [p. 57]). And this applies to a mind quite as much as to an animal organism, for ‘there is no special form to fix the limits of a unity anywhere between the smallest distinct element and the whole mental world’ (p. 55 [p. 57]). And this, though it is true that ‘consciousness is itself an element in mental processes, so that here the form which determines the change knows itself ’ [p. 57]. What we have in consciousness is merely a series of presentations, and this is equally true of the willing consciousness; for will merely means that the presentation of an idea is the cause 9
Caird’s allusion is to Kant’s dismissive remark about the ‘freedom’ of the turnspit which, once wound up unwinds, as we say, ‘of its own accord’: see AK 5: 97, MG 218.
108
1897 examiners’ reports
of the presentation of that of which it is the idea. Even if we can act, as Kant asserts, in view of a law which prescribes what ought to be done, this merely means that the presentation of the law is causally related to the presentation of the action as done. And Kant was confusing the content of the former presentation which is universal, with the form of its individual presentation as an event in the psychical series, when he spoke of moral action, as action determined by a ‘mere conception’.10 A conception, as universal, may be the ground of another conception analytically, but it cannot synthetically be the ground or cause of an existence either in inner or outer sense. And Mr. Moore attempts to show that Kant confused these two things by reference to the way in which he speaks of the relations of practical and transcendental freedom. Now I have already spoken of what I take to be Mr. Moore’s misunderstanding of Kant’s assertion that practical freedom is experienced, but that the explanation of its possibility must be transcendental. We are conscious of being able to act not only in view of the objects of our particular desires or the pleasure they bring us, but in view of the idea of law or of a possible kingdom of ends. But how can we justify this consciousness, or prove that this is not an illusion, involving as it does practical freedom, i.e. the power of acting irrespective of the limitations of our individual being, as parts of a world in space and time essentially related to the other parts? How can we vindicate an experience which seems, if we take it as it presents itself[,] to carry us beyond what can be experienced? Kant’s answer is, that the self or ego is not an object like other objects, but the principle of unity in knowledge, a principle implied in the synthetic functions (by which objects of experience are determined for us as such objects) and which also, as we become conscious of its pure unity, gives rise to the ideas of reason in contrast with which experience is found wanting. So far, however, it might be said, as Kant says, that the self is merely the subject of knowledge which we cannot determine as an object; and, on the other hand, that the idea of an object to which it gives rise, is merely the abstract counterpart of its own formal unity. This might fairly be said, if the idea of a whole, or totality-in-unity, which thus arises, did not become a motive or principle of action. We might then possibly contemplate the play of our passions, as we contemplate the succession of events without us (like the conscious ‘Bratenwender’):11 and then we should not be conscious of freedom, for we should not attribute the acts to ourself at all. For in the theoretical 10
Caird is here discussing pp. 60–1 of Moore’s 1897 dissertation.
11
See note 9.
Edward Caird
109
consciousness, the consciousness of self is essentially the consciousness of the unity to which experience is referred, and which also is the source of the ideas that make us dissatisfied with experience. In our practical consciousness, however, this ideal does not merely come last to make us strive after an impossible completion of experience; it comes first as a consciousness of totality, which is bound up with the consciousness of the acting self. And although in one sense this is a ‘mere conception’, which we cannot objectify without typifying it as a law of nature, yet this does not prevent it from being a motive. I should, however, at once state there is a defect in Kant which partly justifies Mr. Moore’s representation of him, a defect which I have already pointed out. This is that the consciousness of the self, as the unity underlying experience, is taken as purely analytical. But Kant himself, as we have seen, shows us the way out of his difficulty, when he says that it involves a synthesis, which really means that it is a synthesis. For, if the unity on which we return from the synthesis of experience, becomes by that very return conscious of itself as a self, this surely is a synthetic process. The correction to be made in Kant is, therefore, to acknowledge this, and to consider how the acknowledgment of it will transform experience, i.e. how far the fact that experience exists only for a self, gives a new point of view for the interpretation of experience. It is because Kant views the unity of self consciousness as analytical, that he considers it only capable of suggesting a regulative idea, and not an idea that reconstitutes experience for us. And it is for the same reason that he conceives the moral idea as a conception, which may be typified, but cannot be known as realised. Another defect which is the counterpart of this lies in his imperfect doctrine of the inner sense, with which Mr. Moore deals on pp. 66–7 [pp. 66–7]. This doctrine is greatly modified in Kant’s second edition, but he is unable quite to transform it without altering his view as to the analytic character of self-consciousness (I may refer to my discussion of this in my Kant i. 605 seq.) In pp. 74–5 [p. 72] we see how Mr. Moore presents the opposition of analysis and synthesis. Either, he thinks, we have a synthetic relation of causality between presentations or an analytic development of conceptions, and there is no third possibility. This is a dilemma for Kant, just so far as he conceives the judgment of self consciousness with the ideas of reason and the moral law, as all expressing an analytic unity. But just so far as he points to a better idea, and suggests that the view of experience which we get by looking only to the connections of objects with each other,
110
1897 examiners’ reports
is essentially abstract, and that we have to correct this view by the idea that the object is essentially an existence-for-a-self, 12 we can reject both horns of the dilemma. And yet we are not reduced to a mere impossible ‘combination of the notion of causal dependence [ . . . . . . ] with that of logical dependence’ p. 75 [p. 72]. In other words, the truth is not to be reached either by treating time-conditioned experience as absolutely real, or by taking refuge in the mere timelessness of abstract laws; not only when on the one hand we correct the abstractness of ordinary experience and of science by the consideration of its relativity to a unity not in that series; and on the other hand, correct the abstractness of that unity by viewing it as essentially related to, and manifesting itself in, an experience of which time is one of the elements. And this seems to me the main point of importance in relation to the discussion in ‘Mind’. We cannot say simply that reality is, or that it is not, in time. It is not in time in the sense that we cannot rest upon our first conception of reality as a time-series, without considering that time, like objectivity generally, is for a self. On the other hand, we cannot take the unity of the Ego as simply timeless; that is, its unity with itself in knowledge and action cannot simply be opposed to, or severed from, the time process of the objective world which exists for it; for self consciousness is only possible through the consciousness of an objective world. To say simply that it is in time, would be to confuse it with one of its own objects; to say simply that it is not in time, might be taken to mean that it does not relate itself to, or realise itself in, an objective world. In particular, such a view would make it impossible to conceive how a self, which as such, is as Aristotle said a universal dÅnamiv should be also a particular part of the world he knows. From this point of view then, the question whether reality is, or is not in time, cannot be answered by a simple ‘yes’ or ‘no’.a I do not deny that there are great difficulties in working out this view, but it seems to me the only view that offers any kind of solution of the problem. Though I have been too lengthy already, I must notice another difficulty of which Mr. Moore frequently speaks; a difficulty which presses nearly up on Kant, but on which I think some light is thrown by this view. Are a
We might say ‘the whole is not in time, but time is in it’: but this without further explanation, would not help very much.
12
Caird advances this position in vol. i of The Critical Philosophy of Kant: ‘all knowable objects are in necessary relation to a thinking subject; they are essentially objects-for-a-self: and this relativity makes it impossible to treat them as external to the consciousness of such a subject’ (i, p. 12).
Edward Caird
111
we to say that Will, if free, is essentially rational, or that if free, it may be either rational or irrational? I do not think the question can be answered when simply so stated. To Kant, it is a great difficulty that the rational being should have any other motive than reason, and he tries to get over it by speaking of him as ‘taking up’ into his will, the motives of particular desire. But ultimately he is forced to think of these motives – or something that makes us capable of being affected by them, – as taken up into the will, prior to experience, (i.e. as a sort of original sin). According to the view I have suggested above, there would be no purely objective or sensuous motives at all. I never seek to gratify my desires, but always to gratify myself. I always act sub ratione boni, and therefore in all I do, I am in some way determined by the idea of the whole, which is the counterpart of the self. I am free however, in the highest sense, only in so far as I have organised my conflicting desires by this idea which is always involved in them; just as I know in the highest sense, only when I have recognised the unity that underlies all my experience, and have re-organised my knowledge in view of it. Mr Moore, (p. 82 [p. 76]) quite rightly argues that no such transformation, or reconstitution, of our motives could come out of the abstract idea of self-consistency, which seems to be all that is expressed in Kant’s first formula for the moral law. But Kant reinterprets this formula as equivalent to consistency with the ideas of the self, and of a kingdom of ends; and this shows that he is trying to express the idea of the whole, as involved in the consciousness of self, though continually checked by his view of self consciousness as analytic. Agreement of will with its own nature as rational is, if it be taken in this sense, not an empty idea, but one which is capable of being used as a principle to reorganise our desires; and which in fact, always does in a rational being enter into the constitution of them. The worst and the best Will so far agree, as there is no Will of a rational or self-conscious being which is not for the good. This may be a sufficient reply to what Mr. Moore says, p. 83 [p. 77]. We cannot have absolute badness any more than absolute error. The ‘ought’ arises out of the contrast of the particular end sought with the idea of the good or end involved in self consciousness; and the consciousness of this contrast may often precede, and always in the long run, follows, the attainment of ends which are unworthy. It is upon the growing sense of this discord, that moral improvement mainly depends. In the last ten pages of this chapter of the Essay, Mr. Moore seems to be trying to get back some of the ideas which he has rejected. Thus his general theory would lead us to say that, if good can be taken as a predicate of reality at all, it can be so taken only in a sense that excludes all evil;
112
1897 examiners’ reports
what is real is, as such, good. If, on the other hand, we apply the idea of goodness to the phenomenal, we can only say that all phenomena are good or evil according to the point of view we take up; good, as the real appears in them, and evil as it only appears (p. 90 [pp. 81–2]). But in this way there is no room for a better or a worse; and Mr. Moore, as we have seen, has excluded the idea of any relative whole, either in man or animals, as ‘arbitrary’ (p. 54 [p. 57]). But in p. 91 [p. 82] he brings back just this arbitrary conception, and bids us regard a good will as the best of things, and a bad will as the worst, assuming that a systematic unity is realised or realisable in the Will, as an end in itself. For ‘more of the subordinate objective ends contribute in the good will as means to a result, which is also an end in itself, and in the bad will are emptied of the goodness which they possess in their own right, by being united into a whole which is the most complete negation of an end in itself ’ [p. 82]. The idea of the self as an end in itself, is thus supposed to be an organising principle, which brings the natural goods of life – i.e. all particular ends of desire – into a systematic or organic whole. This, indeed, seems to be explained as merely ‘formal completeness,’ or ‘inclusiveness’; but obviously it cannot be meant merely that there is a larger aggregate of particular ends brought together in the good than in the bad will. If further we reject the idea that natural goods are determined as such by the relation to pleasure and pain, as Mr. Moore rejects it (p. 92 [p. 84]) there seems to be nothing left except that each of them is good as real, and all are equally good. Now a ‘sum of pleasures’ may be a doubtful conception, in itself and in so far as it characterises the whole by reference to one element, but a ‘sum of reals’, into which every element entered on equal terms with the others, would be meaningless. And how could the abstract idea of reality or goodness supply an organising principle for such a sum? Mr. Moore thinks we may treat freedom and goodness as conceptions, each of which implies the other (p. 92 [p. 83]), so that in the case of things determined as natural goods, we can argue to their free causation from their goodness, while, in regard to moral goods, we can argue to their good uses from the conception of their cause as free. But as to natural goods, we can only say on Mr. Moore’s principles, that as appearances they are bad, and as appearances of the real they are good; while as to moral goods, we can only say that they imply freedom if we can take the subject for which they exist, as a self, which as self conscious has an idea of the whole, and can use that idea as an organising principle, for the particular ends of desire.
Edward Caird
113
Mr. Moore tries to explain this reciprocal implication of the good and the free, by reference to the Hegelian categories which (p. 93 [p. 83]) he supposes all to imply each other, and all to embrace the whole extension of appearances, though they nevertheless express different degrees of reality in appearance. I think Mr. Moore has not quite seen the bearing of Hegel’s view. In a sense it is true that all the categories which we can apply to any thing, we can apply to everything; just as we cannot understand fully the ‘flower in the crannied wall’, without knowing ‘what God and man is’, according to Tennyson’s epigram. But taking a flower or a stone by itself, apart from its relations to higher forms of reality – which is the quite legitimate abstraction of science, – we can explain them satisfactorily without the use of the higher categories. And the categories sufficient for Mathematics are not sufficient for Physics, and those sufficient for Physics are not enough for Biology; and so on through all the scales of existence. We can of course apply the category of quantity to a man’s body or to his mind, or, if you like, to the absolute; but it becomes increasingly inadequate, and its application (without reference to the higher categories) takes the meaning out of the object in question. Hegel conceives the categories and the existence to which they are applied as graduated, so that we cannot fully explain the lower except through the higher; but the lower does not imply the higher as the higher implies the lower. And in the ascent extension and intension grow together; for the higher category, like the higher object that calls for its application, at once includes and transcends the lower. At the same time, Mr. Moore’s assertion has an element of truth in it as applied to such categories as ‘free’ and ‘good’; for the highest categories – what Hegel calls the categories of the Begriff – involve organic relations between different elements, so that each reciprocally implies the other. In chapter ii, Mr. Moore criticises the basis of Prof. Sidgwick’s Hedonism. He attacks 1) the doctrine that we cannot attribute the character of goodness to anything out of relation to some consciousness; and 2) the doctrine that in relation to consciousness there is ultimately no reason to prefer one thing to another, except as it produces a greater pleasure or a less pain. In regard to the former point, Mr. Moore asserts that, so far as we can abstract from the relation of consciousness at all, we can prefer, or count it good, that e.g. things should be beautiful, though no one has consciousness of them. He has ‘not given any strict argument on these points, as Prof. Sidgwick gives none’ (p. 100 [p. 89]). He puts it simply as an appeal to ‘intuitive judgment’. It seems to me that, if we take the idea of good, as simply expressing the completeness of an organic system, we
114
1897 examiners’ reports
can judge that an object has good uses more or less, without reference to any consciousness; i.e. we can abstract from the relation to consciousness, and still so judge. On the other hand, in so far as neither the world nor any part of it can ultimately be taken as a complete whole without reference to consciousness, and when we speak of good in morals – good that is willed – we are necessarily regarding the whole in that relation, I think that Mr. Sidgwick’s view is, for Ethics, the true one. As to the second point, it seems to me that Mr. Moore is right in saying that, if good be taken in this latter sense, i.e. if we think of the good as Aristotle does in the Ethics, as something perfect and self sufficient, and to which nothing can be added from without, – we cannot take it as pleasure, or as any other particular element of life, though we can say that neither it nor any other element can be left out. The organising principle, however, cannot be found in the mere abstract idea of the Sum of all things, each of which, as real, is an element in the good; it must, as Aristotle urged, be found in some positive conception of the whole, derived from the nature of the self who is to realise it – some conception that puts pleasure and everything else in its place, as a means to, and at the same time element in, the whole. I rather think that Aristotle would justly have directed against Mr. Moore’s view, one of the criticisms which (whether justly or not) he directs against Plato’s idea of the Good; viz. that an abstract idea of the Good, – the idea of that which is common to all things, to which the predicate good is applied – cannot supply a principle of unity in the moral life. At least I do not see how Mr. Moore would answer such a criticism.
The 1898 dissertation The Metaphysical Basis of Ethics by George Edward Moore
There be many that say: Who will shew us any good?1 1
Psalm 4 verse 6.
Preface
The greater part of the Dissertation, which I submitted for examination last year, has been included in the present work. Some omissions and alterations, involving an important change of view have been made; and nearly as much again of new matter has been added. I have followed the suggestions of my examiners in attempting to distinguish more clearly between my own views and those of Kant; and, in deference to the same suggestions, I have added an appendix on the chronology of Kant’s ethical writings. I have consulted the works of Caird and Adamson among English writers on Kant, and of Kuno Fischer, Benno Erdmann, J. H. Erdmann, Cohen and Vaihinger among the Germans. Had I been giving a general account of Kant’s philosophy, I should, no doubt, have had large obligations to acknowledge to all of these writers: as it is, I probably owe to them more than I can estimate. But I have not consciously taken any of my views directly from them. For my own metaphysical views, I am no doubt chiefly indebted to Bradley. But I have come to disagree with him on so many points, and those points of importance, that I doubt if I can name any special obligations. For my ethical views it will be obvious how much I owe to Prof. Sidgwick.2 I could not have devoted so large a space to criticizing his views on certain points, had he not supplied me with much of the common ground on which I take my stand. I am glad to find how largely I agree with his ‘Criticism of the Critical Philosophy’, Mind viii, 29, 313 in my discussion of Kant’s speculative philosophy also. 2
3
As with the 1897 dissertation, most of Moore’s references to Sidgwick’s works are to The Methods of Ethics whose title is abbreviated as ‘ME ’. Moore often just gives a page number, and we have not altered his main text to make it explicit that a reference to ME is intended. As before, however, we have provided further details in a footnote where Moore’s page numbers differ from those of the seventh edition (1907) of ME which is the only edition likely to be available to present-day readers. Moore, we presume, intends to refer here to Sidgwick’s paper ‘A Criticism of the Critical Philosophy’, Mind, 8 (1883), 69–91, 313–37. As will be apparent, however, Moore’s page numbers (‘29, 31’) do not
117
118
The 1898 dissertation
Part of this dissertation has been printed in ‘Mind’ for April of this year (n.s. vii. 26); but my views are no longer precisely those expressed in that paper. Some apology is perhaps needed for the smallness of the space occupied by directly ethical discussions in a Dissertation which professes to take Ethics for its subject. With more time, more would have been said about Ethics. But I found that everything I had to say presupposed general philosophical conclusions, which, since they are not generally accepted, it seemed desirable to work out as fully and consistently as possible. Reference to any modern ethical treatise will show how much their conclusions depend on the meanings given to such terms as ‘rational’, ‘empirical’, ‘necessary’ and many others. And this fact alone would seem to justify the large space devoted to a preliminary discussion of these terms. My references to Kant are in every case to the pages of Hartenstein’s edition (8 vols., Leipzig, 1867–8). R.V. stands for the ‘Kritik der reinen Vernunft (vol. 3); P.V. for the ‘Kritik der praktischen Vernunft’ (vol. 5); G. for the ‘Grundlegung zur Metaphysik der Sitten (vol. 4); and M. for the Metaphysik der Sitten’ (vol. 7).4
4
fit this paper; perhaps ‘329, 331’ were intended, since it makes sense to suppose that Moore agrees with Sidgwick’s discussion on these pages. As with Moore’s 1897 dissertation Moore’s references and abbreviations are retained here, but we have added cross-references to the standard Prussian Academy edition of Kant’s works and to current translations. As before Moore omits here to mention Kant’s ‘Kritik der Urtheilskraft’ (Critique of Judgment) to which he refers twice; he also refers to Kant’s Prolegomena to any Future Metaphysics and we add references to the revised translation of this by Gary Hatfield (Cambridge University Press, 2004).
table of contents 5
Introduction Chapter i. On the meaning of ‘Reason’ in Kant Chapter ii. Reason Chapter iii. The meaning of ‘Freedom’ in Kant Chapter iv. Freedom Chapter v. Ethical Conclusions Appendix on the Chronology of Kant’s Ethical Writings Appendix on Prof. Sidgwick’s Hedonism 5
6
Moore’s
This edition
i–xxiv
pages 120–32
1–52 53–87
133–60 161–80
88–144 145–85 186–97
181–210 211–31 232–7
198–205
238–42
206–20
Not included6
This table of contents occurs in Moore’s manuscript. But the page numbers given here as ‘Moore’s’ do not match those of the pages of his manuscript, which are somewhat chaotic. Presumably these are the numbers of the final typescript which Moore submitted and which does not survive (so far as we know). The only versions of this appendix which we have are from the 1897 dissertation and the appendix is therefore included with that dissertation and not repeated here.
119
Introduction
The scope of ‘Ethics’ has been very variously defined.7 Without prejudging any of the questions which it will be necessary to discuss hereafter, it may be stated summarily that the subject of the present essay is an enquiry into the nature of that which we denote by the terms ‘good’ or ‘what ought to be’. It may, perhaps, be well to confine the term ‘Ethics’, as does Professor Sidgwick, for example (p. 4), to ‘the science or study of what is right or what ought to be, so far as this depends upon the voluntary action of individuals’. Such a view is often roughly expressed by defining Ethics as ‘The Science’ or ‘Art of Conduct’;a and such is the scope of Aristotle’s Ethics, the book from which the term has been derived. ‘Ethics’ would thus take its place beside ‘Politics’, in the sense in which the latter is distinguished by Professor Sidgwick from Political Philosophy on the one hand and Political Science on the other. It is, perhaps, best on this view of its scope, to adopt Professor Mackenzie’s term and call it a ‘Normative Science’;8 for the term ‘Art’ would seem to be most properly confined to the actual pursuit of some end, or group of ends, in so far as such pursuit involves a systematic use of certain definite means, and not to include any statement of, or enquiry into, the rules by which such end or ends may be attained. If, for example, a book be entitled ‘The Art of Music,’ that title seems to denote the subject-matter described, just as does the title ‘Origin of Species’ or ‘The Cray-fish’, not the description or enquiry itself. Such a
Professor Sidgwick’s own view (ME p. 4) that his definition, as including Intuitionist conceptions, cannot be brought under the notion of Art, seems inconclusive, since any intuitionist who pretends to have a system, and therefore an Ethics at all, must admit the subordination of means to end in some cases; and the assumption of several ultimate ends, not intimately connected, seems in no way contrary to the notion of an Art.
7
As will be obvious, the first part of this introduction is taken with few changes from the introduction to the 1897 dissertation. Most of the explanatory editorial footnotes to the 1897 introduction are not repeated here, but a few new ones are added where appropriate. Cf. J. S. Mackenzie, Manual of Ethics, London: W. B. Clive, 1897, p. 4.
8
120
Introduction
121
an enquiry, in so far as it described the rules by which music of certain definite sorts, assumed to be the ends pursued in the Art, is actually to be produced, would rather itself be termed ‘Rules of the Art of Music’ or, in so far as it tried to give a reasoned connection between the rules by subordinating the various assumed ends to one main end, assumed to be the most general definition of the object pursued in the Art, ‘Theory of the Art of Music’. In this sense, the inquiry or the knowledge resulting might be termed a Normative Science, and such a ‘Normative Science’ Ethics as defined by Professor Sidgwick and as expounded by Aristotle, seems to be. The distinction between an Art and Science may seem almost too obvious to need mention; but it seems to be almost uniformly overlooked, and to lead to a misapprehension of the relation of Theoretical to socalled Practical Philosophy. E.g., if Art be defined as a ‘scientific discipline’ whose aim is ‘the moulding of things by man’s activity’ while the aim of Science is said to be knowledge,a it becomes plausible to treat Practical [Philosophy] as a study really coordinate with Theoretical Philosophy. But though Art, as the actual doing of things, in which sense alone the moulding of things can be its direct aim, may really be treated as so coordinate with knowing – the distinction being that between volition and cognition in psychology; when Art is treated as a ‘scientific discipline’ its direct object becomes ‘knowing’ just as much as that of science. The object of Ethics, ‘what ought to be’, is certainly different from that of any science, but in as much as the direct aim of Ethics is to know this and not to do it, it becomes pure theory and is subordinate to the general conditions of knowledge. This common confusion between doing and knowing what to do has not a little connection with Kant’s distinction between Theoretical and Practical Reason. It is from this confusion, too, that Professor Sidgwick was able to sayb that Practical Philosophy including ‘the study of the fundamental principles of Ethics and Politics’ . . . ‘seems to hold a position in reference to Arts in general,.. similar to that which Theoretical Philosophy holds with reference to Sciences in general;’ although he subsequently defines Arts (rightly, as I think) as ‘all departments of human activity, carried out systematically with reasoned adaptation of means to ends, for the attainment of some particular end, other than the knowledge applied c in the Art’. a b
c
Friedrich Paulsen System der Ethik (Hertz, Berlin 1889) p. 1. Printed lecture on ‘Scope of Philosophy’: 1897, p. 11. [This lecture is reprinted in H. Sidgwick, Philosophy: Its Scope and Relations, London: Macmillan, 1902; the passages Moore cites come from p. 28.] Italics are mine.
122
The 1898 dissertation
But it is to be noted that in ‘Normative Sciences’ or ‘Theories of Art’, the end, for the attainment of which rules are given, is itself merely assumed to be good. All the imperatives laid down are merely hypothetical. Aristotle, for instance, assumes that eÉdaimon©a is the end which both the individual and the state ought to pursue; and that Ethics and Politics are therefore only concerned to define what is meant by eÉdaimon©a and to lay down the rules by which it may be attained. Professor Sidgwick similarly makes it an axiom that pleasure alone can be an end in itself. But if the ultimate end to the attainment of which all Arts are only means (the Art of Conduct like the rest) be itself made the object of enquiry; if it is attempted to discover not even what it is we ought to pursue, but why we ought to pursue anything at all; or what is meant by saying that a thing is good or ‘ought to be’ – this enquiry seems to be different in kind from that of a Normative Science, and, since it is the subject-matter of the present essay, I have preferred to entitle it ‘The Metaphysical basis of Ethics’ or ‘Metaphysics of Ethics’, although authority might have been found for calling it ‘Ethics’ simply. The ground to be traversed will thus correspond to that covered by Kant in his Critique of Practical Reason, and to the province which Professor Sidgwick would assign to Practical Philosophy; but it will appear later why I prefer not to use the word ‘Practical’ in the title of my subject. It may perhaps be doubted whether such a subject-matter exists at all; but, pending proof, the meaning of the distinction drawn between Metaphysics of Ethics and Ethics may be elucidated by the analogy of the following distinction between Science and Metaphysics. Science is wholly occupied in ascertaining the laws which obtain among data immediately given as existent. In so far as it answers the question ‘What exists?’, it only points to what may be called the ‘matter of knowledge’, and ascribes to it the predicate of existence, without inquiring what that predicate means in itself. For Metaphysics, on the other hand, approaching its problem through epistemology, the stress must lie on the meaning of the predicate:a its ultimate strength against Scepticism lies in the impossibility of making any proposition that shall not involve an assertion of ‘being’ and the validity of certain logical laws. For it therefore the question ‘What exists?’ must be answered by a reference to the only possible tests of the validity of knowledge – tests which can never guarantee the necessary existence of the ‘matter of knowledge.’ Hence Kant, with his conception of a Ding an Sich, a
All so-called Metaphysics does not perhaps approach its problem through epistemology; but in so far as it can claim to be rational knowledge, it would seem that it must not assume any type of existent as the ultimately real; and, if it proceeds to justification of its assumption, no other method seems open to it.
Introduction
123
and the later Idealisms, have spoken of an ultimate reality, however little they have succeeded in fully defining its content, in comparison of which all the matter dealt with by Science can be condemned as mere Appearance. A similar distinction seems possible in the region of ‘the good’; we can not only ask ‘what is good’ with the meaning ‘Tell us what are the data of experience to which we may apply the term “good”’; but it is at least possible that just as the question ‘what is being?’ cannot be answered by pointing to anything which is, or even to the whole world which seems to be, so the question ‘what is good?’ may involve a metaphysical enquiry to which no identification of the good with any one empirical datum, such as pleasure, or with that which under certain conditions leads to it, can ever by the nature of the case furnish an adequate answer. It is possible pleasure may be good, and that a maximum of pleasure may be what we ought to aim at; and yet nevertheless to identify ‘the good’ with pleasure or ‘what ought to be’ with a maximum of pleasure may involve a fallacy like that of identifying a chair or a table, because they are, with ‘being’, or matter, when it is held to be real, with reality. Indeed such definitions must become tautologous, unless (1) either it be maintained that by ‘A maximum of pleasure is the good’ is meant ‘A maximum of pleasure, or that which is productive of a maximum of pleasure, is what is meant whenever the word good is used’ – a statement which is flagrant contradiction with the facts. For it is impossible to take refuge in the hypothesis that people have erred as to what they meant, since error implies that some definite meaning is attached to each of the terms falsely conjoined. Hence it must be admitted people meant something definite, e.g. by ‘virtue’, when they said it was the good; and so the error, if error there be, can only lie in the identification of it with good – not in mistaking it for maximum pleasure. (2) Or else it must be allowed that good has a meaning of its own, not exhausted by any empirical concept or any definition involving such. The same dilemma may be put with reference to Aristotle’s definition of the good – nerge©a kat ì retn9 etc. Is this a significant proposition? If so, then something other than nerge©a katé retn must have been meant by ‘good’. But it is impossible that that something other can have been any other empirical definition of good; since the significance of Aristotle’s doctrine rests on his refusal to identify nerge©a etc. with any such ordinary conception, e.g. pleasure. Good, in short, must have some meaning of its own, apart from any reference to empirical concepts, unless we are to mean e.g. by ‘pleasure is the good’ only that we shall limit the word good to pleasure; 9
‘Activity of the soul in accordance with virtue’, Nicomachean Ethics 1098a16.
124
The 1898 dissertation
and in that case the question ‘Is pleasure good?’ must be meaningless,10 although the consistent holder of such a view cannot be convicted of error. His view consists in denying that there is such a subject as Ethics at all, as distinguished from Psychology and Lexicography. And if his reason is simply that that he can see no special meaning for the word ‘good’, we may not be able to force him to see it, and his position may be made logically consistent. But on the other hand he is equally powerless against those who profess to see such a special meaning. Perhaps the fallacy involved in all empirical definitions of the good may be quite plainly exhibited as follows. One party, let us suppose, holds that the good is pleasure, another party holds that it is self-realisation, it being admitted that self is to be defined as in psychology by reference to the results of empirical enquiry. Our contention is that if each party is to hold that the other is wrong, each must presuppose ‘the good’ to have some other meaning than that which by their definition they assign to it. Either the controversy must be of the same nature as if one party maintained that light was ether-waves and the other that it was air-waves, in which case it is plain that each must have some direct knowledge of what they mean by light as distinct either from ether-waves or from air-waves; or else it must be sought to maintain that the controversy is merely ‘about words’ – about the definition of the term ‘good’. But in this latter case, either each party must be content to allow that the other may also be right – in which case both propositions become insignificant; or else there must be some reference to a standard by which the right meaning of words can be determined. And this latter contention involves the term that is to be defined: the argument is circular. Nor can this conclusion be escaped by the contention that we should have said the true meaning of words. For mere words have no true meaning, in the only sense in which truth can be established without reference to an ethical standard.11 The true meaning of a word can only be the sense which it ought to convey either absolutely or because it is usually used to convey it. But (1) it is impossible for either party to contend in the present instance that ‘pleasure’ or ‘self-realisation’ are what the term good is usually employed to convey. For in the first instance their propositions would be palpable contradiction with the facts; and in the second place they would not be ethical but purely lexicographical, and in order to bring them into the former sphere they would be forced to consider whether 10 11
The rest of this paragraph is an addition in 1898. It is notable that Moore here deletes the following sentence from p. 10 of the 1897 version: ‘Truth and error consist only in the consistency or inconsistency of the relations thought to hold between real objects, with those which actually obtain.’
Introduction
125
the vulgar notion of good was a true one or not; for which purpose the contention that good is a mere term must be wholly given up. If on the other hand (2) it is maintained that the word ‘good’ ought absolutely to be used in this or that sense; the moral standard implied, if it is not the lexicographical one of ordinary usage, must be that of truth, and truth implies a relation between real things, in which case ‘good’ must have a meaning other than that assigned to it. In other words, if the discussion what a word ought to be used to [be] mean or what it really does mean is to be of philosophical importance, it must always be implied that there is some real object or relation, perhaps not sufficiently disengaged by former analysis, for which a fixed term is required for the purposes of scientific discussion, and to which the term in question has always had reference, although perhaps confusedly. There can be no true meaning of words in a strict sense; but, if we are to discuss the truth about reality, it may be necessary to confine a word to one meaning; and when the word is already in common use, it must be shewn that the meaning proposed has been usually attached to it. It is in this sense that the present essay proposes to discuss the meaning of ‘good’ and ‘ought to be’: and it is maintained that there is a real object or relation corresponding to them, and to which they have always had more or less definite reference, and the identification of which with anything empirical always involves a tautology. We may, then, I think, fairly consider it proved that ‘good’ must have a meaning other than an empirical one; though freely admitting that the discovery of the empirical objects to which the term may be truly applied is highly important. All that we contend is that the proposition ‘This or this directly given object, or that which stands in such and such a uniform relation to this object, is the good’ is not capable of empirical verification. It is not of the nature of empirical or scientific propositions, which uniformly include some particular phenomenon in a larger class – but a class of phenomena only: e.g. when we say man is an animal, where the two notions conjoined are both of them defined by directly given empirical characteristics. To take a final instance: if it be said that the good is only that which pleases, that proposition is certainly significant, for we know what it is to be pleased, and it tells us that all things which cause us to have that feeling of pleasure are good. But if the proposition is to remain significant and not to become tautologous, it must not be maintained that good is a merely empirical conception. It may be attempted by pointing e.g. to a virtuous man: it may be said ‘See, he is good; and he does produce a balance of pleasure over pain.’ But this tells us no more than that the virtuous man is to be included in the class of things that produce more
126
The 1898 dissertation
pleasure than pain: it tells us nothing whatever about the good except that where we apply that predicate it can be shewn that a balance of pleasure is produced. But our whole question was of whether we were justified in applying that predicate and the present answer is a mere petitio principi. The justice of the above criticism seems to be tacitly implied in any ethical system which coordinates ‘what ought to be’ with ‘what is’ as if they were two fundamentally opposed conceptions; but its importance for speculative purposes seems much too generally neglected. It is only if the object of Ethics is conceived to be purely practical, to tell you ‘what you ought to do’, as is definitely announced by Aristotle, that there can be any excuse for assuming an Ethical axiom. If the object of the enquiry is knowledge for its own sake no axiom the denial of which would not involve an affirmation can be admitted; and though the results so obtained may be very inadequate to tell us how to live, quˆa knowledge, their value depends upon their truth and not upon their use. The main object of the present essay, then, is a purely speculative one; though it will be attempted later to shew how the theory here to be developed is connected with the Normative Science of conduct. So far I have been engaged in defining what I mean by Metaphysics of Ethics or the attempt to give a ‘transcendental’ meaning to good. This serves to distinguish the present enquiry from all such as start from an empirical definition of the good, and are only further concerned with the attempt to shew how their principle may be consistently applied and what are the means of attaining their end. It is Kant who seems most clearly to have recognised this distinction and to have done most towards a systematic exposition of the nature of the concept ‘good’;a but since he states the problem in a different way some preliminary statement of my grounds of dissent from him may not be out of place. Kant considers the discussion to fall within the province of Practical Reason, just as Professor Sidgwick would assign it to Practical Philosophy: and this nomenclature has been so universally adopted, that a discussion of it seems necessary, to avoid misunderstanding, at the outset of the present essay. In its more restricted sense Practice seems to have an essential reference to human nature; when some might even think that this exhausted its significance. On this understanding, Practical Philosophy should be confined to the determination of what ‘we ought to do’ in Ethics and Politics. But most would admit that a complete answer to this question a
Pref to P.V. [Presumably AK 5: 9ff. (MG 143ff.) despite the fact that Kant’s account of the good is very different from Moore’s.]
Introduction
127
implied primˆa facie a discussion of ‘what ought to be’, whether we can do it or not.a And when this further problem is taken into account, as by Kant it certainly is, the meaning of ‘Practical Philosophy’ must necessarily be extended beyond the sphere of mere human action, and embrace the discussion of what Nature ‘ought to’ produce, or the ‘end’ of the whole process of things in time. With this widened conception of ‘Practice’, the discussion would however still assume that what ought to be necessarily was not; that it was only something which might be in the future. And this contrast between ‘what is’ and ‘what ought to be’, as involving a distinction in time, is generally assumed to be fundamental. But from the point of view of the present essay no distinction involving Time can be assumed as fundamental. Time seems to be a merely empirical datum which, as such, must be discarded, when a logical test of reality is set up. It seems therefore necessary that a discussion which professes to deal with the most fundamental conceptions of Ethics, should start without prejudging the question whether ‘what ought to be’ really is or not. In other words, the conception of ‘good’ seems logically more fundamental than that of what ought to be if the latter is understood to imply anything empirical – in short, than any conception in which ‘practice’, by the largest extension of its meaning, can be involved. On the other hand it will be attempted to shew that what ‘ought to be’ may rightly be taken as identical with ‘good,’ the limitation of the former term being only due to a confusion between ‘being’ as existence in time and ‘being’ in a transcendental sense; for what only ‘ought’ to exist in time certainly may primˆa facie 12 ‘be’ in the transcendental sense – or, to adopt the language of modern Idealism, be real. It will follow indeed that what ought to be includes also what we ‘ought to do’ or what ought to be done; but that because it is ‘what ought to be’, and not vice versˆa. That this rejection of ‘practice’ is justified seems proved by the mere conceivability of a ‘good’ which is not merely to be: the wider notion is always logically the more fundamental. On our view, then, the distinction between ‘what ought to be’ and ‘what is’ is not such as to justify a fundamental division of Philosophy or of the Reason into the two parts ‘Theoretical’ and ‘Practical’. The persistence of this division is probably to be ascribed to Aristotle, whose Ethics, however, a
cf. Professor Sidgwick p. 33 [ME 7th edn p. 35]. I do not however accept his restriction, that ‘ought’ in the wider sense ‘merely implies an ideal or pattern which I “ought” – in the stricter sense – to seek to imitate as far as possible (see later, p. 218).
12
Moore has here deleted the following passage from pp. 13–14 of the 1897 version: ‘and (as will appear if the present discussion is successful), must’.
128
The 1898 dissertation
are professedly not metaphysical. In fact the distinction between ‘what ought to be’ and ‘what is’ is not as great as the distinction between what is and what exists; and the ‘practical’ is so far from representing ‘what ought to be’ that it is inextricably bound up with mere existence. It does indeed imply the transcendental concept; our very first point is that it necessarily leads thereto: but as soon as it is recognised that this concept to which it leads is of a wider nature than itself, it appears as much a misnomer to call the inquiry into that concept ‘practical’, as to call any branch of science ‘metaphysics’. So far therefore as general philosophical scheme goes, the standpoint here taken up seems to agree most with that of Plato. The ‘good’ is to be considered as an Idea pkeina ts oÉs©as perhaps – (the meaning that can be attached to such a phrase will be subsequently considered):13 but the really essential distinction is between Ànta and gign»mena, and the ‘practical’, by its very notion, belongs to the latter class. Kant, however, is a much more convenient starting-point for a Metaphysic of Ethics than Plato: partly because his exposition is so much fuller and more systematic; partly because he was able to make use of so much more speculation; and partly because his direct influence upon modern thought seems so much greater. And much of what he said seems to lead directly to the view above expressed. He himself says that ‘freedom’ of the will rests upon ‘transcendental freedom’, which latter is a speculative notion, though, according to him, not demonstrable by speculative means; and the fault in his statement seems to proceed largely from his not sufficiently distinguishing between the relation of reason to consequent and that of cause to effect. Our will, as he defines it, is a power of being [a] cause; but what he calls a ‘pure will’ cannot be a ‘will’ at all, since by his own definition of causality as applying only to relations of phenomena it must be put out of court by the mere fact of its ‘purity’. This is the old objection raised against his statement of the relation between phenomena and things in themselves – an objection which seems fatal to the basis of Schopenhauer’s professedly Kantian system. He himself, in answer to a critic of the ‘Grundlegung’, gives his reason for not taking ‘the good’ as the fundamental conception in his Metaphysic of Ethics. This however seems to amount to no more than that ‘the good’ necessarily implies an ‘object’ to which it may be applied.a Now an ‘object’ a 13
P.V. p. 67 [AK 5: 63, MG 191]. This reference to a subsequent consideration of this phrase is carried over from the 1897 dissertation (see p. 14) where Moore does indeed return to this phrase (see p. 83). But in the 1898 dissertation there is no similar later discussion of the phrase.
Introduction
129
must, according to the ‘Pure Reason’, be given in intuition. The ‘good’ must therefore either contain some empirical notion or it must refer to the object of an ‘Idea’, the existence of which we cannot ratify. But in the ‘Practical Reason’, he proceeds to infer the existence of ‘objects’ of ‘Ideas’ from the moral law. His insistence, then, on making the ‘law’ and not the ‘object’ the fundamental notion of his ethical philosophy seems to be due to the fallacy of mistaking the datum from which you infer, for the logical ground of the inference.a There is nothing in what he says to shew that the notion ‘good’ is not logically more fundamental than the moral law: he only insists that if it were not for the moral law we should know nothing of that notion. In short he seems to confuse the ratio cognoscendi with the ratio essendi, a confusion against which he himself gives a warning elsewhere.b It is the latter which in philosophy must always form the basis of distinctions; and therefore it would be more appropriate to distinguish the ‘Critique of Practical Reason’ as ‘philosophy of the good’ from ‘philosophy of that which is’, than by means of the ‘practical law’, which is only the datum from which it sets to work. Enough has perhaps been said in preliminary justification of our subject matter. As Kant himself says, philosophy differs from Mathematics in that it cannot start with definitions, because it has no intuition whereon to base them: a philosophical treatise must justify and explain itself by its whole course. I propose, therefore, in this essay, to criticise Kant’s Metaphysics of Ethics in some detail, as a means of making plainer my own view. I shall try to convict him of inconsistencies, which seem not unimportant, since they seem to have been adopted from him by later writers such as Green, Bradley, and Mackenzie.14 This is, of course, not possible, without some attempt at exposition of his views; and it seems therefore desirable to make some statement of the principles by which that exposition has been guided. It is no doubt a just demand that the historian of philosophy should be impartial. It is his business merely to give a true account of what the philosopher under consideration actually thought, and not to read his own views into his author’s works. But it is not perhaps sufficiently recognized of how little use this principle is. The historian may recognize that he has to give the facts; but in philosophy he has no sure method of determining what the facts are – what his philosopher did actually think. He may no doubt obtain the credit of unexceptionable impartiality very cheaply by giving a b
He seems himself to recognise this, P.V. p. 31 [AK 5: 29, MG 163]. P.V. p. 4 note [AK 5: 5, MG 140].
14
The rest of this introduction is new to the 1898 dissertation.
a
130
The 1898 dissertation
mere summary of the philosopher’s views in the philosopher’s own words. But it may well be doubted whether this is history; for it is far too likely that those words will convey no definite meaning either to himself or to his readers. If his object is to convey information, as presumably it should be, he must be careful that some meaning can be attached to the words he uses. It is impossible to give an account of any man’s ideas, without a knowledge of the facts to which alone his ideas can refer. If the ideas are false, they must at least be composed of references to actual fact, only connected in a wrong way; and the manner of connection must again be one which actually holds in some case, although it does not hold of the particular facts to which it is here applied. What these facts are to which the ideas of a philosopher must refer, is a question which only a philosopher can answer. Hence a historian of philosophy is necessarily involved in all the uncertainties from which philosophy itself has never emerged. He must relate the ideas of his author, if his account is to be anything more than a bare repetition of his author’s words, by reference in the last resort to ideas which seem to him to represent the actual truth. In so far as he fails in this, his account will have no meaning for him. But, as is usual in philosophy, what to him appears to be truth or fact will to others appear as fiction; and when he decides that his author must have meant this, because in fact there was nothing else for him to mean, his opponent may accuse him of partiality on this ground alone. In fact, however, the historian deserves no blame for giving his exposition in the only terms which seem to him to have any sense, however unintelligible they may appear to others: on the contrary it is his duty so to do. The charge of partiality can only be deserved, if he has attributed to his author without sufficient ground, an actual recognition of these facts, which appear to him so clear. To expound Kant in terms of Hegel, is the duty of a convinced Hegelian, but it is unlikely that such an exposition will seem fair to a disciple of Herbert Spencer, however carefully the expositor may have refrained from attributing Hegel’s views to Kant. In fact the follower of Herbert Spencer will probably be right in thinking that the exposition is not impartial; although he will have no ground for thinking any other actual view less biased. For, beyond somewhat narrow limits, there seems no ground for determining what a philosopher actually thought except a judgment of what it is right to think. The other method, put forward now, especially in Germany, as ‘history of development’ rests in the end on this. It is true that much light may be thrown on what a man thought by discovering what his predecessors thought; but our view of what they thought must again be determined largely by our view of what can be thought – of what facts there are altogether to suggest thoughts.
Introduction
131
Though therefore the distinction between clear exposition of a meaning and biased exposition of a meaning is in theory a perfectly sharp one, it would seem impossible in practice to draw the line between the two. In the following account of Kant, I have, I admit, mainly endeavoured to make the meanings I attribute to him precise and clear; for it seems to me that it is in this direction rather than in that of impartiality, that the history of philosophy has hitherto been most at fault. At the same time I try to support my views by careful reference to his own words and even to the ‘history of development’ so far as support may be obtained from such sources. To those who believe that Reason can be Practical, or that the human will is free, I cannot hope to appear perfectly impartial, although I admit that Kant thought so too. For if they are right, my exposition is obviously defective in that it gives too much prominence to those factors of Kant’s thought which I can recognise as true and into which alone I can analyse those other views of his, which I think false. According to them there must in each case be some other ultimate factor, which would probably have been much before Kant’s mind, but which I cannot conceive he should have thought of, because to my mind such a factor is an absolute non-entity. Such an omission may produce the appearance of a biased judgment as much as a complete mis-statement. With regard to the difficulty of founding a view of Kant’s meaning solely on his actual words, it may perhaps be useful to call attention to a passage of his own upon the subject. In his Preface to the Second Edition of the Critique of Pure Reason he says: ‘Every philosophical treatise may be shrewdly hit at separate points, since it cannot enter the lists, so completely armed in proof, as a mathematical work; and yet the structure of the system, treated as a unity, will not run the least risk in such attacks. Only there are but few who possess the necessary versatility of mind to survey it as a whole, if it is new; and still fewer who take pleasure in such survey, inasmuch as all innovation brings them trouble. Even apparent contradictions may be picked out in any writing, especially where the method of exposition is not of scholastic rigidity, if single passages, torn out of their context, be compared together. And in the eyes of him who relies upon another’s verdict, these will throw a prejudicial light upon the work; although by him, who has mastered the idea as a whole, they may easily be resolved.’a We must then suppose that Kant himself does not pretend to have observed a rigid exactitude in his terminology. No meaning can be proved to have been his, by a mere citation of his words in one or even in several passages. He a
R.V. p. 32 [Bxliv, GW 123].
132
The 1898 dissertation
wishes to be judged by the total impression of his work. What this should be, it is obviously impossible to determine by any method of scientific accuracy. We are left more dependent than ever upon a general judgment of what there was for him to mean. And although a judgment formed upon this method, may lead us to agree with what we thus take to be his general meaning; it is inevitable that it should also lead to a contradiction of some of his express statements. If these should seem to the historian to be not only in apparent, but also in real contradiction with his general meaning, he will no doubt be open to the charge of not having understood that general meaning rightly. But this is a charge which cannot be proved. For though Kant himself would not admit an actual contradiction between any of his views, his admission of apparent contradictions argues a certain vagueness or confusion in his thought with reference to some of the topics on which it touches; and we may be rigidly confining ourself to an exposition of his meaning, in shewing that he held contradictory propositions, if only we do not state that he himself saw them to be so. If he had seen it, he would have been forced to choose between them; and it belongs to the historian’s province to determine which he would have chosen, since that would depend upon what he thought to be the real significance of his system. And in this determination the historian must assume that Kant would have been influenced, to some extent, by the same arguments which appear conclusive to him. Failing this assumption, it would be impossible for him to declare what was the essential contention upon which the philosopher really meant to insist.
chapter i
On the meaning of ‘Reason’ in Kant
In all that Kant says on the subject of Ethics in the Critique of Pure Reason and his subsequent writings, reference is constantly made to the notion of a ‘Practical Reason’. From the time when he had conceived his final system of philosophy, generally known as the ‘Critical’ or ‘Transcendental’, ‘Practical Reason’ always appeared to him one of the most fundamental conceptions of Ethical Philosophy. Hence it was in a ‘Critique of Practical Reason’ that he tried to establish those principles upon which he supposed the possibility of an ethical system to be founded; just as he had expressed his views upon the principles and limits of speculative philosophy in a ‘Critique of Pure Reason’. It seems never to have occurred to him that serious difficulty could be found in this conception of a practical faculty of Reason. He assumes its existence in the opening words of his Critique of it, telling us that the objection of his work is to prove, by a criticism of the whole extent of this practical faculty, that there is such a thing as pure practical Reason. Now there can be no doubt that Kant thought there was something in common between Practical and Theoretical Reason. He declares that ‘after all it can be only one and the same reason, to be distinguished solely in its application’a ; and accordingly he often speaks of Reason in its theoretical and practical use, or of the theoretical and practical faculties of reason, as if Reason were one with two different applications, rather than as if there were two different kinds of Reason. What he means by Theoretical Reason will therefore throw some light on what he means by the Practical: they have, as he says ‘a common principle’ (G. 239).15 In his Preface to the First Edition of the Critique of Pure Reason, Kant tells us that his object is to satisfy the demand made upon Reason ‘that she should undertake anew the most troublesome of all her offices, that of self-knowledge, and should set up a court of justice, that should secure her in her just claims, and on the other hand dispose of all groundless pretensions, not by despotic sentences, but according to her everlasting a
G. p. 239 [AK 4: 391, MG 47].
15
AK 4: 391, MG 47.
133
134
The 1898 dissertation
and unchangeable laws.’ ‘This tribunal’ he says ‘is none of other than the Critique of Pure Reason itself ’ (p. 7).16 This is obviously metaphorical language; but it is just this metaphorical personification of Reason, of which I think Kant never cleared his thought. He goes on to say that by a Critique of pure17 Reason, he means a critique of the ‘faculty of Reason in general, in respect of all kinds of knowledge, to which she may aspire independently of experience’. This knowledge independent of experience he generally denotes as a priori. All absolutely a priori knowledge is knowledge of pure Reason. Hence it is pure Reason which is employed in the Critique itself, since that book determines a priori the limits of possible knowledge (p. 9, 28).18 The book is in very truth a judgment passed by pure Reason upon itself. Hence too Logic, Mathematics, and even Physics, in part at least, are knowledge of Pure Reason, in as much as their propositions are known a priori (p. 14).19 This faculty of a priori knowledge is the universal mark of pure reason; and we must therefore examine more closely what a priori knowledge means. Kant himself offers us two marks by which what is a priori may be distinguished.20 ‘A proposition which is thought along with its necessity is an a priori judgment’. And it is ‘absolutely a priori,’ only if it be not deduced from any proposition, that is not itself a necessary proposition. The second mark of the a priori is strict universality. But unfortunately Kant himself seems to admit the invalidity of this as a mark; since he immediately proceeds to state that an empirical universality may hold in all cases (‘for example, in the proposition: All bodies are heavy’) and hence be strictly universal (R.V. p. 35 (ib.)).21 ,22 16 18 20 21
22
17 Moore’s emphasis. Axi–xii, GW 101. 19 Bix–x, GW 107. Axv, GW 102; Bxxxvi, GW 119. Moore’s account of the marks of the a priori is based on Kant’s introduction to the B text, esp. B3–4 (GW 137) which Moore refers to below as R.V. p. 35. At this point Moore’s MSS has the following passage, which has then been deleted:
And he can only support the distinction between such cases and those of absolute universality, by reference to an implication in propositions of the latter kind, that their source is pure Reason. His exposition therefore, so far as he is here concerned to point out how propositions may be recognised as proceeding from pure Reason, is vitiated by circularity. With regard to necessary propositions, Kant would seem to recognise that their necessity is something which cannot be further described but must be immediately recognisable. For instance, he frequently appeals to the fact that Mathematics does give necessary propositions, without troubling to explain in what their necessity consists or to prove that it exists. It would seem possible, according to him, to class any true proposition whatever as a necessary or as an empirical proposition; and this classification lies at the base of his whole work. An empirical proposition is one which merely asserts something about this or that particular thing in space or time (it may be simply their existence), or the connection between such; and if anyone does not understand what is meant. This is the end of page 3 of the MSS of this chapter and the next page is numbered 12. The passage at the bottom of page 3 also occurs in ‘The Nature of Judgment’ (p. 184), and it seems clear that
On ‘Reason’ in Kant
135
It is true Kant states that this empirical universality is merely arbitrary. We ought, he says, to express our proposition in the form: ‘So far as we have yet observed, there is no exception from’ the rule that all bodies are heavy. But it would seem that such a qualification can only affect the truth of our proposition and not its content. It may be questioned whether we have a right to assert universality, but it is universality which we assert. The limitations which Kant points out as belonging to the proposition, can properly be expressed only in the doubt whether we have found a rule at all, not in a doubt whether there are exceptions to it. It may not be true that all bodies are heavy; but whether true or not, it is a universal proposition. There is no difference between this proposition and such as are a priori, in respect of universality. And Kant could hardly wish to assert that the difference lay in its truth. For this proposition, he would admit, may be true; and, if so, then it would be a priori. But he would not admit the suggestion that it may be a priori: he asserts that it is not so. The difference between the empirical and the a priori, if there is a difference, must therefore be in some other mark than this universality, which Kant nevertheless asserts to be ‘by itself an infallible criterion’ (ib., p. 35).23 We may next consider whether such a mark is to be found in ‘necessity’. In this investigation, too, it may be well to examine his example ‘All bodies are heavy,’ since this proposition might seem to have a claim to necessity also, just as it is undoubtedly universal. Kant speaks of it as ‘a rule borrowed from experience’ (ib., p. 34).24 By this language and by his use of ‘Bodies are heavy’ as convertible with it, he would seem to suggest that he would not base its empirical character solely on its extensional interpretation. If, as seems probable, he would allow ‘Body is heavy’ or ‘Man is mortal,’ to be equally empirical propositions, then it is plain that what he calls empirical may involve necessity. It is certain, at all events, that if we are to understand by empirical propositions only such as experience can justify, such a proposition as ‘All bodies are heavy’ cannot be regarded as empirical. It is based on the proposition ‘Body is heavy,’ with which, if it is to be used for purposes of inference, it must be regarded as convertible. I assume, therefore, that Kant would not have refused to regard ‘Body is heavy’ as an empirical proposition. It would seem certainly to come under his class of ‘rules drawn from experience,’ whereas ‘All bodies are heavy,’ regarded solely as extensional, cannot be called a rule. The use of this example
23
Moore used the missing pages for the composition of ‘The Nature of Judgment’. We have therefore inserted at this point the material (pp. 184–9) from ‘The Nature of Judgment’ which directly follows this passage, up to a point where the line of argument diverges from that developed in the rest of this chapter. See the editors’ introduction pp. lxxv–lxxx for more discussion of the issues here. 24 R.V. p. 34 (B2, GW 137). R.V. p. 35 (B4, GW 137).
136
The 1898 dissertation
would seem to lead to important results with regard to the true definition of empirical propositions. But let us first return to ‘All bodies are heavy’; since even this would seem to involve in its very meaning an assertion of necessity. If it be taken purely in extension, it must be resolved into ‘This body, and that body, and that body, ad infinitum, are, have been and will be heavy’. It involves, therefore, the proposition ‘This body is heavy’. But in any proposition of this simple categorical form the notion of substance and attribute is already involved.a Wherever a predicate is asserted of a subject, it is implied that the subject is a thing; that it is something marked by the possession of certain attributes and capable of possessing others. ‘This body is heavy’ presupposes, therefore, ‘Body is a thing, and heaviness is a mere attribute’. For we could not convert the proposition into ‘Heaviness is corporeal’, But that ‘Body is a thing,’ and that ‘Heaviness is an attribute,’ would seem to be necessary propositions. We may indeed be mistaken in supposing that they are true; but if we were ever to find that heaviness was not an attribute, we should be bound to conclude that it never had been and never would be, not that it was so once but had ceased so to be. All such judgments are truly ‘thought along with their necessity’. They are as necessary as that 2 + 2 = 4. The difference between the two forms of proposition lies not in that the former lacks necessity, nor even that it implies the proposition ‘Heaviness exists’; for even if heaviness did not exist, the proposition would be true. The proposition means that heaviness could not be other than an attribute; and hence, if Kant’s words (p. 34)25 are to be taken strictly, it cannot be empirical. In this respect, therefore, it is quite on a level with ‘2 + 2 = 4’; which also would be true even if there were no two things. The difference seems to lie rather in the nature of the concepts of which the necessary relation is predicated. ‘Heaviness’ can exist; it is not meaningless to say ‘Heaviness exists here and now’; whereas ‘attribute,’ ‘two,’ and other like conceptions can only claim a precarious sort of existence in so far as they are necessarily related to these other notions of which alone properly existential propositions can be made. If, therefore, we wish to find propositions involving no necessity,b we must descend to purely existential propositions – propositions which do not involve the notions of substance and attribute. These alone can be truly taught us by experience, if experience, ‘cannot teach us that a thing could not be otherwise’ a b
25
R.V. p. 36 [B6, GW 138]. Even these involve the necessary properties of time; but this point may be reserved for later consideration. R.V. p. 34 (B2, GW 137).
On ‘Reason’ in Kant
137
(p. 34).26 And even these are free from necessity, only if they are understood to assert something with regard to an actual part of actual time. They must involve necessity as soon as the distinction between ‘This is’ and ‘This was’ is disregarded. It would seem, in fact, to be a mark of the sort of existence which they predicate that it is in time. They may affirm ‘This exists,’ or ‘This has existed,’ but if they take the general form ‘This is,’ that must always be understood to mean no more than ‘This always has been, is now, and always will be,’ and can be strictly analysed into as many different judgments as time is divisible into separate moments. If, therefore, the difference between the empirical and a priori lay primarily, as Kant implies, in the nature of the judgment, not in that of the concept, only existential propositions could be empirical. In order to represent even ‘This body is heavy’ as an empirical proposition, it would be necessary to analyse it into the form ‘Heaviness and the marks of body exist here and now’. But this is certainly not its whole meaning. We must, therefore, suppose that in order to obtain a clear definition of what Kant meant by empirical propositions, we must base it upon the nature of the concepts used in them. Empirical concepts are those which can exist in parts of time. This would seem to be the only manner of distinguishing them. And any proposition into which an empirical concept enters may be called empirical. Kant himself does recognise the necessity involved in such a proposition as ‘This body is heavy,’ although, for reasons which will appear hereafter, he states it in a somewhat different way. The main object of his ‘Analytic’ is to show that any such judgment involves a ‘synthesis of the manifold of sense-intuition,’ which is ‘necessary a priori’ (p. 126).27 But he regards this synthesis rather as necessary in order to bring mere perceptions into relation with the ‘unity of apperception,’ than as directly involved in the empirical judgment. Moreover, in order to explain how the forms of synthesis can apply to the manifold, he introduces the inner sense as mediator, and describes the judgment as converting the psychical connexion of the presentations into an objective connexion rather than as applying the categories to a mere manifold, which cannot properly be described as psychical. Accordingly he gives as the ultimate empirical judgment, out of which the application of substance and attribute produces ‘Bodies are heavy,’ the subjective judgment ‘When I carry a body, I feel an impression of heaviness,’ instead of that given above ‘Heaviness and the marks of body exist together.’ 28 He does not seem to see that his subjective judgment already 26 28
27 R.V. p. 126 (B127, GW 225). R.V. p. 34 (B3, GW 137). Moore gives the following reference here: ‘P. 121, cf. also Prol., p. 54n’. The passage he is referring to is clearly R.V. p. 121 (B142, GW 252). The reference to Kant’s Prolegomena is to the footnote to §22 on p. 54 of Hartenstein’s edition, vol. 3 (AK 4: 305; trans. Hatfield, 57).
138
The 1898 dissertation
fully involves the category in question. A statement about my feelings is just as ‘objective,’ in the required sense, as a statement about what is conceived as in space. With the above definition, therefore, it is obvious why ‘Body is heavy’ should be called empirical; whereas, if absence of necessity had been the mark required, it would have been difficult to find a reason. For this proposition does not only involve, like ‘This body is heavy’ or ‘All bodies are heavy,’ the necessary judgments that body is a thing, and heaviness an attribute; it asserts a relation between a ‘heaviness’ and ‘corporeity’ such as no experience can prove or disprove. If we found a body which was not heavy, that would indeed lead us to deny the truth of the proposition; but it would also entitle us at once to the opposite necessary proposition ‘Body cannot be heavy’. And this is just what holds of 2 + 2 = 4. It is perhaps inconceivable to us now that two and two should not make four; but, when numbers were first discovered, it may well have been thought that two and two made three or five. Experience, no doubt, must have been the means of producing the conviction that this was not so, but that two and two made four. The necessity of a proposition, therefore, is not called in question by the fact that experience may lead you to think it true or untrue. The test of its necessity lies merely in the fact that it must be either true or untrue, and cannot be true now and untrue the next moment; whereas with an existential proposition it may be true that this exists now, and yet it will presently be untrue that it exists. The doubt about the truth of ‘Body is heavy’ would seem to proceed chiefly from our uncertainty as to what we mean by ‘Body’ and by ‘heavy’. We cannot recognise instances of them with as great precision as we recognise instances of number; and hence we cannot be sure whether the truth of our proposition may not be overthrown. The proposition is arbitrary solely in this sense. There would seem no doubt that we mean by it to assert an absolute necessity; but between what precise concepts the necessary relation, of which we are certain, holds, we must leave to experience to discover. From the foregoing analysis it would, therefore, appear that the true distinction upon which Kant’s division of propositions into a priori and a posteriori, necessary and empirical, is based, is the distinction between concepts which can exist in parts of time and concepts which seem to be cut off from existence altogether, but which give rise to assertions of an absolutely necessary relation. Kant would seem to include among empirical propositions all those in which an empirical concept is used; whether the proposition asserts a necessary relation between an empirical and an a priori concept, or between two empirical concepts. What it is important to emphasise is that these two kinds of proposition are not distinguished by the absence of the marks which he gives for the a priori;
On ‘Reason’ in Kant
139
they both include both necessity and strict universality. Empirical propositions would therefore include a wide range of propositions, differing very much in the meaning of their assertions. They seem to extend upwards from mere assertions of the existence of this or that, of the type ‘Heaviness exists here and now’; through propositions of the usual categorical form ‘This body is heavy,’ which include necessary propositions in their meaning, but at the same time imply an assertion of existence; to propositions which assert existence at every time, while still retaining the element of necessity included in the last, like ‘All bodies are heavy’; and finally to those propositions, upon which alone the validity of the last class can be based – propositions which assert a necessary relation, without any implication of existence whatever, of the type ‘Body is heavy’. The only common element in all these different classes would seem to be that they all make assertions with regard to some empirical concept, i.e., a concept which can exist in an actual part of time. The second and third classes are mixed and involve necessity, because there is also included in them an assertion with regard to an a priori concept. To all of them Kant would seem to oppose as purely a priori propositions, those which make an assertion solely with regard to a priori concepts and which for that reason can imply no assertion of existence, since an a priori concept is one which cannot exist in the limited sense above explained.29 In the foregoing account we have only considered such judgments as Kant would call ‘synthetic’. These are indeed the class with which we are chiefly concerned, since the main object of the Critique, as he often expresses it, is to discover how synthetic judgments a priori are possible. But Kant does also claim that analytic judgments are a priori. He even holds that the ‘chief business of our reason’ consists in making them (p. 39).30 It is therefore proper to consider how they are to be distinguished. Kant himself says they might be called ‘explanatory judgments’ and holds that they only ‘split up the subject into its component concepts, which we already thought in it, though perhaps confusedly’ (p. 40).31 But it would seem doubtful whether they deserve even so much importance as this account would give them. Let us take Kant’s example: ‘All bodies are extended’. Here, Kant says,32 ‘I need only become conscious of the manifold, which I always think in’ my concept of body, ‘in order to find’ 29
30
At this point we return from ‘The Nature of Judgment’ to p. 12 of Moore’s manuscript. The next paragraph of ‘The Nature of Judgment’ closely resembles a fragment of chapter ii (see p. 170), and the subsequent discussion there does not continue the line of thought developed up to this point; hence it seems reasonable to make a break here. For further discussion see the editors’ introduction, pp. lxxvi–lxxvii. 31 R.V. p. 40 (B11, GW 141). 32 R.V. p. 40 (B11, GW 141). R.V. p. 39 (B9, GW 140).
140
The 1898 dissertation
the predicate of extension ‘included in it’. But this would seem to reduce the supposed analytic proposition rather to a synthetic proposition a posteriori about my own state of mind. In order to establish this, it is necessary to consider what is meant by ‘the concept, which I always associate with body’. Must ‘my concept of body’ be a correct one? If so, then, whenever I think of body, I must think of what body really is; and in that case the analytic proposition will assert that in a real body, and not only in my conception of it, extension is included. That Kant does not mean this is plain as well from his words as from his famous criticism of the ontological argument. That ‘God is good’, though an analytic proposition, does not imply that God exists. I may, therefore, make analytic propositions without having a correct conception of the subject about which I make them. But, if my conception need not be correct, who is to say that a man’s conception of body might not include, for instance, pleasure? In this sense, which is the only sense of ‘my conception of body’ which the instances seem to warrant, it means no more than the ideas which I always associate with the word body. And if I do not happen to know what the word body properly means, i.e. the ideas with which it is usually associated, then, if Kant’s account be just, I may be making an analytic proposition when I say “All bodies are full of pleasure’. I think it will be allowed that Kant would not wish to call this an analytic proposition. The true account of those propositions which he called analytic seems rather to be that they all assert merely the meaning of words. I am certain that ‘All bodies are extended’ solely because I know that the word ‘body’ properly, i.e. customarily, denotes something that is extended. The type of analytic propositions is the definition in a dictionary; and the definitions of a dictionary are certainly empirical, i.e. synthetic a posteriori, except in so far as they may be taken as ethical, i.e. asserting how you ought to use a word. That analytic propositions should have been assumed to have importance in philosophy and to form a class distinct from the synthetic, would seem to be due to the distinction between knowledge already acquired and its discovery. Whenever an analytic proposition is used for philosophical purposes, its usefulness always depends on a synthetic proposition which is implied in it, but which, owing to our familiarity with it, we do not recognise as such. In order to33 the formation of any composite concept whatever, synthetic propositions are required; but when the concept is, in general use, under a single name, the necessary connections between its marks are assumed as 33
Sic; but ‘for’ seems intended.
On ‘Reason’ in Kant
141
parts of a given whole, and this whole is taken as the logical prius34 of the various connections, whereas they are in reality the prius of it. This would seem to explain why Kant’s discovery that mathematical propositions are all synthetic was not made earlier. Four was taken to mean the sum of two and two, or twice two, and if this is the meaning of four then obviously 2 + 2= 4. In fact, however, as Kant points out, the sum of two and two is something different from four. Four is a distinct idea and in order to know that it is equal to the sum of two other numbers a synthetic proposition is required. It is much to be deplored that Kant did not carry his investigations on this matter further. His retention of analysis, in this sense, as an important process distinct from synthesis has done much to confuse his thought in many places.a As an instance of the use he makes of it, may be taken a sentence of his, in the very paragraph in which he discusses the synthetical nature of mathematical propositions (p. 43).35 ‘If ’ he says ‘anybody will not allow’ that mathematical judgments are a priori, ‘well and good[.] I confine my statement to pure Mathematics, the very conception of which entails, that it contain not empirical knowledge, but merely pure knowledge a priori.’ Kant seems to rely on this argument to some extent as proving that there are a priori judgments in Mathematics. It may, indeed, be admitted that if the word ‘pure’ is generally used as equivalent to a priori, the common use of the name ‘pure Mathematics’ proves that most people think there are a priori propositions in Mathematics. But these are all merely empirical propositions, some concerning the general use of words, and others psychological; and all may be denied without any contradiction. Moreover they are quite useless to prove Kant’s point that there are a priori propositions in Mathematics; except in so far as common belief may be taken as proving the truth of what is believed. A philosopher, who is often concerned in proving that common belief is wrong, must shrink from relying on such an argument. The analytic proposition may therefore be dismissed as useful to philosophers only in the sense in which the pens and ink with which they write their works are useful. Analysis does indeed denote a process useful in philosophy as elsewhere, namely experimental observation; but the propositions thus obtained Kant has no hesitation in declaring to be synthetic (p. 40).36 Analysis, in the sense in which it gives a name to analytic propositions, is in reality only a synthesis of words a
34 35
See later (p. . . . ) for a most important instance. [It is not obvious what passage Moore has in mind here – perhaps the critical discussion of the unity of apperception on pp. 146–7.] ‘Prius’ is the neuter form in Latin of ‘prior’; it means ‘that which has priority’. 36 R.V. p. 40 (B11, GW 142). R.V. p. 43 (B14–15, GW 144).
142
The 1898 dissertation
with the objects they denote; while the description Kant gives of it belongs properly to the other sort of analysis as applied to a special subject-matter, namely an individual’s mind, and has no relation whatsoever to so-called ‘analytic propositions’. Having now obtained some account of the distinction between a priori and empirical, we may proceed to consider the bearing of this distinction upon the problem before us – the meaning which Kant assigns to the word ‘Reason’. As has been already stated, Kant regards all a priori propositions as propositions of Pure Reason; and having discovered the nature of a priori propositions, we have now to imagine in what relation Kant thinks them to stand to that to which he calls ‘Reason’. In this connection it is necessary to discuss what Kant seems to have held to be his great achievement in philosophy. In a famous passage from his Preface to the Second Edition of the Critique (pp. 17–19)37 Kant compares his work to that of Copernicus. Just as Copernicus found it possible to explain the motions of the stars, when he replaced the old assumption that they revolved round the spectator, by the assumption that the spectator revolved and the stars remained fixed; so Kant hopes to have solved the problems of Metaphysics by assuming that the objects of knowledge conform to the laws of our cognitive faculty, instead of supposing that our cognitive faculty must conform to them. ‘If ’ says he ‘intuition had to conform to the nature of the objects, I do not see, how anything could be known about it a priori; but if the object conforms to the nature of our intuitive faculty, I can quite well represent to myself this possibility.’38 The same advantage, he thinks, follows from the assumption that the conceptions by which mere presentations are referred to an object also do not conform to the object, but vice versa: this assumption too makes it easy to see how another kind of a priori judgments39 can be made about objects. In short, Kant thinks that he has brought about a revolution in Metaphysics, by which it can be put upon the ‘sure path of a science’,40 in consequence of this ‘happy thought,’41 that ‘in a priori knowledge nothing can be attributed to the objects, except what the thinking subject takes out of itself.’42 The view that this was a happy thought seems to have been taken by very many of Kant’s successors and commentators. Kuno Fischer,a for example, expands the comparison with Copernicus as illustrating the a 37 41 42
Geschichte der neuern Phil. Vol. 3. Ed. 3, pp. 7–9. 38 Bxvii, GW 110. 39 sic. 40 Bxv, GW 110. R.V. pp. 17–19 (Bxiv–xviii, GW 109–11). Bxiii–xiv, GW 109 (GW translates Kant’s Denkart as ‘inspiration’ instead of ‘happy thought’). Bxxiii, GW 113.
On ‘Reason’ in Kant
143
main theme of the critical philosophy, whereby a standpoint is established from which the dogmatic method in philosophy can be finally condemned. Even Lotze is inclined to accept the principle, so far as thought is concerned (Logic trans. 2nd ed. pp. 3–4 and passim).43 On the contrary it appears to me an exceedingly unfortunate suggestion; and in discussing this connection which Kant finds between the a priori and the Reason, I shall hope to shew that there is no reason for asserting it; that it does not serve the purpose for which it was devised; and that it is inconsistent with other contentions of the Critical Philosophy, which appear to me to be true and of great importance. In the passage above referred to,44 Kant only puts the doctrine forward as a hypothesis. He adds in a note that it is proved apodictically in the Critique itself.45 It appears that he thinks the Critique furnishes two separate proofs (1) from ‘the nature of our presentations of space and time and the elementary concepts of the understanding’ (ib.) (2) from the avoidance of contradiction in relation to the notion of the ‘unconditioned’, which is rendered possible only by the distinction between appearances and things as they are in themselves (pp. 19, 20).46 It will be convenient to examine these two proofs separately. (1) The Aesthetic explicitly raises the question at issue in the form: ‘What are space and time? Are they actual existents? Are they only determinations or perhaps relations of things, but of such a sort, that they would belong to things in themselves, even if they were not intuited, or are they such as attach only to the form of intuition, and hence to the subjective nature of our mind, without which nature these predicates can be attributed to no thing whatever?’ (p. 58).47 Kant then proceeds to shew, in the first argument of the metaphysical discussion, that ‘space cannot be borrowed through experience from the relations of outer experience’48 since it is already presupposed in distinguishing certain sensations (Empfindungen) as ‘not only different, but in different places’. To this argument, in a modified form, as shewing that space is logically prior to the objects of experience, no exception need be taken: but it only shews that space is logically prior or a priori, not its required subjectivity. In the transcendental discussion, he argues from the necessity of geometrical propositions to the a priori nature 43
44 45 47
Moore is here referring to R. H. Lotze, Logic, ed. B. Bosanquet, Oxford: Clarendon Press, 1887, though the passage Moore refers to does not express the Kantian doctrine in question. Lotze provides a much clearer expression of his quasi-Kantian idealism in his Metaphysics (ed. B. Bosanquet, Oxford: Clarendon Press, 1887) book i §94, pp. 220–1. The passage Moore has referred to is R.V. pp. 17–19 (Bxiv–xviii, GW 109–11). 46 R.V. pp. 19–20 (Bxx, GW 112). The note occurs on Bxxii, GW 113. 48 B38, GW 157. R.V. p. 58 (B37–8, GW 157).
144
The 1898 dissertation
of space, which is their presupposition: and he then proceeds simply to repeat his ‘hypothesis’, as if it now were certain and carried conviction. It is ‘a manner of explaining the concept of space, which must be presupposed if geometrical knowledge is to be possible’.49 ‘How can there dwell in the mind’ he says ‘an outer intuition which precedes the objects themselves, and in which the conception of these objects can be determined a priori? Obviously no otherwise than insofar as it has its seat in the subject, as the formal nature of the same, whereby it is predisposed to be affected by objects and thereby to obtain immediate presentation, i.e. intuition of them: that is to say only as form of the outer sense in general. Accordingly our explanation alone renders intelligible the possibility of geometry, as a synthetic knowledge a priori.’50 No other argument for the subjectivity of space seems to be exhibited in the whole of the Aesthetic. Kant argues indeed that space and time cannot be properties of things in themselves nor yet represent their relations to one another, because ‘neither absolute nor relative determinations of things can be intuited before the existence of the things, to which they belong, and hence not intuited a priori.’51 But this goes no way to prove their subjectivity, since he has done nothing to shew that the disjunction in his question cited above is exhaustive. And again he seems to have nothing to say against those who maintain the absolute reality of space and time as subsistent, except that their assumption ‘of two everlasting and endless, self-subsistent non-things, which are there, (without being anything actual) only in order to embrace all that is actual’52 is difficult. The difficulty may be admitted; but that does not prove that Kant’s solution of it is the true one. In short, the Aesthetic may be taken as proving that space and time are a priori, not, as Leibnitz’s view entailed, mere empirical abstractions from given objects, like sensations; but in support of their subjectivity it seems to offer no better argument than that of the Preface. Kant’s argument in the Analytic is much more difficult to make out. In his Preface to the First Edition (pp. 9–10)53 he says it may be divided into two parts; the one ‘referring to the objects of the pure understanding and being meant to establish and render intelligible the objective validity of its a priori conceptions,’ the other treating ‘the pure Understanding itself in a subjective connection’. The latter, or ‘subjective deduction’, is not, he says essential to his main object: it is ‘as it were a search for the cause of a given effect’, and has therefore some resemblance to an hypothesis, though it is 49 52
B40, GW 176. B56, GW 166–7.
50
51 B42, GW 159. B41, GW 176. R.V. pp. 9–10 (Axvi–xvii, GW 103).
53
On ‘Reason’ in Kant
145
not really such. But the ‘objective deduction’, even if the other be rejected, retains its full strength, ‘to exhibit which what is said on pp. 92–93,54 can be sufficient by itself.’ In what is said on these two pages (111 and 112)55 no distinct argument to prove the mental origin of the categories occurs. But the argument, such as it is, already given in the Preface and the Aesthetic seems to be presupposed. Kant states that there are two ways in which a ‘synthetic presentation’ and its object may agree: either the object may be necessary to the existence of the presentation or vice versˆa; the first alternative gives empirical knowledge, the second a priori. He goes on to explain, that just as ‘the condition under which alone objects can be intuited, actually lies in the mind to base the objects, so far as their form is concerned’, so it is here the question whether there are not also a priori conceptions as antecedent conditions, under which alone anything can be thought as an object. In all this Kant seems to take it for granted that only if these ‘conceptions’ are present in the mind, whenever it attempts to think an object of experience, is it intelligible that all objects of experience should have certain a priori characteristics. Thus he immediately afterwards attributes Hume’s failure to find a ground for asserting causality, (which, says Kant, he recognised must be a priori) to the fact that it did not occur to him ‘that perhaps the understanding by means of these conceptions could itself be author of the experience, in which its objects are found.’56 To prove that these conceptions could alone attach to objects a priori, if they came from the mind and not from the object, is not however the object of the Transcendental Deduction. This, he says himself, is rather to shew that the categories, if they are to be a priori, must be recognised as a priori conditions of the possibility of experience. That is to say, conceptions which are implied in or necessarily connected with any empirical judgment whatsoever, are themselves necessary. But, though Kant does not here give us an argument for the subjectivity of the conceptions of the understanding, as the Preface seemed to promise, he does throughout assume this subjectivity as a means of proving what he declares to be the object of his argument – that these conceptions are necessarily implied in every empirical judgment. He might, thus, perhaps, if his argument were sound, plead for his hypothesis from its success; since it has helped him to explain how the categories should be a priori, by proving their universal implication. But this at best could only prove 54 55
These are Kant’s own page numbers, in this case A92–3, which occur in the passage quoted from the A text preface referred to in the previous footnote. 56 B127, GW 225. R.V. pp. 111–12 (B125, GW 224).
146
The 1898 dissertation
their subjectivity to be a probable hypothesis; and, as a matter of fact, Kant seems rather to have taken it for granted without inquiry. Thus his language with regard to Locke and Hume (pp. 112–113)57 betrays a curious confusion. He speaks as if Locke, though recognising these conceptions as ‘pure conceptions of the understanding’, nevertheless deduced them from experience. And similarly he says Hume saw that the conceptions ‘were not of themselves bound together in the understanding’; as if Hume had seen that they were at least in the understanding. Though, therefore, in the Preface and again in this passage, he would seem to regard it as a great discovery that these conceptions are contributed to experience by the understanding, he would seem never even to have raised a doubt that they could be regarded as other than ‘belonging to the understanding’. The question he asked was only whether ‘conceptions of the understanding’ could be abstracted from experience. And it is no wonder that, putting the question in this form, he should hardly see the need of argument that what belongs to the understanding should be contributed to experience by the understanding; but should have thought it more in need of proof that what thus belonged to the understanding should have objective validity. Accordingly, in his introductory section (p. 109),58 he tells us we have here a difficulty which did not occur in the case of space and time, namely ‘how subjective conditions of thinking are to have objective validity.’ He here simply assumes that the categories are ‘subjective conditions of thought’, and, reversing the argument we have been led to expect, finds it difficult to conceive how, being so, they should apply to objects at all; whereas in the Preface, he wished rather to shew that they were subjective, on the ground that their a priori nature would thus become intelligible. We can, therefore, hardly expect to find in the Analytic any proof of the point in question – that such conceptions as substance and attribute, cause and effect have their sources in the mind. But it will be useful to give some account of the reasoning employed, in order to see whether its results, so far as they seem valid, could not be just as well obtained without involving this hypothesis of subjectivity. The main stress in the ‘Transcendental Deduction’ lies on what Kant calls the ‘synthetic unity of apperception’. All knowledge, he says, would be impossible, unless the presentations, given through sense, could be united in a consciousness. Any judgment whatsoever implies ‘I think’ along with it; and it is only because of this necessary connection with a subject of knowledge, that we are able to make statements about our presentations. 57
R.V. pp. 112–13 (B127–8, GW 225–6).
58
R.V. p. 109 (B122, GW 222).
On ‘Reason’ in Kant
147
It is, then, because any empirical proposition must stand in relation to the unity of thought, that the conceptions of the understanding, whereby its matter is brought into relation to this unity, are involved in it. That the categories are the only concepts, whereby objective judgments can be made, Kant regards as established by the logical analysis of judgment; and he says it is just as impossible to give a reason why the understanding should have this peculiar property of being able to bring about the unity of apperception only by means of the categories, as to give a reason why space and time are the only forms of possible intuition for us (p. 123).59 On the other hand, that the unity of apperception is involved in any judgment whatsoever, he regards as a merely analytical proposition (pp. 117, 119).60 The whole deduction seems to depend on these two premises. Now that some categories are involved in every judgment seems to me certain. Whether Kant has found them all, or whether all he names are really so involved, may be left doubtful; but, at all events, they will include the two upon which subsequently the chief stress lies – substance and attribute and cause and effect. But the validity and necessity of these, as logically presupposed in empirical judgments, would seem to be capable of proof from the mere consideration of such judgments. Why the unity of apperception should be introduced into such a proof at all, does not seem clear. It could not serve any purpose, unless Kant had already assumed that the categories were conceptions by which the understanding ‘spontaneously’ arranged the matter presented to it, and were therefore, like this unity, primarily necessities of thought and not of fact. And that they are so requires proof. But, even if this were granted, Kant’s ‘analytic’ proposition seems questionable. By its ‘analytic’ character, according to the account given above, nothing can be meant but that in the very definition of knowledge reference to an ‘ego’ must be included. This, in fact, presupposes a synthetic proposition about the nature of knowledge. Now, waiving Kant’s questionable identification of the ‘ego’ with consciousness, it may be admitted that knowledge does imply consciousness. But there remains a question which Kant does not seem to have asked himself. Are we here concerned at all with knowledge? Will it not be sufficient for our purpose, if we can find out what is true? This is a question which seems vitally to affect Kant’s whole theory of the subjectivity of a priori knowledge, and therefore to be most important for a clear understanding of what he means by reason. If truth is something independent of knowledge and therefore of consciousness, no theory which 59
R.V. p. 123 (B145–6, GW 254).
60
R.V. p. 117 (B135, GW 248); R.V. p. 119 (B138, GW 249).
148
The 1898 dissertation
tries to explain the validity of necessary propositions by shewing them to be involved in knowledge or in consciousness can possibly attain its purpose. It may be a true theory, but it cannot explain that which it professes to explain. That truth is independent of our consciousness I shall attempt to shew at some length in my next chapter. It is only possible here to make some remarks with regard to Kant’s contrary supposition, which, while they presuppose my view, may at the same time help to illustrate it. I must first state that ‘knowledge’, in its definition, does properly involve a reference to truth. We cannot strictly be said to know a thing, or to have knowledge of it, unless that which we know is true. But for that very reason I think it would be desirable that the word should be used as little as possible in philosophy. For knowledge does not only imply truth, but also, as admitted above, involves a reference to consciousness. Knowing would seem to be a composite conception, denoting, on the one hand, a specific psychical state, including a distinction of subject and object, and on the other hand that the object, where the object is a proposition, should be true. Where the object is not a proposition, it may be doubted whether the term ‘knowledge’ should strictly be applied, although it undoubtedly is so in every day usage. At all events it is certainly the business of philosophy to make propositions, and if possible to prove that they are true. And my point is that to say that philosophical propositions are true logically implies nothing whatsoever with regard to consciousness, not even that they can be known; although to say that they are known would certainly imply that they are true. When, therefore, Kant argues that necessary propositions can only be explained, if they are contributed to experience by the subject, his theory may indeed be true, but it cannot possibly prove or explain their validity and therefore cannot, as he wished, furnish an answer to Hume’s scepticism. For there may fairly be urged against him the question: How can he prove the proposition, that what is necessary to consciousness is true? His answer must involve a vicious circle. This circle is, in fact, involved in his defence of his Critique as ‘a criticism of reason by itself.’61 It has often been pointed out that such a method would seem to lead to an infinite regress. For the reason which criticises ought, if reason needs criticism at all, to be subjected to criticism in its turn; and so on, ad infinitum, it being impossible ever to find a perfectly sure basis from which to start. I can see no solution of this difficulty. Dr. Caird attempts to solve it in the Introduction to 61
Moore gives no reference here, but Axi–xii (GW 101) in Kant’s first preface fit this comment.
On ‘Reason’ in Kant
149
his ‘Critical Philosophy of Immanuel Kant’62 (pp. 14–16): but his attempt appears to me to rest on that very identification of knowledge with truth, which I cannot accept. ‘We can never’ he says ‘know anything except as it is related to the conscious self within us’. With this proposition I fully concur; but when he goes on to say ‘Whatever we deal with, we are still dealing with our own consciousness of things’, I can no longer follow him. If he can speak of anything as related to our conscious self, does it not follow that that which is so related to our consciousness, is not itself, as he implies, ‘our own consciousness of things’? If we can never deal with anything but ‘our own consciousness of things’, there can be no meaning in talking of something that is only related to our consciousness. He would reply, I imagine, that there is indeed a relation between consciousness and things, only we must be careful not to understand it as an external relation (p. 15). I am not certain that I understand what this can mean. The relation between subject and object in knowledge I would certainly admit to be a unique relation – a relation not to be confused with any of the relations commonly called external: it is not the same, I take it, as that of mind to matter, which Kant sometimes calls external; nor is it the same as that of parts of space to one another; nor finally can it be a causal relation. Further than this I cannot define it: but it certainly does not seem to me to justify an implication that things are the same as our consciousness of things. I am not therefore ‘asking for a criterion within the mind of that of which the one assumed characteristic is that it is without the mind in such a sense that it cannot come into any relation with mind at all’ (p. 15). I admit that of which I desire a criterion must be able to come into some relation with mind; but that does not seem to me to justify the assertions either that it is mind or within it: it seems rather to exclude them both. Nor on the other hand am I a sceptic, for I admit some propositions to be true; indeed, I think Dr. Caird’s refutation of scepticism to be in essentials just. ‘A purely negative position is’ I admit ‘an impossibility: even a question involves an assumption; even a doubt, still more a negative conviction, must have a positive certainty behind it.’63 But it is for this very reason that I doubt that Dr. Caird’s own philosophical position with regard to consciousness; since that seems to me to offer no satisfactory answer to scepticism. I think I can deny that consciousness is the source of truth, without necessarily implying that my own denial is untrue. And no 62 63
Edward Caird, The Critical Philosophy of Immanuel Kant, 2 vols., Glasgow: James Maclehose, 1889. Moore’s references are all to volume i. Ibid., p. 5.
150
The 1898 dissertation
philosophical position seems to me safe, which [can thus be denied without inconsistency].64 I dissent therefore from Kant, in a very essential point in which Dr. Caird seems to agree with him. It has been very generally agreed that some of Kant’s work is vitiated by being merely psychological; and I am unable to see that the doctrines of the unity of apperception and of the transcendental ideality of time and space are any better than psychology. Epistemology (Erkenntnisstheorie) they might perhaps be called, since that word, by derivation at all events, implies a reference to knowledge and hence to consciousness; in fact, too, it has been used, as yet, to cover no small quantity of psychological material. But if epistemology is to be the science which investigates the nature of truth, then the above-mentioned doctrines of Kant do not fall within his epistemology. So far as Kant’s deduction aims at proving that a priori principles are valid of experience, by shewing them to be logically prior to it, his inquiry does belong to ‘epistemology’ or ‘theory of knowledge’, in the sense in which that science is distinguished from psychology. But when he tries to explain their validity, by reference to the knowing subject, this is no longer so. For an explanation of validity must, strictly speaking, be some proposition logically prior to that of which the validity is explained. In this sense Euclid’s axioms explain the validity of his propositions. If the proposed explanation is not such but consists in some proposition, like those about consciousness, not logically prior, it can at best be only a causal explanation of the origin of our knowledge of the principles in question, and cannot explain why they are true. With regard then to Kant’s Copernican revolution, so far as it rests on the arguments of the Aesthetic and Analytic, the following may be said. In the Aesthetic he only repeats the ‘hypothesis’ of the Preface to the Second Edition. His argument here is that the application of a priori judgments to presented objects is only intelligible, if one supposes the mind’s nature to be such that it can only be affected in certain specific ways by those objects. But this argument is unnecessary to prove that the judgments are a priori, since that is already assumed by Kant, in virtue of their obvious necessity, and further proved by their logical priority to empirical judgments. It cannot therefore make their validity intelligible, but can at best only offer a theory of their origin, for which there is no shadow of reason. In the Analytic, on the other hand, he assumes the mental origin of the categories; and supposing that their a priori nature is thereby rendered intelligible, he tries to deduce their objective validity. For this purpose he assumes the necessary 64
This material has been crossed out, but only in pencil, presumably later.
On ‘Reason’ in Kant
151
relation of every judgment to a knowing subject; and for this assumption there is as little reason as for that of the Aesthetic. Here again too, even if every judgment did involve the unity of apperception, that would give no ground for thinking any judgments necessary; since there is nothing to shew that our minds must always think of objects in the same way. Finally the subjective explanations of Aesthetic and Analytic alike are inconsistent with the very necessity which they attempt to establish. For they presume to deduce necessity from a mere fact, namely that our mind is so and so constituted, and this, on Kant’s own principles effectually excludes the propositions deduced from any claim to be absolutely necessary. So far as this explanation is concerned, the Critique must lose its title to a priori demonstration, and thus the very hypothesis by which Kant hoped to establish once for all a scientific philosophy, would, if it were true, render such philosophy impossible. (2) The argument which, as above stated,65 Kant supposes to be furnished by the Dialectic, is of a somewhat different nature. As briefly stated in the Preface to the Second Edition (pp. 19–20),66 it runs as follows. Reason ‘necessarily and rightfully demands the unconditioned thingsin-themselves, as condition of everything conditioned’. Now, if it be assumed that our knowledge conforms to objects, it turns out that the unconditioned cannot be thought without a contradiction; but, when it is assumed that the objects conform to our manner of representation, the contradiction disappears. The details of this argument are worked out, in the Dialectic; but only, apparently, in the section on Rational Cosmology. In the sections on Rational Psychology and Theology, Kant contents himself with showing that the supposed proofs of previous philosophers are baseless, without proving that they lead to contradictions. In the Antinomies, however, it is shewn that contradictions result from the assumption that anything is unconditioned among the objects of experience. The conclusion is therefore that the objects of experience cannot be things-inthemselves, since these latter must be unconditioned. Now, granting that contradictory propositions must be made about the objects of experience, or the world in space and time, and that hence they cannot be things in themselves, does it follow from this that they are appearances, and that their defect lies in their conformity to our faculties of knowledge? Kant thinks so, because he presupposes that truth consists in the correspondence of our knowledge to objects, and conversely that falsehood can only be explained as due to a lack of such correspondence on 65
P. 143.
66
R.V. pp. 19–20 (Bxx, GW 112).
152
The 1898 dissertation
our part (p. 86).67 But if, as seems necessary, truth must be defined quite independently of knowledge, and is indeed presupposed in the definition of the latter, then there seems no reason why falsehood either should be supposed to have its source in our minds. Falsehood may be defined as the character of a proposition which does not correspond to the nature of the world. And there is nothing to shew that the nature of the proposition is due to us who make it. Though therefore the Antinomies prove that our conception of the world, as in space and time, is not a true one; there is no reason to say that this is because it is our conception. Kant’s hypothesis, therefore, that the contradictions result from applying to mere appearances what is only true of things in themselves, is quite unjustified. They may indeed result from applying to one sort of object what is only true of another; but nothing is gained, in explanation of them, by the assumption that the one sort is appearance, or conditioned by our subjectivity. Kant’s acknowledged proofs of the subjectivity of a priori knowledge would seem, therefore, all to be equally groundless. There remains, however, one important presupposition which would seem to have influenced his thought on this point, though he does not expressly put it forward as an argument. This is his theory of sensations as given or received. The justice of this theory he seems never even to have questioned. He uses as his criterion of empirical judgments, that they are concerned with the matter of knowledge, which is in this way given or received. And if this is to be used as the mark of what are commonly called sensations, then the theory that other elements of knowledge have a mental origin, follows naturally. For if these other elements were themselves merely given to the mind, then the empirical elements could not be distinguished from them in this respect. Now Kant confesses he cannot understand how necessary propositions can be merely given; and hence he has to seek for another explanation of their origin. In answer to this form of argument for the connection of the a priori with Reason, two points may be urged. (1) That necessary propositions are themselves in a sense merely ‘given’. (2) That the matter of knowledge cannot be regarded as ‘given’ in the sense which Kant supposes. (1) With regard to the former point it is necessary to combat Kant’s often repeated statement that synthesis by the conceptions of the understanding is an ‘operation’. It is contradictory to his own principles that he should regard it as such, since he declares that ‘operation’ (Handlung) is a conception merely deduced from the category of cause (p. 101),68 and therefore 67
R.V. p. 86 (B82, GW 197).
68
R.V. p. 101 (B108, GW 214).
On ‘Reason’ in Kant
153
cannot presumably be used in explanation of that category or as logically prior to it. He states indeed that the synthesis of the understanding ‘if it be treated by itself, is nothing else than the unity of the operation, of which, as an operation, it is conscious even without sensibility,a by means of which, however, it is capable of determining the sensibility itself internally in view of the manifold, which may be given to it according to the form of sensible intuition.’69 But how it should be possible to be conscious of an ‘operation’, except through sensibility, i.e. under a schematized category, his own philosophy leaves entirely unexplained. Yet throughout the Analytic he uses words which imply this manner of regarding reason. The understanding ‘binds together’ the matter of sense to form an object; the spontaneity of thought is opposed to the receptivity of sense; ‘the power of judgment’ and ‘power of imagination’ are essential factors in the explanation of knowledge; the understanding is a ‘faculty’ (Verm¨ogen); the categories are ‘functions’ of the understanding; synthesis is always regarded as if it were a ‘putting together’ and not a mere name for the necessary connection of two disparate elements. But not only is this whole conception inconsistent with his theory of the application of the categories; it is also liable to the general objection against any attempt to base a philosophy on thought or consciousness. Even if it were true that the mind was active in respect to knowledge, yet this could only be known by a special application of the necessary propositions involved in all experience; and it could not therefore be a logical basis from which their necessity could be deduced. The Ego, whether it be regarded as individual or transcendental, must, if anything is to be known about it, be at all events an object of knowledge, and therefore there must already be involved in any proposition about it, some necessary truth, prior to any which could be discovered by the examination of it. Finally I can find no ground in myself for asserting that the proposition 2 + 2 = 4 is produced by a spontaneous operation of my mind. It presents itself to me as a fact forced upon me from outside just as much as any sensation. I am not conscious of any freedom in my cognition of it. My understanding (if I may be supposed to have any) is certainly not conscious of any act of synthesis with regard to it, whether with or without sensibility. It is no doubt true that the mind is constantly active; and that on every fresh occasion, when I think that 2 + 2 = 4, the thought is a result of its activity. a 69
My Italics. B153, GW 257.
154
The 1898 dissertation
But this does not appear to me to have any reference to the content or validity of the proposition. Is it not true, and of the same nature, whether I am thinking of it or not? Or do I create it and its truth, as often as it comes into my mind? It and all similar necessary propositions are given in the sense that they arise in the mind, without any consciousness or operation on the subject’s part; though, no doubt, they may be causally connected with his previous mental states. But neither of these facts determine their nature or validity. (2) With regard to the special way in which sensations are supposed to be given, it would appear that in the first place this is due to the supposition of their being caused by things in space. The term sensation has original reference only to the bodily organs – the senses of hearing, sight, etc. The reason why such data as colours, sounds, smells, etc. are distinguished as a special kind of element in knowledge, and one peculiarly involuntary, would seem to be that they have an obvious connection with certain motions in corresponding organs of the body, motions caused for the most part by what is external to the body. This, from the ordinary point of view, is the plainest way of distinguishing them from thoughts and propositions for which no external cause can be assigned. Obviously, however, this definition of sensation cannot be used in a philosophical analysis of the elements of knowledge, which attempts to explain the nature of the spatial world: for this definition presupposes the nature of the spatial world as already known. Yet considering how many professed philosophers have involved themselves in this hopelessly vicious circle, it is perhaps not unfair to suppose that Kant’s thought may have been influenced by it. He avoids it, however, in his express statements. With him, what ‘affects our sensibility’ is not the object in space, but an assumed object which he calls Ding-an-Sich. Sensations are, according to him, ‘given’, because they are directly due to some sort of action by the Ding-an-Sich upon our mind; whereas the a priori notions are contributed solely by our mind, only on occasion of sensation. For this theory he has no arguments to offer: it has only ‘never occurred to him to doubt it’.70 The relation between sensations and the Ding-an-Sich, must, as he describes it, be conceived under the category of cause, and yet, according to him it cannot be, since the Ding-an-Sich is no object of experience. It seems doubtful whether he would ever have conceived this contradictory notion at all, had he not started with the assumption that sensations were given or received – mere 70
Moore gives no reference for this passage, but it could be an allusion to the opening sentences of the introduction to the B edition (B1, GW 136).
On ‘Reason’ in Kant
155
presentations (Vorstellungen). And this assumption is probably due to the English philosophers, with whom it undoubtedly originated in the vicious theory that sensations were caused by things in space (as in Locke’s theory of primary and secondary qualities). Kant, since he set out to explain the spatial world, saw that he could not adopt this theory, but he seems nevertheless not to have rejected that which was a mere corollary from it – the receptivity of sense, and hence to have been forced to his own unjustified theory of the Ding-an-Sich.a If it be supposed that sensations are immediately cognised as something given, and that this is the basis of Kant’s theory, I can only repeat that I cannot find myself to be more passive with regard to them than with regard to any other object of consciousness. In the relation of subject and object, the object always appears to me as something merely ‘presented’, merely there, not as something produced by the subject, which contemplates it, and this equally whether it be a sensation, a thought, or a feeling. Moreover there is always a difficulty, as to whether there can be any meaning in talking of an element presented as given. If the ‘givenness’ is to be included with it as object of knowledge, it would seem it must be synthetically related to it; and this relation presupposes those necessary relations which are supposed not to be given. In order, therefore, that anything should be presented as given, it is necessary that a necessary proposition should be given along with it. So that if the ‘givenness’ of sensations is to be based on the evidence of immediate introspection, this argument implies that necessary connections are also given; and thus all point is taken from the desired conclusion. As a final result of all these considerations, it would appear, therefore, that though Kant does connect the notion of Reason with that of a priori propositions, he has no valid reason for doing so. His conception of Reason is an essentially composite one, combining in itself two distinct characteristics, the connection of which there is no reason for asserting; and his assertion of this connection involves his philosophy in hopeless difficulties and inconsistencies. His whole treatment of Reason seems to involve a paralogism of the same kind as that which he exposes in Rational Psychology. He strenuously denies, indeed, that the unity of consciousness is to be conceived as in any way substantial. But when he describes it as the source of a priori propositions, as affected in certain ways by the data of sense, and even as acting in a specific manner upon those data, he would a
By strictures on the Ding-an-Sich in this sense, as cause of sensations, nothing is implied against it in the sense, in which Kant uses it in the Dialectic, = the unconditioned.
156
The 1898 dissertation
seem to be using that very conception the legitimacy of which he denies. He distinguishes ‘mind’ in the transcendental sense, from mind as determined in time, the object of the inner sense; he warns us against the confusion of the ‘transcendental’ with the ‘empirical’ Ego: but he can only conceive the former under categories, which he would restrict to the latter, and he can give no valid reason why the existence of any but an empirical ego should be assumed. There is, in short, no reason for supposing that such a science as has been called ‘Transcendental Psychology’ in distinction from empirical psychology, does exist; or for regarding ‘Reason’ as other than an object of empirical psychology. It is attempted to base the distinction by asserting that Transcendental Reason is a condition for the possibility of knowledge, as indeed the word ‘transcendental’ implies. But when the ambiguities of this description are cleared away, all reason for the distinction would seem to disappear. By ‘knowledge’ what is meant? If ‘truth’, then it is difficult to see that there can be any other condition for a true proposition than some other true proposition. If the mere process of cognition, then does not empirical psychology investigate the conditions for the possibility of this? A similar ambiguity is involved in the word ‘condition’. In what sense a ‘condition’? If an existent be meant, upon the existence of which the existence of something depends, then condition is equivalent to ‘cause’, and both reason and knowledge must be conceived under the category of substance, as in empirical psychology. But if a logical condition be meant, then it must be some true proposition from the truth of which the truth of another can be inferred. If any third kind of condition can be pointed out, then, no doubt, the whole of this criticism must fall to the ground. Meanwhile, it seems inevitable to regard the ‘Ego’, as ‘transcendental condition of knowledge’, as merely a confusion of the two, since it is undoubtedly described in language which is suitable to both. Kant’s notion of Reason seems therefore to be a confused one, since he regards it partly as something which can act and be acted upon, partly as a logical presupposition of true propositions. This confused notion, which he has partly assumed, partly adopted on grounds appearing in the Aesthetic and Analytic, he already applies in a striking manner in the Dialectic. He there speaks of Reason as having ‘wants’ and ‘demands’. This would seem to be the purest personification of Reason. And yet, even here, where she is no longer the source of truths but merely of ideals, Kant still allows her to influence his opinion as to what is true. Her demand for the ‘unconditioned’ is justified in such a sense that we should really be bound to hold contradictory propositions for true, unless a world which satisfied her demands, were possible. The antinomies cannot be solved,
On ‘Reason’ in Kant
157
except on the supposition that the unconditioned really exists, though not in the spatio-temporal world. The confusion of Kant’s conception here appears as clearly as possible. If the unconditioned is really no more than a ‘want’, then there are no antinomies; for the fact that I want to fly is not contradictory to the fact that I cannot fly. But if there are antinomies requiring solution, then that must be because it is true, and not merely desired or desirable, that the world must be unconditioned. Hence Reason here too is regarded as a source of truth and not merely as a creature with wants. Only, as in the earlier part of the Critique the former characteristic was emphasized, so here is the latter. This conception of Reason in the Dialectic makes the transition to Kant’s notion of a Pure Practical Reason more easily intelligible. Reason’s a priori knowledge ‘can’, he tells us, ‘be related to its object in two ways: it may either merely determine this object and its conception (which must be given from some other source), or it may also make it actual. The first is theoretical, the second practical knowledge of Reason’ (Pref. to 2nd ed., p. 14).71 It would at first sight appear surprising, on any view of Reason, that the actualisation of objects should be regarded as part of its cognitive functions. But, when as we have seen, Reason in its speculative use is already supposed actually to produce in the total object of knowledge, all that part which may be described as formal, it is no longer a very big step to conceive it as producing the entire object. If any cognition be regarded as in part at least a spontaneous production of its object, there is no apparent reason why the production of a complete object should not be regarded as cognition. Indeed it would seem to be a mere prejudice on Kant’s part which prevents him from regarding all objects as mere creations of Reason. For, as we have seen, there is no reason to distinguish any one element as more ‘given’ than another. Kant does, however, maintain the distinction between Speculative Reason as ‘determining’ a given object, and Practical Reason as producing an object. This distinction, though it appears to be an utterly groundless one, gives him a verbal formula for the difference between the Understanding and Practical Reason. But the same distinction would seem unable to supply him with a contrast between Speculative Reason, in the restricted sense of the Dialectic, and Pure Practical Reason. For Pure Practical Reason does not produce an object, but merely gives a command to which actions ought to conform. In this respect, then, of being primarily the source of a mere Idea, it would seem indistinguishable from the Reason of the 71
R.V. p. 14 (Bix–x, GW 107).
158
The 1898 dissertation
Dialectic. Both alike, in Kant’s description of them, demand something. To this demand, in the case of Speculative Reason, the title of knowledge is denied; and the same restriction ought therefore to apply to Practical Reason. Why either, then, should be called Reason, and why they should be distinguished from one another is not apparent. As faculties for demanding what is not given, they might be distinguished from the faculty for determining the given or for producing the given (though that is a contradiction in terms) but not from one another. The real distinction between them concerns solely the nature of the propositions to which they are supposed to give rise, and would therefore give no reason for distinguishing them as different faculties, except on the absurd presupposition that there must be faculties as the causes of the propositions, and therefore, since the propositions are different, different faculties cause them. As was pointed out with regard to the Dialectic, Reason’s so-called demand is there a principle, which is objectively valid, though not capable of application to the spatio-temporal world. On the other hand, the command of the Practical Reason, is in fact a statement not that something is, but that something ought to be or that something is good. Kant recognises, indeed, that it is with these two regions of ‘what is’ and ‘what ought to be’ that his two Critiques deal. But owing to his unfortunate supposition that the origin of necessary propositions must be explained, and that this is only to be found in Reason itself, he attempts to deduce the principles of Metaphysics and Ethics respectively from a different faculty of Reason, whereas in reality it is the necessary difference of these principles themselves which alone gives ground, and that a wholly insufficient one, for the assumption of a difference of Reason. Kant’s notion of Practical Reason embraces accordingly not only all the confusions to be found in his conception of Reason in general but also new ones peculiar to itself. (1) Reason is regarded throughout Kant’s Critical Philosophy, as if it were not only the source of a priori propositions but also explained their validity. They are valid, because they proceed from Reason. But in fact a proposition can only be valid, because another proposition is valid: a logical condition is the only possible condition of validity. ‘Reason’ therefore usurps the functions of a logical condition, but at the same time, since it is never identified with a proposition, it is regarded as some sort of substantial entity, and therefore as a causal condition of the a priori. (2) It is already regarded in the Analytic as something active. This notion of it naturally follows from the conception of it as a causal condition of knowledge. Kant’s statement that he is concerned not with the question
On ‘Reason’ in Kant
159
how experiences arises, but with what lies in it (Prol: §21a),72 is directly contradictory to his universal treatment of Reason. That he should think he does not treat the former question, would seem to be due to his failure to appreciate that there is no difference in principle between a simultaneous and a precedent condition, if only the condition be not a logical one but a causal one. He would seem to regard Reason as a condition of knowledge, in the sense in which the constitution of a body is a condition of its response to a given impulse. But this will not defend him from the charge of making empirical psychology his business. And when he speaks of Reason as actually operating, it would seem difficult to see how he could rebut the charge of making it a temporally antecedent cause of knowledge. (3) When he speaks of certain principles as demanded by Reason, he would seem obviously to regard it as their temporal antecedent. And this personification of Reason is rendered more unquestionable by the fact that he does not regard the principles in question as presuppositions of experience, and hence ‘demands’ in the sense of postulates. At the same time his express denial of objective validity for the principles in question, though they have their source in Reason, is contradictory to his explanation of the objective validity of the Forms of Intuition and the Categories, as due to their origin in Reason. This stricture is not met by reference to his theory that the latter have objective validity only in application to a given matter. For he certainly derives their objectivity, in the sense of their validity as necessary propositions, from Reason alone; even though he also gives other explanations of it. On the contrary his use of the principles of Reason in the Dialectic furnishes one refutation of his theory of the limitation of knowledge to given objects. (4) In the Practical Reason, he completes the confusion of Reason with Will. As we have seen, he regards Practical Reason in general as actually producing objects. And whereas the a priori principles of Practical Reason ought, on this view of it, to be merely the principles to which the production of objects conforms i.e. the principles of causality, he only regards them as the principles to which the production of objects ought to conform. He thus utterly destroys any possibility of connection between Practical Reason in general and Pure Practical Reason, although at the same time he generally uses the specific volitional nature of the former as furnishing the distinction between Pure Speculative and Pure Practical Reason. Practical Reason, then, combines the following discordant functions. (1) Some necessary a priori proposition about what is good is needed as 72
Prolegomena to Any Future Metaphysics, §21a (AK 4: 304; trans. Hatfield, 55–6).
160
The 1898 dissertation
a fundamental principle of Ethics. Practical Reason, according to Kant, is what gives this. It must, then, be the source of a proposition and at the same time (2) the condition of its validity. (3) It not only thus furnishes a reason, why a thing should be done, but is also itself the reason, or cause, why such a thing is done. (4) Being once made cause not only of a principle, but also of action in accordance with a principle, it is also in its widest sense cause of action contrary to such a principle – Against this monstrous conception we have to urge (1) That there is no reason for ascribing the fundamental principle of Ethics, to any entity whatever. (2) Such an ascription can only be based on an empirical proposition a posteriori, and if the ascription is to be regarded as establishing its validity, its validity is destroyed by being made dependent on the truth of such a proposition. (3) If the principle be ascribed to Reason, it can only be to Reason in the same sense in which Reason is the source of the fundamental principles of Metaphysics. There is no reason for any distinction between Practical Reason and Speculative Reason. (4) Even if Reason were cause of the principle [it]73 would not thereby become cause of action either in accordance with [or contrary to the] principle; any more than the principle of causality, is [cause of a] natural event.74 73 74
The page is torn at this point and we have reconstructed the words here and below. Moore has deleted the phrase ‘which conforms to that principle’ which originally concluded this sentence. A further point ‘(5)’ then followed, but this was also deleted and the paper is so torn at this point that it is uncertain what the point was.
chapter i i
Reason
The present chapter is intended to expound and support the ultimate philosophical position, which was presupposed in the last chapter. It was there attempted to express several of Kant’s fundamental propositions in other words than those which Kant himself uses. This attempt must have been nugatory, unless some definite meaning had been attached to the words thus substituted for Kant’s. But it is impossible to attach a definite meaning to words, without some definite view as to what are the ultimate facts or truths, which alone can be meant by any words. A word can have no meaning at all, unless its meaning can be expressed in terms denoting some actual fact or relation. It was necessary therefore to analyse Kant’s thought into elements, which he perhaps did not recognise as ultimate, but which seemed to be the only ones to which any idea could have reference. In the course of this analysis more criticism was introduced, than was perhaps necessary for the purposes of mere exposition; but, so long as there is no agreement as to the ultimate data of philosophy, any fruitful exposition would seem to require some defence of the data presupposed as well against those of the philosopher expounded as against other possible objectors. Failing such direct contravention of accepted views, a new view can hardly be made even intelligible: and some view, it must be repeated, is presupposed in any exposition. The present chapter will therefore treat more fully some of the questions raised in the last; and, upon that basis, will proceed to discuss the proper meaning of ‘rational’ in Ethics. In the preceding investigation the word ‘proposition’ was used as an ultimate term. This word, it may be admitted, does naturally imply a mental formulation, if not an actual expression in words. Both these implications were meant to be entirely excluded, and the word was nevertheless used, because there seems no better term to express the meaning intended. ‘Judgment’, which is also sometimes used, seems even worse, since it not only denotes a mental event, and hence implies activity still more openly, but is also commonly used as the name of a mental faculty. Our object will 161
162
The 1898 dissertation
be now to show that, whatever name be given to it, that which we call a proposition is something independent of consciousness, and something of fundamental importance for philosophy. ‘Truth and falsehood,’ says Mr. Bradley (Logic p. 2),75 ‘depend on the relation of our ideas to reality’. And he immediately goes on to explain that, in this statement, ‘ideas’ must not be understood to mean mere ‘states of my mind’. The ideas, he says, on the relation of which to reality truth depends, are76 . . . ‘mere ideas, signs of an existence other than themselves,’ and this aspect of them must not be confused either with their existence in my mind or with their particular character as so existent, which may be called their content. ‘For logic, at least,’ he says, ‘all ideas are signs’ (p. 5);77 and ‘A sign is any fact that has a meaning,’ while ‘meaning consists of a part of the content (original or acquired) cut off, fixed by the mind, and considered apart from the existence of the sign’ (p. 478 ). But Mr. Bradley himself does not remain true to this conception of the logical idea as the idea of something. As such, indeed, it is only the psychological idea, related, indeed, to that which it signifies, but only related to it. Hence he finds it necessary, later, to use ‘idea,’ not of the symbol, but of the symbolised. Ideas, as meanings, not as ‘facts, which have a meaning,’ ‘are,’ he says (p. 8),79 ‘the ideas we spoke of, when we said “Without ideas no judgment”.’ And he proceeds to show that ‘in predication we do not use the mental fact, but only the meaning’; although, where he did say ‘Without ideas no judgment,’ his words were ‘we cannot judge until we use ideas as ideas. We must have become aware that they are not realities, that they are mere ideas, signs of an existence other than themselves.’ It would seem plain, then, that there his doctrine was that we do, in predication, use the mental fact, though only as a sign; whereas here his doctrine is that we do not use the mental fact, even as a sign, but only that which it signifies. This important transition he slurs over with the phrase: ‘But it is better to say the idea is the meaning’. The question is surely not of which is ‘better to say,’ but which is true. Now to Mr. Bradley’s argument that ‘the idea in judgment is the universal meaning’ I have nothing to add. It appears to me conclusive, as against those, of whom there have been too many, who have treated the idea as a mental 75 76
77
F. H. Bradley The Principles of Logic, London: Oxford University Press, 1883. This is the bottom of p. 2 of the MSS; pp. 3–5 are missing, presumably because they were used for the composition of ‘The Nature of Judgment’ which starts with the passages here quoted from Bradley. We have inserted here the material from ‘The Nature of Judgment’ pp. 176–7 which directly continues the passage quoted here. For further discussion, see Editors’ introduction pp. lxxv–lxxx. 78 Ibid., p. 4. 79 Ibid., p. 8. Bradley, Principles of Logic, vol. i, p. 5.
Reason
163
state. But he seems to me to be infected by the same error as theirs, alike in his preliminary failure to distinguish clearly whether it is the symbol or the symbolised of which he is speaking, and in his final description of the ‘idea, as meaning,’ when he has definitely decided in its favour. ‘A meaning,’ he says, as we saw above, ‘consists of a part of the content (original or acquired) cut off, fixed by the mind, and considered apart from the existence of the sign.’ And again, ‘an idea, if we use idea of the meaning, is neither given nor presented, but is taken’ (p. 8).80 If indeed ‘the universal meaning’ were thus simply a part of the content of our own ideas, as mental states, and that, too, a part ‘cut off’ by our own minds, it would be intelligible that ‘truth and falsehood’ should still be said to ‘depend on the relation of our ideas to reality’. It will be our endeavour to show, on the contrary, that the ‘idea used in judgment’ is not a part of the content of our ideas, nor produced by any action of our minds, and that hence truth and falsehood are not dependent on the relation of our ideas to reality. I shall in future use the term ‘concept’ for what Mr. Bradley calls a ‘universal meaning’; since the term ‘idea’ is plainly full of ambiguities, whereas ‘concept’ and its German equivalent ‘Begriff’ have been more nearly appropriated to the use in question. There is, indeed, a great similarity between Kant’s description of his ‘Begriff,’ and Mr. Bradley’s of his ‘logical idea’. For Kant, too, it is the ‘analytical unity of consciousness’ which makes a ‘Vorstellung’ or ‘idea’ into a ‘conceptus communis’ or ‘gemeinsamer Begriff’ (R.V., p. 116 n.).81 It is our object to protest against this description of a concept as an ‘abstraction’ from ideas. Mr. Bradley’s doctrine, as above sketched, presupposes that, when I have an idea (Vorstellung) of something, that something is itself part of the content of my idea. This doctrine, for the present, I am ready to admit; my question now is whether, when I have an idea of something, that something must not also be regarded as something other than part of the82 content of my idea. The content of an idea is, Mr. Bradley tells us, what the idea is; it is ‘a character 80 82
81 B134, GW 247. Ibid. At this point we revert to page 6 of Moore’s manuscript, which survives complete, since its first sentence replicates the first sentence of the passage on p. 177 of ‘The Nature of Judgment’ which directly follows the material taken above from ‘The Nature of Judgment’. Beyond this point there is some divergence between the manuscript and the text of ‘The Nature of Judgment’, and it is clear that Moore rewrote the passage on page 6 when preparing ‘The Nature of Judgment’. We reproduce here the relevant passage from ‘The Nature of Judgment’ for the purpose of comparison.
. . . content of my idea. The content of an idea is, Mr. Bradley tells us, what the idea is; it is ‘a character which is different or distinguishable from that of other’ ideas, treated as mental facts. Now, before I can judge at all on Mr. Bradley’s theory, a part of this character must have been ‘cut off and fixed by the mind’. But my question is, whether we can thus cut off a part of the character of our ideas, and attribute that part to something else, unless we already know, in part at least, what is the character of the idea from
164
The 1898 dissertation
which is different or distinguishable from that of other’ ideas, treated as mental facts. Now in order to know what the content of any idea is, it is necessary, according to Mr. Bradley’s Theory, that I should have an idea of my idea, and should have cut off part of the content of that secondary idea and attributed it to the first idea, whose content is to be known. But similarly it is quite impossible that I should know what the content of my secondary idea is, until I have made it in its turn the object of a third idea, by taking part of this tertiary content. And thus this process would seem to go on ad infinitum without even culminating in knowledge at all. For it is difficult to suppose that knowledge can be explained as the attribution of a part of a content, of the whole of which I am ex hypothesi utterly ignorant. The impossibility of psychological knowledge would then seem to follow from the theory in question; and with it the impossibility of making the theory. For a part of the content of an idea cannot be less an object of psychology than the content of which it is a part; and the theory itself, being a judgment, involves such parts of content. I can hardly cut off and fix a part, still less predicate it of something else, until I know the83 [the character of the idea from which we are to cut off the part in question.] Mr. Bradley’s theory presupposes that I may have two ideas, that have a part of their content in common; but he would at the same time compel us to describe this common part of content as part of the content of some third idea. But what is gained by such a description? If the part of content of this third idea is a part only in the same sense, as the common part of the other two is a part of each, then I am offering an explanation which presupposes that which was to be explained. Whereas if the part, which is used in explanation, which we are to cut off the part in question. If not, then we have already made a judgment with regard to the character of our idea. But this judgment, again, requires, on Mr. Bradley’s theory, that I should have had an idea of my idea, and should have already cut off a part of the content of that secondary idea, in order that I may make a judgment with regard to the character of the primary idea that is in question. And similarly it is quite impossible that I should know what the content of my secondary idea is, until I have made it in its turn the object of a third idea, by taking part of this tertiary content. And so on ad infinitum. The theory would therefore seem to demand the completion of an infinite number of psychological judgments before any judgment can be made at all. But such a completion is impossible; and therefore all judgment is likewise impossible. It follows, therefore, if we are to avoid this absurdity, that the ‘idea used in judgment’ must be something other than a part of the content of any idea of mine. 83
Because page 6 ends in mid-sentence in a way which does not join up with the comparable passage in ‘The Nature of Judgment’ it is not clear exactly how the text of the 1898 dissertation continued after the end of page 6, and the reconstruction here is inevitably speculative; the phrase in square brackets is taken from the relevant part of ‘The Nature of Judgment’, but somewhat out of context. None the less it is reasonable to think that the substantive changes are limited to these paragraphs and that the material which follows on pages 178–84 of ‘The Nature of Judgment’ came relatively unchanged from the 1898 dissertation. For it is in these pages that Moore presents his own theory of propositions, concepts and existents which we know he had presented in the 1898 dissertation. For further discussion, see editors’ introduction pp. lxxviii–lxxix.
Reason
165
is a part in the only sense which will make my explanation significant, i.e., an existent part, then it is difficult to see how that which belongs to one idea can also come to belong to other ideas and yet remain one and the same. In short, the idea used in judgment is indeed a ‘universal meaning’; but it cannot, for that very reason, be described as part of the content of any psychological idea whatever. These difficulties, which are of the same nature as the famous tr©tov nqrwpov urged against the hypostasised Platonic ideas, inevitably proceed from trying to explain the concept in terms of some existent fact, whether mental or of any other nature. All such explanations do in fact presuppose the nature of the concept, as a genus per se, irreducible to anything else. The concept is not a mental fact, nor any part of a mental fact. Identity of content is presupposed in any reasoning; and to explain the identity of content between two facts by supposing that content to be a part of the content of some third fact, must involve a vicious circle. For in order that the content of the third fact may perform this office, it must already be supposed like the contents of the other two, i.e., having something in common with them, and this community of content is exactly what it was proposed to explain. When, therefore, I say ‘This rose is red,’ I am not attributing part of the content of my idea to the rose, nor yet attributing parts of the content of my ideas of rose and red together to some third subject. What I am asserting is a specific connexion of certain concepts forming the total concept ‘rose’ with the concepts ‘this’ and ‘now’ and ‘red’; and the judgment is true if such a connexion is existent. Similarly when I say ‘The chimera has three heads,’ the chimera is not an idea in my mind, nor any part of such idea. What I mean to assert is nothing about my mental states, but a specific connexion of concepts. If the judgment is false, that is not because my ideas do not correspond to reality, but because such a conjunction of concepts is not to be found among existents. With this, then, we have approached the nature of a proposition or judgment. A proposition is composed not of words, nor yet of thoughts, but of concepts. Concepts are possible objects of thought; but that is no definition of them. It merely states that they may come into relation with a thinker; and in order that they may do anything, they must already be something. It is indifferent to their nature whether anybody thinks them or not. They are incapable of change; and the relation into which they enter with the knowing subject implies no action or reaction. It is a unique relation which can begin or cease with a change in the subject; but the concept is neither cause nor effect of such a change. The occurrence of the relation has, no doubt, its causes and effects, but these are to be found only in the subject.
166
The 1898 dissertation
It is of such entities as these that a proposition is composed. In it certain concepts stand in specific relations with one another. And our question now is, wherein a proposition differs from a concept, that it may be either true or false. It is at first sight tempting to say that the truth of a proposition depends on its relation to reality; that any proposition is true which consists of a combination of concepts that is actually to be found among existents. This explanation was indeed actually used above (p. 179),84 as a preliminary explanation. And it may be admitted that propositions with which this is the case are true. But if this constituted the truth of a proposition, concepts too might in themselves be true. Red would be a true concept, because there actually are red things; and conversely a chimera would be a false concept, because no such combination either has been, is, or will be (so far as we know) among existent things. But the theory must be rejected as an ultimate one, because not all true propositions have this relation to reality. For example 2 + 2 = 4 is true, whether there exist two things or not. Moreover it may be doubted here whether even the concepts of which the proposition consists, can ever be said to exist. We should have to stretch our notion of existence beyond intelligibility, to suppose that 2 ever has been, is, or will be an existent. It would seem, in fact, from this example, that a proposition is nothing other than a complex concept. The difference between a concept and a proposition, in virtue of which the latter alone can be called true or false, would seem to lie merely in the simplicity of the former. A proposition is a synthesis of concepts; and, just as concepts are themselves immutably what they are, so they stand in infinite relations to one another equally immutable. A proposition is constituted by any number of concepts, together with a specific relation between them; and according to the nature of this relation the proposition may be either true or false. What kind of relation makes a proposition true, what false, cannot be further defined, but must be immediately recognised. And this description will also apply to those cases where there appears to be a reference to existence. Existence is itself a concept; it is something which we mean; and the great body of propositions, in which existence is joined to other concepts or syntheses of concepts, are simply true or false according to the relation in which it stands to them. It is not denied that this is a peculiarly important concept; that we are peculiarly anxious to know what exists. It is only maintained that existence is logically subordinate to truth; that truth cannot be defined by a reference to existence, but existence only by a reference to truth. When I say ‘This paper exists,’ I must require that this proposition be true. If it is not true, it is unimportant, and I can have no interest in it. But if it is true, 84
‘The Nature of Judgment’, p. 179 = p. 160 here.
Reason
167
it means only that the concepts, which are combined in specific relations in the concept of this paper, are also combined in a specific manner with the concept of existence. That specific manner is something immediately known, like red or two. It is highly important, because we set such value upon it; but it is itself a concept. All that exists is thus composed of concepts necessarily related to one another in specific manners, and likewise to the concept of existence. I am fully aware how paradoxical this theory must appear, and even how contemptible. But it seems to me to follow from premisses generally admitted, and to have been avoided only by lack of logical consistency. I assume Mr. Bradley’s proof that the concept is necessary to truth and falsehood. I endeavour to show, what I must own appears to me perfectly obvious, that the concept can consistently be described neither as an existent, nor as part of an existent, since it is presupposed in the conception of an existent. It is similarly impossible that truth should depend on a relation to existents or to an existent, since the proposition by which it is so defined must itself be true, and the truth of this can certainly not be established, without a vicious circle, by exhibiting its dependence on an existent. Truth, however, would certainly seem to involve at least two terms, and some relation between them; falsehood involves the same; and hence it would seem to remain, that we regard truth and falsehood as properties of certain concepts, together with their relations – a whole to which we give the name of proposition. I have appealed throughout to the rules of logic; nor, if any one rejects these, should I have much to fear from his arguments. An appeal to the facts is useless. For, in order that a fact may be made the basis of an argument, it must first be put in the form of a proposition, and, moreover, this proposition must be supposed true; and then there must recur the dilemma, whether rules of logic are to be accepted or rejected. And these rules once accepted, would seem themselves to offer a confirmation of our theory. For all true inference must be inference from a true proposition; and that the conclusion follows from the premiss must again be a true proposition: so that here also it would appear that the nature of a true proposition is the ultimate datum. Nor is an appeal to the ‘matter’ of the proposition more useful than the former appeal to the facts. It may be true that this matter is given in sensation, or in any other conceivable way. We are not concerned with its origin, but with its nature; and its nature, if it is to enter into a true proposition, must, we agree with Mr. Bradley, be the nature of a concept and no other: and then the old conclusions follow. Nor, finally, is a vicious circle involved in our own attempt to establish conclusions with regard to truth, by rules of logic in which that conception is presupposed. For our conclusion is that truth is itself a simple concept; that it is logically prior to any proposition. But a vicious circle occurs only where a proposition is
168
The 1898 dissertation
taken as prior to a concept, or a more complex proposition (one involving more concepts) as prior to one which is more simple. Valid logical processes would seem to be of two kinds. It is possible to start from a complex proposition and to consider what propositions are involved in it. In this case the latter must always be more simple than the former; and they may be true; although the former is false. Or it is possible to start from a more simple proposition and to deduce one that is more complex, by successive additions of concepts; which is the properly deductive procedure exhibited in the propositions of Euclid: and in this case the premiss must be true, if the conclusion is so. It may be well to state that both procedures are synthetic, in the sense that the results arrived at are different from the premisses, and merely related to them. In a vicious circle, on the other hand, the two procedures are confused. A result arrived at by the former of the two processes just described, is regarded as involving the truth of its premiss. Thus, when we say that the conceptual nature of truth is involved in logical procedure, no vicious circle is committed, since we do not thereby presuppose the truth of logical procedure. But when an existent is said to be involved in truth, a vicious circle is committed, since the proposition ‘Something is true,’ in which ‘Something exists’ is supposed to be involved, must itself be true, if the latter is to be so. It seems necessary, then, to regard the world as formed of concepts. These are the only objects of knowledge. They cannot be regarded fundamentally as abstractions either from things or from ideas; since both alike can, if anything is to be true of them, be composed of nothing but concepts. A thing becomes intelligible first when it is analysed into its constituent concepts. The material diversity of things, which is generally taken as starting-point, is only derived; and the identity of the concept, in several different things, which appears on that assumption as the problem of philosophy, will now, if it instead be taken as the starting-point, render the derivation easy. Two things are then seen to be differentiated by the different relations in which their common concepts stand to other concepts. The opposition of concepts to existents disappears, since an existent is seen to be nothing but a concept or complex of concepts standing in a unique relation to the concept of existence. Even the description of an existent as a proposition (a true existential proposition) seems to lose its strangeness, when it is remembered that a proposition is here to be understood, not as anything subjective – an assertion or affirmation of something – but as the combination of concepts which is affirmed. For we are familiar with the idea of affirming or ‘positing’ an existent, of knowing objects as well as propositions; and the difficulty hitherto has been to discover wherein the two processes were akin. It now appears that perception is to be regarded philosophically as the cognition of an existential proposition; and it is thus apparent how it can furnish a basis
Reason
169
for inference, which uniformly exhibits the connexion between propositions. Conversely light is thrown on the nature of inference. For, whereas it could not be maintained that the conclusion was only connected with the premisses in my thoughts, and that an inference was nothing, if nobody was making it, great difficulty was felt as to the kind of objectivity that belonged to the terms and their relation, since existence was taken as the type of objectivity. This difficulty is removed, when it is acknowledged that the relation of premisses to conclusion is an objective relation, in the same sense as the relation of existence to what exists is objective. It is no longer necessary to hold that logical connexions must, in some obscure sense, exist, since to exist is merely to stand in a certain logical connexion. It will be apparent how much this theory has in common with Kant’s theory of perception. It differs chiefly in substituting for sensations, as the data of knowledge, concepts; and in refusing to regard the relations in which they stand as, in some obscure sense, the work of the mind. It rejects the attempt to explain ‘the possibility of knowledge,’ accepting the cognitive relation as an ultimate datum or presupposition; since it maintains the objections which Kant himself urged against an explanation by causality, and recognises no other kind of explanation than that by way of logical connexion with other concepts. It thus renounces the supposed unity of conception guaranteed by Idealism even in the Kantian form, and still more the boasted reduction of all differences to the harmony of ‘Absolute Spirit,’ which marks the Hegelian development. But it is important to point out that it retains the doctrine of Transcendentalism. For Kant’s Transcendentalism rests on the distinction between empirical and a priori propositions.85 ,86 85
At this point ‘The Nature of Judgment’ (p. 183, bottom) continues with the following passage: This is a distinction which offers a striking correspondence to that between the categorical and hypothetical judgments; and since one object of this paper is to combat the view which inclines to take the categorical judgment as the typical form, and attempts in consequence to reduce the hypothetical judgment to it, it will not be out of place to discuss Kant’s distinction at some length.
86
We take it that this sentence was written specially for ‘The Nature of Judgment’ in order to bridge the transition from the preceding material, which comes from chapter ii of the 1898 dissertation, and the next pages, which come from chapter i. For in ‘The Nature of Judgment’ (p. 184) the next paragraph begins ‘Kant himself offers two marks by which an a priori judgment may be distinguished’ and continues in a way which almost exactly resembles the passage in chapter i of the 1898 dissertation which precedes the gap in the text there (see p. 134). We therefore used the material which follows this paragraph on pages 184–8 of ‘The Nature of Judgment’ to fill the gap in chapter I left by the missing pages of the manuscript for that chapter. The end of that material is indicated by the fragment which now follows (see the next note). For further discussion see the editors’ introduction p. lxxix. Although it has been crossed out, the following passage which fills the lower half of an unnumbered page of chapter ii (the top half has been cut off ) fits well at this point. It closely resembles the following paragraph of ‘The Nature of Judgment’ (p. 189):
170
The 1898 dissertation
[It] was pointed out in the last chapter how this distinction could be based. The line of division there drawn was between propositions involving empirical concepts, and those which involved none such; and an empirical concept was defined as one which could exist in an actual part of time. This division was necessary in order to include all the various kinds of propositions which Kant includes under the term empirical, many of which involve a priori concepts. At the same time it was said that, if the division were to be based on propositions not on concepts, pure existential propositions would alone be regarded as empirical. These do indeed obviously form the basis of the other division; for a concept cannot be known as one which could exist in time except on the ground that it has so existed, is existing, or will exist. It is now necessary to carry our analysis87 . . . But we have now to point out that even existential propositions have the essential mark which Kant assigns to a priori propositions – that they are absolutely necessary. The distinction of time was said to be ultimate for an existential proposition. If this is so, it is obvious that necessary propositions, of the kind which Kant endeavours to establish in the Aesthetic, are involved in them. It was pointed out that a pure existential proposition could only assert the existence of a simple concept; all others involving the a priori concepts of substance and attribute. If now we take the existential proposition ‘Red exists,’ we have an example of the type required. It is maintained that, when I say this, my meaning is that the The line of division, therefore, upon which Kant’s Transcendentalism is based, would seem to fall between propositions involving empirical concepts and those which involve none such; and an empirical concept is to be defined, not as a concept given by experience, since all concepts are so given, but as one which can exist in an actual part of time. This division is necessary in order to include all the various kinds of propositions which Kant includes under the term empirical, many of which involve a priori concepts. If the division were to be based on the nature of the propositions, as such, as Kant pretends to base it, we saw that pure existential propositions alone could be thought to have a claim to form a class by themselves, as empirical propositions. These do indeed obviously form the basis of the other division; for a simple concept cannot be known as one which could exist in time, except on the ground that it has so existed, is existing, or will exist. The fact that the surviving passage has been crossed out is readily explained by the reference there to ‘the last chapter’ which was obviously inappropriate in the context of ‘The Nature of Judgment’; the crossing out indicates that Moore had marked this passage as one which needed some attention. This also explains why this fragmentary passage was separated from the pages he used when composing ‘The Nature of Judgment’, and thus why this fragment survives. As a consequence of all this we have inferred that the material which follows this paragraph in ‘The Nature of Judgment’ is based on the material which, in chapter ii, followed the corresponding paragraph there – i.e. the fragment above. Hence we have used this part of ‘The Nature of Judgment’ (pp. 189–192) to reconstruct the next missing pages of the 1898 dissertation. 87
Quite how the end of the surviving fragment (‘It is now necessary to carry our analysis’) connects with the material imported from ‘The Nature of Judgment’ at this point is uncertain. It may well be that there are a few lines missing here. But we doubt if much of substance is missing.
Reason
171
concept ‘red’ and the concept ‘existence’ stand in a specific relation both to one another and to the concept of time. I mean the ‘Red exists now,’ and thereby imply a distinction from its past and future existence. And this connexion of red and existence with the moment of time I mean by ‘now,’ would seem to be as necessary as any other connexion whatever. If it is true, it is necessarily true, and if false, necessarily false. If it is true, its contradictory is as fully impossible as the contradictory of 2 + 2 = 4. But the necessity thus involved in existential propositions does not do away with the importance of Kant’s distinction between the empirical and the a priori. So far as he attempts to base it upon the fact that what is empirical alone is ‘given in experience’ and may be referred to ‘sense,’ it must indeed be given up; but as against the English philosophers, who held the same view about sense-knowledge, it retains its full weight. The Transcendental Deduction contains a perfectly valid answer to Hume’s scepticism, and to empiricism in general. Philosophers of this school generally tend to deny the validity of any propositions except those about existents. Kant may be said to have pointed out that in any of these propositions, which the empiricists considered to be the ultimate if not the only, data of knowledge, there was involved by the very same logic on which they relied to support their views, not only the uniform and necessary succession of time, and the geometrical properties of space, but also the principles of substance and causality. He does not, indeed, thereby prove the truth of the axioms and principles in question; but he shows that they are at least equally valid with, and more ultimate than, those upon which empiricism builds. Although, therefore, it seems no longer possible to hold, as Kant held, that a reference to existents is necessary to any proposition that is to claim the title of ‘knowledge,’ and that the truth of such propositions can alone claim immediate certainty; although, on the contrary, it seems that existential propositions are only a particular class of necessary proposition: yet the transcendental deduction is still important. A deduction from the ‘possibility of experience’ does not indeed really represent the nature of Kant’s argument. For the possibility of experience presupposes that we have experience, and this again means that certain existential propositions are true: but this does not involve the truth of any particular existential propositions; although its truth is involved in theirs. What Kant really shows is that space and time and the categories are involved in particular propositions; and this work is of greater value than a deduction from the possibility of experience would have been. He does not indeed recognise that the propositions from which he is deducing are themselves necessary, and that there may therefore be other necessary propositions, with a like claim to certainty, not to be deduced from them. He therefore imagines himself to have exhausted the field of knowledge; whereas in fact he has only
172
The 1898 dissertation
shown certain logical connexions within that field. But it is not here proposed to dispute the truth of particular existential propositions; and though, unlike Kant, we admit them to be merely assumed, we may be thankful that he has shown us what can be inferred from them. Moreover, Kant’s distinction between space and time on the one hand, and the categories on the other, also retains its value, though we can no longer describe their general difference as he did. It seems rather to be this: That time alone is sufficient for some sort of experience, since it alone seems to be involved in the simplest kind of existential proposition, e.g., ‘Pleasure exists’; and that again time and space together will suffice to account for the possibility of other pieces of knowledge, without the use of the categories. It is necessary to make a fresh assumption of propositions such as even Hume recognised, and such as are universal in physical science, in order to find the principles of substance and accident and causality implied. In all such propositions time and space are presupposed as well, but these categories are not implied in every proposition involving time and space. The simplest existential propositions are then to be regarded as necessary propositions of a peculiar sort. In one kind the necessary properties of time are involved; in another those of space also. But though this fact, which Kant points out, is very important against empiricists, we cannot regard it with him as establishing the truth of geometry and of the corresponding propositions about time. For existential propositions which are false, as well as those which are true, involve the same propositions about space and time. No existential proposition of any sort seems discoverable, which might not thus be false; not even the famous ‘cogito’ is indubitable. We cannot, therefore, take the ‘possibility of experience,’ in any possible sense, as sufficient warrant for our knowledge of space and time; and we must regard the truths of geometry as independently known for true, just in the same way as some existential propositions are so known. Similarly, those propositions which involve substance and attribute are not sufficient to establish the truth of the propositions thereby involved. The permanence of substance is indeed, Kant shows us, as certain as the empirical propositions which Hume took to be alone certain. But its truth must be known independently of these, since it is involved also in false propositions of this type. It would, in fact, be true, whether any such propositions were true or not. Kant has only taught us that, if any of them are true, it must be so likewise. He failed to see that its truth may be asserted immediately on the same ground as theirs; for he was misled by the previous course of philosophy to suppose that there was something more immediately indubitable in them. Their truth is, in
Reason
173
fact, the last thing which common sense doubts, in spite of its familiarity with erroneous perceptions. Kant’s merit was in pointing out, what he himself did not recognise, that their being undoubted does not prove them to be indubitable; or rather, that the doubt which is cast on some of them proves conclusively, what common sense, in its contentment with rules that have exceptions, does not perceive, that they are highly doubtful.88 Our result then is as follows: That a proposition is universally a necessary combination of concepts, equally necessary whether it be true or false. That it must be either true or false, but that its truth or falsehood cannot depend on its relation to anything else whatever, reality, for instance, or the world in space and time. For both of these must be supposed to exist, in some sense, if the truth of our proposition is to depend upon them; and then it turns out that the truth of our proposition depends not on them, but on the proposition that they, being such and such, exist. But this proposition cannot, in its turn, depend on anything else, for its truth or falsehood: its truth or its falsehood must be immediate properties of its own, not dependent upon any relation it may have to something else. And, if this be so, we have removed all reason for the supposition that the truth and falsehood of other propositions are not equally independent. For the existential proposition, which is presupposed in Kant’s reference to experience or in Mr. Bradley’s reference to reality, has turned out to be, as much as any other, merely a necessary combination of concepts, for the necessity of which we can seek no ground, and which cannot be explained as an attribution to ‘the given’. A concept is not in any intelligible sense an ‘adjective,’ as if there were something substantive, more ultimate than it. For we must, if we are to be consistent, describe what appears to be most substantive as no more than a collection of such supposed adjectives: and thus, in the end, the concept turns out to be the only substantive or subject, and no one concept either more or less an adjective than any other. From our description of a proposition, there must, then, disappear all reference either to our mind or to the world. Neither 88
This is the end of the penultimate paragraph of ‘The Nature of Judgment’ (p. 192). In the final paragraph Moore suddenly uses the term ‘judgment’ where he has previously been talking of ‘propositions’, in a way which does not fit well with the approach taken in this chapter of his dissertation (whose main aim had been announced, in a passage not used in ‘The Nature of Judgment’, as one of showing that ‘that which we call a proposition is something independent of consciousness, and something of fundamental importance for philosophy’ – p. 162). None the less it makes sense to suppose that this paragraph has been adapted from that which occupied the corresponding place in the 1898 dissertation by substituting for the word ‘proposition’ the word ‘judgment’ which is appropriate for a paper entitled ‘The Nature of Judgment’. Even if Moore made other changes which we cannot now detect, it is worthwhile reversing this substitution in the reconstruction which follows above. For more discussion, see the editors’ introduction’ pp. lxxviii–lxxix.
174
The 1898 dissertation
of these can furnish ‘ground’ for anything, save in so far as they are complex propositions. The nature of the proposition is more ultimate than either, and less ultimate only than the nature of its constituents–the nature of the concept or logical idea.89 . . . extended part. Both these propositions seem to be true, and they are contradictory to one another. It is impossible to reject the Law of Contradiction, since otherwise there could be no meaning in saying that any proposition whatever was true. We must therefore suppose that our two propositions are not both true. Yet we wish to maintain that any true proposition with regard to space and time must presuppose them both. – The position, thus stated, is obviously untenable. Propositions which presuppose a contradiction cannot be simply and absolutely true. We can, therefore, only maintain, at most, that some propositions involving space and time are truer or less false than others; but this much has been presupposed throughout our discussion. Can any meaning be given to such comparative truth? It seems only possible to give it the meaning of consistency in falsehood; but that meaning is perhaps enough. If the inconsistencies in our propositions with regard to time and space are reduced to a minimum, then they may be said to be comparatively true. For, granted that all our propositions with regard to time and space are absolutely false, yet these may involve a greater or less number of false presuppositions, according as they are or are not logically consistent with one another in other details than their implication of the above-mentioned contradiction. 89
The 1898 dissertation manuscript now resumes on p. 25, and continues unbroken to the end of the chapter (p. 38 of the manuscript). The topic Moore is discussing at this point concerns the significance of the antinomies concerning space and time which Kant presents in book ii (chapter ii) of the ‘Transcendental Dialectic’ of his Critique of Pure Reason. There is no discussion of this topic in ‘The Nature of Judgment’, which is why there is a very abrupt transition at this point in this reconstruction of chapter ii; a paragraph or two must be missing at this point. Moore had discussed this topic in his 1897 dissertation (see p. 24) and briefly in chapter i of the 1898 dissertation (pp. 151–2). He also discusses the topic in some lectures on ‘Kant’s Moral Philosophy’ which he gave in London early in 1899, and the following sentences from lecture 4 (Cambridge University Library Add. MS 8875 14/2/5) can be used as a way of introducing the line of thought Moore’s discussion here presupposes, although Moore’s initial phrase here – ‘extended part’ – suggests that in the preceding page of the dissertation Moore had in fact been discussing space rather than time, which is the case he presents in his lecture. Kant points out that there is a difficulty with regard to the infinity of events in Space and Time. We need only consider the point with regard to Time, since it is the same with regard to Space. Kant thinks we can say both that there must have been a first event in time, and that there can have been no first event. The first is necessary, because if an infinite amount of time had elapsed before now, we could never have arrived at now. The second is necessary because, however far we go back in time, there must always have been some event earlier than any given one we come to. This is the most plausible way in which I can express his argument. These two propositions, he says, contradict one another: if one is true the other must be false.
Reason
175
But, even if this will suffice to explain our meaning when we say that a geometrical proposition, for instance, is true, we have not yet solved all our difficulties. For the peculiarity of the above-mentioned contradiction is that it seems impossible to decide which alternative is true. Whenever a real contradiction is found, it is necessary that one or other of the two contradictory propositions should be true. This must therefore be the case here also. But both are so equally involved in the whole body of science and in the common view of the world, that there seems no possible ground for deciding which is the true one. This cannot even be done by assuming one by itself, and deducing the consequences of such an assumption. For it seems impossible to discover any necessary connection between either singly and any other proposition. To assume either by itself brings greater inconsistencies into the body of science, than to assume both together. The latter assumption is plainly false, and yet it may be said to have greater relative truth than either of the others, one of which must be true. This fact appears to justify in great part the distinction drawn by Kant between Phenomena and No¨umena. The world of phenomena, we may say, is a world consisting of existential propositions which are all of them false because they all involve time, if not space, and therefore both of two mutually exclusive alternatives; but it is a world of falsehoods otherwise marvellously consistent. The world of no¨umena is a world of true propositions that are consistent both among themselves and with one or other of these two alternatives. Of this world, it may be admitted, we know little yet; but still there are some propositions, such as those of Arithmetic and Logic, which are independent of space and time, and may therefore be said to belong to it. That it is impossible to know more cannot be asserted. It is always possible that some connection, hitherto latent, should appear, and that we should therefore know it, in the same sense in which we now know anything. Kant’s restriction of knowledge to phenomena is unjustified: what he should rather have asserted is that we cannot truly know anything, either phenomenal or no¨umenal. For knowledge, being itself a process in time, cannot be truly said to exist at all. The phenomenal is not, as he thinks, what appears to us; for the no¨umenal also can appear to us, so far as anything can appear at all. It is merely what is false; and that anything appears to us is itself similarly false. But it is also true, that, if anything be held to appear at all, the greater part of what so appears must be held to be thus false; and that it is this part in which we are most interested. Whether the no¨umenal world or any part of it exists, must be held a doubtful question. There would certainly seem to be no ground for assuming the connection of existence, in the special relation necessary to
176
The 1898 dissertation
constitute an existent, with any of those concepts or propositions which we have distinguished as belonging to the world. On the other hand, wherever existence does seem to be connected with a concept or proposition, it appears, as we said above, to be also connected with time. In all the existential propositions, which we daily use, it is presupposed that existence is past, or present, or future. But, although this is so, there seems no ground for asserting that existence cannot be connected with some other concept independently of either space or time. What such a connection should be like, we cannot now imagine, since time appears to us as a presupposition of all existents. But existence is certainly a simple concept, independent, therefore, of space and time; and there seems no reason why we should not even come to know its connection with other concepts than those with which it now appears connected. Such a connection might suffice to give us an interest in the no¨umenal world, such as at present it is very hard to feel. Having thus set a view on some main principles of Speculative Philosophy over against that of Kant, it seems desirable to state how the result of the same principles, when they are applied to Ethical Philosophy, differs from his results; and particularly in what sense it is still possible to claim a rational foundation for Ethics, when the doctrine of the Practical Reason has been given up. It will have appeared from the above discussion that reasoning is not to be considered as owing its chief significance to the fact that it is a process which is conducted by our mind or reason. Logic is not to be regarded primarily as telling us even how we ought to think about a given matter, but rather as giving the relations which do truly hold between certain kinds of propositions, whether or not we do or ought to think such relations. That we ought to think such relations can only be a deduction from the fact that such relations do truly hold, combined with the ethical proposition that we ought to think what is true – a proposition which can certainly not claim unconditional validity. In reasoning, therefore, we simply follow the order of connection, in which propositions do actually stand to one another, and our reasoning is correct or the reverse, according as that objective but not necessarily existent order, which we follow, is true or false. Accordingly the description of man as a rational or reasoning being means primarily that he is capable of apprehending propositions and the connections, whether true or false, which exist between them. To call an ethical principle ‘rational’ will therefore mean, in the first place, that it is a proposition. The question of its origin, whether it be regarded as a precept of Reason, as innate, or derived from experience,
Reason
177
must in any case be wholly irrelevant. For such explanations can, at best, only give a history of how we come to apprehend it, of our knowledge of it: itself can not properly be said to have any origin at all. Nevertheless it may be well to state that the description of it as intuitively known seems nearest to the truth; for, whether it be inferred or immediately apprehended, it is, in either case, itself merely ‘given’, as much as any so-called sense-datum. But though an account of its origin, or rather the description of how we know it, is irrelevant to its rationality; the same is not the case with regard to its logical relations to other propositions. It was explained above that these relations might be of two kinds. Any proposition within the limits of a single subject may either be related to others within those limits, as being involved in them, or as being one in which others are involved. It would seem that a rational principle would always be understood to mean a proposition involved in others and not vice versˆa, – for the reason that, though it must be true, if they are true, it does not follow conversely, that, if it is true, they must be. Only such a proposition could therefore supply the ultimate truth on the subject in question. In Ethics, for example, though it may be true that the action of Rudolf Rassendyll in refusing the crown of Ruritania was a good action, and though that is undoubtedly an ethical proposition, yet hardly anybody would be prepared to admit that we had found in it a rational principle for ethics, whatever consequences we might logically draw from it. It would not be called a rational principle, because any method which attempted to deduce from it the truth or falsehood of all other ethical propositions, would not be a rational method, but would inevitably involve the vice of circularity. Further, a rational principle would generally be understood to mean a true principle. For knowledge of its truth also we must rely in the last resort upon intuition. But in this respect, too, its logical relation to other propositions is of the highest importance. Ethical reasoning does, indeed, generally consist in showing that some one proposition held to be true is inconsistent with some other likewise so held. This process requires that the denial of a principle involved in the one, should be involved in the other. It is not, indeed, thereby shown whether the principle, or its denial, is true. The process is only valuable for the purpose of producing conviction; but for this purpose it has an unlimited usefulness. A principle thus shown to be involved in another proposition may be accepted either on its own merits or on the ground of that other proposition: but, in either case, when once it is accepted, it gives a ground for systematising ethical judgments, which is not supplied by them in the complex form in which they are ordinarily held. Thus, in trying to show that ‘Pleasure is not the good’, we
178
The 1898 dissertation
rely partly on the implication of that principle in other propositions, which are admitted on both sides to be true, partly on its apparent self-evidence. If, however, it were not of the nature of a principle (i.e. involved rather than involving), the latter test only, which is the least convincing, could be applied. But in all subjects, there are also principles of a kind more fundamental than these. Those are such as are involved in any false as well as in any true proposition within the limits of the subject; principles which cannot be denied without affirming them. In Ethics the principle of this sort is that ‘Good is a simple concept’, or, as Professor Sidgwick puts it, that ‘Good is an unanalysable notion.’90 If there is any such subject as Ethics at all, this principle must be true; it is logically involved in any ethical proposition whether true or false. I have attempted to show that this is so, at some length, in my Introduction (pp. 123–6). It is here only necessary to insist that this is the rational principle of Ethics. There is involved in it, as belonging to the nature of a concept, that universality upon which most stress is laid in Kant’s Categorical Imperative. But Kant’s Categorical Imperative, as will presently be shown, includes other elements besides this, the presence of which renders it unfit for an ultimate principle. At the same time it must be admitted that our proposition lacks the complete rationality, which belongs to the ultimate principles of Speculative Philosophy: the Primacy of Speculative Reason must be vindicated against Kant. For, whereas the Ethical Principle involves the notion of truth, the notion of truth does not involve it. Hence, whereas the position of absolute scepticism with regard to Metaphysics is untenable, since a denial of validity involves the validity of denial; the ethical sceptic can rationally hold, not indeed that good is no concept, but that there is no simple concept to be denoted by the word ‘good’. It is obvious, therefore, how largely the rational principle of Ethics must depend, for a perception of its truth, on Intuition. It can indeed be shown to be involved in any ethical proposition, and that is enough for the conviction of most. But there is a large class of propositions, entirely independent of it, and from which therefore its truth cannot be inferred. Any practical maxim must involve this fundamental principle. But by itself this principle is obviously insufficient. If it is asked, what reason is there why I should do this? The only answer, in the first place, must be: Because it is good to do it. The goodness, however, may rest either in the action itself or in some result that it will produce: the action may be 90
See the editors’ introduction on this claim: p. xxxviii.
Reason
179
good either as end or means. In the former case, it must involve some principle of the form: This kind of action is good. In the latter case, the supplementary principle that – it is good to produce a good thing, – is required; and it is of this type that most practical maxims would seem to be. This supplementary principle is therefore fundamental for practice. Its truth would seem to be self-evident, although it cannot be denied that it is not involved in our primary ethical principle. Rational action might therefore be defined as action for which there is a valid reason; and such a reason can only be found in a true principle of the form ‘This kind of thing is good’, supplemented or not, according to circumstances, by the principle ‘It is good to produce a good thing’. But it cannot be denied that ‘rational action’ is commonly used to denote something other than this. In its widest sense it would seem to include any action which I perform, because I think a thing good, whether it be the action itself that I so think, or its effects, and whether or not I have deduced its goodness from any higher principle. In this case an action is said to be rational, not because there is a reason for it, but because its cause, in part at least, is the thought of a reason. It is obvious how widely these two conceptions differ; and yet they are frequently confused, even by philosophers, while common usage seems to fluctuate between the two. For instance, it will frequently be said ‘That was an unreasonable action of yours’, where the meaning is, not that I did not think the action to be right, but that I thought so wrongly, either because it was not a proper means to the end I had in view, or because the end itself was not a good one. On the other hand, if I say ‘I acted rationally; I had a good reason for that action’, I am apt to mean, not merely that the action was rational because my reason was good, but also because I had it. Again, a utilitarian might say ‘That action was rational, because it tended to produce a maximum of pleasure’, a judgment which will apply to the actions of sheep and goats, as well as to so called rational beings. Whereas Kant sometimes says, ‘This action is rational, because I was moved to it by the thought that it was good’, a judgment which will apply as well to the action of any conscientious bungler as to that of the wisest of mankind. It is cases of this latter kind which seem to have given rise to the ambiguity of the term. The principle which must form part of the ‘reason’ for any action, – the principle of the form ‘This kind of thing is good’, – is one upon which no agreement has yet been attained. There is hence almost always a difference of opinion as to whether the reason, the thought of which caused any action, was a valid one or not. It is natural for others, in giving to the agent the benefit of the doubt, and still more natural for
180
The 1898 dissertation
the agent himself, to call an action rational, when the thought that it was so, was its cause. Such a confusion is to some extent excusable in the agent himself, since he cannot, at any one moment, distinguish between what he thinks good and what is so; and hence the action which has a valid reason will often appear to him the same action as that which he did because he had a reason. Moreover it has certainly been widely held that to act because we have reasons is peculiarly meritorious, whether that view is cause or effect of the confusion in question; and in that case such action will, of course, always be rational also in the sense that there are good reasons for it, since it will always be at least partially good. Finally, the principle according to which the action is good, has sometimes been held, as by Kant, to be the cause of our thinking it, – a view perfectly consonant with the common theory that our knowledge is often caused by the object of it; and, if this were true, the cause and the reason would in many cases be the same. In view of these considerations it may perhaps be well to confine the term ‘rational action’ to actions which both have a good reason, i.e. are good, and are also caused by the belief that they have one. It was only necessary to point out here that these two meanings are absolutely distinct, even where the action is the same.
chapter iii
The meaning of ‘Freedom’ in Kant
In beginning a discussion of Kant’s notion of ‘Freedom’ which he himself considers to be essentially connected with his Ethical system, it seems most important to emphasize that fact that, so far as his express statements are concerned, he accepts unconditionally the view of Determinism and rejects that of Freedom, in the only sense in which the two have been generally discussed by English thinkers.91 In ordinary controversies on the subject, no such absolute distinction is drawn between two kinds of ‘causality’, two kinds of ‘determination’ (Bestimmung – the sense which is implied in ‘Determinism’), two kinds of ‘possibility’, or finally an ‘intelligible’ and an ‘empirical’ character, as is drawn by Kant. Professor Sidgwick, indeed, puts the question in such a form that Kant’s answer would probably have to be on the Libertarian side; but this result seems only to be obtained at the cost of the above-mentioned ambiguity. ‘Is the self ’ he says ‘to which I refer my deliberate volitions a self of strictly determinate moral qualities, a definite character partly inherited, partly formed by my past actions and feelings, and by any physical influences that it may have unconsciously received; so that my voluntary action, for good or for evil, is at any moment completely caused by the determinate qualities of this character together with my circumstances, or the external influences acting on me at the moment – including under the latter term my present bodily conditions? or is there always a possibility of my choosing to act in the manner that I now judge to be reasonable and right, whatever my previous actions and experiences may have been?’a Now to the first half of the first alternative, ‘Is the self to which I refer my deliberate volitions a definite character, etc.’ Kant would be compelled to give what Professor Sidgwick considers to be the Libertarian answer of ‘No’, because there seems to be implied in it the alternative of what he a 91
ME pp. 61–2. Pp. 181–5 of this chapter come unchanged from the 1897 dissertation pp. 20–5.
181
182
The 1898 dissertation
would call an ‘intelligible character’; though even here, he would be in some doubt, because it seems implied that the ‘intelligible character’ cannot be ‘of strictly determinate moral qualities’ or ‘definite’. And with this presumption that Professor Sidgwick accepted his distinction, he would probably answer ‘No’ to the second half ‘Is my voluntary action, . . . at any moment completely caused by the determinate qualities of this character’, although, had that question come by itself, his answer would probably have been ‘Yes’, since the sequel shews that when Professor Sidgwick says ‘completely caused’ he is only thinking of what Kant calls ‘natural causality’ (Natur-causalit¨at). So, too, in answer to the second question, he would only say ‘Yes’, on the presumption that Professor Sidgwick might mean by ‘possibility’, intelligible as well as empirical possibility. But when Professor Sidgwick goes on to exemplify the determinate view by reference to the principle of causality as employed in the Natural Sciences; when he says (p. 62) that ‘the substantial dispute relates to the completeness of the causal dependence of the volition upon the state of things at the preceding instant’, there could no longer be any doubt that only that causality was meant of which Kant had been at such pains to prove the universal validity in the Critique of Pure Reason; and only some reason for surprise that reference should have been made to the possibility of a self with any other than a psychological character. Professor Clifforda gives a statement of the doctrine of Free Will as commonly understood, which seems so clear as to be worth quoting. ‘Whenever a man exercises his will, and makes a voluntary choice of one out of various possible courses, an event occurs whose relation to contiguous events cannot be included in a general statement applicable to all similar cases. There is something wholly capricious and arbitrary, belonging to that moment only; and we have no right to conclude that if circumstances were exactly repeated, and the man himself absolutely unaltered he would choose the same course.’ Now this doctrine Kant would absolutely condemn. In fact, if Determinism only means that all men’s actions conform to the laws of nature, and so with the progress of psychology, could ultimately be predicted as certainly as the motions of the planets (and this is what Professor Sidgwick seems obviously to mean, and what is usually meant by it) Kant would have no hesitation in calling himself a Determinist. ‘All actions of man in Appearance’ says he ‘are ‘determined’ (bestimmt) by his empirical character and the other contributory causes according to the order of nature, and if we could investigate all Appearances a
Essay on ‘Right and Wrong’ in ‘Lectures and Essays’ (1886) p. 318. [W. K. Clifford, eds. L. Stephen and F. Pollock, Lectures and Essays, London: Macmillan, 1886; repr. 1901 vol. ii, pp. 126–7. The emphasis in the quoted passage is Moore’s.]
‘Freedom’ in Kant
183
of his choice (Willk¨uhr) to the bottom, there would be no single human action, which we could not foretell with certainty,a and recognise as following necessarily from its preceding conditions.’b Freedom, according to him, is absolutely impossible, if reality is ascribed to events in space and time. ‘Since the thorough-going connection of all Appearances in a context of Nature is a law that admits of no exception, this must necessarily upset all Freedom, if one were determined to cling obstinately to the Reality of Appearances. Hence also those, who in this latter respect follow the common opinion, have never been able to succeed in uniting Nature and Freedom with one another.’c Now the dispute between Libertarians and Determinists is undoubtedly conducted in general by those who do ‘follow the common opinion’ of ascribing reality to what Kant calls Appearances, i.e. matter as treated in Physics and mind as treated in Psychology. In so far as Determinism is regarded as bringing the phenomenon of Will into harmony with the results established by experimental investigation of Nature, it can only be a doctrine concerned with what Kant calls Appearances, and as such the above quotations seem to prove his unqualified adherence to it. It would, in fact, appear absurd to the ordinary champion of Free Will, to declare that ‘actions . . . which never have happened and perhaps will not happen’d are yet ‘necessary’; and yet it is only on this basis that Kant is prepared to defend Free Will. If this be absurd, there is no choice but Determinism. Kant, in fact, uses ‘necessity’ here in a totally different sense from that in which common sense usually understands it. ‘“Ought” expresses a kind of necessity and connection with reasons, which is found nowhere else in the whole of Nature. . . . It is impossible that anything else ought to happen in Nature, than what in all these temporal relations actually is; indeed “ought”, if we only look at the course of Nature has absolutely no meaning.’e If you declare future action to be ‘necessary’ the ordinary man would suppose you must mean ‘it will happen’ – that you are predicting something according to the Laws of Nature; if you do mean that ‘perhaps it won’t happen’, he would say that you are using terms inaccurately: you ought to have said it was only probable or possible. But meanwhile it is sufficient to point out that Kant does say this absurd thing; and that from this second meaning of ‘necessity’ there follows a second meaning of possibility also. Since that which according to the Laws of a b c e
It will be seen that I myself think it necessary slightly to modify this statement. R.V. pp. 380–1 [B577–8, GW 541] [Moore’s emphasis]; cf. also P.V. p. 103 [AK 5: 99, MG 219]. d R.V. p. 380 [B576, GW 541]. R.V. p. 373 [B565, GW 535]. R.V. p. 379 [B575, GW 540].
184
The 1898 dissertation
Nature is only possible can be called necessary, that which according to the Laws of Nature is absolutely impossible, may, from Kant’s point of view be regarded as ‘possible’.a It is only on this supposition of the possibility of the impossible, that Kant could have answered ‘Yes” to Professor Sidgwick’s second question. What, then, if Kant is a Determinist, does he mean by that Freedom, the reality of which he asserts? He approaches this problem in the Critique of Speculative Reason, in discussing the application of that notion of causality, to defend the objectivity of which against Hume was one of the chief motives to the formation of his Critical Philosophy. The principle of causality is a principle by which the existence (Dasein) of one object is necessarily connected with the existence of another. Hume had not denied the necessity of some kinds of relations between objects; for instance he seems never to have suspected any insufficiency in the grounds we have for saying that five apples and seven apples are equal to twelve apples. This, he would have said, is a relation which, while the objects remain the same, must also remain the same, because, whether there ever were twelve apples or not, the mere idea of them would always stand in the same relation to the ideas of seven and of five. He recognises that the certainty of a relation of this sort does not depend on the certainty of any existential proposition. Even if we could never learn that 12 = 7 + 5, unless we had seen twelve objects, yet our ground for saying that 12 = 7 + 5, is not only that we have seen cases in which they are, as, according to him, our only ground for saying that there are apples, is the perception of them. He would entitle us here to the further statement, that if there are twelve objects of any sort, it is impossible but that they should be equal to seven and five of the same sort. Our statement that 12 = 7+ 5, does not only mean, that in all cases which we have observed, this is so; as would (according to him) such a statement, as that, if there are apples, there were apple-blossoms before them. In the former instance, he would say a priori that there never could be a case in which 12 was not equal to 7 + 5; whereas in the latter, he would utterly deny our right to make a generalisation universal. According to him apples are preceded by apple-blossoms, when they are; but that they must be, or that any existent must ever be so connected with any other, is what, he would say, we have no ground to be certain of. a
The words of R.V. p. 379 [B576, GW 540] ‘Now the action must undoubtedly be possible under natural conditions if it is conformed to the “ought”’ must be understood to mean, that any actual action, which was in accordance with ‘ought’, must also have natural possibility i.e. have been capable of prediction according to natural laws: not that for any conceivable action to be moral, it must also be naturally possible. So in M. p. 18 [AK 6: 221, MG 376], is it obvious that the ‘morally possible’, the ‘permitted’, may be something which you cannot actually do.
‘Freedom’ in Kant
185
In fact Hume treats propositions about relations of quantity, as if they were purely analytic – as if 7 + 5 were identical with 12. Kant saw that they were not so. Mathematical propositions are synthetic; and at the same time, as Hume admitted, they are a priori. Kant’s object, therefore, in the Critique of Pure Reason, is to consider the validity of a priori synthetic propositions, as well those mathematical ones, which Hume thought to be analytic, as those expressing a causal relation to which Hume, because he saw that they were not analytic, had denied any a priori necessity. The precise manner of Kant’s Deduction is not here to the point. What the above brief account of the question at issue is designed to bring out, is his distinction between the category of causality and the mathematical categories, which makes him call the former ‘real’ or ‘dynamical’; for it is upon this distinction that he conceives to rest the possibility of Freedom. It is solely in virtue of the dynamical categories that we are able to say that if one thing exists, another must exist. The mathematical categories, he holds, only tell us that between things appearing in space and time, certain relations must hold; they do not allow us to conclude from the existence of any one thing to the existence of any other. Hence, when it appears that, in trying to obtain a complete notion of the world as ‘determined’ by the various categories, we are involved in antinomies, the contradictory propositions resulting from such an application of the mathematical categories must both be false, since both relate solely to the temporal and spatial properties of things. But with the dynamical categories the case is different. These profess to deal with the existence of things. Now Kant always asserts the existence not only of appearances, but also of something of which these are appearances. It is therefore possible that any proposition which concerns the existence and not merely the relations of things, may be true with regard to these Things-in-themselves. And hence, though the contradictory propositions resulting from the attempt to obtain a complete view of the world under the dynamical categories, must both be false, if it is presupposed that we are talking only of things existing in time, it is possible that both may be true, if we are entitled also to take into account the existence of a Thing-in-Itself. It is with regard to the proposition that things have an unconditional or first cause, i.e. a cause which is not itself the effect of something else, that Kant points out this possibility; and the causality of such a first cause would, he says, be Freedom. If, therefore, it can be shewn that this proposition is true, not merely that it may be, then the existence of Freedom is proved.92 92
At this point Moore introduces into the 1898 dissertation some new material which continues until p. 188.
186
The 1898 dissertation
In the Critique of Pure Reason, therefore, we do not arrive at the result that Freedom exists, but merely that it is possible. And even this result is only attained by the assumption that the world of existents, which we know, and within which Determinism holds, is merely phenomenal, and that there exists, beside it, a world of Dinge an Sich, which are in some sense the ground of phenomena. It has been explained in our first chapter how Kant endeavours to support this assumption; and reasons have been given, both there and in Chapter II, for holding the assumption to be invalid. Kant does, in fact, only arrive at the conclusion that Freedom is possible, by presupposing that it is real, for he bases its possibility admittedly on the existence of Dinge an Sich, and on the other hand, his argument in the Dialectic presupposes that Dinge an Sich are unconditioned. Taking, therefore, the two presuppositions together, it would follow that the Ding an Sich both exists and is unconditioned, i.e. free. But this conclusion was concealed from Kant himself, because his reason for asserting its existence was one, which he felt would not justify him in ascribing to it any definite properties; whereas the principle that it must be unconditioned is conversely insufficient to prove that it exists. But, though Kant’s professed reasons for the assertion of Freedom are to be sought not in the Critique of Pure Reason, but in his ethical works, the first Critique throws a great deal of light on the question how Freedom is to be conceived. The notion has, as we have seen, something in common with the conception of causality. That which is free is not, however, a cause in the ordinary sense, since it implies nothing whatever with regard to time. In this respect its relation to its effect is more like that of reason and consequent, from which, in the form of the hypothetical judgment, Kant derives the unschematised category of causality. But, though it differs from a cause, in that it is not to be conceived as preceding its effect in time, nor indeed as being in time at all, it has again a greater resemblance to a cause than to a reason, in that it must be conceived as an existent. The no¨umenon, which is alone capable of freedom, can only be thought as self-subsistent; although in the Critique of Pure Reason Kant professes to be unable to determine that it is exists.93 In this respect, again, it is similar to the Ding an Sich, which is certainly conceived by Kant as an existent. And it is hard to see how the two are to be distinguished either, in respect of their relation to phenomena; since the relation of the Ding an Sich to these, not being that of the schematised category; and yet being described as that of ground 93
Sic – Moore originally wrote ‘it is existent’, but then changed ‘existent’ to ‘exists’ without deleting ‘is’.
‘Freedom’ in Kant
187
or cause, must be conceived as that of reason to consequent, transferred, so far as possible, to a relation of existent to existent. In any case, it would seem that Kant conceives Freedom as something combining the notions of causal and logical priority. The desire to effect such a combination seems, indeed, to have been one main motive of his philosophy; and we may add, one main cause of its confusion. It is this, to which his description of an Intuitive Understanding, the perfect knowing faculty of God, seems to point. This conception combines for him the different conceptions of truth found in rationalism and empiricism. The necessity expressed in the universal judgment it takes from the first, and the assertion of existence expressed in the categorical judgment from the second. Were we possessed of such a knowing faculty, philosophy would not oscillate between the two extremes of saying that our sensations are all that we can know, while the so-called laws by which we attempt to group them are mere abstractions and represent nothing in the world itself; and of asserting that nothing is truly known except universal connections or predicates. With our present knowing faculty we can only attain a union of unity with plurality, or identity with difference, in two ways, each imperfect. Either the unity or identity takes the form of a mere abstract notion or logical universal, which exists only in each given particular; or it is the given particular which persists through change and unites in itself several universals. Kant imagines that for God these two forms of unity would be the same. Each object would be identical with every other in the sense in which the universal is identical with itself in each particular; but at the same time each would be substantial and so unite in itself all the others in the same way as now the particular unites universals. How this should be, he rightly says, is to us incomprehensible: we cannot reduce the ‘individuation’ given by the forms of space and time to the individuation given to concepts by thought, or vice versˆa. But that Kant should have regarded this as the ideal of knowledge seems to explain the central position which Freedom holds in his philosophy, while at the same time it throws light on that conception. For that which is free differs from the Intuitive Understanding, only in that Kant is concerned in explaining its relation to phenomena, instead of to other noumena. Both conceptions are alike in that they must be conceived as existent, and related to other existents, but in a relation of identity, as Kant conceives that of reason and consequent to be. In this, again, too, appears the similarity of the no¨umenon to the Ding an Sich. For the identity of the Ding an Sich with the no¨umenal Ego has been suggested by Kant himself. The two would therefore seem to stand to one another exactly as does the Intuitive Understanding to its object; and the relation
188
The 1898 dissertation
of the one to phenomena could, therefore, hardly be distinguished from that of the other. The Ding an Sich would thus again appear as deserving the ascription of Freedom as completely as the no¨umenal Ego, to which, as we shall see, Kant practically confines it. In view of these considerations, it may fairly be said that the ascription of Freedom to the Ding an Sich results from the general position of the first Critique. In that case, however, every phenomenon alike will be freely caused, because all alike are grounded in the no¨umenal world. Kant himself never expressly draws this conclusion, because of his uncertainty as to how the Ding an Sich is to be regarded. He has asserted its existence, and that prevents him from completely identifying it with the no¨umenon, whose existence he cannot yet assert. Moreover he never fully systematized his doctrine of no¨umena, being chiefly concerned in pointing out his differences from previous philosophers in the systematic treatment of phenomena. Though, therefore, the ‘Pure Reason’ would lead us to expect that Transcendental Freedom would be acknowledged as affecting all phenomena alike, it is not surprising to find that in his express statements on the subject this is by no means the case.94 Kant does not admit that every separate thing may be regarded as a result of intelligible or free causality, just as every separate thing is a result of natural causality. All he claims in his discussion of Transcendental Freedom at the end of the Critique of Pure Reason, is that ‘among natural causes there may also be some, which have a faculty, that is only intelligible.’a And he goes on to explain that by these he means only mankind. ‘In inanimate or merely animal nature we find no reason to suppose any faculty conditioned otherwise than through sense. But man, who knows the rest of nature only through senses, gets knowledge of himself also through mere apperception . . . and is to himself partly, we must admit, a phenomenon, but partly also, namely in view of certain faculties, a merely intelligible object.’b Now it is plain that Kant here regards man as on an absolutely different level from other things, in respect of freedom. Man has freedom and nothing else shares it in any degree. And throughout his ethical works this attitude is maintained. Free causality is attributed to man alone, among the objects of experience.c So that, whereas natural causality applies with absolute universality – to him as well as to all other objects, a c
94
b R.V. p. 379 [B574, GW 540]. R.V. p. 378 [B573, GW 539]. Though also it belongs to any other ‘reasonable beings’, if such there be. Gr. p. 237 [AK 4: 389, MG 45].
Moore here reverts to the 1897 dissertation pp. 37–8, though with several changes.
‘Freedom’ in Kant
189
freedom appears as a sort of miraculous power, whose influence may be traced in some events, but not in others.95 It is with this restriction of freedom to the human will, in the form in which it appears throughout Kant’s ethical works, that we have now to deal in detail. It will appear that his conception of ‘practical freedom’, seems to partake partly of the nature of that transcendental freedom, which he should logically have regarded as ground of every event alike, and partly of the nature of human volition. In it, therefore, the hybrid conception of transcendental freedom, derived partly from a logical, partly from a causal relation, is still further tainted by a second mixture with the causal parent. In it, the discrepancy between the causal and the rational strains bursts out into open incompatibility.96 ‘Practical Freedom’ is something which must belong to all ‘reasonable’ beings, as such.a It is defined negatively as ‘the independence of our choice from compulsion through impulses of sense’;b and positively as ‘a power’ or ‘causality’ of ‘reason’, ‘to begin a series of events entirely of itself ’.c ‘Pure practical reason’ is identified with ‘pure will’; and ‘will’ again, up to the end of the Critique of Practical Reason, seems to be identical with ‘choice’ (Willk¨uhr), though in the Preface to the Metaphysic of Morals,d they are distinguished in a very important manner; for it is there declared that only ‘choice’ can be called ‘free’, ‘will’ being concerned not with ‘actions’, but only, like practical Reason, with the giving of Moral Laws. Kant’s account of the way in which we must conceive ‘practical freedom’ in relation to experience is as follows. Every ‘cause’ (Ursache) has a ‘power’ (Verm¨ogen), which may also be called its ‘causality’ (Causalit¨at) which ‘power’ is necessarily connected with the subsequent appearance of a definite effect (Wirkung); and the law of this connection is called the ‘character’ of the cause. The transition from the ‘causality’ to the ‘effect’, however it be conceived, is called the ‘action’ of the cause (Handlung). Now in ‘natural causation’, the ‘causality’ of every cause is also an effect of some previous cause and so on ad infinitum; and the ‘action’ therefore is merely a transition in time. But for every natural object, we must also suppose there to be an intelligible ground; and there is no contradiction a b c d 95 96
G p. 296 [AK 4: 449, MG 96]. R.V. p. 371 [B562, GW 533]; cf. M. p. 11 [AK 6: 213–14, MG 375]. R.V. p. 372 [B562, GW 534]; cf. M. p. 11 [AK 6: 213–14, MG 375]. M. p. 23 [AK 6: 226, MG 380]. The next paragraph is new to the 1898 dissertation. Moore now jumps to p. 59 of the 1897 dissertation, and pp. 189–206 of chapter iii (1898) come from pp. 59–78 of the 1897 dissertation with only minor changes.
190
The 1898 dissertation
in thinking of this intelligible ground as cause (in another sense) of the ‘causality’ of the natural object. The ‘causality’ of the natural object would thus be effect both of some preceding natural object and also of its intelligible ground. But the intelligible ground is, as such, in no way subject to ‘time-conditions’, and therefore its ‘action’ in producing the ‘causality’ which is its appearance, is not a time-transition. It cannot therefore be said to ‘begin to act’ at any time; although its effect, i.e. the ‘causality’ of the natural object has a beginning. It is thus original cause of an appearance, which is on another side also effect of a conditional cause and in its turn cause of other appearances. It begins ‘of itself’ a series of events in time, without itself beginning to act. But this is an account which would apply to any natural object whatever, whereas Kant applies it in a special sense to human volition. I quoted above (p. 188) a passage of Kant,a in which he says that ‘man knows himself ’ not only through his senses but ‘also through mere apperception, and that too in actions and inner determinations, which he cannot ascribe to the impression of the senses. He is to himself, it must be admitted, partly a phenomenon, but partly also, namely in view of certain faculties, a merely intelligible object, because his action cannot be ascribed to the receptivity of sensibility. We call these faculties Understanding and Reason.’b And he goes on to say that Reason appears preeminently as the faculty of supersensible being. We are next told that the Imperatives expressed by ‘ought’ make it plain that ‘this Reason has a Causality, or at least that we represent it as having one.’ And finally we have the following sentence: ‘Now this “ought” expresses a possible action, of which the ground is nothing more than a mere conception; whereas, on the contrary, the ground of a merely natural action must always be an appearance.’c In this passage I think we have presented the full extent to which Kant’s error of restricting practical freedom to reasonable beings goes, together with the confusions on which that restriction was based. One ambiguity occurs in the last sentence; and it is a very important one, since it seems to have given rise to many false notions of what Kant meant by freedom. This sentence expresses in an antithetical form the difference between ‘free’ and ‘natural’ causality – which he frequently says are the only two kinds of causation possible. The first is distinguished by this that its ground is a mere conception; whereas the ground of the second is always a phenomenon. Now from the account given above of Transcendental Freedom it will a c
R.V. p. 379 [B574, GW 540]. R.V. p. 379 [B575, GW 540].
b
R.V. p. 379 [B574–5, GW 540].
‘Freedom’ in Kant
191
appear in what sense this description of free causality is to be understood. A free cause must necessarily appear to us as a logical reason, and so far as ‘a mere conception’; because it is not, as such, presented to us as an object of intuition. But from what Kant says in the preceding context, as well as from his general account of will elsewhere, I think it is plain that he is not thinking of ‘a mere conception’ in this sense. When our will is singled out as having a special kind of causality, inasmuch as it can be ‘determined to action by the presentation (Vorstellung) of certain lawsa , Kant shews what it is he is thinking of. The ‘mere conception’, in the only justifiable sense for freedom, would be the laws themselves, and not the ‘presentation’ of the laws. Every ‘conception’ may be regarded from two points of view, either as a psychical existent, or from the point of view of its content; and it is this very important (and obvious) distinction which Kant appears to have neglected. If the causation exercised by the presentation of a conception were enough to justify freedom, freedom would be no more than an aspect of every mechanical process; and thus there would not even be an appearance of conflict between it and natural causality. For it is precisely ‘presentations’ to which Kant repeatedly asserts that the objects of experience are reduced, when they are viewed, as he holds they must be, as appearances. An appearance is a ‘mere presentation’, and it is only between such that the causal laws will hold. There would therefore be no difference between ‘an action of which the ground was no more than’ the presentation of ‘a conception’, and an action of which the ground ‘must always be an appearance’: for the presentation of anything whatever is, as such, an appearance. Kant himself would seem to recognise this in a passage of the ‘Canon of Pure Reason’, in which for that very reason he is driven to an almost direct contradiction of what he says in the context quoted above. In this passage (p. 530)97 he says: ‘Practical Freedom can be proved through experience. For not only that which charms, i.e. affects the senses directly, determines human choice, but we have a power to overcome impressions upon our sensual desiderative faculty (Begehrungsverm¨ogen), through presentationsb of what, even in a somewhat remote way, is useful or harmful; and these considerations of that which, in view of our whole state is desirable, i.e. good and useful, are based upon Reason. Hence also Reason gives laws, which are Imperatives, i.e. objective Laws of Freedom, and which tell us, a 97
G. p. 275 [AK 4: 427, MG 78]. R.V. p. 530 (B830, GW 675).
b
My italics.
192
The 1898 dissertation
what ought to happen, even though perhaps it never does happen, and are distinguished in that respect from Natural Laws, which deal only with that, which happens.’ He then goes on to suggest that on a wider view, what here appears as freedom, might be seen to be nature (which would, indeed, with regard to part of his statement, be certainly the case); but this, he says, is a speculative question, irrelevant just here. Finally he comes to this: ‘Accordingly we know practical freedom through experience as one among natural causes, namely a causality of the reason in determination of the will; whereas Transcendental Freedom demands an independence of this reason itself (in view of its causal power to begin a series of appearances) from all determining causes of the sense-world, and so far appears to be contrary to the Law of Nature, and hence to all possible experience; it therefore remains problematic. But for reason in its practical use this problem is irrelevant . . . . . . .The question with regard to Transcendental Freedom concerns solely speculative knowledge. We can set it aside as wholly irrelevant, when we have to do with the practical.’a Now in this passage Kant states very well what is characteristic of human volition; and his definitions of ‘will’ are constantly expressed in the same fashion. Will differs from other instances of natural causation, in as much as in it the idea (to use the common English word for ‘Vorstellung’) of something, which is not yet real, tends to bring about the realisation of that thing; and he may be justified in saying that the process ‘is based upon Reason’, since to have an idea of anything either real or imaginary presupposes that faculty of cognition which distinguishes man from beasts, and still more from inanimate nature. Nay, more than this, in the special instance, which Kant takes to be the only truly ‘moral’ willing, where the idea which acts as cause, is the idea of conformity to a universal law, the content of the idea is so abstract, that it may be confidently asserted that only reasonable beings are capable of having such an idea. But nevertheless the idea is even here still ‘an appearance’, and, as such separated by an impassable gulf from the content, of which it is an idea. And, inasmuch as it is in its character of idea, i.e. as a psychical existent, that it produces an effect, the causation is still merely ‘natural’. This, as we have said, Kant in the present passage fully recognises. But it is only the more remarkable that he should speak of Reason in the same context, as ‘giving laws of Freedom’, as if it were Reason in the same sense, which is the source on the one hand of objectivity, and on the other hand of abstract ideas, whether true or false. In this Kant betrays the psychological standpoint, a
R.V. p. 530 [B830, GW 675].
‘Freedom’ in Kant
193
which has been fully discussed in our first two chapters.98 The distinction between what is true and what is only believed (although only a ‘rational’ being can believe) is one which cannot be either done away or bridged over, however small be the amount of what we may be thought to really know in comparison of what we must be content to believe; and it is this distinction which is here in question. Knowing, the function of Reason, is on one side a natural function, and, as such, it is indistinguishable from believing; but, in so far as knowing is distinct from believing, i.e. in so far as that which is known is true, there are no two words which express a difference more profound. When Kant talks of the only true morality as based upon the laws which Reason gives itself, the whole course of his work shews that he means laws which tell us truly what ought to be done; it is, indeed, only on this condition that he could claim universality for them. In this sense ‘Reason determines the Will’ whenever the idea which is cause of our action, is an idea of what is truly good. But it is only in an utterly different sense that ‘Reason’ can be said to ‘determine the Will’, whenever the idea, which causes our actions, implies the power of knowing concepts. And it is only in this second sense that such determination of the will can be called a ‘practical freedom’ which is independent of ‘Transcendental Freedom’. Accordingly Kant himself, as we have said, recognises elsewhere that ‘the practical conception of freedom is based upon’ the ‘transcendental Idea of Freedom’;a and again, speaking of freedom, ‘as one of the faculties, which contain the cause of the appearances of our sense-world’, i.e. as practical, in distinction from transcendental, freedom, he declares that we cannot hope to establish its actuality in experience, ‘inasmuch as we can never infer from experience to anything, which must not be conceived at all according to laws of experience’.b And this seems sufficient to contradict his statement in the ‘Canon’ that ‘we know practical freedom through experience as one among natural causes’ and that ‘for reason in its practical use’ the problem of transcendental freedom ‘is irrelevant’. Kant has therefore confused the purely natural process of human volition, with the transcendental aspect of it, which would alone entitle us to ascribe to man ‘practical freedom’; and it is solely on this confusion that the special place he assigns to man as a ‘free’ agent seems to be based. It is true that the content of the idea, which acts as cause in volition, is different from the content of any other natural cause; but that content is merely the a 98
R.V. p. 371 [B561, GW 533].
b
R.V. p. 385 [B586, GW 546].
Moore here omits a passage from the 1897 dissertation in which he balances this criticism of Kant with enthusiasm for the ‘enormous services which he did for epistemology’; see pp. 62–3.
194
The 1898 dissertation
form of the cause, and difference of form is something which in no way renders one natural cause more or less of a natural cause than any other. The question is, however, complicated by the fact that we are dealing, in the case of volition, with an ‘appearance of the inner sense’; and as this point also touches the legitimacy of Kant’s assertion that ‘appearances’ are mere ‘presentations’, an assertion which I used in my argument to prove that for him human volition was really a causation by ‘appearances’, something should be said of it. Berkeley had held that everything which we directly experience, must be regarded as a state of our own mind. Kant pointed out that Berkeley’s hypothesis would leave us no means of making a rational distinction between dreams and perceptions – between mere illusion (lauter Schein) and appearance (Erscheinung). He shews that we necessarily conceive things extended in space (the form of outer sense) as forming a connected series, of which the succession is objective, i.e. not existent merely as and when we perceive them. We could, he holds, not know a thing to be a state of ours, except by contrasting it with something not so dependent on us, since we can only know ourselves in contrast to things in space. Our ‘ideas’, therefore, in so far as they are states of ourselves, must be conceived as forming a causal series of their own, different in the order of its connections from that of the content of those ideas, considered as extended in space. The knowledge of the mental series, as such, involves the existence of the series of things in space; but the latter must be conceived as so far independent of the former that it might exist, without being represented in any mental series. Nevertheless it remains true that the spatial series can only be conceived as having identical elements with those which occur in the mental series. And thus it is true, in a sense, that that the spatial world is composed of mental elements; only these mental elements, in so far as they are conceived as existing in one space, and belonging to one series, of which the parts are in a relation of complete causal interdependence, must not be confused with the elements, whose content is partially the same, but which form part of the mental series. Kant’s denomination of the elements of the spatial world as ‘mere presentations’ is justifiable, if it be clearly understood that these ‘presentations’ are the same in content only, with those which belong to the mental world, but not the same in respect of their causal relations with one another. And this, on the whole, he makes plain enough by his distinction of the ‘outer’ from the ‘inner’ sense. But the ambiguity of the name nevertheless lends a handle to confusion, into which he sometimes seems to fall, as, for instance, when he speaks of spatial objects as being actually ‘in our mind’ (in unserem Gem¨uthe). A concrete instance will make the
‘Freedom’ in Kant
195
distinction plain. The ‘presentation’ of a cow munching upon grass, may in the mental series be cause of a feeling of pleasure, or of the recollection of the words ‘cud’ or ‘ruminate’; but the same munching will probably be the cause, in the material series, of digestion, and, perhaps, of such movements in the cow’s brain as correspond to her feeling of pleasure. Yet the latter munching is certainly composed of some elements which are the same as those composing the former. The difference is not so much between the ‘states’ between which, as Kant insists, the causal connection must always be conceived to hold, but between the substances, of which, as Kant also insists, they must be conceived as states. In the case of the material cow, the substance is conceived as extended in a unique space; but in the case of the presentation of her, the substance is mental – not extended in space, but also conceived as persisting through time. This difference of substance adds an element of content to the states of each series, which distinguishes a state conceived as belonging to the one from the same state conceived as belonging to the other. Part of the content, then, of some mental states is necessarily conceived as forming a world by itself. But there are other mental states, as, when we are dreaming imagining, feeling pleasure and pain, or engaged in abstract thinking, of which the content is not so conceived. In these cases the content, in so far as it is conceived as existing, exists only as the form of mental states. The difference between such contents and those which are conceived as constituting the spatial world, is what is marked by the fact that Kant calls the latter contents themselves ‘presentations’. For these latter are not mere content, i.e. merely what distinguishes one mental element from another, but appear to be given as existent and to be bound together in one context of experience, exactly in the same way as the presentations of them, along with the other mental elements, appear to be given as existing and to be bound together in another context of experience. There is exactly the same reason for calling these contents themselves ‘presentations’, as for giving that name to the mental states of which they are contents and to all other mental states. The first group are ‘presentations’ of ‘the outer sense’, the second of ‘the inner’. When a psychologist examines his mind, he has, to use Hume’s terms, ‘impressions’ both of ‘impressions’ and of ‘ideas’; and ‘impression’ is used in both these cases in the same sense, that of something both immediately given, as an object of intuition (Kant’s Anschauung), and as forming part of a unique series. But when his ‘impression’ is of an ‘impression’, the latter impression necessarily appears as part of his mental series, and it was only as such that Berkeley erroneously wished to consider it; whereas when his impression is of an ‘idea’, the ‘idea’ is
196
The 1898 dissertation
really a part of his mental series and nothing more. When Kant, therefore, speaks of the ‘appearance’ to which he confines our knowledge of the real world, as ‘mere presentations’, he means what would be more properly termed ‘contents of presentation’, since they need not always have actually been ‘presented’; but still he means those contents, considered as having the character conferred by presentation, i.e. as absolutely particularised or individuated by the unique time-series, not in their abstract character. Anything whatever, which ‘determines the will’, i.e. which causes action, is a ‘presentation’ in this sense, since it must be an element in the mental series and an ‘object’ (Gegenstand) of the inner sense, either possible (if it be subconscious) or actual. But confusion is apt to arise because some of these ‘presentations’ of the inner sense, are also ‘presentations’ of the outer. And a further confusion also occurs, because the ‘presentations’ of the inner sense, may be ‘ideas’ either of other ‘presentations’ of the inner sense, e.g. the ‘idea’ of my future pleasure, or of presentations of the outer, e.g. of some very good wine, not actually existent. And all these cases are different in respect of the content of the ‘presentation’, from those in which it is some mere concept like the moral law. Nevertheless, what it is important to notice and what is too often left out of sight, is that in all cases alike what causes action is some ‘presentation’ of the inner sense, whether its content be the moral law or the ‘idea’ of another presentation; and that thus in all cases the ‘will’ is subject to the ‘natural’ law of causality, and ‘determined’ by an ‘appearance’. Now it is characteristic of ‘voluntary’ action, i.e. of human volition, as opposed to conation in general and to some forms even of desire, not only that the presentation of the inner sense which is its cause, must always be an actual presentation, as opposed to one that is subconscious, but also that this presentation must have for content a mere ‘idea’ of another presentation of the outer or inner sense, and not an actual presentation either of the one or the other. Thus it is possible to desire a pear, which you actually see, and this desire may be the cause of your eating it; but, if this is all, the action will not be voluntary. To make it voluntary, the desire must not be of the pear, but of the taking and eating it, actions which are not yet actually prevented. It is this which would seem to distinguish actions, which are automatic, as a result of habit, from those which are merely instinctive. Both alike are caused by ‘presentations’ of the inner sense, which are not actually presented but subconscious; only in the first the subconscious ‘presentation’ must be of something itself not actual, whereas in the case of instinct, it may be of an actual presentation of the outer sense. For instance, there seems little doubt, that a dog may be directly
‘Freedom’ in Kant
197
moved by the sight of a man, whom he hates, to attack him. In which case the presentation of the man, is a subconscious presentation from the point of view of the inner sense (for the dog probably does not perceive that he sees the man); but it is not a subconscious presentation of the attack, as it might be in a man who had consciously formed the habit of caning a certain person, whenever he met him. These distinctions are all of them important in considering Kant’s account of volition, as with regard to them he uses at times ambiguous language. What we are concerned with now, however, is their bearing on Kant’s view of practical freedom, as belonging only to man. And I think they serve to explain another confusion in the passage quoted above.a It is not only that ‘conception’ is there used as equivalent to ‘presentation of conception’, or conception as psychical existent; but there seems also to be an ambiguity in the sense in which he speaks of man’s action as characterised by the fact that it ‘cannot be ascribed to the receptivity of sensibility’. This description is opposed to that of ‘lifeless or merely animal’ nature, in which we have no reason to ‘think of any power, as other than sensuously conditioned (sinnlich bedingtb ). Now by this latter expression it seems plain that what Kant means, is that the actions of all other parts of nature have sensible objects for their causes. It is in this sense that it forms a true antithesis to what immediately follows: ‘Man, who knows the whole of nature besides solely through senses, gets knowledge of himself also through mere apperception’; i.e. other animals and inanimate nature are mere objects of the senses, and therefore their actions can only be caused by objects of the senses, which is the type of natural causality; whereas man is something more. But, when Kant goes on to speak of man as recognising actions of his own ‘which he cannot ascribe to the impressions of the senses’, it is hard to believe that those ‘actions which can be ascribed to sense impressions’ are not actions of man’s will, under the aspect in which he calls it ‘sensual’ (sinnlich), a term which he explains as meaning that it is ‘affected pathologically (through motives (Bewegursachen) of sense)’.c Now, if this be so, Kant would seem to be treating three totally distinct notions as all equally relevant for his purpose. Man is sensual (1) in so far as his actions may be regarded as sensible objects, belonging to a context of experience (2) in so far as he receives impressions of the senses, i.e. in so far as he is capable of perceiving sensible objects (3) in so far as the ‘idea’ of sensible objects is the cause of his actions. And again, in all these divisions, a c
R.V. p. 379 [B575, GW 540]. R.V. p. 371 [B562, GW 533].
b
R.V. p. 379 [B575, GW 540].
198
The 1898 dissertation
an ambiguity remains because the ‘sensible objects’ may be objects either of the outer or of the inner sense. Now, when Kant speaks of ‘heteronomy’ of the will, that is a distinction based upon the third sense in which man is ‘sensual’. The will is heteronomous, when that which causes the action is either the ‘idea’ of some sensible object, either of outer or inner sense, or the actual presentation of some object of outer sense. Even here, however, Kant is ambiguous, since he sometimes speaks as if the actual presentation of an object of outer sense, were necessary to constitute heteronomy – such a case as that given above of the desire of a seen pear; and, where he does not appear to limit the conception so narrowly as this, he seems almost always to be thinking of the ‘idea’ of some object of the outer sense, e.g. the taking of the pear, and hardly ever of the ‘idea’ of some object of the inner sense, e.g. the sensations of touch and taste that accompany the eating of the pear, or the pleasure they excite. Yet it is plain that where the cause of action is the ‘idea’ of some future state of the agent’s mind, the action is fully heteronomous. But the confusion which bears on ‘practical freedom’ is that Kant’s language encourages the notion that where action is not heteronomous, i.e. does not fall under (3), it is therefore not naturally caused, i.e. does not fall under (1); which is by no means the case, although all action which is heteronomous must fall under (1). Man’s sensuousness under (2) is irrelevant both to heteronomy and to practical freedom. Kant is here committing the error of regarding the relation of subject and object, in knowledge, as the relation of ordinary causality, which holds only between objects. He traces the possibility of heteronomy quite rightly to the fact that much of our experience is what he calls sensuous. But he seems to regard the effect of objects upon us, whereby we know them, as the same thing as their effect in determining the will; whereas the knowledge of objects cannot be rightly regarded as an effect upon our sensibility at all, and, if it were, it would not be the same thing with the effect which the actual presentation of objects or the ‘idea’ of them, regarded as psychical existents, produces in subsequent states of mind. Yet Kant seems to regard the undoubted fact of sense-perception as a confirmation of the reality of heteronomy, and, through that, of his misleading description of ‘practical freedom’ as something empirically knowable. It is in man’s sensuousness under (1) that we reach what is really relevant to freedom. When this aspect of his nature is clearly distinguished from (2) and (3) it becomes plain that all his actions fall simply under ‘natural laws’. They are all mere ‘appearances’ of the inner sense, and are completely caused by previous appearances of the inner sense. It is true, that, where the cause of an action is the presentation of a law, the law which is presented is
‘Freedom’ in Kant
199
not itself a sensuous object, and herein there lies a real distinction between the actions of men and beasts. But what is important is that the law must be presented to the inner sense, before it can be conceived as causing the action; and as a presentation of the inner sense, it is as truly a sensuous object as any other, though only of the inner sense. Thus we might truly say of man, what Kant says of beasts, that we have no reason to think of any of his faculties as ‘other than sensuously conditioned’. The difference between men and beasts is only that beasts are apparently incapable of knowing concepts, i.e. the content of the presentation which causes their action must always be itself a presentation: in other words, because they know nothing but sensible objects, their action is always heteronomous. But this difference between man and beasts only proves him to be a different sort of natural object from them; it is by no means sufficient to entitle us to ascribe ‘practical freedom’ to him. Man does not, when he acts morally, exhibit any ‘pure activity’.a This ‘pure activity’, in so far as it can be ascribed to him at all, consists only in his being able to conceive the Ideas of Reason;b that, when they have been conceived, they should influence his action, cannot be regarded as an instance of the same activity. And this power of conception itself cannot, without danger of misleading, be called a ‘pure activity’. That is an expression which suggests that the Ego to which Kant attributes it, may from the cognitive point of view, at all events, be regarded as itself an uncaused cause of its pure notions, e.g. the Ideas of Reason. But, as has been pointed out above, Kant could not maintain that the Ding an Sich or the Transcendental Ego really stood in a causal relation to experience: their relation to it is rather that of reason and consequence. To maintain the opposite is to identify the ‘synthesis’ of the Transcendental Ego, with the actual process of judging, and to make the Transcendental Ego itself the psychological subject. If the relation of the premises of a syllogism to the conclusion could be called ‘activity’, there might be nothing misleading in the notion of a ‘pure activity’. But, as a matter of fact, it is rather the relation of the recognition of the premises to the recognition of the conclusion as conditioned by them, that seems to correspond to our notion of cognitive activity; and such a process Kant enables us to set aside, as not what he means by ‘pure activity’, since it involves time. Yet Kant himself would seem to have guarded sufficiently against this misapprehension by the rigour with which he rejects the attempt to conceive as prior in time, that which, if it determine the will, shews that will to a
G p. 299 [AK 4: 451, MG 98].
b
G p. 300 [AK 4: 452, MG 99].
200
The 1898 dissertation
be ‘practically free’. In this rejection he is quite consistent. ‘The action’ he says ‘so far as it is to be imputed to thought (Denkungsart, identified just before with ‘intelligible character’) ‘as its cause, nevertheless does not follow from it at all according to empirical laws, that is, so that the conditions of pure reason [precede it], but only so that the effects of pure reason in the appearance of the inner sense precede’.a In other words that which is to be regarded as the condition, or, as Kant calls it, cause, of the action, in so far as that action exhibits practical freedom, does not precede the action in time. The action is only preceded by the consequence, or, as Kant calls it, effect, of this ‘intelligible’ condition; and hence the action itself may be said to ‘follow from’ the condition as a conclusion follows from premises, but not to follow it in the time-order (cf. above p. 190). Now, in the case of moral action, this ‘effect’, which produces the action, is just the presentation of the moral law; and the intelligible condition of that effect is the moral law itself. Kant himself allows that this effect or presentation, must always be present in human volition; and, what I wish to maintain, is that this is all that the analysis of human volition as such can ever shew to be present. I have examined the confusions, upon which Kant’s contrary view that the law itself is somehow to be obtained by analysis of volition, that it is given by a ‘pure Will’ or ‘practical Reason’, seems to be based; and those confusions seem sufficient to explain the view and to shew that, for Kant at any rate, it was baseless. The point at issue is this: Whether ‘will’ can be understood at all as other than a form of ‘activity’; and whether, if it be an activity, it must not be conceived as essentially conditioned by time, and therefore, in Kant’s language a mere ‘appearance’. If it be a mere ‘appearance’, the conception of a ‘pure Will’ is nonsense; and ‘will’ cannot be ascribed as an attribute to anything real – either to God or to the Transcendent Ego.99 That ‘will’ is a form of ‘activity’ has, I suppose, never been disputed. Kant himself, as we have seen, refers us, for our notion of pure Will, to the pure activity of the Ego. This pure activity must, according to him, be timeless. The question is therefore whether a timeless activity has any meaning. This conception seems again to rest on the notion of something intermediate between the relation of reason to consequent and cause to effect. If the relation of reason to consequent is to be described as a timeless a 99
R.V. p. 381 [B579, GW 542]. The next paragraph is rather different from the corresponding paragraph on pp. 71–2 of the 1897 dissertation.
‘Freedom’ in Kant
201
activity, then we understand what is meant. It may even be admitted that some existent not in time, may stand in a timeless relation to some existent that is in time; and, though there seems no evidence that there is such a timeless existent, and though we can only imagine its relation as similar to that of reason and consequent, this relation too may possibly be meant by a timeless activity. But even here any reason for identification of our supposed existent with Will seems wholly wanting; and activity seems an extremely misleading name for its relation to other existents. My conclusion, then, is this: That ‘will’ is only a special form of natural causality, or rather, a natural causal process, where the cause is of one definite sort. It is a special form of natural causality, just as explosion of gunpowder by a match is one special form of natural causality, and explosion of gunpowder by percussion is another. And, that on which I wish to insist, is that voluntary action, of whatever sort, whether autonomous or heteronomous, exhibits ‘freedom’, in the sense which I have hitherto explained as essential to Kant’s notion, no more and no less than gunpowder explosions or any other natural process whatever. It seems, indeed, strange that this conclusion from his doctrine should have escaped the notice, both of himself and others, to the extent to which it has. For he repeatedly asserts that for every ‘appearance’ we must suppose an intelligible ground (the Ding an Sich) and it is just this dependence of the cause of his actions on an intelligible ground (the Transcendental Ego), which he describes as constituting man’s practical freedom. Moreover even the identity of the Ding an Sich and the Transcendental Ego has been suggested by him and accepted by others; though this would not be necessary to justify the inference, since the dependence on an intelligible ground is by itself sufficient for practical freedom. When this is acknowledged, ‘practical’ freedom disappears altogether, as something intermediate between natural causality and transcendental freedom. For, as Kant himself says, nothing intermediate is possible; only two sorts of causality can be conceived at all. ‘Freedom’, then, for Kant means only ‘transcendental freedom’, and ‘transcendental freedom’ is not ‘practical’, in the sense that it is inseparably connected with ‘action’ alone. It is true that actions are dependent on ‘transcendental freedom’, but that is only because it is the relation which holds between the empirical causes of those actions and the transcendental ground of such causes; whether sensible objects produced effects, and so indicated their right to be considered practical (as they always must), or not, they would equally be results of ‘transcendental freedom’. The degree to which Kant himself was forced to recognise the unpractical nature of his conception of Freedom, is singularly illustrated by a passage
202
The 1898 dissertation
in the ‘Metaphysics of Morals’ to which I have referred above (p. 189; M. p. 23).100 He here declares that ‘Will’, which he has hitherto regarded as identical with ‘pure practical Reason’, and so as that which is alone endowed with ‘Freedom’ in his special sense, cannot be called either ‘free’ or the reverse, because it is not ‘susceptible of compulsion’. This ‘susceptibility of compulsion’ implies subjection to natural law, and, as so subject, he declares that human ‘choice’ (Willk¨uhr) may be called ‘free’. He would seem, therefore here to recognise that ‘action’ can only be conceived as a timeprocess; indeed he says that ‘Will’ does not refer to ‘actions’ (Handlungen); and it is only because he sees that he would be departing too far from the ordinary use of ‘freedom’, if he disconnected it from action, that he now denies freedom to ‘Will’. The fact is that his previous doctrine has already departed from the ordinary usage, further than he himself was fully aware; and hence the inconsistency, with which he now tries to patch up the discrepancy. The true way of meeting the difficulty would have been, as has been pointed out, to insist on his meaning of Freedom as the true one, and to give up the special connection which he had hitherto asserted between it and human volition: to recognise that ‘Willk¨uhr’ was a mere ‘appearance’, and therefore not ‘free’, and that, that which was free, had not even so much connection with volition, as to deserve the name of ‘Will’. But there remains yet another aspect in Kant’s conception of Freedom, involving yet another confusion. Kant deduces the reality of Freedom from the reality of the moral law; and the moral law is what determines the only objective ‘end’ (Zweck). Our reason tells us that we ‘ought’ to do so and so; and it is merely, because we know that we ‘ought’, that we are able to conclude that we ‘can’, i.e. that we are free. ‘If I ought, I can’ does not mean for Kant, that only actions, which will follow upon my choice of them, can be reasonably included in any notion of duty. Kant fully recognises, what I shall try to shew later (p. 219), that the fundamental sense of ‘ought’ is that in which it prescribes an ideal, which need not be physically or psychologically possible.a The connection of ‘ought’ with ‘can’, is one of complete dependence of the latter on the former. What I ought to do, or what ought to be, is ‘morally possible’,b even if, according to natural laws, and so far as experience enables me to predict, it never will happen. This is what Kant means by the ‘Primacy of the Practical Reason’. It is not that ‘Will’, as equivalent to Practical Reason, is to be regarded as a 100
cf. P.V. p. 101 [B579, GW 542]. AK 6: 226, MG 380.
b
P.V. p. 61 [AK 5: 58, MG 186].
‘Freedom’ in Kant
203
the fundamental reality; but that the Practical Reason, as prescribing the moral law, enables us to enlarge our knowledge of reality in a way that the Speculative Reason alone could not do. From the fact that a ‘kingdom of ends’ ‘ought to be’, we can conclude that a ‘kingdom of ends’ is not only possible but necessary, although, if our knowledge were confined to experience alone, we should be bound to conclude that it was neither actual nor possible. Now when Kant takes the moral Imperative, as such, for the primary ‘fact’ in the metaphysics of ethics, he seems to have been influenced by his theory that will is more than a psychological faculty. Will is no doubt the source of the imperative mood as distinguished from the indicative. But, if the above analysis of will has been correct, it can not possibly be the source of the universality, which distinguishes the moral imperative from every other. The will is capable of commanding, as of doing, what is wrong as well as what is right; and so far as it, and therefore the imperative form, is concerned, an injunction to wash your hands is precisely the same thing as an injunction to pursue what is good. The question what ought to be willed, cannot be answered by an analysis of will itself; for such analysis can never tell you more than what is willed, which differs with different times and different persons. Kant rejects the theory that ‘the good’ is the foundation of ethics, precisely because he thinks it can only be directly determined by an appeal to the feelings, which vary subjectively with individuals; and it is the similar subjectivity of the will, which seems to me to condemn his own ‘imperative’. He was perhaps misled by his desire to explain the ambiguity of the term ‘law’ into thinking that the only possible way of distinguishing ‘moral’ from ‘natural’ laws, was the implication of command in the former. What he meant by his conception of a ‘pure’ will, was a will which should command consistency in your view of what ought to be done – a consistency which is necessary to ‘reason’ as such. But all that a ‘pure’ will could really mean, would be, a will which was consistent with itself in the fact that it always commanded, not in respect of that which it commanded. We may admit that the former notion can be obtained by an analysis of will; but it would be by no means sufficient to give a moral principle. It is true that will cannot but be consistent with itself, in that it always commands; but what Kant did not see, was that it was a consistency in the object commanded a that he was after all requiring. It is true that this consistency is directly given as no more than a formal a
cf. P.V. p. 114 [AK 5: 108–9, MG 227]. He does not see that the mere notion of ‘ought’ is an object or matter of a special sort, though, indeed, not an object of intuition.
204
The 1898 dissertation
condition of good willing; but it is a formal condition of good willing, and not of willing in general. Kant’s distinction between ‘formal’ and ‘material’, is therefore wholly misleading. Thus, what really distinguishes a moral law from laws of nature, is that it expresses the connection of natural objects, not with one another, but with the notion of ‘end’ or ‘goodness’. Its primary form is ‘This is good’ or ‘this is an end-in-itself ’ or ‘this ought to be’; the command ‘do this’ is no more than a corollary from such a judgment. It is a ‘law’, because such a judgment, if true at all, expresses a universal truth; whereas subjective maxims, cannot be interpreted into more than a particular statement of fact – ‘I mean to do this’ or at most ‘I think you ought to do it’. To try to obtain this objective validity out of ‘Will’ – ‘this ought to be willed’ out of ‘I or you or all the world, do, or can, or must will this’ – is a procedure similar to that of trying to prove ‘this is true’ from the fact that all the world believe it. The assertion of ‘goodness’ claims rationality in the sense explained above (pp. 192–3), precisely as does the assertion of truth: the first has no more connection with volition, than the latter with cognition; both rest on the same ‘Theoretic Reason’ – if we are to adhere to the custom of ascribing them to ‘reason’ at all. If, therefore, we examine Kant’s first expression of the moral law ‘Act only on that maxim whereby thou canst at the same time will that it should become a universal law’, we get the following results. The significance of the expression rests on the words ‘thou canst will’. But there seem no limits to the empirical possibilities of willing. It is possible to will one thing one day, and the opposite the next. It is only, if the question has been begged, by a previous assumption that will cannot be inconsistent with itself, with regard to the objects willed, (in respect of always being will, it must certainly be consistent) that this expression will serve for what Kant means. Kant assumes that I cannot will a thing to be a universal law, to which I might afterwards wish an exception. Perhaps, I cannot rationally; but what is to compel me to will rationally? It is admitted that most of us very seldom do. To set up this as the fundamental principle of morals is as if one were to find the basis of epistemology in the principle ‘Think that only to be true, which you can at the same time think to be universally true’; whereby it would be assumed that no one can think otherwise than rationally. The insertion of ‘At the same time’ only makes matters worse, by involving the theory (which has sometimes been held) that the principle of contradiction means no more or less than that two contradictory propositions cannot both be true at the same time. Moreover our epistemological principle would have just as much right to the imperative form, as the moral one.
‘Freedom’ in Kant
205
That form is in each case totally irrelevant to the substance of the principle. The reason why Kant uses it is precisely the same fallacy, which he betrays in ‘thou canst will’. He imagines that the objective validity of the principle proceeds from the nature of the will itself, and hence represents it as so proceeding. In fact, however, this reference to will merely conceals the fact that he is dealing with the notion ‘ought’; and the only essential element in the expression is that of universality. Its whole significance can, therefore, be rendered in the proposition ‘You ought to do that only which ought universally to be done.’ Now, if in this way we reject the theory that the principle of Ethics has essential reference to will, we thereby give up the precise connection, which Kant found, between ethics and freedom. Their connection, for him, was that the moral law, necessarily implying will, yet obviously not identical with the law which governs the natural course of human volitions, could have no meaning unless we suppose it to be the guiding principle of another or pure will, which is hence possessed of that ‘intelligible causality’ which he calls freedom. Unless therefore such a free being were theoretically possible, the moral law would actually be contradictory to the course of nature, and we could not say it was reasonable to obey it. The ‘Critique of Pure Reason’, however, shewed it to be possible; and, there being, therefore, nothing against the conclusion to which the moral law would lead us, we can accept its validity unconditionally and along with it the reality of that freedom which could alone make it valid. It is, thus, only because there is a free will, that there is a moral imperative. But if, as I have tried to shew, the principle of ethics is not an imperative, there would seem no need for ‘free will’ or ‘freedom’ to ground it. That principle amounts merely to this, that there is an objective ‘end’, something which ought unconditionally to be, an absolute good. And Kant, though he does express the moral law in terms of these notions, yet thinks them to be merely derivative from that primary imperative, and connected with ‘Freedom’ only through it. There is, he reasons, an unconditional prescription: there must therefore be something unconditionally prescribed, and again an unconditional prescriber. The former is that which is absolutely good, or the sole objective ‘end’, and the latter is pure will. But, says Kant, there is none among the objects of experience which will answer to the former notion, since none is fit to be a universal object of pursuit. The only objective end must therefore be the only object, which has such universality, and whose reality at the same time is vouched by the moral law itself – namely the prescriber or pure will. Hence pure will and its object are the same: as is expressed in the famous saying with which Kant opens his ‘Foundations for the
206
The 1898 dissertation
Metaphysics of Morals’; ‘There is nothing anywhere in the world, nay, nothing at all that is conceivable even outside it, that could be considered good without limitation, except only a good Will ’. Thus ‘good’ or ‘end’ is for Kant nothing but the necessary object of a necessary ‘will’, which object can be none other than that ‘will’ itself.101 ,102 But if we reject, as seems necessary, Kant’s own method of stating and establishing the connection between Freedom and Good, it still remains to ask what sort of relation between the two is implied in his doctrine, and what can be said for the reality of such a relation. Kant’s assertion, stripped of its misleading references to Will, would seem to amount to this: What is good must be the effect of a free cause. This statement of relation is implied in his argument that the Moral Imperative must be taken as the datum from which we can infer the reality of Freedom. By the Moral Imperative we know that something is good, and we can infer from this that something is free, because the notion of freedom is logically involved in that of goodness. It will hence follow that any good action must be free, but it will not follow that any free action must be good. Free stands to good, as does, for example, the notion of triangle to that of isosceles triangle. It is true that if there are isosceles triangles, there must be triangles; but it is not true conversely that if there are any triangles, there must be isosceles. Kant’s argument is of the form: We know (as moral datum) that there are isosceles triangles, and we know (from the first Critique) that there may be triangles, and hence we can infer that there are triangles, but not that all triangles are isosceles. I think this may be fairly said to represent Kant’s general view of the relation between freedom and ethics, when his doctrine of will is left out of account. But when all his various statements are considered, the matter turns out to be by no means so simple, as this would lead us to expect. If this were all, Kant might consistently maintain that bad actions, as well as good, were ‘freely caused’ by the intelligible character, although that contention would need a separate proof. But the inconsistency of Kant’s position is shewn in that he hesitates to maintain this. He does, indeed, assert that the man, who commits a bad action, is free, because he could have committed a good one instead. But he nowhere, I think, contends that the man who does a good action is free, because he could 101 102
Moore adds in the margin: ‘Is this good will “pure”?’ At this point the remainder of chapter iii (1898) diverges from the 1897 passage which Moore has been repeating – not surprisingly, since in the 1897 passage which follows Moore attempts to establish the idealist thesis that ‘That, that which is ultimately real, appearing to us as the reason of whatever happens, is also necessarily “good”, and alone absolutely good’ (p. 78).
‘Freedom’ in Kant
207
have done a bad one.103 This shews that Kant did not hold good and evil actions to bear the same relation to Freedom; and it is for this reason that it seems impossible to accept Professor Sidgwick’s accounta of the discrepancy in Kant’s notion of Freedom. Professor Sidgwick holds that Kant confused the two notions of ‘Neutral Freedom’ or ‘freedom of choice between good and evil’ and ‘Good Freedom’ or the doctrine that ‘man is free in proportion as he acts rationally.’104 But, in seeking to establish the former doctrine as sometimes implied by Kant, Professor Sidgwick relies only on Kant’s assertion that the evil agent might have done otherwise, which only implies that he is free in that he might always do the good, not that his evil action is a result of a free choice. Moreover, as we shall endeavour presently to shew, ‘Neutral Freedom’ is indistinguishable from ‘Capricious Freedom,’ which, as Professor Sidgwick acknowledges, Kant ‘undoubtedly repudiates.’105 Kant’s inconsistency lies, then, in the first place, rather in the fact that he does not hold good and evil actions to be equally results of Transcendental Freedom. The doctrine that they are so, would have resulted from the general account, given above, of what such Freedom must be held to be. And in holding this doctrine, he would not have been committed to ‘Capricious Freedom’. For ‘Capricious Freedom’ is directly contradictory to Determinism, whereas ‘Transcendental Freedom’ [is not]106 , since in it the free cause is not a cause in the natural sense, but something partaking at least as much of the nature of a reason. That Kant does not expressly ascribe bad actions to Freedom, would seem to be due to the fact that he has not clearly defined his position as to its relation to good ones. He certainly has a tendency to hold that only good actions are free, as well as his fundamental proposition that only free actions are good. It seems doubtful, indeed, whether the first proposition can be ascribed to him as definitely as Prof. Sidgwick (loc. cit)107 ascribes it.b There will hardly be found any passages, more decisive for this view, a
b
103 104 106 107
Mind, XIII. 51. [This is Sidgwick’s paper ‘The Kantian Conception of Free Will’, Mind, 13 (1888) 405–12; the paper is included as an appendix to the fifth, and subsequent, editions of ME, pp. 511–16 and Moore’s page reference – ‘51’ – must be a mistaken reference to p. 511 of this edition.] Neglecting, for the moment, a double ambiguity which we shall try to shew to be inherent in Prof. Sidgwick’s conception of ‘rational’ action. Moore adds an X in the margin next to this sentence, probably to indicate that the point made is important. 105 Ibid., p. 512. ‘The Kantian Conception of Free Will’, ME 7th edn, p. 511. As it stands this sentence is constructed oddly, but this emendation makes sense of it. I.e. ‘The Kantian Conception of Free Will’, ME 7th edn, esp. pp. 514–15.
208
The 1898 dissertation
than those which Prof. Sidgwick quotes; and yet all of these seem capable of another interpretation. Take, for instance, the first: ‘a free will must find its (sic) principle of determination in the Law’ (P.V. 30).108 This passage, quoted in full (unless there is some mistake in the reference) runs ‘a free will, independently of the matter of the Law, must nevertheless find aa principle of determination in the Law’. Now from the context it appears that Kant’s argument here is: Supposing that there is a free will, it must be capable of determination, and must therefore be determined by the form of the Law, since we know nothing else fit to determine it. This argument rests on Kant’s negative conception of what is free, as something incapable of determination by appearance. The form of the Law is accordingly the only thing, we know of, fit to determine what is free, since it is the only thing we know of, which is not a mere appearance. This is sufficient for his purpose, in showing that his formal Law is the only true principle of Morals we can find. But he would not consider himself to have proved that nothing else, such as we do not know, can determine a free will. This is plain from the manner in which he proceeds to say that he does not now ask whether the moral law and freedom ‘are really different, and it is not rather true that an unconditional Law is merely the self-consciousness of a pure practical reason, and this (reason) wholly at one with the positive notion of Freedom’. He here admits he has not proved that the moral law is the only object of pure practical reason. He plainly thinks it is, but he maintains, as usual, a cautious reserve. The moral Law does determine a free will, but it may not be the only principle of determination. And this language is quite in harmony with his refusal, in the Metaphysic of Morals (p. 23),109 to admit that freedom can be defined as a power to choose either good or evil actions (which, by the way, he identifies with ‘liberty of indifference’). We can, he says, only know it as a negative property in us; and though, by specially saying that the possibility of evil choice is incomprehensible, he implies that the choice of good is not equally so, he again does not commit himself. Similarly, when Freedom is said to be ‘a causality according to immutable laws’, this need not be taken to mean that is only in accordance with the moral law, but it may also be a causality according to such laws, as will explain those evil choices, which are at present incomprehensible. a 108
My italics. AK 5: 62, MG 162.
109
AK 6: 226, MG 380.
‘Freedom’ in Kant
209
It may be admitted, then, that Kant does not see that there are as good grounds for saying that we are free to do evil as that we are free to do good; although it seems difficult to prove against him the assertion that we are only free to do good. We have now to ask what are his reasons for asserting that we are free to do good. How does he propose to shew that if an action is good it must be free? How, in short, does he establish the existence of Transcendental Freedom on an ethical basis? His argument on this point seems simply to reduce itself to the assertion ‘I can, because I ought’. His view is that, if anything ought to be, it must be possible. It may be admitted that this is true in the widest sense of the word possible, in which to call a thing possible, merely asserts that it is a concept, or, if it is a complex concept, does not involve a contradictory proposition. This kind of possibility Kant calls ‘logical possibility’, and this is not the sense in which he asserts that the good is possible. He states that ‘practical’ and not only ‘theoretical’ ‘sources of knowledge’ can furnish a ground for ascribing ‘objective validity’ or ‘real possibility’ (R.V. p. 23 note).110 At the same time he asserts, as we have seen, that what is morally possible is not necessarily naturally possible. He regards moral possibility, then, as some third kind, which is neither ‘logical’ nor ‘natural’. Now if his proposition ‘I can, because I ought’, is to be significant, moral possibility must have some meaning independent of the meaning of ‘ought’. But Kant does not shew us any such meaning, nor does it seem possible to shew any except those of logical and natural possibility. His moral argument for the reality of freedom would, therefore, seem to reduce itself to the tautology ‘I ought, because I ought’ – a tautology, which it is easy to overlook, because ‘must’, which expresses necessity and therefore includes possibility, is also used in German, as in English, for a synonym of ‘ought’. Accordingly in his actual applications of the notion of ‘moral possibility’, he appears to use it sometimes as equivalent to logical and sometimes as equivalent to ‘natural possibility’. Each of the two meanings becomes in turn the more prominent, according as he lays stress on the no¨umenon as cause, or reason, or freedom as transcendental or practical. With this the whole construction of Kantian Freedom seems finally to fall to the ground. Transcendental Freedom, upon which the main stress of the theory lies, is possible, as a chimaera is possible, but in no other sense. There is no ground for asserting the existence of anything timeless, but it is not contradictory to natural laws to suppose it. Secondly, if any such 110
R.V. p. 23 (Bxxvi, GW 115).
210
The 1898 dissertation
timeless existence is asserted, it cannot, without contradiction, be supposed to have any closer connection with the human will, than with any other phenomenon. And finally, neither the moral principle itself, nor action in accordance with it, can be taken in any way to prove or to depend upon the existence of Transcendental Freedom. Kant fails to prove that there is any other than the Deterministic way of regarding human action; but the basis of Ethics is not thereby, as he thought, endangered.
chapter iv
Freedom
In the common notion of freedom the most universal characteristic seems to be the absence of external constraint, whether asserted to impel or to prevent.111 Where the immediate cause of a motion or change seems to lie in the thing which moves or changes and not in anything outside it, there, in a sense at all events, freedom is predicable. But this is a notion which is obviously not limited to human actions. Many of the movements and changes of animals and plants have their proximate causes in the things themselves; and the same might probably be said of any body in so far as it moved in accordance with Newton’s second Law of Motion. It is thus we seem to talk of ‘free as air’, or of the wheels of a watch moving ‘freely’. But there is an obvious defect in this wide notion, in that the limits, whether spatial or temporal, of any group we may take for our unit or thing, are always more or less arbitrary. A watch may be moving freely when its spring is driving it; but the movement of any one of its wheels is not free, because the wheel is driven by the spring or by another wheel.a And, again, there seems no reason why we should single out the proximate or immediate cause for such pre¨eminence, nor anything to determine how far back in the past a cause ceases to be proximate. It is difficulties of this sort which seem to have gradually tended to restrict the notion of freedom to men; because in man the notion of self is far more striking than elsewhere and the distinction between the internally and externally caused, therefore, primˆa facie more satisfactory. The difference between himself and anything else whatever is more constantly forced on a man’s notice and more practically important to him than any other difference; and it is therefore not unnatural that the notion of free, in the sense of selfcaused, action, whether or not it is originally derived from his experience a 111
cf. on this subject, P.V. p. 100, foll [AK 5: 96, MG 217]. All of this chapter, except for the paragraph on pp. 227–8 and the final paragraph (pp. 230–1), comes with only minor changes from pp. 38–57 of the 1897 dissertation.
211
212
The 1898 dissertation
and transferred anthropomorphically to other things, should at all events be more widely applied and less easy to dispense with in his own case and that of other beings like him, than elsewhere. Now the vulgar doctrine of Free Will, as ‘Liberty of Indifference’, seems to be in the main an attempt to raise this distinction between self and the world entirely above the level of an arbitrary distinction. It was seen that this could not be done, if the self were regarded as a part in the causal chain of events, since it must then be subject to the infinite divisibility inherent in time, and the ultimate causal unit remains as arbitrary as any unity of time. It was therefore maintained that man’s soul was an agent undetermined by previous events in time; it was the absolutely simple unity of Rational Psychology, and as such distinguished from all natural objects, which were always both themselves divisible into parts and also incapable of certain discrimination from an ever wider whole. Such a notion of a finite uncaused cause inevitably follows from the attempt to distinguish within the world of experience cases of purely internal and purely external causation. And there are good reasons why the human will should have been taken as the final instance of a cause which is not also an effect. The progress made in the analysis of mental processes has been very slight in comparison with that made in physical science (1) because of their greater complexity (2) because experiment in psychology must be either indirect or encumbered by the fact that the observed is also the observer and (3) because subconsciousness must be taken into account. And the region of the incompletely known is the favourite abode of a metaphysical monstrosity. In plain language, where facts are not completely understood, some short-sighted metaphysical theory is generally introduced as affording an easy road past the difficulties which stand in the way of thorough investigation. And, secondly, apart from the general difficulty of establishing exhaustive causal laws, which applies in a less degree to physical science also and prevents certain prediction even there, there seems to be a real reason, which from the nature of the case can never disappear, why human volition should produce the illusion of so-called Freedom. It is this, that, in virtue of the deterministic hypothesis itself, the knowledge that a certain course of action was about to be must always exert some influence upon the course actually pursued, and so make the result different from what was foreseen after a consideration of all the other elements that would contribute to it. And even if the fact of this knowledge were taken to account in the calculation, and the prediction modified accordingly, the knowledge of this modification would again introduce a new element, which
213
Freedom
would require a fresh calculation, and so on ad infinitum. This seems to be a difficulty inherent in the double nature of the mind as subject and object – a difficulty which makes it possible to pronounce a priori that complete prediction of the results of mental process must always be impossible. It is a difficulty which does not apply to prediction in the physical world of space, considered, as seems necessary at present, in abstraction from the world of mind. It could only modify our view of that, if the real connection of body and mind were fully discovered. As it is, mental processes, though obviously corresponding to physiological [ones], and useful for their investigation, have only too much the appearance of a totally independent world from the point of view of causality and reciprocity. So that the distinction is justifiable, when we say that the results of human volition, alone among causes, must of necessity remain incapable of prediction. And this fact, along with the greater empirical difficulties of prediction in the case of mind, seems sufficient to account for the illusive belief that the will, at any rate, is free, though it be admitted that nothing else is. The failure to discover a cause in any particular instance of itself encourages a belief in the uncaused; and when to mere failure is added an absolute impossibility of discovery, the case is naturally strengthened. That the belief in uncaused volition is illusory, the progress in scientific method, with the resultant growth of empirical psychology, has rendered it more and more difficult to doubt. Yet this fact by itself would be no argument against Free Will. For, in however many instances causation were proved, though that would, on Hume’s principle, be a cause of our expecting it in others, yet it would not be, by itself, any reason for that expectation. An inductive argument always needs, as empiricists put it, to be supplemented by the assumption of the uniformity of nature. And that this assumption is not in this case an assumption but an a priori necessity, may, I think, be considered to have been sufficiently proved by Kant’s argument in the Analytic. He there shews that every event must have a cause if there is to be an objective succession in time; and such an objective succession is certainly presupposed by all actual experience.112 Accordingly Kant himself fully recognises the a priori certainty of the Deterministic view,a and it seems inevitable to agree with him. As for ‘the immediate affirmation of consciousness in the moment of deliberate action’b . which is asserted to stand against Determinism, great a 112
cf. on this subject, P.V. Pref. p. 12-end [AK 5: 12–end, MG 146–7].
b
ME p. 65.
Moore adds a large question mark in the margin next to this sentence in the manuscript.
214
The 1898 dissertation
care is needed in deciding what it is that consciousness then affirms. Professor Sidgwick expresses it thus. ‘Certainly when I have a distinct consciousness of choosing between alternatives of conduct, one of which I conceive as right or reasonable, I find it impossible not to think that I can now choose to do what I so conceive – supposing that there is no obstacle to my doing it except absence of adequate motive – however strong may be my inclination to act unreasonably, and however uniformly I may have yielded to such inclinations in the past.’a,113 Now with regard to this statement, it is to be noted first that what consciousness is said to affirm as on the Libertarian side in the Free Will controversy is only the possibility of choosing the reasonable course; the note appended seems to be intended to defend this limitation.114 This note runs: ‘It is not the possibility of merely indeterminate choice, of an “arbitrary freak of unmotived willing”, with which we are concerned from an ethical point of view; but the possibility of action in conformity with practical reason.’b The same limitation is also introduced in the passage quoted in our last chapter (p. 181; Methods, p. 62): ‘Is there always a possibility of my choosing to act in the manner that I now judge to be reasonable and right?’ Now, what precisely is meant by ‘the possibility of merely indeterminate choice’ mentioned in the note, I am not quite sure that I understand; especially as, in the text, the choice, of which the possibility is under consideration, is pointedly distinguished as one to which the presence of adequate motive cannot be assumed, so that the question presumably is of the possibility of unmotived choice, in so far as there ceases to be a question, if the motive is alone sufficient to produce the choice. We are considering the possibility of a choice ‘in conformity with practical reason’, where something other than the motive supplied by that conformity is necessary to produce the choice; and where, therefore, the choice if possible must be assumed to be to some extent unmotived. a b
113
114
ME p. 65. Note that ‘action in conformity with practical reason’ seems here to be identical with the ‘doing what I conceive to be reasonable’ of the text. [This note is altered in the seventh edition: instead of the phrase ‘the possibility of action in conformity with practical reason’, Sidgwick now writes ‘the possibility of choosing between rational and irrational motives’.] The phrase ‘supposing that there is no obstacle to my doing it except absence of adequate motive’ which Moore cites does not occur in the seventh edition of ME; Sidgwick here writes instead ‘supposing that there is no obstacle to my doing it other than the condition of my desires and voluntary habits’, presumably in order to emphasise the presence of other motives. It is conceivable that Sidgwick was motivated to make this change by Moore’s discussion here of the earlier version which he will have read in his capacity as one of Moore’s examiners. This is note 1 on ME p. 65.
Freedom
215
But, however this may be, I do not see how it can be denied that ‘we are concerned from an ethical point of view’ with the possibility of a wrong or unreasonable choice – a choice not ‘in conformity with’, but contrary to ‘practical reason;’ and, whether such a choice be an ethical concern or not, Professor Sidgwick has certainly elsewhere stated that the question of its possibility is essential to the Free Will controversy. For instance, in the passage, just quoted, on p. 61,115 he says the question is whether ‘my voluntary action, for good or for evil, is at any moment completely caused’; though, as we have seen, in the following sentence, he drops the ‘for evil’, and inquires only into the possibility of my choosing to act rightly. So too, on p. 58, he says ‘It is obvious that the Freedom thus connected with Responsibility is not the Freedom that is only manifested in rational action, but the Freedom to choose between right and wrong which is manifested equally in either choice’. In fact, when Professor Sidgwick on p. 65 limits the possibility of Freedom to the possibility of choosing rational actions, he seems to have fallen partially into the error of which he accuses Kant and his followers – the error of ‘identifying Rational and Free action’ (p. 57). For he seems to suppose that only rational action can be free, though he does not complete the identification by also implying that only free action can be rational. It is noteworthy, however, that it is only against the first of these two statements, the one which he himself subsequently seems to adopt on the immediate affirmation of consciousness, that he directs his attack, as the supposed view of Kant; whereas it is the latter, which, as has been said (Chapter iii, pp. 190–3) Kant is chiefly concerned to maintain. In fact even the statement which Professor Sidgwick quotes, as of Kant’s disciple, that a man ‘is a free agent in so far as he acts under the guidance of reason’, does not imply, as Professor Sidgwick thinks it does, that irrational action may not be also free; but only that all rational action is free, which is equivalent to ‘only free action is rational’, but not to ‘only rational action is free’. We may therefore use the weapons supplied by Professor Sidgwick himself on pp. 57–59, but there misdirected against a Kantian view, to defend ourselves in considering the possibility not only of our choosing the right course but also of our choosing the wrong one, whatever our previous character may have been. But it seems worth while, in passing, to dwell a little longer on the apparent inconsistency with which after maintaining that wrong must be just as free as right actions he himself seems immediately to 115
Moore is here referring to ME (5th edn) both here and in the subsequent page references. As we have mentioned earlier (see p. 117 note 2) there are many slight differences between the fifth and the seventh editions of ME, but none that is of significance for Moore’s argument here.
216
The 1898 dissertation
adopt the one-sided view which he has been attacking. Thus he concludes the anti-Kantian paragraph on p. 59 with the words ‘Can we say, then, of the wilful wrongdoer, that his wrong choice was ‘free’, in the sense that he might have chosen rightly, not merely if the antecedents of his volition, external and internal, had been different, but supposing these antecedents unchanged? This, I conceive, is the substantial issue raised in the Free Will controversy’; and he does not seem to see that the whole tendency of his foregoing argument is to shew that it would be just as pertinent to ask: ‘Can we say of the wilful rightdoer, that his choice was free, in the sense that he might have chosen wrongly etc.’? Now it is curious that in Aristotle’s discussion of the voluntary, wherever he touches the point, he always emphasizes the remark that vice must be considered to be as voluntary as virtue, not vice versˆa; as if it were a wide-spread notion in his day that virtue was voluntary, but vice necessitated. This, indeed, was probably due to the Socratic doctrine, that all vicious choice is merely due to our ignorance, and since to choose ignorance would itself be a vicious choice, to something beyond a man’s control. In Christianity, on the other hand, just the opposite doctrine prevails, that sin is due to our free will, and virtue solely to the merit of Christ; though, there, indeed, the other one-sidedness also appears, hopelessly entangled with it. But what is to be noted is the general tendency not to accept both doctrines at once, but each separately. For this, I think, whichever be the one taken, is evidence of an almost inevitable feeling that good and evil are somehow not on a par with respect to reality; that good is somehow more natural and therefore not in so much need of notice or explanation as evil. It is the Origin of Evil which has always appeared as a problem; as if the Origin of Good were not primˆa facie equally inexplicable. With the Greek, who has no personal God to start with, as the author of nature, the explanation is found in blind natural necessity: man being the apparent best, it seems less obvious to suppose him cause of the worst. Whereas the Christian’s notion of God forces him to find an explanation of what is contrary to God, and therefore cannot be regarded as directly created by him, in a power approaching nearer to God’s than that of blind nature, and so in man. The one-sidedness of our judgments is also illustrated by the far greater provision made in almost every state of society for punishment than for reward, and the general feeling that we need be much more careful whom we single out for blame than whom for praise: and here again the one-sidedness may be connected with either of the two opposite theories, according as punishment is regarded as corrective and deterrent, which harmonises with the Greek view, or as retributive, as suits the Christian. Nevertheless, it is plain, that this one-sidedness cannot be
Freedom
217
defended on such ethical principles as Professor Sidgwick’s, any more than it could on Aristotle’s. For those who take the world of experience to be real, there can (as Kant saysa ) be no freedom; and similarly, if freedom be nevertheless maintained, there can be no ground for asserting that good is more real than evil, even if there be a greater quantity of things which are good than of things which are evil. The essentially Kantian view that only free action is rational, in the specific sense which he gives to it, would seem, though it cannot be regarded as proved, to hint at a possible explanation of this general prejudice. But, let us now consider, how the admission that wrong choice must be considered by the ordinary Libertarian to be as much the work of Free Will, as right choice – an admission which, as we have seen, Professor. Sidgwick is urgent to justify against Kant – affects Professor Sidgwick’s statement of what he holds to be ‘immediately affirmed by consciousness in the moment of choice’ on the Libertarian side. This statement, as we have seen, runs ‘that I can now choose to do what I so conceive’ (i.e. conceive as right and reasonable) – ‘supposing that there is no obstacle to my doing it except absence of adequate motive.’ Now the condition here asserted in parenthesis implies that if the doing of a thing is impossible for other reasons than the absence of motive, there can be no question of choosing to do it. In other words the only Free Will, which is at all affirmed by consciousness, is only free to choose actions, to the performance of which there is no obstacle but absence of motive. And this restriction to what we may call physically possible actions, seems plausible when the question is also restricted to reasonable actions; for to the ordinary reader it would seem that an action to which there were insuperable physical obstacles could not be called a reasonable one to choose. But we have seen that the restriction of the question to reasonable actions cannot be defended; and, if we recognise this and accordingly put the question as to the possibility of choosing wrong and unreasonable actions, it is no longer so plain that an action, which is physically impossible, is not an unreasonable one to choose. Part therefore of the plausibility of Professor Sidgwick’s statement as to the affirmation of consciousness, seems to be due to his erroneous limitation of the Free Will question to reasonable actions. But this objection is not fundamental. It serves to lead up to another, which I believe to be of great importance, since it is based on a distinction, without recognising which it seems impossible to arrive at a clear decision in the Free Will controversy. It is obvious that my last argument depended a
R.V. p. 373 [B564, GW 535].
218
The 1898 dissertation
largely on the meaning to be given to ‘reasonable’ and ‘unreasonable’. Now Professor Sidgwick has discussed at length in Chapter iii the meaning to be given to ‘reasonable’. He distinguishes two senses of ‘ought’, an ‘ethical’ and a wider sense, and notes (p. 36 n. 4) that it is in the former sense that he will henceforward use the word. This ‘ethical’ sense, when it is said that an action ‘ought to be’ done, implies not only that it is reasonable in the sense in which Reason may apply the unanalysable notion of ‘ought’ to ‘a pattern to which no more than an approximation is practically possible’ (p. 35 n.), but also that the action ‘is thought capable of being brought about by the volition of any individual, in the circumstances to which the judgment applies’ (p. 35). Now it may, I think, be assumed that when an action is termed reasonable, Professor Sidgwick generally means by it that it ‘ought to be’ done in this narrower ethical sense, since his definition of reasonable conduct seems to be that which ‘ought to be done’. There is therefore included in the notion ‘reasonable’ as applied to conduct both that it partakes of the ‘ideal’ and also that it is ‘practically possible’. But unfortunately Prof Sidgwick does not tell us with regard to ‘unreasonable’ action, whether the negative is to be taken as denying both these elements or only one of them. If it is to be taken as denying both, then our previous argument not only destroys the plausibility of Professor Sidgwick’s statement of the Libertarian contention, but renders it manifestly unfair.116 For the contention that ‘I can now choose to do what I conceive to be unreasonable supposing that there is no obstacle to my doing it except absence of adequate motive’ becomes absurd, since to doing what I conceive to be unreasonable, there must always be the obstacle that it is physically impossible; and hence the form of expression Professor Sidgwick uses for the Libertarian contention with regard to reasonable action cannot be used for that contention with regard to unreasonable action. But I imagine that Prof. Sidgwick would adopt the other alternative, that the negative only denies that the action partakes of the ‘ideal’, not that it is thought to be practically possible. In this case, therefore, actions conceived to be reasonable and actions conceived to be unreasonable have this in common that they are both conceived to be practically possible. It is therefore possible that, in asking our question with regard to them, the answer may depend solely on this common property; and hence the fact of their being reasonable or unreasonable would be wholly irrelevant to it. The question would thus be wholly removed from the ethical sphere to which Professor Sidgwick wishes to confine it. 116
The rest of this paragraph and opening sentences of the next one are a modified version of the corresponding passage in the 1897 dissertation (p. 46).
219
Freedom
Our inclusion then of ‘unreasonable’ actions among the possible objects of free choice, if free choice there be, serve to make it plainer that the question whether it is possible to choose a good or a bad action involves a question entirely separate from that of the action’s practical possibility. The appearance of irrationality in an unreasonable action of Professor Sidgwick’s type may be due to the fact that, when I say ‘choose to do’, I do imply that, the choice is made, the action may reasonably be expected to follow from it. But this reasonable expectation cannot be admitted as an ultimate element in the definition of a reasonable action as that ‘which ought to be’. The ‘wider sense’ in which Prof Sidgwick admits ‘I sometimes judge that I ‘ought’ to know what a wiser man would know’ is not only unable to be ‘conveniently discarded in ordinary discourse’ (p. 35), but must be regarded as the fundamental one, if we are to get a clear notion of the issue in the Free Will controversy. It would, indeed, be a ‘futile’ conception, ‘if it had no relation to practice’ (p. 35 n); but we cannot, without begging the question with regard to Free Will, assume that this relation to practice consists in the fact that what is so judged is also thereby judged to be always practically possible. And this I think is also plain from what Professor Sidgwick himself subsequently says in his chapter on Free Will (p. 67). He here admits that, after we have made up our minds on the question whether I can do a thing, if I choose, ‘still the question remains “Can117 I choose what – if I can choose – I judge to be right to do?”.’ Here the ambiguous form ‘Can I choose to do?’ is at last discarded; and Professor Sidgwick explains that what he thinks his consciousness affirms is ‘that I can118 choose’. He adds, it is true, ‘within the limits above explained’. But my point is that by putting the question in this form he has actually removed ‘the limits above explained’. They appear, in the sentence just quoted, in the proviso ‘what – if I can choose 119 – I judge to be right to do’. But does not that proviso make the question into nonsense? For note that the question is, can I choose what I conceive to be right (p. 65). But if, before I can decide whether it is right, I have to decide that I can choose it, this question must be answered before it can be asked: which is absurd. Professor Sidgwick seems here expressly to state a relation of complete mutual dependence between the questions ‘Is this right?’ and ‘Is this a possible object of free choice?’, a dependence which he maintained in a qualified form on p. 35; and, if this holds, it is certainly impossible to answer the Free Will question in the form ‘can I choose what I conceive to be right or reasonable’. Surely the only escape from this dilemma is the view expressed above that from the 117
Sidgwick’s emphasis.
118
Sidgwick’s emphasis.
119
Moore’s emphasis.
220
The 1898 dissertation
definition of ‘reasonable’ as ‘what is in conformity with practical reason’ we must absolutely exclude all consideration of physical possibility: and this, I suspect, does make the Free Will question one of ‘an arbitrary freak of unmotived willing’, and yet a question which must be answered, before what Professor Sidgwick takes to be the question ‘from an ethical point of view’, can be asked. These considerations, I think, sufficiently prove that we cannot separate the Free Will question, in a supposed ethical sense, from the wider question; nor suppose that consciousness affirms the Libertarian doctrine in the former sense, without prejudging the latter. It may be that this makes the question even more unimportant for Ethics than Professor Sidgwick thinks it; but that will be afterwards discussed. There still remains the question: Does consciousness affirm in the moment of choice that I can choose anything which I judge to be good or bad? – if I may state the issue in that way, to get rid of the question-begging ‘choose to do’ and doubly questionbegging ‘reasonable’.120 It is, I think, in the first place, to be noted with regard to this affirmation, that it only concerns the possibility of choosing; that the free will controversy seems to have narrowed itself to the question of free choice. For it is only choice which distinguishes voluntary from nonvoluntary action, and the ordinary Libertarian would hardly maintain that non-voluntary actions could be free. But, on the other hand, the question is also seen to be a wider one than that which is ordinarily discussed – than is stated, for example, by Professor Sidgwick. For since, as has been shewn, the physical possibility of the action, which is the possible object of choice, cannot be considered to be a necessary element in constituting it good or reasonable, in the sense which is fundamental for ‘practical reason’, it seems hardly possible to exclude mere choices, such as that I should have the genius of Shakespeare, though, that I should have it, might be reasonably considered physically impossible. Even such a case as a choice to prevent the sun rising tomorrow can hardly be excluded from the class which Professor Sidgwick recognises as ‘a species of volition’ – ‘the adoption of an object of desire as an end to be aimed at’ (p. 60). For though, perhaps, no one but a madman would make such a choice, yet his choice would prove that it can be made: and we ourselves do often choose through ignorance what is impossible in this sense; the only reason why we do not choose what we also think impossible, seems to be not that we cannot, (either in the deterministic or libertarian sense) but that it does not seem worth while. The question, whether a choice will produce in any 120
Moore has marked out from the beginning of the paragraph to here in the MS, in very light pencil.
Freedom
221
degree the effect chosen, seems to be merely one for experience to decide, and we judge of it just as we judge of the probabilities and possibilities of events in the physical world. It does not seem to be concerned in the Free Will controversy, if the issue of that controversy be clearly stated. Locke and Humea , indeed, agree marvellously in their treatment of Liberty, both asserting that it means simply ‘a power to act, as we choose’. But it would seem to be for this very reason, that they are able to treat the Free Will controversy so cavalierly as they do. If the question were merely as to whether we did not sometimes do what we chose, it would, as they say, be obvious what answer we should give; but it would not, as they also say, be obvious that liberty in this sense was not contrary to ‘necessity’ since the question ‘Can we choose?’ would still remain unanswered. They both, it is true, leave an ambiguity even on the first point by not sufficiently considering what is implied in their notion of ‘power’; but, nevertheless, when they speak of a power to act, as we choose, they would appear to mean only as Locke says (§27)121 that the existence or not of the action is dependent upon our choice. In this part of free action, then, it may be admitted that they leave no room for anything contrary to ‘necessity’; since their notion is that the action is necessitated by the choice. Locke, however, sees that the point in dispute occurs not here but in the question: Are we free to choose? (§22).122 And this question he dismisses as absurd, on the ground that it means: Can a man will, what he wills? But it does mean this, unless his definition of freedom, as power to do what I choose, be already accepted. It would indeed be absurd to ask ‘Can I choose to choose?’, in the sense ‘Am I free to choose which of two alternatives I will choose?’ But Locke has no right to assume that this is meant by the question ‘Am I free to choose?’ That question may mean ‘Am I the original cause of my choice?’ and this he leaves undiscussed. Both Locke and Hume, therefore, neglect the point of the controversy by their definition of freedom. They have, however, done some service to the question, in as much as their treatment of it is a protest against that confusion of freedom ‘to do, if I choose’, and freedom ‘to choose’. Their defect is that they assume that it was an answer to the first only which was really wanted; and hence their contempt of the dispute. As a matter of fact, I am free, in the ordinary political sense, when ‘I do what I choose, because I choose it’, since there the immediate cause of my action lies in myself, i.e. in my choice. But that is not freedom in the a 121
Locke: Essay ii 21, 14 foll. Hume: Treatise iv p. 110 foll. Essay, book ii, ch. 21 §27.
122
Essay, book ii, ch. 21 §22.
222
The 1898 dissertation
sense demanded by Libertarians. What they wish to maintain is that the choice itself is caused only by a self, which is an uncaused entity; and this implies that where alternatives are presented, the choice between them is wholly independent of their previous habits, disposition etc. The question then is: ‘Does consciousness affirm, when alternatives are presented, that I can choose any of them that I think either good or bad’, which would seem to be equivalent to ‘any conceivable presented alternative’? And with this we come to the last ambiguity of statement which seems to me to stand in the way of our giving a clear answer to the question. Professor Clifford (Lectures on Ethics p. 327)123 rather ingeniously urges that, if the deliverance of my consciousness is to be ‘of any use in the controversy’, it must be ‘competent to assure me of the nonexistence of something which by hypothesis is not in my consciousness’ i.e. the subconscious mental elements which the Determinist must suppose to determine the choice. But it seems possible to surmount this objection by maintaining that it is enough, if consciousness can make a positive affirmation as to what is cause of the choice, without requiring it to prove exhaustively that nothing else in the world can be. If the man of science, before he enunciates a law, is always bound to prove, that no other elements besides those, whose constant connection with the effect has been observed by him, really contributed to it, no scientific laws have been discovered yet. If therefore consciousness does affirm that ‘I’ am the cause of the choice, that should be sufficient. But then the question arises what can it mean by ‘I’? Is it quite certain that when consciousness seems to affirm that ‘I can choose so and so’, it means more than ‘it is possible that such and such a choice will take place in my mind’? If it does not mean more than this, its affirmation is not against Determinism; since, as we have tried to shew above (pp. 212–13), even on the Deterministic hypothesis, it must always be entitled to affirm the latter proposition, even if it does not always exercise its right. For by saying that such and such a choice is possible I imagine we can mean no more than that we do not know but that it will happen; and even if ‘the uniformity of nature’ can be proved in such a sense as to justify an assertion with regard to any event whatever that it certainly will not happen, this can never be the case with regard to an event conditioned by a conscious forecast. The cases in which it is ‘on Determinist principles legitimate to conclude it to be certain – and not merely highly probable – that I shall 123
Moore refers here to the essay ‘Right and Wrong: the Scientific Ground of their Distinction’ that he had referred to earlier in fn. a on page 182; the passage Moore refers to occurs on pp. 143–4 of vol. ii of Clifford’s Lectures and Essays, London: Macmillan, 1901.
Freedom
223
deliberately choose to do what I judge to be unwise’ seem to be not merely ‘rare’ as Prof. Sidgwick says (p. 67) but absolutely non-existent. It seems, therefore, that the ‘affirmation of consciousness’ as against Determinism disappears on the attempt to make it precise. The attempt to find a more exact meaning for the vulgar notion of freedom has thrown us back upon the conception with which we started. Instead of free action being the action of an ‘uncaused self’, we have to be content with it as selfcaused action: anything may be said to act freely in so far as the immediate cause of its change lies in itself. We have now to analyse more carefully what is implied in this notion. It was one of Kant’s great meritsa in the Critique of Pure Reason to have pointed out that there is nothing absolutely ‘inner’ in the objects of experience, either of the outer or inner sense, either in nature or in mind. He gave the final blow to the doctrine of ‘essences’ and ‘faculties’, as principles of explanation, by showing that advance in scientific knowledge presupposed the complete interdependence of things; that all we can know for certain about them is their relation to one another. This indeed was one of his motives to his distinction of Phenomena and Dinge an Sich; for he could not avoid the conviction, though he could not justify it, that there must be something self-subsistent somewhere. But his main point was that if you treated natural objects as if they were self-subsistent, you could not escape the most unbearable contradictions. This was the ‘natural dialectic of Reason’. In the Critique of Judgment, however, he began to see that he had overemphasised the doctrine that all we can know is mere relations. He here recognises that a philosophy of nature must take into account the ‘matter of knowledge’ as such, since it too must have some element of necessity. Thus it is not only the categories and the pure forms of Intuition which have an a priori certainty, but the sense-manifold must also be of such a nature that the categories and forms of Intuition will apply to it. It must be of such a nature as to supply terms to these relations. And, though the amount of the nature of objects of experience, which is thus determined a priori, is far from giving them a claim to be considered completely rational, it yet gives them a certain amount of inwardness and self-subsistence. Thus, in considering the course of Nature, it becomes obvious that, though we try to explain what happens, by referring it in each case to something prior in time, and so on ad infinitum, there is also presented another element, left out of account by this method (the only one allowed a
e.g. in the ‘Amphibole’ R.V. p. 235 [B334, GW 375].
224
The 1898 dissertation
by Kant in the Pure Reason), which also helps to explain what happens. This element is the actual qualitative nature of the events we are trying to explain. So far as merely causal connection is concerned, there is no reason why there should be any change in the world whatever, except that which is involved in the lapse of time. Each moment of time is different from the one before, and, if the world were quite without other differences, there would yet be a necessary connection between its state at one moment and its state at the next, exactly fulfilling the type which Kant sought to prove against Hume. For the state of the world at one moment would be a different thing from its state at the next, in the sense in which Hume denied that you were really entitled to infer from the existence of one thing the subsequent existence of another. But, even if this were so, causation would obviously not afford a complete explanation of the course of nature. The world which did thus persist unchanged through time, would still itself be part of the reason of the course of Nature. We could not exhaust our knowledge of each successive state by saying it was such as to have been the effect of the one before and the cause of the one which followed it. It would still remain true that each state was what it was, besides being related to those before and after it; each would have a content – the content in virtue of which each was identical with that of every other; and the nature of this content would require to be taken into account in explaining each state. We can assert a priori not only that each state of the world must be necessarily connected with those that precede and follow it, but also that it must have some definite qualitative nature. It is not only what it is because the previous state was what it is, but because it is what it is. This consideration seems obvious enough, but yet it is one which is very apt to be neglected. It was recognized in the Aristotelian doctrine of ‘formal’ causes; but has been put out of sight by the procedure of modern science, which seeks always for efficient causes, without sufficiently considering that there could be no efficiency unless there were also ‘form’. It is no doubt of more practical importance to establish the relations between things than just to recognise those things; and Kant in his desire to justify the methods of natural science, seems to have been misled by the prominence given in it to the discovery of relations, into an unjustifiable neglect of the qualitative aspect of things. There was also, as has been pointed out above, another reason for the emphasis which he lays on relation – namely his desire to protest against the assumption that the objects of experience were real, or absolutely self-subsistent. And finally, quality, so far as it is necessary, is only
Freedom
225
one: there are not, as in the case of relation, a number of different forms to justify. But still, from a philosophical point of view, it seems to be of equal importance, and is always presupposed by science in discriminating the things between which relations are to be discovered. Things, then, in so far as they must be terms of relations, may be said to have a self. But the degree of selfhood would not suffice to define the notion of freedom. For we are as yet not entitled a priori to infer in the world any differences of quality. And if there were none, as in the case above supposed, there would be no reason to suppose that the causal connection between the successive world-states was in any way dependent on their qualities. The quality would necessarily be taken into account in explaining the series as a whole; but the causal connection might be considered to hold between them purely as existing, in so far as they had matter in the Aristotelian sense, not in virtue of their form. And this, it is to be noted, is all that Kant proves for causality in the Critique of Pure Reason; the necessary connection is between the existence (Dasein) of things. But, as a matter of fact, there are differences of quality in the world of experience, and whatever be the justification for it, there is connected with this difference of quality a most important addition to the notion of causality. Causality in Kant’s sense would not justify any Law of Nature, and yet without these science would be impossible. There is implied in any law, that ‘Like cause has like effect’ and vice-versˆa; and in this conception we have at once the causal relation between things conceived as depending on their qualitative nature (under ‘quality’ is included for this purpose, position in space). It is no longer the thing, considered as individuated merely in time, which is necessarily connected with those preceding and following, but the thing, as distinguished by a particular quality, is considered to have a necessary connection with other things so distinguished. It is not assumed that all the qualities in the world might not be different from what they are; but it is assumed that, given any one quality, it has a unique causal relation with some other one, in the sense that only the thing of which it is a quality can be cause of the thing of which that other is a quality, and only that other thing can be the effect of the first thing. With this we seem to have arrived at the notion of a thing with a distinguishable self, having a distinct efficiency in virtue of that self. And in this conception of the course of nature there is contained the union of Determination with Freedom, in its simplest form. Each thing, marked by a simple qualitative nature, is no doubt determined in that it is the effect of some other thing, and given that other thing, it was forced to appear.
226
The 1898 dissertation
But also it is itself similarly the cause of something else, and free so far as its effect depends upon its own nature. It is nothing against this that its own nature depends in its turn upon something else; for that something else could not by itself have produced the effect which it produces. It is an essential link in the chain, and though the effect is not solely due to it, some part of the effect is due to it and to it only. In this notion, then, we seem to have a quite precise distinction between free and necessary action. Every action is both free and necessary; freedom and necessity are an aspect, under which every event must be necessarily regarded. And thus we have an ultimate and valid meaning for freedom, as appearing in the world of experience; a meaning, which in consequence of the Kantian criticism, substitutes within that world the notion of the selfcaused for that of the uncaused or original, which had been maintained by the pre-Kantian metaphysics. But neither common sense nor Kant himself are satisfied with this. Both wish to drag into the world of experience a notion of freedom which would be inconsistent with complete mechanical determination, as applying only to some of the objects of experience, and not merely as an aspect from which every mechanically determined event alike may be regarded. This is attempted no longer by maintaining the freedom of conscious will, in the sense of ‘Liberty of Indifference’, – the old notion, discarded by Kant, against which I have hitherto directed my attack – but by treating certain groups of qualities, discoverable in the world of experience, as bound together in a so-called ‘organic’ unity, which confers on them a ‘self’ in a special sense, and makes not their parts, but them, ultimate factors in natural processes. The human mind may then be regarded as a special instance of this sort of unity, and so capable of free action as a whole, differing in degree perhaps, but of the same kind as that exercised by any organism. Kant, it is true, in treating this notion of ‘organism’ in the Critique of Judgment, still asserts that it is ‘regulative’ or ‘reflective’ only, not ‘constitutive’ or ‘determinant’. By this he means that our experience would be possible without it. We should still be able to have knowledge of an objective world, if all things in it were only mechanically determined. But nevertheless he does now maintain that the explanation of certain things presented in experience, namely organisms, is absolutely impossible according to mechanical laws. By this he appears to mean that their ‘form’ is wholly ‘contingent’ from a mechanical point of view (p. 372)a ; and this a
[Kritik der Urteilskraft, ed. Hartenstein, vol. 5] p. 372 [AK 5: 360; Critique of Judgment, trans. Meredith, ed. Walker, Oxford University Press, 2007, 188].
Freedom
227
contingent form is such that all the parts of the thing may be regarded as acting for the sake of each other’s ‘form’, i.e. the form of each is end as well as means to all the rest (p. 385)a . Thus, in order to define an organism, Kant is obliged to use, in describing the relation of the parts to one another, the notion of ‘end’, which he admits to be only given by the Practical Reason. He says that the form of the whole organism can only be regarded as the ‘end’ of the parts (and that only ‘regulatively’) on condition that the parts are ends to one another. But apparently his only reason for thinking that they are so, is that their ‘form’ is inexplicable by mere mechanical causation; and this ‘form’ is no longer the systematic unity which it was in the case of the whole, but the presupposition of that. So that it would seem to be undistinguished from ‘form’ in the sense above explained, in which every natural object equally may be said to have a ‘form’. There is therefore no reason primˆa facie, why an organism should be distinguished as having a systematic unity, which would not apply to any group of interacting parts whatever; since in every case, as was shewn above, the action must be conceived as affecting the ‘form’, and likewise in every case the mere principles of the understanding will give no reason why the form should be such as it is. Apart, therefore from the notion of end, which must be separately discussed, Kant’s only reason for saying that in organisms we have presented a whole or systematic unity, such as cannot be explained by natural laws, is the assumption that its parts can be seen to be related in a non-mechanical way. And his ground for assuming this, again, is one that applies equally to all causally related elements. So that he fails, on the whole, to justify even as a regulative principle the special position which he attributes to organic things.124 But the doctrine that organisms may be regarded as non-mechanical unities be125 based on a supposed direct insight into the relation of the parts as mutually ends and means. If this view be taken, it must first of all be explained what ‘end’ means. In the first place it may be taken to mean ‘effect’; but this meaning will obviously not help us in the present instance. The second and commonest meaning has reference to volition: it is something which we desire or the thought of which moves us to action. The third meaning is one commonly confused with this, as, for instance, it was shewn in the last chapter that Kant confused them: and a
124 125
[Kritik der Urteilskraft, ed. Hartenstein, vol. 5] p. 385 [AK 5: 373; Critique of Judgment trans. Meredith, ed. Walker, Oxford University Press, 2007, 201]. The next paragraph is new for the 1898 dissertation. Sic – the sentence originally started ‘If . . . ’, which has subsequently been deleted.
228
The 1898 dissertation
here means what is good in itself. Kant’s view of the sense in which parts of organisms could be called ends seems to have been this confusion of our second and third senses. They must be regarded ‘as if they had been willed’ and were therefore good. But we dismissed this connection of good with will in our last chapter; and there is certainly no ground for saying that the parts of an organism do consciously will one another or that their relation is willed by some external power. It remains, therefore, that if they are ends at all, it must be because they are good in themselves. While at the same time they are certainly means to one another – i.e. causes or necessary conditions of one another’s existence. Moreover, in respect of their relations to the whole, it may be admitted that, as parts of it, they are better than each would be taken in isolation: that two things, which would separately be comparatively worthless, may by a special conjunction become highly valuable as parts of that whole which they combine to form. Organisms do not appear to me to be at all good examples of such wholes; but such wholes there certainly are. Though, however, these have a unity which is not mechanical, they are also always mechanical unities. Their parts are always interrelated as cause and effect; and unfortunately nature does not seem so to respect their combination, as to prevent their dissolution by natural causes, or to allow the whole, as such, to produce an effect, which cannot be accounted for by the individual working of the parts. The reference to organic wholes, as ends-in-themselves is, therefore, wholly useless for our present purpose. On the other hand, if we confine our attention to the relation of the parts, without employing the notion of end, there is something in this relation which cannot be explained by the mere ‘principles of the understanding’. Our knowledge of nature would be impossible, unless things did act according to rules, i.e. unless the form of a thing could be taken to indicate what its effect would be. This is a principle which Kant seems merely to assume (R.V. p. 380: jede Ursache eine Regel voraussetzt . . . und jede Regel eine Gleichf¨ormgkeit der Wirkungen erfordert)126 but it would seem to be a properly constitutive one. In the Critique of Judgment, it is, however, confused with another principle which Kant does declare to be ‘regulative’. It is here asserted that, if we are to have knowledge of nature, the variety of qualities presented in the sense-manifold must not wholly exclude recurrence; and this is taken to be equivalent to the principles of ‘homogeneity, specification and continuity’. These are principles according 126
B577, GW541: ‘every cause presupposes a rule . . . and every rule requires a uniformity in its effects’.
Freedom
229
to which we should always aim at a systematic classification of the Laws of Nature themselves, hoping to find a continuous chain from one, general enough to embrace them all, down through regulative subdivisions, to infinite particularity. Kant’s view is that we cannot know a priori that Nature is capable of such continuous classification, and there do actually appear to be gaps which spoil its symmetry; but that the ideal of science is to reach it, and that scientific progress consists in continual approximation to it. It is therefore a ‘regulative’ but not a ‘constitutive’ Idea. Now, whether this be so or not, it seems extremely important to distinguish these principles, from the notion with which Kant connects them. It is true that there must be some recurrence of identical elements in experience, in order that we may find any Laws of Nature at all; but it does not follow from this that these Laws will themselves be capable of a completely systematic classification. The two principles seem to be on an entirely different level, whereas Kant takes them both indifferently as evidence that Nature as a whole must be regarded as if it had been designedly adapted to our intelligence. The truth is that the one is presupposed by the principle ‘Like cause, like effect’, and so is necessary to a knowledge of nature as mechanically determined; whereas the other may indeed be regarded as a mere ideal. This ideal of perfect classification would indeed presuppose the other; but the inverse proposition does not hold. Kant’s reasons, therefore, for holding the Idea of complete system to be purely regulative, do not invalidate the objectivity of the principle that all the elements in nature reciprocally determine one another’s form, and that, therefore, though the form of Nature as a whole must be regarded as contingent, the form of the parts in relation to one another is necessary. On the other hand, the notion of a systematic whole does not apply objectively either to Nature as a whole, or to any group of parts in nature. We must therefore be content, so far as the doctrine of experience is concerned, with regarding every element in Nature as equally free and determined; but this much we may regard as fact. It is, no doubt, convenient to treat certain groups as wholes, and to ascribe freedom to them in the sense that their parts and not anything external to them are immediate cause of their changes; but this grouping must be admitted to be wholly arbitrary. It is not merely, as Kant thinks, that a mechanical determination of the parts is the only one that can be presupposed as actual; but also that there is nothing to explain in the whole, which is not also in each of the parts, taken by themselves, and which is not presupposed by their mechanical connection.
230
The 1898 dissertation
Now the case of human volition would seem to be precisely the same. This, too, is to be explained mechanically; but is free, in virtue of that very mechanical explanation, in so far as the form of each element in the series may be regarded as determining the form of the next. And there is no more ground for treating the mind as a whole as free, than there is for treating an organism as free in nature. If a precise and ultimate meaning is to be given to such freedom, it must be explained as that which properly belongs to the ultimate mental elements, not as anything which resides in the whole, as such. For there is here again no special form to fix the limits of a unity anywhere between the smallest distinct element and the whole mental world. But, on the other hand, it may be admitted that the activity of mental elements is sometimes directly perceived, in a sense in which that of physical elements never can be. For consciousness itself is an element in mental processes, so that here the form which determines the change knows itself. However, the important point to emphasize is that, from the point of view of explanation of experience, this makes no difference. In the context of inner experience this form plays just the same part as any other form in outer experience. The difference is only to itself; it is not of an objective significance. From the common point of view, then, which takes the world of experience as ultimately real, this, in which every part of that world is alike free and alike determined, is the only sense of freedom, which can withstand criticism, as in no way based on arbitrary distinctions. It is a sense, which would, to most, seem to be the same as that of determination. But it can, I think, be seen to underlie all common uses of freedom: and it is largely to the difficulty of distinguishing it as an irreducible aspect in mechanical causation, that there is to be attributed the mistaken attempt to shew that the notion of freedom is irreducible, by maintaining the existence of uncaused choice.127 It was admitted that there may be a timeless existent, such as [Kant]128 imagines his free no¨umenon to be. If there is such an existent, it must stand in some reasonable connection with the existents of which we know. But what that connection is, no way was found of conceiving it, it cannot be that of cause to effect in the ordinary sense, and any attempt to combine this notion with that of reason and consequent seemed to fail. It cannot therefore be called free in any positive sense, but it may be wholly good. And 127 128
The next, final, paragraph is new for the 1898 dissertation. The need for this addition is obvious.
Freedom
231
if it were so, this might serve to explain in part the common connection of freedom with goodness. For it would be free, in the negative sense that it was not conditioned by causal determination and the necessary connection of anything thus undetermined with goodness would tend to associate the two ideas in a way that is nevertheless not justified by anything that we know as good in the world of space and time.
chapter v
Ethical Conclusions
Kant’s ethical theory as a whole may be best characterised as a fusion of two wholly distinct doctrines.129 On the one hand his work contains an elaborate investigation into the nature and relations of the concept ‘ought’ – an investigation covering the ground of what may be called ‘Metaphysics of Ethics’. And on the other hand, it contains a peculiar and exclusive doctrine of Practical Ethics. Kant himself seems never to have perceived how distinct these two parts of his work were, and how completely the former is independent of the latter. To the metaphysical department belongs Kant’s recognition of the independence of moral law from natural law. What ought to be is something which perhaps never has been and never will be; and never, by considering what has been or what is, will you discover it. In his own language: the moral law is something a priori and no induction from experience. This doctrine at once distinguishes his system from any so-called naturalistic ethics; from the ethics based on psychological hedonism, or from modern evolutionistic systems. And to the same department belongs his recognition that ethics must be based on law. Any ethical principle must be universal in the sense that if it is valid for you now, it must be valid for [everyone] else under the same circumstances. His ‘Fundamental Law of pure practical Reason,’ ‘Act so that the maxim of your will can always at the same time hold as principle of a universal legislation,’ (P.V. 32)130 recognises the proposition that if anything whatever, whether simple or complex, can be truly said to be ‘good’ or ‘what ought to be’, that same something, whenever it recurs, must likewise be held to be good. But Kant also has a practical doctrine, which is completely distinct from this ultimate metaphysical one. He holds that ‘there is nothing conceivable in the world, nay even outside of it, which can be held as good without 129
This chapter is new for the 1898 dissertation.
232
130
AK 5: 31, MG 164.
Ethical Conclusions
233
limitation, except a good will’. This statement might be taken as a mere definition of what is commonly called ‘moral good’, as opposed to other kinds of good; but Kant does not intend it as such. A good will is in his opinion the only thing which can truly be regarded as an end-in-itself; even happiness is only good in conjunction with this, and not to be aimed at for its own sake. This second doctrine of Kant’s involves the proposition that no action is good unless the motive to it is the agent’s belief that it is good. The ‘good will’ is the will, which not only acts in conformity to duty, but also ‘from duty’. And it may be admitted that this disposition to be moved to action by one’s belief that an action is good, is a most important good thing; being, in fact, the mark of what we should generally call a ‘good man’, and the sign of a good character. But Kant tries to elevate it, as has been done by others, into the sole principle of practical ethics. This is the claim which we have now to consider. Kant seems to think that this practical doctrine of his may be deduced from his metaphysical doctrine. On the contrary I shall try to shew that it is inconsistent with it. Two main steps are involved in the deduction. In the first place there is the famous attempt to deduce from the categorical imperative, by the principle of contradiction, what actions are in conformity with duty. A second step is required to prove that only actions, of which the motive is the desire to act thus in conformity with duty, are really good. (1) The fallacy of Kant’s attempt to deduce such a principle as ‘Thou shalt not lie’ from the Categorical Imperative has been frequently exposed. It has been pointed out that the mere principle ‘What is good once is good always’ is not sufficient to shew what is good once. Kant does not indeed attempt a direct deduction of particular duties from the moral law. He uses it rather as a criterion whether a particular maxim suggested is dutiful or not. He does not even say that supposing the maxim suggested is dutiful, its conformity with the moral law will prove it to be so. But he does maintain that, if it be undutiful, the moral law by itself will suffice to show that it is so. This seems to be his meaning in the ‘Metaphysik der Sitten’ (p. 192),131 where he says that the fact that the maxim is not contradictory to the universal law, is merely a ‘negative principle’. This description is, therefore, fully in harmony with that of the ‘Grundlegung’ (pp. 269–272)132 and of the ‘Practical Reason’ (pp. 28–29; 73–4).133 Kant’s 131 133
132 AK 4: 421–4, MG 73–5. AK 6: 221, MG 376. AK 5: 28, MG 161; AK 5: 69–70, MG 196.
234
The 1898 dissertation
contention in all these cases, [is] that the moral law will suffice to detect an undutiful maxim, although maxims in conformity with the moral law are not thereby proved to be dutiful. He maintains that dutiful maxims must be in conformity with the moral law, but he does not exclude the possibility of maxims which are neither dutiful nor undutiful. To refute Kant’s whole position, then, we have only to shew that want of conformity to the moral law will not suffice to detect all undutiful maxims. It cannot, in fact, suffice to detect any by itself alone. For it is only if we already know a certain act to be undutiful, that it will help us to decide that an act of precisely the same kind is also undutiful. This is an important principle for the regulation of conduct and applies as well to the dutiful as to the undutiful; but it always presupposes a second principle – that the act with which we compare the one to be considered is, as the case may be, dutiful or undutiful. Kant, however, thinks that it is sufficient by itself alone to detect the undutiful. His argument presupposes that ‘This is always good’ may involve a contradiction which is not involved in ‘This is good now’; and that, therefore, since the former, as contradictory, cannot be true, the latter must also be untrue. But either ‘This is always good’ presupposes ‘This is good now’, in which case the untruth of the latter does not follow from that of the former; or ‘This is good now’ presupposes ‘This is good always’. But this it obviously cannot directly do. The true relation of them is that they both alike presuppose ‘This is good’; and neither can be true unless this is so. Unless, therefore, there is a contradiction in ‘This is good,’ no contradiction in ‘This is good always’ can prove the untruth of ‘This is good now’. But ‘This is good’ cannot be contradictory, unless either the ‘This’ or ‘good’ are contradictory conceptions like that of a square circle. But the actions Kant is discussing are obviously not of this nature; nor again can he maintain that ‘good’ is such. Accordingly the proposition ‘This is good’ can only be denied as the basis of the proposition ‘This is bad’. And consequently a premiss of this form, as well as the moral law, is necessary to Kant’s argument. (2) There is even less to be said for Kant’s second contention that only action, of which conformity to law is the motive, can be called really good. It is obvious that the actions in conformity to the law, which Kant’s former argument was intended to deduce, have some special quality which it is difficult to call by any other name than ‘good’. Kant’s point seems to be that it is only the goodness of these which can form a moral motive to the doing of them. But in that case we get the curious result that that which is good is not good unless it is done because it is good. Kant only conceals this result from himself by his assumption that the quality in what is willed which makes it a proper motive, is in the end the same as that
Ethical Conclusions
235
of the will which is moved by it. He holds that the moral law is the law of the will, that is moved by proper motives, and hence that in willing it we are willing to be moved by proper motives. He does not see that there is nothing to determine what is a proper motive, unless we allow the same standard to be applied to the objects willed as is applied to the will itself. But if Kant’s doctrine be not understood to involve this vicious circle, it is liable to still graver objections. For, if the value of the willing does not at all depend upon the value of the object willed, but solely on the fact that conformity to the moral law is the motive; then the object willed need not necessarily be in conformity to the moral law, and we shall be acting perfectly well, when we intend to lie and murder, provided only we are mistaken in thinking those actions to be in conformity with the law, and are moved by that goodness which we attribute to them. The root of this vicious consequence, which can certainly be drawn from some of Kant’s expressions, again lies in his doctrine of will. He assumes that we cannot act for duty’s sake, without also acting dutifully, although he recognizes that we can act dutifully, without acting for duty’s sake; and the falsity of this assumption he could hardly have failed to see, but for his representation of the moral law as nothing but the law of will, and consequent conclusion that the will which he thought good could will nothing but what was good. While therefore we accept Kant’s metaphysics of ethics as pointing out the fundamental principle that ‘Good is a concept’, with its consequence that ‘What is good once, is good always’, it seems impossible to accept his theory that any practical principle can be obtained, without a second premiss, not to be deduced from the first, of the form ‘This is good’. His practical doctrine it seems impossible to accept at all. He actually assumes as his second premiss (without admitting it to be such) that we ought to aim solely at acting with a good motive. To this we object that a motive is good (1) if what moves us is the belief that our proposed action is good (2) if the proposed action is also really good. (1) It may be admitted that to act under the belief that our proposed action is good, is itself good; but it cannot be the sole thing which we ought to aim at. (2) If the proposed action is also really good, then it must be good to aim at doing it, whether we believe it to be so or not. But an action itself may be good either as means or end. And, if it is good, as a means, then something else must also be good; and it follows that it will be good to aim at that, and not only at the action which is a means to it. To act with a good motive can therefore be only one among the objects at which we ought to aim: and it is itself better, if the belief, which moves us, is a true one.
236
The 1898 dissertation
It appears, then, that before any practical ethical precepts can be laid down, the principle involved in Kant’s Moral Law, must be supplemented by some other principle or principles declaring that ‘This or that is good’. Nor does there seem any way of proving that any one of these is true or false. It is only possible to offer one in the hope that it may appear to be true, and further to try to produce conviction by shewing it to be more in accordance with complex propositions accepted by your opponent than other possible ones are. I have only to suggest that if any one principle can be found, it seems to me that it must be one connecting good not only with any simple concept but with one that is complex. This seems to me the chief objection that can be waged against the ‘greatest happiness’ principle. I cannot accept the theory that pleasure only is good in itself, for a reason urged by Plato in the Philebus (21).134 Philebus is there at once convinced by Socrates that pleasure is not the sole good, by the consideration that, if it were, the best possible life could be lived, without our either knowing that we were pleased, remembering that we had been, or reckoning that we should be. This argument seems to me as immediately convincing as it did to him. The strength of it is obscured by the fact that in thinking of pleasure we always do think of ourselves as knowing that we have it, and of it as accompanied by other mental states. But that it is difficult to imagine pleasure as existing apart from what are, in the order of nature, its necessary conditions, is no valid objection to the argument. For in considering our ethical principle, as has been insisted, in another connection, we are not entitled to take into account what is physically possible. Pleasure must, if our proposition with regard to it is to have any meaning at all, be conceivable as something separable even from existence; and therefore, for the same reasons for which Prof. Sidgwick insists that we must judge by quantity and not by quality of pleasure, we must also judge its value apart from what goes with it or causes it, but is not it. There seems, indeed, to be obviously included in the Utilitarian principle, in any form in which it is advocated, what is, in truth, a separate principle – that existence must be included in the ultimate end. It is scarcely likely that any one will maintain that non-existent pleasure is what we ought to aim at. Yet pleasure is certainly something distinct from existence. I should therefore certainly include in my conception of an ultimate end existence; and along with that, if pleasure also (but of its necessity I am not convinced), knowledge of the truth, and certain forms of emotion. 134
See Philebus 21a–d; Moore uses this argument from Plato’s Philebus in Principia Ethica (§52: pp. 139–40), where he quotes extensively from Plato’s dialogue.
Ethical Conclusions
237
But, even when we are provided with an ethical principle determining our ideal (and no definite one is here offered), the problem of practical ethics is by no means yet solved. It was pointed out above (p. 179) that in order to get a practical precept, we must, unless, as seems to me impossible, we decide that our ideal consists in our actions themselves, accept the supplementary principle that to do what will produce our ideal is good. But by this principle we are at once involved in a consideration of the physically possible. We have to decide what will produce our ideal, before we can say ‘It is good to do that’ or ‘You ought to do that’. There would, therefore, be necessary to a complete system of practical ethics, the completion of psychology and all the other sciences necessary to prediction. But, granted that we do not know, and seem unlikely ever to know, completely what will happen, it will obviously be said that we must try to obtain our practical precepts by considering what is probable. A rational precept will be one which bids you do what will probably have the best results. In attempting to make a rational choice you will, therefore, have to consider (1) What is good (2) What will most probably bring it about. It may certainly happen that what you think good is something which it seems impossible to effect by anything which you can do. In this case rational choice is obviously impossible. But the same would seem to hold even if this is not the case. For you cannot determine with certainty what the result of any choice will be. By your moral principle you are bound, if you would be rational, to choose the best you know of. But, on the other hand, if your choice is to be practical, you must choose something which will possibly happen. But this possibility ranges through all degrees from the lowest up to the highest probability. The question, therefore, arises whether a greater good that is less probable is to be preferred to a lesser good that is more probable. And this seems to be a question susceptible of no certain solution. Thus, even supposing the degrees of good accurately fixed (a problem which does not seem insoluble), no rational practical conclusion could still be drawn from them, since the degree of probability would have to be weighed against the degree of goodness. Practical Ethics has, therefore, essentially the nature of a compromise. Its principle that you should choose the best possible attempts unsuccessfully to combine the two irreconcilable principles that you should choose what is best, and that you should choose what will happen. Either of these by itself is insufficient as a rational guide to practice; and yet each claims to be absolute.
Appendix on the chronology of Kant’s ethical writings
The writings on which the above account of Kant’s ethical views is based were published in the following order.135 Kritik der reinen Vernuft (first edition) Grundlegung zur Metaphysik der Sitten Kritik der reinen Vernuft (second edition) Kritik der praktischen Vernuft Religion innerhalt des Grenzen der blosen Vernuft Metaphysik der Sitten
1781 1785 1787 1788 1793 1797
I have treated all these works, without regard to their differences of date, as if Kant’s view on the fundamental points of ethical philosophy had not changed during the period they cover. But some writers have maintained that a change of view is exhibited in the later writings as contrasted with the earlier ones, and hence that they are not all to be taken as equally representative of Kant’s views. It seems, therefore, well to give some account of the reasons which lead me to dissent from this contention. In the first place, no ground can be given for thinking that Kant had consciously changed his view from one work to another. He nowhere gives any hint that he has; and, on the other hand, it is notorious how vehemently he insists to the last that his Critical system is completely self-consistent and needs no alteration. Moreover he refers in his later ethical to his earlier ones, as if he were completely satisfied with the results there established; and to his pre-critical writings, which really expressed different views, there is nowhere any such reference. Thus, in the Preface to the Practical Reason (p. 8)136 he refers to the Grundlegung, explaining how the two works are related to one another, and is even concerned to defend the position taken up in the earlier work against criticisms which had been passed on it. Had 135
This appendix is new for the 1898 dissertation.
238
136
AK 5: 8, MG 143.
Chronology of Kant’s Ethical Writings
239
he been conscious that he had modified his views in any important point, he would surely have mentioned it here. Similarly he begins his Preface to the ‘Metaphysik der Sitten’ (the last work of all), by a reference to the ‘Critique of Practical Reason’, explaining that his present work merely contains the System based on the principles there established.137 Finally, apart from the frequent references in all the ethical works, to the results of his ‘Critique of Pure Reason’, it is incredible that if the ‘Grundlegung’ had caused him to modify the views on Freedom explained in the ‘Pure Reason’, there should have been no reference to this in the second edition, which was published two years later than the Grundlegung. As it is, the passages on Freedom and Morality in the ‘Aufl¨osung der kosmologischen Idee von der Totalit¨at der Ableitung der Weltbegebenheiten aus ihren Ursachen’138 (pp. 370–385),139 and in the ‘Kanon der reinen Vernuft’140 (pp. 526–540)141 remain unaltered in the second edition. We may, then, I think, fairly assume that Kant himself was not conscious of any change of view on Ethics, during his ‘Critical’ period, and that he would have been quite content that his ethical views should be tested by a reference to any one of the above-mentioned writings. But it may still be maintained that his view changed unconsciously; and that therefore some one of his works is much superior to the others. It must, however, be insisted that, even if this were so, it could not properly be said that the chosen work expressed Kant’s ethical views any more than the rest. It would, at most, be true that Kant thought this at one time, and something else at another. Only if it could be shewn that certain self-consistent principles were common to all the works, whereas some only were vitiated by inconsistencies, could it be plausibly maintained, that the works which were free from such inconsistencies, expressed the essence of Kant’s doctrine. Even in this case it cannot be proved that Kant would have accepted the corrections of his would-be supporters; but it may be admitted that there is some probability that he would. But any such view will be finally disposed of, if the inconsistencies, which are supposed to be peculiar to one work, can be shewn to be inherent also in those, for which a freedom from them is claimed. The ‘Grundlegung’ is the work, which has been most canvassed in this respect. It has not, so far as I know, been maintained that the ‘Religion’ or the ‘Metaphysik’ show any marked difference from the ‘Practical Reason’. 137 138 139
AK 6: 205, MG 365. ‘Resolution of the cosmological idea of the totality of the derivation of occurrences in the world from their causes’. 140 ‘Canon of Pure Reason’. 141 B823–47, GW 672–84. B560–86, GW 532–46.
240
The 1898 dissertation
But it has been maintained that there are striking differences between the ‘Grundlegung’ and the ‘Practical Reason’; and the ‘Pure Reason’ has been supposed, on the one hand, to agree with the ‘Practical’, where it differs from the ‘Grundlegung’, and on the other hand to differ from both, but to be still further from the ‘Practical Reason’ than from the ‘Grundlegung’. (1) It is pointed out that in the ‘Grundlegung’ the categorical imperative is expressed in the form ‘Act only according to that maxim, by means of which you can at the same time will, that it be made a universal law’ (p. 269),142 whereas in the ‘Practical Reason’ it runs ‘Act so that the maxim of your will can at the same time always hold (gelten) as principle of a universal legislation’.143 It might, certainly at first sight, be thought, that the formula of the ‘Grundlegung’ would cover cases, which that of the ‘Practical Reason’ would not. Indeed the ‘Grundlegung’ immediately goes on to consider two classes of cases, in one of which the maxim could not serve as a universal law because it could not without a contradiction be even thought as a universal law of nature (p. 272),144 whereas in the other class it would only be contradictory to will it. And the first of these classes only is used in the first illustration in the ‘Practical Reason’ (p. 28).145 But in the ‘Typik’146 (p. 73)147 an instance of the second class reappears, as if it were precisely on a level with the first: and all alike are to be tested by the question ‘Could you regard this maxim as possible through your will?’ and this question again is represented as equivalent to ‘Would your will agree to be in such an order of things?’ So that, if the ‘Grundlegung’ is inconsistent with the ‘Practical Reason’ in this respect, the latter would appear to be equally inconsistent with itself. In fact, however, with Kant’s derivation of the moral law from will, the two formulae are identical. (2) It is maintained that in the ‘Grundlegung’ we are free only to will the good, whereas in the ‘Pure Reason’ we are free to will the bad as well. This point was touched above in Chap. iii. It was there said that Kant hesitates to ascribe bad actions directly to Freedom, although his general position would allow him to do so; and that, on the other hand, though he shews a corresponding tendency to say that Freedom can give rise to good alone, on this point he is not quite express. It must be admitted that the ‘Grundlegung’ gives more colour than the other works to the contention that this latter was his view. Kant does here treat Freedom solely as source of the moral law; and this is explicable from the object of the work, which starting from the ordinary view of moral goodness, tries 142 144 146
143 AK 5: 30, MG 164. AK 4: 421, MG 73. 145 AK 5: 27, MG 161. AK 4: 424, MG 75. I.e. ‘Of the Typic of Pure Practical Judgment’.
147
AK 5: 69, MG 196.
Chronology of Kant’s Ethical Writings
241
to shew how first the moral law and then freedom are involved in it, and is therefore not concerned to state the relation of Freedom to wickedness. Yet even in this work he states that ‘man is not responsible for his inclinations and impulses and does not ascribe them to his proper self, i.e. his will, but he is responsible for the indulgence a which he might accord to them, if he allowed them influence upon his maxims to the prejudice of the rational laws of his will’ (pp. 305–6);148 a statement which comes as near as any to the view that some sort of badness, i.e. moral wickedness, is to be ascribed to Freedom. And, conversely, while in the Pure Reason, since he is there dealing directly with Freedom, not with the moral law, he lays chief weight on the evidence furnished by our bad actions, saying that one may ‘look upon a wicked action as altogether unconditioned with regard to the agent’s previous state, as if the agent began therewith, entirely of himself, a series of consequences’ (pp. 383–4);149 yet even here he goes on to explain that ‘This reproach is grounded on a law of reason, whereby reason is regarded as a cause which could and should have determined man’s conduct otherwise’,150 thus avoiding the downright statement that the wickedness is due to Freedom, and inferring Freedom rather from the fact that he might have acted well; and in the Canon (pp. 533–4)151 use is made of the idea of a moral world, which presupposes that a free will, when freed from all hindrances of nature, would do the good only. (3) On the above point the ‘Practical Reason’ seems to hold an intermediate position between the ‘Pure Reason’ and the ‘Grundlegung’; a fact which is again easily explicable from its object, which necessitates a fuller treatment of Freedom than was suitable to the latter, and a greater insistence on the moral law than was in place in the former. But it has been said that it and the ‘Grundlegung’ both shew an advance on the ‘Pure Reason’, in that they give up the expression of the moral law as ‘Do that, whereby you become worthy to be happy’ (R.V. p. 534).152 This formulation is said to shew an Eudaimonism; a contention which is too obviously absurd to need refutation. And the supposed advance in the Grundlegung is almost as easily refuted: for Kant there explains that the ‘mere deserving to be happy’ interests us, apart from any empirical interest, because it is a simple deduction from the importance of the moral law; that, though, therefore, it is not a primary form of the moral law, yet the principle ‘aim at being worthy of happiness’ involves this and nothing else, and is only explicable under a 148 151
My italics. AK 4: 458, MG 104. B836, GW 678.
152
149 B583, GW 544. B808, GW 679.
150
B583, GW 544.
242
The 1898 dissertation
the assumption of autonomy, not of heteronomy (pp. 297–8).153 Similarly in the ‘Practical Reason’ the principle that virtue does make you worthy of happiness, and that nothing else can, is vigorously maintained and used as a basis for the deduction of the celebrated postulates of ‘Immortality’ and ‘God’ (pp. 116–8 passim).154 On these three grounds, then, which seem to be the most important named, there seems no case for the contention of a change of view within Kant’s Critical Ethics. The most important inconsistency [in]155 his system seems rather to be that noticed above, as to the relation of practical to transcendental freedom, which appears [in] most acute form as between two parts of one work, the ‘P[ure Reason,] but can be shewn to exist in all the other works too. This is, in [fact], inseparable from his general treatment of ‘will’, as source of m[oral] principles and actions alike, a doctrine which is only verbally av[oided] by the strict distinction of ‘Wille’ and ‘Willk¨uhr’, first made in the ‘Metaphysik der Sitten’ (p. 23).156 153 155 156
154 AK 5: 110–13, MG 228–30. AK 4: 450, MG 97. The page is torn at this point, and the words in square brackets in the next lines are our best guesses of the missing text. AK 6: 226, MG 380.
Examiner’s report on the 1898 dissertation
Report on a Dissertation entitled ‘The Metaphysical Basis of Ethics’ by Mr. G. E. Moore Bernard Bosanquet
The Dissertation which has been submitted to me presents, in my opinion, an exceedingly difficult problem. I will state in general terms the question on which, if I am right, the electors must make up their minds, and will then give some reasons for my opinion, in order to expose it to criticism. In considering the claim of undergraduate candidates for a scholarship or for a place in the class list, it is enough I suppose to be satisfied of the ability and knowledge displayed in their work. It is not expected that they shall make a serious contribution to the science with which they are dealing. If they shew what is called ‘promise’ to a marked degree, their claim is taken to be made out. But I presume that a Dissertation offered to the electors for a Fellowship is judged by a somewhat higher standard. The electors, I imagine, would desire to be satisfied that the candidate has begun to turn his promise into performance, and has not merely shown even a brilliant capacity for a certain kind of study, but has taken up a line of work which has begun to be fruitful in his hands. Of course it must be borne in mind that in philosophy a critical or skeptical attitude is no bar to the achievement of highly valuable results. In the present case, to the best of my judgment, the above distinction applies in its fullest force. It would hardly be possible, within the limits of the subject chosen, to display more knowledge, ingenuity, and power of continuous persistence in a line of argument than the writer of the dissertation has displayed. As a piece of controversial pleading his work would do credit to any living author. And I do not mean to imply that he is deficient in strictly philosophical acumen. On particular points, especially in the discovery of difficulties and discrepancies, his insight is remarkably keen. And his conception – to make one observation in detail – of the As mentioned in the editors’ introduction’ (vii pp. lxvii) Ward was the other referee for Moore’s 1898 dissertation. But we have been unable to find his report (if he ever wrote one). Bosanquet’s report is in the Wren Library at Trinity College, catalogued as Add. MS A. 247 2(2). It comprises seven single small demy sheets, in typescript and corrected by hand by Bosanquet himself.
245
246
1898 examiner’s report
duties of a historian of philosophy seem to me to be thorough and just. I was therefore all the more disappointed when I found myself almost wholly unable to appreciate the theoretical point of view which the author has adopted. It appears to me to lie beyond the limits of paradox which is permissible in philosophy. It may be suggested that a disagreement with his positive views should be no bar to an independent judgment of the criticism of Kant, of which, as explained and supported by the author, the Dissertation consists. But here, as I am glad to point out, it is by his own merit that he suffers. His just conception of the duty of the historian of philosophy has the effect of marrying his positive views so closely to his critical standpoint, that it is wholly impossible to estimate them separately. His view of what Kant can have meant is controlled by his view of what there was for Kant to mean, and he analyses what he takes to be Kant’s errors and misapprehensions into what he takes to be the truths at which Kant is aiming. Thus the whole weight of the work rests upon the author’s theoretical position; and if this position seems altogether inadequate, the value of the exposition of Kant, which depends on it, is destroyed. This observation only applies to matters of principle in philosophy, and not to what may be called the philological aspect of the author’s criticism. I believe that so far as regards the citation and immediate interpretation of passages from Kant his comprehensiveness and candour are unimpeachable. This then is the general statement of the problem which in my view lies before the electors. If they are of the opinion that ability and intellectual distinction, with great knowledge, and, I may add, with great earnestness in the pursuit of truth, are sufficient qualifications, then I think they will be absolutely safe in electing the author of the dissertation; and it is quite possible that they may secure in him a considerable philosopher. But it is my duty to point out that I have never met with a stronger case of the paradox that a man may be exceedingly able and devoted, and yet, for some inexplicable reason, may fail to take up a sound position in his science. As to what future development may bring with it I can see no sure ground for predicting. Only it occurs to me that constant work under high pressure at a limited subject may have had to do with the one-sidedness which I complain of. If this is so, the author of the dissertation might be able in the future to justify himself in a way which does not now appear. It is obvious that my conclusion1 is based on a difference of philosophical opinion between the author and myself. When we get beyond the 1
The word ‘opinion’ has been crossed out here.
Bernard Bosanquet
247
mere estimate of ability and promise, and begin to consider the value of contributions to philosophy, I do not know how this is to be avoided. All I can do is to indicate some specific points to which my opinion refers, so that the electors may know how to discount my judgment. Of course I am only referring to points in the Dissertation, and am not giving an account of them which would be complete if taken by itself. In sum, the intellectual motive of the Dissertation, as I read it, is to dissociate Truth from the nature of Knowledge, and Good from the nature of Will, so as to free Metaphysic from all risk of confusion with Psychology. The theory of the proposition and the concept which harmonises with the dissociation of Truth from the nature of Knowledge is set out in chapter 2. I confess that I feel a difficulty in regarding it as serious. It is necessary no doubt to distinguish, in the processes and products of cognition, between their nature as knowledge and their psychological genesis. But the theory here propounded seems to reduce the world of truth to an immutable framework of hypostasised ‘propositions’ or ‘concepts’ in relations, which are indeed possible objects of thought, but are entities not dependent on thought, nor partaking of any character which distinctively belongs to thought. Truth and falsehood depend on the nature of the relation between the ‘concepts’ which constitute a ‘proposition’ (neither of these entities necessarily implying mental formulation); and their nature, (that of truth and falsehood), cannot be further defined, but must be immediately recognised. Here it seems to me clear that ‘the child has been thrown away in emptying the bath.’ To get rid of mere psychology, the essential idea of consciousness and cognition as an endeavour towards unity has been abandoned, and relational truth has been hypostasised as a self-subsistent form of Reality. The hostile criticism upon Kant’s Copernican attitude, and the inability to find an interpretation for his ‘personification’ of Reason as ‘a creature with claims’, or to treat the rationality of ethical action in any adequate way, are the natural consequences of this abandonment. The divorce between the ‘good’ and the nature of the will, as also between Reason with its law and this same nature, seems to me to indicate an equally hopeless surrender of the most important connections. It seems a verbal question whether the idea which acts as cause in volition is to be counted as a natural cause, and whether Reason can be Practical; but it would appear that to disconnect the nature of actual will from that of its object and its law must make any account of the ethical self or of ethical freedom impossible. The impossibility of finding a meaning for Kant’s Autonomy,
248
1898 examiner’s report
except in a sense in which it applies to all natural objects, results from this disconnection, which, as I understand it, is a mere case of the author’s general separation between the nature of consciousness on the one hand, and Truth and Goodness, as simple or unanalysable concepts, on the other. To make my meaning clear, I ought to add that I do not complain of the rejection of the Free Will of Indifference, either in itself, or as ascribed to Kant, and the writer’s defence against objections based on moral responsibility seems to me successful. But I do think it a serious matter that views should be adopted by an interpreter of Kant which wholly preclude him from giving a positive significance to the idea of Freedom which underlies Kant’s whole philosophy. I put my complaint in this form, because I agree with the writer that his views leave him no alternative so far as his representation of Kant is concerned. In making Kant consistent, I suppose, some of his statements must in any case be rejected or explained away, and the interpreter’s views of what it was possible for him to mean decide which are to be so treated. I ought in fairness to call attention to a reservation under which the author seems to propound his views throughout. He is speaking as ‘from the common point of view, which takes the world of experience as ultimately real’. And he maintains this ground, I gather, as an argument ad homines; as he rightly considers it to be the ground taken by both Determinists and Indeterminists, in their controversy. It may be that his views would be better able to justify themselves if they were liberated from this assumption, than they appear to be while it is accepted. I have not allowed this Report to expand into a complete philosophical discussion of the Dissertation, because such a discussion would involve the electors to act as judges in a philosophical controversy. I have stated the opinion which with great reluctance I have been obliged to form, and the question, which, if I am right, is presented to the electors for decision. I have also, I think, given sufficient indications of the grounds for my opinion to enable the electors to allow for any bias which may have affected it. The above are my principal objections. If I have exaggerated their importance, my estimate of this dissertation is not trustworthy. It may illustrate the situation if I state what my treatment of the work would have been, if it had been sent me for review by ‘Mind’, I should have treated it respectfully as a brilliant essay by a very able writer, but should have endeavoured to point out that its positive stand-point and consequently its treatment of the subject were hopelessly inadequate, that is to say that the writer was not successful, to any appreciable extent, in
Bernard Bosanquet
249
representing the real nature and interconnection of the factors involved in the problem with which he was concerned. I do not think that I can make my judgment any clearer by2 any further observations. 2
This final sentence has been added by hand (by Bosanquet) to the typescript, and at this point the writing is indistinct; but interpreting Bosanquet’s mark as ‘by’ is a reasonable hypothesis.
Index
analytic truth lviii, 139, 140, 141 Apostles Society xv Bosanquet, Bernard lxx on 1898 dissertation xviii, lxx–lxxi, 245–9 on Moore as Research Fellow xxii Bradley, Francis Herbert xiii, xxvi ideas xiii, xxviii, xxxiii, liv, 162, 163 Moore’s attitude to xiii, xxiv, xxvii, 4, 117 Brentano, Franz xxxi Caird, Edward xlvii, lxvii on 1897 dissertation lxvii–lxix, 99–114 Moore’s attitude to 3, 148 evil 81, 83 freedom connection with ethics 74, 77 human 37, 38, 57 libertarian notion of 21, 41, 44 practical liii, 59, 73 as self-caused action 54 transcendental lii, 35, 37, 58, 73, 78, 209 vulgar notion of 38, 39 Frege, Gottlob xxxii, xxxiii good connection with freedom 206 connection with reality 14, 78, 83, 127 definable or not xxxviii, 9, 10 fundamental to ethics xlix, 13, 178, 235 intuitive knowledge of lvi transcendental xlix, 12 unanalysable xxxviii Green, Thomas Hill xxiv on empiricism xxv as idealist xxv Moore’s attitude to xxxvii Prolegomena to Ethics xxiv, xxxvii, xlvii
Hume, David 23, 27, 48, 52, 146 intuitive knowledge a priori lvii, lxiii, 152, 153 ethics lvii judgment content vs. object xxxi Moore on 161 Kant, Immanuel antinomies 33, 151 appearances 27, 30, 32 on a priori lvii, 134 categorical imperative lxiv, 75, 76, 232, 233 Copernican Revolution xlviii, lviii, lxii, 142, 150 on freedom xii, 62, 73 intelligible character 21, 36, 37 intellectual intuition 187 practical reason liv, 59 on reason xiii, lxi, 133, 155 transcendental deduction 146, 171 transcendental idealism lviii, 30, 64, 143 works discussed by Moore x, xii, lxvi, 3, 238 Locke, John 48, 146 McTaggart, John xvii metaphysics of ethics 8, 12, 235 of science 8 Moore, George Edward life xiv–xxiii works ‘The Elements of Ethics’ xx ‘Freedom’ xii, lxxv, 118 ‘The Nature of Judgment’ xiii, xxxiv, lxxv, 134, 162 Principia Ethica xvi, xxi, xxii, xliv, xlv, lxv
250
Index ‘The Refutation of Idealism’ xxi, xxii, xxxiii on a priori lviii as Bradleian idealist xxvi and ‘the common point of view’ 57, 230 concepts xxix, xxxii, xxxiii, xxxv, 163, 165, 168, 173 on ethics xlix, 6 existence xxxii, xxxiii, xxxiv, 166, 168, 173 free will and prediction xl on Hegel 33, 78, 83, 85, 169 on history of philosophy li, 129–32 on Kant on a priori lvii, 137, 138, 170 on Copernican Revolution 143 on determinism xxxix, 20, 21, 41 on free causality 25, 34, 36, 37 on freedom xlvii, li, 58, 64, 186, 187, 189, 209 general attitude to xlvii, 32, 62, 150, 169, 192 on practical reason liv, 62, 63, 157, 158–60 on reason xlviii, 156, 179 practical ethics lxiv, 232, 236 on organic unities xlv on Sidgwick on free will xxxix, 50 on hedonism xli–xlv, 87–91 perception 168 pleasure 92, 236 propositions xxix, xxxii, xxxiii, xxxiv, 161, 165, 168, 173 reason 176 truth 166, 173 and belief 76 and consciousness 148 and knowledge 147, 152, 156 organic unity 54, 226, 228 Plato l, 14, 236
reality and appearance 35 and transcendental freedom 58 Russell, Bertrand early career xvi, xvii on Moore xvi, xxiii, xxxiv sensations lxii, 152, 154, 155 Sidgwick, Henry on 1897 dissertation xxxvii, 97–8 on free will xxxviii, 41, 47 on hedonism xli, 87 on Kant on Freedom 207 The Methods of Ethics x, xxxvi Moore’s attitude to xxxv, 4, 117 on pleasure 91 Stout, George Frederick Moore’s attitude to xxx on psychology xxx–xxxi thing-in-itself 31, 186 and appearances 24, 31 and freedom 25, 34 as unconditioned 30 as ultimate ground 26, 27 time unreality of xvi, xlviii, 13, 71, 127, 174 Trinity College Prize Fellowship competition xvii Twardowski, Kazimierz xxxi Ward, James xii, xiii, xviii, xxii will free will 47, 72 good will 78, 233 human 39, 57 illusion of uncaused will 40, 47 and moral law 76, 234 as natural cause 72 and prediction 40, 50
251