Descartes

DESCARTES Edited by JOHN COTTINGHAM

OXFORD UNIVERSITY PRESS 1998

-iii-

Oxford University Press, Great Clarendon Street, Oxford OX2 6DP Oxford New York Athens Auckland Bangkok Bogota Bombay Buenos Aires Calcutta Cape Town Dar es Salaam Delhi Florence Hong Kong Istanbul Karachi Kuala Lumpur Madras Madrid Melbourne Mexico City Nairobi Paris Singapore Taipei Tokyo Toronto Warsaw and associated companies in Berlin Ibadan Oxford is a registered trade mark of Oxford University Press Published in the United States by Oxford University Press Inc., New York Introduction and selection © Oxford University Press 1998 First published 1998 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press. Within the UK, exceptions are allowed in respect of any fair dealing for the purpose of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, or in the case of reprographic reproduction in accordance with the terms of the licences issued by the Copyright Licensing Agency. Enquiries concerning reproduction outside these terms and in other countries should be sent to the Rights Department, Oxford University Press, at the address above This book is sold subject to the condition that it shall not, by way of trade or otherwise, be lent, re-sold, hired out or otherwise circulated without the publisher's prior consent in any form of binding or cover other than that in which it is published and without a similar condition including this condition being imposed on the subsequent purchaser British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data Descartes / edited by John Cottingham. --(Oxford readings in philosophy) Includes bibliographical references and index. 1. Descartes, René, 1596-1650. I. Cottingham John, 1943II. Series. B1875.D3615 1998 194--dc21 98-2535 ISBN 0-19-875183-4 ISBN 0-19-875182-6 (pbk.) 1 3 5 7 9 10 8 6 4 2 Typeset by Best-set Typesetter Ltd., Hong Kong Printed in Great Britain on acid-free paper by Bookcraft (Bath) Limited Midsomer Norton, Somerset -iv-

CONTENTS Abbreviations and References Introduction John Cottingham I. Descartes and the Metaphysics of Doubt Michael Williams II. The Cogito and its Importance Peter Markie III. Clearness and Distinctness in Descartes Alan Gewirth IV. Foundationalism, Epistemic Principles, and the Cartesian Circle James Van Cleve V. Descartes on the Will Anthony Kenny VI. Descartes' Theory of Modality Jonathan Bennett VII. The Epistemological Argument for Mind-Body Distinctness Margaret D. Wilson VIII. Descartes and the Unity of the Human Being Geneviève Rodis-Lewis IX. Descartes' Theory of the Passions Stephen Gaukroger X. Descartes' Treatment of Animals John Cottingham

-v-

vii 1 28 50 79

101 132 160

186 197 211 225

XI. Descartes' Method and the Role of Experiment Daniel Garber XII. Descartes' Concept of Scientific Explanation Desmond M. Clarke XIII. Force (God) in Descartes' Physics Gary C. Hatfield Notes on the Contributors Bibliography Index of Citations of Descartes' Works Index of Names

-vi-

234 259 281 311 313 319 325

ABBREVIATIONS AND REFERENCES STANDARD EDITIONS AT C. Adam and P. Tannery (eds.), Œuvres de Descartes ( 12 vols., rev. edn., Paris: Vrin/CNRS, 1964-76). CSM J. Cottingham, R. Stoothoff, and D. Murdoch (eds.), The Philosophical Writings of Descartes, Vols. I and II ( Cambridge: Cambridge University Press, 1985). CSMK J. Cottingham, R. Stoothoff, D. Murdoch, and A. Kenny (eds.), The Philosophical Writings of Descartes, Vol. III, The Correspondence ( Cambridge: Cambridge University Press, 1991). To assist the reader, whenever references to the above editions were not supplied in the original printing of the articles included here, they have been added, in square brackets [ ], by the editor of the present volume. OTHER EDITIONS AG E. Anscombe and P. T. Geach (eds.), Descartes, Philosophical Writings ( London: Nelson, 1969). AM C. Adam and G. Milhaud (eds.), Correspondence de Descartes ( Paris: Presses Universitaires de France, 1936-63). CB J. G. Cottingham (ed.), Descartes' Conversation with Burman ( Oxford: Clarendon Press, 1976). Hall T. S. Hall (ed.), Descartes, Treatise on Man ( Cambridge, Mass.: Harvard University Press, 1972). HR E. S. Haldane and G. R. T. Ross (eds.), The Philosophical Works of Descartes ( 2 vols., Cambridge: Cambridge University Press, 1911). K A. Kenny (ed.), Descartes: Philosophical Letters ( Oxford: Oxford University Press, 1970).

-vii-

Olscamp P. J. Olscamp (ed.), Discourse on Method, Optics, Geometry and Meteorology ( Indianapolis: Bobbs-Merrill, 1965). Note: In the original printing of the articles, references to Descartes' individual works ( Discourse, Meditations, Regulae, etc.) were given in a number of different ways (e.g. in Latin, or French, or in various abbreviated forms), and to avoid confusion these have been changed to a consistent format throughout the volume. In the case of works other than by Descartes, references in the notes to frequently cited books or articles have been shortened to author and title alone; full publication details are given under the relevant heading in the Bibliography. Finally, additional explanatory notes have occasionally been inserted by the editor of the present volume; these are indicated by a letter suffix (e.g. '3a'); footnotes with a number alone are by the original authors.

-viii-

INTRODUCTION JOHN COTTINGHAM I Descartes is one of those very few philosophers whose ideas changed the shape of the subject. Whether for good or ill is at first hard to say. His reputation is a strangely ambivalent one: on the one hand, the revered 'father of modern philosophy'; on the other hand, the reviled source of such dangerous errors that the label 'Cartesian', 1 by the end of the twentieth century, has for many philosophers become almost a term of abuse. 2 But despite the frequent modern denunciation of his (actual or supposed) doctrines, Descartes is still perhaps the most widely studied philosopher in the canon. The scenario of his philosophical masterpiece, the Meditations, where the lonely thinker sits by the fire and systematically dismantles his previous beliefs in order to make the laborious journey from doubt to certainty, is familiar to countless undergraduates; and the first truth discovered along the road, cogito ergo sum, 'I am thinking, therefore I exist', 3 is perhaps the best-known saying in philosophy. Why should we study Descartes? A superficial, though true, answer would be that he is important for the 'history of ideas'. Yet Descartes' significance for today's philosophers is of a quite distinct kind from, for example, that which his great contemporary Galileo has for today's scientists. Galileo is respected as a brilliant early theorist and researcher, whom ____________________ 1 'Cartesian': pertaining to Descartes, from 'Cartesius', the Latin version of his name. Many of Descartes' writings were published in Latin, which was still the language most commonly used in the seventeenth century by those who wished to reach an International audience. 2 Compare the following remark: ' Frege exposed Cartesian errors about thought while retaining . . . Cartesian errors about ideas. It is as if the residual Cartesian poison running through the philosophical systems of the day was gathered by Frege . . . Into a single virulent boil ready to be lanced by Wittgenstein' ( Anthony Kenny, Frege (Harmondsworth: Penguin, 1995), 212). 3 The actual wording of the famous Latin phrase does not occur in the Meditations ( 1641), but may be found in the Principles of Philosophy ( 1644); its French equivalent, 'je pense done je suis', occurs in the earlier Discourse on the Method ( 1637).

-1-

present-day physicists can look back to as a pioneer of modern dynamics and astronomy. But no modern scientist spends time seriously debating the truth of Galileo's views: what was wrong has long since been discarded; what was on the right lines, long since incorporated. With Descartes' relationship to modern philosophy, things are very different. In the first place, there is still a small but by no means negligible minority of modern philosophers whose outlook in various key areas of the subject might fairly be described as Cartesian. 4 In the second place, even among that larger number of philosophers who regard most of Descartes' philosophical doctrines as suspect, those doctrines still carry a powerful resonance: so far from their being of merely 'historical' interest, some of the twentieth century's most original thinkers have expended vast amounts of energy in combating them. In the third place, and most strikingly of all, Descartes' thinking, irrespective of whether or not his conclusions are accepted, has had a pervasive and enduring effect on the very structure of the philosophical agenda. When philosophers ask about what counts as certain and reliable knowledge, about the relationship between mind and matter, the respective roles of reason and experience in scientific inquiry, the nature of subjectivity and how it relates to the objective world of science--the way in which these central problems are formulated would be largely inconceivable without the particular stamp which Descartes placed upon philosophy. René Descartes was born in 1596 in the small town between Tours and Poitiers which now bears his name. His mother died when he was just over 1 year old, and he was brought up by his maternal grandmother; his father remarried when he was 4. At the age of 10, he was sent away to boardingschool at the Jesuit College of La Flèche (between Angers and Le Mans), where he remained till he was 18. He was educated in the tradition now known as 'scholasticism'--the comprehensive body of philosophy based on the teachings of Aristotle, as systematically adapted to the demands of the Christian faith by the great thirteenth-century philosopher and theologian Thomas Aquinas. The scholastic tradition had dominated the universities and schools of Europe for many centuries by the time Descartes was ____________________ 4 For direct supporters of the Cartesian conception of the mind as a non-material substance, see e.g. Richard Swinburne, The Evolution of the Soul ( Oxford: Clarendon Press, 1986), and Michael Foster, The Immaterial Self ( London: Routledge, 1991). Philosophers who reject Cartesian dualism, but retain the Cartesian notion of a mental domain inaccessible from the objective perspective of physical science, include T. Nagel, in Mortal Questions ( Cambridge: Cambridge University Press, 1979), ch. 12, and M. Lockwood, in The Mind, the Brain and the Quantum ( Oxford: Blackwell, 1989), ch. 16. For the influence of Descartes' thought in the theory of language use and acquisition, see Noam Chomsky, Cartesian Linguistics ( New York: Harper and Row, 1966), esp. 59-72. -2a pupil, and he studied such recent exponents of the genre as the Spanish Jesuit Francisco Suarez (whose vast Metaphysical Disputations was published in 1597) and the Feuillant

scholar Eustachius a Sancto Paulo, whose elegant Compendium of Philosophy in Four Parts, first published in 1609, became a highly popular textbook; it aimed to provide a definitive summary of the complete corpus of established philosophical knowledge, covering logic, physics, metaphysics, and ethics. But the old order was crumbling. Early in 1610, in Padua, Galileo turned his newly constructed telescope on Jupiter, and discovered four satellites orbiting round it--the first really hard piece of evidence against the traditional Ptolemaic view that the Earth was the centre of all motion; he also observed mountains on the Moon, a decisive refutation of the hallowed Aristotelian thesis that all celestial bodies are perfectly spherical. What the young Descartes, then 14, made of these events, we cannot guess, but it is known that he was one of those chosen to take part in a spectacular ceremony held at La Flèche that summer to mark the death of the College's royal benefactor, Henry IV; in the course of the solemnities a sonnet was recited hailing the discovery of the moons of Jupiter, which 'brightened the gloom of the King's death'. 'I was educated', Descartes subsequently wrote, 'in one of the most famous schools of Europe, where I thought there must be learned men if they existed anywhere on earth'; nevertheless, he wryly observed, 'I found myself beset by so many doubts and errors that I came to think I had gained nothing from my attempts to become educated but increasing recognition of my ignorance' ( A T VI4-5: C S M I113). When, as a young man, Descartes decided to seek for knowledge in 'the great book of the world' ( A T VI9: C S M I115), he found a vastly more stimulating mentor in the Dutchman Isaac Beeckman (seven years his senior), whom he met accidentally in Breda, on 10 November 1618. The scholastic curriculum Descartes had followed at school firmly separated physics from mathematics, the latter being supposed (because it dealt with purely abstract objects) to have no relevance to the study of actually existing bodies. Beeckman, by contrast, grasping what was to be the key to the seventeenth-century revolution in science, aimed for an approach which 'combined mathematics and physics in an exact way'. 5 'You alone', Descartes wrote to him a few months later, 'roused me from my state of indolence . . . If perhaps I should produce something not wholly to be despised, you can rightly claim it as your own' ( A T X163: C S M K4). ____________________ 5 The phrase comes from Beeckman journal for December 1619; see Journal tenu par Isaac Beeckman de 1604 à 1634, ed. C. de Waard et al. ( The Hague: Nijhoff, 193953), 244. -3Exactly a year after meeting Beeckman, at the age of 23, Descartes found himself in Ulm, on the Danube; on 10 November, he remained closeted all day in a 'stove-heated room' where, he later wrote, 'I was completely free to converse with myself about my own thoughts' ( A T VI11: C S M I116). The long hours of intense meditation were followed by a night of powerfully vivid dreams, involving (amongst other things) a discussion about a line from the poet Ausonius, 'Quod vitae sectabor iter?' (What road in life shall I

follow?), and a vision of an 'encyclopaedia', which seemed to symbolize 'all the sciences collected together'. 6 Descartes awoke convinced he was destined to found a new philosophical system. After further travels in Europe, interspersed with time spent in Paris, Descartes decided to emigrate to the Netherlands in 1628; it was to remain his home (though with frequent changes of address) for the rest of his life. His earliest major philosophical work (begun soon after his night of troubled dreams, but abandoned unfinished before he left for Holland) was the Regulae ad directionem ingenii ('Rules for the Direction of our Native Intelligence'). The Regulae emphasizes the importance of the perceptions (or 'intuitions') of a clear and attentive mind, purged of the fluctuating impressions of the senses, and typified by the kind of mathematical vision which enables us to grasp with perfect certainty that 'a sphere is bounded by a single surface' ( A T X368: C S M I14). But Descartes proposes to extend the simplicity and accuracy of mathematical knowledge to cover all subjects whose properties can be expressed in abstract terms. He envisages a 'discipline which contains the primary rudiments of human reason and extends to the discovery of truths in any field whatsoever . . . a general science which explains all the points that can be raised concerning order and measure, irrespective of the subject-matter' ( A T X374, 378: C S M I17, 19). Once arrived in Holland, Descartes worked at a number of scientific and mathematical projects which had interested him for some time: he developed a system for representing geometrical problems algebraically, laying down the foundations of what we now know as co-ordinate geometry; he did research on the mathematical analysis of problems in optics, anticipating Snell's law of refraction; and he applied mechanical principles of explanation to a number of problems in meteorology. The results of this work were eventually to be published in 1637 as a collection of three ____________________ 6 The story of the dreams is told by Descartes' early biographer Adrien Baillet, and is based on material from Descartes' notebooks, which were found among his papers shortly after his death. See Baillet, Vie de Monsieur Des-Cartes, Bk. I, ch. 1; cf. Cottingham, Descartes, 161 ff. -4essays, La Dioptrique, Les Météors, and La Géométrie. By 1633, Descartes also had ready for publication a full statement of his theories of physics and cosmology--Le Monde ('The World', or 'The Universe'); in a concluding portion of the work, the Traité de l'homme ('Treatise on the Human Being'), Descartes dealt with physiology and the nervous system, using the same mechanical principles he had employed to explain the origins and formation of the stars and planets. The reduction of biology to physics, and indeed the idea that all corporeal phenomena could be explained simply by reference to the movement of particles of a certain size and shape, is one of the central planks of Cartesian scientific method; in place of the Aristotelian idea that different areas of science require different standards of precision and different principles of explanation,

Descartes offers a single uniform set of simple mathematico-mechanical principles to cover all known phenomena. To the last sentence there belongs a crucial qualification: except for the phenomena of human thought and consciousness. In the Discours de la méthode ('Discourse on the Method'), 7 prefixed to the three scientific essays of 1637, Descartes insisted that while all animal and a considerable amount of human behaviour could be explained 'according to the [mechanical] disposition of the internal organs', like a 'clock consisting only of wheels and springs', the special functions of thought and language require a 'rational soul', which 'unlike other things, cannot be derived in any way from the potentiality of matter, but must be specially created' ( A T VI 59: C S M I141). This is the famous, or infamous, thesis of mind-body 'dualism': the view that the mind (or 'soul'--Descartes uses the two terms interchangeably) 8 is an independent substance, wholly distinct from the body: 'This "I"--that is, the soul by which I am what I am--is entirely distinct from the body, and indeed is easier to know than the body, and would not fail to be whatever it is, even if the body did not exist' ( A T VI 33: C S M I127). The Discourse reaches this striking conclusion by applying Descartes' so-called method of doubt: 'I decided to . . . reject as absolutely false anything in which I could imagine the least doubt, or order to see if I was left believing anything that was entirely indubitable.' It is possible, Descartes argues, to pretend that everything of which I am aware has no more reality than 'the illusions of my dreams'; yet 'while I was trying in this way to think that everything was false, immediately I noticed that it was necessary that ____________________ 7 The full title says 'Discourse on the Method of rightly conducting reason and reaching the truth in the sciences'. The Discourse, together with the three essays designed to illustrate the method, was published anonymously in Leiden in June 1637. 8 cf. Synopsis to Meditations, French version ( 1647) ( A T IXA10: C S M II10 n. 3). -5this truth, I am thinking, therefore I exist, was so firm and sure that even the most extravagant suppositions of the sceptics were incapable of shaking it' ( A T VI32: C S M I127). 'Je pense, donc je suis' (or, in Latin, 'Cogito ergo sum'): this is the first indubitable discovery along the road to the 'long chain' of interconnected knowledge that Descartes proudly advertised.9 Descartes now swiftly proceeds to reason that he remains unable to doubt his own existence, even while doubting that he has a body, or that there is any physical world at all; hence, he concludes, 'I am a substance whose whole essence is simply to think . . . and does not depend on any material thing in order to exist' (ibid.). In the Meditations on First Philosophy ( 1641) published in Latin four years after the Discourse, Descartes develops the brief and schematic argument of the earlier work in much more detailed and sophisticated terms. 'First Philosophy' was the traditional term for metaphysics; for Aristotle it was the study of 'being qua being', with special reference to the nature of substance (that which exists in its own right); later, in the Middle Ages,

metaphysical inquiries tended to be very closely bound up with discussion of the supreme substance, God. Descartes followed broadly in this tradition, but observed to his editor Marin Mersenne that in the Meditations 'the discussion is not confined to God and the soul, but treats in general of all the first things to be discovered by philosophizing' ( A T III 235: C S M K157). As the meditator branches out from awareness of his own existence to proving the existence of God, the argument aims to establish the 'clear and distinct' perceptions of the intellect, purged of the confused deliverances of the senses, as the foundation of systematic and reliable knowledge (what Descartes calls scientia)-knowledge both of the three-dimensional world around us ('the whole of that corporeal nature which is the subject-matter of pure mathematics'), and also of the nature of the mind, a substance that Descartes takes to be indivisible, unextended, and entirely distinct from the body. In 1644, Descartes published, in Latin, a lengthy four-part compendium of his metaphysical and scientific views, the Principia philosophiae ('Principles of Philosophy'). In a preface to the French translation, which was issued in 1647, Descartes used a famous metaphor to express his ambition to construct a complete system of knowledge: 'the whole of philosophy is like a tree: the roots are metaphysics, the trunk is physics, and the branches emerging from the trunk are all the other sciences, which may be reduced to three principal ones, namely medicine, mechanics and morals' ( A T IXB14: C S M I186). Descartes had always presented his philosophy as ____________________ 9 Cf. Discourse, part ii ( A T VI19: C S M I120) and part v ( A T VI40-1: C S M131). -6primarily practical rather than speculative, having the clear potential to improve the human lot--in mechanics, by controlling the environment and making ourselves 'masters and possessors of nature', and, in medicine, by 'freeing us from innumerable diseases and perhaps even from the infirmity of old age' ( Discourse, part vi, A T VI62: C S M I143). The place of morals in Cartesian science is, however, more problematic, since the question of how we should live relates to our uniquely human nature--a nature which, for Descartes, is fully explained neither in terms of the mechanisms of the body (in principle the same as those to be found anywhere in 'corporeal nature') nor, on the other hand, in terms of the purely intellectual and volitional operations of the mind, which are 'entirely independent' of the body. Despite the catch-phrase so often attached to his views, Descartes regarded the human being as more than a 'ghost in a machine' 10 (an incorporeal spirit manipulating the mechanisms of the body); rather, it is what he called a 'real substantial union of mind and body', characterized by a rich repertoire of properties-feelings, sensation, emotions, passions--that are not reducible either to pure thought or to physical extension. The attempt to understand our human nature, and to use that understanding to complete his science of morals, occupied much of Descartes' time during the 1640s, culminating in his last book, Les Passions de l'âme ('The Passions of the Soul'), which appeared in 1649.

Within four months of its publication, the author was dead, a victim of pneumonia, which he contracted in Stockholm on a prolonged visit to the court of Queen Christina of Sweden. By the time of his premature death, the 'new philosophy' of Descartes was already being widely discussed throughout Europe, and despite numerous attempts to suppress his theories in the latter half of the seventeenth century, and despite the subsequent eclipse of his physics by that of Newton, Descartes' ideas and presuppositions increasingly became part of the standard framework of philosophical inquiry. Philosophy, of course, is never a static affair, and our understanding of many of the problems Descartes raised has been reshaped and transformed by the work of those who came after him-thinkers such as Leibniz, Hume, and Kant, to name but three of the most important. Yet all these three, and many of their successors, can be seen, in different ways and in varying degrees, as inheritors of the Cartesian tradition in philosophy. Today, as we began by observing, things are very different; but the fact that so much contemporary philosophy feels compelled to insist on its emancipation from Cartesian paradigms is in itself a remarkable tribute ____________________ 10 See Gilbert Ryle, The Concept of Mind ( London: Hutchinson, 1949), ch. 1. -7to the pervasive influence of the philosophical revolution which Descartes inaugurated. II In embarking on the metaphysical inquiries designed to yield the foundations of his new philosophy, Descartes, as we have seen, deploys his celebrated 'method of doubt' to sort out what is reliable in his previous beliefs from what must be discarded. As Michael Williams puts it, in the opening essay of this collection, the Meditations follows 'a journey from prephilosophical common sense to metaphysical enlightenment, each step of which is taken in response to an encounter with scepticism' (p. 28 below). Descartes himself, of course, was not a sceptic: 'no sane person', he wrote in the Synopsis to the Meditations, 'has ever doubted that there really is a world, and that human beings have bodies, and so on' ( A T VII16: C S M II 11). But he uses the weapons of the sceptic to encourage us to weed out the 'prejudices' or 'preconceived opinions' (Latin praejudicia) 11 that we may have uncritically accepted from parents, teachers, and other authorities; what is left after the weeding process, what survives the doubt, will provide 'foundations for the sciences that are stable and likely to last' ( A T VII17: C S M II12). But although the First Meditation presents a series of doubts in the voice of the ordinary 'pre-philosophical' person, Michael Williams points out that 'the doubts are much less natural, much less metaphysically non-committal, than they are made to seem' (p. 29 ). Arguments for doubting our ordinary beliefs had been around for a long time before Descartes: many of the traditional lines of reasoning used by the ancient Greek sceptics were presented in the works of Sextus Empiricus, who wrote (in Greek) around AD 190; and Latin editions of Sextus, making his work available to a much wider audience, had been published in the 1560s, not long before Descartes' birth. But Descartes' reasons for doubt, though drawing

on some of the traditional materials, are structured in a very different way. In the first place, as Williams notes, the classical presentation of scepticism is 'flat': a variety of arguments of roughly equal status are deployed to impugn a variety of types of belief. But Descartes' arguments are directed particularly against the material world: like Plato and Augustine before him, Descartes sees true knowledge as relating not to the fluctuating world of the senses, but to a more abstract, intellectual domain. And in the second place, Descartes' doubts are what Williams calls ____________________ 11 For the role of praejudicia or consuetae opiniones ('habitual opinions'), see A T VII22: C S M II15. -8stratified. Exposing first the liability of the senses to optical illusions and other errors, next our lack of 'sure signs' to distinguish waking from sleeping experiences, and finally even the certainty of our belief in the existence of those 'general kinds of things' out of which our dreams are composed, the Cartesian meditator probes 'ever more deeply into the bases of our convictions' (p. 34 ). This differentiation of levels of doubt is one o f Descartes' most characteristic innovations. And it has three important implications. First, by exploring the suggestion that dreams can, in effect, reproduce the content of waking experiences, Descartes prepares the way for a radical kind of doubt never raised in classical scepticism: 'the sky, the air, the earth, colours, shapes sounds and all external things' may be no more than 'the delusions of dreanis'--snares which the imaginary 'malicious demon' has 'devised to ensnare my judgement' ( A T VII22: C S M II15). Secondly, '[t]he progressive development of Cartesian doubt insinuates, without ever directly arguing for, a foundational conception of knowledge' ( Williams, pp. 37-8 below); in other words, Descartes is inviting us to accept the assumption that, however extreme our doubts about what exists 'out there', there will be a more fundamental level of certainty, relating to inner consciousness and its contents. Williams draws the striking conclusion that the sceptical arguments of the First Meditation are already 'loaded' in favour of a conception in which the 'thoughts', or mental contents, of the subject are taken to enjoy a special sort of immediacy and directness. Williams describes the 'loading' involved here as 'metaphysical', though the reader may initially feel that it might more properly be called 'epistemic' (and certainly Descartes himself distinguished the 'order of discovery' followed in the Meditations from the metaphysical order of reality that underpins it). If Williams is right, however, Descartes' procedure assumes from the start a certain picture of reality, a new conception of the 'mental' whereby the 'sensory contents' of the mind can be abstracted from the (bodily) senses that give rise to them. So far from being a natural presentation of the ordinary pre-philosophical outlook, the story told in the Meditations is a highly artificial one. '[T]he scepticism whose limits are explored involves the special dubitability of beliefs about the external world, and the external world only emerges as a distinct problem once the important metaphysical concessions have already been made' (p. 48 ).

The first foothold of certainty which the meditator attains, as he tumbles around in the 'deep whirlpool' of doubt (first paragraph of Second Meditation), is knowledge of his own existence. Even given the extreme doubt -9represented by the scenario of the malicious demon, bent on deceiving him in every possible way, Descartes recognizes one indubitable truth: let him deceive me as much as he can, he will never bring it about that I am nothing so long as I think that I am something. So after considering everything very thoroughly, I must finally conclude that this proposition, I am, I exist, is necessarily true whenever it is put forward by me or conceived in my mind. ( A T VII25: C S M II17) Although this line of thought has an immediate and intuitive appeal for most people, the precise status of 'the Cogito, as it has come to be called, has been minutely analysed and debated by generations of philosophical critics and commentators. Descartes tended to respond rather brusquely to those in his own day who wished to debate it: 'this knowledge is not the work of your reasoning, or information passed on to you by your teachers; it is something your mind sees, feels and handles' ( A T V138: C S M K331). But what exactly does this come down to? As Peter Markie points out in his essay (Chapter II), Descartes often talks in ways which suggest that there is an immediate self-evident inference from one's thinking to one's existence. Yet the inference is not, as it were, wholly self-contained and self-sufficient: to make it, Descartes conceded, you have to know what thought is, what existence is, and also the 'simple notion' that it is impossible to think without existing ( Principles of Philosophy, part i, art. 10). This is not a matter (as Descartes explained elsewhere) of deducing one's existence from one's thinking by means of a syllogism, starting with the major premiss 'whatever thinks exists' ( A T VII140: C S M II100); rather, given simply a knowledge of the meanings of the terms, and the selfevident relationship between them, one is directly aware in one's own case that for as long as one is thinking, one must exist. Yet, what is it that gives thinking a privileged certainty in the first place? 'Ambulo ergo sum' (I am walking, therefore I exist) will not do for Descartes, since I might be mistaken about the truth of 'ambulo' (I might be only dreaming I am walking, or the malicious demon might be making me hallucinate that I am walking 12 ). But if a deceptive god can trick me about physical actions in this way, why should he not also be able to trick me about the things I think I see clearly with my mind's eye, such as that I am thinking (compare Descartes' own remarks about this in the Third Meditation, A T VII36: C S M II25). This notion of 'seeing with the mind's eye', elsewhere called 'intuition' or 'clear and distinct perception', is, according to Markie, 'the least clear and distinct concept' in Descartes' ____________________ 12 Or, to use a fantastic scenario beloved of present-day philosophers, I might be a

surgically excised brain, kept alive in a vat of nutrients by an insane scientist who delights in stimulating the cerebral cortex so as to give me all the sensations of walking. -10philosophy: '[h]e never adequately explains what this mental vision is, or why apprehending a proposition by it is sufficient to make belief in the proposition very reasonable' (p. 71 ), A further difficulty, as many commentators have pointed out, is that the referent of the 'I' in 'I think' is not something Descartes is really entitled to be sure of. '[W]hat is it about Descartes' self-awareness when he clearly and distinctly perceives that he thinks that makes his awareness an awareness of him?' ( Markie, p. 73 below). Descartes takes himself to be aware of an individual, unitary self, but never quite explains how that self is to be individuated ('How do I know that I am not ten thinkers thinking in unison?', asks Elizabeth Anscombe, echoing a worry first voiced by one of Descartes' contemporaries 13 ). Finally, there is a problem about how far Descartes' certainty about his thinking extends. Sometimes Descartes uses the term cogitatio ('thought') in a very broad way, which seems to include volitions, desires, and items of sensory awareness (e.g. 'I long to hear her voice'; 'I seem. to smell perfume'), as well as intellectual activities like doubting and cogitating; yet does this mean that the meditator can be indubitably certain that he is angry or depressed or in love ( Markie, p. 76), and that any of these mental states would do as a basis for the certain knowledge of his own existence? There are complex questions to be explored here, some of which have still to be answered satisfactorily by defenders of Descartes. 'I am certain that I am a thinking thing,' says Descartes, taking stock at the start of the Third Meditation; and he proceeds to ask whether, in virtue of knowing this, he can identify what is required for being certain about anything. The answer he comes up with is what has come to be known as the 'truth rule': 'I seem to be able to lay it down as a general rule that whatever I perceive very clear and distinctly is true' ( A T VII35: C S M II 25). We might feel inclined to object (using a phrase of Wittgenstein's in a rather different context): 'how can he generalize so irresponsibly from a single case?' Moreover, as we have already seen, there are problems about the very notions of clarity and distinctness. What exactly are these supposed features of our mental perceptions? How are they defined, and how do we identify them? In his classic paper on this topic, Alan Gewirth (Chapter III) starts by pointing out that the immediate objects of our perception are, for Descartes, ideas,: 'I take the word "Idea" to refer to whatever is immediately perceived by the mind' (Second Replies, A T VII ____________________ 13 Anscombe is quoted by Markie, p. 75 below. For a seventeenth-century anticipation, see the letter to Descartes from 'Hyperaspistes' of July 1614 ( A T III404). -11-

181 : C S M II127). Ideas have (to use the modern jargon 14 ) 'intentionality': they represent or refer to things. But since 'we can have no knowledge of things except via ideas' ( A T III476: C S M K202), clearness and distinctness, Gewirth points out, must be 'qualities internal to ideas and perceptive acts' (p. 82 below). Now this fact is often interpreted as implying that Descartes 'privatized' ideas, or regarded them as inhabiting an entirely subjective domain; and in this case, clearness and distinctness would be purely psychological (as opposed to logical) aspects of our thought, which makes it hard to see how they could be indicators of truth. Gewirth, by contrast, argues that Descartes characterizes clearness and distinctness in normative terms. For an idea to be clear and distinct, Gewirth suggests, there has to be an "equality" of its direct and interpretative contents' (p. 87 ), and (as explained by Gewirth) this turns out to be a logical, not a psychological, relation--one which depends on an analysis of the representational content of the idea, as opposed to subjective features like, say, vividness. This, however, still does not indicate how we tell whether an idea is clear and distinct. Here Gewirth explains Descartes' method by invoking his example of the wax in the Second Meditation: starting with an obscure and confused conception of the wax, based on sensory qualities (like smell and colour) which turn out to vary under different conditions, the meditator slowly reaches towards a more stable and precise conception by discarding every feature that is non-essential to the wax, until (as Gewirth puts it) 'a direct content is attained which survives every reductive device, remaining so long as the object can be conceived, and without which the object can no longer be conceived' (p. 96 below). A further problem remains, however: even after this careful and precise procedure, there still seems to be a gap between the 'idea' directly perceived by the mind and the real 'thing' it represents. It is in order to ensure that ideas match reality that the role of God in Descartes' system is so crucial: in Gewirth's words, 'the methodological orientation must be supplemented by a metaphysical one, culminating in the divine guarantee' (p. 98 ). Descartes' invoking of God as the guarantor of human knowledge is a pervasive feature of his metaphysics: 'God is supremely perfect and cannot be a deceiver on pain of contradiction; hence every clear and distinct ____________________ 14 The term (not used by Gewirth) is in fact a revival of a medieval Latinism (from intentio-a reaching out for or pointing at something); the label, common though it has become, is not a very happy one, since it risks confusion with the ordinary meaning of 'intentional'. Descartes himself avoids the term, but uses an even more opaque piece of jargon, 'objective reality' ( A T VII41: C S M II28). It would be more perspicuous to talk of the 'representational' nature of ideas, or their 'aboutness'. -12perception is undoubtedly true' (Fourth Meditation, A T VII62: C S M II 43). But this at once raises the thorny problem of the 'Cartesian Circle' (first pointed out by Descartes' acute contemporaries, Marin Mersenne and Antoine Arnauld 15 ): if we can be sure of the

truth of our clear and distinct perceptions only once we have established God's existence, how are we entitled to rely on the clear and distinct perceptions needed to prove his existence in the first place? The problem seems to be a devastating one for Descartes' procedure, since it suggests that the Cartesian journey to knowledge can never get started. In his analysis of the issue, James Van Cleve (Chapter IV) discusses and rejects several proffered solutions, including the view that in the early stages of his inquiry Descartes regards it as sufficient that he is irresistibly convinced of the truth of certain propositions, God being invoked at a later stage to provide a metaphysical guarantee of their objective truth. Van Cleve rejects this solution on the grounds that Descartes is playing for 'higher stakes' (p. 106 ): in seeking the foundations for his new science, he demands the kind of certainty that implies not just psychological conviction, but also objective truth. Moreover, such certainty must be able to survive every 'epistemically possible' doubt (p. 108 )--including what Descartes himself called the 'hyperbolical' scenario of the malicious demon. Van Cleve's own solution trades on the idea that, to get started, Descartes does not require a general guarantee of the truth of all clear and distinct perceptions; it is enough that for any given proposition, if he clearly and distinctly percieves it, then he is certain of its truth (what Van Cleve calls 'principle A'). Now if we accept that the meditator does clearly and distinctly perceive a stock of premisses (such as 'If I think, I exist'), which he knows for certain to be true, the crucial point is that, to have such knowledge, the meditator does not need to apply or invoke principle A; he simply needs for his initial inquiries to be covered by that principle. The issues here reach far beyond Cartesian scholarship, since, as Van Cleve shows, they are crucially relevant to debates in contemporary epistemology between 'foundationalist' and 'coherentist' theories of knowledge. Descartes (who may be regarded as the father of foundationalism) is often criticized on the grounds that he cannot (on pain of circularity, or an infinite regress) justify the fundamental epistemic principles he employs. Van Cleve argues, by contrast, that this criticism rests on a misunderstanding, since 'a subject need not know that an epistemic principle is true in order for the circumstance [it specifies] to give him knowledge' (p. 120 ). The debate continues. ____________________ 15 See Second Objections ( A T VII125: C S M II89) and Fourth Objections ( A T VII214: C S M II150). -13III The central role of God in Descartes epistemology and metaphysics presents the Cartesian inquirer with a philosophical problem somewhat analogous to the ancient 'problem of evil' in Christian theology. The theologian asks why, if creation is the work of a God who is supremely good, there is evil in the world; the Cartesian seeker for knowledge asks why, if his mental powers were bestowed on him by a perfect God, he is none the less so often subject to intellectual error. Like some of his theological

predecessors, Descartes identifies human freedom as the culprit; or, to be more precise, he takes our liability to error to be a function of the relationship between the intellect and the will. Our intellect, though limited (there are many things we finite beings do not perceive clearly) is none the less a perfect instrument, in so far as whatever we do clearly and distinctly perceive is guaranteed to be true. But our will, as Descartes puts it, 'extends wider than the intellect' (Fourth Meditation, A T VII58: C S M II40)--that is, we have the freedom to jump in and affirm or deny a proposition even when we do not clearly and distinctly perceive its truth or falsity. The route to truth thus becomes a matter of restraining the will, and remembering to give assent only to what is clearly and distinctly perceived. In his analysis of the Cartesian theory of the will, Anthony Kenny (Chapter V) argues first of all that in his early writings Descartes followed scholastic orthodoxy in treating judgement (affirming or denying a proposition) as an intellectual act, but that by the 1640s he had arrived at the position that judgement should be considered as an act of will, like pursuing or shunning. Kenny proceeds to ask why Descartes should have lumped judgement together with desire and aversion in this way, separating it from intellectual perception; and he suggests as a possible justification that assenting or dissenting to a proposition can be regarded as a form of 'commitment', subject, like other commitments, to being appraised as 'sincere or insincere, rash or cautious, right or wrong' (p. 144 ). But if this is the rationale for Descartes' position, Kenny argues that it is fundamentally flawed, since the 'onus of match' applying to beliefs is quite the opposite to that applying to desires: if I assert it to be a sunny day, then things go wrong if my assertion fails to fit the weather; but if I desire it to be a sunny day, then things go wrong if the weather fails to match my wishes. Having dealt with these conceptual preliminaries, Kenny goes on in the second half of his essay to expound a long-standing and vitally important distinction between. two conceptions of freedom: firstly, 'freedom of indif-14ference', which is a two-way, contra-causal power to X or not to X; and secondly, 'freedom of spontaneity', which is the freedom I have when I do something simply because I want to. Now the situation of the meditator when he focuses on some clearly and distinctly perceived proposition is, according to Descartes, that he cannot but assent to its truth: 'from a great light in the intellect', says the Fourth Meditation, 'there follows a great inclination of the will' ( A T VII59: C S M II41). When my intellect is illuminated by the 'light of reason', I spontaneously, and irresistibly, assent to the truth of the relevant proposition. My assent here is completely determined, but I am none the less free in the second sense indicated above; as Descartes puts it, I have a 'spontaneous' belief which is 'uncompelled by any external force' ( ibid. ). But what of Descartes' attitude to wholly undetermined freedom, or freedom of indifference? This is complicated by the fact that when he speaks disparagingly of 'indifference' in the Fourth Meditation (as 'the lowest grade of freedom'), he means the situation when I am 'wavering', with insufficient information to enable me to decide. There is another passage, however (from a letter to

Mesland of 9 February 1645 16 ), where Descartes appears to claim that we have freedom of the will in the traditional strong sense of 'indifference'--a 'positive faculty of determining oneself to one or other of two contraries' ( A T IV 173: C S M K245). This at first sight suggests an inconsistency in Descartes, since if we always enjoy freedom of indifference in this indeterministic sense, this implies that the will is not constrained by the light of reason, but can after all resist assenting even to clearly and distinctly perceived truths. Kenny argues convincingly, however, that Descartes holds such 'liberty of perversion' (withholding assent from a clearly perceived truth) to be available only subsequently to a given perception (if, for example, we allow the relevant proposition to slip out of focus); but at the time we are actually perceiving it, we cannot but assent. Descartes' pronouncements on freedom are not always easy to decipher, but Kenny makes a strong case for the consistency of his position. As for whether 'liberty of spontaneity' is a sufficiently robust notion to support the concepts of genuine human autonomy and responsibility, this is something that remains in hot dispute among philosophers even today. Descartes' attribution to the will of powers often previously assigned to the intellect is part of a tendency in his philosophy which is sometimes called 'voluntarist'. Another striking example is his doctrine that the 'eternal truths' (the necessary truths of logic and mathematics) are creations of the ____________________ 16 As Kenny notes, the date and addressee of this letter are in doubt. -15will of God. The dominant tradition in philosophy regards such truths as propositions which are 'true in all possible worlds', or which 'could not have been otherwise'. Descartes, by contrast, asserts that they are the result of a divine fiat, just like the creation of this world: God 'was free to make it not true that the radii of the circle are equal--just as free as he was not to create the world' ( A T I152: C S M K 25). This 'allegedly strange, peculiar, curious, and incoherent doctrine' ( Jonathan Bennett, p. 160 below) first appears in a number of letters to Mersenne in the early 1630s, and although it is not explicitly asserted in the Discourse, Meditations, or Principles, there are other texts that make it plain that Descartes continued to maintain it. The problem with Descartes' allowing this 'worm of contingency' 17 into the centre of his system is that it seems clearly inconsistent with the 'absolute certainty' that he sometimes claimed for his philosophy. 18 Indeed, if nothing is absolutely necessary or impossible, if God could even have made contradictory propositions both true (as Descartes sometimes goes so far as to assert-- A T IV118: C S M K 235), then the logical foundations of all valid reasoning seem threatened. In his interesting attempt to rescue Descartes from some of these problems, Jonathan Bennett (Chapter VI) points out, first of all, that the passages where Descartes discusses his doctrine of the divine creation of the eternal truths tend primarily to emphasize the greatness of God--the total dependence of all things on him--rather than his

'omnipotence' (in the sense of an ability to do absolutely anything). The main impulse behind his discussions seems to be one of piety and respect for the Deity, a refusal even to contemplate that there might be anything at all that does not depend on his will. Moreover, even in the passages where Descartes apparently suggests that God could do anything, he always couples this with a warning that we should not attempt to grasp what he could or should do; 19 and as Bennett points out, this 'cuts both ways', for '[a]nyone who asserts "God can do everything" implicitly claims to grasp what God can do' ( Bennett, p. 164 below). It is important to remember that Descartes frequently stresses that God is 'incomprehensible': we can know he exists, and understand some of his attributes, but we cannot fully 'comprehend' or grasp him--'just as we can touch a mountain, but not put our arms round it' ( A T I152: C S M K 25). There is, however, for Bennett, a purely philosophical rationale for Descartes' theory of the dependence of the eternal truths on God's will, ____________________ 17 See Cottingham, 'The Cartesian Legacy'. 18 'Absolute certainty arises when we believe it is wholly impossible for something to be other than we judge it to be' ( Principles of Philosophy, part iv, art. 206, French version). 19 Compare the letter to Mesland of 2 May 1644 ( A T III118-9: C S M K235). -16one that is independent of considerations of religious piety, and concerns instead the analysis of 'modal concepts' (that is, concepts involving the notions of necessity and possibility). Bennett's claim is that ' Descartes held . . . that our modal concepts should be understood or analysed in terms of what does or does not lie within the compass of our ways of thinking' (p. 167 ). The necessity of a proposition like '2 + 2 = 4', in other words, consists in our being unable to conceive otherwise. Moreover, God's 'making' it necessarily true amounts to his creating our nature such that we cannot understand any other possibility than that two and two should equal four. If the Descartes who emerges on this interpretation appears as a distinctly 'modern' figure, 20 this is a bullet Bennett is quite prepared to bite: 'there was in Descartes a wide, deep, vivid streak of subjectivism or pragmatism about truth--a willingness to treat results about the settlement of belief as though they were results about how things stand in reality' (p. 171 ). Bennett does not claim that his interpretation can be squared with everything Descartes wrote; indeed, he discerns a tension between this subjectivist stance and residual objectivist hankerings which often lead Descartes into 'notoriously shaky and murky' arguments. But Bennett's resounding conclusion is that the 'best and most disciplined parts' of Descartes' epistemology and metaphysics belong to this subjectivist strand in his thought. IV We turn now to the part of Descartes' philosophy that has generated the most sustained and intense opposition among twentieth-century thinkers, and indeed was searchingly

questioned by critics of his own time--his philosophy of mind. The Cartesian claim that the conscious self (at least in respect of its two main faculties of understanding and willing 21 ) is incorporeal--entirely independent of the body and capable of existing without it-is harshly out of tune with most present-day work on the nature of mind. The point perhaps needs qualifying: certainly there are many philosophers ____________________ 20 One might say (though Bennett himself does not use the term) that Descartes emerges on this interpretation as in some respects a 'proto-Kantian'. Kant, more than 100 years later, was to suggest a 'Copernican' revolution in philosophy, whereby certain features hitherto supposed to belong to nature 'in itself' are construed as necessary categories in terms of which our human experience and understanding is structured (just as Copernicus had proposed that the supposed motion of the celestial sphere is in fact to be explained in terms of the motion of our own planet). 21 Feeling, emotion, and sensory awareness are, for Descartes, rather different, as will emerge later. -17today who would agree with Descartes that there is something about the nature of consciousness that resists complete explanation in terms of the quantitative language of physics; but today's prevailing view is that such conscious activity, at the very least, depends on certain physical processes--namely, brain processes, or perhaps (for example, in alien life forms) some other kind of physical processes capable of doing the same job. Descartes' standard view, by contrast, is that the mind is an immaterial substance--a separate entity in its own right, which 'does not require any place, or depend on any material thing, in order to exist' ( Discourse, part iv, A T VI33: C S M I127). Margaret Wilson (Chapter VII), in her careful analysis of the main argument Descartes uses to establish the distinctness of mind and body, begins by pointing out that in establishing that he is a 'thinking thing' (res cogitans) in the Second Meditation, Descartes does not claim to have justified the conclusion that he is essentially distinct from anything physical; he does, however, claim that he has a distinct conception of body as an extended thing, and that this conception is separate from that of thought ( Wilson, p. 188 below). This now provides the groundwork for the main argument for mind-body distinctness in Meditation Six (the 'epistemological argument', as Wilson calls it). The reasoning (to summarize) is that since I have a clear and distinct understanding that extension belongs to the essence of body, and thought to the essence of mind, and since I also have a clear and distinct conception of body as non-thinking, and of mind as nonextended, it follows that I can clearly and distinctly understand body apart from mind, and vice versa. And since whatever I clearly and distinctly understand can be brought about by God, it follows that God can bring it about that body and mind are apart (or separate); hence the two can indeed exist apart, and are really distinct (cf. A T VII78: C S M II54).

After mentioning, and refuting, three common but ineffective criticisms of this reasoning (p. 190 below), 22 Wilson proceeds to examine the objection first put forward by Antoine Arnauld (following a hint from Caterus): in perceiving himself as a thinking thing, might not Descartes have perceived only part of his nature; might not physical extension, unbeknownst to him, also be implicated in his essence? The best strategy for the Cartesian at this point (suggested by some remarks of Descartes himself, ____________________ 22 Perhaps the most important of these is the objection that all the argument shows is that mind and body could exist apart, not that they are distinct. But Wilson correctly points out that two things are distinct for Descartes merely if it is possible for them to exist separately. (It follows, incidentally, that those many modern critics who maintain that Cartesian mind-body dualism is false, while admitting that it is at least a logical possibility, are in fact conceding the core of Descartes' theory of the mind.) -18unpacked by Wilson) is to argue that, in conceiving of mind as a thinking thing, I have a clear and distinct recognition that thinking is by itself sufficient for me to exist: in conceiving myself as able to exist 'with thought alone' and no other attribute ( A T VII223: C S M II157), I have what Descartes calls a 'complete' conception of thinking substance. And such complete knowledge, Wilson argues, is 'sufficient for the epistemological argument to go through'. This is not to say, of course, that Descartes' reasoning is free from all problems; but if Wilson is correct, it is a lot less suspect than has been supposed by those who have attempted to dismiss it out of hand. It is one of the paradoxes of his account of the relationship between body and mind that, having spent great energy devising arguments to establish the distinction between them, Descartes took almost equal pains to underline their intimate union: 'There is nothing my own nature teaches me more vividly than that I have a body, and that when I feel pain there is something wrong with the body, and that when I am hungry or thirsty the body needs food and drink.' Or again, 'I am not merely present in my body as a sailor in a ship, but am very closely joined and as it were intermingled with it, so that I and the body form a unit' ( A T VII80-1: C S M II56). In her essay, the great French Cartesian scholar Geneviève Rodis-Lewis (Chapter VIII) examines the status of this 'unit' in Descartes' philosophy. Descartes variously called it a 'single whole', a 'single person who has both mind and body together', an 'entity in its own right', and a 'true substantial union' ( A T IX64, III694, III460, III508: C S M II56, C S M K228, 200, 209). Philosophers in the Anglophone tradition have often tended to ignore, or swiftly dismiss, the Cartesian theory of the mind-body union, taking their cue from Descartes' own warning that there is something mind-boggling in conceiving two things at the same time as both distinct and united. 23 RodisLewis, however, by carefully setting the Cartesian theory in its proper historical context, shows that, in contrast to the 'angelism' that was a popular position of the time (regarding humans as spiritual beings temporarily trapped in animal bodies), Descartes was concerned to do justice to the 'exceptional' and 'unique' character of our human nature. The key to

____________________ 23 In the letter to Elizabeth of 28 June 1643, Descartes says he reckons the human mind is 'not capable of conceiving distinctly and at the same time the distinction between soul and body and their union; for to do this one must conceive of them as a single thing and at the same time to conceive of them as two things, and these two conceptions are mutually opposed' ( A T III693: C S M K227). The last phrase (in the original: ce qui se contrarie) should (pace C S M K) probably not be taken to imply that there is an absurdity or a contradiction in regarding body and mind as both distinct and united, but merely as pointing out that one cannot grasp both the distinction and the union in a single mental act. -19grasping the special character of the human mind-body union, which Descartes insisted was better experienced than intellectually analysed ( A T III692: C S M K 227), is to appreciate the significance of the sensations and passions, reducible neither to pure cogitationes nor to purely bodily events, but the result of the interaction between the two. Rodis-Lewis points out that the majority of Descartes' successors, including his socalled disciples, all showed a tendency to bypass the 'union', and replace genuine mind-body interaction with some kind of more abstract correlation or concomitance, thus missing the distinctive core of Cartesian 'anthropology', 24 which Descartes took to have a crucial importance for his science of morals. Descartes' schematic list of the modes of awareness that are unique to the mind-body union comprises: First, appetites like hunger and thirst; secondly, the emotions or passions of the mind which do not consist of thought alone, such as the emotions of anger, joy, sadness and love; and finally all the sensations such as those of pain, pleasure, light, colours, sounds, smells, tastes, heat, hardness and the other tactile qualities. ( Principles, part i, art. 48: A T VIIIA23: C S M I209, emphasis added) It is the second (italicized) category on this list that has special importance for Cartesian ethics. The passions, which 'do not consist of thought alone', originate in physiological processes; as a result of psycho-physical associative responses, some innate, others the result of conditioning in early childhood and subsequently formed habits, we find ourselves in the grip of emotional states which we often understand only imperfectly, and whose effect on us, and our tendency to act and react in certain ways, is often quite out of proportion to our reflective thoughts about what we should be pursuing or avoiding. The problem of the mastery of the passions was of course a very ancient one in moral philosophy, but what Descartes brings to it is a systematic exploration of the physiological aetiology of the various emotional responses and, in consequence, a new kind of recipe for dealing with this potentially most rewarding, but often most dangerous, part of our human nature.

In his discussion of Descartes' views on the passions, Stephen Gaukroger (Chapter IX) begins by pointing out that moral, medical, and what may be called 'psychotherapeutic' considerations are inextricably intertwined in renaissance and early modern ethics. Moreover, Descartes' approach to the role of the passions in human life has its roots in two ____________________ 24 'The study of the human being' (from the Greek anthropos). Though the term 'anthropology' does not make an explicit appearance in Rodis-Lewis's article, the book of collected essays of which it forms the second chapter is entitled L'Anthropologie cartésienne. -20ancient traditions in moral philosophy--the intellectualist tradition of the Stoics on the one hand (who construed the passions as false beliefs) and the Augustinian tradition on the other, emphasizing the crucial role of the will as the key to the moral status of our emotional responses. A third influential element, as Gaukroger shows, is the detailed analysis and classification of the passions offered by Aquinas, in the Summa theologiae. Gaukroger proceeds to explore the central features of Descartes' own account of the passions, which invokes the well-known Cartesian identification of the pineal gland in the brain as the organ of interaction between neural impulses (the 'animal spirits') on the one hand and the activities of the mind or soul on the other. Despite the 'ghost-in-themachine' caricature of Descartes' position, with the animal spirits pushing into the pineal gland from one side and the 'volitions' of the soul trying to counteract them on the other (see Spinoza Ethics, part v, preface), Gaukroger rightly points out that although 'the will cannot halt this [the motion of the animal spirits and the resulting stimulation of a desire] directly', nevertheless, 'it can represent objects to itself so vividly that, by the principle of association, the course of the spirits will gradually be halted' (p. 221 below). Descartes' emphasis, like Aristotle's, is on the importance of systematic training and habituation for the cultivation of virtue; but, unlike Aristotle, Descartes insists on treating the subject en physicien, which implies that we need a comprehensive understanding of the role of the body, as well as that of the mind, in our mental and moral life, and also of the interactions between the mental and physical that arise from their 'substantial union'. The direction and regulation, but not suppression, of the passions is for Descartes the key to living a good life; and 'good', Gaukroger aptly concludes, involves both 'being ethical' and also 'being fulfilling', 'for the two are inseparable in Descartes' account' (p. 223 ). One of the points about sensations and passions that Descartes had stressed as early as the Meditations is that they are, in general, of signal benefit to the preservation of life: there is an obvious utility in the fact that we feel thirsty when the body needs fluid. One might have supposed that exactly the same is true of cats and dogs, but Descartes is generally held to have denied that non-human animals have any sensation whatsoever: this is the notorious Cartesian doctrine of the 'bête-machine'. Descartes certainly asserted that animals are 'automata', and that the animal (and indeed the human) body may be regarded

as a 'machine made by the hand of God'); but in my essay on Descartes' treatment of animals (Chapter X) I point out that, properly construed, neither of these notions logically -21implies that animals lack feelings. What Descartes does insist on (in his famous discussion of animals in part v of the Discourse on the Method) is, first, that animal utterances do not amount to genuine language, but are instead a set of programmed responses to stimuli; and secondly, that all the movements and behaviour exhibited by animals can be explained micromechanically, by reference to the movements and shapes of the particles out of which the various organs are composed. Neither of these claims in itself says anything about the truth or falsity of statements like 'Fido is frightened' or 'Felix is hungry'. There are several passages, moreover, where Descartes' argument about animals lacking genuine language is linked to such claims as that all their utterances are 'merely expressions of their fear, their hope, or their joy', so that 'they can do these things without any thought' (emphasis supplied; see letter to Newcastle, quoted below, pp. 230 -1). This appears not only not to deny, but actually to assert, the truth of statements like 'Fido is frightened'; and in general it would seem that Descartes wishes to explain such truths micro-mechanically, not to deny them (cf. A T VII230: C S M II161). He is, in other words, a reductionist about such phenomena, not an eliminativist. But what about the following 'knock-down' argument: 'the passions arise (as seen in the previous section) from the "union" of body and soul; animals (on Descartes' own insistently reiterated view) lack souls; therefore animals cannot have passions.' Well, that is no doubt what Descartes ought to have said; but what he in fact said was, for example, that 'if you train a magpie to say "good day" to its mistress when it sees her coming, [this will not be genuine language because] all you can have done is to make this word the expression of one of its passions' ( A T IV574: C S M K 303, emphasis added). What may be called the 'orthodox' interpretation of Descartes' conception of animals regards passages like this as unfortunate slips on Descartes' part; on this view, all he really means to attribute to the bird is 'purely physiological' happenings, not the 'actual consciousness' of pain, hunger, or whatever. This kind of response reveals, I think, the extent to which many people have become more 'Cartesian' than Descartes himself: believing that pain is defined by some inner conscious occurrence, accessible only to the subject, they take this characteristic inner occurrence to be always present in those who are 'really' in pain, and proceed to take Descartes to task for denying it occurs in animals. 25 I believe there are ____________________ 25 In a letter to Mersenne of 11 June 1640, Descartes observes, 'je n'explique pas sans âme le sentiment de la douleur' (I do not explain the sentiment of pain without reference to the soul) ( A T III85: C S M K148), which might seem to support the view that he denies that animals can 'really' be in pain. However, the remarks that follow make it clear that sentiment here does not,

-22more confusions in this line of thought than can be comfortably exposed in an introductory essay; but for the present purpose I will simply observe that in talking about the distinction between humans and animals Descartes never uses the term 'consciousness' (indeed, the term rarely appears anywhere in his writings); and he certainly never uses phrases like 'the actual consciousness of hunger'. What he denies to animals is thought and language; what he accepts (and scientifically explains) is the truth of statements like 'Fido is angry', 'Felix is hungry', 'Birdie expects a titbit'. If there is something more that needs to be said about animals, which Descartes' theory denies or does not allow for, then the onus is on his critics to explain, in a philosophically coherent way, what it is. V Despite the pervasive and (for all the recent efforts of philosophers) still continuing influence of Cartesian ways of thought on our modern culture, Descartes' scientific views (in the narrow sense)--his theories of astronomy, physiology, and so on--are now widely regarded as of purely historical interest. Descartes was right to agree with Harvey about the circulation of the blood, but spectacularly wrong to insist that the blood was forced along the arteries during the expansion (diastole) rather than the contraction (systole) phase of the heart ( Discourse, part v). He was right to agree with Galileo about the revolution of the planets, including the Earth, round the Sun, but quite mistaken in denying the possibility of a vacuum ( Principles, part ii, art. 16); and his notion that the planets were driven round by a kind of celestial whirlpool or 'vortex' ( Principles, part iii. art. 30) was soon to be swept away by Newton's theory of universal gravitation. In general, much of Cartesian physics now appears naïve and simplistic--for example, in its lack of the concepts of mass and density, and its attempt to make do with the 'purely geometrical' notion of volume or 'quantity of matter' defined entirely in spatial terms (cf. Principles, part ii, arts. 6-11). There are, nevertheless, great philosophical riches to be found in Descartes' general theory of scientific inquiry, especially once he is acquitted of the long-standing but unfair charge of being a pure 'apriorist', attempting to spin out the whole of physics from the armchair without any reference to experience. The starting-point of Daniel Garber's essay on ____________________ as might at first appear, refer to an 'inner qualitative sensation', but rather to something in the intellect (l'entendement)--in other words, a conceptual or cognitive episode; it is this that is denied to creatures without souls. -23' Descartes' method and the role of experiment' (Chapter XI) is the growing recognition that in many areas of his science Descartes was a 'serious experimenter'. But if

experiment and observation are essential to the Cartesian scientific enterprise, how does this square with Descartes' standard account of knowledge as derived from intellectual intuition and subsequent deduction? Garber points out that in the Regulae, where Descartes first presents his intuition-deduction model of knowledge, much attention is devoted to outlining a 'workable procedure for discovering an appropriate intuition'; such procedures involve the reduction of complex questions to more and more basic questions (p. 239 ). But how exactly does experiment fit in here? Examining Descartes' later work on the rainbow (in the Optics), Garber argues that experiment helps the Cartesian scientist perform the reductive part of the method, by suggesting the sequence of steps that leads from a question to an intuition: 'It may not be obvious to us at first just how we can go deductively from the nature of light to the rainbow, but poking about with water droplets, flasks, and prisms may suggest a path our deduction might follow' (p. 251 ). This important auxiliary and preparatory role for observation may seem to clash with Descartes' reiterated complaints (especially in the Meditations) about the reliability of observation and the senses in general. Garber reminds us that the 'hyperbolic' rejection of the senses is considerably tempered by the end of the Meditations: certain spontaneous sense-based beliefs (e.g. in the existence of physical objects) are, for Descartes, acceptable, provided they are corroborated by careful rational scrutiny--'they may be trusted to some extent and in some circumstances, but only after they have been given a clean bill of health by reason' (p. 255 ). The overall conclusion is that reason and experience are both important for Descartes, though in different ways--the latter in preparing the ground for the acquisition of knowledge, the former in allowing the relevant phenomena and causal dependencies to enter the body of genuine science (cf. pp. 256 ff). This complex interdependence reveals just how misleading it can be to classify Descartes' scientific thought in terms of the traditional contrast between 'rationalists' and 'empiricists'. That Descartes can be 'excused of the most characteristic theses of the paradigm rationalist' (p. 259 below) is the starting-point of Desmond Clarke's study of the Cartesian concept of scientific explanation (Chapter XII). One of the key points to emerge is that 'there is an implicit but significant complicity on Descartes' part with a kind of crude empiricism which avoids the theoretical in favour of scaled-down pictures of the objects of our ordinary perception' (p. 269 ). This is not to deny that in -24formulating the general laws of his physics (e.g. in the 'rules of impact' set out in Principles, part ii, arts. 45-52), Descartes adhered to his programme for the 'mathematicization' of physics, invoking only the quantifiable functions of size, shape, and motion. Nevertheless, Descartes recognizes, according to Clarke, 'that he has no way of dealing with the multiplicity of factors which determine the results of collisions between moving particles'; and '[i]n default of such a mathematically exact description of the parts of matter in motion, [he] falls back on ordinary experience as a source of explanatory concepts'. Here the use by Descartes of models, or 'analogies', drawn from our ordinary experience is crucial: wine-vats, tennis-balls, walking-sticks, whirlpools,

screws, and threads, and a host of others. These 'scaled-up versions of [supposed] microevents', as Clarke aptly terms them, play a fundamental role in Descartes' explanatory strategy: 'it would hardly be an exaggeration to say that Descartes' whole scientific project is one of imaginatively constructing descriptions of the motions of particles which might explain natural phenomena and our experience of them' (p. 275 ). Although the charge of a certain crudeness in Descartes' procedures may be justified, the fundamental issue he is grappling with here is the need to build a bridge, as it were, between the abstract and very general principles of mathematical physics and a satisfying set of explanations of the actual mechanisms operating in the real world. And the cluster of problems arising here is still very much a subject of debate among contemporary philosophers of science. The final essay in this collection examines one of the most fascinating and problematic concepts in the philosophy of science, that of force--a notion which preoccupied some of the greatest philosophers of the early modern period, notably Malebranche, Leibniz, and Hume, but whose status in Cartesian physics is often difficult to pin down. Some of Descartes' phrases suggest that matter possesses causal powers--for example, to transmit or to resist motion (compare the talk of particles 'imparting' and 'transferring' motion in Principles, part ii, art. 40); but the impression from other passages is that matter is wholly inert and that all causal agency must be attributed to God (cf. art. 36). In his essay, Gary Hatfield (Chapter XIII) favours the latter interpretation, focusing particularly on Descartes' 'laws of nature', which anticipate what we now call the principle of inertia--the tendency of a moving body to continue in motion in a rectilinear direction unless acted upon. Though Descartes sometimes talks of a body's possessing 'its own force of going on in a straight line', Hatfield argues that 'in explaining its tendency to persevere, Descartes does not appeal to the force that a body has by virtue of its motion'. The relation, indeed, is just -25the reverse. For Descartes defines 'the force of a body to act' as 'simply the tendency of everything to persist in its present state as laid down by the first law of nature' ( Principles, part ii, art. 43); and the reference to the first law makes it clear, as Hatfield points out, that the 'tendency' in question 'does not follow from any property of matter, but rather from an attribute of God' (p. 296 ). Hatfield's conclusion is that ' Descartes was constrained by his geometrical treatment of matter in motion to exclude all causal agency from matter itself . . . Motion is fundamental to Descartes' system of nature, but it is not by itself causally fundamental. God is the cause of motion' (p. 305 ). This may, to the modern ear, make Descartes' physics sound antique or quaint, like the 'occasionalism' developed later in the century by the so-called disciples of Descartes who regarded the impact between bodies as merely the 'occasion' for God, the sole true causal agent, to exercise his power. Hatfield certainly sees the occasionalists as direct inheritors, rather than distorters, of Descartes' outlook; and if the resulting picture of Cartesian physics is a profoundly metaphysical one, that is only appropriate: ' Descartes' physics is unintelligible without recourse to his

metaphysics' (p. 310 ). Though this is surely right, we may perhaps end by observing that there is a rather more 'modern' way of construing Descartes' banishing of inherent forces and powers from the purely natural world: namely, that he is moving towards the Humean insight that we have no real handle on what such 'forces' inhering in matter might be; all the natural scientist can do is to characterize the (mathematically describable) regularities that in fact obtain. For Descartes, of course, there is an ultimate cause of these regularities, namely God, whereas for Hume 'the ultimate springs and principles' of nature are 'totally shut up from human curiosity'. 26 Yet even this difference is not as striking as might at first appear, given Descartes' insistent declaration that 'God is a cause whose power exceeds the bounds of human understanding' ( A T I150: C S M K25). In this area, as in so many others, Descartes' reflections form a bridge between the medieval and the modern world-views. If there is a single impression to emerge from the essays collected here, it is a sense of the extraordinary scope of Descartes' thought and the diverse nature of his appeal to a wide variety of present-day philosophical interests. The time is long gone when philosophical discussion of Descartes' work was confined to the kinds of fanciful question that get philosophy a ____________________ 26 David Hume, An Enquiry Concerning Human Understanding [ 1748], sect. IV, part 1, ed. Selby-Bigge, 30. -26bad name, such as 'How do I know I am really awake?' The vigour of the continuing debates over Cartesian themes, in metaphysics, epistemology, the philosophy of mind, psychology and ethics, and the philosophy of science, testifies to the enduring power of Descartes' legacy. Though modern philosophy is no longer in thrall to the Cartesian outlook, there are still many central areas of the subject where Descartes' ideas exert their fascination, and continue to form a distinctive point of departure. It is likely to remain so for a long time to come. -27I DESCARTES AND THE METAPHYSICS OF DOUBT MICHAEL WILLIAMS I Descartes' Meditations 1 recounts a journey from pre-philosophical common sense to metaphysical enlightenment, each step of which is taken in response to an encounter with scepticism. The first-encounter induces a provisional doubt, clearing Descartes' mind of prejudices that would blind him to new truths. This allows subsequent encounters to lead

him to new certainties, by which the provisional doubt is gradually overcome, though not of course in a way that simply returns him to his initial position. According to this story, the problem of scepticism emerges within the framework of common sense, to be solved within that of Cartesian metaphysics. The First Meditation doubts must thus be 'natural', in the sense of metaphysically non-committal, or at least non-controversial. They must be developed out of resources that common sense can recognize as its own and must not depend, either for their intelligibility or power to command attention, on prior acceptance of controversial metaphysical ideas, particularly those which seemingly emerge in response to them. It is essential to Descartes' project that the story be told this way. The promise of a response to sceptical problems will not be a point in favour of a system of philosophical ideas if those ideas help generate the problems in, the first place, especially since the problems, once grasped, tend to be more compelling than any particular solutions. But, for Descartes, there is also a question of method. If, as is surely the case, the Meditations is meant to exemplify inquiry conducted according to Cartesian precepts, its argument ____________________ Michael Williams, "'Descartes and the Metaphysics of Doubt'" from Amelie O Rorty (ed.): Essays on Descartes' Meditations, 117-39. Copyright © 1986 by The Regents of the University of California, by permission of the author and of the publisher, the University of California Press. 1 In The Philosophical Works of Descartes: subsequent references to this work are cited as HR. -28must follow the order of reasons, at each stage depending only on considerations that have been explicitly avowed and validated. This means that the metaphysics is precluded from playing a covert role in generating the initial sceptical problems. I think, however, that Descartes' doubts are much less natural, much less metaphysically non-committal, than they are made to seem. In this essay, I want to explain why this is so, and how the illusion of naturalness is created. We are so familiar with Cartesian scepticism that its unnaturalness may not always be easy to see. We need something to compare it with. Classical Greek scepticism comes to mind. Descartes represents, and perhaps understood, his own sceptical arguments as deriving from the classical tradition, implying that he had discovered new solutions to commonly recognized problems. And it is true that, with the exception of the evil deceiver, the materials of the First Meditation are all traditional. 2 Descartes nevertheless understates his own originality. Myles Burnyeat has argued, convincingly, that the central Cartesian problem, the problem of the external world, was never posed by the ancients. 3 Somehow,

in Descartes' writings, traditional sceptical topics, such as madness and dreaming, take on a new significance. Burnyeat thinks that Descartes saw that the traditional materials sustain a more radical doubt than the classical sceptics realized. He was able to do so because, for him, the significance of scepticism is exclusively methodological. Scepticism is encountered in the course of a project of pure, theoretical inquiry, for the purposes of which all practical constraints are set aside. 4 By contrast, the classical sceptics saw scepticism as a path to happiness, a way of living in the world, whose existence can therefore hardly be the subject of sceptical doubt. There is something to this. Descartes does insist on the theoretical and methodological character of his doubts. Part of the significance of meditating is that doing so involves withdrawal from the world. But Descartes does not extend traditional scepticism; he transforms it. Cartesian and classical scepticism are radically different in kind, which is why it is ____________________ 2 The deceiver is the only feature of his development of scepticism for which Descartes claims originality [ A T V147: CSMK 333]. See J. G. Cottingham (ed.), Descartes' Conversation with Burman, 4, subsequent references to this work are cited as CB, followed by page number. 3 Myles Burnyeat, "'Idealism and Greek Philosophy: What Descartes Saw and Berkeley Missed'", Philosophical Review, 91 ( Jan. 1982), 3-40. For making me think about the novelty of Descartes' scepticism, I am grateful to Avner Cohen. 4 In emphasizing this aspect of Descartes' project, Burnyeat is influenced by Bernard Williams. See his Descartes: The Project of Pure Enquiry ( Harmondsworth: Penguin, 1978), esp. ch. 2. -29illuminating to compare them. Doing so will provide a way of bringing out the distinctive commitments of Cartesian scepticism, the metaphysics of Cartesian doubt. II In the First Meditation, Descartes induces an intense, if provisional, doubt by reflecting on a carefully arranged sequence of epistemological problems. Perhaps by now his procedure seems too familiar to deserve comment. Yet, for the classical sceptics, neither this connection with doubt nor this exclusive dependence on epistemological arguments was a feature of scepticism. As Sextus tells it, the sceptic begins as a doubter, but does not end as one. 5 The sceptic's capacity for suspension of judgement (epochē) is his remedy for doubt, not a means of inducing it.

The sceptic-to-be is a 'talented' individual who finds his peace of mind, which he prizes more than anything, disturbed by his awareness of various problems and puzzles. His response is to look into matters for himself, so as to determine the truth. This shows, I think, that his initial state can quite properly be described as one of doubt or uncertainty. Doubt is the state in which we want to know the truth but cannot decide where it lies; and, judging by his reaction to his disquiet, this is just the state the scepticto-be starts out in. His inquiries, however, produce an unexpected result. Instead of resolving his uncertainties, they produce a lively sense that no disputes are ever resolvable, that it is possible to argue with roughly equal plausibility (isosthenia) for and against anything, and thus they lead him to make a practice of suspending judgement. To his surprise, this is the point at which the sceptic achieves the tranquillity (ataraxia) he has sought all along. 6 Why doesn't his failure to bring his inquiries to a definite conclusion intensify his disquiet? The answer can only be that acquiring the distinctive sceptical capacities for isosthenia and epochē attenuates, or even eliminates, the urge to really know how things are. This is confirmed by the way the accomplished sceptic lives. In large measure, he follows what Descartes calls 'the teachings of nature', to distinguish them from the results of ____________________ 5 Sextus Empiricus, The Outlines of Pyrrhonism, Bk I, no. 12. Subsequent references are cited as OP, followed by book and marginal number. The edition I have used appears as Sextus Empiricus, vol. 1, with translation by R. G. Bury, Loeb Classical Library ( London: Heinemann; Cambridge, Mass.: Harvard University Press, 1933; repr. 1967). 6 OP 1, 25-9. -30informed rational judgement. Suspending judgement does not lead, as Hume notoriously supposes, to slow suicide. Rather, having got beyond the need to justify and explain, the sceptic lives by his spontaneous observations and impulses, the customs he has grown up with or come to feel comfortable with, and the practices of his trade or profession. He lives by how things strike him, acknowledging his opinions as just his opinions. He calls this living by 'appearances'. 7 Obviously, nothing could be more distant from Descartes' outlook. Descartes starts out in something like the classical sceptic's position: dissatisfied with the current state of his knowledge and hoping to do better. But whereas scepticism creeps up on the classical sceptic, as the result of his persistent inability to bring disputes to a definite conclusion, scepticism for Descartes is a theoretical problem that he chooses to confront. From the outset, scepticism is under his control: it is not, as it is for the classical sceptic, something with a life of its own, which might affect his motives in unforeseen and, perhaps initially, even unwelcome ways. Since Descartes' confrontation with scepticism takes place in a context where his fundamental motive, the urge to really know, is unaffected by sceptical reflections, a context where even the teachings of nature are minimally insistent, such

reflections serve to intensify the initial uncertainty rather than eliminate it. Scepticism becomes wedded to doubt. But for the purposes of this essay, the differences between the ways Descartes and the classical sceptics reach their respective sceptical states are even more significant than the differences between the states themselves. Though classical suspension of judgement is not doubt, it is reached by means that involve an extension of ordinary, non-philosophical doubting, the sort of doubting that comes as a result of exposure to conflicting opinions, each plausible in its way. The classical sceptic achieves epochē by mastering the art of meeting any given argument or thesis with a countervailing argument or thesis of equal plausibility. Since ethical arguments and theses will be met by countervailing ethical arguments and theses, physical by physical, logico-philosophical by logico-philosophical, no special importance attaches to arguments derived from theoretical considerations in epistemoloay. This explains the extreme discursiveness we find in Sextus (and in Montaigne, the true heir to the classical tradition). The accomplished sceptic has to master a potentially endless number ____________________ 7 OP 1, 23-4. Such 'appearances' are evidently not sense-data. For an interesting discussion of the Pyrrhonian notion of 'appearance', see Myles Burnyeat, "'Can the Sceptic Live his Scepticism?'", in Burnyeat (ed.). The Skeptical Tradition ( Berkeley, Los Angeles, and London: University of California Press, 1983), 117-48. -31of conflicting arguments, not just a short list of epistemological problems: he cultivates a kind of learned ignorance. If there is a problem here, it is not yet that of solving a canonical list of sceptical problems, derived from theoretical considerations in epistemology, though this is what the problem of scepticism became after Descartes. The classical sceptics are not, of course, innocent of epistemological arguments. But the arguments they use, and the way they use them, grow out of the practice of isosthenia, which remains the basis of classical scepticism. These arguments are summarized by the famous 'modes'. 8 To make sense of the Ten Modes, the first set Sextus presents, we should begin by separating out the ninth and tenth. The tenth simply illustrates how sceptics practise isosthenia in ethical disputes, whereas the ninth makes the point that whether an object is counted amazing or precious seems to depend more on its scarcity than on its intrinsic character. With these modes set aside, it becomes apparent that those remaining exemplify a single pattern of argument: all are versions of the so-called problem of the criterion. The first step of this argument is given in general form by the eighth mode: how things are judged to be is always pros ti, relative to something or other: the person judging, the

conditions under which the judgement is made, and so on. Since, the argument continues, we inevitably face conflicting opinions, we can avoid suspension of judgement only if we have a criterion by which to determine where the truth lies. To take an example from conflict among the senses, if sight tells us that an object is smooth and touch that it is rough, shall we believe sight or touch? The sceptic finds that attempts to answer such questions invariably fail. If we make one sense the judge of truth, we simply beg the question against the other and, in any case, reason in a circle. If we propose an independent criterion, our hypothesis will be met by a conflicting hypothesis, yielding a new dispute to be adjudicated and threatening us with either a circle or a regress. Whatever we try, all roads lead to epochē. The Five Modes, given as discrepancy, regress, relativity, hypothesis, and circular reasoning, present in a more abstract form the strategy illustrated diversely by the Ten. Discrepancy and/or relativity suggest the need for a criterion; but any attempt to establish one will lead either to brute assertion (hypothesis), circular reasoning, or an infinite regress, each of which in one way or another leads to epochē. The Two Modes are even more abstract, presenting an entirely general ____________________ 8 OP 1, 31 ff. -32form of the regress problem. Any attempt to justify anything can be made to lead either to circular reasoning or an infinite regress. This suggests that the problem of the criterion can be regarded as a special case of the regress problem: that problem as it arises in connection with epistemological standards. The important point is that the modes are not presented as theoretical arguments for the impossibility of knowledge. They are offered as techniques for heading off attempts to escape the effects of isosthenia by establishing a criterion for resolving disputes. The context in which to apply them is always provided by discrepancy and conflict between opinions: if there were no disagreements, there would be no occasions to apply the modes. The marshalling of conflicting opinions, therefore, remains the basis of classical scepticism. This extension of ordinary doubting needs no basis in epistemological theory. Nor does use of the modes, which are deployed defensively against anti-sceptics who themselves admit the need for a criterion. The sceptic himself need have no views one way or another. In contrast, Cartesian scepticism is the outcome of exclusively epistemological arguments. True, there are echoes of isosthenia in the Meditations, particularly in the discussions of God and the deceiver, where Descartes propounds competing accounts of his origin and relation to Providence which he cannot yet rule out, and some of which call in question the veracity of his previous beliefs. By balancing these disturbing possibilities against his deep-rooted but unexamined convictions, he achieves a general doubt. But this

is not the fully-fledged classical technique. If Descartes turns to theological speculation, it is not, as it would be for Sextus, in order to suspend judgement on religious matters, but to raise the question of whether our beliefs arise in a way that makes them likely to be true. The theological trappings derive their importance from their epistemological implications. Epistemological argument remains the basis of Cartesian scepticism. We see, then, that Descartes does not simply put the resources of traditional scepticism to new uses. He reduces and redirects them. So our first question is: why? But we have to ask not only why Descartes makes scepticism a matter of purely epistemological argument, but also whether the epistemological considerations he exploits are the same as those exploited by the classical tradition? There are reasons for suspecting that they cannot be. In its limited use of epistemological arguments, classical scepticism makes exclusive use of what I shall call the 'formal' problems of the regress and the criterion. They are formal in the sense that they represent -33questions that can be asked of any anti-sceptic who thinks that his beliefs can be justified. They depend on no substantive theories of justification: for example, theories about what kinds of beliefs can legitimately be cited in justification of what other kinds. And they are always fully general; they are applied indifferently to beliefs of all kinds and all subjectmatters. As a result, classical scepticism is 'flat'. It never communicates a sense that some beliefs may be more problematic than others. In contrast, Cartesian doubt is, intentionally, differentiated. Descartes wants to show that we can, in a provisional way, entertain doubts about anything, but especially about material things. 9 It is also stratified: the sequence of arguments given in the First Meditation conveys a strong sense of probing ever more deeply into the base of our convictions. We need to investigate how this sense of differentiation and stratification is inculcated, and what epistemological considerations a stratified doubt implies. III Like Descartes, Bacon believes that the reconstruction of scientific knowledge must be preceded by a provisional doubt. But whereas Descartes favours the use of very general epistemological problems to empty our minds, Bacon recommends that we counteract the force of received opinion by filling our minds with all the new facts and theoretical possibilities we can find. 10 This straightforward adaptation of classical scepticism shows that putting scepticism to a methodological use does not by itself lead to the reduction of scepticism to a purely epistemological problem. But there are several reasons why the classical technique of isosthenia would not serve Descartes' purposes.

In a letter to Mersenne of 1641 Descartes writes: . . . I may tell you between ourselves that these six Meditations contain all the foundations of my Physics. But please do not tell people, for it might make it harder for supporters of Aristotle to approve them. I hope that readers will gradually get used to my principles, and recognise their truth, before they notice they destroy those of Aristotle. 11 The classical sceptic would induce epochē by balancing the new principles against the old. But in Descartes' view the new and the old never balance. To the prejudiced mind, under the spell of the senses, the new principles will seem absurd; to the mind properly prepared, they will be compelling. ____________________ 10 Bacon, Novum Organum, Bk 2, aphorism 32; in J. Spedding and R. E. Ellis (eds.), Philosophical Works ( London: Longmans, 1879), iv. 11 A. Kenny (ed.), Descartes: Philosophical Letters, 94 [ A T III298: C S M K173]. 9 See the opening of the Synopsis of the Meditations ( H R I140 [ A T VII12: C S M II9]). -34To get the new principles a fair hearing, then, will require tact and indirection. Too early and too definite an identification of the substantive questions at issue must be avoided. Accordingly, the provisional doubt is made to turn on general epistemological considerations alone. But perhaps even more significant is the fact that Cartesian doubt necessarily pretends to a kind of systematicity and closure that has no place in classical epochē. The provisional doubt of the First Meditation is to be resolved by finding new principles, which will modify our understanding of whatever previous beliefs are eventually reinstated. Descartes' original opinions, as originally understood, have been definitively undermined, made dubious beyond recall. In contrast, classical scepticism is open-ended, and represents a continuing engagement, rather than a completable project. It is comprehensive, in that the sceptic will try to deal with all theses and arguments that come his way. But it is not closed: the classical sceptic is always ready to deal with new claims or attempts to revive old ones. He never claims to have dealt definitively and in advance with everything a particular kind of dogmatist might say, for this itself would be a kind of dogmatism. Descartes' definitive doubt is the mirror image of his definitive certainty. Having raised, as he claims, 12 all possible doubts, he will be able subsequently to claim that whatever principles survive his sceptical scrutiny have been established with metaphysical finality. Classical scepticism, even if used as a methodological device, could support no such claim. With the emergence of a new idea, the balance, even if at present dramatically tipped, might always be restored, or even tipped the other way. Ordinary doubting, and its

sophisticated extension, classical isosthenia, are always contingent on the current state of knowledge. They offer no test for absolute certainty. To make plausible the idea of a definitive sceptical doubt, hence of definitive metaphysical certainty, Descartes must represent the credibility of his former opinions as resting ultimately on certain highly general epistemological principles that can be exhaustively examined for possible sources of error. But Descartes does not acknowledge the theoretical considerations that mandate this reduction of scepticism to a theoretical problem in epistemology. Rather, he represents the reduction as a condition of making his project practicable. To effect a 'general upheaval' of his former opinions, he writes: . . . it will not be requisite that I examine each in particular, which would be an endless undertaking, for owing to the fact that the destruction of the foundations of ____________________ 12 C B 4 [ A T V 147: C S M K 333]. -35necessity brings with it the downfall of the rest of the edifice, I shall only in the first place attack those principles upon which all my former opinions rested. 13 This is misleading. It is not simply 'not requisite' to examine each opinion on its specific merits; it is requisite not to. The epistemological systematicity implied by the metaphor of 'foundations' allows beliefs to be undermined in a 'topic-neutral' way: that is, independently of their particular content, the current state of collateral knowledge, and all concrete problem situations and contexts of inquiry in which they might be examined. But the possibility of this kind of doubt is a requirement, not a discovery. There is no argument anywhere in Descartes' Meditations to show that common sense recognizes the conception of justification, embodied in the metaphor of foundations, on which Descartes' definitive doubt depends. For all he has shown, common-sense doubting, hence common-sense certainty, may always be to some degree context-bound. Classical scepticism, which is in a way 'an endless undertaking', allows for this possibility. But Descartes excludes it by projecting into common sense a theoretical conception of justification disguised as a matter of practical convenience. What makes this projection seem more innocent than it is, is Descartes' extreme vagueness as to the nature of the 'principles' on which his former opinions rested. All he tells us is that these opinions were derived by way of the senses. Though there may be some sense in which this is uncontroversially true, to generate the sceptical doubt he requires, Descartes will have to interpret it in a way that is far from uncontroversial. But before we can look into how the required interpretation is introduced, and how its

theoretical character is disguised, we must deal with the question of how Cartesian doubt, unlike classical epochē, conveys the sense of stratification we noted earlier. IV In the first four of the Ten Modes, Sextus presents and rebuts a sequence of progressively more specific responses to the problem of the criterion. The First Mode questions the least common denominator of dogmatic solutions to the problem, the assumption that men are the judges of truth. Why not other animals, who in some respects seem to have the advantage over us? The sequence of questions then continues: if men, which particular man or men, given that men disagree? If some particular man or men, ____________________ 13 HR I 145 [ A T VII18: C S M II12]. -36by what sense or faculty, given that various senses and faculties often suggest different judgements? And if by some particular sense or faculty, under what conditions or in what circumstances, given that what is thought to be true is influenced by conditions and circumstances? At each 'stage' in the sequence the same argument is presented, the argument we are already familiar with. The first step is invariably to claim that the proposed criterion either itself produces discrepant opinions or is in conflict with other prima-facie equally plausible criteria. Subsequent steps argue that any attempt to remedy the situation will lead to brute assertion, circular reasoning, or an infinite regress. Each failed attempt to resolve the problem of the criterion leads to a comprehensive suspension of judgement. Now contrast this with Descartes' procedure in the First Meditation. The crucial feature of the way Descartes develops his doubt is that it is not complete at every stage. Rather, at each stage some beliefs are called in question, but others survive, pending the deployment of more powerful sceptical considerations. The first point Descartes makes is that he cannot trust his senses without qualification, because they have often deceived him about objects that are barely perceptible or very far away. But this leaves untouched beliefs about objects close by and in plain view. To call these in question, he needs the dreaming argument. But even the dreaming argument, as Descartes understands it, leaves unscathed beliefs about things that are 'very simple and very general', and to undermine the credibility of these, he has to raise questions about his origin, nature, and relation to Providence, a line of thought encapsulated in the conceit of the evil deceiver. And even this final, 'hyperbolical' doubt seems implicitly to concede Descartes some knowledge. For he does not end the First Meditation in a state of complete confusion. He knows at least how things seem to him, even if as yet he has no idea how they really are.

This stratification of doubt imposes a corresponding stratification of knowledge. Through the progressive development of his doubt, Descartes effects a context- and subjectmatter-independent partitioning of his beliefs into broad epistemological classes, ordered according to how difficult it is to doubt them. First in the order come the beliefs that are never doubted, subsequently to be identified as those that involve Descartes' immediate knowledge of his own 'thoughts', whose exemption will be retrospectively justified on the grounds of their supposed incorrigibility. Moreover, the order of justification, which Descartes calls 'the order of reasons', must reverse the order of doubt. The progressive development of Cartesian doubt insinuates, without ever directly arguing for, a -37foundational conception of knowledge, the view of knowledge that sees justification as constrained by just the sort of context- and subject-matterindependent order of epistemic priority that is implicit in Descartes' stratified doubt. The scepticism that develops within a foundational conception of knowledge contrasts sharply with classical scepticism, even in its most purely epistemic form. Acceptance of the notion of a context-independent order of epistemological priority makes scepticism turn on problems about inference: the sceptic challenges us to show how beliefs of a given kind can be justified on the basis of beliefs that, in virtue of their place in the epistemic hierarchy, constitute the proper evidence for them. The classical sceptic's undifferentiated epochē thus gives way to a form of scepticism organized around specific sceptical problems: the external world, induction, other minds, and so forth. This transformation of scepticism is the legacy of Descartes' Meditations. It is not immediately apparent that the First Meditation will issue in a radical transformation of scepticism. At the first stage of the doubt, Descartes could be taken to be following Sextus, proposing a rather vague criterion of truth (the senses) and, having found it wanting, replacing it with something more refined (the senses operating under ideal conditions). 14 But with the argument that begins by mentioning madness and modulates into a discussion of dreaming, his divergence from the classical tradition becomes increasingly evident. Whereas dreaming is of crucial significance for Descartes, Sextus mentions it only incidentally as one of the 'conditions or dispositions' that influence judgement. (The others include madness, old age, sickness.) Here is his discussion in full: Sleeping and waking, too, give rise to different impressions, since we do not imagine when awake what we imagine in sleep, nor in sleep what we imagine when awake; so that the existence of our impressions is not absolute but relative, being in relation to our sleeping or waking condition. Probably, then, in dreams we see things which to our waking state are unreal, although not wholly unreal; for they exist in our dreams, just as waking realities exist, although nonexistent in dreams. 15

Notice the emphasis on the difference in content between waking experience and dreams. Because of this difference, dreaming offers one more way into the familiar problem of the criterion, which is why Sextus attaches no special significance to it. Various textual echoes suggest the second book of Cicero Academica ____________________ 14 Burnyeat interprets Descartes this way in 'Idealism and Greek Philosophy' (see pp. 32 ff.). A similar reading can be found in Frankfurt, Demons, Dreamers, and Madmen, ch. 5. 15 OP 1, 104. -38as the likely source for Descartes' discussion of dreaming. As in the First Meditation, dreaming is discussed in connection with the question of whether the senses are a source of certainty. Lucullus, the anti-sceptic in the dialogue, has claimed that the delusions of dreams, drunkards, and lunatics are feebler than, hence easily distinguished from, veridical sense experiences. Cicero replies that this need not be so, since men are sometimes fooled by dreams and hallucinations. Whatever makes normal experience convincing can be mimicked by dreams, which means that normal experience does not convince us in virtue of anything that guarantees its truthfulness. But, just as in Sextus, the argument stresses the discrepancy between the content of waking perception and that of dreams, the point being to ask why we take the one and not the other to be veridical. 16 Like Sextus and Cicero, Descartes begins by associating dreaming with madness and pointing up the discrepancy between waking perception and dreams. Thus: I am a man, and . . . consequently in the habit of sleeping, and in my dreams representing the same things, or sometimes even less probable things, than do those who are insane in their waking moments. 17 However, he promptly executes an abrupt volte-face, continuing: How often has it appeared to me that in the night I dreamt that I found myself lying in this particular place, that I was dressed and seated near the fire, while in reality I was lying undressed in bed? 18

Suddenly the emphasis is on the 'fact' that dreams can reproduce the content of the most commonplace waking experiences. This shift of emphasis allows Descartes to pose a question that never occurs to the classical sceptics: how can Descartesever be sure that he is not dreaming? Might not his entire conscious life be a kind of dream? Cicero evidently sees no such general problem, for he writes: As if anyone would deny that a man that has woken up thinks that he has been dreaming, or that one whose madness has subsided thinks that the things that he saw during his madness were not true. 19

____________________ 16 Cicero, Academica, in Cicero, vol. 19, with translation by H. Rackham, Loeb Classical Library ( London: Heinemann; Cambridge, Mass.: Harvard University Press, 1933; repr. 1979), Bk 2, sect. 27. Margaret Wilson interprets Descartes' dreaming argument in a way that has Descartes arguing along what may seem to be the same lines as Cicero, and noting Cicero as his probable source would strengthen her case. But though I think that Descartes had the Academica in mind when composing the Meditations, his version of the dreaming argument differs significantly from Cicero's, as I show in section 6, where I discuss Descartes' adaptation of Academic materials. For Wilson's interpretation, see her Descartes, ch. 1. 17 HR I145 [ AT VII19: CSM II13]. 18 HR I145-6 [ AT VII19: CSM II13]. 19 Cicero, Academica, 579. -39While dreaming, we may not realize it. But when we are awake, we know it. Descartes, however, though awake and in possession of his senses, is I almost capable of persuading' himself that he is dreaming. Not surprisingly, since the situation he recalls having dreamt himself to be in corresponds exactly to the situation he finds himself in at the moment. The shift of attention from the discrepancy between waking perception and dreaming to the possibility of their convergence also serves to detach dreaming from madness. This is essential to the way Descartes develops his doubt. Associating dreaming with drunkenness and lunacy suggests that the dream-state may be one of mental confusion. But Descartes holds that our capacity to judge, hence potentially to know, is unaffected by our being in a dream rather than awake. This allows him to continue his stratification of knowledge, arguing that there are still beliefs unscathed by the dreaming argument. For example: 'whether I am awake or asleep, two and three together always form five and the square can never have more than four sides'. 20 Dreaming and waking do not, therefore, involve radically different kinds of experiences. Rather, the difference between them must be entirely extrinsic, and Descartes hints more than once that it has to do with causation. As the Meditations progresses, 'dreaming' will be reinterpreted to mean 'experience, apparently of objects, but without the usual corresponding external causes', and eventually, 'experience without any physical causes whatsoever'. There are several steps to be taken before we confront the nightmare fantasy of a being sans corps, sans monde. Descartes takes the dreaming argument to leave open the possibility that our experiences, dreams though they may be, are caused by, and in some suitably abstract way correspond to, external physical objects. Only after he has questioned his relation to God, alternatively the deceiver, is he able to ask: 'how do I know that he has not brought it to pass that there is no earth, no heaven, no extended body, no magnitude, no place, and that nevertheless I possess the perceptions of all these things and that they seem to exist just exactly as I now see them?' 21 Nevertheless, the materials that make it possible to raise the general sceptical problem about the external world are substantially in place by the end of the dreaming argument. The problem

emerges as that of inferring, from what we know, whether we are awake or dreaming (i.e. how things seem), the existence and nature of the supposed external causes of our perceptions, which is why it is essential that the Cartesian dreamer retain intact his ability to think. ____________________ 20 HR I147 [ AT VII20: CSM II14]. 21 HR I147 [ AT VII21: CSM II14]. -40We see, then, that the progressive doubt of the Meditations is informed throughout by a foundational conception of knowledge. Descartes' distinctive sceptical problems, and the metaphysical framework that makes it possible to raise them, are introduced together. If they seem compelling, it is through the power of Descartes' narrative. There is no argument for accepting the foundational conception, even for the purposes of a philosophical reconstruction of knowledge, still less for projecting it into unreconstructed common sense. Cartesian scepticism is not, like classical scepticism, an extension of common-sense doubting. It is dependent from the outset on a metaphysic as essential as it is unavowed. It remains for us to explore the devices by which the theoretical character of the doubt is both clarified and disguised. V Gassendi dismissed the First Meditation as contrived and unnecessarily prolix. Only when he came to comment on the Third did it begin to dawn on him what Descartes had been up to: You seem afterwards to make it doubtful not only whether any ideas proceed from external things, but also whether there are any external objects at all. And you seem thence to infer that although there exist in you the ideas of things said to be external, those ideas nevertheless do not prove that the things exist, since they do not necessarily proceed from them, but may be due to yourself or to some other cause, I know not what. It was for this reason I fancy that you previously continued to say: That you had not previously perceived earth, sky, or stars, but the ideas of earth, sky and stars, which might possibly be a source of delusion. 22 Here we see Gassendi interpreting the First Meditation in the light of things that have become clear to him in the course of the Second and Third, though the objection he goes on to level might make us wonder whether, even at this point, he has really grasped the structure of Descartes' problem. 23 This belated and retrospective understanding of the First Meditation doubt is, nevertheless, entirely appropriate to the way the Meditations is written. The First

Meditation suggests far more than it says. But in the Third Meditation Descartes finally identifies both his former certainties and the 'principles' on which they rested: They were the earth, sky, stars, and all other objects which I apprehended by means of the senses. But what did I clearly and distinctly perceive in them? Nothing more ____________________ 22 HR II154-5 [ AT VII282: CSM II196-7]. 23 e.g. he asks why one born blind has no idea of colour. -41than that the ideas or thoughts of these things were presented to my mind. And not even now do I deny that these ideas are to be met with in me. But there was yet another thing which I affirmed, and which, owing to the habit which I had formed of believing it, I thought I perceived very clearly, though in truth I did not perceive it at all, to wit, that there were objects outside of me from which these ideas proceeded, and to which they were entirely similar. And it was in this that I erred. 24 This brings out very clearly what is involved in the sceptical problem about the external world. Our basic knowledge is of our own thoughts or ideas: the question is how to justify inferences from this basic knowledge to beliefs about the existence of the supposed external, physical causes of those ideas. What was only implicit in the First Meditation is finally brought into the open.This process of clarification and interpretation is resumed in the Sixth Meditation. Again, Descartes represents his beliefs about bodies and their qualities as depending on inferences from 'the ideas of all these qualities which presented themselves to my mind, and which alone I perceived properly or immediately'. 25 The grounds for such inferences, however, were the vividness of the ideas of sense and their independence of his will, reasons which seem wholly inadequate when he reminds himself that I have never believed myself to feel anything in waking moments which I cannot also sometimes believe myself to feel when I sleep, and as I do not think that these things which I seem to feel in sleep, proceed from objects outside of me, I do not see any reason why I should have this belief regarding objects which I seem to perceive while awake. 26 Again, the metaphysical presuppositions of the doubt are made clear in a way they never were when it was first developed. These are, to summarize: 1. The assumption that there is a universal, context- and subject-matterindependent order of reasons; that knowledge conforms to a foundational structure. 2. The identification of our 'thoughts', the contents of our 'minds', as all that we perceive 'properly or immediately'. 3. The assimilation of sensations to 'thoughts' as characterized above. This assimilation

allows Descartes to entertain the possibility that his 'sensations' might be just what they are even if he had no senses--if there were no physical world at all--while, by trading on the common-sense notion that the (physical) senses are what put us (embodied persons) in touch with external objects (i.e. objects outside our bodies), preserving ____________________ 24 HR I158 [ AT VII35: CSM II24]. 25 HR I187 [ AT VII75: CSM II52]. 26 HR I189 [ AT VII77: CSM II53]. -42the idea that 'sensations' are the sine qua non of knowledge of external reality. 'Sensations' thus become the ultimate, if prima-facie noncommittal, evidence on which inferences to beliefs about that reality must be based. 27 It is clear why Descartes cannot be open about these presuppositions from the start: not until later in the Meditations does he have the necessary theses and distinctions in hand. To take just one example, common sense does not recognize the Cartesian analysis of sensation into mental and physical components, if for no other reason than that it does not recognize the Cartesian conception of the mental. To make it appear as if the doubt about the external is generated out of resources that common sense does recognize as its own, Descartes must project back into his prephilosophical position metaphysical considerations that are properly available only after common sense has been left far behind. This projection is accomplished, and disguised, in the Third and Sixth Meditations, where, as in the passages just cited, he interprets the First Meditation doubts, which are deliberately left somewhat vague and indeterminate, under the guise of recalling them for the purpose of further examination. But this 'recollection' is only a pretence. 28 Consider again Descartes' question about the bodies he used to take himself to perceive: 'what did I clearly and distinctly perceive in them? Nothing more than that the ideas or thoughts of these things were presented to my ____________________ 27 On the novelty of Descartes' conception of sensation, see Wallace Matson, "'Why Isn't the Mind-Body Problem Ancient?'", in Paul Feyerabend and Grover Maxwell (eds.), Mind, Matter, and Method ( Minneapolis: University of Minnesota Press, 1966), 92102. Bernard Williams ( Descartes, 57 f.) sees an important difference between the doubt based on the dreaming argument, which allows us to question any given perceptual belief, and the hyperbolical doubt based on the deceiver, which calls in question all beliefs at once, thus raising the sceptical problem about the external world. Moreover, he notes that the latter doubt has built into it 'a causal model of perception'. The hyperbolical doubt thus poses a problem for Williams's attempt to defend the

reasonableness and metaphysically non-committal character of Cartesian doubt by portraying it as the natural outcome of transposing common-sense ideas about knowledge into the context of pure theoretical inquiry. Williams is inclined to cope with this difficulty by suggesting that the hyperbolical doubt is not strictly necessary for Descartes' methodological purposes. There are two things to say about this. First, a causal model of perception does not by itself create the possibility of the hyperbolical doubt: indeed, understood a certain way, it may even block the road to the sceptical problem of the external world, as my discussion of the Academics in section VI shows. Thus the metaphysical presuppositions of Cartesian doubt are much more extensive and have much less to do with common sense than Williams implies. They include at least epistemological foundationalism and a new conception of mind. Second, they cannot be brushed aside, for the specal dubitability of the very existence of the external world, hence the hyperbolical doubt, is essential to the purposes Descartes has in mind for his doubt. Who can doubt that the reconstruction of knowledge would proceed very differently if this special dubitability were never argued for? 28 Paul Feyerabend finds a similar rhetorical technique in Galileo. See Against Method ( London: New Left Books, 1975), ch. 7. -43mind?' 29 But the answer ought to be simply 'Nothing'. Descartes was in no position to perceive clearly and distinctly that the ideas of these things were presented to his mind because he had not, at that stage, thought his way through to a clear and distinct conception of either 'thought' or 'mind'. Descartes must pretend that he always in some way recognized that only his ideas were properly and immediately known, for only in that way can he suggest that his previous beliefs rested on a dubious inference from these basic data. But his question would be more properly put, not as what did he clearly and distinctly perceive when he was mired in common sense, but as what could he have come to perceive that way if he had had the relevant distinctions already in hand? Put that way, however, the question would subvert, rather than reinforce, the illusion that the doubt about the external world was generated out of the resources of common sense alone. This pretence of recollecting, while in fact interpreting, is aided considerably by Descartes' practice of playing back and forth between the common-sense understanding of 'external object', 'dream', and so on, and the senses that such terms take on in the light of his own metaphysical innovations. Thus consider the following passage, where Descartes is considering whether his ideas need have any 'external' cause, even a nonphysical one: 'perhaps there is in me some faculty fitted to produce these ideas without the existence of external things, even though it is not yet known by me; just as, apparently, they have hitherto always been found in me during sleep without the aid of external objects'. 30 In its first. occurrence, 'external' means outside the Cartesian mind; in the second, it bears the familiar meaning of spatially external to the body. In a similar way, dreams are sometimes experiences we have while asleep, a state in which an implication of embodiment exists: but when the evil deceiver is said to produce 'illusions

and dreams', dreams are the experiences of a bodiless, worldless self. Such equivocations allow Descartes to read metaphysical innovations back into his common-sense startingpoint without his appearing to do so. VI From the argument so far, one thing at least should be clear: that there is no such thing as the argument from dreams. So I am not claiming that any ____________________ 29 HR I158 [ AT VII35: CSM II24]. 30 HR I161 [ AT VII39: CSM II27]. -44sceptical argument which appeals to the fact that we dream is metaphysically loaded in just the way Descartes' argument is. Equally, it is no objection to claim that a version of the dreaming argument can be developed in a more metaphysically non-committal way. What would need to be shown is that a version could be developed that would lend itself to the same uses that Descartes puts his version to. We can clarify the metaphysically loaded character of the Cartesian argument from dreams by returning briefly to the ancients. Doing so will allow us to determine more precisely what ancient materials Descartes makes use of and how he transforms them. I argued above that, whereas Descartes appeals to the (alleged) fact that dreaming can mimic waking to the point of being indistinguishable from it, the ancients stress the discrepancy between what we experience in dreams and what we experience when awake. This point needs to be made with some care, however. It applies without qualification to the Pyrrhonians. But the Academics argue somewhat differently. Indeed, they argue in ways that, superficially, are more than a little reminiscent of Descartes. Pyrrhonian scepticism exploits the 'formal' epistemological problems of the criterion and the regress of justification. Central to Academic scepticism, however, is a quite different argument: that, since non-veridical perceptions can be identical to, or at least indistinguishable from, veridical perceptions, no perception guarantees its own veracity. Consequently, since all knowledge begins with perception, there is no guaranteeing the veracity of anything we accept: that is, nothing can be known. 31 When we add to this the fact that the Academics cite dreams and hallucinations-perceptions of objects that have no real existence--as examples of false perceptions that may be indistinguishable from true ones, it may seem that we must credit them with the very argument that Descartes uses to pose the problem of our knowledge of the external world. It may further seem that, if they did not pose this problem themselves, they must simply have failed to see the implications of their own arguments, as Burnyeat has claimed about the ancient sceptics in general.

However, on closer examination, a different picture emerges. We can find in the Academic sceptics the elements of Cartesian doubt, but not in ____________________ 31 Cicero, Academica, 519f., 573f. Sextus, Against the Logicians, I, 409f., in Sextus Empiricus, vol. 2. The first passage from Cicero, where Lucullus is expounding the Academic arguments, is especially striking for containing other ideas that will become important to Descartes: God's omnipotence and the mind's power to 'move itself' (p. 527). But these ideas are not put to Cartesian uses. Lucullus's replies to the arguments he has expounded show that his remarks about false perceptions being potentially indistinguishable from true show that they are to be interpreted along the lines I indicate in the text. -45their distinctively Cartesian alignment. Properly understood, the Academics confirm, rather than create difficulties for, the argument I have been developing. The first thing to notice is that the Academic arguments exploit two quite distinct ways in which true perceptions might be indistinguishable from false. The first, as in Descartes' version of the dreaming argument, has to do with content; or, more precisely, with the objects presented (visa). The second has to do, not with what is presented, but with the mode of presentation: in particular, with whether veridical perceptions have a distinctive persuasiveness, a feeling of reality or substantiality that nonveridical perceptions lack. Now, and this is the crucial point, perceptions that are indistinguishable in respect of content are always thought of as perceptions of (= caused by) indistinguishable objects: the standard examples are two eggs, and a man and his twin brother. As a result, this kind of indistinguishability is not what is at issue when the discussion turns to dreams and hallucinations, where there are no real objects causing the perceptions. When these false perceptions are brought up, it is always to make the argument that, since the dreamer can be taken in by his dreams, and the madman by his delusions, whatever makes true perceptions persuasive may be equally characteristic of false ones: in this sense, perceptions to which no real objects correspond may be indistinguishable from true perceptions. Accordingly, we find opponents of the Academics making two quite distinct replies: that no two real objects are perfectly indistinguishable, and that dreams and hallucinations are not persuasive in quite the same way as ordinary waking perception. 32 Though there can be little doubt that the sources of Cartesian scepticism are primarily Academic, we see that Descartes' dream argument is the product of uniting considerations that the Academics keep distinct. To effect the union, Descartes must dissociate indistinguishability of perceptions with respect to content from causation by nearidentical objects. His new conception of the mental, which allows him to think of 'sensations' in abstraction from the senses, makes this possible. Lacking this, the Academic sceptics think of sensation in partly causal-physical terms, as an affection of the living organism. 33 Accordingly, they think of similar perceptions as perceptions of

(caused by) similar objects. This precludes uniting their arguments so as to raise the question of whether there is any external world at all. Not that they would have wished to unite them. Academic scepticism with respect to the senses seems to have developed in response to the Stoic ____________________ 32 Cicero, Academica, 529-41. 33 Sextus, Against the Logicians, I, 162. -46attempt to explain a 'cognitive' impression (phantasia kataleptikē) as an impression that, in virtue of being caused in a certain standard way, is both incapable of falsity and distinctively persuasive. 34 Against this Stoic solution to the problem of the criterion, the Academics argue that the existence of indistinguishable objects shows that normal causation does not guarantee true opinions. Present successively even a Stoic sage with two eggs, and he will be unable to say whether he has seen two eggs or the same egg twice. He can go wrong, even though the conditions of perception are ideal. Notice that the whole point of this argument turns on associating indistinguishability of objects presented with causally normal perception. That we sometimes 'perceive' things that do not really exist makes a different but complementary point. Where the causal background is ex hypothesi deviant, as in dreams and madness, there is no reason to suppose that ordinary experience will ever be exactly duplicated. Indeed, the point the Academics want to make becomes more striking given the commonsense idea that strange things are seen in madness and in dreams. For the point is that, since dreamers and madmen can be taken in by their delusions, not even the wildest experiences need differ from ordinary perceptions in respect of persuasiveness, however else they may differ. Thus, not merely does normal causation not guarantee truth, deviant causation does not guarantee unpersuasiveness. The Stoic theory of cognitive impressions fails on two counts. So, though the elements of Cartesian scepticism are Academic, the way they are put together is Descartes' own. His tacit metaphysical presuppositions are what make possible this realignment of traditional materials. VII We may conclude that Descartes' doubts are not natural, not metaphysically noncommittal. On the contrary, they are informed from the outset by the metaphysical considerations that are supposed to emerge in response to them. Only Descartes' artful method of exposition--which involves developing his doubts in a way that is initially rather vague, but also reminiscent of traditional scepticism, and subsequently clarifying and interpreting them, under the guise of recollecting what has already been agreed-prevents this from being immediately apparent.

____________________ 34 See Michael Frede, "'Stoics and Skeptics on Clear and Distinct Impressions'", in Burnyeat, Skeptical Tradition. -47This means that the story told by the Meditations cannot be taken at face value. In this story, metaphysical discoveries are made in the course of exploring the limits of scepticism. But the scepticism whose limits are explored involves the special dubitability of beliefs about the external world, and the external world only emerges as a distinct problem once the important metaphysical concessions have already been made. The unnaturalness of the Cartesian doubt is, however, perhaps not always easy for a modern reader to see. When we approach the First Meditation for the first time, most of us are either already vaguely familiar with some central Cartesian ideas, or we are taken through the Meditations by a teacher who is thoroughly familiar with them. This is why the uncomprehending reactions of intelligent contemporaries like Hobbes and Gassendi are worth pondering. 35 In his Dictionary article "'Pyrrho'", Bayle records a doubtless fictitious conversation between two priests, one hostile to scepticism, the other sympathetic. The hostile priest expresses surprise that there are any sceptics left. He can understand how there might have been pagan sceptics, for the pagans lacked the guidance of revelation; but he cannot understand expressions of scepticism from men of his own time, who live in the light of the gospels. But the sympathetic priest replies that it is even easier for modern men to be sceptics than it was for the ancients. Not merely is Christian doctrine itself a fertile source of paradoxes (think of the Trinity), but arguments deriving from the new science of bodies, which tend to show that some qualities we used to ascribe to bodies, colour and heat, for example, are really ideas in the mind, are readily extended to all qualities, with the result that we must wonder whether we know anything about physical reality. If my argument in this essay is correct, and we interpret their remarks rather freely, Bayle's priests are both right. The one hostile to scepticism is pointing to the impossibility of a modern man's being a classical sceptic, and, whether he was right that this was an impossibility in religion, it was fast becoming an impossibility in science. In Sextus's time, it may have been possible to argue with roughly equal plausibility for, say, Aristotle's conception of matter on the one hand, and Democritus's on the other. Indeed, even Montaigne, writing in the second half of the sixteenth century, could still argue in the manner of Sextus, parading the endless diversity of philosophers' opinions with a view to inducing in the reader a sense of the futility and ridiculousness of ever claiming to know the truth about anything. But the time was fast approaching when no one could seriously ____________________ 35 Hobbes saw nothing new in Descartes' arguments for scepticism, which he referred to slightingly as 'those matters of ancient lore'. See HR II60 [ AT VII171: CSM II121].

-48claim to be able to argue as convincingly for Aristotle's physics as for Newton's. The classical route to scepticism through the balancing of conflicting opinions was being closed off, at least for large areas of inquiry. The persistence of irresolvable conflict was on its way to becoming an indication that the dispute was merely verbal, as it is in Hume. 36 But, by a curious irony, metaphysical reflections connected with the very scientific advances that killed classical scepticism were in the process of giving rise to a new form of scepticism, in its way far more disturbing. This is what Bayle saw: indeed, what he saw, correctly, as the legacy of Descartes. ____________________ 36 Hume takes the capacity to be subverted by 'a due contrast and opposition' to be the hallmark of the vacuous opinions that are produced by the 'capricious' principles of the imagination. See his Treatise on Human Nature [ 1739-40], 225. -49II THE COGITO AND ITS IMPORTANCE PETER MARKIE The basic story is well known. Descartes goes looking for something absolutely certain, beyond even the slightest, most unreasonable doubt, to serve as the permanent foundation for his knowledge. He dismisses the propositions evidenced by his senses. The traditional sceptical worries about hallucinations, madness, dreams, and deceiving gods convince him that there is no certainty there. He lands on a bedrock certainty capable of withstanding even his worries about a deceptive god: he exists. But I have convinced myself that there is absolutely nothing in the world, no sky, no earth, no minds, no bodies. Does it now follow that I too do not exist? No: if I convinced myself of something then I certainly existed. But there is a deceiver of supreme power and cunning who is deliberately and constantly deceiving me . . . ; and let him deceive me as much as he can, he will never bring it about that I am nothing so long as I think that I am something. So after considering everything very thoroughly, I must finally conclude that this proposition, I am, I exist, is necessarily true whenever it is put forward by me or conceived in my mind. (Second Meditation ( A T VII25: C S M II16-17)) Descartes quickly includes claims about his mental state in his list of certainties. He is certain that he thinks, doubts, imagines, wills, and the like: 'But what then am I? A thing that thinks. What is that? A thing that doubts, understands, affirms, denies, is willing, is unwilling, and also imagines and has sensory experiences' ( A T VII28: C S M II19). He is also certain of what he seems to perceive: 'For example, I am now seeing light, hearing

a noise, feeling heat. But I am asleep, so all this is false. Yet I certainly seem to see, to hear, and to be warmed. This cannot be false; what is called "having a sensory perception" is strictly just this, and in this restricted sense of the term it is simply thinking' ( A T VII29: C S M II19). Descartes goes on to define the central problem of epistemology for the next 300 years. How can we move from our certain knowledge of the ____________________ Peter Markie, "'The Cognito and its Importance'" from John Cottingham (ed.): The Cambridge Companion to Descartes ( 1992), 140 -73, by permission of the author and the publisher, Cambridge University Press. -50content of our experience to a knowledge of its cause? How can we know whether our experience is caused by an external world that is basically the way the content of our experience represents it as being or by an external world that is radically different, say one that contains a deceptive god who manipulates our mind? Descartes solves the problem by appeal to an omnipotent, omnibenevolent God. Subsequent philosophers have made more plausible, but generally less elegant, moves. I want to concentrate on Descartes' initial claim to certainty about his thought and existence. On the surface, the dictum generally used to sum up this position--'I am thinking, therefore I am'--seems as obvious and as uninteresting as the claim that fish do not need bicycles. The dictum is not even new with Descartes, since Augustine anticipates him. 1 Descartes admits that his position is obvious, telling us his dictum is 'so simple and natural that it might have occurred to any writer' (letter of Nov. 1640 ( A T III248: C S M K159)). 2 Yet, Descartes' claim to certainty about his thought and existence is extremely important for both his epistemology and his metaphysics, and once we get beyond a superficial reading of the text, his account of how he gains this certainty turns out to be one of the most confusing aspects of his philosophy. I shall briefly discuss the importance of Descartes' claim to certainty, and then I shall develop an interpretation of his position that clears away the main points of confusion. Finally, I shall consider some general philosophical problems that are raised by Descartes' position. I. THE IMPORTANCE OF DESCARTES' POSITION Descartes' claim to certainty about his thought and existence is central to his general programme in epistemology. He wants to answer scepticism, and he wants to do so within foundationalism, the view that all our knowledge begins with some self-evident beliefs which are not evidenced by any others but yet provide our justification for all the rest we know. To succeed in this programme, Descartes must define the set of selfevident ____________________

1

See Augustine, De Trinitate, Bk. X. ch. 10. Descartes claims that he moves beyond Augustine's point by seeing that his certainty of his thought and existence and his uncertainty about his body provide the basis for a defence of the distinction between himself and his body: A T III247: C S M K159. For more on the relation between Descartes' position and Augustine's, see Elizabeth Anscombe, "'The First Person'", in S. Guttenplan (ed.), Mind and Language ( Oxford: Oxford University Press, 1975), 4565, Curley, Descartes Against the Skeptics, and Noonan, "'Identity and the First Person'", in C. Diamond and J. Teichman (eds.), Intentions and Intentionality ( Ithaca, NY: Cornell University Press, 1979). 2 The addressee of this letter is a matter of debate; cf. A T V660 n. -51beliefs, and show that its membership is both certain and extensive enough to support the rest of our knowledge about the world. His claim to certainty about his thought and existence is the initial move in his attempt to do so. Descartes also thinks that his claim to certainty of his thought and existence plays an important role in his metaphysics. In the letter where he says that the dictum 'I think, therefore I am' is so obvious that it might have come from anyone's pen, he observes that its real value is that it can be used to 'establish that this I which is thinking is an immaterial substance with no bodily element' ( ibid. ). The idea that he can use his initial certainty about his thought and existence, and his initial uncertainty about his body, to establish that he is an immaterial substance distinct from his body is a continuing theme in Descartes' philosophy. He hints at it in the Rules: 'Again there are many instances of things which are necessarily conjoined, even though most people count them as contingent, failing to notice the relation between them: for example the proposition, "I am, therefore God exists", or "I understand, therefore I have a mind distinct from my body"' (Rule XII ( A T X421-2: C S M I46)). Just as we can derive God's existence from our own, we can somehow derive the distinctness of our mind and body from the fact that we understand. Descartes' comments become more informative as his philosophy develops. In The Search for Truth, he sketches an argument from premisses about his knowledge of himself and his ignorance of his body to the conclusion that he is distinct from his body: Indeed, I do not even know whether I have a body; you have shown me that it is possible to doubt it. I might add that I cannot deny absolutely that I have a body. Yet even if we keep all these suppositions intact, this will not prevent me from being certain that I exist. On the contrary, these suppositions simply strengthen the certainty of my conviction that I exist and am not a body. Otherwise, if I had doubts about my body, I would also have doubts about myself, and I cannot have doubts about that. I am absolutely convinced that I exist, so convinced that it is totally impossible for me to doubt it. ( A T X518: C S M II412)

In the Discourse on the Method, Descartes begins with the premiss that he is certain of his thought and existence but uncertain of his body: Next I examined attentively what I was. I saw that while I could pretend that I had no body and that there was no world and no place for me to be in, I could not for all that pretend that I did not exist. I saw on the contrary that from the mere fact that I thought of doubting the truth of other things, it followed quite evidently and certainly that I existed; whereas if I had merely ceased thinking, even if everything else I had ever imagined had been true, I should have had no reason to believe that I existed. And he infers that he is distinct from his body: -52From this I knew I was a substance whose whole essence or nature is simply to think, and which does not require any place, or depend on any material thing, in order to exist. Accordingly this 'I'--that is, the soul by which I am what I am--is entirely distinct from the body, and indeed is easier to know than the body, and would not fail to be whatever it is, even if the body did not exist. ( A T VI32-3: C S M I127) In the Meditations, Descartes says that thought is the only thing he knows to be part of his essence prior to proving God's existence, and that 'it follows from the fact that I am aware of nothing else belonging to my essence, that nothing else does in fact belong to it' ( A T VII8: C S M II7). In the Principles, he reports his certainty of his thought and existence and his uncertainty of his body, and then says that 'this is the best way to discover the nature of the mind and the distinction between the mind and the body' (part i, art. 8 ( A T VIIIA 7: C S M I195)). The importance of Descartes' claim to certainty of his thought and existence extends beyond the role it plays in his programmes in epistemology and metaphysics. Understanding and evaluating what he has to say require us to come to grips with some basic philosophical issues that he himself ignores or treats only in passing: e.g. issues about how we think of ourselves and distinguish ourselves from other objects in the world. Descartes' position thus provides us with the occasion for more philosophic work and discovery. It is time to gain a better appreciation of Descartes' position. I shall begin with an interpretation that captures many, but not all, of his statements. II. THE SELF-EVIDENT INTUITION/IMMEDIATE INFERENCE INTERPRETATION Descartes presents intuition and deduction as his only sources of certainty in the Regulae:

But in case we in turn should slip into the same error, let us now review all the actions of the intellect by means of which we are able to arrive at a knowledge of things with no fear of being mistaken. We recognize only two: intuition and deduction. (Rule III ( A T X386: C S M I14)) Deduction is 'the inference of something as following necessarily from some other propositions which are known with certainty' ( A T X369: C S M I15). Intuition is the faculty by which we gain the initial certainties that make deduction possible. By 'intuition' I do not mean the fluctuating testimony of the senses or the deceptive judgement of the imagination as it botches things together, but the conception of a -53clear and attentive mind, which is so easy and distinct that there can be no room for doubt about what we are understanding. Alternatively, and this comes to the same thing, intuition is the indubitable conception of a clear and attentive mind which proceeds solely from the light of reason. ( A T X368: C S M I 14 ) Intuition is distinguished from deduction by the fact that it does not involve a movement of thought through a series of inferences and by its immediate self-evidence: 'Hence we are distinguishing mental intuition from certain deduction on the grounds that we are aware of a movement or a sort of sequence in the latter but not in the former, and also because immediate self-evidence is not required for deduction, as it is for intuition' ( A T X370: C S M I15). Descartes decides that while self-evident propositions are known only by intuition, and conclusions derived from them in several intuited steps are known only by deduction, propositions immediately inferred from self-evident intuitions may be described as known by either intuition or deduction, depending on our perspective: 'It follows that those propositions which are immediately inferred from first principles can be said to be known in one respect through intuition, and in another respect through deduction. But the first principles themselves are known only through intuition, and the remote conclusions only through deduction' ( A T X370: C S M I15). When we immediately infer a conclusion from an intuited self-evident premiss, we are not aware of any movement of thought through a series of premisses, so we may describe our knowledge of the conclusion as intuitive. No extended series of intuitions leads us to the conclusion; there is just one mental act in which the self-evident premiss is intuited and the immediate conclusion is drawn. Yet, we are also inferring a conclusion from a premiss, so we may also describe our knowledge as deductive. Descartes says that we gain knowledge of our thought and existence by intuition: 'Thus, everyone can mentally intuit that he exists, that he is a thinking thing, that a triangle is bounded by just three lines, and a sphere by a single surface, and the like. Perceptions

such as these are more numerous than most people realize, disdaining as they do to turn their minds to such simple matters' ( A T X368: C S M I14). Descartes would presumably extend his appeal to intuition to account for his certainty about his particular mental states: e.g. that he doubts, wills, imagines, seems to see light, hear noise, feel heat. Descartes' appeal to intuition is unclear. The set of intuited propositions includes both self-evident propositions not inferred from any others and propositions immediately inferred from self-evident premisses. On which side of this distinction does Descartes place the propositions about his -54mental state and the proposition that he exists? Descartes seems to regard the propositions about his mental state as self-evident ones that are not inferred and the proposition that he exists as one that is immediately inferred from premisses about his mental state. His knowledge of his mental state is intuitive in the primary sense that it is self-evident and not inferred. His knowledge of his existence is intuitive in the extended sense that he immediately infers his existence from intuited premisses about his mental state. Descartes tells a correspondent: You will surely admit that you are less certain of the presence of the objects you see than of the truth of the proposition: I am thinking, therefore I exist? Now this knowledge is not the work of your reasoning, or information passed on to you by your teachers; it is something that your mind sees, feels and handles; and although your imagination insistently mixes itself up with your thoughts and lessens the clarity of this knowledge, it is, nevertheless, a proof of our soul's capacity for receiving from God an intuitive kind of knowledge. ( A T V137-8: C S M K331) Descartes presents the inference that he thinks and therefore exists, and he says that his knowledge is intuitive and not a product of his reasoning. How can his knowledge both involve the inference 'I think, therefore I am' and be intuitive? The answer is that Descartes intuits the self-evident proposition that he thinks and simultaneously immediately infers that he exists. His knowledge that he thinks is intuitive in the primary sense of being self-evident and entirely non-inferential; his knowledge that he exists is intuitive in the extended sense of being immediately inferred from the simultaneously intuited premiss that he thinks. He makes the same point to Mersenne: And when we become aware that we are thinking beings, this is a primary notion that is not derived by means of any syllogism. When someone says, 'I am thinking, therefore I am, or exist,' he does not deduce existence from thought by a syllogism, but, recognizes it as something self-evident by a

simple intuition of the mind. This is clear from the fact that if he were deducing it by means of a syllogism, he would have to have had previous knowledge of the major premise 'Everything which thinks is, or exists'; yet in fact he learns it from experiencing in his own case that it is impossible that he should think without existing. (Second Replies ( A T VII140: C S M II100)) Descartes again presents the immediate inference from his thought to his existence, and he says that his knowledge is not deductive but a simple intuition of the mind. His point again seems to be that his knowledge of his thought is intuitive since it involves his grasping a self-evident, noninferred premiss, and his knowledge of his existence is intuitive since it involves his immediately inferring that he exists from the simultaneously -55intuited premiss that he thinks. Descartes' remarks in the Discourse and Principles further support this interpretation: And observing that this truth 'I am thinking, therefore I exist' was so firm and sure that all the most extravagant suppositions of the sceptics were incapable of shaking it, I decided that I could accept it without scruple as the first principle of the philosophy I was seeking. ( Discourse ( A T VI32: C S M I127)) For it is a contradiction to suppose that what thinks does not, at the very time when it is thinking, exist. Accordingly, this piece of knowledge--I am thinking, therefore I exist--is the first and most certain of all to occur to anyone who philosophizes in an orderly way. ( Principles, part i, art. 7 ( A T VIIIA 7: C S M I195)) In each passage, Descartes presents his immediate inference from his thought to his existence as a single piece of knowledge; it is the first principle of his philosophy. His point seems to be that in one act of intuition, he grasps the premiss and immediately infers the conclusion. It is important to note two other points. First, Descartes says that the initial premiss that he thinks can be replaced by other claims about his mental state, e.g. that he seems to see: For if I say 'I am seeing, or I am walking, therefore I exist', and take this as applying to vision or walking as bodily activities, then the conclusion is not absolutely certain. This is because, as often happens during sleep, it is possible for me to think I am seeing or walking, though my eyes are closed and I am not moving about; such thoughts might even be possible if I had no body at all. But if I take 'seeing' or 'walking' to apply to the actual sense or awareness of seeing or walking, then the conclusion is quite certain, since it relates to the mind, which alone has the sensation or thought that it is seeing or walking. ( Principles, part i, art. 9 ( A T VIIIA 7: C S M I195))

Second, Descartes' talk of intuition and deduction from intuitions as our two sources of knowledge in the Rules gives way to talk of clear and distinct perception in the Discourse, Meditations, and Principles. He never announces that the faculties are the same, but their equivalence is strongly suggested by the fact that he designates them by similar descriptions: 'the light of reason' and 'the light of nature'. We are told in the Rules that 'intuition is the indubitable conception of a clear and attentive mind which proceeds solely from the light of reason [rationis luce]' (Rule III ( A T X368: C S M I14)) and in the Principles that 'the light of nature [lumen naturæ] or faculty of knowledge which God gave us can never encompass any object which is not true in so far as it is indeed encompassed by this faculty, that is, in so far as it is clearly and distinctly perceived' (part i, art. 30 ( A T VIIIA 16: C S M I203); consider too Meditations ( A T VII38-9: C S M II26-7)). 3 ____________________ 3 Descartes uses 'natural reason' (ratio naturalis) and 'the natural light' (lumen naturale) interchangeably in Comments on a Certain Broadsheet ( A T VIIIB 353: C S M I300). -56We may, then, state Descartes' explanation of his certainty of his thought and existence in terms of clear and distinct perception: all his clear and distinct perceptions are certain; he directly, non-inferentially, clearly, and distinctly perceives the propositions about his thought; and he clearly and distinctly perceives that he exists by immediately deriving that claim from a clearly and distinctly perceived premiss about his thought. As he puts it when he reflects on his knowledge that he is a thinking thing at the start of the Third Meditation: 'In this first item of knowledge there is simply a clear and distinct perception of what I am asserting' ( A T VII35: C S M II24). I shall call this interpretation 'The Self-Evident Intuition/Immediate Inference Interpretation'. I now want to present some passages that cause problems for it. Then we can see how the problems can be solved. III. OUR PRIOR KNOWLEDGE OF THE GENERAL PRINCIPLE THAT WHAT THINKS MUST EXIST One problematic passage is in the Principles: And when I said that the proposition I am thinking, therefore I exist is the first and most certain of all to occur to anyone who philosophizes in an orderly way, I did not in saying that deny that one must first know what thought, existence and certainty are, and that it is impossible that that which thinks should not exist, and so forth. But because these are very simple notions, and ones which on their own provide us with no knowledge of anything that exists, I did not think they needed to be listed. (Part i, art. 10 ( A T VIIIA 8: C S M I196))

Descartes says that prior to knowing that he thinks and therefore exists, he must know, not only what thought, existence and certainty are, but also the general proposition that it is impossible for what thinks not to exist. His point seems to be that his inference from his thought to his existence uses the general proposition as a suppressed premiss. It is not 'I am thinking, therefore I exist'; it is 'I am thinking and whatever is thinking must exist, therefore I exist'. We might try to accommodate this passage by modifying our interpretation of Descartes' inference. His explanation of his certainty would then be that he clearly and distinctly perceives the self-evident proposition that he thinks and the self-evident proposition that whatever is thinking must exist, and he deduces that he therefore exists. His clear and distinct perception of the premisses and the conclusion is enough to make them certain. We cannot get out of trouble this easily. Recall Descartes' comment to Mersenne: -57And when we become aware that we are thinking beings, this is a primary notion that is not derived by means of any syllogism. When someone says, 'I am thinking, therefore I am, or exist,' he does not deduce existence from thought by a syllogism, but, recognizes it as something self-evident by a simple intuition of the mind. This is clear from the fact that if he were deducing it by means of a syllogism, he would have to have had previous knowledge of the major premise 'Everything which thinks is, or exists', yet in fact he learns it from experiencing in his own case that it is impossible that he should think without existing. (Second Replies ( A T VII140: C S M II100)) Descartes explicitly denies that his inference from his thought to his existence is a syllogism, using the general premiss that whatever thinks must exist. He also says that we learn that we think and therefore exist prior to learning that whatever thinks must exist. He repeats the point in response to one of Gassendi's objections: The author of the Counter-Objections claims that when I say 'I am thinking, therefore I exist' I presuppose the major premise 'Whatever thinks exists, and hence I have already adopted a preconceived opinion. . . . the most important mistake our critic makes here is the supposition that knowledge of particular propositions must always be deduced from universal ones, following the same order as that of a syllogism in Dialectic. Here he shows how little he knows of the way in which we discover the truth. It is certain that if we are to discover the truth we must always begin with particular notions in order to arrive at general ones later on (though we may reverse the order and deduce other particular truths once we have discovered general ones). (Appendix to the Fifth Objections and Replies ( A T IX205-6: C S M II271))

If we modify our interpretation so that Descartes' inference uses the general premiss that whatever thinks must exist, we shall be in conflict with his replies to Mersenne and Gassendi; but if we do not modify our interpretation in this way, how are we to account for his claim in the Principles that prior to knowing that we think and therefore exist, we must know that whatever thinks must exist? 4 ____________________ 4 Williams, Descartes, 91-2, suggests that Descartes' position should be understood relative to two ways of interpreting the claim 'Whatever thinks must exist'. We can read the claim so it presupposes that there are things that think, or we can read it so it does not. When Descartes says he must know the general claim in order to know that he exists, he is thinking of the nonexistential version. His inference from his thought to his existence is 'I think, whatever thinks must exist; therefore, I exist', where the general premiss is read so it does not presuppose the existence of thinking things. When Descartes says that his knowledge of the general claim is based on his knowledge of the particular claims that he thinks and he exists, and when he says that he does not use a syllogism in which the general claim is a premiss to learn that he exists, he has the existential version of the claim in mind. Yet this suggestion still runs up against Descartes' assertion that his knowledge of his thought and existence does not involve reasoning and is intuitive; for we have seen that he applies the term 'intuition' only to what is either selfevident or immediately inferred from the selfevident. -58Our Awareness of Substances The Principles and Third Objections and Replies contain other problematic passages: However, we cannot initially become aware of a substance merely through its being an existing thing, since this alone does not have any effect on us. We can, however, easily come to know a substance by one of its attributes, in virtue of the common notion that nothingness possesses no attributes, that is to say, no properties or qualities. Thus, if we perceive the presence of some attribute, we can infer that there must also be present an existing thing or substance to which it may be attributed. ( Principles, part i, art. 52 ( A T VIIIA 25: C S M I210)) If I may briefly explain the point at issue: it is certain that a thought cannot exist without a thing that is thinking; and in general no act or accident can exist without a substance for it to belong to. But we do not come to know a substance immediately, through being aware of the substance itself; we come to know it only through its being the subject of certain acts. (Third Objections and Replies ( A T VII175-6: C S M II24))

Descartes seems to be saying the following. Each object consists of qualities and an underlying substance in which the attributes inhere. All we immediately observe when we are aware of an object is some of its attributes, and from the existence of the observed qualities and the general principle that every observed quality is in some substance, we infer the existence of the underlying subject. This general position implies an account of self-knowledge that is at odds with the one we have attributed to Descartes. When we turn our attention inward and reflect on ourselves, all we are immediately aware of is our thoughts. Our initial knowledge is not correctly reported by the statement 'I am thinking', but by the statement 'Thought is taking place'. We cannot immediately infer our existence from this knowledge. The best we can infer is that, since every observed quality is in some substance, some substance thinks. We can reason, 'Thought exists and whenever any observed quality exists there is a substance that has it, so there is a thinking substance.' Our knowledge is inferential, but the inference is syllogistic rather than immediate, and it has a different beginning and end than the inference 'I am thinking, therefore I exist'. We cannot, of course, solve our problem by deciding that Descartes adopts the syllogistic inference just considered and offers it as his account of how he gains certain selfknowledge through clear and distinct perception. What would we then make of his explicit denials that his inference from thought to existence is a syllogism? What would we make of the fact that he states his inference from thought to existence so it includes an -59explicit reference to himself in particular? It is 'I am thinking, therefore I exist'; not 'Thought is taking place, every observed quality is in some substance, therefore some thinking substance exists'. 5 The Uncertainty of Clear and Distinct Perception The Self-Evident Intuition/Immediate Inference Interpretation attributes to Descartes the principle that clear and distinct perception always produces certainty, even in the face of reasons for doubt like the Deceptive God Hypothesis. Descartes is certain that he thinks and therefore exists, because he clearly and distinctly perceives that fact. Yet, while Descartes seems to make clear and distinct perception a sufficient condition for certainty in the Rules, he seems to change his mind when he subjects his faculties to a more critical examination in the Meditations and Replies. In the beginning of the Third Meditation, he writes: But what about when I was considering something very simple and straightforward in arithmetic or geometry, for example that two and three added together make five, and so on? Did I not see at least these things clearly enough to affirm their truth? Indeed, the only reason for my later judgement that they were open to doubt was that it occurred to me that

perhaps some God could have given me a nature such that I was deceived even in matters which seemed most evident. And whenever my preconceived belief in the power of God comes to mind, I cannot but admit that it would be easy for him, if he so desired, to bring it about that I go wrong even in those matters which I think I see clearly with my mind's eye. ( A T VII36: C S M II25) Descartes claims that a version of the Deceptive God Hypothesis gives him a reason to doubt simple truths, even when he sees them with his 'mind's eye'. His reference to his mind's eye surely seems to be one to clear and distinct perception. Descartes' replies to his critics contain further indications that he does not take clear and distinct perception to be a sufficient condition for certainty. When Mersenne observes that 'an atheist is clearly and distinctly aware that the three angles of a triangle are equal to two right angles, but so far is he from supposing the existence of God that he completely denies it' (Second Objections ( A T VII125: C S M II89)), Descartes replies: The fact that an atheist can be 'clearly aware that the three angles of a triangle are equal to two right angles' is something I do not dispute. But I maintain that this ____________________ 5 Descartes sometimes goes out of his way to stress the first-person element in his claim to certainty of his thought and existence by adding the pronoun ego, which is superfluous in Latin. Twice in the Principles and once in the Second Objections and Replies, he writes: 'ego cogito, ergo sum' ( A T VIIIA 7: C S M I195; A T VIIIA 8: C S M I196; A T VII140: C S M II100). -60awareness of his is not true knowledge, since no act of awareness that can be rendered doubtful seems fit to be called knowledge. Now since we are supposing that this individual is an atheist, he cannot be certain that he is not being deceived on matters which seem to him to be very evident (as I fully explained). (Second Replies ( A T VII141: C S M II101)) Descartes denies that the atheist has 'true knowledge' on the grounds that the atheist is uncertain of whether he is deceived by some god. Prior to proving God's existence and non-deceptive nature, Descartes is just as uncertain as the atheist about the existence of a deceptive god. His clear and distinct perceptions should not produce certainty for him either. 6 It is difficult to see how the Self-Evident Intuition/Immediate Inference Interpretation can be modified to take these passages into account. We might try basing the interpretation on a weaker claim about clear and distinct perception: not all clear and distinct perceptions produce certainty, only a proper subset of them do, and Descartes' immediate

inference that he thinks and therefore exists falls in that subset. What, then, is this proper subset? In the passages above from the Third Meditation, Descartes subjects even very simple propositions perceived utterly clearly by the mind's eye to the doubt raised by the hypothesis of a deceptive god. The Irrelevance of Clear and Distinct Perception A final problem for the Self-Evident Intuition/Immediate Inference Interpretation is raised by Descartes' Second Meditation discussion of his certainty of his existence: But I have convinced myself that there is absolutely nothing in the world no sky, no earth, no minds, no bodies. Does it now follow that I too do not exist? No: if I convinced myself of something then I certainly existed. But there is a deceiver of supreme power and cunning who is deliberately and constantly deceiving me; and let him deceive me as much as he can, he will never bring it about that I am nothing so long as I think that I am something. So after considering everything very thoroughly, I must finally conclude that this proposition, I am, I exist, is necessarily true whenever it is put forward by me or conceived in my mind. ( A T VII25: C S M II1617) Descartes claims certainty of his existence, but he does not once mention clear and distinct perception or an immediate inference from thought to existence. His explanation instead seems to be that he is certain he exists ____________________ 6 Of course, in replying to Mersenne, Descartes refers explicitly only to clarity and not to both clarity and distinctness. Yet, while he only mentions clarity, he does not go on to deny Mersenne's claim that an atheist can have both a clear and a distinct perception. Instead he goes on to deny that an atheist can have the certainty required for scientific knowledge. It hardly seems likely that he would do this if his argument with Mersenne was over whether an atheist could have both a clear and a distinct perception. -61because he has no reason to doubt that belief, and he has no reason to doubt it, because every hypothesis that might give him a reason to doubt it, such as the hypothesis that some god deceives him, simply entails, and so affirms, it. Descartes' point about his belief in his existence can be extended to his belief that he thinks. Reasons for doubt like the Deceptive God Hypothesis simply entail that his belief is true. Yet, if this is all there is to Descartes' position, what are we to make of his other claims to certainty about his mental state: e.g. his claim to be certain that he seems to see light? Reasons for doubt, like the Deceptive God Hypothesis, do not entail that he seems to see light. He must exist and think to be deceived; he does not need to seem to see light. What

are we to make of Descartes' references to clear and distinct perception (intuition) as the source of his certainty and to his immediate inference 'I am thinking, hence I exist'? It has been suggested that the point of Descartes' inference is just that every reason for doubt entails his existence by entailing that he thinks. The problem with this suggestion is that Descartes says the premiss in his inference can be any claim about his mental state; he may just as well reason, 'I seem to see; hence I exist'. The point of this inference surely is not that every reason for doubt entails Descartes' existence by entailing that he seems to see. 7 We need to account for Descartes' point that his certainty of his thought and existence results from the fact that every potential reason for doubt affirms that he thinks and exists, but we need to do so in a way that still lets us account for his claim to certainty about his other mental states, his reference to clear and distinct perception, and his immediate inference from his mental state to his existence. IV. TOWARD AN IMPROVED INTERPRETATION We have found several passages that conflict with the Self-Evident Intuition/Immediate Inference Interpretation, even though we initially developed that interpretation on the basis of strong textual evidence. Should we just decide that Descartes is wildly inconsistent, or, more charitably, that ____________________ 7 See Curley, Descartes Against the Skeptics, for the suggestion of how to capture Descartes' inference. See Hintikka, "'Cogito ergo sum: Inference or Performance'", and "'Cogito ergo sum as an Inference or a Performance'", for an interpretation of Descartes' position that relies heavily on the Second Meditation passage at hand, to the point of paying inadequate attention to Descartes' references to clear and distinct perception and to an immediate inference from his mental state to his existence. For discussions of Hintikka's interpretation, see Feldman, "'On the performatory Interpretation of the Cogito'"; Frankfurt, Demons, Dreamers, and Madmen; Kenny, Descartes; and Wilson, Descartes. -62his brilliance causes him to see several ways to explain his certainty of his thought and existence, his open-mindedness keeps him from being able to choose between them, and his charity makes him leave the choice to us? I think not. We can modify the Self-Evident In tuition/Immediate Inference Interpretation to account for some, though not all, of the problematic passages. The rest have alternative readings consistent with the modified interpretation. To develop this new interpretation, I must first examine two basic concepts of Descartes' epistemology, his concept of certainty and his concept of a reasonable belief. Descartes' epistemology contains two degrees of epistemic appraisal; that is to say, two degrees of justification relative to which beliefs are assessed. One is the top standard of

certainty. The other is a lesser degree of justification, which Descartes describes as highly probable or very reasonable belief. In the First Meditation, after he decides that his sensory evidenced beliefs about the external world are not certain, he notes that they are none the less very reasonable: My habitual opinions keep coming back, and, despite my wishes, they capture my belief, which is as it were bound over to them as a result of long occupation and the law of custom. I shall never get out of the habit of confidently assenting to these opinions, so long as I suppose them to be what they are, namely highly probable opinions--opinions, which, despite the fact that they are in a sense doubtful, as has just been shown, it is still much more reasonable to believe than to deny. ( A T VII 22: C S M II15) 8 It is plausible to think Descartes would accept a few basic principles about these two grades of epistemic appraisal. First, all the beliefs that meet the demands of certainty for him, such as his beliefs about his thought and existence in the Second Meditation, are also very reasonable, but some of his very reasonable beliefs, such as his sensoryevidenced beliefs about the external world in the First Meditation, are not certain for him. Second, which degree of epistemic appraisal a belief meets is determined by his evidence for the belief. The sensory evidence Descartes has for his external world beliefs in the First Meditation makes those beliefs very reasonable but not certain. The evidence Descartes has for his belief in his existence in the Second Meditation makes that belief both very reasonable and certain. When a belief is self-evident, Descartes' evidence ____________________ 8 The certain and the very reasonable are not the only grades of epistemic appraisal in Descartes' epistemology. For example, early in the Sixth Meditation, he takes the claim that the body exists to be probable on the grounds that it provides the best explanation of some data about the imagination ( A T VII73: C S M II51). The degree of appraisal involved here is lower than either the certain or the very reasonable (the highly probable). The data about his imagination do not support the existence of body to the point of making it either certain or very reasonable. -63for it consists of his act of clearly and distinctly perceiving it. When a belief is not selfevident, Descartes' evidence for it consists of those beliefs that constitute his reason for believing it.Third, the difference between what is merely very reasonable and what is certain is that Descartes has a slight reason to doubt the former. Descartes' reason for doubt must be slight, since the beliefs are very reasonable--as he puts it, his reason for doubt is 'metaphysical and exaggerated' ( A T VII460: C S M II308)--but, even a slight reason for doubt keeps a belief from being certain. A hypothesis gives Descartes a reason to doubt one of his beliefs just when it is a possibility he has not ruled out, and it indicates how his belief might be false despite his evidence. The hypotheses that he is dreaming and that some god deceives him are possibilities he has not ruled out in the

First Meditation, and they indicate how his very reasonable beliefs about the external world might be false despite his sensory evidence for them. Commentators have offered competing accounts of how a reason for doubt is a possibility that Descartes has not ruled out. The one most in keeping with Descartes' remarks is that a reason for doubt is a possibility he has not ruled out in the sense that he is not certain it is false. In the First Meditation, Descartes is not certain he is not dreaming and not being deceived. Once he decides he is certain of these points--at the end of the Sixth and Third Meditations, respectively-he rejects the Dream and Deceptive God Hypotheses as reasons for doubt. The fact that any hypothesis that has not been ruled out with certainty is capable of serving as a reason for doubt is just what makes Descartes' doubt 'exaggerated', as he puts it. It is also what makes his reasons for doubt so difficult to rule out. 9 Relative to these points, we can better understand Descartes' claim to certainty about his thought and existence. Descartes' claim has two parts: he has evidence for these beliefs that makes them very reasonable, and that evidence resists even the slightest, most exaggerated reasons for doubt, so that his beliefs are certainties; no hypothesis he has yet to rule ____________________ 9 Descartes sometimes uses the terms 'metaphysical certainty' and 'moral certainty', for example, in the Discourse ( A T VI37-8: C S M I129-30). and the Seventh Objections and Replies ( A T VII471: C S M II317). As I understand him, 'metaphysical certainty' refers to what I have been calling 'certainty', and 'moral certainty' refers to what I have been calling 'reasonableness' and 'high probability'. Descartes also writes of a form of certainty that is best termed 'psychological certainty', since it has to do, not with the strength of our evidence for a proposition, but with our inability to doubt it. See his remarks in the Meditations ( A T VII65: C S M II45; A T VII 69-70: C S M II48). Note that the points I make in the text are somewhat independent of these issues; e.g. one can accept the points I make in the text without also accepting my view that morally certain beliefs are to be equated with very reasonable or highly probable ones. For more on all three forms of certainty, see Feldman, "'On the Performatory Interpretation'"; Gewirth, "'The Cartesian Circle'"; Curley, Descartes Against the Skeptics; and Markie, "'The Cogito puzzle'" and Descartes's Gambit. -64out with certainty indicates how his beliefs might be false despite his evidence for them. Since Descartes' claim to certainty is complex, his explanation of it must be equally complex. He must explain what makes his beliefs very reasonable for him, and why no hypothesis he has yet to rule out with certainty indicates how they might be false. Now that we better understand the form Descartes' explanation must take, let us return to the Self-Evident Intuition/Immediate Inference Interpretation and see how it can best be modified. The Modified Self-Evident Intuition/Immediate Inference Interpretation The first thing Descartes must do is explain what makes his beliefs in his thought and existence very reasonable. This is where his frequent appeals to clear and distinct

perception and an immediate inference from his thought to his existence come into play. Descartes takes his belief in his thought to be very reasonable, because the proposition that he thinks is a self-evident one he clearly and distinctly perceives to be true. His act of clear and distinct perception is the 'evidence' that makes his belief that he thinks very reasonable. The same may be said of his other beliefs about his mental state. His belief in his existence is very reasonable, because he immediately infers it from a very reasonable belief about his mental state. Yet, what makes these beliefs so reasonable as to be certain? Descartes' answer is that he has no reason to doubt them. Now his observation about how reasons for doubt just affirm his thought and existence comes into play. Consider once again how he puts the point: 'But there is a deceiver of supreme power and cunning who is deliberately and constantly deceiving me; and let him deceive me as much as he can, he will never bring it about that I am nothing so long as I think that I am something' (Second Meditation ( A T VII25: C S M II17)). The Deceptive God Hypothesis does not give him a reason to doubt his beliefs that he thinks and exists; for it entails them, and so fails to indicate how they might be false despite his clear and distinct perception. The same is true of the Dream Hypothesis. Two aspects of this part of Descartes' position need development. First, the Deceptive God Hypothesis and the Dream Hypothesis clearly fail to cast doubt on his beliefs that he thinks and exists, but why is it that no other hypothesis can do so? Second, why do not these or other hypotheses cast doubt on such beliefs as that he seems to see light, which are sometimes used as premisses for the Cogito? The Deceptive God Hypothesis does not entail that he seems to see light; could not a god deceive him about what he seems to see? -65Descartes might deal with the first issue by adopting three plausible principles about reasons for doubt. First, no hypothesis casts doubt on a contingent belief it entails. 10 Second, a hypothesis indicates to Descartes how one of his beliefs might be false only if it entails the proposition he would express by 'I exist'. The idea is that Descartes must relate a hypothesis to himself before it gives him a reason to doubt, and he does that by making his first-person belief in his existence part of the hypothesis. The hypothesis that Descartes would express by 'Some god deceives the greatest seventeenth-century philosopher' does not give him a reason to doubt his beliefs, but the one he would express by 'Some god deceives the greatest seventeenth-century philosopher and I am that philosopher' does. Third, a hypothesis indicates how one of Descartes' beliefs might be false only if it entails the proposition he would express by 'I think'. The idea is that each reason for doubt must indicate how Descartes' intellectual abilities are leading him astray, due to their own intrinsic limitations or to his mishandling of them, and any hypothesis to that effect will include the information that he thinks. These three principles entail that no hypothesis casts doubt on Descartes' reasonable beliefs in his thought and existence. The first principle requires that a reason to doubt those beliefs must not entail them; the second and third principles require that a reason to doubt must entail them. No hypothesis meets all three requirements.

It is more difficult to fill the gap in Descartes' account of why no hypothesis casts doubt on such mental state beliefs as that he seems to see light. Could not a deceptive god make him think he seems to see light when he really does not, or, perhaps, more plausibly, make him think he seems to see red when he really seems to see orange, or has a pain when he really has an itch? The best way to fill this gap in Descartes' explanation may be a fourth principle: a hypothesis indicates how one of his contingent beliefs might be false only if it is possible for Descartes to have the belief while the hypothesis is true and the belief is false. The idea is that a hypothesis indicates to Descartes how one of his contingent beliefs might be false only if it shows how he could actually have the belief and be mistaken. 11 Relative to this principle, Descartes might argue that he has no reason to ____________________ 10 The principle is restricted to logically contingent propositions to avoid problems caused by the fact that, as logical entailment is strictly defined, every hypothesis entails every proposition that is a logically necessary truth. A proposition p entails a proposition q just when it is logically impossible that p be true and q be false. Some may prefer a more vague principle that is not formulated in terms of the logical relation of entailment between the hypothesis and the belief, and so does not have to be restricted to contingent propositions: no hypothesis h casts doubt on a believed proposition p, if h 'contains' p. The notion of containment is undefined. The next two principles may be modified accordingly. 11 Note that this principle is also restricted to beliefs in logically contingent propositions: i.e. to ones that might have been false. Beliefs in propositions that are necessarily true cannot satisfy -66doubt his beliefs about the contents of his mental states, since it is impossible for him to have those beliefs and for them to be false. They are all incorrigible for him. It is impossible for him to believe falsely that he seems to see red or that he is in pain. To believe that one seems to see red is, in part, to seem to see red. To believe that one is in pain is, in part, to be in pain. 12 It is important to appreciate how we have modified the Self-Evident Intuition/Immediate Inference Interpretation. We have retained the view that, according to Descartes, he clearly and distinctly perceives the selfevident proposition that he thinks and immediately infers that he exists. Yet, we have retained this as Descartes' explanation of why his beliefs in his thought and existence are very reasonable for him. We have appealed to Descartes' statements about how reasons for doubt affirm his thought and existence to develop an explanation of why his beliefs are so reasonable as to be certain. It is now time to see how this modified interpretation deals with the passages that cause problems for the initial one. Some of the passages fit under the modified interpretation quite nicely; the rest can be reinterpreted so they do not conflict with it.

Evaluation of the Modified Self-Evident Intuition/Immediate Inference Interpretation Our modified interpretation easily avoids one of the problems we have examined. Our initial interpretation is inconsistent with Descartes' claim, in the Meditations and Replies, that some clear and distinct perceptions, specifically ones of very simple mathematical truths, are made doubtful by the Deceptive God Hypothesis. Our modified interpretation is consistent with Descartes' claim. It says all clear and distinct perceptions are very reasonable, and only those that concern our thought and our existence are certain. The difference between the certain clear and distinct perceptions and the merely very reasonable ones is that the former resist reasons for doubt like the Deceptive God Hypothesis. Such reasons for doubt do not ____________________ the principle, since it is impossible for them to be false. Again, some may prefer a more vague principle that does not use logical entailment: a hypothesis indicates how one of Descartes' beliefs might be false only if the state of affairs of the hypothesis being true and Descartes having the belief does not 'contain' the state of affairs of the belief being true. 12 Some commentators have noted that Descartes' beliefs about his particular mental activities are not immune to reasons for doubt in exactly the same way as his beliefs that he thinks and exists are. They have taken this to indicate that either Descartes is not serious in claiming certainty of his particular mental states or he is not to be taken literally in his claim that the premiss that he thinks may be replaced by such claims as that he seems to see. See Cottingham, Descartes, 38-42. If I am on the right track, there is no need for such interpretative gymnastics. -67indicate how our beliefs in our thought and existence might be false despite the clear and distinct perceptions that support them. 13 A second problem is our initial interpretation's inability to account for Descartes' comments about how potential reasons for doubt entail that he thinks and exists. The interpretation makes these comments irrelevant, by reducing Descartes' position to just two claims: all clear and distinct perceptions are certain, and he clearly and distinctly perceives his thought and existence. Our modified interpretation avoids the problem. Descartes' references to clear and distinct perception explain the reasonableness of his beliefs about his thought and existence. His comments about how potential reasons for doubt entail his thought and existence help explain why those reasonable beliefs are certain. Both sets of comments are essential to Descartes' explanation. A third problem concerns whether Descartes' inference from his thought to his existence is immediate or a syllogism. Recall the passage from the Principles, quoted above under Section III, which insists on the importance of a prior knowledge of the general principle that whatever thinks must exist. The way to account for this passage is to pay close

attention to what Descartes says he must know prior to knowing his thought and existence. He must know that what thinks must exist, and he must also know what knowledge, thought, existence, and certainty are. His point, surely, is not that all this information must be added to his inference from thought to existence to bridge the gap between his initial premiss and his conclusion. He does not need to add definitions of thought, existence, and certainty to his argument to get from 'I think' to 'I exist'. His point is this: he must have some of this information to understand the propositions that he thinks and that he exists and the rest to understand his account of why they are certain for him. He cannot understand the propositions, unless he knows what thought and existence are. He cannot understand his account of why they are certain for him unless he knows what certainty is. He cannot understand his account of why they are certain unless he knows that what thinks must exist; for part of his account is that his belief that he thinks immediately entails, and so makes reasonable, his belief that he ____________________ 13 Some commentators think we should deal with the problematic passages by interpreting them so they do not contain the claim that some clear and distinct perceptions are uncertain. Discussion of this issue is mainly found in debates on the "'Cartesian Circle'". See Cottingham, Descartes; Curley, Descartes Against the Skeptics; Doney, "'The Cartesian Circle'" and "'Descartes's Conception of Perfect Knowledge'"; Feldman, "'Epistemic Appraisal and the Cartesian Circle'"; Frankfurt, Demons, Dreamers, and Madmen; Gewirth, "'The Cartesian Circle'", "'The Cartesian Circle Reconsidered'", and "'Descartes: Two Disputed Questions'"; Kenny, Descartes, and "'The Cartesian Circle and the Eternal Truths'"; Van Cleve, "'Foundationalism, Epistemic Principles and the Cartesian Circle'" [Ch. IV below]; and Markie, Descartes's Gambit. -68exists. Descartes does not offer the general principle that what thinks must exist as a suppressed premiss in his inference from his thought to his existence. He offers it as something he must know to understand why his thought and existence are certain for him. Moreover, it is sufficient that this general principle is reasonable for Descartes; it need not be certain. When Descartes claims to be certain of his thought and existence in the Second Meditation, he does not offer that claim to certainty--'I am certain about my thoughts and existence'--as a certainty. He presents it and his explanation of why it is true as reasonable beliefs about his epistemic state. 14 The last problem with our initial interpretation concerns Descartes' account of our awareness of substances. The trouble comes from the passages in the Principles and the Third Objections and Replies, quoted above on p. 59, where Descartes seems to say that all we immediately observe when we are aware of an object is some of its qualities, and from the existence of the observed qualities and the general principle that every observed quality is in some substance, we can infer the existence of the underlying subject. When we reflect on ourselves, then, all we immediately observe is our thought. Our initial knowledge is correctly given by the statement 'Thought is taking place', and from this we

can at best reason: 'Thought exists and whenever any observed quality exists there is a substance that has it, so there is a thinking substance.' We do not gain initial knowledge of ourselves by an immediate inference from the self-evident premiss 'I am thinking' to the conclusion 'I exist'. There is a better interpretation of these passages, one that makes them irrelevant to the logic of Descartes' Cogito inference. Take the Principles passage first. Descartes is not concerned to make a general point about the content of our thought when we try to gain knowledge about substances. He is concerned with how we can know that a particular thing is a ____________________ 14 For more on this point see my work in Descartes Gambit, esp. chs. 2 and 5. It is also worth taking note of a passage from the Conversation with Btirman that relates to the Principles passage. Burman reports ( A T V147: C S M K333) that Descartes explains the relation between the Principles passage and his insistence that his knowledge of his thought and existence is intuitive by drawing a distinction between explicit and implicit knowledge. In the Second Meditation, he explicitly knows that he thinks and therefore exists, but he only implicitly knows that whatever thinks must exist. The difference between explicit and implicit knowledge escapes me; it may or may not be related to the solution I have offered. For more on these issues see, Curley, Descartes Against the Skeptics; Frankfurt, "'Descartes'" Discussion of his Existence'; Williams, Descartes; and Wilson, Descartes. Some commentators seem to think that Descartes claims to be certain that he is certain of his thought and existence, which would require him to be certain of his explanation. See Gueroult, Descartes selon l'ordre des raisons, 51, and perhaps Cottingham, Descartes, 41-2, 69-70. I know of no textual support for this position. Descartes does not claim to be certain of his certainty about himself, and nothing he says commits him to such a view. -69substance. His point is that we do not just intuit or observe the fact that a particular thing is a substance; as he puts it, the mere fact that something is a substance 'does not of itself have any effect on us'. We learn that a particular thing is a substance by first observing that it has some qualities and then inferring that it is a substance, by the premiss that whatever has observed qualities is a substance: 'if we perceive the presence of some attribute, we can infer that there must also be present an existing thing or substance to which it may be attributed.' Descartes may be interpreted as making the same point in the passage from the Third Objection and Replies. When he says that 'we do not come to know a substance immediately, through being aware of the substance itself; we come to know it only through its being the subject of certain acts', his point is that we do not just directly observe or intuit that a particular thing is a substance; we infer that fact from the information that it has some observed qualities, and that everything with observed qualities is a substance. In Descartes' own case, then, he does not intuit the proposition he

would express by 'I am a substance', and he does not immediately infer that proposition from any of the self-evident ones about his mental state. He learns that he is a substance by reasoning: 'I think, whatever has an observed quality is a substance; therefore, I am a substance.' This position is consistent with Descartes' claim to know that he thinks and exists by intuiting that he thinks and immediately inferring that he exists. 15 The modified version of the Self-Evident Intuition/Immediate Inference Interpretation thus avoids the textual problems with the initial version. 16 It is time to consider some objections to Descartes' position. They will help ____________________ 15 There is another way to interpret the passages so that they are consistent with our modified interpretation. Instead of being concerned with our knowledge that something is a substance, Descartes may be concerned with our knowledge of those things that are substances. His point may be that we never just know a substance per se. We only know a substance by knowing propositions about it to the effect that it has some attributes. This may be what he has in mind when he writes in the Third Replies that 'we do not come to know a substance immediately, through being aware of the substance itself; we come to know it only through its being the subject of certain acts'. In his own case, Descartes does not just know himself per se; he knows himself by knowing propositions that attribute qualities to him. He is never just aware of himself; he is always aware of himself as having some attribute. This reading again makes the passages consistent with our modified interpretation. It has also been suggested by Wilson, Descartes, 66-7, though she seems to have more reservations about it than I do. Two other passages may also be read in the same way; see Principles, part i, art. ii ( A T VIIIA 8: C S M1196); and Fourth Objections and Replies ( A T VII222: C S M II156). For a treatment of Descartes' position that assumes that he does reason, 'Thought is taking place, every attribute is in a substance, therefore, some substance exists,' see Sievert, "'Descartes's Self-Doubt'" and "'Sellars and Descartes . . .'"; and Kenny, Descartes; see also my discussion of Sievert and Kenny in "'Cogito Puzzle'" and Descartes Gambit. 16 I have not discussed a Third Meditation passage that might still be regarded as problematic for even the modified version of the Self-Evident Intuition/Immediate Inference Interpretation. Descartes writes: 'And since I have no cause to think that there is a God at all, any reason for -70us appreciate some of the underlying philosophical issues that Descartes leaves as exercises for his readers. V. PROBLEMS FOR DESCARTES If we grant that Descartes can transfer his reasonable belief from the proposition that he thinks to the proposition that he exists by an immediate inference, we still should object to his account of how he gains the very reasonable belief that he thinks to begin with. 17

He says he clearly and distinctly perceives that he thinks. His concept of clear and distinct perception is the least clear and distinct concept in his philosophy. He never adequately explains what this mental vision is, or why apprehending a proposition by it is sufficient to make belief in the proposition very reasonable. 18 An especially perplexing point is that Descartes appeals to acts ____________________ doubt which depends simply on this supposition is a very slight, and, so to speak, metaphysical one. But in order to remove even this slight reason for doubt, as soon as the opportunity arises I must examine whether there is a God, and, if there is, whether he can be a deceiver. For if I do not know this, it seems that I can never be quite certain about anything else' ( A T VII36: C S M II25; my emphasis). Descartes' remark implies that until he knows that God exists and is not a deceiver, he cannot be certain of even his own thought and existence. He thus rejects by implication the very claim to certain self-knowledge that we have been trying to understand. Descartes later modifies his position so that it does not contain this implication. Mersenne points out the implication to him: 'It follows from this that you do not yet clearly and distinctly know that you are a thinking thing, since, on your own admission, that knowledge depends on the clear knowledge of an existing God; and this you have not yet proved in the passage where you draw the conclusion that you clearly know what you are' (Second Objections, A T VII125: C S M II89). Descartes replies: 'When I said that we can know nothing for certain until we are aware that God exists, I expressly declared that I was speaking only of knowledge of those conclusions which can be recalled when we are no longer attending to the arguments by means of which we deduced them. Now awareness of first principles is not normally called "knowledge" by dialectitians' (Second Replies, A T VII140: C S M II100). Descartes is mistaken about what he 'expressly declared' in the Third Meditation, but the important point is that, upon consideration, he rejects the view that he must know God to be certain of his own thought and existence. 17 Some commentators object to Descartes' attempt to infer his existence immediately from the premiss that he thinks. Wilson, Descartev, 55; and Kenny, Descartes, 169-70, object that the immediate inference is not valid in first-order quantification theories without existential presuppositions. Hintikka, "'Cogito ergo sum: Inference or Performance'", 114 -15, objects that it is question-begging. See my "'Cogito Puzzle'" and Descartes Gambit for replies to both objections; see Wilson, Descartes, for a reply to Hintikka. To some extent, the criticisms of Descartes' inference are encouraged by his own criticisms of syllogistic reasoning. See Curley, Descartes Against the Skeptics; and Markie, Descartes's Gambit, for discussions of Descartes' criticisms. 18 Descartes gives his most formal definition of clear and distinct perception in Principles, part i, art. 45 ( A T VIIIA 21-2: C S M I207-8). The issue is complicated by the fact that Descartes writes of his clear and distinct perception of propositions (e.g. Discourse ( A T VI33: C S M I 127)), but also of his clear and distinct perception of ideas (e.g. Principles, part i, arts. 45-6 ( A T VIIIA 21-2: C S M I207-8)), of clear and distinct propositions (e.g. Principles, part i, art. 30 -71-

of clear and distinct perception to account for both his knowledge of contingent claims about his mental state and his knowledge of simple necessary truths: 'Thus everyone can mentally intuit that he exists, that he is a thinking thing, that a triangle is bounded by just three lines, and a sphere by a single surface, and the like. Perceptions such as these are more numerous than most people realize, disdaining as they do to turn their minds to such simple matters' ( Regulae, Rule III ( A T X368: C S M I14)). Our knowledge of our mental states is hardly the same as our knowledge of simple necessary truths. It is plausible to say that we learn that every sphere is bounded by a single surface in a mental vision in which we just grasp that the idea of the sphere includes the idea of being bounded by a single surface. Yet, this is not how we learn that we think. We do not learn that we think by perceiving a relation of containment, identity, diversity, or the like between some ideas. 19 Descartes leaves us wondering exactly how our beliefs about our mental state and existence become reasonable. Note that we will not improve matters by simply cutting Descartes' appeal to clear and distinct perception out of his explanation of his certainty of his thought and existence. We will then be left with only the second part of his explanation of his certainty and, in effect, with the observation that each potential reason for doubt entails that he thinks and exists. Descartes will be open to a criticism nicely stated by A. J. Ayer: What Descartes thought that he had shown was that the statements that he was conscious, and that he existed, were somehow privileged, that, for him at least, they were evidently true in a way which distinguished them from any other statements of fact. But this by no means follows from his argument. His argument does not prove that he, or anyone, knows anything. It simply makes the logical point that the one sort of statement follows from another. 20 Descartes needs both parts of his explanation of his certainty. He needs an account of what makes his beliefs in his thought and existence very reasonable and an account of why those reasonable beliefs resist every reason for doubt. Unfortunately, the first part of his position is basically uninformative. ____________________ ( A T VIIIA 17: C S M I203)), and of clear and distinct ideas (e.g. Meditations ( A T VII46: C S M II31)). For further discussion of this topic see Frankfurt, Demons, Dreamers and Madmen; Gewirth, "'Clearness and Distinctness'" [Ch. III. below]; Kenny, Descartes; and Markie, Descartes's Gambit. 19 I assume that Descartes' assertion 'I am thinking' does not have the same content as 'Thought is taking place'. 20 Ayer, "'I Think Therefore I Am'", 82. Feldman ( "'On the Performatory Interpretation'") makes a similar objection to Hintikka's interpretation, which, as we have noted, stresses the logical relations between particular claims and downplays the role of clear and distinct perception. -72-

Descartes also says very little about the content of his beliefs about his mental state and existence. He takes the content of his mental state beliefs to be the propositions he would express by 'I am thinking' and 'I am in pain', rather than those he would express by 'Thought is taking place' or 'Pain is occurring'. The former propositions are about him in particular; they entail his existence. Yet, what exactly is the content of these propositions by virtue of which they are about him? To put the point another way, what is it about Descartes' self-awareness when he clearly and distinctly perceives that he thinks that makes his awareness an awareness of him? Is he directly acquainted with himself in the same way that he is directly acquainted with an idea, like a pain sensation? Is he aware of himself by virtue of conceiving a particular concept of himself? If so, what concept? Descartes seems committed to the view that he is not directly acquainted with himself. He thinks of himself by conceiving an idea of himself. In the Meditations, he writes that: Undoubtedly, the ideas which represent substances to me amount to something more and, so to speak, contain within themselves more objective reality than the ideas which merely represent modes or accidents. Again, the idea that gives me my understanding of a supreme God, eternal, infinite, 〈immutable〉, omniscient, omnipotent and the creator of all things that exist apart from him, certainly has in it more objective reality than the ideas that represent finite substances. ( A T VII40: C S M II28) He writes in the Second Replies that 'Existence is contained in the idea or concept of every single thing, since we cannot conceive of anything except as existing. Possible or contingent existence is contained in the concept of a limited thing, whereas necessary and perfect existence is contained in the concept of a supremely perfect being' ( A T VII166: C S M II117). His general position seems to be that we think of-substances by grasping ideas of them, and he never indicates that his thoughts about himself are an exception. What then is the idea by which Descartes thinks of himself when he knows for certain that he thinks and exists? It cannot be a concept in which he conceives of himself relative to some of his non-mental traits, for he doubts whether he has any such traits in the Second Meditation. It might be a concept in which he conceives of himself relative to some of his mental traits. Descartes sometimes writes as though he conceives of himself in this way. But immediately I noticed that while I was trying thus to think everything false, it was necessary that I, who was thinking this, was something. ( Discourse ( A T VI32: C S M I127); my emphasis) -73[I]t is easy for us to suppose that there is no God and no heaven, and that there are no bodies, and even that we ourselves have no hands or feet, or indeed any body at all. But we cannot for all that suppose that we, who are

having such thoughts, are nothing. ( Principles, part i, art. 7 ( A T VIIIA 7: C S M I194-5)) Descartes describes himself relative to his thoughts, but his point is unclear. He may be saying that his concept of himself is the concept of a thing with these thoughts. He may be giving us another version of the dictum 'I am thinking, therefore I am' by telling us that he has some thoughts and the fact that he has them entails that he exists. Descartes is in trouble if he believes that he individuates himself relative to his thoughts. To begin with, his explanation of why he is certain of his thought and existence will need to be revised. He does not just 'intuit' that he thinks and immediately infer that he exists. He first discovers a mental attribute, determines that it is a thought, decides that one and only one thing has it, and then concludes, 'I think', or, more properly, 'The thing with this thinks', where 'this' refers to the thought. Only then does he immediately infer 'I exist', or, more properly, 'The thing with this exists'. 21 The view that Descartes individuates himself relative to his thoughts is open to serious objections. Suppose that he considers one of his ideas, does not yet know whether he has produced it or whether God has produced it in him, and decides that whatever has produced it is perfect. Suppose too that he is the source of the idea. Clearly, the thought that Descartes would express by 'The thing that produced this is perfect', where 'this' refers to the idea, is not the same as the one he would express by 'I am perfect'. He believes the former, but he may not believe the latter. The difference between the thoughts is that, although Descartes thinks of himself in each thought--he is the referent of both 'the thing that produced this' and of 'I'--In the first thought he only thinks of himself and in the second he thinks of himself as himself. This difference between the two thoughts is lost if we analyse the thought Descartes would express by 'I am perfect' as ____________________ 21 Note the difference between this account of the process by which Descartes gains certainty of his existence and the position considered earlier according to which Descartes reasons: 'Thought is taking place, every observed quality exists in some substance; therefore, some substance thinks.' Descartes would express different propositions by 'Thought is taking place' and 'The thing that has this thinks', where 'this' refers to a thought of which he is immediately aware; the former proposition is not about him in particular, while the latter is. See Van Cleve, "'Conceivability and the Cartesian Argument for Dualism'", for an interpretation according to which Descartes individuates himself relative to his thoughts. Zemach, "'De Se and Descartes'", also attributes that position to Descartes, and then revises it to address some contemporary issues about self-reference. -74the one he would express by 'The thing that has this is perfect', where 'this' again refers to the idea. The difference between Descartes' thinking of himself and his thinking of

himself as himself is surely not that between his thinking of himself as the cause of an idea, 'The thing that caused this is perfect', and his thinking of himself as the thing that has the idea, 'The thing that has this is perfect'. 22 A second objection is contained in Elizabeth Anscombe's question, 'How do I know that I am not ten thinkers thinking in unison?' 23 Suppose Descartes observes his pain and says to himself, 'I am in pain.' He also observes his sadness and says to himself, 'I am sad.' He then takes note of both beliefs and infers, 'I am in pain and sad.' His third belief is justified by the first two, but it is hard to see how that can be, if each belief involves his individuating himself relative to his ideas. His inference becomes 'The thing with this' is in pain, and the thing with thisb is sad, so the thing with thisc is in pain and sad.' The demonstratives refer to his pain (a), sadnes's (b), and the combination of his pain and sadness (c), respectively. He is justified in believing his conclusion on the basis of his two premisses only if he is justified in believing the additional premiss: the thing with thisa is identical to the thing with thisb. The additional premiss is not justified for him. He has no reason to believe that the subject of the one sensation is identical with the subject of the other, given that all he is aware of is the sensations themselves. In all, then, Descartes leaves us wondering how he thinks of himself when he forms his certain beliefs about his thought and existence. He seems committed to the view that he thinks of himself by conceiving some concept of himself. Yet, no adequate concept of him seems to be available. He does not think of himself through a concept that identifies him by his physical traits. He does not think of himself by one that identifies him relative to his mental traits. What is left? 24 Another problematic aspect of Descartes' claim to certainty of his thought and existence is raised by the question of just how far his certainty about his thought extends. Descartes tells us that he is certain of such claims about his mental activities as that he doubts and imagines; he tells us that he is certain about such claims about his particular mental contents as that he seems to see light, hear noise, and feel heat. Yet, exactly how far ____________________ 22 For more on this point, see Markie, Descartev's Gambit, esp. ch. 3. 23 Anscombe, "'First Person'", 58. 24 Descartes might say that the concept by which he thinks of himself is just the concept of being him. For a contemporary statement of this approach to self-reference, see R. Chisholm, Person and Object ( London: Allen & Unwin, 1986); see too my discussion in Descartes's Gambit. -75can he go? Can he be certain whether or not he is angry, depressed, jealous, or in love? He does not say. One way to fill this gap in Descartes' position is to return to a suggestion we considered earlier about why the Deceptive God Hypothesis fails to cast doubt on such beliefs as that

he seems to see light. The hypothesis fails to indicate how these beliefs might be false, because it is impossible for him to have the beliefs, for the hypothesis to be true and for the beliefs to be false. The beliefs are incorrigible. Descartes cannot believe that he seems to see light unless he actually seems to see light; having the belief in the experience includes having the experience. Descartes might take a similar approach to explaining the extent of his certainty: he is certain of those beliefs about his mental state that are incorrigible. If his beliefs about whether he is angry, depressed, jealous, in love, and so on are not incorrigible for him, then they are not certain for him. If Descartes' position is developed in this way, it is open to arguments that have been offered against the incorrigibility of even such mental state beliefs as his beliefs that he seems to see light and feel pain. In part, the issue concerns how we form our beliefs about our mental state. Do we always form them by an act of self-evident intuition, or can we also form them on the basis of an inductive inference from some other beliefs, just as our external world beliefs are formed? If they can be formed inductively, there is room for error, as critics of the incorrigible, such as Keith Lehrer, are quick to point out. One might believe one is having a sensation S, a pain for example, because one is having a-different sensation, S*, an itch for example, and one has mistaken S* for S, that is one has mistaken an itch for a pain. How could this happen? It might happen either because of some general belief, to wit, that itches are pains, which one has been led to believe by some authority, or one may simply be misled on this occasion because one has been told by some authority that one will experience a pain. In short, one might have some false belief which together with the sensation of an itch produces the belief that one is in pain. Beliefs about sensations can be inferential, and one can infer that one is in a conscious state that one is not in by inferring from some false belief that this is so. 25 The issue of the incorrigibility and certainty of mental state beliefs also takes us back to the question of the content of those beliefs. According to some critics of Descartes' position, these beliefs are corrigible and uncertain, because they involve the classification of an experience and the act of classification can be mistaken. A. J. Ayer puts the point in this way: The fact is that one cannot in language point to an object without describing it. If a sentence is to express a proposition, it cannot merely name a situation; it must say ____________________ 25 See K. Lehrer, "'Why not Scepticism'", Philosophical Forum, 2 ( 1971), 351-2, and also T. Parsons, "'Mistaking Sensations'", Philosophical Review, 79 ( 1970). -76-

something about it. And in describing a situation, one is not merely 'registering' a sense-content; one is classifying it in some way or other, and this means going beyond what is immediately given. 26 Ayer applies this observation about language even to beliefs about what we seem to experience. The content of our belief that we seem to see white is the proposition that we are having an experience similar in colour to others we, and perhaps others, have called 'white'. This classification of our present experience relative to others can be mistaken and, hence, is uncertain. [E]ven if we exclude all reference to other people, it is still possible to think of a situation which would lead me to suppose that my classification of a sense-content was mistaken. I might, for example, have discovered that whenever I sensed a sensecontent of a certain quality, I made some distinctive overt bodily movement; and I might on one occasion be presented with a sense-content which I asserted to be of that quality, and then fail to make the bodily reaction which I had come to associate with it. In such a case I should probably abandon the hypothesis that sensecontents of that quality always called out in me the bodily reaction in question. But I should not, logically, be obliged to abandon it. If I found it more convenient, I could save this hypothesis by assuming that I really did make the reaction, although I did not notice it, or, alternatively, that the sense-content did not have the quality I asserted it to have. The fact that this course is a possible one, that it involves no logical contradiction, proves that a proposition which describes the quality of a presented sensecontent may as legitimately be doubted as any other empirical proposition. 27

It is unclear how Descartes would respond to these arguments. He might avoid Lehrer's by limiting his claim to certainty to those mental state beliefs that are intuited rather than inductively inferred from some evidence. To meet Ayer's, he might reject the initial claim that each mental state belief involves classifying an experience; he will then have to give an alternative account of the content of those beliefs. In all, then, Descartes' account of his certainty about his thought and existence leaves a number of important questions unanswered. We are left wondering what clear and distinct perception is, how he conceives of himself, which of his beliefs about particular mental states resist every reason for doubt, what he takes the content of those beliefs to be, and how he would defend their incorrigibility, assuming that that is a partial source of their certainty. Descartes says just enough to raise these important philosophical issues; that he raises them is part of what makes his position interesting and important. In conclusion, we have examined the role that Descartes' claim to ____________________ 26 A. J. Ayer, Language Truth and Logic ( New York: Dover, 1986), 91.

27

Ibid. 92-3. -77-

certainty of his thought and existence plays in his philosophy and the interpretative and philosophical issues that are raised by his claim. We have settled the main interpretative issues. The philosophical issues remain as part of Descartes' legacy to us. 28 ____________________ 28 I am indebted to John Cottingham and Margaret Wilson for their written comments on an earlier draft of this chapter. A version of this chapter was presented to the philosophy department at St Mary's College, Maryland; participants in the discussion, especially Reg Savage, made several helpful comments. -78III CLEARNESS AND DISTINCTNESS IN DESCARTES ALAN GEWIRTH I Descartes' general rule that 'whatever is clearly and distinctly perceived is true' has traditionally been criticized on two closely related grounds. As Leibniz, for example, puts it, clearness and distinctness are of no value as criteria of truth unless we have criteria of clearness and distinctness; but Descartes gives none. And consequently, the standards of judgement which the rule in fact evokes are purely subjective and psychological. There must hence be set up analytic, logical 'marks' by means of which it can infallibly and without arbitrariness be recognized whether any ideas or propositions are or are not clear and distinct. 1 In a general sense, these criticisms had been laid before Descartes himself. In the Fifth Objections Gassendi writes: Why are there so many and so varied opinions among men? Each one thinks that he perceives clearly and distinctly the opinion which he defends. And lest you say that many are either being partisan or pretending, I want you to notice that there are those who even face death for the sake of the opinions which they hold, even though they see others facing it for the sake of the opposite opinions. Hence, Gassendi concludes, Descartes' main concern should have been to 'propound a method which should direct us and teach us when we are deceived and when not deceived in thinking that we clearly and distinctly perceive something.' 2 In the face of

this objection, Descartes' reply is significant. The instance of men who face death on behalf of their opinions, he writes, ____________________ Alan Gewirth, "'Clearness and Distinction in Descartes'" from Philosophy, Vol. 13 ( 1943), 17-36, by permission of the author and the publisher, Cambridge University Press. 1 Cf. L. Coutourat, La Logique de Leibniz ( Paris: Alcan, 1901 [repr. Hildesheim: Olms, 1961]), 196, 202-3, with texts there cited. For a more recent version of this criticism, Cf. C. S. Peirce , Collected Papers ( Cambridge, Mass.: Harvard University Press, 1931-60), v. 391. 2 Fifth Objections ( A T VII278, 318 [ C S M II194, 221]). -79proves nothing . . . because it can never be proved that they clearly and distinctly perceive that which they pertinaciously affirm. And as for what you then add, that the concern must be not so much with the truth of the rule as with a method to discern whether or not we are deceived when we think that we clearly perceive something, I do not deny this; but I contend that this very thing has been accurately set forth by me in its place, where first I discarded all prejudices, and then enumerated all the principal ideas and distinguished those which were clear from the obscure or confused. 3 Descartes is here referring, of course, to the procedure followed in the first three Meditations. But before examining the relevant details of that procedure, it should be noted what he considers to be the general character of the 'proof' of clearness and distinctness such as he mentioned to Gassendi. Whereas Leibniz demands that the criterion be 'palpable', 'mechanical', and lacking in 'even the least difficulty', and declares that 'there is no need for prolix discussions concerning our prejudices', 4 Descartes states already in the Discourse that 'there is some difficulty in noting well what are the things which we conceive distinctly', 5 and tells Gassendi that 'I do not believe that those who are so little concerned with the uprooting of prejudices that they complain that I have not spoken of them "simply and in few words", will readily perceive the method whereby we can distinguish that which is really perceived clearly from that which is only thought to be clearly perceived'. 6 The setting forth of criteria of clearness and distinctness, then, will for Descartes necessarily involve a psychological discipline. Among his basic comments concerning formal logic and the 'synthetic' method of demonstration, both of which he opposed to his own method, were that the former permits the mind to 'go on a holiday from the evident and attentive consideration of the inference itself', 7 and the latter 'wrests assent from the reader, no matter how unwilling and pertinacious he may be'. 8 Both of these methods, in their formal character, thus realize Leibniz's ideal, but only, Descartes holds, at the ____________________ 3 Fifth Replies ( A T VII361-2 [ C S M II250]). (Italics here and in all other quotations

are mine.) An anonymous exponent of Gassendi took up the debate at this point, 'denying' that Descartes had set forth a method for distinguishing the really from the apparently clear and distinct (['Hyperaspistes'] to Descartes, July 1641 ( A T III402)). Unfortunately, however, his examples of men firmly convinced of the clearness and distinctness of their perceptions, and nevertheless in contradiction to one another, were all taken from theological controversy, so that Descartes was able to say in rebuttal merely that 'reply to them would be very easy for one who distinguishes the light of faith from the natural light, and sets the former before the latter' (to ['Hyperaspistes'], Aug. 1641 ( A T III426 [ C S M K191])). 4 Coutourat, La Logique de Leibniz, 100 nn. 2 and 3, 203 n. 2. 5 Discourse, part iv ( A T VI33 [ C S M I127]). 6 Fifth Replies ( A T VII379 [ C S M II260]). 7 Regulae, Rule IV ( AT X405-6 [ C S M I36]). 8 Second Replies ( A T VII156 [ C S M II111]). -80expense of losing all heuristic value. His own method, on the other hand, is dedicated not to exhibiting the formal interrelations of ideas and propositions, but to research into subject-matters the truth concerning which is not yet known to the mind, so that the necessity which it achieves cannot be a merely formal one but must derive from the impact of the subjectmatter itself upon the mind. Hence, even when the results of the method are set down in writing, but exhibiting the same procedures as were followed in the original investigation, the method 'has nothing whereby it may impel an inattentive or unwilling reader to believe: for if the least bit of what it propounds be not attended to, the necessity of its conclusions is not apparent'. 9 When this consideration is put within the framework of a theory of representative perception, where only ideas can be directly perceived by the mind, the reason for the psychological orientation of Descartes' method becomes evident. Yet, if Descartes' criteria of clearness and distinctness will not be logical in the formal, 'mechanical' sense demanded by Leibniz, neither will they be so radically psychological as to be divorced from all logical considerations. At least there is a difference, Descartes insists, between the evidentness of clear and distinct perception and the 'precipitation', 'pertinacity', and 'persuasion' of minds which cannot prove what they assert, or whose assertions vary according to arbitrary whim. 10 And the accomplishment of the difference comes through a certain prudence acquired by habituation: 'only the prudent correctly distinguish between that which is clearly and distinctly perceived, and that which only seems or appears so'; 11 for 'there are few who are accustomed to clear and distinct perceptions'. 12 This difference is accentuated from the side of the subject-matter itself; for if Descartes insists upon the ease and infallibility of the deduction prescribed by his method, 13 he also warns that the deduction must first be 'prepared for' and 'discovered', and the order of the deduction 'excogitated', 14 so that the clearness and distinctness exhibited by the process will have had to meet the tests presented by these preliminary operations upon the specific problem in hand.

____________________ 10 Cf. Discourse, part ii ( A T VI18 [ C S M I120]); Second and Third Replies ( A T VII146, 192 [ C S M II104, 135]); Regulae, Rule II ( A T X363 [ C S M I11]); letter to Regius, 24 May 1640 ( A T III65 [ C S M K147]); letter [to 'Hyperaspistes'], Aug. 1641 ( A T III430-1 [ C S M K194-5]); Principles, part i, arts. 50 and 68; Notae in programma [Comments on a Certain Broadsheet] ( A T VIIIB 352 [ C S M I299]). 11 Seventh Replies ( A T VII462 [ C S M II310]). 12 Ibid. ( A T VII511. [ C S M II348]). 13 Regulae, Rules II, III, IV, XII ( A T X365, 369, 372, 429 [ C S M I12, 14-15, 16, 50-1]). 14 Regulae, Rules IV, XII, XIV, XV ( A T X372, 428, 439-40, 451 [ C S M I, 16, 50, 57, 64]). 9 Ibid. ( A T VII155-6 [ C S M II110]). -81II But let us now examine the basis and development of these considerations, following the Cartesian order from the simple to the complex. The direct object of the mind's act of perception is for Descartes always an idea. 15 Ideas, however, have a double status: on the one hand, they are themselves existents, 'formal essences', modes of thought; on the other hand, they are significances, 'objective essences', representative of things other than themselves. 16 In order that ideas be susceptible of truth, they must be viewed not in the first way, or 'materially', but in the second way, or 'formally'. 17 The total cognitive situation for Descartes thus contains three factors: perceptive act, idea (the direct object or content perceived), and thing purportedly represented by the idea. From this characterization, it is at once apparent that the clearness and distinctness of an idea cannot directly consist in the relation of 'correspondence' or 'conformity' between the idea and the thing. For since 'we can have no knowledge of things otherwise than through the ideas which we conceive of them', 18 it follows that if clearness and distinctness were qualities consisting in some direct relation between the idea and the thing it purports to represent, the mind, not knowing the thing without the idea, could never know whether it had attained these qualities, so that the first precept of Descartes' method, to accept as true only what is clearly and distinctly perceived, could lead only to obscurity and indecision. Clearness and distinctness, then, cannot in their essential nature be the same as truth; it remains that they are qualities internal to ideas and perceptive acts. And yet the significatory character of ideas plays an important part in their being clear and distinct. It is precisely from the relation between these two aspects of ideas, the internal and the representative, that the difficulties which have traditionally been found in Descartes' doctrine arise. To understand the nature of those difficulties, we must first consider Descartes' general description of clearness and distinctness: A clear perception I call that which is present and open to the attending mind; just as we say that those things are clearly seen by us which, being present to the

____________________ 15 '. . . ostendo me nomen ideae sumere pro omni eo quod immediate a mente percipitur' [I make it clear that I am taking the word 'idea' to mean whatever is immediately perceived by the mind] (Third Replies ( A T VII181 [ C S M II127])). 16 Third Meditation ( A T VII40 ff. [ C S M II28 ff.]). Cf. A T VII37 [ C S M II25]: 'Quaedam ex his (cogitationibus) tanquam rerum imagines sunt, quibus solis proprie convenit ideae nomen' [Some of these thoughts are as it were the images of things, and it is only in these cases that the term 'idea' is strictly appropriate]. 17 Fourth Replies ( A T VII222 [ C S M II156]). 18 Letter to Gibieuf, 19 Jan. 1642 ( A T III476 [ C S M K202]). -82regarding eye, move it sufficiently strongly and openly. But that perception is distinct which is not only clear but is so precise and separated from all others that it plainly contains in itself nothing other than what is clear. 19 It will be noted that the word which Descartes here uses is 'perception', not 'idea'. But that he means by it the perceived rather than the perceiver, the object rather than the act, is apparent from the analogy which he draws with sight: 'perception' is given the same relative status as 'things seen by us'. Since, however, idea and perceptive act are correlative, the requirements indicated for either will to a certain extent belong to the other as well. In order to be clear, then, an idea must be (1) 'present' to the mind, (2) 'open' to the mind, and (3) the mind in turn must be 'attending' to the idea. Some of the problems of Descartes' doctrine are already apparent in this formulation. It might seem that the first and third requirements are really the same, that an idea cannot be 'present' to the mind at all unless the mind is 'attending' to it, for 'there can be no thought in us of which we are not conscious at that same moment at which it is in us'. 20 Moreover, it might be held that there cannot even be an idea unless these two requirements are fulfilled, since an idea is 'all that which is immediately perceived by the mind'. In these respects, difficulty might be found with Descartes' analogy of vision. For an object can be locally 'present' to the eye without the eye's 'regarding' the object; moreover, the objects of vision can presumably exist even while they are not being seen; but an idea is constituted both as to its very existence by the mind's act of perception, and as to its significance or representative character by precisely that about which the mind is thinking. The case is the same with Descartes' other requirement, that the idea be 'open' to the mind. The analogous phrase on the side of vision is that things 'move the eye sufficiently strongly and openly'. The 'sufficiency' to which Descartes is here referring obviously is the ability of the mind to recognize the object which is seen. But here again, does not the mind recognize any idea, since an idea is precisely that, and all that, of which the mind is at any moment thinking? At least, such would seem to be the tenor of definitions like these: 'I call generally, by the name "idea", all that which is in our mind when we conceive a thing, in whatever manner we conceive it'; 21 and 'by the name "idea" I understand that

form of any thought, through the immediate perception of which I am conscious of that same thought'. 22 In the Nouveaux Essais, Leibniz protests against just this ____________________ 19 Principles, part i, art. 45. 20 Fourth Replies ( A T VII246 [ C S M II171]). 21 Letter to Mersenne, July 1641 ( A T III392-3 [ C S M K185]). 22 Second Replies, Definition II ( A T VII160 [ C S M II113]). -83definition of an idea when, to Philalethes' question, 'Is it not true that the idea is the object of thought?', he has Theophilus reply, I grant it, provided you add that it is an immediate internal object, and that this object is an expression of the nature or qualities of things. If the idea were the form of thought, it would arise or cease with the actual thoughts which correspond to it; but since it is their object, it can be anterior or posterior to the thoughts. 23 Now for Descartes it is also correct to say that ideas, at least those which are clear and distinct, represent the essences of things. But apart from the difference in the interpretation of this statement owing to fundamental metaphysical dissimilarities from Leibniz, the statement itself cannot be used by Descartes in a methodological context; i.e. he cannot consult the essences of things in order to specify the clearness of an idea, because things themselves cannot be directly perceived. It is only after the nature of clearness and distinctness has been determined, and the veracity of God demonstrated, that Descartes is able to assert that clear and distinct ideas are true. And this difference in the initial interpretation of the status of ideas also prevents a solution of our general problem along the lines given by Leibniz. The problem to which the above considerations have pointed is, of course, how can any idea whatsoever fail to be clear and distinct, i.e. how can clearness and distinctness in any way constitute a normative requirement for ideas when, on Descartes' formulation, every idea seems ipso facto clear and distinct? For if what has been said about ideas makes all ideas clear, then it also makes them distinct, since the mind perceives only that which it is in fact perceiving, and nothing 'other' than that. In the later discussion of the same problem by Locke and Leibniz, 24 the solution of the former, based upon the relation of ideas to their names, was amended by the latter to rest upon the relation of the idea to the thing it purports to represent: an idea is not clear if it does not permit recognition of its object. But in the Cartesian context whereby things in any case can be perceived only through ideas, this solution cannot be accepted, at least without serious qualification. III The difficulties which we have been canvassing arise out of the need to find in Descartes' doctrine distinctions, analogous to Locke's between idea and

____________________ 23 Leibniz, Nouveaux Essais, II, i, 2. 24 Locke, An Essay Concerning Human Understanding, II, xxix, 5-6; Leibniz, Nouveaux Essais, II, xxix, 5. For Kant's echo of the same difficulty, see Kritik der reinen Vernunft (Critique of Pure Reason), B414-15n. -84name, and Leibniz's between idea and thing, which will explain how clearness and distinctness can be normative qualities and yet capable of ascertainment in a methodological context, within the sphere of ideas and perceptive acts. The Cartesian distinctions are forthcoming when it is seen that the idea and the perceptive act are less simple than our above interpretations have indicated. Let us examine some instances where Descartes calls ideas obscure and confused; for the considerations on which he bases this characterization should help to reveal the nature of the opposite qualities as well. Sense perceptions are the best-known class of Descartes' obscure and confused ideas. Yet this is not the only characterization which he applies to them. They are called clear and distinct in two different contexts: (1) 'when they are viewed only as sensations or thoughts', 25 and (2) when they are used as 'signifying to the mind what is helpful or harmful to the composite of which it is a part'. 26 They are called obscure and confused, however, when viewed as representing the essence of material things outside the mind. 27 From this example we may infer that the clearness and distinctness, or obscurity and confusion, of an idea are neither intrinsic to the idea nor explicable in terms of a simple relation between idea and perceptive act: for the same ideas of sensation may be both clear and obscure, distinct and confused. It seems hardly possible to say that the ideas involved in these different instances are in no way the same; at least Descartes applies to all of them the common name 'sense perceptions'. There must, then, be some additional factor whose combination with the two already indicated gives to the idea or perceptive act one or another of these qualifications. This third factor is revealed in Descartes' significant expression that the sensations are 'viewed as' (spectantur ut). For the situation in which there enter clearness and distinctness involves not merely the passive apprehension of a directly perceived content, but also some interpretation with regard to that content. In each of the three instances given above of the perception of ideas of sensation, the same content was directly perceived; but in the first case the ideas (e.g. colours) were 'viewed' or interpreted simply as sensations, in the second case as having some biologically symbolic function, and in the third as being 'similar' or cognitively adequate to external material things. It would hence seem that it is in the relation of direct and interpretive contents or perceptions that the clearness and distinctness, or obscurity and confusion, of an idea are to be found. ____________________ 25 Principles, part i, art. 68; cf. Third Meditation ( A T VII35 [ C S M II24]); Regulae, Rule XII ( A T X423 [ C S M147]). 26 Sixth Meditation ( A T VII83 [ C S M II57]). 27 Principles, part i, arts. 66, 70; Sixth Meditation ( A T VII 83 [ C S M II57-8]).

-85The same conclusion emerges from consideration of other ideas which do not involve the sharp distinction between essential and non-essential representation found in Descartes' account of sense perception. Thus the ideas of the modes thought and extension, Descartes writes, can be clearly and distinctly understood if they be viewed not as substances, or things separated from other things, but only as modes of things. . . . But if, on the other hand, we were to consider them without the substances in which they are, by that very fact we would view them as subsisting things, and so we would confuse the ideas of a mode and of a substance. 28 Here again the same content is directly perceived, and it is from the relation of that content to the way in which it is 'viewed' or interpreted that clearness and distinctness, or confusion, result. Descartes' doctrine admits, however, of greater generality than is envisaged in these two examples. For in each of these examples, the direct content (sense qualities, or 'the modes thought and extension') is held constant, while the interpretative content varies. But Descartes also discusses clearness and distinctness in situations where the direct content varies while the interpretative content remains fixed. For example, not every idea which is interpreted to be representative of God is clear and distinct. The idea will have these qualities only if 'we do not put anything fictitious into it, but note those things alone which are really contained in it, and which we evidently perceive to pertain to the nature of the most perfect being'. 29 Among these 'real' attributes Descartes lists 'eternal, omniscient, omnipotent, source of all goodness and truth, creator of all things, and in sum having in himself all those things in which we can clearly note some perfection which is infinite, or terminated by no imperfection'. 30 On the other hand, the idolaters' 'idea of God', which in its direct content representatively includes, together with perfection, such attributes as vindictiveness and corporeality, is obscure and confused. 31 We can now see the sense in which Descartes' description of clear and distinct perception in the Principles is a normative characterization, and not one belonging to any idea whatsoever. For in every perceptual situation there is some content 'which is present and open to the attending mind' (e.g. sense qualities, or 'vindictive corporeal being'), but this content 'need not be an integral and complete perception of the thing', 32 i.e. what is directly perceived may be other, or less, than what the mind interprets ____________________ 28 Principles, part i, art. 64. 29 Ibid. art. 54. 30 Ibid. art. 22. 31 Second and Fourth Replies ( A T VII138-9, 223-4 [ C S M II99, 163]). 32 Principles, part i, art. 34.

-86(or wishes to interpret) 33 itself to be perceiving (e.g. 'matter', or 'God'). Thus the direct content which, in Descartes' statement, is required to be 'present and open to the attending mind' is not any content whatsoever (since every perceptual situation fulfils this condition), but rather that content which shall be 'integral and complete' in relation to the mind's interpretation of it. When the interpretative content is taken as basic, then, an idea is clear if its direct content (e.g. 'incorporeal', 'good') includes all that which is included in the interpretation ('God'), and distinct if the direct content includes nothing other than this. And similarly, when the direct content is taken as the point of reference, an idea is clear if its interpretative content (e.g. 'having biological utility') includes all that which is included in its direct content (sense qualities), and distinct if the interpretation includes nothing else. 34 In most general terms, then, the clearness and distinctness of an idea may be said to consist in the 'equality' 35 of its direct and interpretative contents. This equality is, of course, a logical rather than a narrowly quantitative relation. In the context of science, where the concern is with the essences of things, the connection between the two contents must be necessary. The minimum requirement for an idea to be clear, then, is that whichever content be taken as basic, the other include what in the Replies is called the 'formal nature' (ratio formalis) 36 of its object, and what in the Principles is called the object's 'leading property, which constitutes its nature and essence'. 37 Thus the equality between the square of the base and the square of the other two sides of a triangle 'is not clearly understood unless in a right triangle'; 38 i.e. the idea which is interpreted to be representative of that equality is not clear unless what is essential to such a proportion, its inherence in a right triangle, be included in the idea's direct content. Similarly, the minimum requirement for an idea to be distinct is that ____________________ 33 Thus Descartes writes that 'the ideas I have of heat and cold are so little clear and distinct that I cannot learn from them whether cold is only a privation of heat, or heat a privation of cold, or each is a real quality, or neither' (Third Meditation ( A T VII43-4 [ C S M II30])). It is hence in relation to such contemplated interpretation of their direct content, in which one wishes to 'learn from them' the nature of the qualities they represent, that these ideas are lacking in clearness and distinctness. On the other hand, an interpretation of them in terms of biological utility finds the same Ideas, i.e. the same direct contents, clear and distinct. 34 From this it can readily be seen that when the direct content is the basis of evaluation (as in Descartes' example of the clear but confused perception of pain in Principles, part i, art. 46), ideas are usually clear, but their distinctness may come into question; on the other hand, when the interpretative content is made the basis, ideas may often be lacking in clearness, but if they are clear, they will usually be distinct as well. 35 Regulae, Rule XIV ( A T X439-40 [ C S M157]). 36 First, Second, Third, and Fifth Replies ( A T VII113, 147, 175, 368 [ C S M II81, 105, 123-4, 253]). 37 Principles, part i, art. 53. 38 Fourth Replies ( A T VII225 [ C S M II159]).

-87nothing contradictory to the essence of its object be included in it; it is in this sense that Descartes defines a distinct idea as one which 'contains nothing other than what is clear'. Thus a right triangle 'cannot be distinctly understood if the proportion of equality between the squares of its sides and of its base be denied', 39 for then the contradictory of that proportion, an essential property of the triangle, would be contained in the idea of the triangle. Similarly, 'an atom can never be conceived distinctly, since the very signification of the word implies contradiction, viz. to be body and to be indivisible'; 40 and 'as for a true part of matter, the determinate quantity of space which it occupies is necessarily included in the distinct thought which one can have of it'; 41 i.e. if the idea in which a direct content is interpreted to represent a part of matter is to be clear and distinct, that direct content must representatively include the occupancy of a determinate quantity of space. IV But has not our discussion been confusing idea with judgement? It is perhaps a sign of the correctness of the above formulation that this same objection was made to Descartes himself by Arnauld. 42 For although Descartes insists that neither ideas nor perceptive acts, but only judgements, may incur error, 43 he holds that there is a 'material falsity in ideas, when they represent as real what is not real'. 44 It will be noted that just as perceptive acts 'view as', so ideas 'represent as': in each case the orientation is interpretative rather than consisting in simple apprehension. Now if ideas be viewed normatively, from the standpoint solely of what they 'really' represent, it is nonsense to speak of ideas not 'representing' their objects 'as' those objects are. Thus Descartes tells Gassendi that 'the idea represents the essence of the thing, and if something be added to or subtracted from it, it forthwith becomes the idea of a different thing'. 45 And from precisely the same standpoint Arnauld objects to Descartes: 'that idea of cold, which you call materially false, what does it exhibit to ____________________ 39 Fourth Replies ( A T VII225 [ C S M II158]). 40 Letter to Mersenne, 30 Sept. 1640 ( A T III191 [ C S M K154]); cf. Second Replies ( A T VII 152 [ C S M II108]); Conversation with Burman ( A T V160 [ C S M K343]). 41 Letter to Mersenne, 30 Sept. 1640 ( A T III192 [ C S M K154]). 42 Fourth Objections ( A T VII206 [ C S M II145]). 43 Third Meditation ( A T VII37 [ CSM II26]); Fourth Meditation ( AT VII56-8 [ C S M II 39-40]). 44 Third Meditation ( A T VII43 [ C S M II301). 45 Fifth Replies ( A T VII371 [ C S M II256]). -88your mind? A privation? Then it is a true idea. A positive being? Then it is not the idea of cold.' 46 Descartes' reply stresses that the issue involves not merely a normative

conception of ideas, but a mixture of direct content with interpretation: 'I cannot decide whether what it exhibits to me is something positive outside my sense, or not.' 47 Such interpretation does not, of course, make the idea or perceptive act judgemental; for judgement involves an act of will in addition to perception. Hence, any object of the understanding, no matter how complex, is an idea so long as there enters no volitional act. Now a proposition or inference is not of itself volitional; it may, and in this context it indeed must, be an object of perception in the same sense as the simplest idea. For if it is simply by the act of willing an idea or perception that the judgement comes into being, must not the idea or perception have previously been a proposition or group of propositions? If the idea were logically equivalent only to a term, the addition thereto of the act of will could in no way result in a judgement. What one affirms or denies is not a term but a proposition which, having previously been simply a perceptual interpretation of a significant content, is transformed into a judgement by the volitional act of affirmation or denial. That the interpretative perception is not a judgement is, indeed, indicated explicitly by Descartes in one of his letters: What men judge frequently disagrees with what they understand. . . . Thus, when they judge that space, which they call empty, is nothing, they none the less understand it as a positive thing. Thus, too, when they think that accidents are real, they represent them to themselves as substances, although they do not judge them to be substances. 48 To 'represent as' or to 'understand as' is hence different from judging; and the idea which results from such an interpretative operation takes the logical form of a proposition or inference. 49 ____________________ 46 Fourth Objections ( A T VII207 [ C S M II146]). 47 Fourth Replies ( A T VII234 [ CSM II164]). 48 To ['Hyperaspistes'], Aug. 1641 ( A T III430 [ C S M K 194 ]). 49 For a statement of indifference as to whether ideas are expressed as terms or as propositions, cf. letter to Mersenne, July 1641: 'Car, qu'elles (les idées) s'expriment par des noms ou par des propositions, ce n'est pas cela qui fait qu'elles appartiennent à l'esprit ou à l'imagination; les unes et les autres se peuvent s'exprimer de ces delix manières' ( A T III395). ['For it is not whether they are expressed by terms or by propositions which makes them belong to the mind or the imagination; they can both be expressed in either way' ( C S M K186).] In another letter to Mersenne ( 22 July 1641), Descartes writes: 'Je n'entends pas bien la question que vous me faites, savoir si nos idées s'expriment par un simple terme; car les paroles étant de l'invention des hommes, on peut toujours se servir d'une ou de plusieurs, pour expliquer une même chose. . .' ( A T III417). ['I do not understand your question whether our ideas are expressed by a simple term. Words are human inventions, so one can always use one or several to express the same thing . . .' ( C S M K187).]

-89The non-formal orientation of Descartes' method bears out these considerations. For even the 'simple natures' are described as 'propositions', 50 and of course intuition, which is the elementary act of perception, is stated to have as its object now propositions, now the simple natures and their necessary connections, now consequences or series of propositions. Both the simple natures and the act of intuition are by Descartes opposed primarily to judgements, in virtue of the volitional character of the latter, which may result in 'formal' falsity. But since ideas may be propositions, there is no absurdity in calling them 'materially' false. V Once an idea has what we have seen to be the minimum of clearness and distinctness, it can become clearer if, while it is still interpreted to be representative of the same object, its direct content comes to include additional attributes necessarily connected with the interpretative content. 'The more attributes we apprehend in the same thing or substance, the more clearly do we know it.' 51 And by the same token, the idea will become more distinct, for the richer its content, the more is it distinguished from what is other than it: 'A concept is not made more distinct by the fact that we comprehend fewer things in it, but only by the fact that those things which we do comprehend in it we accurately distinguish from all others.' 52 This variation in degree reveals another distinction, in addition to that between direct and interpretative contents, required to give clearness and distinctness a normative basis. This other distinction is between the explicit and implicit contents of an idea. We have seen that although a clear and distinct perception is 'integral and complete', it nevertheless can become more clear, i.e. even more 'complete'. The limit of such increase is an 'adequate' idea, in which are representatively 'contained absolutely all the properties which are in the thing known'. 53 Only God can have such adequate knowledge; the human mind is incapable of it, although it may have 'complete' knowledge. Thus man may have a complete and hence a clear and distinct idea of God although this idea is by no means adequate to God's infinite perfection, just as 'we do not doubt that one unlearned in ____________________ 50 Regulae, Rules V, VI, XI, XII ( A T X379, 383, 409, 410, 428 [ C S M I20, 22, 38, 39, 50]). 51 Principles, part i, art. 11. Cf. Second Meditation ( A T VII33 [ C S M II22]). 52 Principles, part i, art. 63. 53 Fourth Replies ( A T VII220 [ C S M II155]). Cf. Second Replies ( A T VII140, 152 [ C S M II 100, 108]). -90geometry has the idea of a whole triangle, when he understands it to be a figure comprehended by three lines, although many other things can by geometers be known

about that triangle, and be noted in the idea of it, of which he is ignorant'. 54 For a direct content not to 'contain absolutely all the properties which are in the thing known' is possible only because of the logical orientation of Descartes' view of ideas. Metaphysically, of course, an idea refers beyond itself to a thing which the mind can never perceive directly; but in denying adequacy to human cognition Descartes is not relying exclusively, or even primarily, upon this basic fact; rather, he is referring to what can be ascertained in ideas themselves. Now for a strictly psychological position, an idea is precisely and exhaustively that content of which the mind is at any time aware, so that (putting Descartes' definition of adequate knowledge in terms of ideas and perceptive acts) to speak of a perception which is not aware of everything contained in an idea is a contradiction in terms. For Descartes, on the other hand, a distinction is possible between those ideas which are 'fictitious', i.e. arbitrarily compounded by the mind itself, and those which the mind merely discovers without adding to their 'objective reality'. It still remains true that every idea not only depends upon the mind for its existence, but consists in that very content of which the mind is directly aware. But once thus constituted as a direct object of perception, an idea of the latter sort is, with regard to its further significance, an independent logical entity containing within itself a system of implications, of simple natures and their relations, which deduction may gradually reveal, and which indeed it is the task of science progressively to discover. Thus, to take again Descartes' favourite example of the idea of a triangle, when I imagine a triangle, although perhaps such a figure exists nowhere in the world outside my thought, and never has existed, yet it has forthwith a determinate nature, or essence, or form, immutable and eternal, which has not been made up (efficta) by me, and does not depend upon my mind; as is shown by the fact that various properties can be demonstrated of this triangle, viz. that its three angles are equal to two right angles, that its greatest side is subtended by the greatest angle, and the like, which willynilly I now clearly perceive, even if I have in no way thought of them previously when I imagined a triangle, so that they have not been made up by me. 55 Hence, even though the properties revealed by such deductive development have not previously been suspected by the mind, these operations do not basically 'augment' the idea, but merely 'render it more distinct and ____________________ 54 Fifth Replies ( A T VII368 [ C S M II254]). 55 Fifth Meditation ( A T VII64 [ C S M II44-5]). Cf. Regulae, Rule XII ( A T X422 [ C S M I46]); Conversation with Burman ( A T V151-2 [ C S M K337]). -91explicit, because all these properties must have been contained in that idea which was had at first'. 56

It is important to note, however, that this variation in the degree of clearness and distinctness is possible only because an idea for Descartes involves an interpretative as well as a direct content. The 'idea' which remains fixed and unchanged throughout the process in which properties previously unknown are discovered in 'it' is not merely the initial direct content, since this undergoes obvious increase; it is rather the interpretative content which, equated at the outset with the direct content, is found to be necessarily connected with the 'new' properties, since these are found to be necessarily connected with the initial direct content, so that the mind interprets each succeeding property to be representative of the same object. It can readily be seen how this interpretative perception is the initial methodological basis of Descartes' doctrine of substance and essence. 57 VI Thus far we have been discussing the clearness and distinctness of ideas simply in terms of the characteristics which the ideas themselves must possess. But we must now turn to consider just how these characteristics of ideas are to be ascertained, i.e. what is that 'method' of which Descartes spoke to Gassendi, whereby that which is really clear and distinct can be distinguished from that which is only thought to be so. In terms of the preceding discussion, this question has a special urgency. For on the one hand, Descartes has said that things can be known only through the ideas of them; on the other hand, in calling sense perceptions, or the idea of corporeality, obscure and confused when they are viewed as representing, respectively, the essence of material things and God, he seems to have assumed that the mind has some independent insight, apart from ideas purporting to represent them, into the 'real' essence of things. The question of how the clearness and distinctness of ideas are to be ascertained is thus basically the same as the question of how the mind, within the methodological context of ideas and perceptive acts, can ascertain that its definitions of objects, which it employs as a basis for evaluating the clearness and distinctness of the equating of direct and interpretative contents, ____________________ 56 Fifth Replies A T VII371 [ C S M II256]; cf A T VII368 [ C S M253-4]. 57 Cf. Second Replies, Definition II ( A T VII161 [ C S M II113]), and especially Fourth Replies ( A T VII222 [ C S M II156-7]), where Descartes shows that substance is that which is interpreted to be the subject of directly perceived ideas. -92are real and not merely nominal or arbitrary. This problem, it must be emphasized, is different from the ultimate metaphysical question of whether clear and distinct ideas are true. The divine guarantee enters only to give final sanction to the decisions already arrived at by methodological means. These means are set forth in the Rules and in the second part of the Discourse. It will be recalled that in telling Gassendi of his 'method', Descartes emphasized first the removal of 'prejudices'. Now a prejudice is a judgement made prior to clear and distinct perception of the content which is judged. 58 Does this mean that clearness and distinctness must be

employed to attain these very qualities? The answer is given by the distinction between the perceptive act and the idea or content perceived. For it is possible that direct and interpretative contents be 'equal' or necessarily connected, while the full perception of the grounds for this be lacking. This would be the case if the idea were not 'open' to the mind, i.e. if its elements and the relations between them were not revealed to the extent required to justify, or to permit the deduction of, this necessary connection. In this perceptual, as against the preceding logical, sense, an idea is obscure and confused not because of an actual inequality between its direct and interpretative contents, but 'only because there is contained in it something which is unknown'. 59 Indeed, it is from this latter obscurity and confusion that ideas become contradictory, i.e. confused in the logical sense. 60 To make an idea 'open' to the mind, to render known its elements and to see how the necessary connection of its direct and interpretative contents follows from them, is thus the purpose of Descartes' basic methodological precept whereby 'we reduce involved and obscure propositions step by step to simpler ones', until we come to an 'intuition of the simplest ones of all'. 61 The ultimate conceptual elements of ideas are called by Descartes 'simple natures'. They include, on the side of material things, such concepts as 'figure', 'extension', 'motion', and on the side of mental phenomena, 'cognition', 'doubt', 'ignorance', 'volition'. Unlike composite ideas, in which it is possible to discriminate from one another not only direct and interpretative contents, but also various parts of the direct content, the necessity of whose connection with one another in the idea is not ____________________ 58 Cf. Principles, part i, art. 27. 59 Second Replies ( A T VII147 [ C S M II105]). 60 Second Replies ( A T VII152 [ C S M II108]). 61 Regitlae, Rule V ( A T X379 [ C S M I20]). The de omnibus dubitandum ['one must doubt everything'] of the First Meditation is just such a systematic reduction of ideas received 'from the senses or through the senses' to the thoughts which are their elements. -93self-evident, the simple natures cannot be misinterpreted, for it is impossible to discriminate in them a direct and an interpretative content. To think of these simples at all is to think of them completely, and hence clearly; similarly, their very simplicity makes it difficult for the mind to confuse them with, i.e. interpret them as, anything 'other' than themselves, so that they are perceived distinctly as well. 62 Since all composite ideas involve these simples, it follows that to attain a clear and distinct perception of any composite idea requires the reduction of the idea to these self-evident elements and then the perception of the precise way in which they are combined in a necessary nexus to form the idea originally in question. This reduction does not, of course, result in an 'adequate' idea. For in order to be clear and distinct, the idea must in its direct content represent only the 'formal nature' of the

object which it is interpreted as representing, but not the object or 'matter' itself--the idea of infinity, for example, must not itself be infinite, just as the idea of obscurity need not be obscure 63 --so that the reduction will attain not the total system of implications in which an idea is involved, but those elements which are 'sufficient' 64 to establish (or refute) the necessary connection of the idea's direct and interpretative contents. The various 'relations or proportions' in which the proposition was initially involved will hence have been so 'reduced' that the equality between its direct and interpretative contents will itself be 'clearly seen'. 65 The reduction attains this perceptual clearness and distinctness because it enables the mind, within the methodological context of ideas and perceptive acts, to 'distinguish that which pertains to the true and immutable essence of a thing, from that which is attributed to it only through a fiction of the understanding'. 66 The contents of a fictitious idea, such as that of a winged horse or of a triangle inscribed in a square, can be clearly and distinctly conceived in separation from one another; i.e. even if one part be explicitly denied of the other, each can still be fully conceived without in any way contradicting the significance which it presents to the mind. An idea will be representative of a true and immutable nature, then, if the connection of its contents ____________________ 62 Regulae, Rule XII ( A T X418-20, 422 [ C S M I44-5, 46]). The further refinements which Descartes makes on this doctrine of the simple or primitive 'notions' ( Principles, part i, arts. 47ff., and letter to Elizabeth, 21 May 1643 ( A T III663 [ C S M K217])), since they involve essentially the same principle as the discussion of the Regulae, although in a broader context, are here passed over. 63 First and Second Replies ( A T VII113, 147-8 [ C S M II81, 105]). 64 Compare Second and Fourth Replies ( A T VII140, 152, 221 [ C S M II100, 108, 156]). 65 Regulae, Rule XIV ( A T X440 [ C S M I57-8]). 66 First Replies ( A T VII116 [ C S M II83]). -94is necessary, not contingent, so that they cannot be sundered by way of a 'real' distinction: 'although one can think of the one without paying any attention to the other, one cannot, however, deny it of that other when one thinks of both.' 67 Those ideas which do not contain true and immutable natures, but only fictitious ones compounded by the understanding, can be divided by that same understanding not only by abstraction, but by a clear and distinct operation, so that those which the understanding cannot thus divide have undoubtedly not been compounded by itself. 68 If, then, the mind finds that it cannot deny the parts of ideas of one another and still be true to the meaning which each part directly presented to it, such ideas are representative of objects which are essential natures independent of the mind for their being what they are. The direct contents of these ideas are seen, as a result of this 'method of difference',

to represent the essence, or at least part of the essence, of the objects which they are interpreted as representing, so that the direct and interpretative contents are equal to one another. It will have been noted, however, that the criterion of 'immutability' stated in the second quotation given above was the inability of ideas to be divided 'by a clear and distinct operation', which Descartes goes on to specify as meaning 'rightly understanding what I say'. Here again, as previously in the discussion of the discarding of 'prejudices', the logical sense of clearness and distinctness, as characteristics of ideas, is stated to be ascertained by clearness and distinctness as characteristics of the perceptive act which, resulting from the reductive process, finds necessity of connection within the ideas. An excellent example of this process is the famous operation upon the wax in the Second Meditation. It is required to ascertain the essential nature of matter. A direct content which is interpreted to be representative of matter is chosen, consisting in the various sense qualities which the wax initially presented. Then these qualities are put through a series of reductions to see whether they and the interpretative content 'so depend upon one another, that one can in no way be changed while the other remains unchanged'. 69 It is found that the sense qualities are changed, although the wax still 'remains the same'. Hence, those qualities are essentially 'other' than the wax, and the interpretative perception of them as representing the essence of the wax was not distinct. This 'identity' of the wax throughout the changing of the direct content, ____________________ 67 Letter to Gibieuf, 19 Jan. 1642 ( AT III475 [ C S M K202]). Cf. Sixth Replies ( A T VII443 [ C S M II298-9]). 68 First Replies ( A T VII117 [ C S M II83-4]). 69 Regulae, Rule XII ( AT X429 [ C S M I51]). -95which is greatly emphasized by Descartes, 70 can be understood only through the interpretative aspect of his doctrine of ideas. It is because the interpretative perception is held constant that the reductive process can be viewed, in the methodological context, as going from accidental to essential attributes of the same object, and not from one set of ideas to another set wholly unrelated thereto. The process consists, then, in gradually divesting the direct content interpreted as representing a certain object of the 'forms external' to that object, i.e. of those qualities with whose denial the object can still be conceived, so that there is no necessary connection between the object and those qualities. The end of the process comes when a direct content is attained which survives every reductive device, remaining so long as the object can be conceived, and without which the object can no longer be conceived. This direct content is hence necessarily connected with the interpretative content whereby the mind thinks of the object in question, and constitutes the essential definition of that object. And only in virtue of such a reductive process is the resultant perceptive act clear (as attaining the essence of the object, in that the direct content which is actually representative of the essence of the object which it is interpreted as representing has been made 'present and open to the

attending mind') and distinct (as excluding everything 'other than' what is essential). The wax is thus ascertained to consist essentially of extension and mobility, not of any peculiar colours, sounds, and tastes; hence, at the conclusion of the reductive process whereby the sense qualities are removed, Descartes writes that the perception of the wax 'can be either imperfect and confused, as it was before, or clear and distinct, as it now is, in so far as I attend less or more to the things of which it consists'. 71 The physical manipulation which the procedure of the wax involves is the most graphic, but not the only, form which the reductive operation leading to this perceptual clearness and distinctness may take. The culmination of the process, Descartes points out, consists in an 'inspection of the mind alone'. The same operation in principle occurs when infinite perfection is found to be essential to God: ultimately it is the impossibility which the mind finds in conceiving the latter without the former that methodo____________________ 70 'Remanet adhuc eadem cera? Remanere fatendum est; nemo negat, nemo aliter putat. Quid erat igitur in ea quod distincte comprehendebatur? Certe nihil eorum quae sensibus attingebam; nam . . . mutata jam sunt: remanet cera' (Second Meditation ( AT VII30)). ['Does the same wax still remain? We must admit it does; no one denies this, no one thinks otherwise. So what was it in the wax that I understood with such distinctness? Evidently none of the features which I arrived at by means of the senses; for . . . all these have now altered--yet the wax remains' ( C S M II20).] Cf. Principles, part ii, art. 4. 71 Second Meditation ( A T VII31 [ C S M II21]). -96logically shows the idea equating them to be a real definition, representative of a 'true and immutable nature'. 72 These essential definitions consist, of course, not in genus and differentia but in various combinations of simple natures. 73 Once attained, they can then be used to test the clearness and distinctness of all other ideas which are interpreted to be representative of the objects defined. It is thus that the idolaters' 'idea of God', where the direct content includes 'vindictiveness' and corporeality', is shown to be confused; for both these qualities imply imperfection, the former because it is a tendency away from goodness, and hence toward non-being, the latter because whatever is corporeal is divisible, and 'it is a greater perfection not to be divided than to be divided', so that 'if you mean (by "God") only what is most perfect in the nature of body, this is not God'. 74 The procedure is now deductive, and consists in 'comparison' of the essential definition with other ideas, through as many intermediate steps as are necessary. 75 It is in this sense that Descartes mentioned to Gassendi, as the second part of his method of ascertaining what is really perceived clearly and distinctly, after the removal of prejudices by the process of reduction, that he 'enumerated all the principal ideas and distinguished those which were clear from those which were obscure or confused'; i.e. he set forth which ideas constituted the real definitions of God, the mind, material objects, and which not, so that deductive comparison could then ascertain

____________________ 72 Fifth Meditation ( A T VII67, lines 21-4 [ C S M II46, bottom four lines]). It is because Descartes holds that he has shown in this way that the idea of God as an infinitely perfect being represents a true or real essence, and is not merely a nominal definition, that he maintains that his ontological argument, unlike that of St Anselm as reported by Thomas Aquinas, is neither verbal nor fallacious. Cf. First Replies ( A T VII115-19 [ C S M II82-5]), and letter to Mersenne, 16 June 1641 ( A T III383 [ C S M K183-4]). 73 Definitions by genus and differentia, as viewed in the Aristotelian tradition, presuppose essentially different kinds of things, so that to define any given species requires showing the genus to which the species belongs, and then the form which differentiates it from other species of the genus and constitutes it what it is. Such definitions are hence basically additive: the differentia adds a characteristic which is essentially other than that of any other species. The Cartesian definition by simple natures, on the other hand, adds no 'new' characteristic, but consists in a different arrangement or 'mixture' of the same basic elements possessed by all other things; such definitions are hence 'mechanistic' (cf. the example of a right triangle to exhibit what is meant by 'species', Principles, part i, art. 59). The only point at which the Aristotelian differentia would enter into the Cartesian scheme is in the initial separation of the 'material' from the 'Intellectual' natures. There is, however, a certain analogy between the two methods; just as for Aristotle, if the definition of a species states the genus alone, it has not differentiated the species in question from other species, i.e. the definition is not 'distinct', but if it states the differentia, it also includes the genus (cf. Metaphysics, VII 12, 1037b, 29 ff.); so for Descartes if a definitory idea contains that mixture of simple natures which represents the 'formal nature' of its subject, but other ingredients not necessarily connected therewith, the idea is clear but not distinct, but If it is distinct, it is also clear (cf. Principles, part i, art. 46). 74 Second Replies ( A T VII138 [ C S M II991). Cf. Fifth Meditation ( AT VII67-8 [ C S M II 46-7]); Conversation with Burman ( A T V161 [ C S M K344]). 75 Cf. Regulae, Rule XIV ( A T X439 [ C S M I57]). -97what further ideas interpreted as representative of these objects would be clear and distinct. The discovery of what ideas represent the essences of things is thus effected by operations internal to ideas and perceptive acts. It is precisely because ideas are viewed by Descartes as being representative of things which are not themselves perceived directly, that clearness and distinctness, characteristics emerging within these operations, are not in their immediate nature the same as the transcendental relation of conformity between ideas and things which Descartes calls 'truth', 76 so that the methodological orientation must be supplemented by a metaphysical one, culminating in the divine guarantee. This addition, which later rationalists, like Spinoza, were to find superfluous, indicates from the side of things the circumspection which similarly controls Descartes' internal approach to clearness and distinctness themselves. It is by subjecting the mind to

the coerciveness exercised by the internal meanings of ideas, rather than by subjecting those ideas to the whims of a dogmatic mind, that these qualities are ascertained. That this is Descartes' real emphasis has been overlooked by the traditional interpretation of his method because of the opposition, insisted upon in that interpretation, between the rationalistic and the experimental approaches to science. Yet both the formulae themselves of clearness and distinctness and Descartes' actual application of them show that these qualities pertain as much to the experimental methods of agreement and difference which he employed in his physical operations as to the operations dealing with the 'intellectual' natures in which sense experience plays no part. The purpose of physical science for Descartes is to explain the phenomena of material nature by a deductive system of ideas whose principles are the simple concepts of extension, figure, and motion, which represent the essence of all bodies. The clearness and distinctness of the ideas or propositions of physics will require necessity of connection, then, not only within the deductive system itself but also between each explanatory proposition and the class of phenomena which it is designed to explain: the phenomena, or effects, or direct contents, and the explanations, or causes, or interpretative contents, will similarly have to be 'proved so to depend upon one another that the one can in no way be changed while the other remains unchanged'. The ascertainment of this mutual dependence, in turn, involves recourse to 'crucial' experiments whereby it is determined which of various possible explanations, each a certain 'composition' or 'mixture' of the material simple natures logically compatible with the prior ____________________ 76 Letter to Mersenne, 16 Oct. 1639 ( A T II597 [ C S M K139]). -98propositions of the system, is in fact necessarily connected with the phenomenon in question: 'in order to be able to ascertain which of these causes is the true one, it is necessary to consider experiments which cannot agree both with the one and with the other'; 77 or, in another statement, 'to seek out some experiments which will be such that their outcome will not be the same if it is in one of these ways that the effect is to be explained, as if it is in another'. 78 A clear and distinct idea in physics, then, especially when detailed problems are being dealt with, must be an experimentally verified idea: a proposition purporting to be explanatory of a given phenomenon must be 'equal' to that phenomenon and not 'other' than it, but the ascertainment of this requires recourse to experiment, because science must be of the actually existing world, whose objects, as existing, the senses alone reveal determinately to the mind, whereas the pure understanding would be unable to differentiate which of its ideas represented actually existing material things and which only possible ones. Nor is this appeal to phenomena and experiment a departure from the intra-ideational orientation of Descartes' method. For the phenomena, as dealt with in the method, are themselves 'ideas' directly perceived through the senses, and representative, although confusedly, of the existing material things which the scientific ideas represent clearly and distinctly. A scientific idea, then,

finds its place as clear and distinct within the system of physics in virtue of a triple necessary connection: (1) with the basic principles of extension, motion, and figure, which, as the example of the wax showed, represent the essence of material things; (2) with the ideas preceding it in the deductive system; and (3) with the particular phenomena physically caused by and hence representative of the particular material existents which are to be explained. 79 The basic emphasis of the method of clearness and distinctness is thus upon the testing and confirmation of scientific ideas: the principle is that 'for right philosophizing . . . the greatest care must be taken not to admit anything as true which we cannot prove to be true'. 80 And the factor determining which ideas are representative of the essences of various objects, and hence clear and distinct, consists in that which the ideas themselves compel the mind to perceive after it has reduced them to their elements and tried to separate and combine them in various ways. There is, indeed, a psychological aspect of the method also, as is shown, for ____________________ 77 Description du corps humain, sect. xviii ( A T XI242 [ Description of the Human Bod), C S M I317]). 78 Discourse, part vi ( A T VI65 [ C S M I144]). 79 Cf. Second Replies ( A T VII135 [ C S M II97]). 80 Fifth Replies ( A T VII354 [ C S M II245]). -99example, in Descartes' description of intuition as 'a pure and attentive mind's conception, so facile and distinct that there remains no doubt concerning that which we understand'. 81 But this facility and indubitableness are regarded by him as effects, rather than as causes, of clearness and distinctness in the logical and perceptual senses, as culminating the process whereby the mind recognizes the contents of ideas to be of such sort that it is unable to perceive them in any way other than the connection before it. It is in this sense, not in a narrowly psychological one, that Descartes can say, as Burman reports, that 'whether or not perceptions are clear we shall know best from our own consciousness, and for this it is of the greatest help to know all those elucidations of things which the author set forth in the first book of the Principles'. 82 The method of clearness and distinctness as Descartes develops it hence exhibits no merely psychological or subjective criteria, but rather logical and perceptual ones which the mind is to use in order to ascertain the conformity of its thought with a world external to it. ____________________ 81 Regulae, Rule III ( A T X368 [ C S M I14]). 82 Conversation with Burman ( A T V160 [ C S M K344]). Cf. Principles, part i, arts. 47 ff. -100-

IV FOUNDATIONALISM, EPISTEMIC PRINCIPLES, AND THE CARTESIAN CIRCLE JAMES VAN CLEVE The problem of the Cartesian Circle is sometimes treated as though it were merely an exercise for scholars: Descartes fell into it, and their job is to get him out of it. But more is at stake than extricating Descartes. In its generalized form, the Cartesian Circle is none other than the Problem of the Criterion, a problem that any epistemology must face. Moreover, to solve the problem of the Circle one must answer questions about epistemic principles that are pivotal in contemporary debates between foundationalists and coherentists. There is reason to hope, therefore, that by examining Descartes' problem, we can throw light on problems of our own.This paper is divided into two parts. In Part One I examine solutions to the problem of the Circle that are possible within Descartes' own framework. In Part Two I show how what we learn in Part One may be used to resolve some contemporary disputes that hinge on the status of epistemic principles. PART ONE I The problem of the Cartesian Circle arose for Descartes because he appeared to commit himself to each of the following propositions: 1. I can know (be certain) that (p) whatever I perceive clearly and distinctly is true only if I first know (am certain) that (q) God exists and is not a deceiver. ____________________ From Philosophical Review, 88 ( 1979), 55-91, Copyright © 1979 Cornell University, by permission of the publisher and the author. -1011.

I can know (be certain) that (q) God exists and is not a deceiver only if I first know (am certain) that (p) whatever I perceive clearly and distinctly is true.

Obviously, if (1) and (2) are both true, I can never be certain of either p or q. To be certain of either, I would already have to be certain of the other. Yet Descartes said he was certain of bothp and q. How can this be possible? 1 Any adequate solution to the problem of the Cartesian Circle will plainly have to deny either (1) or (2). In the next section I consider a famous solution that denies (1). II The solution I have in mind is the Memory Gambit, according to which God is called upon to guarantee not the truth of clear and distinct perceptions, but the accuracy of our memories. The most able recent defender of this solution is Willis Doney, who cites a number of passages that seem to show that this solution was Descartes' own. 2 In these passages Descartes says that if I remember 3 clearly and distinctly perceiving something that I do not now clearly and distinctly perceive, I can be certain of it if and only if I

know that God exists and is not a deceiver. He also says that an atheist can know theorems of geometry if he is clearly and distinctly perceiving them at the time, but warns that doubts may arise later that only knowledge of God's veracity can remove. Doney concludes that the function of God is to guarantee the accuracy of memory, 4 and that the atheist's plight is that in his ignorance of God's guarantee, he cannot be sure that he really did clearly and distinctly perceive what he remembers so perceiving. ____________________ 1 In presenting the problem this way, I follow Doney, 'The Cartesian Circle'. Cf. Arnauld 'only remaining scruple' in HR II92 [AT VII214: CSM II150]. The more generalized form would be this: how can I know any epistemic principles unless I first know some other propositions from which to derive them? But how can I know those other propositions unless I first know some epistemic principles? See Roderick M. Chisholm, The Problem of the Criterion, The Aquinas Lecture, 1973 ( Milwaukee: Marquette University Press, 1973). I shall not address myself to this problem explicitly, but what I would say about it will become clear in Part Two. 2 Doney, 'Cartesian Circle'. Passages often cited in this connection are HR I184; HR II38, 39, 114-15, and 245 [AT VII70, 140, 246, 428: CSM II48, 100, 171, 289]. 3 Throughout this essay I use 'remember' in the sense of 'ostensibly remember', so as not to ensure by definition that what is remembered is true. 4 Perhaps Doney does not wish to say that God guarantees that whatever we remember is true, but only that whatever we remember having perceived clearly and distinctly is something that we did perceive clearly and distinctly. -102Limited to what he can clearly and distinctly perceive at each moment, his knowledge will be 'meager and fugitive'.Although the Memory Gambit has some textual plausibility, Harry Frankfurt has convincingly argued that it is neither the solution Descartes intended, nor a very satisfying solution in its own right. 5 I shall not repeat his case here, but I do want to point out an alternative explanation of the passages that make the Memory Gambit tempting. Consider the following sequence of propositions: 1. I remember clearly and distinctly perceiving p. 2. So, I did clearly and distinctly perceive p. 3. So, p is true. Descartes says that the atheist cannot argue from (1) to (3). According to the Memory Gambit, this is because he cannot take the step from (1) to (2). But another possible explanation is that he cannot take the step from (2) to (3). And if this is what Descartes had in mind, then he must have felt that a divine guarantee for clear and distinct perception was needed after all. 6 III

I pass now to solutions that deny (2). Interesting solutions of this type have been offered by Alan Gewirth and Fred Feldman. Gewirth sums up his basic strategy as follows: ' Descartes's argument is not circular, for, while it is by the psychological certainty of clear and distinct perceptions that God's existence is proved, what God guarantees is the metaphysical certainty of such perceptions.' 7 Psychological certainty is a subjective affair, implying only an irresistible compulsion to believe. 8 Metaphysical certainty, on the other hand, is an objective affair, implying truth. 9 If we let 'certain' in (1) and (2) express metaphysical certainty, then Gewirth would deny (2). I do not need to have metaphysical certainty that clear and distinct perceptions are true ____________________ 5 Frankfurt, 'Memory and the Cartesian Circle'. 6 This point is well made by A. K. Stout in "'The Basis of Knowledge in Descartes'", Mind, 38 ( 1929), 330-42 and 458-72; repr. in Doney (ed.), Descartes, 91. See also Frankfurt, 'Memory and the Cartesian Circle'. 7 Gewirth, 'The Cartesian Circle', (emphasis added). I shall also refer to two other articles by Gewirth: 'The Cartesian Circle Reconsidered' and 'Descartes: Two Disputed Questions'. 8 Gewirth 'Cartesian Circle', 374. 'Clear and distinct perceptions are so coercive in their effect upon the mind that the mind cannot help assenting to them as true at the time it has such perceptions' (ibid. 383). 9 Ibid. 378, 394. -103before I prove God's existence; it suffices if clear and distinct perception gives me psychological certainty. 10 But once I am psychologically certain that God exists, I can use this fact to establish the metaphysical certainty of clear and distinct perceptions.The big question prompted by this approach, of course, is this: how can mere psychological certainty about God possibly give rise to metaphysical certainty about clear and distinct perceptions? Gewirth's answer to this question will emerge from the following reconstruction of his account of Descartes' programme: 11 1. I perceive clearly and distinctly that the premisses of Descartes' theological arguments are true, and that their conclusions follow from them. I thereby arrive at clear and distinct perception (and thus psychological certainty) that God exists and is no deceiver. 2. A proposition P is metaphysically certain if and only if there is no proposition R that is a reason for doubting P. 12 3. R is a reason for doubting P only if R is itself clearly and distinctly perceived (and thus something I am psychologically certain of). 13 4. The only reason for doubting the truth of clear and distinct perceptions is the hypothesis that God is a deceiver. 5. If I clearly and distinctly perceive that God exists and is no deceiver, then I do not clearly and distinctly perceive that God is a deceiver.

6. 7. 8.

I do not clearly and distinctly perceive that God is a deceiver. ((1) and (5)) The hypothesis that God is a deceiver is not a reason for doubting anything. ((3) and (6)) All clear and distinct perceptions are metaphysically certain. ( (2), (4), and (7))

In the beginning clear and distinct perceptions are only psychologically certain. But as soon as we have used them to prove that God exists and is ____________________ 10 Throughout the remainder of this essay I shall use 'clear and distinct perception' (singular) for the act or faculty of perceiving clearly and distinctly, and 'clear and distinct perceptions' (plural) for the propositional objects of such perception. That the objects of clear and distinct perception are always propositional is well argued by Frankfurt in Demons, Dreamers, and Madmen, ch. 12. 11 My sources for what follows are the first two Gewirth articles mentioned in n. 7. The steps I list are all present in the first article, but they are more explicitly set forth in the second. 12 By a 'reason to doubt' P Descartes does not mean a reason to disbelieve P (i.e. believe its negation), but a reason to withhold assent from P (provided you want to assent only to what is certain). 13 Gewirth says that R need only purport to be clearly and distinctly perceived. But since he later disqualifies propositions from being reasons to doubt, on the ground that they are not in fact clearly and distinctly perceived, he needs the stronger premiss here. -104no deceiver, they become metaphysically certain. 14 This is because by proving the existence and veracity of God we eliminate the only possible reason for doubting clear and distinct perceptions. But according to (2), if there is no reason to doubt something, it is metaphysically certain. In this way our original psychological certainties get parlayed into metaphysical certainties.As it stands, this argument is open to two serious objections. 15 In the first place, premiss (3) sets the requirements that must be satisfied by a reason to doubt absurdly high. Descartes himself certainly does not require that reasons to doubt be clearly and distinctly perceived. He says that the reasons to doubt he brings forth are 'very slight', 'metaphysical', and 'themselves doubtful'. 16 In the second place, premiss (5) seems to presuppose that we can never clearly and distinctly perceive each of two mutually inconsistent propositions. But this is a thesis one would rather see as a conclusion than as a premiss of the Cartesian enterprise. 17 Fortunately, however, these objections can both be avoided by a single change in the argument. We need only replace premiss (3) by the following premiss: 1. R is a reason for doubting P only if it is not the case that its negation, not-R, is clearly and distinctly perceived. (I shall abbreviate the consequent by 'R is not excluded by clear and distinct perceptions'.) Here we no longer require that reasons to doubt be upheld by clear and distinct perception, but only that they not be condemned by it. This enlarges the class of

permissible reasons to doubt, thus mitigating the objection to premiss (3). Moreover, we may now dispense altogether with the ____________________ 14 Gewirth says that the existence and veracity of God is the first metaphysically certain proposition ( 'Cartesian Circle', 394). But in fact the logic of his argument requires that all things we were previously psychologically certain of (because we perceived them clearly and distinctly) become metaphysically certain at one stroke the moment that proposition is established. 15 I pass over a third: why should the hypothesis of a deceiving God be the only reason for doubting clear and distinct perceptions? 16 HR I159; HR II277 [AT VII36, 474: CSM II25, 319]. 17 This point is made by Anthony Kenny in 'The Cartesian Circle and the Eternal Truths'. In 'Descartes: Two Disputed Questions', Gewirth makes the following reply: Descartes is entitled to the premiss that clear and distinct perceptions are mutually consistent, because consistency is an internal matter of relations among ideas, not an external matter of correspondence to fact. But this reply mistakenly assumes that Descartes' doubt is rooted in his representationalism, and arises only where there is a gap to be bridged between mental ideas and extra-mental facts. In fact, his doubt is far more radical. To see this, notice three things: (i) the propositions we can clearly and distinctly perceive are limited to propositions about what is going on in our own minds and propositions about the relations among abstract entities; (ii) Descartes does not hold that mental happenings and abstract entities are known only via representatives-here there is no representationalist gap; yet (iii) he still finds room to wonder whether clear and distinct perception is a guarantee of truth. -105objectionable premiss (5), since the conclusion, (8), follows from (1), (2), (3′), and (4) alone.There remains, however, an objection that is fatal to Gewirth's whole approach. The revised argument establishes only that clear and distinct perceptions are metaphysically certain in the sense jointly defined by premisses (2) and (3′). That sense amounts to this: a proposition is metaphysically certain if and only if every reason for doubting it is excluded by clear and distinct perceptions. Now it was assumed initially that clear and distinct perceptions are only psychologically certain. What is added when, at the end of the argument, we say that God's veracity and other things clearly and distinctly perceived are metaphysically certain? Just this: that we are psychologically certain not only of those propositions themselves, but also of the falsehood of every reason for doubting them. Thus, we have not advanced to a new kind of certainty at all. We have merely extended the class of psychological certainties. Descartes played for higher stakes. The certainty he sought was certainty in a sense entailing both maximal evidence and truth. Despite what Gewirth says, metaphysical certainty in his sense entails neither. 18 It remains at bottom a purely psychological notion.In the hope of obtaining a conclusion that is epistemologically more nourishing, let us turn to Feldman. 19 His reconstruction of Descartes is identical in structure with the one we obtained by revising Gewirth, but there is an important difference: he replaces the concept of psychological

certainty with the epistemic concept of practical certainty. Practical certainty is the sort of certainty involved in ordinary knowing of the justified-true-belief variety. Being practically certain of something, unlike being psychologically certain of it, entails having some justification for believing it. Feldman's argument may be set out as follows: 20 1. By means of Descartes' theological arguments, I attain practical certainty that God exists and is no deceiver. 2. A proposition P is metaphysically certain if and only if there is no proposition R that casts metaphysical doubt on P. 3. R casts metaphysical doubt on P only if it is not the case that its negation, not-R, is practically certain. ____________________ 18 Gewirth wants to say both that the absence of psychologically compelling reasons to doubt is sufficient for metaphysical certainty and that metaphysical certainty is sufficient for truth. But he cannot have it both ways. 19 Feldman, 'Epistemic Appraisal and the Cartesian Circle'. 20 I have recast his argument somewhat, but every premiss I attribute to him is logically equivalent to a premiss or definition he states in his article. I have also dropped references to persons and times. -1061. 2.

The only proposition that casts metaphysical doubt on any clear and distinct perception is the hypothesis that God is a deceiver. All clear and distinct perceptions are metaphysically certain.

((1)-(4)) The logic of this argument is the same as Gewirth's. Before we have proved the veracity of God, clear and distinct perceptions are only practically certain. They are not metaphysically certain, because doubt is cast on them by the hypothesis that God is a deceiver. But as soon as we become practically certain that God is not a deceiver, that hypothesis, by (3), is no longer eligible to cast doubt. By (4), however, it is the only candidate, so nothing any longer casts doubt on clear and distinct perceptions. Therefore, by (2), they are henceforth metaphysically certain. In light of (2) and (3), Feldman's conclusion is equivalent to the following: we are practically certain not only of clear and distinct perceptions, but also of the falsehood of every proposition that would cast doubt on them. This is an improvement over Gewirth's conclusion, since it implies that clear and distinct perceptions have something going for them epistemically. But my basic misgiving remains. Feldman's conclusion, like Gewirth's, provides no guarantee that clear and distinct perceptions are true. For a Cartesian this is not enough. 21 Gewirth and Feldman fall short of giving us what Descartes wanted for the same reason. They both define metaphysical certainty in terms of the absence of reasons to doubt. Then they make it very hard for anything to qualify as a reason to doubt. As a result, they

make it very easy for things to qualify as metaphysically certain. Their standards for certainty are thus set too low. Descartes, on the other hand, was much more liberal about what could count as a reason to doubt. For him, a proposition could function as a reason to doubt just so long as it was (in Feldman's terms) a 'metaphysical possibility'. 22 This makes his standards for certainty very high. In the next section I defend Descartes' high standards. ____________________ 21 Roderick Firth "'The Anatomy of Certainty'", Philosophical Review, 76 ( 1967), 3-27, divides senses of 'certain' into three classes: truth-evaluating senses, warrantevaluating senses, and test ability-evaluating senses. In terms of this scheme, Descartes' sense of 'certain' is both truth-evaluating and warrant-evaluating; Feldman's is warrant-evaluating but not truthevaluating; and Gewirth's is neither warrantevaluating nor truth-evaluating, but only (to extend Firth's classification) beliefevaluations. 22 Feldman considers the objection that a reason to doubt need only be a 'metaphysical possibility', i.e. something of whose falsehood we are not metaphysically certain. (I prefer the term 'epistemic possibility' here.) He gives both a textual and a strategic reply. The textual reply is that Descartes himself required that reasons to doubt be practically possible. But the passages Feldman cites in support of this are both inconclusive in themselves and outweighed by other passages, such as HR II266 [AT VII460: CSM II308], where Descartes says that the 'very least ground of suspicion' may engender doubt. The strategic reply is that by doing things his way 'we -107IV According to H. A. Prichard, 'We can only be uncertain of one thing because we are certain of something else.' 23 And according to Wittgenstein, 'The game of doubt presupposes certainty. If you tried to doubt everything, you would not get as far as doubting anything.' 24 A similar point was made in Descartes' own day by Bourdin, who asked concerning the reasons Descartes gave for doubting, 'If they are doubtful and replete with suspicion, how can they have brought any force to bear upon you?' 25 Descartes was unimpressed with this objection. In reply to Bourdin he said, 'We may well enough be compelled to doubt by arguments that are in themselves doubtful.' 26 Descartes was right about this. Reasons to doubt need not be certain: they need only be epistemically possible. To vindicate Descartes against Bourdin and company, I shall now construct an argument that purports to show that one thing is uncertain without presupposing that anything else is certain. There are three preliminaries. First, I define epistemic possibility as follows: if P is a proposition that S is considering at t, then P is epistemically possible for S at t if and only if S is not certain at t of not-P. Second, by the Demon Hypothesis I mean the following: an evil demon brings it about that whatever seems evident to me is false. Finally, let T be the proposition that 2 + 3 = 5. (If there is any proposition the reader finds more evident than 2 + 3 = 5, he may let T be that one instead.) Now here is the argument:

1.

The following proposition is epistemically possible for me: T seems evident to me, and the Demon Hypothesis is true. 2. If P entails Q and P is epistemically possible for me, then Q is epistemically possible for me. (In other words, epistemic possibility is transmitted by entailment. This is the analogue of a theorem in modal logic.) 3. (T seems evident to me, and the Demon Hypothesis is true) entails (T is false). ____________________ help to provide a conceptual framework within which a solution to the problem at hand may be found'. But I shall show that a solution may also be found within a framework that sets the standards for certainty higher. 23 H. A. Prichard, 'Descartes's Meditations', in Knowledge and Perception ( Oxford: Clarendon Press, 1950), 71-104; repr. in Doney (ed.), Descartes, 140-68. 24 Ludwig Wittgenstein, On Certainty ( New York: Harper & Row, 1969), sect. 115. 25 HR II273 [AT VII469: CSM II315]. 26 HR II 277 [AT VII473: CSM II319]. -1081. (T is false) is epistemically possible for me. ((1), (2), and (3)) 2. I am not certain that. T is true. ((4), definition) Here we have a valid argument that leads to the conclusion that I am not certain that 2 + 3 = 5. Yet nowhere in the argument is anything claimed to be certain. The Demon Hypothesis, the most powerful of Descartes' reasons to doubt, is said only to be epistemically possible.'Ah, yes, but wait just a minute,' I hear an objector say. 'It is true that in your premisses nothing is claimed to be certain. But in order for the argument to do its job, you who advance it must be certain of the premisses; that is, you must be certain that they are true.' The objector is mistaken. If the premisses of my sceptical argument are merely true, whether known to be so or not, the conclusion will also be true. And in that case I will not be certain that 2 + 3 = 5. But the premisses could all be true without my being certain of anything. Therefore, there are conditions sufficient for my being uncertain about one thing that do not require my being certain about something else. And this is just what the argument was supposed to illustrate.Nevertheless, there is another and more cogent objection to the argument. Premiss (2) is false, and derives only spurious support from the analogy with modal logic. To see this, note that (2) is equivalent to the following proposition: whatever has logical consequences that are uncertain for me is itself uncertain for me. But this is clearly false. If you present me with a remote theorem of number theory, I may well be uncertain of it, but I will be certain of the axioms just the same. Putting this point in terms of epistemic possibility, the negation of the theorem may be epistemically possible for me, but the negation of the conjunction of the axioms will not be. The reason for this discrepancy, of course, is that I may not be certain that the axioms do entail the theorem. This suggests that we modify premiss (2) as follows: 1. If (P entails Q) is certain for me, and P is epistemically possible for me, then Q is epistemically possible for me. In other words, epistemic possibility is transmitted by entailment when I am certain that the entailment holds.The revised premise is undoubtedly true, but it brings a new

difficulty along with it. To make the resulting argument valid, we shall have also to modify premise (3), as follows: 1. I am certain that the following entailment holds: (T seems evident to me, and the Demon Hypothesis is true) entails (T is false). -109Now we are claiming certainty in one of our premisses. So what becomes of my attempt to vindicate Descartes against Prichard and Wittgenstein?The answer is that it still succeeds. Prichard and Wittgenstein thought that grounds for doubt must be certain. Here they were wrong, and Descartes was right. Grounds for doubt need only be epistemically possible, as my revised argument still illustrates. But something must be conceded to Prichard and Wittgenstein none the less. Although a ground for doubt need not be certain, one must in such a case be certain about the logical relation in which the ground stands to the dubitandum. This is the moral of our having to incorporate (3′) into the revised argument. We may sum things up thus: doubt presupposes that something is certain; so far, Prichard and Wittgenstein were right. 27 But the ground on which one bases a doubt need not be certain; on this point Descartes was right.I want to make one observation before moving on. Look at my revised sceptical argument and I think you will agree that the only challengeable premiss is (1). I think you will also agree that (1) is false only if the Demon Hypothesis is not epistemically possible. Thus, you can challenge the argument only by challenging the epistemic possibility of the Demon Hypothesis. And if you do this, you are claiming to be certain that the Demon Hypothesis is false. To overthrow the demon of scepticism, you must take him boldly by the horns. V In this section I shall present what I think is the most promising solution to the problem of the Cartesian Circle. Not only is it the best solution for Descartes, but it also has applications to contemporary epistemology.The key to the solution I advocate is a distinction similar to one Anthony Kenny draws between the following propositions: a. For all P, if I clearly and distinctly perceive that P, then I cannot doubt that P. b. I cannot doubt that (for all P, if I clearly and distinctly perceive that P, then P). 28 ____________________ 27 A corollary they draw is that universal doubt is impossible, but one should not make too much of this. Although one cannot have reasons for doubting everything, one could have reasons for doubting almost any arbitrary proposition. Though not universal scepticism, this is sceptlcism enough. 28 This distinction is anticipated in Kenny book Descartes by the distinction on pp. 183-4 between first-order doubt (which would correspond to the denial of (a)) and secondorder doubt (which would correspond to the denial of (b)). It is drawn explicitly in his article 'The Cartesian Circle and the Eternal Truths'. -110The distinction I shall use is analogous but different. For Kenny's indubitability, which does not obviously entail either evidence or truth, I want to substitute certainty. This gives us the following pair of propositions:

(A) For all P, if I clearly and distinctly perceive that P, then I am certain that P. (B) I am certain that (for all P, if I clearly and distinctly perceive that P, then P). The difference is that (A) says that whenever I clearly and distinctly perceive any proposition, I will be certain of it (the proposition in question), whereas (B) says that I am certain of a general principle connecting clear and distinct perception with truth. Clearly, (A) could be true even though (B) were false. (B) requires that I have the concept of clear and distinct perception, but (A) does not. Moreover, even if I did have this concept, I might be uncertain about the general connection between clear and distinct perception and truth, yet certain of every proposition I did clearly and distinctly perceive. 29 Unfortunately, the same English sentence--'I am certain of the truth of clear and distinct perceptions'--may be used to express either (A) or (B). Perhaps because of this, the distinction is often missed. But it is crucial to the Cartesian enterprise. I shall briefly indicate how it enables us to make sense of two otherwise puzzling passages in Descartes, and then I shall show how it provides an escape from the Circle. The first passage is the notorious fourth paragraph in the Third Meditation, where Descartes appears to oscillate inconsistently between saying, on the one hand, God or no God, I am certain of things when I clearly and distinctly perceive them, and, on the other hand, I can doubt even the truth of clear and distinct perceptions if I do not know there is a veracious God. The appearance of inconsistency is removed if we see Descartes as being uncertain not of particular propositions that he clearly and distinctly ____________________ One claim Kenny makes in connection with the distinction is incorrect. In the book he says that second-order doubt is doubt of particularly propositions 'in a roundabout manner' by referring to them 'under some general heading, such as "what seems to me most obvious"' (pp. 183 -4). In the article he says that the doubt implied by the denial of (b) touches particular clearly and distinctly perceived propositions 'only through referentially opaque wrappers' (p. 689). But referential opacity prevents any 'touching' at all. One who doubts that what he perceives clearly and distinctly is true is not doubting particular clearly and distinctly perceived propositions in any manner, however roundabout, nor through any wrappers, however opaque. What he is doubting is whether there is any connection between clear and distinct perception and truth. 29 It seems also to be the case that (B) could be true even though (A) were false. I might be certain that clear and distinct perception guarantees truth, yet not certain that a given proposition I was clearly and distinctly perceiving was true, provided I was unaware that I was clearly and distinctly perceiving it. -111perceives, but only of the general connection between clear and distinct perception and truth. What he shows us in this paragraph is that at this stage in the Meditations (A) is true of him, but (B) is not. 30

The second thing our distinction enables us to understand is the epistemic advantage Descartes claims over the atheist. He concedes that even the atheist can be certain that the three angles of a triangle are equal to two right angles if he is clearly and distinctly perceiving this at the time. (A) is thus true of atheist and Descartes alike. But if at a later time both men merely remember having a clear and distinct perception of that theorem, Descartes will still be certain of it, but the atheist will not. This is not because (as the Memory Gambit would have it) Descartes can trust his memory and the atheist cannot. It is rather because Descartes can be certain (after he has proved the veracity of God) that anything he once clearly and distinctly perceived is true, whereas the atheist cannot. So (B) is true of Descartes, but not of the atheist. 31 Now let us see how our distinction enables us to break out of the Circle. The first thing to notice is that (B) need not be true at the beginning of the Cartesian enterprise. I do not have to be certain that all clear and distinct perceptions are true before I prove that God exists. Proposition (2), which describes the lower are of the Circle, is false. Although (B) is false at the outset, however, it does not follow that (A) is false at the outset. And if (A) is true at any time, then anything I perceive clearly and distinctly at that time will be something I am certain of. Clear and distinct perception will thus provide me with an initial stock of premisses I know for certain to be true. According to Descartes, these premisses will include I think; if I think, then I exist, and a cause must ____________________ 30 This analysis of the paragraph is essentially Kenny's; see his 'Cartesian Circle and the Eternal Truths', 689. It must be said, however, that the final sentence of this paragraph-'Without a knowledge of these two truths [God exists and is not a deceiver] I do not see that I can ever be certain of anything'--is an embarrassment for almost any interpretation of Descartes. Here he digs himself into a pit so deep, there can be no climbing out. Some interpreters--e.g. George Nakhnikian in "'The Cartesian Circle Revisited'", American Philosophical Quarterly, 4 ( 1967), 251 -5--regard this sentence as an aberration. 31 Descartes' advantage is not as great as he thinks. Merely remembering that he once had a clear and distinct perception of God's veracity will not (on pain of circularity) assure him now that anything he previously clearly and distinctly perceived is true. So Descartes must acquire a present clear and distinct perception of God's veracity by going through the theological proofs afresh. This means that he has an advantage over the atheist only in those cases where the theological proofs are shorter or more easily called to mind than the geometrical proofs the atheist must reconstruct. It could be replied that the content of Descartes' recollection is not 'I once clearly and distinctly perceived that God exists and is no deceiver', but just 'God exists and is no deceiver'. But if so, the atheist is entitled to say that the content of his recollection is not 'I once clearly and distinctly perceived that the angles of a triangle are equal to two right angles', but just 'The angles of a triangle are equal to two right angles'. -112-

contain at least as much reality as its effect, among others. From these first principles I can go on to prove other things, including the existence of God and, eventually, the principle that all clear and distinct perceptions are true. 32 Clear and distinct perception and proposition (A) play crucial roles in my proposal that must not be misunderstood. The fact that I clearly and distinctly perceive a proposition does not serve as a ground for accepting it. It is a source of knowledge, but not a ground. Nor does proposition (A) serve as a ground. Rather, it is a fact that enables knowledge to get started. (We can authenticate this fact later if we wish, but need not do so in the beginning.) Prichard misunderstands Descartes in just the way I am warning against. According to him, Descartes thinks he can arrive at certainty about a proposition P only by running through an argument of the following sort: Whatever I perceive clearly and distinctly is certain. I am perceiving clearly and distinctly that P. Therefore, P is certain. 33 Here proposition (A) and my clearly and distinctly perceiving P both appear as grounds. As grounds, they cannot contribute to knowledge unless they are themselves known. But the question how I know them obviously has no satisfactory answer if an argument like the one above must stand behind any answer I give. Prichard's interpretation of Descartes thus leads to disaster (as he himself is quick to point out). In opposition to Prichard's interpretation, I maintain that, in order to become certain of a proposition, I do not need to know that I am clearly and distinctly perceiving it, nor that whatever I so perceive is either certain or true. It is enough that I do clearly and distinctly perceive the proposition. (A) says that this is enough. For (A) says that perceiving something clearly and distinctly is sufficient to render me certain of it. It follows that nothing else is necessary, unless it is also necessary for the occurrence of clear and distinct perception in the first place. But neither knowledge of (A) nor knowledge of the fact that I am clearly and distinctly perceiving something is necessary for such perception to occur. 34 ____________________ 32 Of course, I do not really believe that Descartes knew for certain everything he said he did: e.g. his causal maxims. But this does not detract from the soundness of the general plan I am attributing to him. 33 Doney (ed.), Descartes, 145. 34 Kenny hits upon an important part of the solution I am advocating when on p. 194 of Descartes he writes: 'When [ Descartes] passes from the clear and distinct perception of something to the affirmation of its truth, he does not do so by tacit appeal to a suppressed major premise; his affirmation is based directly on the intuition and not on a deduction derived from -113-

The point I have been insisting upon could be summed up as follows: (A) is not a principle I have to apply in order to gain knowledge; I need only fall under it. The solution I am proposing does not require us to deny (1), the proposition describing the upper are of the Circle. We can side with Descartes if we like, and hold that in order to know that whatever is clearly and distinctly perceived is true, we must first know that God exists. 35 But this is not to say that before God's existence is known, clear and distinct perception affords us no certainty. On the contrary, it does, and this is what lets knowledge get started. Those who observe that we must be certain of some clear and distinct perceptions independently of our knowledge of God are therefore correct. But to conclude from this that (1) must be denied would be to confuse (A) with (B). Moreover, to say (as I do) that we must be certain at the outset of some clear and distinct perceptions is not to say (as I don't) that we must be certain at the outset of the proposition some clear and distinct perceptions are true. Nor is it to say that we must be certain of a more specific proposition of the form those clear and distinct perceptions that are F (e.g. bathed in the light of nature) are true. In either of these cases we would be entertaining a proposition about the epistemic powers of clear and distinct perception, but no such thing is required of us. We have seen that in virtue of (not by appeal to) proposition (A) we acquire an initial stock of certainties. Where do we go from here? Descartes' route is all too familiar. Among the initial certainties are premisses that entail the existence of God. In the Third Meditation he clearly and distinctly perceives this entailment, thus becoming certain that God exists. In the Fourth Meditation he clearly and distinctly perceives, and thus becomes certain, that God could not be a deceiver. He goes on to infer (and become certain) that whatever he perceives clearly and distinctly is true. It is important to recognize at this point that, despite his usual formulation of it, Descartes' rule of clearness and distinctness (the 'C&D Rule', I shall call it) is not merely a rule of truth. It is also a rule of evidence and, indeed, of certainty. Many passages make this clear. When Descartes first ____________________ a general proposition about the truthfulness of intuitions.' But three pages later he betrays this insight with the following claim: 'If every other certainty is to be built upon the certainty afforded by clear and distinct perception, then it is essential if there is to be any certainty at all of the type Descartes sought, that one should be able to be certain that one is clearly and distinctly perceiving something.' On the contrary, if you don't need 'Whatever is perceived clearly and distinctly is true' as a major premiss, then you don't need 'I am clearly and distinctly perceiving that P' as a minor premiss. 35 Though in fact I would deny this, of course. -114-

introduces the C&D Rule, it is by means of the question, 'Do I not . . . know what is requisite to render me certain of a truth?' 36 A few sentences later he refers to clear and distinct perceptions as 'matters in which I believe myself to have the best evidence'. 37 Elsewhere he says that clear and distinct perceptions are 'true and certain'. 38 Moreover, he often tells us that when he clearly and distinctly perceives something, he cannot help but believe it. 39 What is clearly and distinctly perceived, then, is not only true, but also maximally evident and believed. It is therefore known for certain. The C&D Rule thus turns out to be equivalent to proposition (A). Descartes' procedure could be summed up thus: by falling under proposition (A) (i.e. the C&D Rule), he becomes certain of premisses from which he eventually derives proposition (A) itself. 40 But since he does not have to use proposition (A) at any step along the way, there is no circle. Once Descartes knows that proposition (A) is true, of what use is the information? Well, for one thing, it enables him to cast off the atheist's handicap: he can now be certain of things he merely remembers having clearly and distinctly perceived. But more importantly, it enables him to vindicate his starting-point. It gives him an answer to the critic who says, 'I grant that your procedure is not circular, but I don't see how you can escape the charge of arbitrariness in your first premisses. What is the justification for starting from just the premisses you did?' After he has proved proposition (A), Descartes can give the following reply: 'Those premisses are things I knew for certain. The proof of this is that I perceived them clearly and distinctly, and whatever I so perceive is certain.' To make the last point clearer, let me review the successive stages of the Cartesian enterprise. Here are the things of which Descartes is certain, listed in the order in which he becomes certain of them: (1)

(2)

1 think, the causal maxims, etc. God exists, God is no deceiver.

Propositions known because they are clearly and distinctly perceived. Propositions known because they are clearly and distinctly perceived to follow from premisses at level (1).

____________________ 36 H R I 158 [ A T VII35: C S M II24]. 37 37Ibid. 38 H R I184; H R II41 [ A T VII70, 144: C S M II48, 103]. What Descartes says here is that clear and distinct perceptions are certain after God's existence is proved. For my purposes, we must interpret him to mean: 'From a knowledge of God, I can prove that whatever is perceived clearly and distinctly is certain', rather than 'if I clearly and distinctly perceive something and I know that God exists, then I will be certain of that thing'. The latter makes knowledge of God a prerequisite of certainty, but the former does not. 39 H R I158, 176, 183; H R II42, 266 [ A T VII36, 59, 69, 145-6, 460: C S M II25, 41, 48, 104, 309].

40

Once he becomes certain of (A), (B) is true, too--but not until then. -115-

(3) Whatever I perceive clearly and distinctly is certain. (4) I perceive clearly and distinctly that I think, etc. (5) I am certain that I think, etc.

Principle known because it is clearly and distinctly perceived to follow from propositions at level (2). New premisses, one corresponding to each premiss at level (1). 41 Propositions known because they are clearly and distinctly perceived to follow from propositions at levels (3) and (4).

The propositions at stage (1) are Descartes' first premisses; he accepts them without any supporting grounds. This is what incurs the charge of arbitrariness (or dogmatism). But that charge may be answered as follows. If proposition (A) is true, then Descartes' first premisses are immediately justified; that is, they are justified simply in virtue of being clearly and distinctly perceived, not because they inherit justification from other propositions. From these immediately justified beginnings Descartes goes on to derive (at level (3)) proposition (A) itself, which serves as a reason for the higher-order propositions (at level (5)) to the effect that his initial premisses were justified (indeed, certain). This shows that the initial premisses were not arbitrary. The strategy I am imputing to Descartes is nicely described in a recent article by William P. Alston. He writes: For any belief that one is immediately justified in believing, one may find adequate reasons for accepting the proposition that one is so justified. The curse (of dogmatism) is taken off immediate justification at the lower level, just by virtue of the fact that propositions at the higher level are acceptable only on the basis of reasons. 42 This fits Descartes perfectly. His initial premisses are immediately justified and certain, but the higher-order proposition that says they are certain is justified by appeal to reasons. 43 ____________________ 41 For a discussion of how premisses of the form 'I perceive clearly and distinctly that P' are known, see Frankfurt, Demons, Dreamers, and Madmen, ch. 13. Some critics of Descartes hold such knowledge to be problematic. It is an advantage of my interpretation that the need for It is postponed until level (4). 42 William P. Alston, "'Two Types of Foundationalism'", Journal of Philosophy, 73 ( 1976), 165-85.

43

Alston himself does not attribute this strategy to Descartes. He thinks Descartes is an iterative foundationalist, i.e. one who believes that some propositions of the form 'P is immediately justified for S' are themselves immediately justified for S (ibid. 182 n.). But as I interpret Descartes, he is what Alston calls a 'simple foundationalist', i.e. one who believes that for any epistemic subject, some propositions are immediately justified, but that no propositions of the form 'P is immediately justified for S' are immediately justified. Such higher-order propositions are justified only mediately by reference to the veracity of God. -116-

I shall consider one more objection to Descartes' procedure. This one charges that proposition (A) is arbitrary because one could have chosen any rule of evidence and justified it by the method I attribute to Descartes. My reply is twofold. First, proposition (A) is not arbitrary; it is the conclusion of an argument. One may wish to quarrel with the argument, but that would be a separate objection. Second, it is simply not true that any principle could be justified in the same manner as (A). It has at least to be a true principle, or else no first premisses will be justified in virtue of it. This completes my solution to the problem of the Cartesian Circle. Needless to say, I do not wish to endorse all the details of Descartes' reasoning. But I do maintain that in general outline it is sound, and in Part Two I shall use what is of value in it to throw light on contemporary issues. PART TWO VI In contemporary epistemology there is a movement away from foundationalist theories of justification toward coherentist theories. Wilfrid Sellars and Keith Lehrer, two of the leaders of this movement, have criticized foundationalism by raising doubts about its ability to justify its own epistemic principles. In this part I shall draw on what we have learned about Descartes in Part One to show how these criticisms may be met. Two doctrines are essential to foundationalism: (i) there is a class of propositions--the 'foundations'--that are self-evident or immediately justified; and (ii) every proposition that is justified is so at least partly in virtue of standing in certain relations to the foundations. 44 Certain other doctrines are often associated with foundationalism, but they are not entailed by (i) and (ii). For example, Descartes was a foundationalist who held (iii) ____________________ 44 By a self-evident proposition I do not mean one that derives its evidence from itself (whatever that might mean), but one that does not derive its evidence from any other propositions. (Similarly, when theologians speak of God as self-caused, they do not usually mean that God causes his own existence, but rather that his existence is not caused by any other being. See Caterus's remarks on this point at H R II4 [ A T VII95:

C S M II68-9].) Perhaps a better term is 'Immediately justified', which I shall generally use instead. It should be borne in mind that a justified proposition is not necessarily one that an epistemic subject has gone through the procedure of 'justifying'. -117that the foundational propositions must be not only immediately justified, but also certain and indubitable, (iv) that they are limited in scope to simple necessary truths and propositions about one's own mental states, and (v) that the superstructure is related to the foundations by deductive relations exclusively. One can defend (i) and (ii) without being committed to (iii), (iv), and (v). Foundationalists often set forth principles specifying the conditions under which propositions of various types are justified. Usually called 'epistemic principles', they fall into two groups: principles that tell us that propositions of certain types are justified independently of their logical relations to other propositions, and principles that tell us that if some propositions are already justified, then any propositions that stand in suchandsuch relations to them are also justified. Principles of the first sort I call 'generation principles', since they are principles whereby justification is generated in the first place, and those of the second sort I call 'transmission principles', since they are principles whereby justification is transmitted from some propositions to others. Generation principles are used to lay the foundations, transmission principles to erect the superstructure. The general form of an epistemic principle is 'If . . . then P is justified for S'. The antecedent of a sentence expressing a transmission principle will contain terms of epistemic appraisal (such as 'evident', 'certain', and so on), since it must mention other propositions and specify their epistemic status. But the antecedent of a sentence expressing a generation principle will not contain terms of epistemic appraisal. 45 Descartes' C&D Rule is a generation principle. It tells us that if someone is clearly and distinctly perceiving a proposition--a state we can describe without using epistemic terms--then that proposition is certain for him. Another example of a generation principle is Chisholm's principle (A), which says that if a subject is in any of a designated group of 'selfpresenting states', then it is evident to him that he is in the state in question. 46 ____________________ 45 Some coherentists espouse principles that are neither generation principles nor transmission principles, but a sort of hybrid between the two. An example would be 'If P coheres with the system of propositions accepted by S (or the scientists of our culture circle, etc.), then P is justified for S'. This is like a generation principle in that its antecedent contains no epistemic terms, but like a transmission principle in that its antecedent specifies relations to other propositions. It seems to me, however, that an adequate epistemology must recognize at least one fullblooded generation principle.

'Credibility may be transmitted from one statement to another through deductive or probability connections; but credibility does not spring from these connections by spontaneous generation' ( N. Goodman, "'Sense and Certainty'", Philosophical Review, 61 ( 1952), 160-7). 46 Roderick M. Chisholm, Theory of Knowledge ( Englewood Cliffs, NJ: Prentice-Hall, Inc., 1966), 24-37 and 44. -118An obvious example of a transmission principle is the principle that deduction transmits justification--more precisely, if P is justified for S, and the proposition that P entails Q is true and justified for S, then Q is justified for S. This is the only transmission principle Descartes allowed. 47 But most foundationalists countenance several others, including, perhaps, principles whereby justification is transmitted to propositions about the physical world and propositions about the past. The best-known list of such nondeductive transmission principles is Chisholm's. 48 Critics of foundationalism look upon its epistemic principles with suspicious eyes. 'What is the justification for them?,' they ask. Sellars and Lehrer contend that foundationalists have no satisfactory answer to this question, and urge us to adopt a coherentist view instead. Coherentism denies (i) and (ii), maintaining that a proposition may be justified in virtue of belonging to a coherent system of propositions none of which is immediately justified.It will be convenient to sum up the Sellars-Lehrer critique in three theses: 49 (I) Epistemic principles must themselves be known (or justified) if knowledge (or justified belief) is to arise in accordance with them. 50 (II) There is no way to justify epistemic principles within a foundationalist framework. (III) There is a way to justify epistemic principles within a coherentist framework. If (I) and (II) are both true, foundationalism leads to scepticism. If (III) is true, coherentism does not. So the upshot of the Sellars-Lehrer critique is this: if we wish to avoid scepticism, we must reject foundationalism in favour of coherentism. In the remaining sections, I shall explore ____________________ 47 The principle Descartes introduces in the Sixth Meditation--that beliefs in physical objects prompted by our sensory ideas must be true, else God would be a deceiver--is a rule of truth, not a rule of evidence. I say more about this distinction in Section X. 48 See principles (B)-(I) in Chisholm, Theory of Knowledge, ch. 3. Some readers may question my classification of principles (B)-(F) as transmission principles. They have the form of generation principles, since their antecedents contain no terms of epistemic appraisal. I suspect, however, that Chisholm thinks of them as transmission principles. If they were generation principles, then he would be committed to the view that some propositions about the physical world and some propositions about the past are immediately justified; but Chisholm sides with the Cartesian tradition in holding that such propositions are only mediately justified. 49 Neither Sellars nor Lehrer presents exactly this argument, but the materials for it are

contained in their writings. See Keith Lehrer, Knowledge ( Oxford: Clarendon Press, 1974), 143-4, and Wilfrid Sellars, "'Givenness and Explanatory Coherence'", Journal of Philosophy, 70 ( 1973), 612-24, esp. sects. V-VII. 50 For convenience in what follows I shall often use the term 'knowledge' where justification is all that need be at issue. -119possible foundationalist responses to this challenge, concentrating on (1) and (11). VII Premiss (1) of the Sellars-Lehrer critique is widely taken for granted, but acceptance of it rests on a misunderstanding of how epistemic principles function. I have already discussed this misunderstanding in Part One. There we saw that Descartes did not need to know that the C&D Rule was true in order for clear and distinct perception to give him knowledge. For just the same reason, we can say in general that a subject need not know that an epistemic principle is true in order for the circumstance mentioned in its antecedent to give him knowledge. The argument for this is very simple. An epistemic principle has the form 'If . . . then P is justified for S'. In other words, it says that the obtaining of whatever condition is specified in its antecedent is sufficient for P's being justified for S. 51 Now it is a logical truth that if X is sufficient for Y, then there is no other condition Z that is necessary for Y, unless Z is also necessary for X. But knowledge of an epistemic principle is not necessary for the obtaining of its antecedent. Therefore, knowledge of an epistemic principle is not necessary for knowledge to arise in accordance with it. The first premiss of the SellarsLehrer critique is false. VIII The argument that shows that knowledge of epistemic principles is not required also shows that knowledge of their antecedents is not required. This undermines a criticism Sellars levels against Chisholm's principle (A). That principle, recall, says that if a subject is in any of a designated group of psychological states, then it is evident to him that he is in whatever state it is. Sellars says that this principle 'seems to point to' arguments of the following form: ____________________ 51 Sometimes epistemic principles are formulated so as to allow for the possibility that the justification arising in accordance with them may be defeated or overridden. In these cases the obtaining of the antecedent is not by itself sufficient for justification to arise; what is sufficient is this plus the absence of any overriding circumstances. But this complication does not affect the point I am making. (Incidentally, this same complication calls for a qualification in my earlier characterization of generation principles: a clause in the antecedent stipulating that there are no overriding

circumstances may, in its specification of those circumstances, use terms of epistemic appraisal; but the positive clause in the antecedent will not use such terms.) -120It is a fact that I am F. So, it is reasonable to believe that I am F. 52 He then observes: 'In order for any such argument to do the job, its premise would have to have authority, it would have to be something which it is reasonable to believe.' 53 Presumably the difficulty with this is that if the conclusion were in question, the argument could not be expected to put it out of question. To raise this objection is to misunderstand Chisholm in just the way that Prichard misunderstood Descartes. Chisholm's principle is not supposed to function as a suppressed major premiss under which a subject must subsume himself. The point is rather that just as having a clear and distinct perception of something puts one in a condition of knowing it, so being in a self-presenting state puts one in a condition in which it is evident to one that he is in that state. 54 There is a standard objection to coherentism that can be answered in the same way. Coherentists say that a proposition is justified if it coheres with a system of propositions of the right kind--for example, those already accepted by the subject, or by 'the scientists, of our culture circle', and so on. Foundationalists have been wont to object by asking, 'How do you know which propositions belong to the system? And how do you know that a given proposition does cohere with them?' But if the coherentist principle is true (and one may, of course, wish to question this), one need not know those things. IX In Section VII we saw that epistemic principles need not be known in order for knowledge to arise in accordance with them. This does not mean, however, that the question 'What justifies your principles?' is one that the foundationalist can brush aside. There are two reasons for this. In the first ____________________ 52 W. Sellars, "'Epistemic Principles'", in H.-N. Castañeda (ed.), Action, Knowledge, and Reality ( Indianapolis: Bobbs-Merrill, 1975), 332-47; see sects. 18-20. 53 Ibid. 54 There are, of course, other objections that one could raise against Chisholm's principle. Sellars' most challenging one is this: a necessary condition of P's being evident to anyone is his having learned a sentence that means that P; but being in one of the states Chisholm calls 'selfpresenting' is not sufficient for having learned any sentences; therefore, being in one of these states is not sufficient for anything's being evident to you. Compare the argument on pp. 1312 of "'Empiricism and the Philosophy of

Mind'", in Sellars' Science, Perception, and Reality ( London: Routledge and Kegan Paul, 1963). The big issue raised by this objection, of course, and one too big to discuss here, is the relation of thought to language. -121place, although some knowledge arises even if the principles are not known, there may be other knowledge that does depend on knowing the principles. In Descartes' system, for example, one must have knowledge of the C&D Rule in order to obtain knowledge of propositions clearly and distinctly perceived at an earlier time. One must also have knowledge of the rule in order to arrive at higher-order knowledge--knowledge that one knows. In the second place, even if knowledge of epistemic principles is not required for either of the two purposes just mentioned, we may seek it for its own sake when we embark upon epistemological inquiry. If it turns out that epistemic principles are justified at all, foundationalism will owe us an account of how they are justified. Otherwise it will not be a complete theory of justification. I turn, therefore, to examination of premiss (II) of the Sellars-Lehrer critique. If epistemic principles are justified within a foundationalist framework, this must be in one of two basic ways: either they are immediately justified, lying at the foundation, or they are mediately justified, resting upon the foundation. Within the latter alternative there are two sub alternatives, which are best explained with the help of a term borrowed from Alston: an epistemic proposition is a singular proposition that attributes evidence, certainty, or some other epistemic characteristic to another proposition. 55 Now on the first subalternative, epistemic principles are justified after epistemic propositions, thus occupying a higher storey in the edifice of knowledge, and on the second subalternative epistemic principles are justified before epistemic propositions, thus occupying a lower storey in the edifice of knowledge. 56 This gives us three alternatives in all: the justification of epistemic principles may be immediate; it may be mediate and posterior to that of epistemic propositions; or it may be mediate and prior to that of epistemic propositions. I shall discuss all three alternatives, beginning with the last. X The first pattern to be discussed--mediate justification of epistemic principles prior to the justification of epistemic propositions--follows this general sequence: ____________________ 55 Alston, "'Two Types of Foundationalism'", 169. 56 A third possibility, of course, is that epistemic principles and epistemic propositions occupy the same storey of the edifice. But this is not very promising. There is some prospect of justifying epistemic principles if you can appeal to epistemic propositions, and some prospect of justifying epistemic propositions if you can appeal to epistemic principles, but little prospect of justifying each independently of the other. See Chisholm, "'Problem of the Criterion'", where the same point is made in different terms.

-1221. 2. 3. 4. 5.

Propositions known immediately. Further propositions inferred from propositions at level (1). Epistemic principles inferred from propositions at level (2). Instantiations of antecedents of epistemic principles. Epistemic propositions inferred from propositions at levels (3) and (4).

The reader who refers back to Section V will see that this is precisely the pattern I attributed to Descartes. At level (1) we have the Cogito and other things known immediately, at level (2) the existence and veracity of God, at level (3) the C&D Rule, at level (4) subsumptions of level (1) propositions under the rule, and at level (5) epistemic propositions attributing certainty to the propositions at level (1). I argued in Section V that Descartes' procedure is neither circular nor arbitrary. What is problematic about it is simply whether there are indeed valid inferences from level (1) to level (2) and from level (2) to level (3). As Hume observed, to have recourse to the veracity of God in order to prove the certainty of our perceptions is to make 'a very unexpected circuit'. 57 Contemporary epistemologists who use Descartes' pattern will no doubt want to avoid the circuit through theology, replacing his level (2) propositions by something more in keeping with a naturalistic world-view. Let us therefore inquire into the prospects for a Cartesian epistemology naturalized. 58 Let us begin by taking a look at an attempt by Sellars to give a naturalistic derivation of Chisholm's principle (C), which reads as follows: 'If there is a certain sensible characteristic F such that S believes that he perceives something to be F, then it is evident to S that he is perceiving something to have that characteristic F, and also that there is something that is F. 59 Sellars's derivation of this principle occurs as part of a larger coherentist strategy--in true Hegelian fashion, he is trying to concede the approximate truth of the foundationalist's principles while showing that they find their rationale only within his own coherentist system. None the less, aspects of his derivation might be appropriated by foundationalists. The idea, then, is to derive principle (C) from naturalistic facts--facts of the sort establishable by scientific inquiry. What might these facts be? Sellars's candidates are facts about concept formation and language ____________________ 57 David Hume, An Enquiry, Concerning Human Understanding, sect. xii, part 2 [ed. Selby- Bigge , p. 153]. Hume was referring to sense perceptions rather than clear and distinct perceptions, but the same point holds for the latter. 58 That naturalistic epistemology and Cartesian epistemology may be viewed as sharing the pattern (1)-(5) was suggested to me by Stephen Leeds. 59 Sellars, Theory of Knowledge, 47. -123-

learning. Certain sentences--e.g. 'Here is a red apple'--are learned as directly conditioned responses to states of affairs obtaining in one's immediate vicinity. Now to say that a response R is conditioned to a stimulus S is to say that for the most part R occurs when and only when S occurs. Therefore, one who has learned how to use the sentence 'Here is a red apple' will tend to utter it when and only when a red apple is present to his eyes. Hence, the utterance of that sentence by one who knows the language is a good indicator of the presence of a red apple. And one who finds himself uttering (or spontaneously inclined to utter) 'Here is a red apple' can therefore take this utterance (or inclination) as evidence for the truth of the belief expressed by those words. We thus arrive (in Sellars's words) at 'something very like' Chisholm's principle (C). 60 But just how like Chisholm's principle is it? What Sellars's argument really shows is at most this: if S utters (or is spontaneously inclined to utter) the words 'Here is an F' (where 'F' is any predicate directly conditioned to non-verbal stimuli), then the proposition expressed by those words is likely to be true. This differs from principle (C) in several ways, one of which for our purposes is crucial: where (C) contains the term 'evident', Sellars's principle contains the phrase 'likely to be true'. 61 It is plain that the latter must be taken in a statistical sense--the import of Sellars's principle is that most beliefs of a certain sort are true. Now the mere fact that most beliefs of a certain sort are true does not suffice to make those beliefs evident (or even justified to any degree) for the persons who hold them. After all, any true belief belongs to at least one class of beliefs most members of which are true, but not all true beliefs are justified. What our Sellarsian considerations establish, therefore, is a statistical principle, not an epistemic principle in the proper sense of the term. 62 This may account for Sellars's allegiance to premiss (I). We saw in Section VII that when epistemic principles are understood in the way I suggested, (I) is demonstrably false. But to someone who really has statistical rather than epistemic principles in mind, (I) will seem eminently plausible. The statistical fact that most beliefs of a certain sort are true does not make any of those beliefs evident; but if a believer knew this ____________________ 60 This argument occurs both in "'Epistemic Principles'" and "'Givenness and Explanatory Coherence'". For a somewhat similar account see W. V. Quine and Joseph Ullian, The Web of Belief ( New York: Random House, 1970), 33-6, where it is suggested that a 'force for veracity' can be found 'in the very mechanism of language learning'. 61 Other differences, not crucial for our purposes, are these: in Sellars's principle the values of 'F' are not restricted to sensible characteristics, and the talk of belief has been transposed into talk of utterance and Inclination. 62 I classify as statistical any principle that is concerned with frequency of truth rather than justification. Thus, even a principle to the effect that all beliefs of a certain sort are true would count as statistical. -124-

statistical fact, and knew in addition that one of his beliefs belonged to the sort in question, then (if no evidence pointed the other way) he would be justified in holding that belief. It does seem that statistical principles, unlike epistemic principles, can contribute to knowledge only if they are themselves known. 63 Whether this explains Sellars's acceptance of (I) or not, the point remains that his naturalistic derivation does not yield an epistemic principle, but only a statistical principle. Moreover, given a certain assumption that I have merely left tacit until now, I think it is safe to say that any attempt at a naturalistic derivation of epistemic principles would meet the same fate. The assumption is this: epistemic principles involve concepts that are irreducibly epistemic. That is to say, they involve concepts like evidence and certainty, and these concepts can only be defined with the help of other epistemic concepts; they cannot be defined solely in terms of logical and empirical concepts such as truth, probability, causation, and belief. 64 If this assumption is correct, it is hard to see how epistemic principles could ever be derived from propositions established by scientific inquiry. The difficulties involved would be analogous to those involved in trying to derive 'ought' statements from 'is' statements, or observation statements from theoretical statements (without correspondence rules). In conclusion, then, without a naturalistic reduction of epistemic concepts, there cannot be a naturalistic derivation of epistemic principles. 65 We must look elsewhere for their justification. 66 ____________________ 63 Sellars and Chisholm appear to agree in holding that statistical principles can be of epistemological significance only if they are themselves known. See Sellars, Science, Perception, and Reality, 167-8, and Chisholm, Perceiving ( Ithaca, NY: Cornell University Press, 1957), 27. 64 In "'Empiricism and the Philosophy of Mind'" Sellars seems to agree: 'The idea that epistemic facts can be analyzed without remainder--even "in principle"--into nonepistemic facts . . . is, I believe, a radical mistake--a mistake of a piece with the socalled 'naturalistic fallacy' in ethics' (p. 131 ). 65 This point by itself, however, does not undermine naturalistic programmes in epistemology. Naturalists might propose simply to dispense with epistemic principles, maintaining that statistical principles are all epistemology needs. To this it might be objected that knowledge of statistical principles presupposes the truth of at least one epistemic principle. After all, statistical principles must be inferred from data, surely we need a principle in virtue of which the data are justified and a principle in virtue of which the inference is justified. The naturalist can reply that this is true only if knowledge is analysed in terms of justification. It is not true if knowledge is given an analysis from which the concept of justification is eliminated, as in Goldman's 'causal' analysis ( Alvin Goldman, "'A Causal Theory of Knowing'", Journal of Philosophy, 64 ( 1967), 357-72) or Armstrong's 'reliability' analysis ( David Armstrong, Belief, Truth, and Knowledge ( Cambridge: Cambridge University Press, 1973)). I do not believe that these analyses are adequate, but I lack the space to criticize them here. 66 Unfortunately, the criticisms raised in this section apply to Descartes' enterprise, too.

Although it appeals to a supernatural entity, that enterprise is naturalistic in the sense that it tries to derive epistemic principles from non-epistemic facts. -125XI The next pattern to be discussed--mediate justification of epistemic principles posterior to the justification of epistemic propositions--is also present, although not prominent, in the Meditations. At the beginning of the Third Meditation Descartes introduces his C&D Rule in the following way: I am certain that I am a thing which thinks; but then do I not likewise know what is requisite to render me certain of a truth? Certainly in this first knowledge there is nothing that assures me of its truth, excepting the clear and distinct perception of that which I state, which would not indeed suffice to assure me that what I say is true, if it could ever happen that a thing which I conceived so clearly and distinctly could be false; and accordingly it seems to me that already I can establish as a general rule that all things which I perceive very clearly and very distinctly are true. 67 Here Descartes starts with an epistemic proposition--'I am certain that I am a thing which thinks'--and moves to an epistemic principle--'All things which I perceive very clearly and distinctly are true'. 68 In the paragraphs immediately following, however, Descartes says that his principle is subject to doubts that can be removed only by establishing the existence of a veracious God. Evidently, then, he views the passage I have quoted as belonging merely to the context of discovery, not to the context of justification. In it the C&D Rule is first brought to light, but not yet established as known.But why not regard the passage as a justification of the C&D Rule? We can extract from it the following argument: 1. I am certain that I am a thinking thing. 2. The only possible source of this certainty is clear and distinct perception. 3. Therefore, clear and distinct perception is a source of certainty (i.e. whatever I perceive clearly and distinctly is certain). This argument fits the pattern Chisholm has labelled 'critical cognitivism'. Such arguments contain one premiss affirming that we do have knowledge, certainty, or justification with respect to propositions of a given type, another premiss ruling out all possible sources of this knowledge except one, and a conclusion affirming that the remaining possible source must therefore be the source of the knowledge in question. Epistemic principles are thus justified by appeal to epistemic propositions. 69 ____________________ 67 H R I158 [AT VII 35 : C S M II24]. 68 Keep in mind that instead of 'true' he should say 'certain'. 69 See Chisholm, Theory of Knowledge, 59-61. There is a strain of critical cognitivism in Descartes'reply to the Second Objections. There he justifies the principle that 'Every

idea needs -126Critical cognitivism arouses misgivings in many. I shall spend the rest of this section discussing the two strongest objections to it I have encountered. The first comes from Alston, who in the following passage is criticizing the view that epistemic propositions may be immediately justified, but whose remarks can be adapted against critical cognitivism as well: In taking a belief to be justified, we are evaluating it in a certain way. And, like any evaluative property, epistemic justification is a supervenient property, the application of which is based on more fundamental properties. . . . Hence in order for me to be justified in believing that S's belief that p is justified, I must be justified in certain other beliefs, viz., that S's belief that p possesses a certain property, Q, and that Q renders its possessor justified. (Another way of formulating this last belief is: a belief that there is a valid epistemic principle to the effect that any belief that is Q is justified.) Hence in no case can an epistemic belief that S is justified in believing that p, itself be immediately justified. 70 Alston is claiming that in order to come to be justified in believing an epistemic proposition, I must first be justified in believing an appropriate epistemic principle. If he is right about this, the justification of epistemic principles cannot be posterior to that of epistemic propositions, contrary to what is envisaged by critical cognitivism. Things would have to be the other way around. Let us grant that justification is a supervenient property. 71 There are still two lines of reply to Alston's argument. Using 'Jp' to abbreviate 'S is justified in believing that p', we can symbolize his main premiss as follows: JJp requires ∃ Q{JQP & J[(p)(QP → JP)]}. Now the first reply is this: why is it not sufficient for JJp to have J{∃ Q[QP & (p)(QP → JP)]}? In other words, why is it not sufficient to be justified in believing that p possesses some property Q that renders its possessor justified without knowing which property it is? The job of the critical cognitivist's argument would then be to identify the property Q. I find this reply quite plausible. Sometimes I reflect that I know a certain proposition, then ask myself how I know it (what makes me justified). I can ____________________ to have some really existing cause of its objective reality' by saying 'The admission of

this axiom is highly necessary for the reason that we must account for our knowledge of all things, both of sensuous and of nonsensuous objects, and do so by means of it alone' ( H R II56 [AT VII 165 : C S M II116]. 70 Alston, "'Two Types of Foundationalism'", 170. 71 What I am granting is that justification is supervenient in the sense that its instantiation depends on that of non-epistemic properties. If Alston's point in calling justification supervenient is that its warranted ascription is always based on nonepistemic properties, I am challenging it. -127be justified in my initial reflection even before I have successfully answered the question it provokes. (The analogue in ethics would be knowing that an act is right without yet knowing what makes it right.) The second reply is this: if it is a true epistemic principle that (p)(Qp → Jp), then a sufficient condition for JJp would be QJp, which does not itself require either JQp or J[(p)(Qp → Jp)]. Hence, in order to defend his main premiss, Alston needs the additional, but unstated, premiss that no epistemic proposition can ever possess a property like Q. In short, he must rule out QJp. But why can't we have QJp? Suppose that Q is the property of being clearly and distinctly perceived; why can't I clearly and distinctly perceive not only that 2 + 3 = 5, but also that it is evident to me that 2 + 3 = 5? 72 Unless this alternative can be ruled out, Alston's argument is inconclusive. The other objection to critical cognitivism is that it seems unsatisfyingly arbitrary and ad hoc. I do not know how to allay this misgiving, except by pointing out, as Chisholm does, that the conclusion of a critical cognitivist argument is (sometimes, at least) the consequence of premisses that are individually quite plausible. For example, isn't it certain that I am (at least) a thinking thing? And what else could render me certain of it, if not the fact that I clearly and distinctly perceive it? Before moving on, I want to make three comments about the status of epistemic principles that are justified by critical cognitivist arguments. First, since knowledge of epistemic principles is not on this view a prerequisite for knowing that you know, there may be few occasions, if any, on which knowledge of them is called for. But this is not to say that they are idle; on the contrary, they have to be operative if there is to be knowledge at all. Second, every critical cognitivist argument has one premiss that affirms that all possible sources but one of a given type of knowledge are barren. But it is hard to be sure that one has considered all possible sources, and harder still to be sure that one has considered all possible ways in which knowledge might arise from a given source. Therefore, the justification of epistemic principles afforded by critical cognitivism is bound to be somewhat tentative and conjectural. Third, it is sometimes suggested that epistemic principles are synthetic and a priori. 73 But a critical cognitivist should not say this, for the argument he offers ____________________

72

Russell makes a similar suggestion on pp. 381-2 of Human Knowledge ( New York: Simon and Schuster, 1948): 'The degree of credibility attaching to a proposition is itself sometimes a datum. I think we should also hold that the degree of credibility to be attached to a datum is sometimes a datum.' 73 e.g. by Chisholm in ch. 7 of Perceiving. But Chisholm also tends to favour critical cognitivism. Perhaps one could combine these positions by saying that epistemic principles are discovered through critical cognitivism, but justified by immediate intuition. -128makes essential use of the a posteriori premiss that we do know various things. If epistemic principles are a priori for him in any sense at all, it can only be in the Kantian sense that they are necessary presuppositions of knowledge. XII The last possibility to be discussed is that epistemic principles are immediately justified. The historical Descartes did not countenance this possibility, but I can imagine another Descartes meditating to himself as follows: 'By reflecting on my condition when I clearly and distinctly perceive that 2 + 3 = 5, I can see that my condition is one of knowing. Moreover, I can see that any state of clear and distinct perception would have to be a state of knowing. There could no more be a state of clear and distinct perception that was not a state of knowing than there could be adjacent mountains that did not enclose a valley.' For this Descartes, the C&D Rule is an immediately justified necessary truth. Other foundationalists have claimed a similar status for their epistemic principles. 74 Lehrer objects strenuously to the imputation of immediate justification to epistemic principles. 75 Such a manœuvre, he says, has the following disadvantages: (i) it makes the choice of epistemic principles 'arbitrary'; 76 (ii) it 'entirely begs the question in favor of the foundation theorist', and 'lacks dialectical cogency'; 77 and (iii) it 'opens the door to the most rampant forms of speculation'. 78 What can the foundationalist say in reply? i. To say that the choice of principles is 'arbitrary' is to say that there is no justification for choosing one set of principles rather than another, and Lehrer's ground for saying this is presumably that the foundationalist offers no reason in support of his choice. But if this is his ground, then Lehrer is presupposing that all justification is mediate-that nothing can be justified unless there is some further proposition that supports it. This, of course, begs the question against the foundation theorist. ____________________ 74 e.g. Chisholm in ch. 7 of Perceiving. 75 He also objects to the claim that epistemic principles are necessary truths, suggesting that a sceptic who denied them would not be contradicting himself. This is true if it means that a formal contradiction cannot be deduced from the sceptic's denial by appeal to logic and meanings alone. But not all necessary truths have denials that are self-contradictory in this sense. On the other hand, if (as he seems to be) Lehrer is using 'self-contradictory' in the broad sense in which it is synonymous with 'necessarily false', then it is no longer obvious that the sceptic's denial is not selfcontradictory. 76 Lehrer, Knowledge, 143-4.

77 78

Ibid. 121. Ibid. 152-3. -129-

ii. In laying down a generation principle, a foundationalist affirms that the propositions specified in its consequent are immediately justified whenever its antecedent is true. If a coherentist, who denies that any propositions are immediately justified, asks why he should accept such a principle, he will naturally be unsatisfied if he is told that it is itself immediately justified. So Lehrer is right: the foundationalist does beg the question, and his reply is dialectically ineffective. But let us see if the coherentist can argue with greater dialectical effectiveness. He, too, espouses an epistemic principle, one that takes the following form: if a proposition coheres with a system of propositions of kind K, then it is justified. 79 Why should we accept a principle of this type? The coherentist dare not say that it is self-evident, lest he make a fundamental concession to the other side. It is more likely that he will say that it is justified by its coherence--in which case he is as guilty of begging the question as the foundationalist. I mention this not to indulge in a tu quoque, but in order to bring out the fact that in the foundationalism--coherentism dispute, as in most matters of fundamental disagreement, it may be impossible for either side to support its view without begging the question against the other. It is hard for either side to get a dialectical grip on the other. One should not draw sceptical or relativist conclusions from this, however. One side may be in the right-and know that it is--even if it is incapable of demonstrating this to the other. iii. Let us consider, finally, the objection that foundationalism opens the door to speculation. If the foundationalist claims that his principles are immediately justified, then what is to prevent, let us say, a revelationist from claiming the same status for a principle to the effect that if S has an ostensible revelation that P, then S is justified in believing that P? The answer is that there is nothing to prevent this, but the foundationalist need not agree that the revelationist's principle is justified. Some claims to immediate justification are spurious. Now let us see how speculation fares within the coherence theory. It is a consequence of this theory that any belief might be justified just so long as the believer enlists an appropriate cast of supporting beliefs. 80 Suppose, for instance, that someone entertains the notion that the world is made of cottage cheese, and that the rest of his beliefs form a system of kind K including the following items: the cosmos flowed forth from the teats of a cow and curdled; when I sink my spade into the earth, up come mounds of ____________________ 79 Ibid. 54. Different ways of specifying what the coherence relation is and what a system of kind K is yield different coherentist principles, but the point I am making applies to them all.

80

And, in Lehrer's version, is a disinterested truth-seeker. See Knowledge, 189-90. -130-

creamy white stuff; and so on. What is to prevent this? Again, nothing. Moreover, there is an important difference between this situation and the one described in the last paragraph: the foundationalist did not have to agree that the revelationist's beliefs were justified; but the coherentist does have to agree that the beliefs of our eccentric cosmologist are justified (for him, at least), because they do, after all, satisfy the antecedent of the coherence principle. If anything, then, coherentism opens the door to speculation wider. I conclude that there is nothing inherently objectionable about claiming that epistemic principles are immediately justified. Of course, this status can be claimed more plausibly for some principles than for others. Perhaps few would hesitate to classify as immediately justified the principle that deduction transmits justification. And I would not hesitate to add to the list two generation principles: to cover the first truths a posteriori, Chisholm's principle (A), and to cover the first truths a priori, a version of Descartes' C&D Rule, modified so as to be a rule of prima-facie justification only. But to classify non-deductive transmission principles as immediately justified does not come as easily. XIII The results of Part Two may be summarized as follows. Coherentists have objected that epistemic principles must be justified, but cannot be if foundationalism is true. I have argued for just the opposite: epistemic principles need not be justified, but can be if foundationalism is true. I considered three ways in which this might occur. The first, naturalism, is perhaps the most exciting, but also, alas, the least promising. It overlooks the distinctively epistemic dimension of epistemic principles. The third, immediate justification, incurs no objection in principle, but very few epistemic principles can be claimed to be justified in this way. The second, critical cognitivism, is in some ways unsatisfying, but it, too, incurs no objection in principle, and may be the only alternative for non-deductive transmission principles. 81 ____________________ 81 But there are other approaches I have not discussed: e.g. according to the theory of meaning defended by John L. Pollock in Knowledge and Justification ( Princeton: Princeton University Press, 1974), epistemic principles would turn out to be true in virtue of the meanings of their constituent non-epistemic concepts. For the idea of combining Cartesian and contemporary issues in one paper I am indebted to Stephen Leeds, who suggested it to me in conversations and in a note ( "'Foundationalism and the Cartesian Spiral'"). For helpful criticisms of an earlier version, I am indebted to Ernest Sosa. -131-

V DESCARTES ON THE WILL ANTHONY KENNY The problems I wish to discuss concern two developments which took place, or are alleged to have taken place, in Descartes' teaching concerning the relation of the intellect and the will. The first is this. In the Regulae ( A T X420 [ C S M I45]) Descartes says that we must distinguish the faculty of the intellect by which it intuits and knows things from the faculty by which it makes affirmative and negative judgements. 1 In the same book he says that belief in revelation, by contrast with science, 'is not an act of the mind but of the will'. 2 On the other hand, in the Principles of Philosophy we are told that, while sensation, imagination, and pure intellection are modes of perception, desire and aversion, affirmation, denial, and doubt are different modes of willing. 3 This classification of affirmative and negative judgements as an act of will is anticipated in the Fourth Meditation, where he distinguishes the intellect or faculty of knowing (facultas cognoscendi) from the will or faculty of choosing (facultas eligendi), and says that the intellect merely perceives ideas for judgement 4 and that judgements are acts of the will;5 the cause of erroneous judgement is the fact that our will extends further than our intellect.6 In the same passage of the Meditations occurs the sentence which introduces the second of the two contrasts I want to discuss. 'I could not refrain from judging', Descartes says, 'that what I so clearly understood was true . . . because from a great light in my intellect there followed an inclina____________________ Anthony Kenny, "'Descartes on the Will'" in R. J. Butler (ed.): Cartesian Studies ( Blackwell, 1972), 1-31, by permission of Blackwell Publishers. 1 'Distinguamus illam facultatem intellectus per quam res intuetur et cognoscit ab ea qua judicat affirmando vel negando.' 2 'Non ingenii actio sit sed voluntatis' ( A T X370 [ C S M I15]). 3 'Sentire imaginari et pure intelligere sunt tantum modi percipiendi; ut et cupere aversari, affirmare, negare, dubitare sunt diversi modi volendi' ( A T VIIIA17 [ C S M I204]). 4 'Per solum intellectum percipio tantum ideas de quibus iudicium ferre possum' ( A T VII56 [ C S M II39]). 5 'Illos actus voluntatis, sive illa iudicia, in quibus fallor' ( A T VII60 [ C S M II42]). 6 'Latius pateat voluntas quam intellectus' ( A T VII58 [ C S M II40]). -132tion of will.' 7 The assertion that the will is determined by the intellect is generalized in the geometrical exposition of the Meditations which follows the Second Objections. 'The

will of a thinking substance is impelledvoluntarily and freely, since that is of the essence of the will, but none the less infallibly--towards a good clearly known to it. 8 And in the reply to the same objections he gives examples of propositions which are so clearly perceived by the intellect that we cannot think of them without believing them to be true ( A T VII145 [ C S M II104]). In a letter perhaps written to Mesland in 1645, on the other hand, Descartes wrote as follows: 'It is always open to us to hold back from pursuing a clearly known good, or from admitting a clearly perceived truth, provided we consider it a good thing to demonstrate the freedom of the will by so doing.' 9 We have, then, two contrasts. In the Regulae, Descartes treats judgement as an act of the intellect; in later works he treats it as an act of the will. In the Meditations Descartes says that clear perception determines the will; in the letter to Mesland he says that clear perception can be rejected by the will. I want to take each of these contrasts in turn, to see how far they represent a real change of mind in Descartes. I. DESCARTES' THEORY OF JUDGEMENT To the modern philosopher, the statement of the Regulae that judgement is an act of the intellect, seems more natural than the theory of the Principles that judgement is an act of the will. A practical judgement, a decision what to do, may perhaps be regarded as an act of the will; but a speculative judgement, a decision that such and such is the case, an assent to a proposition rather than to a proposal: this seems, if we are to talk of faculties at all, to belong to a cognitive, rather than to an appetitive, faculty. Such too was the opinion of the scholastics of Descartes' time. To see what he is likely to have been taught by the Thomists of La Flèche, we may consider the following text of St Thomas, one of many quoted in Gilson Index Scolastico-cartésien. In Summa theologiae, Ia IIae 17, 69 St Thomas inquires whether the act of reason can be commanded by the will. He replies as follows: ____________________ 7 'Non potui non iudicare illud quod tam clare intelligebam verum esse ... quia ex magna luce in intellectu magna consequuta est propensio in voluntate' ( A T VII59 [ C S M II41]). 8 'Rei cogitantis voluntas fertur, voluntarie quidem et libere--hoc enim est de essentia voluntatis--sed nihilominus infallibiliter, in bonum sibi clare cognitum' ( A T VII166 [ C S M II 117]). 9 'Semper enim nobis licet nos revocare a bono clare cognito prosequendo, vei a perspicua veritate admittenda, modo tantum cogitemus bonum libertatem arbitrii nostri per hoc testari' ( A T IV173 [ C S M K245]). -133Since reason reflects on itself, it can order its own acts just as it can order the acts of other faculties; and so its own act can be commanded. But it must be observed that the act of reason can be considered in two ways. First we may consider the exercise of the act: in this sense the act of

reason can always be commanded, as when someone is told to pay attention and use his reason. Secondly, we may consider the object of the act; and in this connection there are two different acts of reason to be considered. The first is the apprehension of truth about something; this is not in our power but comes about in virtue of a natural or supernatural light; and so in this respect the act of reason is not in our power and cannot be commanded. But there is another act of reason which consists in assent to what is apprehended. Where what is apprehended is something like the first principles to which the intellect naturally assents, to assent or dissent is not in our power but in the order of nature; and so, strictly speaking, it is at the command of nature. But there are some things apprehended which do not so convince the intellect as to take away its power of assent or dissent; these leave it free at least to suspend its assent or dissent for some cause; and in such cases assent and dissent are in our power and subject to command. 10 Any reader of Descartes will notice great similarities between his doctrine and that of St Thomas. Descartes distinguishes frequently between what we perceive or understand and what we judge: out of many passages, we can quote the letter to Hyperaspistes: 'I have frequently observed that what men judge to be the case differs from what they really understand to be the case.' 11 Descartes' distinction between perception and judgement corresponds to Aquinas's between apprehension and assent. Like Aquinas, Descartes thinks that there are some truths which are perceived in such a way as to force assent; Aquinas calls these 'first principles', Descartes calls them 'clear and distinct perceptions'. As he wrote to Regius in 1640, our mind is of such a nature that it cannot fail to assent to what is clearly understood. 12 Like Aquinas, Descartes thinks that in other ____________________ 10 'Respondeo dicendum quod quia ratio supra seipsam reflectitur, sicut ordinat de actibus aliarum potentiarum, ita etiam potest ordinare de suo actu; unde etiam actus ipsius potest esse imperatus. Sed attendendum est quod actus rationis potest considerari dupliciter: uno modo quantum ad exercitium actus, et sic actus rationis semper imperari potest, sicut cum indicitur alicui quod attendat, et rationi utatur. Alio modo quantum ad objectum, respectu cujus duo actus rationis attenduntur: primo quidem, ut veritatem circa aliquid apprehendat; et hoc non est in potestate nostra; hoc enim contingit per virtutem alicuius luminis vel naturalis vel supernaturalis. Et ideo quantum ad hoc actus rationis non est in potestate nostra, nec imperari potest. Alius autem actus rationis est, dum his quae apprehendit assentit. Si igitur fuerint talia apprehensa, quibus naturaliter intellectus assentiat, sicut prima principia, assensus talium vel dissensus non est in potestate nostra, sed in ordine naturae; et ideo, proprie loquendo, naturae imperio subjacet. Sunt autem quaedam apprehensa, quae non adeo convincunt intellectum, quin possit assentire vel dissentire, vel saltem assensum vel dissensum suspendere propter aliquam causam; et in talibus assensus vel dissensus in potestate nostra est, et sub imperio cadit.' 11 'Frequenter animadverti ea quae homines iudicabant ab iis quae intelligebant dissentire ( A M V52 [ A T III430: C S M K194]).

12

Mens nostra est talis naturae, ut non potest clare intellectis non assentiri' ( A T III64 [ C S M K147]). -134-

cases the mind is free to assent or dissent; but whereas Aquinas says that assent or dissent in such a case is at the command of the will, Descartes, as we have seen, regards assent as being, not just commanded by the will, but as itself an act of the will. I know of no scholastic prior to Descartes who held this view, and I am unimpressed by the attempts of Gilson and Koyré to find precedent for it in Augustine, Thomas, and Scotus. 13 When he wrote the Regulae, Descartes still held the orthodox Thomist view, as we have seen. There is, it is true, one passage which Leslie Beck sees as presupposing the later view 'that judgement whether in its pure or practical use is an assent or dissent, an act of will'. 14 But the passage is most naturally interpreted as applying only to practical matters. Descartes exhorts us, in studying, 'to think solely of increasing the natural light of reason, not with a view to solving this or that scholastic problem but in order that in all the happenings of our life, our intellect may show our will what alternative to choose'. 15 In this passage it is choice, and not judgement, not even practical judgement, which is attributed to the will; though if Descartes had called the choice following deliberation 'judgement', he would not have been departing from scholastic usage. 16 So pace Beck, the Regulae do not differ from the Thomist doctrine that judgement is an act of the intellect. Some time, then, between 1628 and 1640, Descartes changed his mind about the nature of judgement. It is not easy to discover when or why he did so. The Discourse on the Method is not helpful: it hardly mentions the will. In the Third Section, while urging the importance of following men's practice rather than their preaching, Descartes observes that many people do not know what they really believe. 'For the mental act of believing a thing is different from the act of knowing that one believes it; and the one act often occurs without the other.' 17 In his commentary Gilson cites here ____________________ 13 Koyré quotes a passage from Augustine in which a judgement is attributed to the will; but this concerns the act of faith, and not regular speculative judgement ( Essai sur l'idée de Dieu et les preuves de son existence chez Descartes ( Paris, 1922), 78). Gilson's best text for the assimilation between St Thomas and Descartes is the following: 'Conclusio syllogismi quae fit in operabilibus ad rationem pertinet, et dicitur sententia vel iudicium, quam sequitur electio; et ob hoc ipsa conclusio pertinere videtur ad electionem tanquam ad consequens.' This is far from Descartes' identification of judgement and election; and in any case applies only to practical reasoning (in operabilibus) ( ST Ia IIae, 13, 1 ad 2). 14 Beck, The Method of Descartes, 17. 15 'Cogitet tantum de naturali rationis lumine augendo, non ut hanc aut illam scholae difficultatem resolvat, sed ut in singulis vitae casibus intellectus voluntati

praemonstret quid sit eligendum' ( A T X361: C S M110]). 'Iudicium est quasi conclusio et determinatio consilii. Determinatur autem consilium primo quidem per sententiam rationis, et secundo per acceptationem appetitus ... et hoc modo ipsa electio dicitur quoddam iudicium a quo nominatur liberum arbitrium' ( ST I, 84, 3 ad 2). 17 'L'action de la pensée par laquelle on croit une chose, étant differente de celle par laquelle on connait qu'on la croit, elles sont souvent l'une sans l'autre' ( A T VI23 [ C S M I]22]). 16

-135a passage from Regis: 'According to Descartes, the mental act by which we judge something to be good or bad is a function of the will, and the action by which we know that we have judged thus is a function of the intellect. It is no wonder if two functions, one of the intellect and one of the will, are different and can occur apart.' 18 Regis's annotation shows that what is here said is coherent with Descartes' mature theory of judgement; but it does not establish that he already held it. First of all, the passage concerns practical and not speculative judgement; secondly, even someone who thinks that both belief and the knowledge that one believes are acts of the intellect can think it possible for one of them to occur without the other. 19 We may turn next to Descartes' unpublished writings for a clue in this matter. In 1630, in letters to Mersenne, Descartes put forward his famous doctrine of the creation of the eternal truths. It was the common scholastic doctrine that the truths of logic and mathematics were necessary in such a way that not even God could change them: they were not altogether independent of him, because they depended for their truth upon his essence or nature; but they did not depend on his free will in the way that the existence of the world did. Descartes argued that the scholastics talked of God 'as if he were Jupiter or Saturn, subject to Styx and Fate': in contrast, he insisted that it is God who has established these laws in nature just as a king establishes laws in his kingdom. The eternal truths 'are true or possible because God knows them as true or possible; they are not, contrariwise, known to God as true as though they were true independently of him . . . in God knowing and willing are but one thing; so that from the very fact of his willing something he knows it, and for this reason alone is such a thing true'. 20 Attempts have been made to find sources for this doctrine in Scotus and Ockham; like the attempts to find scholastic precedents for the theory of judgement, they are unconvincing. 21 Both the doctrine of the creation of ____________________ 18 'Exactement commenté par Pierre-Silvain Regis: 'Car il faut savoir que, selon M. Descartes, l'action de l'esprit par laquelle nous jugeons qu'une chose est bonne ou mauvaise est une fonction qui appartient à la volonté et que l'action par laquelle nous connaissons que nous avons jugé ainsi est une fonction qui appartient à l'entendement. Ou, ce n'est pas une grande merveille que deux fonctions, dont l'une appartienent à l'entendement et l'autre à la volonté soient différentes, et que l'une puisse être sans

l'autre' ( Gilson, Index Scolasticocartésien, 238). The difficulty is to reconcile what Descartes says here with Passions of the Soul, art. 19. 20 'Sunt tantum verae aut possibiles quia Deus illas veras aut possibiles cognoscit, non autem contra veras a Deo cognosci quasi independenter ab illo sint verae . . . en Dieu ce n'est qu'un de vouloir et de connaître; de sorte que ex hoc ipso quod aliquid velit, ideo cognoscit, et ideo tantum talis res est vera' ( A T I149 [ C S M K24]). 21 There is no real evidence for Koyré''s view that Descartes read Scotus in the 1620s. Even according to Koyré, Scotus held only that God could change moral laws such as the Decalogue, not logical or mathematical truths. 19

-136eternal truths and the theory that judgement is an act of the will are, of course, examples of a 'voluntarist' tendency--a tendency to attribute to the will (human or divine) things which might be attributed to something else (the intellect, or the nature); and such a tendency is to be found in Scotus and Ockham (the happiness of the blessed resides primarily in the will; good and evil are as they are because God so wills). But the resemblance seems to end there. And the connection, if there is one, between the two Cartesian doctrines, is fairly tenuous. It is true that if in God knowing and willing two and three to make five is one and the same act, then we have an act which can be regarded as at the same time an act of the intellect and of the will. But that this is the case with God, whose nature is simple and undivided, does not tell us anything about what is the case in man, where intellect and will are distinct. Descartes does not even use the word 'judgement' about God in this context, though no doubt if he did, he would say that in God judging, just like willing and creating, was identical with seeing, knowing, or understanding. Mersenne informed Descartes that his doctrine resembled that of the Oratorian P. Gibieuf, who published in 1630 his De Libertate Dei et Creaturae. Descartes, on receiving the book from Mersenne (cf. A T I153, 174, 220) had pleasure, as Baillet says, to find wherewithal to authorize what he conceived of indifference and freewill. But though Gibieuf's views on the liberty of God, and on the Jesuit doctrine of liberty of indifference, were very close to those of Descartes, there is nothing in his De Libertate to suggest that speculative judgement is an act of the will. On the contrary, the book reaffirms the traditional doctrine on this point. It is to be observed that liberty is an appetitive, not a cognitive faculty; because it is a faculty whose object is the end or the supreme good, which is an object of appetite not of intellect. It is no objection to this that it is called the faculty of free decision or judgement, and that judgement or decision is an act of reason. For it is called the faculty of free decision, both because it is moved by the free decision or judgement, and because free decision or judgement, when it is perfect and, as they say, practico-practical, includes its acceptance; not that its actual act is a Judgement or decision. 22

Altogether, I can find no passage in Descartes' letters prior to the writing of the Meditations which clearly teaches that judgement is an act of the will. This makes it the more surprising that when it is put forward there, ____________________ 22 'Observandum secundo, libertatem esse facultatem appetitivam, non cognoscitivam: quia est facultas cuius objectum est finis, sive summum bonum, quod est appetitus non intellectus. Nec refert quod vocetur facultas liberi arbitrii sive iudicii, iudicium autem sive arbitrium sit actus rationis. Vocatur enim facultas liberi arbitrii, tum quia movetur a libero arbitrio sive iudicio, tum quia liberum arbitrium sive iudicium (quando perfectum est et practice practicum, ut vocant) eius iam acceptationem includit; non autem quod actus eius elicitus sit iudicium ipsum sive arbitrium' ( Gibieuf, De Libertate, 355). -137it is not presented as a novel thesis which needs to be argued for, but is presupposed and applied before being stated in so many words. Only Gassendi seems to have objected, and that not in the Fifth Objections, but in his Instances, of which one is thus summarized: 'To avoid confusion the intellect and the will should be so distinguished that whatever concerns cognition and judgement should be attributed to the intellect, and whatever concerns appetition and choice should be attributed to the will.' 23 Being unable to find a historical source from which Descartes might have borrowed the doctrine that judgement is an act of the will, we must ask what philosophical considerations may have led him to devise it for himself. The one which first suggests itself is the fact that judgement, even speculative judgement, is, often at least, a voluntary matter. What we believe is influenced by our desires; rash judgement or stubborn incredulity is blamed as a moral fault; courage and effort may be required to retain rational conviction in the face of emotional pressures. In the controversy with Regius there is some evidence that this consideration was the origin of Descartes' theory. In the Notes on a Programme, Descartes objects to Regius dividing understanding into perception and judgement. I however saw that, over and above perception, which is required in order that we may judge, there must needs be affirmation or negation to constitute the form of judgement, and that it is often possible for us to withhold our assent, even if we perceive a thing. I attributed the act of judging, which consists solely in assent, that is in affirmation or negation, not to the perception of the understanding, but to the determination of the will. 24 In his doctoral thesis, La Doctrine Cartésienne de la liberté, Etienne Gilson argued, on the basis of this and other texts, that the origin of Descartes' theory of judgement was to be sought in his desire to adapt Aquinas's theodicy to his own purposes. The problem of evil presented itself to Descartes above all as the problem of error. There existed a set of

arguments in Aquinas to show how God could be exonerated from blame for human sin. By making judgement an act of the will, Descartes assimilated erroneous judgement to sinful volition. Thus he was able to use Aquinas's arguments to exonerate the author of nature from blame for ____________________ 23 'Vitandae confusionis gratia debere intellectum et voluntatem ita distingui ut quicquid cognitionis et judicii est, ad intellcetum pertineat;, quicquid appetitionis electionisque, ad voluntatem' ( A T VII404). 24 'Ego enim, cum viderem, praeter perceptionem, quae praerequiritur ut iudicemus, opus esse affirmatione vel negatione ad formam iudicii constituendam, nobisque saepe liberum esse ut cohibeamus assensionem, etiamsi rem percipiamus: ipsum actum iudicandi, qui non nisi in assensu, hoc est, in affirmatione vel negatione consistit, non retuli ad perceptionem intellectus, sed ad determinationem voluntatis' ( A T VIIIB363 [ C S M I307]). -138human fallibility. 'The problem of sin is the theological form of the problem of error and the problem of error is the philosophical form of the problem of sin.' 25 Two objections may be made to Gilson's thesis, one sound and the other unsound. The unsound objection runs as follows. It is just not the case that all judgement is voluntary in the sense of being avoidable. There are many judgements, as Descartes is the first to admit, that we cannot help making. Even the Notes on a Programme merely say 'nobis saepe liberum esse ut cohibeamus assensionem', and it appears disingenuous for Gilson to paraphrase this, as he does, 'nous savons par expérience que cette affirmation et cette négation sont toujours en notre pouvoir' ( La Doctrine Cartésienne, 276). This objection is unsound because Descartes does not consider it necessary for a judgement to be voluntary that it should be avoidable. Like most scholastics, Descartes was willing to call an act voluntary if it was in accordance with the agent's desires, whether or not it was avoidable; indeed, unlike most scholastics, he was prepared to call an unavoidable, but welcome, action 'free' as well as 'voluntary' ( A T IV116 [ C S M K 234]). 26 Moreover, it is undoubtedly true, as Gilson says, that in the Fourth Meditation Descartes does use, in the interest of theodicy, arguments very parallel to those of Aquinas. The crucial objection to Gilson's thesis is that it was not necessary, for Descartes to be able to exploit Aquinas's arguments, that he should have made judgement an act of the will; it was sufficient for him to make it a voluntary act of the intellect. In scholastic terminology, he did not need to regard judgement as an actus elicitus voluntatis; it was perfectly sufficient for him to regard it, as Aquinas himself did, as an actus imperatus a voluntate. 27 Not all voluntary acts are acts of the will: walking, for instance, may be a voluntary act, but it is an act of the body rather than of the will. Descartes might reject this example because, as he says often, nothing is completely in my power but my

thoughts (e.g. Discourse, part iii; letter 154). But imagination and intellectual thought are under the control of will--we can decide what we are going to think about--but are not acts of the will. ____________________ 25 'La problème du péché est la forme théologique de celui de l'erreur et la problème de l'erreur est las forme philosophique de celui de la péché' ( Gilson, La Doctrine cartésienne, 284). 26 Moreover any erroneous judgement was for Descartes voluntary in the sense of avoidable. 27 Descartes does not, so far as I know, use the pair actus elicitus, actus imperatus. But he frequently uses the terminology of eliciting acts (e.g. Fourth Meditation ( A T VII60 [ C S M II 42]); Comments on a Certain Broadsheet ( A T VIIIB363 [ C S M I307]), and the terminology of actus imperatus is implied in letter 463 [AT IV 162ff.: CSMK 241 ff.]. -139Our desires are of two sorts: one of which consists in the actions of the soul which terminate in the soul itself, as when we desire to love God, or generally speaking, apply our thoughts to some object which is not material. . . . When our soul applies itself to imagine something which does not exist, as when it represents to itself an enchanted palace or a chimera, and also when it applies itself to consider something which is only intelligible and not imaginable, e.g. to consider its own nature, the perceptions which it has of these things depend principally on the act of will which causes it to perceive them. Such perceptions, then, are voluntary; but they are perceptions of the intellect, not inclinations of the will ( Passions, arts. 18, 20). The problem, we might say, is not that error belongs to philosophy and sin to theology; it is that the object of the intellect is truth, and that of the will is goodness; that error is a matter of falsehood, and sin of badness. And this problem Gilson's theory is impotent to solve. Put in less scholastic terms, the problem is why, and with what justification, Descartes should lump judgement together with desire and aversion and separate it from perception and imagination. One reason might be that judgement and desire are, on Descartes' theory, the only acts which we perform if and only if we want to perform them. Walking is something which we do only if we want to; but not every time we want to walk do we succeed in walking. If we want to imagine something, on the other hand, we succeed in doing so; but we often have thoughts in our imagination which we do not will to be there. Neither of walking nor imagination, therefore, is it true that they are acts which we perform if and only if we want to. Judgement and desire, of which this is true, are therefore voluntary in a special way.

But there is a further reason for regarding judgement as an act of the will: and light may be thrown on this from an unexpected quarter. In modern times Frege has taught us to make a sharp distinction between the sense of a sentence, which remains the same whether a sentence appears as a complete unit of communication or as a hypothetical clause in a longer sentence, and the assertion of a sentence, which he marked by a special sign whose function was to indicate that the reference, or truth-value, of what follows it is 'the true'. 28 Many other writers have followed Frege, notably R. M. Hare, who in The Language of Morals made a distinction in sentences between a phrastic (which contains the descriptive content of the sentence) and a neustic (which marks the mood of a sentence, and of which Frege's assertion sign would be an example). 29 The two sentences ____________________ 28 The Philosophical Writings of Gottlob Frege, ed. P. T. Geach and M. Black ( Oxford: Blackwell, 1952), 62 ff. 29 R. M. Hare, The Language of Morals ( Oxford: Clarendon Press, 1952), 18 ff. In his more recent work Hare distinguishes what he called a neustic into neustic, tropic, and elistic. These -140'You will shut the door' and 'You, shut the door' have something in common--the state of affairs which would verify the prediction is the same as the state of affairs which would constitute obedience to the command: namely, your shutting the door. But as sentences in different moods they differ; and the similarity and differences might be brought out by rephrasing them Your shutting the door in the immediate future, please. Your shutting the door in the immediate future, yes. It is not difficult to see a similarity between Descartes' theory of judgement and the theory of Frege and Hare. The perceptions of the intellect, it might be said, are concerned with the unasserted phrastics; an affirmative judgement is, as it were, the mental attachment of the neustic 'yes' to the phrastic presented by the intellect, the mental attachment of the assertion sign to the Fregean 'sense' which is the object of perception. There is of course the difference that neither Frege nor Hare has a negative neustic; negation is regarded not as the polar opposite of assertion, but as the assertion of a phrastic with a negative sense, containing within itself the logical constant for negation. In two passages Descartes seems to make the contrast between phrastic and neustic in the scholastic terminology of matter and form. In the Third Meditation, having said that only those of his thoughts that are like pictures really deserve the name of 'idea', he goes on to say: 'Other thoughts have other forms in addition: when I will, am afraid, assert, or deny, there is always something which I take as the subject of my thought; but my thought comprises more than the likeness of the thing in question; of these some are termed volitions or emotions, others are termed judgements.' 30 The word 'subjectum' suggests to a modern reader the translation 'topic'; but in fact it is used in scholastic terminology as a

synonym for 'materia' in contrast to 'forma', which is explicitly used to refer to what differentiates a judgement from a pure idea. 31 In this passage judgements are contrasted with volitions rather than classified as a species of volition; this makes clear that Descartes uses 'volition' in a narrow sense as well as a broad sense, volitions, strictly so ____________________ refinements are not necessary for a comparison with Descartes, who does not make any analogous distinctions. 30 'Aliae vero alias quaedam praeterea formas habent; ut cum volo, cum timeo, cum affirmo, cum nego, semper quidem aliquam rem ut subjectum meae cogitationis apprehendo, sed aliquid etiam amplius quam istius rei similitudinem cogitatione complector; et ex his aliae voluntates, sive affectus, aliae autem iudicia appellantur' ( A T VII37 [ C S M II25-6]). 31 Cf. also Second Replies: 'distinguendum est inter materiam sive rem ipsam cul assentimur, et rationem formalem quae movet voluntatem ad assentiendum' [ A T VII147: C S M II105]. -141called, being a species of a genus of acts of will which includes also judgements. 32 In the other passage which uses scholastic terminology we are told very explicitly that judgement is an act of the will. This is in the passage already quoted from the Notes on a Programme, where Descartes is objecting to Regius's classification of mental phenomena. Then he divides what he calls the intellect into perception and judgement, which does not accord with my view. I observed that besides the perception which was required for judgement there must also be an affirmation or negation to constitute the form of judgement; and that it is often open to us to withhold our assent even if we perceive a thing. And so I attributed the act of judging, which consists purely in assent, i.e. affirmation and negation, not to the perception of the intellect, but to the determination of the will. ( A T VIIIB363 [ C S M I307]) The statement that the intellect is concerned with the unasserted phrastics needs some qualification; for Descartes uses 'intellect' no less than 'will' in two senses. In one sense the intellect is the possession of the power to recall and combine ideas; it is in this sense that every judgement presupposes an act of the intellect--since judgements must concern ideas, neustics must be attached to phrastics. In another sense the intellect is the faculty which produces clear and distinct ideas and intuits their truth; it is in this sense that Descartes can explain error by saying that the faculty of judging extends farther than the faculty of understanding. He explained this to Gassendi in the Fifth Replies ( A T VII376-7 [ C S M II259]).

When you judge that the mind is a rarefied body, you can understand that it is a mind, that is, a thinking thing, and you can understand that a rarefied body is an extended thing; but you do not understand that one and the same thing is both thinking and extended; this is something you merely will to believe because you believed it before and you do not like changing your mind. When you judge that an apple, which happens to be poisoned, is suitable food, you understand that its odour and colour etc. are pleasant, but not that it is a good thing for you to eat; but because you want it so, you judge it so. And so I agree that we do not will anything about which we understand nothing at all; but I deny that we understand as much as we will; because we can, about one and the same thing, will much and know very little. 33 ____________________ 32 On the basis of this passage Brentano argued that Descartes did not really regard judgement as an act of the will at all; he explains away principle 32 and the Notes on a Programme by saying that Descartes means judgement is an actus imperatus voluntatis and not an actus elicitus voluntatis. He does not seem to have noticed that in the Fourth Meditation Descartes speaks of 'eliciendos illos actus voluntatis, sive illa iudica, in quibus fallor' ( A T VII60 [ C S M II42]). The French makes even clearer the identity of the act of the will and the judgements: 'Dieu concourt avec moi pour former les actes de cette volonté, c'est à dire les jugements dans lesquels je me trompe' ( A T IX48). See Brentano, The True and the Evident ( London: Routledge, 1966), 28-32. 33 'Ita cum iudicas mentem esse tenue quoddam corpus intelligere quidem potes, ipsam esse mentem, hoc est, rem cogitantem, itemque tenue corpus esse rem extensam; unam autem et eandem esse rem quae cogitat et quae sit extensa, profecto non intelligis, sed tantummodo vis credere, quia iam ante credidisti. nec libenter de sententia decedis. Ita cum pomum, quod forte -142The ambiguity of Descartes' 'intellectus' can be paralleled in the English word 'understanding', which is the nearest word to Descartes' French 'entendement'. When we use the word 'understanding', like Locke, as a name for a very general faculty, we might say that the belief that eighteen is a prime number is an operation of the understanding; but of course in another sense no one can understand that eighteen is a prime number, since it is not. So, in this passage of the reply to Gassendi, Descartes restricts the verb 'intelligere' to his adversary's correct performances understanding what mind is, and what body is--and refuses to apply it to his mistaken idea that mind is a kind of body. There is a difference between phrastics and the 'materia' of Descartes' judgements, in that phrastics are unambiguously composite, propositional, containing argument and function; whereas the matter of Cartesian judgement is ideas, and ideas may be simple (e.g. the idea of mind) or composite (e.g. the idea 'horse with wings'), and composite ideas seem sometimes to be expressed nominally (idea of a horse with wings) and sometimes propositionally (idea that a horse has wings). Sometimes Descartes writes as if even a

non-propositional idea can be asserted; whereas of course it would be impossible to attach a neustic to a name standing alone. In such a case, presumably the assertion amounts to the assertion of an extramental existence of the thing represented by the idea. 34

The work of the intellect in the strict sense involves not only the understanding of ideas, but also seeing the combination between ideas (as, that thought is linked with existence). The intellect, in the wide sense, includes the imagination, whose function is to combine together the ideas of various bodily objects (forming, say, the idea of a goat-stag out of the idea of a goat and the idea of a stag). But it is not clear, in Descartes' system, what faculty is responsible for linking together non-corporeal ideas which do not belong together in reality: e.g. what links the ideas together in the idea that mind is a rarefied body? In Gassendi, one might think, it is the will that links these ideas together, just as it is the will which judges the composite idea so formed to be true. But this will not apply in the case of Descartes, whose will makes no such judgement, and who yet, in order to reject the judgement, has to put the two ideas together in the sentence 'The mind is not a rarefied body'. ____________________ venenatum est, iudicas tibi in alimentum convenire, intelligis quidem eius odorem, colorem, et talia grata esse, non autem ideo ipsum pomum tibi esse utile in alimentum; sed quia ita vis, ita iudicas. Atque sic fateor quidem nihil nos velle de quo non aliquo modo intelligamus; sed nego nos aeque intelligere et velle; possumus enim de eadem re velle permulta et perpauca tantum cognoscere' ( A T VII377 [ C S M II259). 34 Cf. Third Meditation ( A T VII36 [ C S M II26]); and the letter to Mersenne of 22 July 1641, in which Descartes says that all ideas not involving affirmation or negation are innate [ A T III 148: C S M K187]. -143The comparison between Descartes' perceptions and Hare's phrastics, then, though illuminating, needs qualification. Let us now turn to the other element, the neustic. Does a consideration of this throw any light on why Descartes considered judgement an act of the will? If the command 'Jones, shut the door' can be rewritten 'Shutting of the door by Jones, please', it seems that Jones's acceptance of, or assent to, this command, might be expressed by 'Shutting of the door by Jones, yes'. Elsewhere I have suggested that wishes, desires, and other pro-attitudes could be similarly expressed artificially by a unit consisting of a phrastic describing the approved state of affairs, and a neustic indicating the attitude of approval. Now of course when Jones agrees to the order 'Shut the door' by saying 'Yes', he means 'Yes I will', not 'Yes that is the case'. None the less, it is a striking fact that we can give an affirmative response not only to propositions and questions, but also to commands and projects, by the same word 'yes'. Our attitudes to both assertions and proposals may be described in terms of affirmation and negation; both may be characterized as 'assent' or

'dissent'; both as forms of commitment. Assent to both a proposition and a proposal may be sincere or insincere, rash or cautious, right or wrong. It is this, I think, which provides the main justification for Descartes' treatment of judgement as an act of the will. For what is it, after all, to ascribe particular actions to one or other faculty? It is to group those actions together in virtue of common features of description and assessment which apply to them. If we take together all those mental activities which can have rightness or wrongness ascribed to them, we will find that they include all those activities which Descartes ascribed to the will and exclude those which he ascribed to the intellect. But this justification of Descartes' procedure suggests immediately an objection to it. It may be wrong to think that the Earth is larger than the Sun, and wrong to have vengeful desires; but the wrongness in the one case consists in falsehood and in the other case in evil. The right, we might say grandly, is a genus of which the species are the true and the good; and Descartes' classification emphasizes the unity of the genus at the cost of ignoring the diversity of the species. It would be open to Descartes to make the following reply. It is indeed the case that judgement, unlike desires, can be classified as true or false. But it is not true that judgements cannot be classified as good and evil. Believing that the human mind, properly used, was infallible, Descartes believed that every erroneous judgement was a moral fault. 'What theologian or philosopher', he asked, 'or indeed what rational man has ever -144denied that we are in less danger of error the more clearly we understand something before assenting to it, and that it is a sin to make a judgement before the case is known?' 35 Moreover, the truth and falsehood which belongs to a judgement, Descartes might have said, belongs to it not in so far as it is an assent, but in so far as what is assented to--what is presented by the intellect--corresponds or does not correspond to reality. Erik Stenius has pointed out that unasserted phrastics possess a truth-value independently of being asserted; what is contained in an if-clause, for example, either is or is not a description of what is the case, even though, since it occurs in an if-clause, it is not being put forward as such a description. 36 This is a fact which is presupposed in the truth-tabular definition of the logical constants. The truth of assertions might be regarded as parasitic on this: an assertion is true if and only if what is asserted is true, i.e. is a description which corresponds to reality. In fact, Descartes does not answer along these lines. Instead he says: Ideas considered in themselves and not referred to something else, cannot strictly speaking be false; whether I imagine a she-goat or a chimera, it is not less true that I imagine the one rather than the other. Again, falsehood is not to be feared in the will or the emotions; I may desire what is evil, or

what does not exist anywhere, but it is none the less true that I desire it. Only judgements remain: it is here that I must take precautions against falsehood. 37 This, as I have remarked elsewhere, is a strange argument. 38 One could as well argue that judgements in themselves could not be false, on the grounds that whether what I judge is true or false, it is none the less true that I judge. I think this reveals a genuine confusion in Descartes. His theory of judgement involves an important insight which he failed to follow up. 39 The point which Descartes has missed is what we may call--to adapt an ____________________ 35 'Quis unquam vel Philosophus vel Theologus, vel tantum homo ratione utens non confessus est eo minori in errandi periculo nos versari, quo clarius aliquid intelligimus, antequam ipsi assentiamur, atque illos peccare qui causa ignota ludicium ferunt' (Second Replies ( A T VII147 [ C S M II105]). 36 E. Stenius, Wittgenstein's 'Tractatus' ( Ithaca, NY: Cornell University Press, 1960), 165 ff. 37 'Quod ad ideas attinet, si solae in se spectentur, nec ad allud quid illas referam, falsae proprie esse non possunt; nam sive capram, sive chimaeram imaginer, non minus verum est me unam imaginari quam alteram. Nulla etiam in ipsa voluntate, vel affectibus, falsitas est timenda; nam quamvis prava, quamvis etiam ea quae nusquam sunt, possim optare, non tamen ideo non verum est illa me optare. Ac proinde sola supersunt iudicia, in quibus mihi cavendum est ne fallar' ( A T VII37 [ C S M II26]). 38 Kenny, Descartes, 117. 39 A further example of the same confusion occurs in the reply to Gassendi: 'Cum autem prave iudicamus, non ideo prave volumus, sed forte pravum quid; nec quidquam prave intelligimus, sed tantum dicimur prave intelligere, quando iudicamus nos aliquid amplius intelligere quam revera intelligamus' ( A T VII377 [ C S M II259]). -145expression of J. L. Austin's--the 'onus of match'. 40 If we express assent to a proposition or a project in the phrastic-neustic form, each expression will contain a description of a possible state of affairs, plus an assent indicator. But let us suppose that the possible state of affairs does not, at the relevant time, obtain. Do we fault the assent, or the state of affairs? Do we condemn the original assent as a false assertion, or do we complain about the subsequent state of affairs as an unsatisfactory outcome? Elsewhere, I have tried to clarify this point by considering the different relation of an architect's plan, and a plan in a guidebook, to a building. 'If the building and the plan do not agree, then if the plan is in a guidebook, it is the plan which is wrong; if the plan was made by an architect, then there is a mistake in the building.' 41 In the relation between the guidebook and the building, the onus of match is on the plan; in the relation between the architectural drawing and the building, the onus of match is on the building. So, in general, in assenting to a proposition, we place an onus on a phrastic to match the world; in

assenting to a command or project, we place an onus on something non-linguistic (primarily, our own actions) to match a phrastic. Descartes, in lumping together affirmation and desire, negation and aversion, confounds the different onus of match involved in the two different kinds of assent and dissent. This, it seems to me, is the fundamental defect in his theory of judgement as an act of the will. The absence of the notion of onus of match at this point is the more surprising, as a very similar notion plays a fundamental part in Descartes' moral theory. 'My third maxim was to try always to conquer myself rather than fortune; to change my desires rather than the order of the world' ( Discourse, part iii [ A T VI25: C S M I123]). II. THE EVOLUTION OF DESCARTES' DOCTRINE OF FREEDOM Throughout the history of philosophy there have been two contrasting methods of expounding the nature of human free will. The first is in terms of power: we are free in doing something if and only if it is in our power not to do it. The second is in terms of wanting: we are free in doing something if and only if we do it because we want to do it. This is the distinction which ____________________ 40 J. L. Austin, "'How to Talk'", Philosophical Papers, ( Oxford: Oxford University Press, 1961) 190. 41 Kenny, "'Practical Inference'", Analysis, 26, Mar. 1968. The point was first made by Elizabeth Anscombe ( Intention [ Oxford: Blackwell, 1957] 56), who modestly but incorrectly attributes it to Theophrastus. -146Hume made when he urged us to distinguish 'betwixt the liberty of spontaneity, as it is call'd in the schools, and the liberty of indifference; betwixt that which is oppos'd to violence, and that which means a negation of necessity and causes' ( Treatise, III. ii. 2). Liberty defined in terms of wanting is liberty of spontaneity; liberty defined in terms of power is liberty of indifference. As Hume observed, the former, but not the latter, is compatible with causal determinism. In their accounts of human freedom most philosophers have combined both elements, and Descartes is no exception. In the Fourth Meditation we read: 'Freewill consists simply in the fact that we are able alike to do and not to do a given thing (that is, can either assert or deny, either seek or shun); or rather, simply in the fact that our impulse towards what our intellect presents to us as worthy of assertion or denial, as a thing to be sought or shunned, is such that we feel ourselves not to be determined by any external force.' 42 This appears tantamount to saying: 'Freewill consists in liberty of indifference, or rather in liberty of spontaneity.' One immediately wants to ask: what is the force of the 'or rather' here? Does it mark second thoughts, so that Descartes is withdrawing the statement that free will consists in liberty of indifference and replacing it with the more correct statement that it consists in liberty of spontaneity? Or does it mean that liberty of

indifference, properly understood, is identical with liberty of spontaneity so that the 'vel potius' means something like 'or, in other words'? The answer, I think, is not quite either of these: it is rather that Descartes thinks that free will often does consist in liberty of indifference, but that sometimes it consists only in liberty of spontaneity, and that is all that is essential to it. He goes on: 'There is no need for me to be impelled both ways in order to be free; on the contrary, the more I am inclined one way--either because I clearly understand it under the aspect of truth and goodness, or because God has so disposed my inmost consciousness--the more freely do I choose that way.' 43 In this passage there is a difficulty in the translation of the phrase 'in utramque partem ferri posse'. If this is taken to mean 'there is no need for me to be able to go both ways'--i.e. to act either way--then the sentence contains an outright denial that liberty of indifference is necessary for free will. Geach, however, takes the passive ____________________ 42 'Voluntas, sive arbitrii libertas, . . . tantum in eo consistit quod idem vel facere vel non facere--hoc est affirmare vel negare, prosequi vel fugere--possimus, vel potius in eo tantum quod ad id quod nobis ab intellectu proponitur affirmandum vel negandum, sive prosequendum vel fugiendum, ita feramur, ut a nulla vi externa nos ad id determinari sentiamus' ( A T VII57 [ C S M II40]). 43 'Neque enim opus est me in utramque partem ferri posse, ut sim liber, sed contra, quo magis in unam propendeo, sive quia rationem veri et boni in ea evidenter intelligo, sive quia Deus intima cogitationis meae ita disponit, tanto liberius illam eligo' (ibid.). -147sense of ferri seriously, and translates 'there is no need for me to be impelled both ways' [ A G96]--i.e. to have reasons on both sides. Taken this way, the sentence is not incompatible with the view that liberty of indifference is essential to genuine freedom; for a full-blooded liberty of indifference would be a freedom to act either way even though the reasons for acting might be all on one side. I think that Geach's rendering is correct: it is borne out by the French version of the Duc de Luynes which, we are told, was revised by Descartes himself. This reads: 'Il n'est pas nécessaire que je sois indifférent à choisir l'un ou l'autre des deux contraires.' At first sight, this too looks like a denial of the need for liberty of indifference; but in fact, when Descartes uses the word 'indifference', he does not mean what Hume and the scholastics meant by indifference. This point is made explicitly in the correspondence with Mesland, which we shall consider later; but it is clear enough from what follows in the Meditations: 'The indifference that I am aware of when there is no reason urging me one way rather than the other, is the lowest grade of liberty.' 44 But the indifference which is the balance of reasons is not the indifference which is the ability to act either way. The present text does not by itself tell us whether Descartes believed such an ability to remain when all the reasons are on one side. 'If I always saw clearly what is good and true, I should never deliberate as to what I ought to judge or choose; and thus, although entirely free, I could never be indifferent.' 45 The fact that I would not have to deliberate ('je ne serais jamais en peine de déliberer', as the French has it) if I always saw what was good

does not establish that I would always do what was good. So the indifference which is here said to be inessential to freedom is the indifference which consists in the balancing of reasons, and not the indifference which is the ability to act either way. However, shortly afterwards, in the case of the Cogito, Descartes expressly denies that such an ability exists. 'I could not but judge to be true what I understood so clearly; not because I was compelled to do so by any external cause, but because the great illumination of my understanding was followed by a great inclination of the will; and my belief was the more free and spontaneous for my not being indifferent in the matter.' 46 A truth ____________________ 44 'Indifferentia illa, quam experior cum nulla me ratio in unam partem magis quam in alteram impellit, est infimus gradus libertatis' [ A T VII58: C S M II40]. 45 'Si semper quid verum et bonum sit clare viderem, nunquam de eo quod esset iudicandum vel eligendum deliberarem; atque ita, quamvis plane liber, nunquam tamen indifferens esse possum' ( A T VII58 [ C S M II40]). 46 'Non potui quidem non iudicare illud quod tam clare intelligebam verum esse; non quod ab aliqua vi externa fuerim ad id coactus, sed quia ex magna luce in intellectu magna consequuta est propensio in voluntate, atque ita tanto magis sponte et libere illud credidi, quanto minus fui ad istud ipsum indifferens' ( A T VII58-9 [ C G M II41]). -148so clearly seen, then, cannot but be judged to be the case; so the ability not to judge, which in this case would constitute liberty of indifference, is lacking. Where there is no such clarity, however, indifference remains, and this is true not only where there are no reasons, or equal reasons, on either side, but wherever the reasons on one side fall short of certainty. For the thought of their uncertainty itself constitutes a reason on the other side. 'However much I may be drawn one way by probable conjectures, the mere knowledge that they are only conjectures and not certain and indubitable reasons is enough to incline my assent the other way.' 47 God, we are told, has 'given me the liberty to assent or not to assent to things of which he put no clear and distinct perception in my understanding.' 48 The Fifth Meditation and the Second Replies make clear that God has given me no such liberty in cases where I do have clear and distinct perception. 'There are some things which are so clear and simple that we cannot think of them without believing them to be true.' 49 The Seventh Axiom, quoted earlier, says: 'The will of a thinking thing is impelled, voluntarily of course and freely, since this is of the essence of the will, but none the less infallibly, towards a good clearly known to it.' 50 The Principles repeats and expands the doctrine of the Meditations. But when free will is first mentioned in principle 37, it looks as if Descartes is attributing liberty of indifference to the assent of clear truths. He writes: 'It is a supreme perfection in man to

act voluntarily or freely, and thus to be in a special sense the author of his own actions, and to deserve praise for them. . . . It is more to our credit that we embrace the truth when we do, because we do this freely, than it would be if we could not but embrace it.' 51 He goes on to say that in many cases 'we have power to assent or not assent at our pleasure'. 52 During the exercise of Cartesian doubt 'we were ____________________ 47 'Quantumvis enim probabiles conjecturae me trabant in unam partem, sola cognita quod sint tantum conjecturae, non autem certae atque indubitabiles rationes, sufficit ad assensionem meam in contrarium impellendam' [ibid.]. 48 'Mihi libertatem dederit assentiendi vel non assentiendi quibusdam, quorum claram et distinctam perceptionem in intellectu meo non posuit' ( A T VII61 [ C S M II42]). 49 'Quaedam sunt tam perspicua, simulque tam simplicia, ut nunquam possimus de iis cogitare, quin vera esse credamus' ( A T VII145 [ C S M II104]). 50 'Rei cogitantis voluntas fertur, voluntarie quidem et libere--hoc enim est de essentia voluntatis--sed nihilominus infallibiliter, in bonum sibi clare cognitum ( A T VII166 [ C S M II117]). 51 'Summa quaedam in homine perfectio est quod agat per voluntatem, hoc est libere, atque its peculiari quodam modo sit author suarum actionum, et ob ipsas laudem mereatur. . . . Magis profecto nobis tribuendum est, quod verum amplectamur, cum amplectimur, quia voluntarie id agimus, quam si non possemus non amplecti' ( Principles, part 1, art. 37). At first sight this appears to mean that when we embrace truth clearly seen, we are free not to embrace it; but from the sequel it is clear that this is not so. Perhaps Descartes means that in such a case the credit goes not to us but to the author of our nature, as the credit for the precise operation of a machine goes to its maker. 52 'Multis ad arbitrium vel assentiri vel non assentiri possimus' [art. 39]. -149conscious of freedom to abstain from believing what was not quite certain and thoroughly examined'. 53 However, this freedom does not hold in all cases, as soon transpires. It does not extend to things which are certain and examined: because there were some things which even to a Cartesian doubter were beyond doubt. The impossibility of withholding assent from clearly perceived truths is explicitly reasserted in principle 43. This (principle) is imprinted by nature on the minds of all in such a way that as often as we perceive something clearly, we spontaneously assent to it, and we cannot in any way doubt that it is true. 54 Despite a superficial impression, therefore, there is no difference of doctrine between the Meditations and the Principles. The Principles were published in 1644. On the 2nd of May of the same year, Descartes wrote to the Jesuit Denis Mesland, then in his final year as a theology undergraduate at La Flèche, a letter which contains his fullest treatment of the problem of free will. The

most important part of the letter is a commentary on the passage from the Meditations 'ex magna luce in intellectu sequitur magna propensio in voluntate' [From a great light in the intellect there follows a great inclination in the will]. Descartes agrees with Mesland that one can suspend one's judgement; but only by distracting one's attention; one cannot refrain from desiring a good clearly seen to be good. If we see very clearly that something is good for us it is very difficult--and on my view impossible, as long as one continues in the same thought--to stop the course of our desire. But the nature of the soul is such that it does not attend for more than a moment to a single thing; and so as soon as our attention turns from the reasons which make us know that a thing is good for us, we can call up before our mind some other reason to make us doubt of it, and so suspend our judgement, or perhaps even make a contrary judgement. 55 This is in perfect accord with the Meditations theory; indeed, it is simply an application to the will's function of pursuing the good of a principle expli____________________ 53 'Hanc in nobis libertatem esse experiebamur, ut possemus ab iis credendis abstinere, quae non plane certa erant et explorata' [ibid.]. 54 'Ita omnium animis a natura impressum est, ut quoties aliquid clare percipimus, ei sponte assentiamus, et nullo mode possimus dubitare quin sit verum' ( A T VIIIA21 [ C S M I207]). 55 'Voyant très clairement qu'une chose nous est propre, il est très mal aisé, et même, comme je crois, impossible, pendant qu'on demeure en cette pensée, d'arrêter le cours de notre désir. Mais, parce que la nature de l'âme est de n'être quasi qu'un moment attentive à une même chose, sitôt que notre attention se détourne des raisons qui nous font connaître que cette chose nous est propre, et que nous retenons seulement en notre mémoire qu'elle nous a paru désirable, nous pouvons représenter A notre esprit quelque autre raison qui nous en fasse douter, et ainsi suspendre notre jugement, et même aussi peut-être en former un contraire' ( A M VI144 [ A T IV116: C S M K2334]). -150citly stated in the Fifth Meditation, and more clearly in the Second Replies, about the will's other function of judging the truth. I am indeed so constituted that I cannot but believe something to be true at the time of perceiving it clearly and distinctly. But I am likewise so constituted that I cannot fix my mind's eye constantly on the same object so as to perceive it clearly; and the memory of a previous judgement often comes back to me when I am no longer attending to my arguments for having made it. Consequently, other arguments might now be adduced which would readily upset my view if I had no knowledge of God. 56

In the Meditations Descartes did not explain how the will falls into sin as explicitly as he explained how the will falls into error. This, he told Mesland, was because he wanted to stay within the limits of natural philosophy and not to involve himself in theological controversies. In this private letter, he is willing to be explicit. 'If we saw clearly (that what we are doing is evil) it would be impossible to sin as long as we saw it in that fashion; that is why they say that whoever sins does so in ignorance.' 57 This was no novelty, but something which he had said in private as early as 24 April 1637. Defending against Mersenne the statement in the Discourse that in order to do well it was sufficient to judge well, Descartes had adopted a familiar scholastic viewpoint. 'The will does not tend towards evil except in so far as it is presented to it by the intellect under some aspect of goodness--that is why they say that everyone who sins does so in ignorance. So that if the intellect never presented anything to the will as good without its actually being so, the will could never go wrong in its choice. But the intellect often presents different things to the will at the same time.' 58 This passage clearly implies that the will cannot go against the intellect unless the intellect itself is somehow on both sides of the fence at the same time. In such a case, of course, the perception of the intellect would be confused rather than clear and distinct; and so once again we can ____________________ 56 'Etsi enim eius sim naturae ut, quamdiu aliquid valde clare et distincte percipio, non possim non credere verum esse, quia tamen eius etiam sum naturae ut non possim obtutum mentis in eandem rem semper defigere ad illam clare percipiendam, recurratque saepe memoria iudicii ante facti, cum non amplius attendo ad rationes propter quas tale quid iudivavi, rationes aliae possunt quae me, si Deum ignorarem, facile ab opinione deiicerent' ( A T VII69 [ C S M II48]). 57 'Si nous le voyions clairement, il nous serait impossible de pçcher, pendant le temps que nous le verrions en cetter sorte; c'est pourquoi on dit que omnis peccans est ignorans' ( A M VI 145 [ A T IV117: C S M K234]). 58 'Voluntas non fertur in malum, nisi quatenus ei sub aliqua ratione boni repraesentatur ab intellectu, d'oú vient ce mot: omnis peccans est ignorans; en sorte que si jamais l'entendement ne représentait rien à la volonté comme bien, qui ne le fút, elle ne pourrait manquer en son élection. Mais il lui représente souvent diverses choses en même temps' ( A T I366 [ C S M K56]). Most commentators seem not to have noticed-and perhaps Descartes himself was not aware-that the dictum he here quotes approvingly, 'omnis peccans est ignorans' is a quotation from his adversary Aristotle (agnoei oun pas ho mochtheros, N. Eth. III. 1110b28). -151draw the conclusion that the will cannot resist the clear and distinct perceptions of the intellect. In the letter to Mesland Descartes ventures so far into theology as to discuss the merits of Christ. 'A man may earn merit, even though, seeing very clearly what he must do, he does it infallibly and without any indifference, as Jesus Christ did during his earthly life.' 59 How is this to be reconciled with the teaching of the Principles that we deserve no praise

for what we cannot but do? Descartes explains that the praise is for paying attention. 'Since a man has the power not always to attend perfectly to what he ought to do, it is a good action to pay attention and thus to ensure that our will follows so promptly the light of our understanding that it is in no way indifferent.' 60 The doctrine then is clear. In the face of clear and distinct perception, freedom to act in a contrary sense is possible only by inattention. In this letter Descartes makes a comparison between his terminology and that used by the scholastics, especially Jesuit scholastics such as those who taught Mesland. For him, indifference does not mean complete absence of knowledge; but the more the known reasons balance each other out, the more indifference there is. You regard freedom as not precisely indifference (in this sense) but rather as a real and positive power to determine oneself; and so the difference between us is a merely verbal one, since I agree that the will has such a power. However, I do not see that it makes any difference to the power whether it is accompanied by indifference, which you agree is an imperfection, or whether it is not so accompanied, when there is nothing in the understanding except light, as in the case of the blessed who are confirmed in grace. And so I call free whatever is voluntary, whereas you wish to restrict the name to the power to determine oneself only if accompanied by indifference. But so far as concerns names, I wish above all to follow usage and precedent. 61 ____________________ 59 'On ne laisse pas de mériter, bien que, voyant très clairement ce qu'il faut faire, on le fasse infailliblement, et sans aucune indifférence, comme a fait Jésus-Christ en cette vie' ( A M VI145 [ A T IV117: C S M K234]). 60 'Car l'homme pouvant n'avoir pas toujours une parfait attention aux choses qu'il doit faire, c'est une bonne action que de l'avoir, et de faire, par son moyen, que notre volonté suive si fort la lumière de notre entendement qu'elle ne soit point du tout indifférente' ( ibid. ). 61 'Ainsi, puisque vous ne mettez pas la liberté dans l'indifférence précisément, mais dans une puissance réelle et positive de se déterminer, il n'y a de différence entre nos opinions que pour le nom; car j'avoue que cette puissance est en la volonté. Mais, parce que je ne vois point qu'elle soit autre, quand elle est accompagnée de l'indifférence, laquelle vous avouez être une imperfection, que quand elle n'en est point accompagnée, et qu'il n'y a rien dans l'entendement que de la lumiére, comme dans celui des bienheureux qui sont confirmés en grâce, je nomme généralement libre, tout ce qui est volontaire, et vous voulez restreindre ce nom à la puissance de se déterminer, qui est accompagnée de l'indifférence. Mais je ne désire rien tant, touchant les noms, que de suivre l'usage et l'example' ( A M V144 [ A T IV116: C S M K234]). -152-

Indeed, in treating 'voluntary' and 'free' as synonymous, Descartes was following the precedent of Gibieuf De Libertate Dei et Creaturae. But Gibieuf was consciously going against the prevailing scholastic tradition which made a distinction between the two. According to most scholastics, the saints in heaven loved God voluntarily (because they did so willingly and not reluctantly) but not freely (since, clearly seeing the goodness of God, they could not do otherwise). 62 On this view, everything free was voluntary, but not everything voluntary was free; and the will, as such, was the capacity for voluntary action, and so not synonymous with the free will or liberum arbitrium. Thus far, Descartes' doctrine of liberty is all of a piece. But there remains one crucial document to consider. This is the letter listed by Adam and Milhaud as being written to Mesland on 9 February 1645; it is number 463 in their collection [ A T IV173: C S M K244]. M. Alquié, in his La Découverte metaphysique de l'homme chez Descartes, 63 and in the notes to the Garnier edition of Descartes' works, regards this as marking a decisive break in Descartes' thought. Now at last, in this letter, according to Alquié, Descartes admits that one can reject an evident perception at the moment of perceiving it. 'Il est donc possible, selon Descartes, (contrairement à l'avis de presque tous les commentateurs) de nier l'évidence en presence de l'évidence même de se détourner du bien sous le charme même de son attrait' [It is thus possible, according to Descartes (contrary to the view of almost all the commentators), to deny what is evident even in the presence of the evident, and to turn away from the good even when under the very influence of its attractive power] ( La Découverte, 289). Hitherto, it was only by ignorance or inattention that Descartes allowed the possibility of sin or error; in this letter, Alquié believes, Descartes' doctrine 'permet de refuser l'évidence et le bien en connaissance de cause' [allows the possibility of rejecting the evident and the good in full knowledge of the cause]. Alquié quotes from the letter: 'It is always open to us to hold back from pursuing a clearly known good or from admitting a clearly perceived truth, ____________________ 62 Gibieuf, De Libertate, 56. 'Voluntati qua natura inest sua libertas . . . Video responderi posse rationem liberi non esse rationem voluntatis, latiusque patere voluntatem quam libertatem: quippe voluntatem ad omne bonum se extendere, libertatem autem ad is tantum quod possit amari vel non amari cum indifferentia . . . sed si radix libertatis attente consideretur, facile erit non deprehendere solum sed convincere voluntatem nihil esse nise libertatem' [To the proposition that liberty is by its very nature present in the will, some may object that the scope of freedom is not that of the will, since the will extends further than freedom; for the will extends to everything good, but liberty merely to that which can be loved or not loved with indifference. In my view, however, if we carefully consider the basis of liberty, it will be easy to grasp with total conviction that the will Is nothing else but freedom]. 63 2nd edn. ( Paris: Presses Universitaires de France, 1966). -153provided we consider it a good thing to demonstrate the freedom of our will by so doing.' And he concludes:

Il ne s'agit pas, en la lettre de 9 février, de la faiblesse d'une attention se pouvant malaisement fixer sur un object unique, ni de l'élan qui, dans les Méditations, empêchait la conscience de se limiter à des objects finis et la portait vers l'infinie luimême par un perpétuel dépassement. Ce qui nous détourne du bien, c'est la mauvaise foi que commande l'égoïsme, et sans doute ce désir d'être Dieu qui, dans la Bible, apparaissait dé'jà comme la source première du péché. Ou'est en effet ce libre arbitre qu'il s'agit d'attester, sinon précisement nous-mêmes? [In the letter of 9 February, it is not a question of finding it hard, through the weakness of our mental attention, to keep focused OD a single object; nor is it a case of that impulse which, in the Meditations, stopped our awareness limiting itself to finite objects and carried us towards the Infinite itself by a continuous forward movement. What turns us away from the good is the kind of bad faith that is rooted in egoism, no doubt combined with the desire to be God, which the Bible has already revealed as the original source of sin. For when our free will is affirmed in this passage, what does it amount to if not our very selves?] No one, says Alquié, has tried to expound Descartes' doctrine of freedom in this letter 'sans essayer d'en affaiblir le tragique' [without trying to play down its tragic element]. I fear I must range myself with the commentators who have been insensitive to the tragedy. I observe, first, that to base on this letter a theory of an evolution in Descartes' thought is to build on sand. No one knows for certain to whom this letter was written or when. It is given by Clerselier in French as part of a composite letter to Mersenne whose other parts date from 1630 and 1637; but as it alludes to the Meditations, it must be later than 1640. Adam and Tannery printed it in their third volume in French as a letter to Mersenne with the hypothetical date of May 1641. Alquié says that at this date the letter 'serait incompréhensible, les affirmations qu'elle contient ne pouvant se situer qu'au terme d'une évolution de pensée comprenant elle-même les Principes et la lettre de 2 mai 1644, contemporaine de leur impression' [would be incomprehensible, since the claims which are made there could only arise at the end of a shift in Descartes' thought, to be found in the letter of 2 May 1644 and the Principles of Philosophy, published that month]. But as the only evidence for this evolution is Alquié's interpretation of this very letter, the progress of the evolution cannot be used to date it. Adam and Tannery later found a Latin text of the letter in a manuscript of the Bibliothèque Mazarine which gave it as a continuation of the letter in French to Mesland of 9 February 1645; accordingly they inserted it in their fourth volume after this letter, and it is retained in this place by Adam and Milhaud in their collection. On internal evidence there seems little doubt that this Latin text is more likely to be the original than the French -154text given by Clerselier; 64 but the attachment to the letter to Mesland is very dubious. All Descartes' letters to Mesland, as almost always to French-speaking correspondents, are in French, not in Latin. If this fragment belongs to the letter of 9 February, we have a change of language in the middle: why should a letter begun in French end in Latin?

Moreover, there is no illusion to the previous letter on free will to Mesland, though some of the same points are covered. It seems most likely that the compiler of the Bibliothèque Mazarine collection put together letters on transubstantiation and liberty, most but not all of which were to Mesland, rather in the way that Clerselier put together the composite letter to Mersenne from various draft documents he found among Descartes' papers. We must resign ourselves to the fact that we know neither the date nor the destination of this letter. Whatever the date of the letter, there is in fact no contradiction between its teaching and that of the earlier letter to Mesland. The passage on which M. Alquié builds his theory can easily be explained in accordance with Descartes' regular doctrine; and there are other passages which flatly contradict the interpretation put on the letter by Alquié. The letter is so short and so important that I propose at this point to insert a translation of the whole of it. Descartes (to Mesland) 9 February 1645 As for the freedom of the will, I entirely agree with what the Reverend Father here wrote. Let me explain my opinion more fully. I would like you to notice that 'indifference' seems to me to mean here the state of the will when it is not impelled one way rather than another by any perception of truth or goodness. This was the sense in which I took it when I said that the lowest degree of liberty was that by which we determine ourselves to things to which we, are indifferent. But perhaps others mean by 'indifference' a positive faculty of determining oneself to one or other of two contraries, that is to say to pursue or avoid, to affirm or deny. I do not deny that the will has this positive faculty. Indeed, I think it has it not only with respect to those actions to which it is not pushed by any evident reasons on one side rather than on the other, but also with respect to all other actions; so that when a very evident reason moves us in one direction, although, morally speaking, we can hardly move in the contrary direction, absolutely we can. For it is always open to us to hold back from pursuing a clearly known good, or from admitting a clearly perceived truth, provided we consider it a good thing to demonstrate the freedom of our will by so doing. It must be noted also that liberty can be considered in the actions of the will before they are elicited, or after they are elicited. Considered with respect to the time before they are elicited, it entails indifference in the second sense but not in the first. Although, when we contrast our own ____________________ 64 'A note in the Institut copy of Clerselier mentions both that the original of the fragment given by Clerselier is in Latin, and that the date and destination must be considered

unknown ( A T III378). -155judgement with the commandments of others we say that we are freer to do those things which have not been prescribed to us by others and in which we are allowed to follow our own judgement, we cannot similarly make a contrast within the field of our own judgements and thought and say that we are freer to do those things which seem neither good nor evil, or in which there are many reasons pro but as many reasons contra, than in those in which we see much more good than evil. For a greater liberty consists either in a greater facility in determining oneself, or of a greater use of the positive power which we have of following the worse although we see the better. If we follow the course which appears to have the most reasons in its favour, we determine ourselves more easily; if we follow the opposite, we make more use of that positive power; and thus we can always act more freely in those cases in which we see much more good than evil than in those cases which are called adiaphora or indifferent. In this sense too the things which are commanded us by others, and which we would not otherwise do spontaneously, we do less freely than the things which are not commanded; because the judgement that these things are difficult to do is opposed to the judgement that it is good to do what is commanded; and the more equally these two judgements move us the more indifference, in the first sense, they confer on us. But liberty considered in the acts of the will at the moment when they are elicited does not entail any indifference either in the first or second sense; because what is done cannot remain undone once it is being done. But it consists simply in ease of operation; and at that point freedom, spontaneity and voluntariness are the same thing. It was in this sense that I wrote that I took a course more freely the more reasons drove me towards it; because it is certain that in that case our will moves itself with greater facility and force [ A T IV173-5: C S M K244-6]. In the first part of this letter Descartes makes explicit a distinction between two senses of 'indifferent' which was implicit in his 1644 letter when he said that indifference did not imply ignorance, and that wherever there was occasion for sin there was ignorance. When he said that, Descartes clearly did not mean that there could only be sin where the reasons for acting were equally balanced on either side; consequently, he must have meant that there was an indifference which consisted in the possibility of acting against the weight of reason. It is this ability--which we might nickname 'the liberty of perversion' 65 --which Descartes now explicitly distinguishes from indifference in the sense of a balance of reasons. Adding this distinction to the distinction between two kinds of liberty which we saw in the 1644 letter, we get the following table. Liberty of spontaneity Liberty = voluntas

Perversion Liberty of indifference

Balance ____________________ 65 'Potestas . . . sequendi deteriora, quamvis meliora videamus' ( A M IV198 [ A T IV174: C S M K245]). -156So far the two letters are perfectly compatible. According to Alquié, the two letters differ crucially because, according to the 1644 letter, we do not enjoy liberty of perversion while we have a clear perception of good, whereas in letter number 463 we do so. However, it is perfectly possible to reconcile the two. When Descartes says in letter 463 that it is always open to us to hold back from pursuing a clearly known good, or from admitting a clearly perceived truth, he need not mean that we can do this at the very moment of perceiving the good and the true. Rather, we must distract our attention, as he said in the 1644 letter. One way of doing this would be to dwell on the thought that it would be a good thing to demonstrate our free will by perversity. This would provide a reason in the contrary sense, without which the will could not act; and eo ipso this would render the perception of truth and goodness unclear; we would, as he said in the 1644 letter, 'merely see confusedly that what we are doing is bad, or remember that we judged it so in the past'. In the 1644 letter he says that we can suspend our judgement 'by representing to our mind some reason to make us doubt of the truth'; letter number 463 suggests a reason one could use. 66 Alquié sees the possibility of such an interpretation, but says that it would satisfy a merely conceptual demand and would ignore the conflict of our life. But it is possible that Descartes shared a concern for merely conceptual demands; and after all, this is not a minor point in his system. To abandon the theory that clear and distinct perception necessitates the will is to call in question the whole validation of reason in which the Meditations culminates. The mark of a clear and distinct idea is that it is one which, however much we may exercise our free will, we cannot doubt; the only way to find truth is to stick to clear and distinct ideas; the only way to find out which ideas are clear and distinct is to do our damnedest to doubt them and fail to do so. But if clear and distinct ideas can be doubted at the moment they are intuited, we should never have genuine and certain knowledge of anything; we would be back in the morass of doubts of the First Meditation. The interpretation I have suggested is confirmed by the passage of the letter which immediately follows that on which Alquié rests his case. Descartes distinguishes liberty before the will's act, and liberty during the act. This suggests that we must make a further distinction in the chart we drew, a distinction between simultaneous and subsequent perversion. Thus we have ____________________ 66 Without some such reason, doubt would be impossible, as Descartes explained to Gassendi ( A T IXA204 [ C S M II270]).

-157Spontaneity Liberty = voluntas

Subsequent Perversion Indifference

Simultaneous Balance

Combining together the data of the two letters, we get the following results. The blessed in heaven and Christ on earth enjoy liberty of spontaneity, but no liberty of indifference, of any kind, not even the liberty of subsequent perversion. An ordinary man with a clear and distinct idea of what is true and what is good enjoys liberty of spontaneity and liberty of subsequent perversion, but not liberty of simultaneous perversion nor indifference in the sense of balance. An ordinary man with a confused idea of what is true and good enjoys liberty of simultaneous perversion; but only a man who sees no reason to one side rather than another enjoys the full indifference. Such a man, Descartes says, does not enjoy the liberty of spontaneity enjoyed by the others. 67 His argument for this will be considered in a moment. The final paragraph of letter 463 is reminiscent of an argument which occurs in Gibieuf De Libertate Dei against the Jesuits who define liberty in terms of an absolute indifference to act or not to act. 68 If that is what liberty consists in, Gibieuf argued, then a man never acts less freely than when he acts freely. For when a man acts, he is not indifferent with regard to acting, but is determined by his very act. To say that it is enough that he could not act when he was on the point of acting, Gibieuf argued, is to say that liberty is only for future acts qua future. On Gibieuf's own view, a man was free if he acted for the sake of the supreme good; and this could be true of him while he was actually acting. 69 ____________________ 67

Blessed C&D Prob. Balance Liberty of spontaneity Yes Yes Yes No (Yes) Liberty of subsequent perversion No Yes Yes Yes (No) Liberty of simultaneous perversion No No Yes Yes (No.) Liberty of balance No No No Yes 68 This is one of several indications that the Reverend Father mentioned at the beginning of the letter is, as Baillet said, Gibieuf. Another is that the definition of indifference Descartes attributes to the Reverend Father is the one Gibieuf uses, and not the Jesuit one which Mesland accepted and Gibieuf attacked. This suggests that Adam and Tannery's first thoughts on the dating of the letter were perhaps better than their second, since it was in the 1630s that we know Descartes was interested in Gibieuf's book. 69 'Quarta probation ducitur ex contradictione aperta quam includit natura libertatis, ut eam exponere consueverent per indifferentiam absolutam ad agendum et non

agendum. Si enim ea sit conditio libertatis, homo nunquam minus libere agit quam cum libere agit. Qui enim agit, non est indifferens ad agendum, sed determinatur actu suo' [The fourth argument is based on the plain contradiction that arises in the nature of liberty as commonly defined in terms of an absolute indifference to act or not to act. For if this is a necessary condition of liberty, a man -158I think that the argument of Gibieuf and Descartes is confused, though it is not easy to explain just where it goes wrong. I do not think, as I once did, that it depends simply on a fallacious inference of modal logic from It is not possible that both p and not-p to: If p, then it is not possible that not-p. In addition to modality, the argument involves subtle points about tense and action which it would take us too far round to investigate. But the paragraph does contain a point which tells against Alquié. Once one acts, Descartes says, the notions of liberty and spontaneity collapse into each other. This would be altogether untrue on Alquié's view, because in the case of someone with a clear and distinct idea acting with simultaneous perversity, there would be free action, but not spontaneous action; since the man would not be acting with that preponderance of reasons on his side which makes the operation easy and constitutes spontaneity. Finally, I wish to consider the penultimate paragraph of the letter. There is something very dubious about Descartes' argument here to show that indifference of balance is the lowest degree of liberty. He argues that a man enjoys least liberty when the reasons are balanced, because he then enjoys less liberty of spontaneity than when he acts in accord with the greater array of reasons, and less liberty of perversity than when he acts in accord with the lesser array of reasons. He could just as well have argued that a man was most at liberty when the reasons were balanced, since he then enjoys more liberty of perversity than when the majority of reasons are on his side, and more liberty of spontaneity than when the majority of reasons are against him. In fact, having once distinguished liberty of spontaneity from liberty of perversity, Descartes should have said that in the one sense of liberty, action in indifferent matters was freer than action upon clear reasons, and in another sense it was less free. But then he could not have said, what he also wanted to say, that there was a single scale of freedom on which the liberty of indifference occupied the lowest place. This shows that Descartes' theory of freedom will not do as a philosophical account; but it is an incoherence which was present in the theory from the beginning. The doctrine of the Meditations, the Principles, and the letters is all of a piece. I see no reason for thinking that at the age of 49 Descartes underwent a spectacular conversion from rationalism to existentialism. ____________________ never acts less freely than when he acts freely. For he who acts is not indifferent with respect to acting, but is determined by his act] ( De Libertate, 13; cf. also 165 ).

-159VI DESCARTES' THEORY OF MODALITY JONATHAN BENNETT I. INTRODUCTION Descartes propounded the allegedly strange, peculiar, curious, and incoherent doctrine that necessary truths are made true by God's voluntary act. 1 It seems to imply that God could have made necessary truths false, which entails that they are not necessary after all. Some of Descartes, interpreters have taken him in that way, as firmly implying that nothing is absolutely necessary or impossible. 2 We all know, however, that if this doctrine runs free through Descartes' philosophical work, it will do untold damage: many of his arguments have to be protected somehow from the thesis that so-called necessary truths are really contingent. Discussing an argument of Descartes' for the real distinction between body and mind, for example, Curley writes: 'If we were to invoke the doctrine of the creation of eternal truths, we might say that a really omnipotent being could cause the mind and body to exist apart even if that were not logically possible. But in the Meditations Descartes is careful not to invoke that extravagant conception of omnipotence, and we would do him no service by bringing it in.' 3 That typifies the kind of thing Cartesian scholars have felt forced to say, charitably shielding Descartes from his own splatter. This 'creation' or 'voluntarism' doctrine does not appear in the Meditations, the Discourse on the Method, or the Principles of Philosophy. Descartes first declared it in three private letters to Marin Mersenne, most of a decade before his first published work appeared. It glows luminously ____________________ Jonathan Bennett, "'Descartes' Theory of Modality'" from Philosophical Review 88 ( 1979), 63967. Copyright © 1979 Cornell University, by permission of the publisher and the author. 1 'Strange': Jolley, The Light of the Soul, 32. 'Peculiar': ibid. 49, and Louis Loeb, From Descartes to Hume: Continental Metaphysics and the Development of Modern Philosophy ( Ithaca, NY: Cornell University Press, 1981), 68. 'Curious': ibid. 67. 'Incoherent': Edwin Curley, Behind the Geometrical Method: A Reading of Spinoza's Ethics ( Princeton: Princeton University Press, 1988), 42. 2 See esp. Frankfurt, 'Descartes on the Creation of the Eternal Truths', 57. 3 Curley, Descartes Against the Skeptics, 198. -160just behind something Descartes wrote in reply to the Second Objections to the Meditations, as we shall see; and he announced it openly in his response to ( Gassendi's)

Fifth Objections. Descartes had described the essences of things as 'immutable and eternal'; Gassendi had demurred at giving that label to anything but God; and Descartes replied--unlimbering his voluntarism doctrine--that God causes everything that is immutable and eternal. The authors of the Sixth Objections, who had seen the Fifth Replies, challenged Descartes on his voluntarism, and he responded forthrightly. Furthermore, when first announcing the doctrine, he urged Mersenne to 'assert and proclaim [it] everywhere'. His reason for keeping the doctrine out of sight in the major published works was evidently not that he was unsure of its truth. How, then, can it be all right to attribute to him arguments and opinions which obviously conflict with it? Yet that is what Curley does, and what I have often done in lectures, crediting Descartes with thoughts which conflict with his 'voluntarist' thesis about necessary truths, and arguments which collapse under its weight. We have thereby implied things like this: Descartes offered an argument from which he concluded that his mind is one thing, his body another; if voluntarism is true, the argument fails; at the time of writing the argument, Descartes believed in voluntarism. This implies that Descartes behaved dishonestly. His reason for not revealing his voluntarism doctrine in the Meditations, whatever it was, cannot excuse his writing things that are condemned by it. ' Descartes offered his allimportant argument for "the real distinction" while believing that it falls to pieces when the whole truth about modality comes to bear on it'--no one should believe this. In fact, Descartes showed no awareness, ever, of voluntarism's threatening the rest of his work. Either he had a blind spot in this direction, or we have misunderstood his doctrine about necessary truths. I shall defend the latter option, showing how to read the voluntarism texts so that they do not obviously undermine Descartes' other work. The reading I shall offer has been adumbrated by others, but nobody has developed it explicitly and accepted it outright, as I shall do. 4 It fits nearly everything in the texts, one of which positively requires it; and the philosophical doctrine it attributes to Descartes harmonizes as well as possible with the rest of his philosophy, and also--dare I add?--has a chance of being true. My interpretation fits nearly everything in the voluntarism texts: this late ____________________ 4 For a useful discussion of previous work, ending with a hint of the interpretation to be offered here, see Lilli Alanen, 'Descartes, Omnipotence, and Kinds of Modality', in Peter H. Hare (ed.), Doing Philosophy Historically, ( Buffalo: Prometheus Books, 1988), 182-96. Another hint--more than a hint, really--will be reported at the end of Section III below. -161in the day nobody will discover that the texts consistently express a single view; nothing can clear them of the charge of wavering sometimes. Still, rereading them in the light of my interpretation, I have been struck by how little my Descartes has to apologize for.

Descartes' voluntarism on my account of it harmonizes as well as possible with the rest of his philosophy; I should explain that too. When wrestling with scepticism in the Meditations, Descartes had two understandings of what his problem was, and two lines of thought as solutions. He needed to choose one problem and one solution, but he did not do so, evidently because he was not properly aware of this doubleness in his thought. Now, voluntarism as I understand it nicely fits in with one of those two lines of thought; no coherent doctrine could harmonize with both. I shall explain all this in Section VI. II. GREATNESS AND OMNIPOTENCE The voluntarism doctrine comes in part from a religious impulse. Roughly and briefly: as a Christian believer wishing to magnify God as a being who is not limited in any way, Descartes went to the extreme of denying that there are even logical limits to what God can do. The other half of the story, which concerns a philosophical grounding for voluntarism, interests me more. Still, we should try to get the religious part right, as that will help us to understand the other. Some writers on this topic assume that the governing theological idea was that of God's omnipotence. Margaret Wilson, for example, gives to her very interesting section on this matter the title 'A God who can do anything'. 5 Geach also takes Descartes to be relying on the concept of omnipotence, and scolds him for it. The claim 'God can do anything', he says, is acceptable as piety, a vague way of magnifying God's greatness, but not as the premiss of any argument. 6 Descartes did not use it as a premiss. One of the voluntarism texts describes God in passing as all-powerful (tout-puissant), but not so as to derive voluntarism from the premiss that God can do anything. 7 The texts draw their energy from a thought of God not as omnipotent but as 'great' in such a way that everything that exists depends upon him. There are therefore--so the thought goes--no independent principles that constrain his actions, and no independent stand____________________ 5 Wilson, Descartes, (120 ff. The phrase comes from the First Meditation ( C S M II14) ('omnipotent'), A T VII21 (qui potest omnia)). 6 P. T. Geach, 'Omnipotence', Philosophy, 48 ( 1973), 7-20. See also Richard L. La Croix, 'The Impossibility of Defining "Omnipotence'", Philosophical Studies, 32 ( 1977), 181-90. 7 Third letter to Mersenne, ( A T I152: C S M K25); i.e. the third of the three letters in which voluntarism is launched [ 27 May 1630]. -162ards by which he can be judged. This lack of constraint is akin to the thought of God's omnipotence, but differs from it in not being open to Geach's criticism. The emphasis on everything's depending on God, this being demanded by piety and respect, shines out from the first letter to Mersenne: 'The mathematical truths that you call eternal have been laid down by God and depend on him entirely, no less than the rest

of his creatures. Indeed, to say that these truths are independent of God is to talk of him as if he were Jupiter or Saturn and to subject him to the Styx and the Fates.' 8 The emphasis remains in the second letter, three weeks later: The eternal truths . . . are not known as true by God in any way which would imply that they are true independently of him. If men really understood the sense of their words they could never say without blasphemy that the truth of anything is prior to God's knowledge of it. . . . So we must not say that if God did not exist nonetheless these truths would be true; for the existence of God is the first and most eternal of all possible truths and the one from which alone all others derive. It is easy to be mistaken about this because most people do not regard God as a being who is infinite and beyond our grasp, the sole author on whom all things depend. 9 This does not say or even hint that God can do anything. The same holds for a passage in the response to Gassendi: 'If anyone attends to the immeasurable greatness of God he will find it manifestly clear that there can be nothing whatsoever which does not depend on him. This applies not just to everything that subsists, but to all order, every law, and every reason for anything's being true or good.' 10 Here is a voluntarism text in which, unusually, Descartes mentions omnipotence: I do not think we should ever say of anything that it cannot be brought about by God. For since every basis of truth and goodness depends on his omnipotence, I would not dare to say that God cannot make a mountain without a valley, or that one and two should not be three. I merely say that he has given me such a mind that I cannot conceive a mountain without a valley, or an aggregate of one and two which is not three, and that such things involve a contradiction in my conception. 11 The phrase 'depends on his omnipotence', however, is driven by the thought of dependence, not that of omnipotence in the sense of the ability to do anything. The passage would not collapse if we replaced that phrase by 'depends on him'. The first sentence might seem to involve omnipotence, but it does not. It says For no x ought we to say that God could not ____________________ 10 Replies to Sixth Objections. ( A T VII435 f. C S M II293 f.). 11 Letter to Arnauld ( A T V224: C S M K358 f.). 8 First letter to Mersenne ( A T I145: C S M K23) [ 15 April 1630]. 9 Second letter to Mersenne ( A T I149 I: C S M K24) [ 6 May 1630]. -163-

do x, which differs from For all x. God could do x. I shall expound the difference in Section VII. The 'nothing independent' thought also led Descartes, as it has others, to hold that there are no moral standards that are independent of God's will. We have just seen the moral and the modal running in a single harness ('truth and goodness'), as they do again here: It is self-contradictory [repugnat] to suppose that the will of God was not indifferent from eternity with respect to everything that has happened or will ever happen; for it is impossible to envisage anything's being thought of in the divine intellect as good, or worthy of action or omission, prior to the decision of the divine will to make it so. For example, God did not will the creation of the world in time because he saw that it would be better this way than if he had created it from eternity; nor did he will that the three angles of a triangle should be equal to two right angles because he recognized that it could not be otherwise. On the contrary, it is because he willed to create the world in time that it is better this way; and it is because he willed that the three angles of a triangle should necessarily equal two right angles that this is true and cannot be otherwise. 12 The moral half of this has no direct link with 'God can do anything', and Descartes cannot have thought that it does. This further confirms that the same holds for the logical half. What fuels the passage is purely the view that everything true or real depends on God's will. Indeed, did Descartes believe that God can do anything? He warns us against 'arrogantly supposing our own minds to be so powerful and wise that we can attempt to grasp and set limits to what God can or should perform', 13 and that cuts two ways. Someone who is sure that God is omnipotent might well enjoin us not to 'set limits' to what God can do, but why should he say that we cannot 'grasp' what he can do? And why should he go on-as Descartes does--to warn us that 'God is infinite and beyond our comprehension'? Anyone who asserts 'God can do everything' implicitly claims to grasp what God can do. Furthermore, if Descartes thought that God is omnipotent, he was absurdly understating his position every time he said that 'God can bring about whatever we clearly perceive in a way exactly corresponding to our perception of it'. 14 III. GOD'S DOINGS AND OUR LIMITS In his voluntarist doctrine Descartes offered two things at once: a thesis about the greatness of God and a philosophical analysis of modal concepts. ____________________ 12 Replies to Sixth Objections ( A T VII431 f.: C S M II291), quoted with some. omissions. 13 Meditations, Preface to the Reader ( A T VII9: C S M II8). 14 Replies to Second Objections ( A T VII169: C S M II119).

-164I have discussed the former and the religious impulse that underlies it. The philosophical analysis may have been driven partly by that same impulse: Descartes' analysis of modality entails that if God created the world, then he is not limited by logic, so Descartes may have been drawn to it by its yielding a result that he wanted on religious grounds. But the analysis is independently plausible, and might be favoured even by an atheist; 15 so Descartes could have been confirmed in his 'no independent limits' theology by having a good-looking philosophical analysis that leads to it. I shall come to the analysis through this fact: in several voluntarism texts, including one already quoted, Descartes juxtaposes (i) a thesis relating necessary truths to God with (ii) a thesis relating them to us. If we illustrate (i) with It is not impossible for God to make an uphill without a downhill, then (ii) is illustrated by We cannot conceive of an uphill without a downhill. These two evidently struck Descartes as natural companions but why? What did he think they have to do with one another? Most previous writers accept this pairing without comment, as though its rationale were obvious. I do not find it so. If we understood (i) as entailing that there could be an uphill without a downhill because God could create one, the companion thesis could be (ii) that we (wrongly) think there could not be an uphill without a downhill because we cannot conceive of it. On this reading, (i) says there is only one modality, and (ii) explains why we think there are more. This links the two all right, but at the prohibitive price of a reading of (i) that turns voluntarism into a rogue elephant crashing destructively through the rest of Descartes' work. Rejecting that, let us ask again: What is (ii), and how does it connect with (i)? If we construe (i) as saying that God made it impossible for there to be an uphill without a downhill, and decline to infer that really it is not impossible at all, then we can reasonably conjoin this with (ii) understood as a thesis about modal epistemology. Thus, (i) says that God set up the modal truths, and (ii) says that he gave us limits to conception as a guide to what the modal truths are. The story runs like that in Margaret Wilson's treatment of this matter: 'God has created our minds in such a way that we cannot directly conceive the opposite of things he has willed to be necessary or eternal.' 16 In Wilson's view, (ii) God set limits to what we can conceive, as his way of giving us epistemic access to the modal facts, and (1) he established those facts in the first place. These two propositions sit side by side, merely conjoined. ____________________ 15 For a non-theistic version of something like it, see my 'On Being Forced to a Conclusion', Proceedings of the Aristotelian Society, supp. vol. 35 ( 1961), 15-34. 16 Wilson, Descartes, 127. Page references in the text to the end of this section are to this book. -165-

Wilson takes them to be parts of a single item which she calls 'the creation doctrine'. She remarks that (ii) could be involved in Descartes' treatment of the 'Deceiving God Hypothesis' in the Third Meditation, putting the point like this: Descartes's creation doctrine in itself requires him to think of the human understanding as limited and constrained by God in certain respects. And this naturally if not logically leads to the consideration that God, if perhaps malevolent, could in general be manipulating our thought to deprive us of access to the truth. (pp. 127 -8) She acknowledges, though, that this appeals only to (ii) God's way of informing us about what the modal truths are, and owes nothing to (i) his creation of them. Descartes does not need voluntarism in order to conclude that God constrains our understandings, and could through those constraints mislead us. Anyway, at this stage in the Third Meditation, Descartes did not need to assert that proposition, but only to say that he did not know it to be false. In this part of the argument, voluntarism sits idle. ( Wilson also finds another role for it in the treatment of scepticism; I shall come to that in my final section.) A bit before this, however, Wilson sketches a treatment of voluntarism which relates (i) integrally to (ii). She writes: ' Descartes did regard the "necessity" we perceive in mathematical propositions as in some sense and degree a function of the constitution of our minds--themselves finite "creatures"' (p. 125 ). So, necessity is a function of how the human mind is; the mind is a creature; so necessity is a creature too. That would unite the theses (i) that God created modal truths and (ii) that we cannot conceive of impossibilities. They would come together in the doctrine that God made us unable to conceive of certain things, and this was his making those things impossible. We could say that God created modal truths by making us unable to conceive of impossibilities, but let us be careful not to think of this in terms of means to ends. I shall start to develop that line of thought in my next section. Wilson herself does not make anything of it. Rather, she continues in a way that seems to change the subject: [This] position was far from wild, or excessively idiosyncratic. From Hume and Kant onward it has been widely held that alleged perceptions of 'necessity' cannot be taken for granted, and that we must in some sense or other have recourse to the structure and workings of our own minds to give an account of these 'perceptions'. (p. 125 ) These remarks ostensibly about our perceptions of necessity were probably meant to stay with the theme of the necessity that we perceive. Obviously, we did not need Hume or Kant to tell us that our perceptions of necessity -166depend on the constitution of our minds. Be that as it may, Wilson does nothing more with her passing suggestion that Descartes had a subjectivist theory about the nature of

necessity--not perceptions of necessity but the thing itself. I shall now pick up that suggestion and run with it. IV. A CONCEPTUALIST ANALYSIS OF THE MODAL CONCEPTS Descartes held, I submit, that our modal concepts should be understood or analysed in terms of what does or does not lie within the compass of our ways of thinking. Roughly speaking: 'It is absolutely impossible that P' means that no human can conceive of P's obtaining while having P distinctly in mind; and similarly for P's possibility and its necessity. In each of these analyses, 'no human can' must be understood in causal, psychological terms, and not as involving the absolute or logical modalities that are being analysed. On this account, the statement (ii) about what we can conceive provides all the content we are entitled to give to our modal statements; and that is why (i) we ought not to think of modal truths as a part of what God has to reckon with. The authors of the Second Objections said that Descartes' a priori argument for God's existence requires the further premiss that the concept of God which it employs is a possible one. His response provides the strongest evidence I have for my reading of the voluntarism texts: If by possible you mean what everyone commonly means, namely whatever does not conflict with our human concepts, then it is manifest that the nature of God, as I have described it, is possible in this sense because . . . [etc., etc.]. Alternatively, you may well be inventing some other kind of possibility which relates to the object itself; but unless this matches the first sort of possibility it can never be known by the human intellect, and so it . . . will undermine the whole of human knowledge. 17 Where I put 'inventing', C S M has 'imagining' (the verb is fingere). Either way, Descartes is treating the 'possibility which relates to the object itself' as a contrivance, something faked up for purposes of argument rather than part of our natural conceptual repertoire. He denies it a life of its own: if the objective concept does not keep in step with the subjective one, he says, it will be direly subversive. He describes the subjective concept of possibility, which makes it a relation to our concepts, as the common meaning of the term 'possible'. So he is offering an analysis of modality, a ____________________ 17 Replies to Second Objections ( A T VII150f.: C S M II107). -167conceptualist analysis--taking 'concepts' to be aspects of the human condition, of course, not entities belonging to a Fregean third realm. This analysis provides a solid basis for Descartes' modal epistemology. He frequently moved from something's being distinctly conceivable to its being possible, as when he

wrote: 'Everything which I clearly and distinctly understand is capable of being created by God so as to correspond exactly with my understanding of it.' 18 We have already seen one striking instance of this--run contrapositively--in the passage about God's lack of moral limits, where Descartes calls something self-contradictory because he cannot envisage it. That conforms to the spirit of his writing that '[w]e should think that whatever conflicts with our ideas is absolutely impossible and involves a contradiction'. 19

In the Meditations and elsewhere, Descartes got from conceivability to possibility through trust in God, rather than through the conceptualist analysis of modality. Because he chose to keep his voluntarism out of sight in those works, he could not expect to be understood or believed if he relied upon it in any of his arguments; so he had to settle for getting 'possible' from 'conceivable' theologically. This is compatible with his having, ultimately, an analytic basis for the move. Theology provides one reason why the move is safe; philosophical analysis provides another. Descartes' analysis of modality also leads to voluntarism. Given that all modal truths are at bottom truths about what we can conceive, and given that God made us how we are (this being a truism for Descartes), it follows that God gives modal truths their status as truths. He made it necessarily true that 2 + 2 = 4 by making us unable to conceive otherwise. Put the conceptualist analysis of modality together with the truism that God made our nature, and out rolls the central tenet of voluntarism in the form of a theological doctrine. The analysis provides a philosophical grounding for the theology, and is not a distraction from it or a rival or alternative to it. V. THE BOOTSTRAPS PROBLEM The doctrine that God created all the modal truths sits at an awkward angle to Descartes' thesis that God's existence is absolutely necessary. It seems to imply that a peculiar bootstrapping procedure has gone on: God ____________________ 18 Sixth Meditation ( A T VII78: C S M II54). 19 Letter to Gibieuf ( A T III476: C S M K202). -168selected the principles of necessity, which rewarded him by guaranteeing his existence! This has led some to suppose that Descartes' doctrine about necessary truths was not meant to apply to propositions about God's own nature and existence. 20 The texts provide little basis for this exclusion, but Curley argues that Descartes might have had a philosophical reason for it, so that it is not a purely ad hoc gerrymander. 21 Anyway, the exclusion is better than what seemed to be the only alternative--namely, voluntarism creates the bootstraps problem, and Descartes alone fails to see this. Now, however, we see a third option: there is no bootstraps problem. Necessarily God exists; that is, when we are thinking clearly, it is inconceivable to us that God should not exist. Because God

gave us our limits, it is he who has made it inconceivable to us that he should not exist, but there is no paradox or circle or other problem in that. When voluntarism is based on the conceptualist analysis of modality, the bootstraps problem vanishes. You may want to object: 'You have extricated Descartes from the tangle of bootstraps by weakening his theology from the momentous and cosmic proposition (a) that it is absolutely necessary that God exists, to the humdrum and local truth (b) that we cannot conceive that God should not exist.' That objection implies that (b) is weaker than (a), which amounts to rejecting the conceptualist analysis of modality. Even if you are right to reject it, however, my hypothesis that Descartes accepted it is all I need to abolish the bootstraps problem. If Descartes believed the conceptualist analysis of modality, and asserted voluntarism on that basis, he was not threatened with any paradox in relating the latter to the thesis that necessarily God exists. Q.E.D. A slightly different objection: 'It is natural--indeed, it is almost inevitable--that one should think of God's existing necessarily as explaining why God exists. Your Descartes, however, cannot see it in that way. Nobody could think that God's existence is explained by there being certain limits to what we can conceive--let alone by his having set them! That shows how radically your Descartes has departed from some natural ways of thinking, and so it is some evidence that you have misunderstood the real Descartes.' For this objection to have much force, there must be evidence that Descartes was one of those philosophers who have thought that 'Why is it the case that P?' might be answered by 'Necessarily P'. There are such philosophers, and they may feel their position to be natural and even inevitable; but Descartes was not of their number, at least with ____________________ 20 Curley, "'Descartes on the Creation of the Eternal Truths'", 592-7; Gueroult, Descartes selon l'ordre des raisons, 26-9. 21 Such textual basis as there is for it is presented by Wilson, Descartes, 123f. -169respect to the existence of God. His a priori argument for God's existence occurs not in the metaphysical context of 'Why does God exist?', but in the epistemological context of 'How can we be absolutely sure that God exists?' When led by his critics to address the metaphysical question, Descartes never answered it in terms of necessity. To Caterus he wrote that God 'does not need' an efficient cause, because he 'possesses such great and inexhaustible power that he never required the assistance of anything else in order to exist'. 22 In a 'geometrical' arrangement of his views, one of the axioms says this: 'The immensity of God's nature is the cause or reason why he needs no cause in order to exist.' And in reply to Arnauld he wrote that 'the reason why God does not need any efficient cause in order to exist depends on . . . the very immensity of God'. In these passages that face the question 'Why does God exist?', the idea of his existing necessarily is vividly absent. VI. NECESSITY AND TRUTH

A third objection, which seems like a strengthened version of the second, goes like this: 'The position you credit Descartes with taking does not let us infer that God exists from its being necessary that he does. More generally, it does not license the inference from NecessarilyP to P. Descartes would surely have been aware of that and would have found it intolerable.' Not so fast! The subjectivist analysis of modality with which I have credited Descartes does secure this result. If necessarily P, then no human being can believe that not-P. Someone in a muddle might have an affirmative attitude to a sentence which means that not-P, but such a person cannot be thinking clearly, cannot really have his mind around the thought that not-P and be assenting to it. While we are thinking clearly, then, the move from NecessarilyP to P is settled, an immovably built-in part of our scheme of things. Renewed objection: 'Stop pretending not to understand! You point out that on that theory of modality nobody can disbelieve a necessary proposition, because nobody can even understand the state of affairs in which it is false. But you were supposed to address the question about how P's necessity relates not to our acceptance of P but rather to P's truth.' This objector relies on a distinction which Descartes often implicitly rejected. Running through his thought was an intensely subjectivist strand, in which ____________________ 22 Replies to First Objections ( A T VII108 f.: C S M II78). The next two quotations are from, respectively, Replies to Second Objections ( A T VII165: C S M II116); Replies to Fourth Objections ( A T VII231 f.: C S M II162). -170issues about what is really the case are displaced by, or even equated with, issues about what to believe or about what can be believed. When writing in this vein, Descartes • launched the Meditations as a pursuit of beliefs that were 'stable and likely to last'; • highlighted propositions which 'we cannot ever think of without believing them to be true'; • attended to the situation of someone who 'is certain that he is not being deceived, and is compelled to give his assent to' a given proposition; • accorded a privilege to 'certain common notions' of whose truth our mind, 'for as long as it attends to them, is completely convinced'; • used the phrase 'may appear false to God or an angel' as a preferred alternative to '[may be], absolutely speaking, false'; • brushed aside a suggestion about falsehood with the words 'What do we care about this [alleged] absolute falsity, since we neither believe in it nor have even the smallest suspicion of it?'; • said of a similar threat that it was 'no objection' because 'the evident clarity of our perceptions does not allow us to listen to anyone who makes up this kind of story'. In short, there was in Descartes a wide, deep, vivid streak of subjectivism or pragmatism about truth--a willingness to treat results about the settlement of belief as though they were results about how things stand in reality, or as though the former mattered and the

latter did not. 23 When in that frame of mind, he would equate our finding not-P inconceivable with our discovering that P, and would have the only kind of warrant for 'If necessarily P, then P' that interested him. The best and most disciplined parts of Descartes' struggle with scepticism in the Meditations belong to this subjectivist strand in his thought. But there was also an objectivist or realist strand, in which he tried to argue from indubitability to truth, with the latter understood in a realist way. The places where this happens, notably the 'proof' of the truth rule, are notoriously shaky and murky parts of the work; but they do exist; both strands are there. That is why I said in Section I that the subjectivist analysis of modality harmonizes as well as anything can with the rest of Descartes' work: it fits the subjectivist strand in the Meditations; nothing could fit both strands. ____________________ 23 For details and references, see my "'Truth and Stability in Descartes'" Treatment of Scepticism'. A similar view is expressed in Louis Loeb, "'The Priority of Reason in Descartes'", Philosophical Review, 99 ( 1990), 30-43, and was adumbrated by Jaegwon Kim, "'What is Naturalized Epistemology?'" ( 1988), repr. in his Supervenience and Mind ( Cambridge: Cambridge University Press, 1993), at 219. -171By keeping voluntarism out of that work, Descartes helped to hide from himself the split in his thought. Had he let it in, it would have virtually forced him to follow the tracks of the objector I have been answering in this section, and thus compelled him to become explicitly clear about how indubitability relates to truth. Perhaps Descartes was subliminally aware of this, that being why voluntarism does not show up in the Meditations or either of its cousins--the Discourse on the Method and the Principles of Philosophy. 24 VII. 'NOT IMPOSSIBLE' AND 'POSSIBLE' The conceptualist analysis of modality, conjoined with the truism that we are as God made us, yields Descartes' theological voluntarism, and satisfies the religious impulse which underlies the latter. The religious idea was that God in his greatness never confronts any rule or standard or principle that holds good independently of him, and this had to be maintained in face of the threat that even God must conform to the laws of logic. The analysis removes that threat. According to the analysis, its being impossible that P is P's relating thus and so to human minds; so God in contemplating how to act does not encounter any limiting truths of the form 'It is impossible to bring it about that Q'. The analysis gives Descartes everything he wants in this part of his theology. If voluntarism were powered by the idea that God can do everything, it would be committed to giving examples of extraordinary things God can do, saying, for instance, that God could make two plus two equal five. That would imply that two plus two could equal five, which would generalize to the conclusion that there are no impossible

propositions. That would be a disaster from any point of view, and the conceptualist analysis of modality condemns it. The modal status of the proposition that two plus two equals five is secured by how this relates to our conceptual capacities, but the proposition that two plus two could equal five relates to those capacities in just the same way, and therefore has the same modal status. Fortunately, ____________________ 24 There is evidence that when Descartes wrote the Meditations he had voluntarism in mind, and tried not to contradict it. I do not see how else to explain this: '[T]here is no doubt that God is capable of creating everything that I am capable of perceiving in this manner; and I have never judged that something could not be made by him except on the grounds that there would be a contradiction in my perceiving it distinctly' (Sixth Meditation, A T VII71: C S M II50). A statement about what God can do would naturally be followed by one about what he cannot do. Descartes, however, says nothing about that, and writes only of what Descarteshas thought God cannot do. On the face of it, this is a strange performance. The only way I can make sense of it is to suppose that even here Descartes is conscious of his commitment to the view--which is part of his voluntarism--that there is nothing that God cannot do. -172Descartes' actual religious concern--as distinct from the omnipotence idea which so many people wrongly attribute to him--does not push him to this disastrous extreme. Where the thought that God can do anything leads on to 'for instance he can make the sum of two and two equal five', the thought that nothing independent of him limits him leads only to 'for instance he does not confront the impossibility that two plus two should equal five'. How do these differ? I have credited the analysis with leading to the likes of this ('NI' for 'not impossible'): (NI) It is not the case that God absolutely could not have made two plus two equal five, and congratulated it on not leading to the likes of this ('P' for 'possible'): (P) God could have made two plus two equal five. How can this be? From something's not being impossible, does it not follow trivially that it is possible? How could any coherent theory entail NI without also entailing P? Hostile answer: 'It couldn't! The inference of P from NI is trivially valid, and thus unblockable; a theory that purports to entail NI but not P must be incoherent. Voluntarism is in as much trouble on your reading as on any other.' Before looking into this, let us notice what a wide swath it cuts. The problem about NI and P arises for every theory of the form:

Because Basis obtains, the modal concepts apply as they do, where Basis is a contingent proposition. It seems all right to infer from such a theory that if Basis had not been the case, the modal concepts would not have applied as they do, so that, for instance, (NI) it would not have been impossible that two plus two equals five. Then how can we block the inference to (P), 'If Basis had not been the case, it would have been possible that two plus two equals five' and thence to 'It is possible that two plus two equals five'? If we cannot, then no theory of the type in question can be right; every such theory implies that so-called necessary truths are contingent, thereby demolishing modality rather than analysing it. This proves a priori that the nature and applicability of our modal concepts cannot be grounded in contingent facts. The last sentence of that paragraph casts doubt on the rest of it. In philosophy such large results do not often come so easily! Is it likely that an argument as short as that suffices to destroy, for example, all of Wittgenstein's work on modality? -173The doubt is justified, because the line of thought is faulty, as I now explain. The inference which it calls trivially valid is indeed so when the modal concepts are being taken for granted as part of the working apparatus. In one special kind of context, however, we should not take them for granted, should not use them to conduct inferences that would ordinarily be valid. I refer to contexts where we are considering what has to be the case for the modal concepts to be usable. A statement of the form 'If Basis were not the case, (NI) it would not be impossible that Q' understates something whose full strength is 'If Basis were not the case, our modal concepts would be inapplicable'; and this obviously does not entail that if Basis were not the case (P) it would be possible that Q. The inference from NI to P fails just precisely when we are exploring a theory such as Descartes' conceptualist analysis of modality. Analogously, a theory might say what must be the case if any spatial concepts are to be applicable, thus entailing things of the form if Basis did not obtain, nothing would have the same size as anything else, but obviously it would not follow that If Basis did not obtain, everything would differ in size from everything else. Residual discontent: 'Something funny is going on here. You are trying to talk about what would have obtained in the absence of (Basis) our de facto basis for applying modal concepts. When opponents claim to infer an absurdity from this, you reply that they are employing a conceptual apparatus which would not be usable if Basis did not obtain. Then you ought to say that nothing is then available, so that nothing can be said about what would obtain if Basis did not. So your Descartes, having spoken of our limits, ought to be silent about God's situation, not even allowing himself to say that God, lacking our limits, does not confront independently grounded impossibilities.' This, though plausible, is wrong. If we hold that our concepts apply as they do because of certain contingent facts about ourselves, we can safely infer from this that in the absence of those facts the

concepts would not apply. We risk incoherence only if we say too much about what positively would be the case if those facts did not obtain. Consider the related idea that we understand some questions the answers to which we could not understand. Apply this first to creatures other than ourselves. We might discover language-using animals, simpler than us, who have arithmetic which they use well, counting and computing up to quite high numbers. They have never yet run out of numbers--the resources seem limitless--but they are anxious about this. 'Has the number pool really no bottom?', they want to know; and it troubles them that they cannot find an answer. Knowing all about their brains and our own, we -174know that they lack the neural structures needed for the thought of a recursion, an endlessly repeatable operation such as going from a number to its successor. (They are capable of recursions, but not of the thought of them). So we can see that they have a clear question the answer to which lies beyond their intellectual limits; we can tell them this, and they might understand what we say, and even believe it. That, I submit, is clearly something we might discover. So we should acknowledge that some of our questions, too, may have answers which we are profoundly, biologically incapable of grasping. We have no trouble understanding this, even though it is the possibility of our being unable to understand something; and we can even say what it is that we would not understand--namely, the true answers to some of our questions. I offer this as refuting the general thesis, which I guessed might be brought against the conceptualist analysis of modality, that we cannot say anything about our own limits without transcending them. VIII. DESCARTES' HANDLING OF THESE DIFFICULTIES In the passages quoted above, Descartes did not put a foot wrong. Applying his conceptualist analysis of modality to his theological problem, he said only that God does not confront any absolute impossibilities. A passage quoted in Section II handles the conceptualist analysis of modality with striking accuracy. Here is a shortened version of it: I do not think we should ever say of anything that it cannot be brought about by God. I would not dare to say that God cannot make it be the case that one and two are not three. I merely say that he has given me such a mind that I cannot conceive an aggregate of one and two which is not three, and that this involves a contradiction in my conception. Descartes does not say here that one plus two could fail to equal three; he merely refuses to assert that God could not make this be the case. He does assert that he cannot conceive of one plus two not making three, and that by his standards it is contradictory. This is exactly right when a religious concern with God's greatness is being helped by a conceptualist analysis of modality. The threatened modal limit on God is removed; but it

continues to limit our powers of conception. Descartes calls it contradictory, but reminds us that this merely relates it to us and implies nothing about how it relates to God. Equally flawless is this, written to Mersenne: 'In general we can assert that God can do everything that is within our grasp, but not that he cannot do what is beyond our grasp. It would be rash to think that our -175imagination reaches as far as his power.' 25 This makes the general claim that God can do things that are unintelligible to us; but it does not say that God can do everything, nor does it fall into the trap of saying, for some value of Q, that God could make Q obtain although we cannot conceive its doing so. On rare occasions, however, Descartes did go wrong in that manner, as when he wrote to Mesland: 'The power of God cannot have any limits. . . . [This] shows us that God cannot have been determined to make it true that contradictories cannot be true together, and therefore he could have done the opposite.' 26 The last part of this is wrong if Descartes' basis was the conceptualist analysis of modality. So this fragment of this letter counts somewhat against my interpretation, and for a reading which makes voluntarism entail that everything is possible. Much more of the evidence counts in my favour, however; and we shall see shortly that another fragment of this same letter also supports a third reading of voluntarism! The letter is a hot potato. Replying to the Sixth Objections, Descartes said something which illustrates how hard it can be to know for sure what is going on in his voluntarism. After a discourse that fits my interpretation well, Descartes continued: 'There is no point in asking by what means God could have brought it about from eternity that it was not true that twice four make eight, and so on; for I declare that this is unintelligible to us'. 27 This is worrying. Descartes had been denying, in effect, that God encountered the necessity that twice four should make eight; and his theory about the nature of our modal concepts implies that denial. But the displayed sentence seems to suggest that God could have made it false that twice four makes eight, which makes a specific claim about what would have obtained if our modal concepts had been different. My Descartes, I repeat, was not entitled to do that, for two reasons: (i) according to him, God did not come up against modality in any way at all--he did not encounter either the impossibility or the possibility that twice four should equal nine, and (ii) if we say 'God could have brought it about that twice four was not eight', we claim to make sense of something of which we cannot make sense, purporting to pull into our domain of understanding something that lies beyond its boundaries. With those worries in mind, now look back at the sentence. It does not ____________________ 25 First letter to Mersenne [15 April 1630] ( A T I146: C S M K23). 26 Letter to Mesland ( A T IV118: C S M K235; emphasis added). 27 Replies to Sixth Objections ( A T VII436: C S M II294). I replace C S M's 'there is no

need to ask' by 'there is no point in asking', and C S M's 'I admit' by 'I declare'. Each change is well supported by the Latin (for the former, consult opus est in Lewis and Short Latin Dictionary, not in the Oxford Latin Dictionary), and each fits its context better. -176actually say that God could have made it false that twice four makes eight. It brushes off the question 'How could he have done this?' as pointless, for a reason having to do with unintelligibility. Furthermore, what the sentence describes as unintelligible to us is not the means whereby God could have made twice four differ from eight. Where I have 'means', Descartes had the noun ratio and the 'this' which he declared to be unintelligible is referred to by the pronoun id; the noun is feminine, the pronoun neuter; the two cannot be linked. What he was calling unintelligible, therefore, must be twice four's being different from eight. So Descartes was dismissing the question 'How could God have done things differently?' on the grounds that we are not intellectually equipped to cope with any aspect of the 'differently'; and this fits my interpretation well. The passage continues with this: 'It would have been easy for God to ordain certain things such that we men cannot understand the possibility of their being otherwise than they are.' This innocently quantifies over the 'certain things', not purporting to state any of them. Still, perhaps not all is well. One naturally hears that last bit as meaning 'It would have been easy for God to ordain certain things--such as twice four's not being eight--Which men cannot understand . . .'; and that involves hearing the passage as implying something to which Descartes, on my account of his views, was not entitled. A little later in the same paragraph, he continued: 'Hence we should not suppose that eternal truths 'depend on the human intellect or on other existing things'; they depend on God alone, who, as the supreme legislator, has ordained them from eternity.' The affirmation after the semicolon fits my story, but the preceding denial does not, because my Descartesdid hold that eternal truths depend in a certain way on the human intellect. This, however, does not seriously embarrass my interpretation. Descartes was responding to his anonymous critics who, in the Sixth Objections, had picked on something in his Replies to the Fifth Objections, and had written: How can the truths of geometry or metaphysics . . . be immutable and eternal and yet not be independent of God? . . . How could he have made it untrue from eternity that twice four makes eight, or that a triangle has three angles? Either these truths depend solely on the intellect while it is thinking of them, or on existing things, or else they are independent, since it seems that God could not have brought it about that any of these essences or truths were not from all eternity. 28 The issue concerns the eternity of the truths in question. The critics are saying that one must choose: either (a) relinquish the idea that they are

____________________ 28 Sixth Objections ( A T VII417f.: C S M II281). In the final phrase, the Latin means simply 'were not', presumably in the sense of 'did not exist or obtain'. CSM's 'were not as they were' cannot be right. -177eternal, and say that they hold only for as long as someone thinks them, or that in some other way they are at the mercy of contingent things; or (b) stick to the view that they are eternal, in which case you must say that they are independent of God, because even he could not give something a certain status for all eternity. The options grouped together in (a) are taken to be outright denials of the eternity of the truths; this is rammed home by 'depend solely on the intellect while [dum] it is thinking of them', which CSM mistranslates as 'depend solely on the intellect that is thinking of them'. Descartes in his reply needed only to deny that eternal truths depend in that way on anyone's thinking of them--a way that would give them no more durable backer than the individual thinker (whom he assumed to be human, though his critics had not said so). That positioned him to choose the other option--the truths are eternal--and to assert bluntly that God could indeed make them so from all eternity. Nothing in that conflicts with the conceptualist analysis of modality. The statement 'Eternal truths [do not] depend on the human intellect' conflicts with the analysis, but what Descartes needed to say does not--nor, I suggest, does what he meant. (I shall return to the eternity of necessary truths in Section XI.) Still, there are lapses, of which I have cited one and a possible second. They are few and minor, however, and it is impressive how seldom Descartes implies that everything is possible, and how often his theological statements of voluntarism are faithful to the analysis of modality that I attribute to him. Curley, alone among previous writers on these matters, has noticed how seldom Descartes took his voluntarism to entail that God could have made twice four not equal eight. Perhaps Descartes meant to say ( Curley suggests) not that God could have made necessary truths false, but only that he could have made them contingent. 29 The only direct support for this occurs in the 'hot potato' letter to Mesland, from which I quoted early in this section. Descartes alludes there to 'things which God could have made possible, but which he has nevertheless wished to make impossible'. This text is Curley's prime exhibit, just as the passage about two 'sorts of possibility' in the Replies to Second Objections is mine. The position that Curley attributes to Descartes has a kind of coherence: there are, as Curley points out, workable systems of logic which allow for a proposition to be possibly possible but not actually so. Whether Descartes' overall philosophy would be much helped by his taking this option is not clear, however, and Curley himself seems unsure about it. ____________________

29

Curley, 'Descartes on the Creation of the Eternal Truths'. -178-

Comparing the doctrine that he attributes to Descartes with the thesis that all propositions are possible, Curley says that his interpretation avoids the horrendous systematic consequences' of the latter, 30 but a page earlier he says that his is 'not much better' than the other; and he concedes that even the position he attributes to Descartes must not be applied to propositions about God's own nature and existence. Those concessions make Curley's reading less inviting than mine, I suggest. Its textual basis is also thinner. In the paragraph where Descartes wrote about 'things which God could have made possible', he also wrotealas!--that God 'could have done the opposite', thereby implying that the truth-value, and not merely the modal status of the seemingly necessary proposition, could have been different. Anyway, voluntarism on Curley's account of it does not minister to Descartes' theological concerns. According to Curley's Descartes, God could make it be only contingently false that two plus two equals five, but he could not make it true; for if he could make it true, then it is possible, and by parity of reasoning, so is every proposition, which is the 'horrendous' result that Curley rightly wants to avoid. This is puzzling. Making it possible for two plus two to equal five looks hard, but making a possible proposition true seems easy for a god. Curley's Descartes holds that God can perform the hard task but not the easy one. This sounds like untenable theology, whether one thinks of God as omnipotent (as Curley's Descartes thinks of him) or merely as unlimited by anything independent of him (as mine does). IX. ESSENCES In one place Descartes speaks of God as decreeing that 'the essences of things, and the mathematical truths which we can know concerning them' shall be eternal and immutable. 31 I shall come to eternity soon, but first let us think about essences. Descartes ordinarily takes essences to be the ground or source of necessary truths, and one wonders how that relates to voluntarism. At first glance it seems to fit in smoothly: God creates necessary truths by creating the essences from which they flow. Loeb understands the situation in that way: he writes of Descartes' 'peculiar doctrine that the eternal truths are dependent on God-willed essences', implying that voluntarism speaks of the creation of essences from which necessary truths somehow arise. 32 If that were Descartes' view of it, though, he would ____________________ 30 Ibid. 593. 31 Replies to Fifth Objections ( A T VII380: C S M II261). 32 Loeb, From Descartes to Hume, 68. -179-

be faced with a nasty question: for God to make it false that twice four is eight, would he have to give a different essence to one of those numbers? If Descartes answered yes, conceding that God could not make twice four not be eight while leaving those essences intact, he would gravely offend against the spirit of voluntarism. If, on the other hand, he answered no, allowing that God could falsify an arithmetical truth either by altering an essence or in some other way, that would cast an eerie light on the claim that God does in fact settle modal truths by creating essences. What reason could there be for this? What point in believing it? Indeed, what could it mean? The former option is intolerable, the latter unattractive; so Descartes needs to stop the question from arising. He could do that by maintaining that 'essence' does no work of its own in the context of voluntarism, so that 'changing the essences' and 'changing the necessary truths' would merely be two names for a single act. The doctrine that essences are a ground for or source of necessary truths, he should say, though all right in its way and at its level, cannot be sustained when we are on the metaphysical ground floor where voluntarism comes into play. That best option is the one that Descartes actually took. The term 'essence' seldom occurs in the voluntarism texts. When it does, Descartes is responding to a critic who has used it, and his reply does not put the term to work in a doctrinal way. For example, he never says in these contexts that eternal truths depend on, or are grounded in, essences. The only voluntarism text where essences are mentioned more than passingly is in a letter to Mersenne, where Descartes equates God's role as 'the author of the essence of created things' with his 'establishing the eternal truths', on the grounds that 'this essence is nothing other than the eternal truths'. 33 X. INNATENESS In the first letter to Mersenne, Descartes said that mathematical truths 'have been laid down by God' and a little further on, that they 'are all inborn in our minds'. Jolley infers from this that ' Descartes saw a connection between innate ideas and the creation of the eternal truths'. 34 Did he? ____________________ 33 Third Letter to Mersenne [ 27 May 1630] ( A T I151 f.: C S M K25). 34 Jolley, Light of the Soul, 51. This is Descartes "first known reference to innate ideas', Jolley writes in an uncharacteristic lapse. The letter does not mention innate ideas; what it calls innate are laws and truths, and Descartes spoke of those as innate several years earlier. I should add that Jolley's reading of voluntarism, though also influenced by Margaret Wilson, differs from mine in a couple of respects. -180In the letter a brief sketch of voluntarism is followed by the statement that necessary truths are knowable by us because God has imprinted them on us; but it looks to me as though Descartes is here moving on from one topic to the next, from voluntarism to

modal epistemology. I cannot see these two themes in the letter as organically connected; nor can I see why Jolley thinks they are. Loeb flatly denies that the second topic--the epistemic use of innateness--has any place anywhere in Descartes' thought. He writes: ' Descartes would have had no use for innate ideas were it not for his [voluntarism].' 35 The negative part of this ignores the many texts in which Descartes connects innateness with knowability in quite Leibnizian ways, with voluntarism being nowhere in sight. (Some were written before Descartes first announced voluntarism, and presumably before he had thought of it.) I am mainly concerned, however, with the positive claim that voluntarism could motivate a doctrine of innate ideas or knowledge. The claim is right, of course. According to voluntarism as I understand it, for P to be necessary is for P to relate thus and so to our intellectual limits; and those limits are innate, in that God gave them to us as part of our human birthright. This humdrum tie between voluntarism and innateness, though solid enough, does not help us with any of the passages where Descartes discusses innateness for its own sake. While insisting that all these passages should be seen in the light of voluntarism, Loeb gives no reason why any of them should be. XI. THE ETERNITY OF NECESSARY TRUTHS Necessary truths, according to our ordinary notion of them, are eternally necessary; so Descartes had reason to want his analysis to imply that they are so. Indeed, he sometimes used 'eternal truths' as a mere label for them; he never called them 'necessary'. He did once refer to 'the mathematical truths that you [ Mersenne] call eternal'; but if he meant to suggest that perhaps they are not eternal, it is the only time that he did so. Their eternity creates a problem for Descartes on any of the more usual readings of his voluntarism. 36 It may also seem to threaten my reading: if ____________________ 35 Loeb, From Descartes to Hume, 68. From this Loeb infers that Descartes' innatism is differently motivated from Leibniz's--which is one of his reasons for denying that the so-called rationalists form a bloc. 36 For some discussion of the difficulties. see Curley 'Descartes on the Creation of the Eternal Truths', 576 ff. -181the necessity that twice two should make four is ultimately a contingent fact about our capacities, then might we not someday find that this arithmetical truth has turned false on us? We wake up one morning to find that we have two noses, rectangular fingers, a hatred for sugar and . . . minds such that twice two no longer makes four.

This version of the problem can be solved, however, though I doubt if Descartes saw how. The prospect with which we are threatened is no threat at all, because facing it means entertaining the thought of four's not being twice two, and we cannot do that. This belongs to the general point discussed in Section VII: our conceptual limits could have been different, but we cannot have a specific thought about any such difference. That applies to the past as well as to the future. Descartes' analysis does not entitle us to say: 'Although it is necessary (for us) now that twice two is four, there may have been--or there may someday be--a time when it was not or will not be.' Anything we say or think now about the modal status that a proposition had or does or will have, or would have if . . . , must be determined by our actual present intellectual limits. So Descartes' analysis secures the eternity of necessary truths. Objection: 'No! On your account, the analysis implies that we must think of necessary truths as eternally true, but not that they really are so.' This, like an earlier objection, rejects the conceptualist analysis of modality: it purports to make sense of something which the analysis declares to be senseless. We are not discussing the truth of the analysis, but rather its powers. We were asking whether it can secure the eternity of necessary truths, and the right answer is: yes, it can do so, according to its understanding of what it is for a proposition to be eternally true. That, however, does not put the eternity matter to rest. Granted that the analysis will not let us think of a change in the modal value of any particular proposition, it does allow us the general thought that some day the distribution of modalities across propositions may change. If it did not allow that thought, it would imply that the analysis itself is unthinkable. Now, even that general thought seems to threaten the eternity of necessary truths, and may well be found troubling--especially by one who values tranquility as much as Descartes did. This worry might be allayed in a Kantian manner, by arguing that mental identity depends on how our thoughts hook into one another along the time line, so that no thinker could survive a conceptual change of the threatened kind. However, one would not expect that from Descartes, who helped himself to the notion of continuant thinkers without considering what it takes for such a thing to last through time. Lacking Kantian solace, -182he dealt with the eternity problem differently: 'I do not think that mathematical truths . . . are independent of God,' he wrote. 'Nevertheless I do think that they are immutable and eternal, since the will and decree of God willed and decreed that they should be so.' 37 This gives to the eternity of necessary truths a theological basis--an imperfect treatment of the topic, but not bad in a way that makes it count against my reading of voluntarism as based on a conceptualist analysis of modality. In the remark last quoted, by the way, 'immutable' does not mean 'unchangeable'. but only 'unchanging'. This is standard for Descartes, who elsewhere equates 'God is immutable'

with 'God acts always in the same way', and who conducts this inference: 'Since they are always the same, it is right to call them immutable and eternal.' 38 XII. ETERNITY AND THE MEDITATIONS Some philosophers have seen the eternity or at least the durability of necessary truths as central to a famous problem that arises out of Descartes' treatment of scepticism in the Meditations. He propounds a 'truth rule' according to which anything that I perceive distinctly (while finding it indubitable) is true; then later he appeals to God's benevolence as a guarantor against some disaster in the life of the mind. Here is the problem: if we securely have the truth rule, what further intellectual benefit can we get from the assurance of God's wishing us well? Two wrong answers to this have had some currency. One says that God's benevolence secures us against failures of memory, and thus against wrong beliefs of the form 'I did distinctly perceive that P'. Nobody now understands Descartes in this way; nothing in the texts gives it any support. The other wrong answer connects with voluntarism, and I shall discuss it. It has turned up at intervals down the years, never gaining ascendancy, but never quite dying out either. I shall call it 'the Bréhier view'. 39 According to it, the truth rule serves at a given time to establish the truth ____________________ 37 Replies to Fifth Objections ( A T VII380: C S M II261). 38 The World, 7 ( A T XI43: C S M196); Replies to Fifth Objections ( A T VII381: C S M II262). 39 It is propounded in Émile Bréhier, 'The Creation of the Eternal Truths in Descartes's System' ( 1937), repr. in Doney (ed.), Descartes: A Collection of Critical Essays, 192208; Norman Kemp Smith, New Studies in the Philosophy Qf Descartes ( London: Macmillan, 1952), 188, 273 ff.; and John Etchemendy, 'The Cartesian Circle: Circulus ex tempore'. It is also hinted at in Wilson, Descartes, 128. -183of P at that time, and God's existence and nature assure us that P will continue to be true thereafter. 'The only problem is to know how I can foresee with certainty that . . . what is intuited will continue to be true.' 40 Most of the propositions that we are initially assured of by the truth rule are necessary truths; so the Bréhier view of God's role requires that even necessary truths be prima-facie unstable, variable, shifty; and voluntarism seems to satisfy that requirement. In a nutshell: God chooses to make twice two equal four, and his goodness towards us assures us that once he has done that, he will never undo it. That was indeed Descartes' view of the matter; but does it inform the Meditations? Nothing in that work points unequivocally to the Bréhier view. When Descartes there makes use of 'was'/'is' and 'then'/'later', he is contrasting an inability at one time to doubt P with a later ability to doubt it, not contrasting P's truth at one time with its falsehood

later. Consider also the terms in which Descartes announces that the truth rule needs help from theology: My nature is such that so long as I perceive something very clearly and distinctly I cannot but believe it to be true. But my nature is also such that I cannot fix my mental vision continually on the same thing, so as to keep perceiving it clearly; and often the memory of a previously made judgement may come back, when I am no longer attending to the arguments that led me to make it. And so other arguments can now occur to me which might easily undermine my opinion, if I did not possess knowledge of God. 41 This counts decisively against the Bréhier view. If Descartes' concern were that a once true proposition might have become false, his repeated emphasis on 'my nature' would be pointless, as would his highlighted contrast between two different intellectual states that I may be in. He goes on to introduce the thought that 'there have been frequent cases where I have regarded things as true and certain, but have later been led by other arguments to judge them to be false'. This creates the worry which a belief in a veracious God will supposedly vanquish; and it is, clearly and explicitly, the thought that on the past occasion I was wrong: that is, that the proposition in question was false. The Bréhier view is a non-starter. Voluntarism casts no useful light on those aspects of the Meditations that have received most attention: the truth rule, divine veracity, the relation between those, the Cartesian Circle. The real connection between the two is the one indicated in Section VI above: voluntarism and the treatment of ____________________ 40 Bréhier, 'Creation of the Eternal Truths in Descartes's System', 201. 41 Fifth Meditation ( A T VII69: C S M II48); see also Principles, part i, art. 13 ( A T VIIIA 9f.: C S M I197). -184scepticism in the Meditations are both shaped by a subjectivist tendency in Descartes' thinking--a willingness to understand concepts such as those of truth and necessity in terms of what they mean to us, how they figure in our lives. 42 ____________________ 42 Comments by Margaret Wilson and by two referees for the Philosophical Review have enabled me to improve this paper considerably. I am grateful. -185VII THE EPISTEMOLOGICAL ARGUMENT FOR MIND-BODY DISTINCTNESS

MARGARET D. WILSON Descartes' mind-body dualism is the aspect of his philosophy that has been most often cited and discussed in recent philosophical writing. Yet there has been, it seems to me, surprisingly little serious effort to gain an accurate understanding of his position. In another paper 1a I have tried to show that Cartesian dualism, as Descartes himself understood it, differs in both content and motivation from the view sometimes called 'Cartesian dualism' in recent discussions of the mind-body problem. The differences derive, especially, from Descartes' conception of the possibilities and limitations of mechanistic physical explanation, and his peculiar contention that 'the brain can be of no use to pure understanding'--as opposed to the faculties of imagination and sense, which are more dependent on body ( A T VII358 [ C S M II248]). It is true, on the other hand, that in trying to establish his dualism, Descartes himself places greatest stress on an argument for the immateriality of mind or self which does not seem to depend on these differences. This argument, which I call the 'epistemological argument' for the distinctness of mind from body, is presented in the Sixth Meditation--and, in somewhat different versions, in various other works. Descartes' epistemological argument constitutes the principal bridge between historical Cartesianism and contemporary (i.e. twentieth-century) discussions of the mind-body relation. I believe that even this argument, taken by itself, has not been correctly represented or criticized in the recent literature. The epistemological argument of the Sixth Meditation has its roots in arguments developed in the Second Meditation concerning knowledge of the self as a thinking thing, and knowledge of body as something ____________________ Margaret Wilson, "'The Epistemological Argument for Mind-Body Distinctness'" in Nous, Volume X ( 1976), 3 - 15, by permission of Blackwell Publishers. 1a M. D. Wilson, "'Cartesian Dualism'", in Hooker (ed.), Descartes: Critical and Interpretive Essays. -186'extended, flexible, movable'. Having used the demon hypothesis in the First Meditation to bring into 'doubt' the existence of body, Descartes argues in the Second Meditation that this 'doubt' does not extend to his own existence ('if he deceives me I exist'). He then considers what attributes can be ascribed with certainty to himself at this stage of his reasoning. He concludes that even certain properties traditionally associated with the soul or vital principle--for example, nutrition--must be presently excluded as part of the doubt of body. There is only one, he finds, that is not called into question on this basis: To think? Here I find it: thought [it] is; this alone cannot be separated from me. . . . I do not now admit anything except what necessarily is true. I am therefore strictly only a thinking thing, that is mind, or soul, or understanding or reason. . . . I am however a true and truly existing thing;

but what sort of thing? I have answered, a thing which thinks. ( A T VII27 [ C S M II18]) This passage may seem to imply that Descartes thinks he has already established the conclusion of the epistemological argument: that he is nothing essentially but a thinking thing, and as such is distinct from anything physical: 'thought alone cannot be separated from me'. However, he is careful to cancel any such implication in the immediate sequel of the passage: But possibly it happens that these very [corporeal] things [such as the human body] which I supposed were nothing because they are unknown to me, are in the real state of things [in rei veritate] not different from this me which I know. I do not know, I do not dispute about this matter now, I can only give judgement on things that are known to me. ( Ibid. ) Descartes does not, then, wish to claim on the basis of the Second Meditation reasoning alone that he knows that only thought and nothing corporeal pertains to his nature. On the other hand, he is not at this point of the argument restricting himself to an epistemically provisional conclusion like 'as far as I now know I am a thinking thing and only a thinking thing'. For instance, he is implicitly claiming to know, not merely that he thinks, but that thought pertains to his nature or essence: it 'cannot be separated from me'. 1 Also, he explicitly maintains that reasoning concerning the indubitability of his own existence (the 'cogito reasoning') has brought him to the conclusion that he is a true and truly existing thing (res vera et vere existens). The importance of this statement should become clear later. The Second Meditation contains at least one other assertion that is important to the epistemological argument: that Descarteshas a clear and ____________________ 1 Descartes does not seem to offer justification for the transition from 'I think' to 'Thought belongs to my nature or essence'. -187distinct idea of himself as a thinking thing (apart from any concept of the corporeal). He begins to hint at this point immediately after the statements already cited. And at the end of the Second Meditation, after arguing that his best knowledge of a typical physical object--a piece of wax--is derived from reason rather than sense, he concludes: What, however, shall I say of this same mind, or of myself? For so far I do not admit that there is in me anything except mind. What, I ask, [of] I who seem to perceive this wax so distinctly? Do I not then know myself not only such more truly, much more certainly, but even much more distinctly and evidently? ( A T VII33 [ C S M II22])

Additionally (as this passage also suggests), Descartes claims in the Second Meditation that he has a distinct conception of body as an extended thing--which conception is separate from that of thought. These claims about distinct perception are important because of Descartes' very consciously held position that only clear and distinct perceptions or conceptions will suffice as the basis for positive affirmations about the nature of a thing (see especially Notes on a Programme, A T VIIIB 351-2 [ C S M I299], and A T III215 [ C S M K155]). I won't try to elucidate in any detail the distinction between clear and distinct conception and 'mere' conception. The distinction can, however, be partly brought out by the example of a geometrical proof. Call the conclusion of a given proof T. After one has examined (or constructed) the proof, one distinctly conceives or perceives that T. Before one has examined (or constructed) the proof, one will, very likely, have been able to conceive that T: that is, one will have been in the state of thinking that it might be the case that T. Being able (merely) to conceive that T does not in any way preclude also being able (merely) to conceive that not-T. Clearly and distinctly conceiving that T, on the other hand, does preclude being able clearly and distinctly to conceive that not-T. Between the Second and Sixth Meditations Descartes 'validates' his distinct perceptions by setting forth 'proofs' of the existence of an omnipotent and benevolent creator who would not permit him to be deceived in what is most evident. Descartes is, then, so far from concluding rashly from what he can conceive to what is the case, that he even finds it necessary to present God as a bridge from what he can distinctly conceive to what is the case. I We may now turn to the epistemological argument itself. The Sixth Meditation begins with the observation that God is capable of bringing about or -188distinct idea of himself as a thinking thing (apart from any concept of the corporeal). He begins to hint at this point immediately after the statements already cited. And at the end of the Second Meditation, after arguing that his best knowledge of a typical physical object--a piece of wax--is derived from reason rather than sense, he concludes: What, however, shall I say of this same mind, or of myself? For so far I do not admit that there is in me anything except mind. What, I ask, [of] I who seem to perceive this wax so distinctly? Do I not then know myself not only such more truly, much more certainly, but even much more distinctly and evidently? ( A T VII33 [ C S M II22]) Additionally (as this passage also suggests), Descartes claims in the Second Meditation that he has a distinct conception of body as an extended thing--which conception is separate from that of thought. These claims about distinct perception are important

because of Descartes' very consciously held position that only clear and distinct perceptions or conceptions will suffice as the basis for positive affirmations about the nature of a thing (see especially Notes on a Programme, A T VIIIB 351-2 [ C S M I299], and A T III215 [ C S M K155]). I won't try to elucidate in any detail the distinction between clear and distinct conception and 'mere' conception. The distinction can, however, be partly brought out by the example of a geometrical proof. Call the conclusion of a given proof T. After one has examined (or constructed) the proof, one distinctly conceives or perceives that T. Before one has examined (or constructed) the proof, one will, very likely, have been able to conceive that T: that is, one will have been in the state of thinking that it might be the case that T. Being able (merely) to conceive that T does not in any way preclude also being able (merely) to conceive that not-T. Clearly and distinctly conceiving that T, on the other hand, does preclude being able clearly and distinctly to conceive that not-T. Between the Second and Sixth Meditations Descartes 'validates' his distinct perceptions by setting forth 'proofs' of the existence of an omnipotent and benevolent creator who would not permit him to be deceived in what is most evident. Descartes is, then, so far from concluding rashly from what he can conceive to what is the case, that he even finds it necessary to present God as a bridge from what he can distinctly conceive to what is the case. I We may now turn to the epistemological argument itself. The Sixth Meditation begins with the observation that God is capable of bringing about or -188making the case whatever I am capable of clearly and distinctly perceiving: 'And I never judged that anything could not be brought about by him, except for the reason that it was impossible for me to perceive it distinctly' ( A T VII71 [ C S M II50]). The first application of this principle is to establish the possible existence of 'physical things conceived as the object of pure mathematics'--since previous Meditations have held these to be distinctly conceivable. The second application is in the epistemological argument. Because I know that all that I clearly and distinctly understand can be brought about by God as I understand it, it is enough that I can clearly and distinctly understand one thing apart from another [unam rem absque altera], for me to be certain that one is different from another, because they can be placed apart [seorsim poni] at least by God; and it doesn't matter by which power this is done, in order for us to judge them to be different; and thus, from this very fact, that I know I exist, and that meanwhile I notice nothing else to pertain to my nature or essence, except this alone that I am a thinking thing, I rightly conclude that my essence consists in this one [thing] that I am a thinking thing. And although

probably (or rather, as I will afterward say, certainly) I have a body, which is very closely conjoined to me, because nevertheless on the one hand I have a clear and distinct idea of myself, in so far as I am only a thinking thing, not extended, and on the other hand I have a distinct idea of body, in so far as it is only an extended thing, not thinking, it is certain that I am really distinct from my body and can exist apart from it. ( A T VII 78 [ C S M II54]) In lieu of detailed analysis of this passage--for which there is no space here--I will merely propose a provisional reading of Descartes' argument which seems to me natural: 1. If A can exist apart from B, and vice versa, A is really distinct from B, and B from A. 2. Whatever I can clearly and distinctly understand can be brought about by God (as I understand it). 3. If I can clearly and distinctly understand A apart from B, and B apart from A, then God can bring it about that A and B are apart (separate). If God can bring it about that A and B are apart, then A and B can exist apart (and hence, by (1), are distinct). 4. I am able to clearly and distinctly understand A apart from B, and B apart from A, if there are attributes Φ and Ψ, such that I clearly and distinctly understand that Φ belongs to the nature of A, and Ψ belongs to the nature of B, and I have a clear and distinct conception of A which doesn't include Ψ, and a clear and distinct conception of B which doesn't include Φ. 5. Where A is myself, and B is body, thought and extension satisfy the above conditions on Φ and Ψ, respectively. -189Hence, by (5), (6), (3), and (4), I am really distinct from body (and can exist apart from it). What, if anything, is wrong with this argument? Let me first mention some commonly heard objections to Descartes' position on the distinctness of mind that are not in fact effective against it. 2 Sometimes Descartes' mind-body dualism is taken to rest on (or partly on) the so-called argument from doubt--which is universally recognized to be fallacious. The argument from doubt is supposed to go something like this: My mind (or self) is distinct from all body. For something true of all body (that I can doubt it exists) is not true of myself (mind). But A and. B are the same only if everything true of the one is true of the other. We need not dwell here on the problems with this argument, for (I trust) it is perfectly obvious that the argument we have quoted from the Sixth Meditation is not a version of it. Whatever may be the connection in Descartes' mind between his inability to doubt his own existence while doubting the existence of body, it is not successfully captured by this unsound reasoning.

According to another objection, Descartes' argument can show at best that mind and body are possibly or potentially distinct (would be distinct if God should choose to separate them)--not that they are distinct. This objection fundamentally misses Descartes' point. Descartes holds that 'two' things are really distinct if it is possible for them to exist in separation. On this view, actual distinctness does not entail actual separateness. A third common criticism of Descartes' treatment of the distinctness of mind derives from the claim that, under sufficient conditions of ignorance, one can conceive almost anything. Thus, the fact that we can conceive that p does not entail that p is even possible: all that follows (at best) is that we have not yet noticed any contradiction in p. But, as our previous discussion indicates, Descartes would turn this objection aside by pointing out that his argument is not based on mere conceivability, but on clear and distinct conceivability. One cannot ignore this crucial distinction without radically misunderstanding his position. ____________________ 2 Versions of the first objection considered here are found in Kenny, Descartes; Leibniz, "'Observations on the General Part of Descartes'" Principles', in Philosophischen Schriften, iv; and Malcolm, "'Descartes's, Proof that his Essence is Thinking"; and of the third in Kenny, Descartes, and Malcolm, "'Descartes's Proof'". These objections have also been discussed by Michael Hooker in unpublished writings. The second objection is one that I have often heard in discussion, though I do not know of any source in the recent Descartes literature. -190I do not wish to claim that the appeal to the distinction between clear and distinct perception and mere perception raises no problems of its own. It raises, of course, the important question of how one recognizes clear and distinct perceptions. I will not attempt to evaluate this problem here. Instead I will turn to a criticism of Descartes' use of the notion of distinct perception in the epistemological argument that is, unquestionably, more directly relevant than the objections mentioned above. II The author of the first set of Objections to the Meditations, whose name was Caterus, found fault with Descartes' attempt to reason from the fact that A and B are distinctly and separately conceived to the conclusion that A and B can exist apart. He writes: Here I match the learned man against Scotus, who says that for it to be the case that one [thing] is conceived distinctly and separately from another, a distinction of the sort called formal and objective--intermediate between a real [distinction] and one of reason--is sufficient. And thus he distinguishes God's justice from His mercy; for, he says, they have concepts [rationes] formally diverse before all operation of the understanding, thus that even then the one is not the other; and

nevertheless it does not follow: justice can be conceived separately from mercy, therefore can also exist separately. ( A T VII100 [ C S M II72-3]) Caterus here does not exactly follow Descartes' 'clear and distinct conception' terminology; nevertheless, he has put his finger on a problem that Descartes must come to terms with. For Descartes himself holds that such 'simple natures' as extension, figure, and motion can each be clearly and distinctly conceived in itself; yet at the same time they are not really distinct; figure cannot exist apart from extended body, and so forth. (See especially Rules for the Direction of the Mind, Rules II and XIV, A T X 410 ff. [ C S M I39].) In his doctrine of simple natures Descartes appears to be squarely committed to the negation of the principle that what can be clearly and distinctly conceived in separation can exist separately. Descartes replies to Caterus by stressing a distinction between complete and incomplete beings. As to the matter of formal distinction . . . I briefly say that it does not differ from a modal one, and extends only to incomplete beings, which I have accurately distinguished from complete [beings], for which [distinction] it indeed suffices that one [being] is conceived distinctly and separately from another by intellectual abstraction from a thing inadequately conceived, not however so distinctly and separately that we understand one or the other [being] as an entity in itself [ens per se] and -191distinct from all others. But for the latter to be the case a real distinction is always required. ( A T VII120 [ C S M II85-6]) Descartes goes on to give precisely the sort of example we would expect, in view of the doctrine of simple natures. Thus, he says, The distinction between the motion and the figure of the same body is a formal one; and I can quite well understand the motion apart from the figure, and the figure apart from the motion; and I abstract both from the body: but nevertheless I cannot understand motion completely apart from a thing in which the motion is, nor the figure apart from a thing in which the figure is, nor motion in a thing in which figure cannot be, or figure in a thing incapable of motion. ( Ibid. ) The same point, Descartes says, applies to the example brought forward by Caterus. With these cases Descartes contrasts the mind-body case: But I completely understand what body is [French version: that is to say, I conceive of a body as a complete thing] merely by thinking that it is extended, figured, mobile, etc., and denying of it all those things which

pertain to the nature of the mind; and vice versa I understand the mind to be a complete thing, that doubts, understands, wills, and so forth, although I deny that any of those things contained in the idea of body are in it. The gist of this passage seems to be that we can conceive body and mind not only distinctly, but as complete things, while denying of each whatever pertains to the nature of the other. Justice and motion, on the other hand, while perhaps capable of being understood distinctly 'in separation', are not thereby capable of being understood 'completely'--i.e. as complete beings. I do not know what passage or passages Descartes may have in mind when he says he has 'accurately distinguished' complete from incomplete beings. Certainly this distinction does not seem to be made explicit in the Meditations. 3 Further, we have seen that the argument, as Descartes states it, begins with the unrestricted claim that: 'It is enough that I understand one thing clearly and distinctly apart from another, to know that one is different from another, for they can be placed apart, at least by God . . .'. This statement must now be rephrased. In order to be able to conclude that A is different from B in the relevant way--i.e. really distinct--one must be able to conceive A clearly and distinctly and completely (as a complete being) apart from B. Also, we can now see that Descartes' further statement in the argument, that he has a clear and distinct conception of himself in so far as he is 'only a thinking thing, not extended', must be given a different reading than that reflected in premiss (5), above. He must be ____________________ 3 Although in the Sixth Meditation ( A T VII78, line 8 from bottom [ C S M II54]) he does speak of understanding himself distinctly as a 'whole' without the faculties of imagination and sense. -192saying both that the concept of himself as a thinking thing comprises no notion of extension and that in thus conceiving himself as a thinking thing he clearly and distinctly conceives of himself as a complete being. III In the Fourth Objections, Antoine Arnauld picks up on Descartes' remarks to Caterus about the need for 'complete knowledge' as a basis for the mind-body distinctness argument. Arnauld reads this as an acknowledgement that the argument will go through only if our knowledge of ourselves as thinking things is, demonstrably, complete in the sense of being exhaustive. He further observes that nothing in the Meditations seems to bear at all on this problem except the argument in the Second Meditation that one can be certain of one's own existence as a thinking thing while doubting or denying the existence of body. But, he concludes,

. . . all I can see to follow from this, is that a certain notion of myself can be obtained apart from [the] notion of body. But it is not yet quite clear to me that this notion is complete and adequate, so that I am certain that I am not in error when I exclude body from my essence. ( A T VII201 [ C S M II141]) According to Arnauld, then, Descartes is not entitled to conclude that extension does not belong to his essence merely from the observation that he clearly and distinctly perceives that thought is essential to him while he 'notice[s] nothing else to pertain to [his] nature'. For perhaps in perceiving himself as a thinking thing, he is perceiving, so to speak, only part of his essence. Arnauld is in effect taking issue, specifically, with the following statement from the epistemological argument--which signals the transition from the conclusions of the Second Meditation to those of the Sixth: 'From this very fact, that I know I exist, and that meanwhile I notice nothing else to pertain to my nature or essence except this alone that I am a thinking thing, I rightly conclude that my essence consists in this one [thing] that I am a thinking thing.' Here Arnauld adduces the case of a man who clearly and distinctly conceives that a given triangle is right-angled, yet lacks a perception of the proportion of sides to hypotenuse. Because his knowledge of the triangle is in this respect incomplete, the man is able to doubt, and even deny, that the sum of the squares on the sides is equal to the square on the hypotenuse. According to Arnauld, this man would be in a position to reason, in a way parallel to Descartes, that since the clear and distinct idea of a right triangle does not include the notion of Pythagorean proportion, God can make a right triangle with some other -193proportion among the squares. This conclusion, however, is false. 4 So the epistemological argument must be invalid. Arnauld has misunderstood Descartes' use of the distinction between complete and incomplete knowledge in his reply to Caterus. However, as Descartes seems to recognize, clearing up this misunderstanding is not all that is necessary in order to answer Arnauld. Descartes rightly takes Arnauld's main question to be: 'Where did I begin to demonstrate how it follows from the fact that I know nothing else to belong to my essence . . . except that I am a thinking being, it follows that nothing else does truly belong to it?' ( A T VII219 [ C S M II154]). And he answers: Surely where I have proved that God exists . . . who can do all that I clearly and distinctly know to be possible. For although much exists in me which I do not yet [at this stage of the Meditations] notice . . . yet since that which I do notice is enough for me to subsist with this alone, I am certain that I could have been created by God without other [attributes] which I do not notice. ( Ibid. ; emphasis added)

Hence, these other attributes may be judged not to belong to my essence, since 'none of those [properties] without which a thing can exist is comprised in its essence'. (There is a suspicion of 'could-can' sloppiness in the Latin, which I won't try to evaluate here.) Descartes further explains that when he spoke, in the First Replies, of the need for 'complete knowledge', he did not mean exhaustive knowledge of the subject--as Arnauld seems to have assumed. Rather, he meant 'knowledge of a thing sufficient to know it is complete', i.e. 'endowed with those forms or attributes, which are sufficient that from them I recognize that it is a substance'. He concludes: Mind can be perceived clearly and distinctly, or sufficiently so for it to be considered a complete thing, without any of those forms or attributes by which we recognize that body is a substance, as I think I have sufficiently shown in the Second Meditation; and body is understood distinctly and as a complete thing without those which pertain to mind. ( A T VII223 [ C S M II157]; emphasis added) He goes on to observe that Arnauld's triangle example is not effective against him, since it 'differs from the case at hand' in making no use of the notion of 'complete knowledge' in the sense that Descartes originally intended. Arnauld's basic objection was that, for all Descartes knows, some other attribute, such as extension, might be necessarily implicated in his essence together with the known attribute of thought; the only way of eliminating ____________________ 4 4 Descartes' doctrine of the creation of the eternal truths is not brought up in this context. -194this possibility is to establish that one knows all the properties of the self. Descartes' position, however, is just that since he recognizes that thought is sufficient 'for me to subsist with [it] alone', he thereby knows no other attribute is necessary. To claim that thought and extension are different, and that either is sufficient to determine a complete or true thing, is already to deny the possibility of some 'hidden' necessary dependence of a thinking thing on the attribute of extension. Thus a 'complete knowledge' in Descartes' originally intended sense is sufficient for the epistemological argument to go through.We may now obtain a clearer understanding of the intended relation between the Second Meditation and the Sixth--indeed, Descartes seems finally to make this relation explicit in the important passage I have quoted from the reply to Arnauld. The Cogito reasoning and its immediate sequel are intended to establish, precisely, that 'mind can be perceived clearly and distinctly, or sufficiently so for it to be considered a complete thing, without any of those forms or attributes, from which we recognize that body is a substance . . .'. I think this explains, for example, Descartes' insertion into the Second Meditation of the statement that he knows he is a true and truly existing thing, merely in conceiving himself as thinking. The role of the epistemological argument in the Sixth Meditation is merely to

establish that the perception of the mind argued for in the Second Meditation (clearly and distinctly perceived as a complete thing in virtue of having the property of thought) is sufficient ground for the conclusion that the mind is really a distinct thing. What is primarily needed, besides the conclusions of the Second Meditation, is the validation of clear and distinct perceptions as reliable guides to reality. 5 Discussion of the objections of Caterus and Arnauld has shown the need for some changes in the analysis of the epistemological argument offered above. I suggest, finally, the following reading: If A can exist apart from B, and vice versa, A is really distinct from B, and B from A. Whatever I clearly and distinctly understand can be brought about by God as I understand it. If I clearly and distinctly understand the possibility that A exists apart from B, and B apart from A, then God can bring it about that A and B do exist in separation. ____________________ 5 I think these observations also help one understand Descartes' notorious tendency, ill works other than the Meditations, to pass without visible transition from 'cogito ergo sum' to 'sum res cogitans'. -1951. 2.

3. 4.

If God can bring it about that A and B exist in separation, then A and B can exist apart and hence, by (1) they are distinct. I can clearly and distinctly understand the possibility of A and B existing apart from each other, if there are attributes Φ and Ψ, such that I clearly and distinctly understand that Φ belongs to the nature of A, and that Ψ belongs to the nature of B, and that Φ ≠ Ψ, and I clearly and distinctly understand that something can be a complete thing if it has Φ even if it lacks Ψ (or has Ψ and lacks Φ). Where A is myself and B is body, thought and extension satisfy the conditions on Φ and Ψ respectively. Hence, I am really distinct from body and can exist without it.

How good (or bad) is the epistemological argument when interpreted in the way that (as I maintain) Descartes intended it? Well--to mention only one problem--it is at the very best no better than the distinction between clear and distinct perception and 'mere' perception. And while I have made some attempt to clarify this distinction, I must admit to distrusting it very radically. (Though I do not know whether recent essentialists' appeals to intuition are on any better ground.) On the other hand, the argument seems to me stronger and much more carefully thought out than Descartes' critics--contemporary or recent--have generally recognized. In particular, Descartes' reply to Arnauld is so direct and apposite that there can be, I think, no justification for repeating and endorsing Arnauld's objection without giving serious, systematic consideration to the reply. 6 ____________________ 6 Here I mean to take issue with, e.g. Kenny treatment in Descartes91 ff. In my introduction to Wilson (ed.), The Essential Descartes, I, too, implied it was questionable whether Descartes' reply to Arnauld is at all cogent--a suggestion that I

now wish to retract. -196VIII DESCARTES AND THE UNITY OF THE HUMAN BEING GENEVIÈVE RODIS-LEWIS Descartes is often accused nowadays of having destroyed the unity of the human being. In the Aristotelian tradition, of course, that unity is fundamental--so much so that we find Aquinas saying that the reuniting of soul and body at the resurrection is a return to the 'normal' state, in contrast to that in which the soul is separated from the body. 1 In reducing the essence of the soul to thought alone, and the life of the body to pure mechanism, Cartesianism was bound to encounter the thorny question of the relationship between two entities that were supposed to be perfectly distinct and of opposing natures. Occasionalism, parallelism, and (in some respects) pre-established harmony, albeit postCartesian solutions, were all conditioned by the ontological dualism of the Cartesian system. As for Cartesian anthropology, it appears to reencounter the difficulties that had beset Platonism. The incorporeal 'I' of the Cogito has no intrinsic need to make use of images; on the contrary, in order to make himself 'rather better known and more familiar' to himself, Descartes adopts the asceticism of the philosopher who strives to escape from this mundane world: 'I will now shut my eyes, I will stop my ears, I will lay aside all my senses; I will eliminate from my thoughts all images of bodily things, or rather, since this is hardly possible, I will regard all such images as vacuous and false.' 2 Is this ____________________ Geneviève Rodis-Lewis, "'Descartes and the Unity of the Human Being'" translated especially for this volume by the editor, John Cottingham, English translation Copyright © John Cottingham 1998, from "'Le paradox cartésian'", in 'L'anthropologie cartésienne by G Rodis-Lewis ( Presses Universitaires de France, 1990), 85-99. Copyright 1990 by Presses Universitaires de France, by permission of the publishers. 1 In taking over the Aristotelian definition of the soul as 'form of the body', St Thomas managed to preserve the 'spirituality' of the soul in an original manner. Cf. Gilson, Études sur le rôle de la pensée médiévale dans la formation du système cartésien, part ii, ch. 8, p. 245, where the notions of Thomist and Cartesian 'anthropology' are introduced. 2 Start of the Third Meditation ( A T IX27 (French) [ A T VII34 (Latin): C S M II24 (English)]). [In the original article, the author refers mainly to the 1647 French translation of the Meditations; references to the original Latin and to the standard modern English version are supplied for the reader's convenience. Ed.] -197not indeed a kind of 'angelism'? 3 And, as we all know, 'he who desires to be an angel . . .' 3a

Moreover, since the body can perform all the vital functions without being animated by the mind, animals turn out to be merely wellconstructed machines; and 'if you now take it as established for certain that the brutes are nothing but mere automata, are you not afraid that one day we shall believe that human beings too are nothing but mere machines?' 4 The question was posed in 1671, to be answered in 1747, with the publication of La Mettrie L'Homme machine; and many materialists were to hail Descartes as the one who had done away with all spiritual entities by exposing their obscurity and uselessness for the explanation of biological phenomena. Now, this historical summary is accurate enough in outline, if we adopt the perspective of hindsight; but should the actual views of Descartes himself be interpreted as containing the seeds of these subsequent developments? We cannot of course go back in time, but we can at least glimpse the originality of the Cartesian outlook if we consider it in relation to the views of Descartes' contemporaries; in this way we shall avoid the distortions of the Cartesian system which arise when one examines it from the perspective of the systems that succeeded it--systems which diverged very considerably from the source of their initial inspiration. I In the early seventeenth century many objections were being raised against scholasticism, but among the most vexed issues were those connected with the problem of soul and body: the theory of substance, immaterialism and associated debates on immortality, and, lastly, the status of biology. On the scientific level, there was general opposition to the 'forms' and 'qualities' which the Peripatetics invoked to explain natural phenomena. ____________________ 3 This term is used by J. Maritain, who provides the following summary of the 'Cartesian legacy' on this issue: 'Cartesian dualism separates the human being into two complete substances, mysteriously joined: a body that is mere geometrical extension, and a soul which is mere thought. Man becomes an angel inhabiting a machine and directing its operations by means of the pineal gland' (paper collected in Le Songe de Descartes ( Paris, 1932) [trans. as The Dream of Descartes ( New York: Philosophical Library, 1944]). 3a 'He who wishes to act the angel ends up playing the beast' (Qui veut faire l'ange fait la bête') ( Pascal, Pensées, no. 678). 4 Jacques Rohault [ 1620-72], a Cartesian whose reputation was based principally on his Physics, raises this objection in his Entretiens sur la philosophie [ 1671], ii. 191. His own answer to the objection is that thought cannot deny itself. -198The new mechanist movement had a wide range of followers, 5 including, apart from Descartes, thinkers as diverse in outlook as Gassendi and his circle, the physicists who

gravitated around the Pascals, father and son, and Father Mersenne, an indefatigable intermediary between the various researchers of the time. This was already a century of engineering, and machines of various kinds were being constructed everywhere. 6 Descartes' achievement, bold though it was, was simply to provide a systematic account of mechanistic explanation--one which relegated organic teleology to the status of a result as opposed to a cause. 7 The recent discovery of the circulation of the blood provided the prototype for this approach: an arrangement of pipes going from the visible vessels to the imperceptible network of nerves where the 'animal spirits' circulated. The latter were reminiscent of the medical theories of Galen only in respect of their name; Descartes' 'animal spirits' are merely eminently subtle matter passing from the brain through the nerves, where depending on the varying amounts which enter, or merely tend to enter, some nerves more than others, they have the power to change the shape of the muscles in which the nerves are embedded, and by this means to move all the limbs. Similarly, you may have observed in the grottoes and fountains in the royal gardens that the mere force with which the water is driven as it emerges from its source is sufficient to move various machines, and even to make them play certain instruments or utter certain words depending on the various arrangements of the pipes through which the water is conducted. ( Treatise on Man ( A T XI130 [ C S M I100])) The image becomes an explanatory schema: when the brain relaxes during sleep, this is because it is moistened with blood, and hence requires less force from the animal spirits, 'just as the wind does not need to be so strong in order to fill the sails of a ship when they are wet as when they are dry' ( A T XI197). And in the absence of a clear interpretation of chemical phenomena, the rudimentary character of unadorned mechanical explanation is compensated for by another comparison, deployed to account for the transformation of food into chyle and blood--'just as the juice of black grapes, which is white, changes into red wine when it is allowed to ferment in the wine-press'. 8 All these physical phenomena exemplify modifications ____________________ 5 Compare R. Lenoble illuminating study Mersenne et la naissance du méchanisme ( Paris: Vrin, 1943). 6 Cf. J. Baltrusaitis, Anamorphoses ou perspectives curieuses ( Paris, 1955), ch. 3, p. 37, which discusses the work of the engineer-architect Salomon de Caus, and draws a comparison with some of the descriptions found in Descartes' L'Homme ( A T XI 130 2), to be discussed later on. 7 The neo-Epicurean mechanist Gassendi, by contrast, retained final causes in the case of living beings; compare the Fifth Set of Objections to the Meditations ( A T VII309 [ C S M II215]). 8 A T XI123. Compare the explanation of fermentation on p. 121 : 'just as happens with new hay in the barn, when it is tossed around to dry', the heat separates the various parts 'just as ordinary water does to quick lime'.

-199of bodily structure arising solely as a result of various local motions; the qualitative processes of alteration which figured in scholasticism become redundant. Indeed, to avoid recourse to a 'pulsing faculty' in the heart, Descartes actually distorted Harvey's observations; 9 this is very like the way his successors were to rise up against the occult character of Newton's 'attractive force'. The rearguard defenders of scholasticism, by contrast, would in turn object to the Cartesian mechanical paraphernalia of wheels and pipes: 'you are trying to take a school of natural philosophy and turn it into a locksmith's shop!' 10 Naïve and archaic as these various reactions may seem, they none the less transcend the framework of a mere chapter in the history of science. For the mechanism developed by Descartes was the corollary of a metaphysical revolution. As a young man, Descartes was at first greatly taken up with a variety of different problems, like the other 'physicomathematicians' referred to in his friend Isaac Beeckman's journal (cf. A T X52). Nevertheless, Cartesianism proper begins only with the demonstration of a complete distinction between mind and body; what emerges next is the uselessness of 'forms' and 'qualities' for the solution of questions in physics; and finally, the progress made in scientific explanation in turn confirms the dualistic starting-point. 11 The concern which Descartes showed to graft his natural philosophy on to its metaphysical roots was a source of surprise to the other mechanists, who were quite happy to ignore the speculations of the scholastics. 12 In order to establish his system as a rival to that which dominated the universities of the time, Descartes adopted, but transmuted, the traditional notions of 'degrees of being' and the various distinctions between substances, attributes, and modes. 13 But we should not exaggerate either his conformity or his opposition to philosophical tradition: the scholasticism which he had studied in many respects owed more to Suarez than to Aquinas. 14 The dominant element was essentialism, 15 and a 'real ____________________ 10 Lettre d'un philosophe à un cartésien de ses amis, attributed to P. Rochon ( Paris, 1672), no. 78, p. 182. 11 Cf. Sixth Set of Replies ( A T IX238-9 [ A T VII442-3: C S M II298]). 12 This emerges clearly from Lenoble Mersenne et la naissance du méchanisme. As H. Gouhier's review of the book ( Revue philosophique, 1944) makes clear, its results go against the common thesis that Descartes' metaphysics is merely designed to protect his physics: 'In fact the audience to which his physics was addressed would have been only too happy to avoid metaphysical questions completely.' 13 Descartes develops these points above all in the first part of his Principles of Philosophy, which he hoped to see adopted as a university textbook. 14 Cf. Gilson Index scolastico-cartésien, and Études, etc. See also G. Rodis-Lewis, L'Individualité selon Descartes ( Paris, 1950), ch. 1. 15 Cf. E. Gilson, L'être et l'essence ( Paris, 1948), ch. 5, pp. 149-52, and, for the influence on Descartes, 155 -9.

9

Cf. Gilson, Études, part i, ch. 2, pp. 80-95. -200-

distinction' 15a between two items tended to be interpreted in terms of their being actually separated. 16 Moreover, alongside the tradition which flourished in the Schools, there was also a revival of Platonism, which attacked the materialist interpretations of Aristotle's elliptical texts on the subject of immortality. 17 Descartes invoked the exhortation of the Lateran Council of 1513 urging Christian philosophers to support the faith by providing rational demonstrations; 18 and the proof which he supplied of the real distinction between soul and body, a proof which in its fundamentals was an original one, reverted to the conception of the human being as a 'conjunction of two completely distinct natures'--so distinct, in fact that Du Pont, author of a work that was very widely read at the time, went so far as to dub the human being an 'animal-angel'. 19 The Augustinian thesis that mankind is 'something intermediate between the beasts and the angels (medium quoddam inter pecora et angelos)' 20 was at the time generally understood in the sense of a juxtaposition: 'an angel, an animal', repeats Bérulle; 21 and his successor at the Oratory, Senault, offers a curious synthesis of this dualism and the theory of knowledge taught in the Schools. By then the theme was so familiar that, as Senault puts it, 'all philosophers admit that mankind is composed of body and mind, and that, when divine Providence formed him, it united beast and angel'. 22 But in the case of Adam alone was ____________________ 15a Distinctio realis: a distinction between two things (Latin res). 16 Cf. H. Matthieu, F. Suarez, sa philosophie ( Paris, 1921), 113, 120-1, 169, 286. 17 Pomponazzi and other interpreters of the school of Padua insisted that the passage at the end of [chapter 5 of] Aristotle De anima, Book 3, amounted to a denial of individual immortality, since the close relationship between soul and body (the former being related to the latter as form to matter) made it impossible. But alongside this denial of immortality on the basis of reason, they juxtaposed its affirmation on the basis of faith (a kind of 'two truths' theory which the Church regarded as suspect). Cf. J.-R. Charbonnel, La Pensée italienne au XVIe siècle ( Paris, 1919); H. Busson, La Pensée religieuse française de Charron à Pascal ( Paris, 1933). 18 See Dedicatory Letter to the Meditations ( A T IX5 [ A T VII3: C S M II4]). The title of the first Latin edition (possibly supplied by Mersenne in 1640) announced a 'demonstration of the immortality of the soul'; Descartes, however, explained that he could not provide a complete demonstration [ A T VII 153 : C S M II 109 ], and in the second ( 1642) and subsequent editions the title reads instead: ' Meditations. . . in which are demonstrated the existence of God and the distinction of the human soul from the body (in quibus Dei existentia et animae humanae a corpore distinctio demostrantur)'. 19 Du Pont, La Philosophie des esprits, 243-4. The book was published in 1602, and reissued in 1606, 1612, 1628, and 1646. 20 Augustine, De civitate Dei, ix. 13.

21

Bérulle, Œuvres de piété, 114, 1644 edn., p. 969. For this kind of juxtaposition, cf. Dagens, Bérulle et les origines de la restauration catholique ( Paris, 1952), 227: 'a nothingness, a miracle' etc.--a series of antinomies which recall Pascal. But the theme of the 'animal-angel' was by then so common that the famous phrase quoted in Pascal Pensées, 'ni ange ni bête' is undoubtedly not meant to be attributed to a specific author. 22 Senault, L'Homme criminel ou la corruption de la nature par le péché, 1st edn., 1644, traité I, discours 1 (pp. 5 - 6 of 1647 edn.). The phrase is also found in another work of Senault, L'Usage des passions ( 1641; p. 73 of 1653 edn.): 'il est ange et bête tout ensemble'. -201it true that 'when God created this masterpiece, he brought the two elements that make up the creature into such apt harmony that the body was ready to obey the mind, and the angel was in charge of the beast' ( L'Homme criminel, 6). As a result of sin, however, the soul became the slave 'of its body, and, in particular, 'since its rebellion it is the slave of the senses, and cannot discover the truth except by means of them' ( L'Homme criminel, traité II, discours 2, passim, and discours 3, p. 39). So Senault concludes, with Aristotle, nihil in intellectu quod non prius fuerit in sensu. . . 22a This historical excursus will put us in a better position to grasp the distinctive position of Descartes. In his case, there is no question of speculating on the consequences of original sin. As we know, rational reflection puts an end to methodical doubt by reaching the certainty of the Cogito: I cannot think without existing, or be ignorant of this thinking thing which is immediately aware of itself--this is an indubitable reality, even though the existence of the body is still in doubt. Moreover, in the development of the Meditations, which is more complex than the summary version we find in the Discourse on the Method, the mutual independence of the two clear and distinct ideas which I have of mind and of matter becomes an ontological dualism of two heterogeneous substances only with the guarantee provided by God; this ensures the truth of what reason makes me see as evident. Thus, the distinctness of the two ideas corresponds to the real distinctness of the entities themselves. Now the paradoxes begin to multiply. The soul can think without the body, despite the trouble it has dispensing with images ('this can only happen with difficulty'); and so it can encompass by means of pure thought, and without any brain traces, spiritual entities at least. Since thought constitutes its entire essence, thinking activity never ceases, from the time of its creation, and hence even before birth, from the moment of its infusion into the body; moreover, this thought is always accompanied by the spontaneous awareness that defines it. 23 Thus the subject apprehends itself directly as object: 'Cogito, ergo sum res cogitans'; and even when I am contemplating another being--a piece of wax, a chimera, or whatever--the correlate of the cogitans, the thinker, is first of all posited as a cogitatum, a mode of my thought. Descartes takes as his starting-point ____________________

22a

'Nothing in the intellect which was not previously in the senses'. Descartes' definition of thought as awareness--in French conscience--introduces the term 'conscience' in its philosophical sense into the French language. Its nature (in contrast to the scholastic view) is not [necessarily] reflexive, and this makes possible a subconscience [a level of thought below that of reflective awareness]: a 'confused' feeling of which we lose all memory, occurring in such limited states as sleep, fainting, or prenatal life. See further G. Rodis-Lewis, Le Problème de l'inconscient et le cartésianisme ( Paris, 1950), ch. 1, pp. 37-103.

23

-202the most subjective kind of idealism, but only so that he may better attain to real objects, through knowledge of the Being who created him, and of the various essences and existences. The 'idea' remains one of my thoughts, but it discloses a true and immutable nature: the idea of the infinite, exceeding the capacity of my finite nature, allows me to be sure of the existence of God. 24 Further, the idea of extension, differentiated by shape and motion, the source of all the properties to be demonstrated by geometry and mechanics, is too different from thought to be directly derived from it; and if God himself produced these modifications which I spontaneously take to be the result of the impression of external bodies, or if he allowed them to be placed in me by some other spirit, 'I do not see how one could acquit him of deceit' ( Sixth Meditation ( A T IX63 [ A T VII 80: C S M II55])). Thus the provisional idealism of the cogitata (objects of thought) gives way to a realism so profound that even truths are entities that are wholly dependent on the Creator. Here we have a fresh paradox which the majority of Cartesians were to reject, preferring to retain the traditional doctrine of eternal, intrinsically necessary truths, inhering from the very beginning in the divine wisdom. In the later history of ideas we can see the consequence of the gap originally opened up by Descartes between on the one hand the cogitata or objects of thought, indubitable in their very subjectivity, and, on the other, the supposed entities corresponding to them. The result was to shut thought in on itself by reifying it. When Descartes spoke of a res cogitans, he was referring to a primitive reality that was essentially active, as indicated by the verbal form cogitans (the present participle active); but the items that are 'thought of' (pensées, past participle passive), in all their diverse forms, appear as 'contents of awareness', and this opens the way to psychologism. The ambivalence of the Cartesian theory of the idea, on the one hand a mode of my thought, on the other hand pointing to an extra-mental reality, 25 was destined to feed the long dispute between Descartes' two 'disciples', Arnauld and Malebranche. The former, more faithful to the letter of the Third Meditation, insisted on the idea as a mode (or, as we would put it, 'state of awareness'), and was accordingly accused of subjectivism. But by adding that an idea is essentially representative of ____________________ 24 Descartes concedes to the empiricists the possibility of my constructing the idea by extrapolating from finite perfections; but absolute perfection is unitary, and it is this

movement to the limit, the faculty of passing from the progressive indefinite to the positive infinite, which could only have been bestowed on me by a Being that is in reality infinite and perfect. 25 The Third Meditation terms the first aspect 'formal reality' and the second 'objective reality'--what is pointed to or represented by the idea; further, corresponding to the objective reality of a true idea, there is another 'formal reality', that of the external object. -203something else, he also grafted some elements of the scholastic theory of intentionality on to Cartesianism. He was, however, insufficiently critical of the deliverances of mental 'awareness', reducing the concept of self-evidence to a kind of 'feeling' which carried conviction in virtue of its force or liveliness. The subsequent, empiricism of the eighteenth century and the mentalistic eclecticism of the nineteenth were also to draw on Cartesian sources. Malebranche, by contrast, insists on the objectivity of essences, but construes them as an 'intelligible world', a world 'in God' that is directly accessible to our thought. As for the immaterial 'I', in so far as it no longer contains ideas, it is more than ever cut off from material entities. The realism of ideas is linked to an affirmation of real bodies, but only faith assures us that God has created them; it is he who has established, without any special intervention whatsoever, a perfect correspondence between the two domains of ideas and of things. Thus the immateriality of the soul, a result of Descartes' Cogito, now no longer allows the soul to be conjoined to any body--even the body that belongs to it: it is 'on the occasion' of a material change that the mind, experiences such and such a feeling, without the latter being a real effect of the former. II Descartes himself, however, even if some of his expressions lent themselves to this kind of occasionalism, 26 had always affirmed the substantial union of soul and body, and the essential unity of the human being. Here again, his successors destroyed that fragile balance between the distinction between soul and body on the one hand and their union on the other. For it was not just to avoid the thunderous disapproval of the Dutch scholastics (all too readily provoked by other affronts to their views) that Descartes condemned Regius for reducing the human being to an 'accidental entity'. 27 The real union, in fact, dates from the very first presentation of Descartes' metaphysics, and is one of the fundamental themes of the Sixth Meditation. 28 Upon satisfying himself of the existence of the external world, Descartes comes up against the special status of 'this body' which ____________________ 26 Cf. H. Gouhier, La Vocation de Malebranche ( Paris, 1926), ch. 3, pp. 83 ff. 27 Letter to Regius, mid-Dec. 1641 ( A T 111460 [ C S M K200]).

28

Compare the admirable analysis of its almost musical compositional structure in Gueroult, Descartes selon l'ordre des raisons, ii. -204-

'by a certain particular right I call mine'; I am 'not merely lodged in it like a pilot in his ship' ( A T IXA 60, 64 [ A T VII76, 81: C S M II52, 56]). This latter condition would be that of an angel if it were conjoined to a body, and Descartes, who hardly ever speculates on pure spirits, invokes this limiting case to underline the difference. 29 I am 'conjoined very closely and so mixed and intermingled with the body that I make up as it were a single whole with it' ( A T IXA 64 [ A T VII81: C S M II56]); I am 'a one single person who has both a body and a mind together'. 30 Neither angel nor beast-machine, the human being constitutes something that is unique in all creation. This organized fragment of matter, 31 its body, enjoys a privileged unity, in so far as it is 'informed' by the soul, and 'in this sense it is indivisible', since 'we do not think that someone who has an arm or a leg amputated is less a human being than any other'. These statements from letters to Father Mesland 32 are not in any way a concession to the traditional vocabulary, 33 or a residual formula from previous ways of talking, ineptly tacked on to the surface of Descartes' dualism. On the contrary, they refer to the reality of a substantial form, the 'only' authentic example of the genre; 34 and this reveals the genesis of the classic error of the, Aristotelians--an anthropomorphic generalization, where everything is, as it we're, 'animated' by little entities. 35 Furthermore, at the same time as he sets up his mechanistic model of scientific explanation, by excluding all qualities from matter, Descartes discovers the true meaning of qualitative impressions: the sphere of 'feeling' does not teach us anything about objects as they are in themselves, but is of vital 'importance' for us. For it turns out that sensations and passions affect us 'only in virtue of the various ways' in which objects can harm or benefit us, or be important to us in general' ( Passions of the Soul, art. 52). Thus, since soul and body constitute 'a single whole', they co-operate, by ____________________ 29 Letter to Regius, Feb. 1642 ( A T III493 [ C S M K206]): 'If an angel were united to a human body, it would not have feelings like us, but would merely perceive the movements caused by external objects, and in that respect would differ from a genuine human being.' 30 Letter to Elizabeth, 28 June 1643 ( A T III694 [ C S M K228]). 31 Descartes readily admits the 'disposition of the organs' necessary for the functioning of the automaton--a disposition which presupposes the ingenuity of the Author of Nature, 'acting throughout in accordance with the exact laws of mechanics . . . imposed by God' (to Mersenne, 20 Feb. 1639 ( A T II525 [ C S M K134])). 32 Letter of 9 Feb. 1645 ( A T IV166 [ C S M K243]). Compare the letter of 1645 or 1646: 'the numerical unity of the body of a human being does not depend on that of its matter, but on its form, which is the soul' ( A T IV346 [ C S M K279]). 33 The letters in question, it should be noted, are concerned with reconciling the mystery

of the Eucharist with the new Cartesian account of matter. To Regius, Jan. 1642: 'agnoscitur sola esse forma substantialis; exemplo animae, quae est vera forma substantialis hominis' ( A T III503, 505 [ C S M K207, 208]). 35 Cf. Gilson, Études, part ii, ch. 1, no. 2, 'La psychologie de la physique aristotélicienne'. 34

-205means of a certain internal teleology, 36 in the conservation of the ensemble; but the natural independence of the two elements frees the destiny of the soul from the accidents that may interrupt the life of the body. 37 Not to fear death is one of Descartes' maxims, but he is also 'one of those who love life the most'. 38 And the whole of his final work is devoted to the phenomenon which is so characteristic of the close interaction between soul and body-the passions. For it is the passions which, if well directed, allow us 'to taste the greatest sweetness in this life' (art. 212). Undoubtedly, it is easier to experience the union between soul and body by actually living it, 39 than it is to explain it intellectually. 'It does not seem to me', Descartes wrote, 'that the human mind is capable of forming a very distinct conception both of the distinction between the soul and the body and of their union; for to do this it is necessary to conceive of them as a single thing and, at the same time, to conceive of them as two things; and the two conceptions are mutually opposed. 40 He also warned us of the temptation to picture or localize the interaction between the two substances: this would involve making the mind into something material, by representing the reciprocal action involved as entailing contact between two things. 41 Moreover, the privileged function of the 'pineal gland' as the central point of convergence for the incoming impressions, and the point of emergence for outgoing impulses, is merely a physiological explanation; it needs to be understood as complementing the fact that 'the soul is united to all the parts of the body conjointly' ( Passions, art. 30). Despite all this, the majority of Cartesians, 42 after vain attempts to ____________________ 36 The point was first made by J. Laporte, "'La Finalité chez Descartes'", Revue d'histoire de la philosophie, 1928, 366-98, and also taken up (despite considerable differences of interpretation) by Gueroult, Descartes selon l'ordre des raison, ii. 177-94. 37 Cf. Synposis of the Meditations ( A T IX10 [ A T VII11: C S M II10]): 'It suffices to show clearly enough that the death of the soul does not follow on the corruption of the body, and thus to give people the hope of a second life after death'--a hope that is confirmed by faith ( Second Replies ( A T IX120 [ A T VII153: C S M II108-9])). 38 To Huygens, Oct. 1642 ( A T III798 [ C S M K216]); cf. letter to Mersenne, 9 Jan. 1639: 'One of the points of my moral system is to love life without fearing death' ( A T II480 [ C S M K131]). 39 Cf. letter to Elizabeth, 28 June 1642: 'It is the ordinary course of life and conversation, and abstention from meditation and from the study of the things which exercise the imagination, that teaches us how to conceive the union of the soul and the body' ( A T

III692 [ C S M K 227 ]). Same letter ( A T III693 [ C S M K227]). 41 To Clerselier, Reply to Counter-Objections ( A T IX213 [ C S M II275]): the alleged problems about the interaction between the two elements arise from a supposition 'which is false and which cannot in any way be proved, namely that since the soul and the body are two substances of different natures this prevents their being able to act on each other'. Cf. Fifth Replies, points raised against Sixth Meditation, no. 4: 'it is not necessary for the mind to be of the same order and nature as the body, despite its having the power or virtue to move the body' ( A T VII389 [ C S M II266]). 42 Except Regis, who insisted that 'the soul's action is always dependent on the body to which it is united' ( Système de philosophie ( 1690), Bk. II, part i, ch. 4 ( Paris, 1960 edn., i. 161)). See also 40

-206clarify the relationship between the two conceptions which Descartes himself declared to be 'mutually opposed', ended up rejecting the 'third primitive notion', which is 'known only obscurely by the intellect alone, or by the intellect aided by the imagination' ( letter to Elizabeth of 28 June 1643 ( A T III692 [ C S M K227])). Yet, in reducing the union to a correspondence between two substances, without any interaction, 43 they abandoned one of the most original and most promising points of Descartes' theory: the practical significance of the sensations and passions. Thus, for Malebranche and Cordemoy, faith alone assures me that I have a body, and this is as much because of the Incarnation of Jesus Christ 44 as because I need a 'victim to sacrifice'. 45 But for all that, they appreciated that (in Malebranche's words) 'when we carefully consider the senses and the passions of the human being, we find they are so well adapted to the end for which they were given us, that we cannot agree with those who say that they are entirely corrupted by original sin'. 46 The internal teleology of our nature is thus preserved; and the development of this Cartesian theme by Cordemoy heralds some of the conclusions of our own contemporary psychology: the soul does not relate its sensations to what immediately causes them, for if that were the case, it is certain that, since all sensations reach the soul on the occasion of an internal movement of the brain, it would refer the sensations to the relevant brain areas. But on the contrary, it is good for the soul to refer these sensations to the original sources of the relevant stimuli, and beneficial that the brain can be stimulated from afar by means of the subtle bodies which are interposed between it and external objects, and thus be disposed to avoid or pursue these objects depending on their degree of utility. 47 Accordingly, when they are not too far away, the soul 'has sensations of objects at a distance, and it relates its sensations to the distant objects, in ____________________ the quotations from Desgabets in G. Lewis, Le Problème de l'inconscient et le

cartésianisme ( Paris, 1950), 150-4. Clauberg 'reciprocal traffic', and the occasionalism developed by La Forge and Cordemoy (cf. Gouhier, La Vocation de Malebranche) and systematized by Malebranche, are all genuinely Cartesian expressions of this line of thought. By contrast, although their approaches to the problem owe something to Cartesianism, the solutions of Spinoza (parallelism of attributes, but with a rejection of substantial dualism) and of Leibniz (pre-established harmony, but with a dynamic and idealist conception of 'body'), depart radically from those of the Cartesians. 44 Invoked by Géraud de Cordemoy in the Sixth Discourse of the Discernement du corps et de l'âme ( 1666), 144; and by Malebranche, Recherche de la vérité, éclaircissement VI. 45 Malebranche, Méditations chrétiennes, ix ( Œuvres complètes. x. 104, 19). 46 Malebranche, Recherche de la vérité, i. ch. 5 ( Œuvres complètes, i. 69). Books I and V develop this interpretation of the importance of the senses and the passions. 47 Cordemoy, Discernement, discours 6, pp. 212-13. 43

-207order to avoid harmful things before they get too near, and to be able to go after things that may be useful, when they are some way off'. 48 Even among those disciples of Descartes who were so excessively zealous in maintaining the 'distinction between the soul and the body' that they were apparently incapable of subsequently uniting them, the function of one's own body and the special nature of the domain of the senses were thus perfectly well grasped, without the least temptation towards 'angelism'. III To conclude: the principal difficulty for the Cartesians was not to contrast the angel and the beast in the human being, but to do justice to the absolutely exceptional character of this unique being that is neither one nor the other. The special character of the human being is amply shown by the role played in our thinking by sensations, the imagination, and the passions. And right from the point at which he arrives at the Cogito, Descartes is obliged, when enumerating the principal manifestations of 'thought', to take account of this dimension, even though at this stage in the argument he is still ignorant of its relationship to our body. 49 But, as for animals, who display the same passions as us, and make the same use of sense-organs in order to grab what is beneficial or escape from what is harmful to life-why do they not have the least bit of sensation or feeling? 50 Would not such feeling, which has a vital function, correspond to what we currently call 'instinct' in human beings? 51 Faced with ____________________

48

Cordemoy, 225. Compare the definitions of res cogitans in the Second and Third Meditations ( A T IX22, 27 [ A T VII28, 34: C S M II 19, 24 ]): 'and that also imagines and has sensory awareness'. The French version of the second of these texts adds 'that loves, that hates'. 50 Descartes recognizes that 'the movements of the passions can in the case of the brutes be even more violent than they are in human beings, without it being possible to infer from this that they have thoughts' ( to Newcastle, 23 Nov. 1646 ( A T IV573-4 [ C S M K302-3])). Cf. Passions, art. 138: 'all the animals devoid of reason conduct their lives simply through bodily movements similar to those which, in our case, usually follow upon the passions'. But they do possess sensus, in so far as this depends on bodily organ, that is to say, an excitation which provokes a reaction ( letter to More, 5 Feb. 1649 ( A T V278 [ C S M K366])). To render the Latin sensus as the French sentiment gives a very equivocal translation; animals do not feel anything. As Descartes puts it, 'I do not explain the feeling of pain in the absence of a soul' ( to Mersenne, 11 June 1640 ( A T III 85 [ C S M K148])). The opponents of the doctrine of 'animal-machines' were quick to charge the Cartesians with cruelty. 51 Cf. letter to Mersenne, 16 Oct. 1639: 'I distinguish two sorts of instincts: one is in us qua human beings, is purely intellectual, and is the natural light . . .; the other is in us qua animals, 49

-208countless protestations by the opponents of the concept of the 'animalmachine', the Cartesians simply repeated that since the soul was by nature immaterial, to grant a soul to animals would be to confer immortality on them. 52 So it is God alone who looks after the perfection of the animal mechanism, 53 and the relevant mechanical response is all the more impressive in that it operates entirely without hesitation or delay. 54 In the case of humans, by contrast, the gift of consciousness is linked to the possibility of a real choice with respect to the messages transmitted by the senses. And so the hypothesis of a 'human-machine' turns out to be strictly unthinkable, not just because each of us cannot mistake the evidence of his or her own consciousness, but Also because the behaviour of others, in virtue of its complexity, together with the employment of meaningful language, reveals to us the existence of genuine thought that corresponds to our own. 55 Perhaps this analysis of the Cartesian outlook itself has a paradoxical flavour, in so far as it is somewhat one-sided. To straighten a curved stick, says Descartes, we have to bend it back in the other direction; and to rediscover the true balance of Cartesianism, it has been necessary not to ____________________ and is a kind of natural impulse towards the preservation of our body . . .' ( A T II599 [ C S M K 140]). 52 Letter to Newcastle, 23 Nov. 1646 ( A T IV576 [ C S M K304]): 'Since the organs of their bodies are not very different from ours, it may be conjectured that there is

attached to these organs some thought such as we experience in ourselves, but of a very much less perfect kind. To this I have nothing to reply except that if they thought as we do, they would have an immortal soul like us. This is unlikely, because there is no reason to believe it of some animals without believing it of all, and many of them, such as oysters and sponges, are too imperfect for this to be credible.' Rohault, in the second of his Entretiens sur la philosophie, takes up this argument, while Martin (Ambrosius Victor), Poisson, and Malebranche find in St Augustine theological arguments against the possession of a soul by the beasts (the link between pain and sin, and so on). The famous article 'Rorarius' in Bayle Dictionnaire historique et critique ( 1697) marks out the principal tenets of the two camps, following a host of works on this topic which appeared in the second half of the seventeenth century. Whilst maintaining that the traditional view is the more probable, the article offers the following (malicious) comment: 'it is unfortunate that the view of Monsieur Descartes is so difficult to sustain, since it is in other respects very advantageous to the true faith.' 53 'A machine . . . made by the hand of God is incomparably better ordered than any machine that can be devised by man' ( Discourse on the Method, part v ( A T VI56 [ C S M I139])). 54 Discourse on the Method, part v ( A T VI57 [ C S M I140]): 'Whilst reason is a universal instrument which can be of use in all kinds of situations, these organs need some particular disposition for each particular action.' 55 The phenomenon of language, to which Descartes returns on several occasions, illustrates one form of non-rigidity in the responses of human beings. Cordemoy, in his Discours physique de la parole ( 1668), shows how these 'institutional signs' of language are quite distinct from spontaneous expressions of passion, and presuppose reason (p. 23 ); he also notes that humans are different from the beasts in that they make things which 'have nothing to do with themselves or their preservation' (p. 6 ). -209pay excessive attention to the isolated, immaterial 'I'. For all that he is regarded as the father of idealism in its many forms, Descartes would probably have disavowed the heavy weight of this subsequent philosophical tradition. But we are best placed to judge the issue on the basis of Descartes' own texts, by situating them alongside the work of his contemporaries and immediate successors. This has been the sole object of the present essay. -210IX DESCARTES' THEORY OF THE PASSIONS STEPHEN GAUKROGER Discussions of the nature of the relation between mind and body in the seventeenth century generally occurred in the context of accounts of the passions, and even without

Elizabeth's prompting [in the sequence of letters beginning 21 May 1643] Descartes would naturally have been drawn into the general question of the passions. They were associated with bodily conditions, bringing them under the purview of medicine and physiology, but they were also given ethical meanings (sadness--tristitia--was a sin for many centuries, for example), bringing them under the purview of moral psychology and theology. Treatises on the passions traditionally shared this combination of concerns, and Petrarch De remediis, for example, a compendium of Stoic techniques for 'healing the passions', can be read as a treatise on morals or psychotherapy: the distinction is simply not there to be made. 1 Much the same can be said of the whole tradition of writing about the passions, including Descartes' Passions de l'âme, which takes us through the physiology of the passions and the nature of melancholia, as well as pointing out that 'all good and evil in this life depend on [the passions] alone'. 2 By the term 'morals', Descartes tells us in the 1647 preface to the Principes, we are to understand 'the highest and most perfect moral system, which presupposes a complete knowledge of the other sciences [viz. medicine and mechanics] and is the ultimate level of wisdom'. 3 We clearly have to be extremely careful about how we characterize the boundaries of ethical discourse in the seventeenth century; 4 for ____________________ Stephen Gaukroger, 'Descartes' Theory of Passions' from S. Gaukroger, Descartes, An Intellectual Biography ( 1995), 394-405. Copyright © Stephen Gaukroger 1995, by permission of the author and Oxford University Press. 1 See Letizia A. Panizza, "'Stoic Psychotherapy in the Middle Ages and Renaissance: Petrarch's De remediis'", in M. Osler (ed.), Atoms, Pneuma, and Tranquillity. ( Cambridge: Cambridge University Press, 1991), 39 - 66. 2 Passions, art. 212 ( A T XI488 [ C S M I404]). 3 A T IXB 14 [ C S M I186]. 4 We have to begin by asking the Socratic question that Bernard Williams urges us to ask: 'what part might be played by knowledge of the sciences; how far purely rational enquiry might -211the practical aim of treatises on the passions is, generally speaking, the 'healing of the soul', something which involves moral and psychotherapeutic considerations, and these require an understanding of the bodily conditions associated with various afflictions of the soul which may give rise to anything from immorality to madness. It is in this light that we must understand Descartes' remark in the Discours, quoted earlier, that if we seek to make men wiser, we must look for the answer in medicine. There is undeniably an element of reductionism here, but it is not the crude reductionism it might at first seem; rather, it reflects what is in fact a complex nexus of considerations that make up an understanding of the nature of wisdom and goodness in this era. The traditional disputes over the nature of the passions had a number of features, but one basic polarity pervading them is that between what can broadly be termed Stoic and Augustinian conceptions of the passions. 5 The Stoics treated the passions as false

judgements, and, following an already strong tradition of intellectualist ethics in Greek thought, they identified virtue and knowledge. Augustine was motivated by a number of theological problems that had concerned the Alexandrian Church Fathers generally; Christological problems about how Christ's 'agony in the garden' was possible if he was God, for example, had led to a study of the nature of the passions, with Athanasius attributing the agony in the garden to Christ's body alone, and Clement distinguishing between bodily passions, which are necessary for the preservation of life, and passions of the soul. 6 A problem which was not at first immediately related to this, but turned out to be a key question because of the nature of Augustine's answer to it, was how there could be evil in the world if God was good. His solution was to make human beings responsible for evil, and the philosophical tool that he uses to achieve this is the notion of free will. Earlier theories of moral behaviour had invoked the notion of responsibility for action, but had not introduced the notion of the will in this context. Augustine did, and thereby explained the existence of evil. But the theory ____________________ take us; how far the answer to the question might be expected to be different if it is asked in one society rather than another; how much, at the end of all that, must be left to personal decision' ( Ethics and the Limits of Philosophy ( London: Fontana/Collins, 1985), 3). Were more historians of ethics sensitive to such questions, there would not be such a hiatus in histories of ethics between the Stoics and Kant's immediate predecessors in the eighteenth century. Indeed, more attention to this period might teach us something about the nature of our own dominant narrow ethical discourse, so effectively criticized by Williams. 5 The Stoic/Augustinian polarity extends beyond the question of the passions, and is a key one in Renaissance thought. See William J. Bouwsma, 'The Two Faces of Humanism: Stoicism and Augustinianism in Renaissance Thought', in H. Oberman and T. Brady (eds.), Itinerarium Italicum ( Leiden: Brill 1975), 3-60. 6 See A. Levi, French Moralists ( Oxford: Clarendon Press, 1964), 1.5-19. -212of the will had ramifications for his views on the intellect and sensation, freedom and determination, and the moral evaluation of purpose and action. 7 In particular, the ethical consideration of the passions now comes to be formulated in terms of the action of the will. The affections, which he treats as the soul's motions, cannot simply be referred to a criterion of rationality, as the Stoics had urged, but must be assessed in terms of the act of will from which they arise, and 'if the will is wrongly directed, the emotions will be wrong; if the will is right, the emotions will be not only blameless but praiseworthy'. 8 So, for example, a virtuous form of sadness or despair--tristitia--is ruled out on the Stoic view, but quite possible on Augustine's conception. 9 The fundamental character of the will, which guides its inclinations, is love, and virtuous affections are to be distinguished from vicious ones in terms of the moral quality of the love that rules in the will. Once the topic of the affections had been formulated in terms of the action of the will, a long exercise of classification and categorization of them began, the most systematic

account being that provided by Aquinas in Part II. i of the Summa theologiae. Drawing extensively on Aristotle Rhetoric, Aquinas' account was both comprehensive and ambitious. Like Augustine, he is opposed to the Stoic doctrine, although his criticisms of it are more systematic and penetrating, pointing out that the Stoics confuse sensitive and rational appetites, and hence the passions and the acts of will which alone are the seat of moral good and evil. 10 He begins by distinguishing 'passions of the body', such as physical suffering, which originate in the body and terminate in the soul, from 'passions of the soul', which originate in the soul and terminate in the body. The various passions of the soul generally recognized are then classified--following Plato's division of the soul into its rational, concupiscible, and irascible parts--into concupiscible affections (desires and appetites) and irascible affections (the passions strictly speaking), the former regarding good or evil absolutely and directly, the latter regarding them 'ratione ardui', that is, as something to be attained or avoided only with difficulty. The classification distinguishes between concupiscible affections directed towards a good object--love, desire, and pleasure or joy; concupiscible affections directed towards an evil object--hate, aversion, and pain or sorrow; irascible affections directed towards a good object--hope and despair; and irascible affections ____________________ 10 Aquinas, Summa theologiae, q. xxiv, arts. 1-2. 7 On the difference between Augustine's account of moral responsibility and that of his predecessors, see the exemplary account in Albrecht Dihle, The Theory of the Will in Classical Antiquity ( Berkeley: University of California Press, 1982). 8 Augustine, De Civitate Dei, XIV. ch. 6. 9 Ibid. ch. 8. -213directed towards an evil object--fear, courage, and anger. 11 But the point of the exercise goes beyond mere classification of affective states. Aquinas's aim is to provide a genealogy for the passions whereby they can all be derived from four: pleasure, pain, hope, and fear. Pleasure and pain are the termini of all the passions, in that all the passions result in one or other of these, and hope and fear are the termini of the movements of the appetite: to take Aquinas's example, love passes through desire to hope, hate through aversion to fear. 12 This classification and genealogy of the passions provided the model for subsequent discussion, and despite the fact that the two most influential later writers on the passions, Vives and Descartes, rejected it, it continued well into the eighteenth century. 13 Aquinas's account of the physiology of the passions reflects the traditional treatment, a treatment that was bequeathed to the seventeenth century. A passion arises when something good or evil is apprehended, exciting an appetite, which induces a bodily change. The passion is constituted by the dual movement of the appetite and the bod. 14 Among the causes of the passion are the bodily 'complexion', which predisposes the body in various ways, and the particular bodily accompaniments of the passion which characterize that passion. But this does not mean that the resulting state of mind is no

more than an effect of the bodily changes, a mere consciousness of them; for the bodily changes are merely an embodiment of the apprehension of good or evil seeking satisfaction. As Aquinas puts it, 'the affections of the soul are not caused by changes in the heart, but rather cause them . . . A man does not seek vengeance because the blood about the heart is inflamed; this is what disposes him to anger, but the anger itself comes from the appetite of vengeance.' 15 Gardiner sums up the doctrine well, noting that while Aquinas makes bodily changes essential to the constitution of a passion, the sensation of those changes is nowhere said to be essential to its psychical constitution. Aquinas's 'general representation is that of a process of apprehension and appetite on the one side resulting in expressive bodily movements on the other, the relation of the two being conceived in Aristotelian terms as one of "form" to "matter", so that the phenomenon may be described by either, but is best described, of course, by the determining factor, the form'. 16 ____________________ 11 Ibid., q. xxiii, art. 4. 12 Ibid., q. xxv, art. 4. 13 H. M. Gardineret al., Feeling and Emotion ( New York, 1937), 110. 14 Aquinas, Summa theologiae, q. xlv, art. 3. 15 Quoted in Gardineret al., Feeling and Emotion, 115. 16 Ibid. 116. -214The most significant break with the Thomist account of the passions comes in the work of Vives. 17 Vives' conception of the passions differs significantly from that of Aquinas, for he considers only the more violent emotions to be passions, and the mind does not so much register the experience as share in it. In his De anima et vita ( 1538), Vives abandons the distinction between concupiscible and irascible passions, and offers a different classification. But the key feature of Vives' account for our purposes is the way he makes the intellect and the will into autonomous faculties whose acts are mutually independent. As Levi has noted, 'behind his theory of the passions there is discernible a breakdown in the traditional scholastic psychology, a breakdown which was to be completed by the Neostoic moralists but which, as early as Vives, sets the stage for the debates of the moralists of the following century'. 18 Although the will should be guided by reason, which has as its object the rationally perceived good, the reason does not actually cause the will to choose in a particular way, for the will is essentially spontaneous in its liberty to choose. As Levi points out, what resulted was not only a blurring of the distinction between the passions and the virtues, but also a separation of the reason and the will 'such that it is difficult to see how an act can at the same time be both rational and free'. 19 The ultimate upshot of this was that the will and the reason were gradually prized further apart, so that in the Neostoic revival of the sixteenth and seventeenth centuries (of which we may take Montaigne and Lipsius as early representatives) 20 it began to be urged that the reason could be relied upon to the exclusion of the will. One might expect disputes pursued along clear Neostoic versus Augustinian/Thomist lines to follow from this, focusing on the question of whether the

will did or did not play a part, but this is not what happened. The issue was complicated by the fact that, although he regarded them as mutually dependent, Aquinas had defended a 'real' distinction between the intellect and the will as two faculties of the soul, one having the true as its object, the other having the good. There is a fundamental instability or unclarity on the question of the relation between the will and reason, and what one tends to find on the anti-Stoic side in the late sixteenth- and early seventeenth-century literature on the passions, whether scholastic, devotional, or secular, is a basic adherence to a generally Augustinian/Thomist ____________________ 17 The most comprehensive account of Vives' contribution is Carlos G. Noreña, Juan Luis Vives and the Emotions ( Carbondale, Ill.: University of Southern Illinois Press, 1989). 18 Levi, French Moralists, 26. 19 Ibid. 27. 20 On this revival see Léontine Zanta, La Renaissance du stoïcisme au XVIe siécle ( Paris: Champion, 1914). -215position tempered to a greater or lesser degree by elements drawn from Stoicism. The deep problems in the literature can be glimpsed by considering Eustache Summa philosophiae ( 1609). Although this was not as influential as the Tableau des passions ( 1630) of Coëffeteau, Descartes does not mention Coëffeteau, whereas he had originally intended to print Eustache's textbook alongside his own Principia when that work was at an early draft stage. What we find in both Coëffeteau and Eustache is a sharp distinction between the higher and lower parts of the soul. 21 Eustache holds that the will is related to the intellect in the higher part of the soul as the imagination is related to the sensitive appetite in the lower part. Aquinas had treated reason, sense, and imagination as part of a single cognitive process, but the Stoic tendency to use the term 'imagination' pejoratively, combined with a move to think of imagination as being concerned with knowledge of material objects and reason as being concerned with knowledge of immaterial objects (universals, etc.), had a significant impact on later scholastic thinking; the imagination became insulated from reason, and began to be treated as the source of error. This opens up the possibility of an act of will being at variance with the sensitive appetite and exercising no control over it. More generally, what results is an extremely unstable amalgam of elements taken from the scholastic tradition and Stoicism, and the price paid is a fragmentation of the soul which has no obvious benefits. This opened the door to an abandonment of the scholastic account in favour of Stoicism, notwithstanding the wellknown difficulties with the Stoic account; for, as Levi remarks, 'as in antiquity, the great asset of Stoic theory is that it restores unity to a fragmented soul'. 22 This, then, is the context within which Descartes was writing. Where the physiology of the passions was treated, there was an almost irresistible tendency to apply properly mental and corporeal attributes interchangeably, and even to conceive of animal spirits in

a vitalistic fashion; and where the faculties were treated, there was an increasing fragmentation of the soul or mind with various 'lower' parts blending into the corporeal functions. And this occurred with no discernible improvement in our understanding of the nature of human affective states. Descartes' aim in producing a systematic account of the passions was more than anything else to restructure the whole question of the affective states around a clear understanding of the distinction between mind and body, ____________________ 21 What follows is derived largely from the exemplary discussion in Levi, French Moralists, ch. 6. 22 Ibid. 155. -216and on the basis of such an understanding to formulate the appropriate notion of a substantial union needed to account for the source and nature of affective states. We have three sources for Descartes' account. The first is his correspondence, that with Elizabeth being the most important, but that with Chanut, from 1646 onwards, also being of significance. Chanut usually acted as an intermediary between Descartes and Queen Christina of Sweden, although there is one letter of importance to Christina herself. 23 The second is his Passions de l'âme, written during the winter of 1645-6, but published only a few weeks before his death. The third is the preface to Picot French translation of the Principia, the Principes de philosophie, which appeared in 1647. With the resumption of correspondence with Elizabeth in the middle of 1645, Descartes moves away from the somato-psychic account that he had favoured earlier to one in which psychosomatic factors are stressed. In the letters to Elizabeth, for example, he provided a psychosomatic treatment of melancholia, whereby the mind is forced to concentrate on agreeable things, this having the effect of relaxing the heart and allowing freer circulation of the blood. This may be a change in emphasis rather than a change in position, and he does not give up the idea that one's bodily disposition may aid or impede the will, pointing out that an indisposition in the body may prevent the will from acting freely. 24 Indeed, he makes it clear that what he has in mind is a reciprocal relation between mind and body: Bodily health and the presence of agreeable objects greatly assist the mind by chasing from it all the passions which partake of sadness and making way for those which partake of joy. And, conversely, when the mind is full of joy, this helps greatly to cause the body to enjoy better health and to make the objects present to it appear more agreeable. 25 In the Passions, which was evidently completed over the winter of 1645-6, 26 Descartes sets out to provide a comprehensive account of the various ways in which mind and body interact. The three parts of the text are designated by Descartes himself as providing a

general account of the mind-body relation and the general nature of passions (part i), a classification of the passions (part ii), and an account of particular passions (part iii), ____________________ 23 This correspondence is collected in English translation, with a very useful introductory essay, in John J. Blom, Descartes: His Moral Philosophy and Psychology ( Hassocks: Harvester, 1978). 24 Descartes to Elizabeth, 1 Sept. 1645 ( A T IV282 [ C S M K262-3]). 25 Descartes to Elizabeth, [ Nov. 1646] ( A T IV529 [ C S M K296]). 26 See Descartes to Chanut, 15 June 1646 ( A T IV442 [ C S M K289]). -217although from near the end of part ii the discussion shifts to the moral/ therapeutic questions surrounding the passions. The importance of beginning the Passions with an account of the general questions surrounding the mind-body relation is brought out rather well by a comment to Chanut in 1646, that the Principia does not in fact get one very far as regards morals. 27 One should remember that the projected fifth and sixth parts of the Principia were supposed to be 'on living beings and on man', respectively. Descartes now realizes that there is a big gap between the metaphysically grounded natural philosophy that he had set out in the Principia and the kind of things that one needs to establish to provide an account of human behaviour. Certainly the idea of a medically based ethics, which would have connected the natural philosophy of parts ii, iii, and iv with the account of human beings in the projected part vi, via a consideration of physiology in the projected part v, no longer looks viable, and in the 1647 preface to the Principes, morals is listed along with mechanics and medicine as one of the three fundamental sciences. 28 The foundations of morals lie no longer in medicine (if they ever did), but rather in an account of the substantial union of mind and body. A clue to the orientation of Descartes' whole approach is given in his statement at the beginning of the Passions that he writes not as 'an orator, nor as a moral philosopher, but as a physicist [physicien]'. 29 This means two things. First, his treatment of the passions is dependent upon the metaphysical foundations of natural philosophy set out in the Principia, and in this respect, the passions have the same ultimate foundation as the two other basic sciences, mechanics and medicine. It is an egregious error to imagine that Descartes is maintaining that one could actually discover substantive truths by deriving them from first principles, and this holds for the passions just as much as for natural philosophy generally. The first principles provide a framework within which the exposition of doctrine must proceed if any systematic certainty is to attach to one's conclusions. So, the first point is that, writing as a 'physicien', Descartes is providing an account of the passions which aspires to some degree of certainty, in contrast to rival accounts. Secondly, as Rodis-Lewis points out, in saying he does not write as a moral philosopher, Descartes is setting himself against 'those for whom this is the point of

departure: the Stoics saw in pathos, passion, a pathological phenomenon, which the sage was required ____________________ 27 Ibid. A T IV441 [ C S M K441]. 28 A T IXB 14, 17 [ C S M186,188]. 29 Passions, reply to second letter, A T XI326 [ C S M 1327]. I am especially indebted in my account of the Passions to the introduction and notes in Geneviève Rodis-Lewis (ed.), Descartes: Les Passions de 1',Inze ( Paris 1970), and to Levi, French Moralists, chs. 9 and 10. -218to quash in aspiring to apatheia'. 30 It is not just that Descartes disagrees with this account of the passions, but with this whole approach to them. Unless one has a proper understanding of the faculties of the mind, bodily physiology, and how the substantial union of mind and body functions, one cannot even begin to investigate the nature of the passions. Even if such an understanding yielded the conclusion that the passions are something that must be overcome, the Stoics would not be vindicated, because they have approached the question in a completely wrong-headed fashion. It is tempting to draw an analogy with Descartes' rejection of the Aristotelian account of perception here: Aristotle simply assumed that our perceptual image must resemble the object perceived, and built his account of the transmission of light and the physiology of perception around this assumption, instead of investigating the optics and physiology first. Similarly, the Stoics can be accused of attempting to provide a moral theory of the passions without any investigation of their psychological, and especially their physiological, basis. Descartes begins part i of the Passions by noting that whether something is called an action or a passion depends simply on whether it is considered with respect to the mind or the body, so the crucial thing is to start with the difference between the soul and the body. 31 As we have seen, contemporary accounts of the passions were very vague on the key question of what the relation between the mind and the body is, and in this context Descartes places himself on firm ground by arguing that any serious discussion of the passions must begin with this question: and, of course, on home ground, for his account of both physiology and the functions of the mind were far more elaborate than anything available in the contemporary literature. Above all, he can avoid the common error of thinking of the action of the mind on the body as being like the action of one body on another. 32 Articles 7 to 16 set out in a summary way the mechanistic physiology of L'Homme, and the Passions contain the only exposition of his physiology published in his lifetime. We are then provided with a division of the functions of the soul into two: actions and passions. Actions comprise volitions which either terminate in the soul, as 'when we will to love God', or in the body, as when we move our legs by willing to walk. They also include those perceptions which have their origin in the soul, as when we reflect upon our own existence. Perceptions which have their origin in the body, on the other hand, are passions. The treatment of the passions then proceeds, from article 21 onwards, in terms of functions of

____________________ 30 In her introduction to Voss (ed.), René Descartes: The Passions of the Soul, pp. xvixvii. 31 Passions, art. 2 ( A T XI328 [ C S M I328]). 32 See Descartes to Elizabeth, 21 May 1643 ( A T III666 [ C S M K218]). -219the soul which depend on its union with the body. Perceptions which do not derive from the soul itself can be caused either by external bodies acting upon us, or from natural appetites of the body, such as hunger, which we sense through bodily organs, or they can be felt 'as in the soul itself', in which case no immediate cause is evident. 33 These last are the 'passions of the soul' to which Descartes' account is devoted, and he is concerned with their phenomenology rather than their causes; for while we may be deceived about their causes--they may be experienced whether we are awake or asleep, for example--we cannot be deceived about their existence or specific nature. 34 They are defined as being 'caused, maintained, and strengthened by a movement of the spirits', and take the form of 'excitations of the soul', as do volitions; but, unlike volitions, they do not have their source in the soul. Articles 30 to 47 then provide a psychophysiology of the soul in terms of Descartes' doctrine that the pineal gland is the seat of the soul, much along the lines of L'Homme. Two features of this account are worth noting. First, in article 36 he explains how different passions can arise in different people who are apparently stimulated in the same way. Clearly, any treatment of the passions is going to have to account for what we might call differences in temperament. But the explanatory value of his account is minimal, to say the least; for he simply translates differences in response into differences in the disposition of the brain. The idea is that the spirits are reflected differently in different people depending on the initial disposition of their brain, and hence pineal gland, something which results in different responses. 35 Here, the poverty of Descartes' tennis-racquet account of the action of the pineal gland, whereby it acts simply by redirecting animal spirits at different angles depending on its disposition, is clear. Second, in article 47 he uses the doctrine of the pineal gland being the seat of the soul to undermine the prevalent account of the passions in terms of a conflict between higher and lower parts of the soul: All the struggles that people customarily imagine between the lower part of the soul, which is called sensitive, and the higher or 'rational' part, or between the natural appetites and the will, just consist in the opposition between the movements which the body, by its spirits, and the soul, by its will, tend to excite simultaneously in the gland. For there is only a single soul in us, and this soul has within itself no diversity of parts; one and the same soul is sensitive and rational, and all its appetites are volitions. 36 ____________________ 33 Passions, art. 25 ( A T XI347 [ C S M I337]). 34 Passions, art. 26 ( A T XI348-9 [ C S M I338]).

35

Passions, art. 36 ( A T XI357-8 [ C S M1342]). Passions, art. 47 ( A T XI364 [ C S M I343-4]).

36

-220The conflicts that we experience are, then, conflicts between the soul and the body, for there is no sense in which they can be either conflicts between higher and lower parts of the soul, or conflicts between different powers of the soul. The spirits can move the pineal gland in a particular way, stimulating a desire for something, and while the will cannot halt this directly, it can represent objects to itself so vividly that, by the principle of association, the course of the spirits will gradually be halted. In these circumstances the soul will be impelled 'almost simultaneously' both to desire and not to desire the same thing; but the 'almost' is important, and this will not be a genuine conflict in the soul. A parallel case occurs in the body, when the passions cause the organs or limbs to act in a certain way, and the soul attempts to stop this, causing conflict in the body. What is required in both cases is mastery of one's passions, which derive from 'firm and determined judgements'. 37 There is no question of using one passion to offset another, and even less of trying to live without passions: they are crucial for fortifying and sustaining individual acts of will, and those who have no inclination for the passion of wonder, for example, are 'usually very ignorant'. 38 Part ii of the Passions deals with their classification. The basis for this is different from that of a writer like Aquinas, who attempts to provide a systematic genealogy in terms of an account of primitive passions from which others can be derived. Moreover, rejecting any distinction between parts of the soul, Descartes rejects the distinction between concupiscible and irascible appetites which depends on this, thereby removing the basis of the traditional classifications. 39 Although he lists six primitive passions-wonder, love, hate, desire, joy, and sadness--and although, except for the first, these are part of the standard scholastic listing, Descartes takes a thoroughly functional approach, for the basis for identifying passions is in terms of the importance their perceivable objects have for us, how much difficulty we have in obtaining them, whether they are harmful or of benefit to us, and so on. Moreover, the full listing--as opposed to the six primitive passions--is explicitly open-ended, for there are, Descartes tells us, an indefinite number of passions. In the course of his discussion of joy and sadness in the Passions, Descartes introduces an important distinction between joy and intellectual joy, and sadness and intellectual sadness. Intellectual joy and sadness are not passions properly speaking, for they 'come into the soul by the action of the soul itself', 40 and not by the action of the body. In the Principia, ____________________ 37 Passions, art. 48 ( A T XI367 [ C S M I347]). 38 Passions, art. 75 ( A T XI384 [ C S M I354-5]). 39 Passions, art. 68 ( A T XI379 [ C S M I352]). 40 Passions, art. 91 ( A T XI397 [ C S M I360-1]).

-221Descartes had mentioned that, when we hear good news, 'it is first of all the mind that makes a judgement about it and rejoices with that intellectual joy which occurs without any bodily disturbance and which, for that reason, the Stoics allowed the wise man to experience'. 41 Such 'inner excitations', 42 as Descartes calls them, come to play an increasingly important role as the Passions progresses, and at the end of part ii he introduces a fuller account of their nature: Here I shall merely add one further consideration which, I believe, helps to prevent us from suffering any discomfort from the passions. This is that our well-being depends principally upon inner excitations which are excited in the soul only by the soul itself. In this respect they differ from the passions of the soul, which invariably depend on some movement of the spirits. Although these excitations of the soul are often joined with those passions which are similar to them, they may also frequently be found with others, and they may even originate in those to which they are opposed. For example, when a husband mourns his dead wife, it sometimes happens that he would be sorry to see her brought to life again. It may be that his heart is constricted by the sadness excited in him by the funeral display and by the absence of a person to whose company he was accustomed. And it may be that some remnants of love or pity occur in his imagination and draw genuine tears from his eyes, in spite of the fact that he feels at the same time a secret joy in his innermost soul, and the excitation of this joy has such a power that the concomitant sadness and tears can do nothing to diminish its force. Again, when we read of strange adventures in a book, or see them represented on stage, this sometimes arouses sadness in us, sometimes joy, or love, or hatred, and generally any of the passions, depending on which objects are presented to our imagination. But along with this we have pleasure of feeling them aroused in us, and this pleasure is an intellectual joy which can just as easily originate in sadness as in any of the other passions. 43 Here Descartes begins to show some influence of neo-Stoicism, for he goes on to tell us that such inner excitations affect us more intimately than the passions, and hence, so long as our soul always has the means of happiness within itself, anything external is powerless to harm it. 44 At this point, considerations of the physiology of the passions recede into the background, as ethical questions are construed in terms of the true worth of things, which the passions exaggerate in one direction or the other. The traditional contrast between reason and passion now appears as a contrast between an inner excitation and a passion. So, for example, in a letter to ____________________ 41 Principles, Part iv, art. 190 ( A T VIII317 [ C S M I280-1]). 42 I follow Voss in translating ~motions intérieures as inner 'excitations' rather than inner 'emotions': see Voss (ed.), Descartes, Passions, 138.

43

Passions, art. 147, ( A T XI440-1 [ C S M 1381]). The first example suggests to me a low view of women on Descartes' part: I cannot imagine him putting the example the other way round, so that it is the wife who feels a secret joy at her husband's death. 44 Passions, art. 148 ( A T XI441-2 [ C S M I381-2]). -222Chanut of 1 February 1647, Descartes distinguishes between 'the love which is purely intellectual or rational and the love which is a passion'. 45 The former consists in the movement of the will that accompanies the knowledge that a good is possessed, the latter in the experience of possessing the good. Intellectual love is clear, love as a passion confused. 46 Our judgement of the worthiness of the object determines the degree of our intellectual emotion, whereas the extent of the passion depends on our degree of esteem for the object. Since the second must be proportional to the first, it is the intellectual or rational version of the emotion, the 'inner excitation', that provides the key to how we should behave. The direction and regulation of the passions enables us to live a good life: 'good' both in the sense of being ethical and in the sense of being fulfilling, for the two are inseparable in Descartes' account. Consider, for example, his discussion of generosity, probably the most important concept in Descartes' ethics. Although the word 'generosité' had much the same core meaning in ordinary seventeenth-century French as it has for us, the French moralists had added a connotation of nobility, and 'generosité' is very close to the notion of 'gloire' defended in the personalist ethics of Balzac, Corneille, and others, in part as an attempt to elaborate a true gentilhomme morality. Descartes defines generosity as follows: I believe that true generosity, which causes a man to esteem himself to the greatest degree which is legitimate, consists solely in this: partly in his understanding that there is nothing which truly belongs to him except his free control of his volitions, and the only grounds for praise or blame are that he uses it well or badly; and partly in his feeling within himself a firm and constant resolution to use it well--that is, never to lack the will to undertake and carry out whatever he judges to be the best. To do this is to pursue virtue perfectly. 47 In other words, it is ultimately our degree of self-esteem which will determine the worth of what we do. Hence the importance of the therapeutic aspects of Descartes' moral programme, such as those set out in the letter to Elizabeth in the middle of 1645, which recommends exercises for ridding oneself of melancholia--the affliction of those drawn to intellectual reflection--by 'directing the imagination'. 48 To put the matter in more modern terms, unless one is properly 'sorted out', questions of moral responsibility (or at least some questions of moral responsibility) do not arise, but 'sorting oneself out' is not something independent of morals: it is a pre-condition for morally responsible behaviour, and for Descartes and

____________________ 45 A T IV601 [ C S M K306]. 46 A T IV603-4 [ C S M K307]. 47 Passions, art. 153 ( A T XI445-6 [ C S M1384]). 48 Descartes to Elizabeth, May or June 1645 ( A T IV219-20 [ C S M K250]). -223any of his contemporaries this meant that it was an integral part of moral philosophy. Such an approach to questions of morality captures a psychological dimension of morality which later accounts ignore, and it has a good deal to commend it. -224X DESCARTES' TREATMENT OF ANIMALS JOHN COTTINGHAM I To be able to believe that a dog with a broken paw is not really in pain when it whimpers is a quite extraordinary achievement, even for a philosopher. Yet, according to the standard interpretation, this is just what Descartes did believe. He held, we are informed, the 'monstrous' thesis that 'animals are without feeling or awareness of any kind'. 1 The standard view has been reiterated in a recent collection on animal rights, which casts Descartes as the villain of the piece for his alleged view that animals merely behave 'as if they feel pain when they are, say, kicked or stabbed'. 2 The basis for this widely accepted interpretation is Descartes' famous doctrine of the 'animal-machine' (bête-machine), a doctrine that one critic condemns as 'a grim foretaste of a mechanically minded age' which 'brutally violates the old kindly fellowship of living things'. 3 But if we look at what Descartes actually says about animals, it is by no means clear that he holds the monstrous view which all the commentators attribute to him. In fact, the traditional rubric ' Descartes's doctrine of the "bête-machine"' is vague and ambiguous; it needs to be broken down into a number of distinct propositions if we are to sort out what Descartes said, and what he is implicitly committed to, from what he neither said nor implied.Consider, then, the following assertions: 1. Animals are machines. 2. Animals are automata. ____________________ John Cottingham, "'Descartes' Treatment of Animals'" originally published as "'A Brute to the brutes? Descartes' Treatment of Animals'" in Philosophy, Vol. 53 ( 1978), 551-9, by permission of the author and the publisher, Cambridge University Press. 1 N. Kemp Smith, New Studies in the Philosophy of Descartes ( London: Macmillan, 1952), 136 and 140. 2 T. Regan and P. Singer (eds.), Animal Rights and Human Obligations (Englewood Cliffs, NJ: Prentice-Hall, 1976), 4. 3 A. Boyce Gibson, The Philosophy of Descartes ( London: Methuen, 1932), 214.

-2253. 4. 5. 6. 7.

Animals do not think. Animals have no language. Animals have no self-consciousness. Animals have no consciousness. Animals are totally without feeling.

Proposition (7) is the 'monstrous thesis' with which Descartes is so often credited. I shall argue that Descartes held theses (1) to (5), but that there is no evidence that he held (7), and even some positive evidence that he regarded (7) as false; however, fuzziness about (6) and its distinction from (5) (together with certain general features of his metaphysics) laid him open to being interpreted as committed to (7). II Thesis (1) is not explicitly asserted by Descartes in this form, but he commits himself to it in so many words in the famous passage on animals in part v of the Discourse, where he says the body may be regarded 'comme une machine qui, ayant été faite des mains de Dieu, est incomparablement mieux ordonnée . . . qu'aucune de celles . . . inventées par les hommes'. 4 Thesis (1) in fact forms part of Descartes' general scientific 'mechanism', and, roughly translated, means that all animal behaviour is subsumable under physiological laws, which, for Descartes, are ultimately derivable from mathematical principles. Essentially, when Descartes says that 'all the motions of animals originate from the corporeal and mechanical principle', 5 he is concerned to promulgate a scientific animal physiology which seeks explanation in terms of efficient, rather than final, causes. 6 Now from none of all this does it follow that when Descartes calls something a 'mechanism' or 'machina', he is automatically ruling out the presence of sensations or feelings; Boyce Gibson's claim that Descartes 'uses the term [mechanism] explicitly to exclude . . . feeling' is not supported by any evidence. 7 In fact, it is important to notice that the human body is, for Descartes, a machine in exactly the same sense as the animal body: 'God made our body like a machine, and he wanted it to function like a universal instrument, which would always operate in the same way in accordance ____________________ 4 A T VI56: C S M I139. 5 Letter to More, 5 Feb. 1649 ( A T V276: C S M K365). 6 Principles, part i, art. 28 ( A T VIII15: C S M1202). See further A T V158 and Cottingham (ed.), Descartes' Conversation with Burman, 85 f. [ C S M K341]. 7 Nor does Gibson cite any: Philosophy of Descartes, 211. -226with its own laws.' 8 The phrase 'bête-machine' can thus be rather misleading, since the mechanical physiology Descartes has in mind operates equally in the case of Homo sapiens. Of course it is true that in the human, but not the animal, case there is the extra

dimension of a 'soul' (I shall come back to this); but this is a separate point. To deny that X has a soul is a separate claim from the claim that Xs movements can be explained by mechanical principles, and is not strictly entailed by it. Proposition (2) is implied frequently by Descartes, and is stated explicitly in a letter to More of 5 February 1649: 'it seems reasonable since art copies nature, and men can make various automata which move without thought, that nature should produce its own automata much more splendid than the artificial ones. These natural automata are the animals.' 9 It is Descartes' use of the term 'automaton' more than any other that has led critics to convict him of holding the monstrous thesis (thus, Kemp Smith speaks of the Cartesian view that animals are 'mere automata . . . incapable of experiencing the feelings of wellbeing or the reverse, hunger or thirst . . ." 10 ). But the inference from 'X is an automaton' to 'X is incapable of feeling' is a mistaken one. Webster's dictionary gives the primary meaning of 'automaton' as simply 'a machine that is relatively selfoperating'; and neither this nor the subsidiary meaning ('creature who acts in a mechanical fashion') automatically implies the absence of feeling. 11 Even today, then, to regard total insensibility as part of the meaning of 'automaton' would seem to be an error; and this seems to have been even more true in the seventeenth century, where 'automaton' probably carried no more than its strict Greek meaning of self-moving thing'. Thus Leibniz, defending his claim that we possess 'freedom of spontaneity', speaks of the human soul as 'a kind of spiritual automaton', meaning no more than that its actiongenerating impulses arise solely ab interno, and produce their effects without the intervention of any external cause. 12 What fascinated Descartes' generation about machines ranging from clocks to the elaborately contrived moving statues to be found in some of the royal fountains was simply this: the complex sequences of movements which to primitive ____________________ 10 Kemp Smith, New Studies, 135. 11 Webster's Seventh New Collegiate Dictionary ( Springfield, Mass.: Merriam, 1963). 12 Leibniz, Theodicy, part i. art. 52. 8 A T V163-4; cf. Cottingham (ed.), Conversation with Burman, 29 [ C S M K346]. 9 'Deinde quia rationi consentaneum videtur, cum ars sit naturae imitatrix, possintque homines varia fabricare automata, in quibus sine ulla cogitatione est motus, ut natura etiam sua automata, sed artefactis longe praestantiora, nempe bruta omnia, producat' ( A T V277: C S M K 366). This is a development of material found in Discourse, part v. -227(or medieval) man might have appeared as certain proof of some kind of inner motive 'force' or 'spirit', could all be explained quite simply by reference to internal mechanical structure--cogs, levers, and the like ( Descartes mentions as an example a statue of Neptune which would threaten with his trident the approaching onlooker who had

unwittingly stepped on a button). 13 The point Descartes is concerned to make over and over again about the behaviour of 'natural automata' like dogs and monkeys is that the mere complexity of their movements is no more a bar to explanation in terms of inner mechanical structure than is the case with the responses of the trident-brandishing 'Neptune'. 14 III So far, then, I maintain that Descartes' characterization of animals as 'machines' and 'automata' is of itself quite insufficient to allow us to conclude that he thinks that animals lack feelings. When we get on to the remaining propositions in our list, things are not so simple. It is, Descartes asserts, in principle possible to mistake a cleverly contrived artificial automaton for an animal. But we could never mistake an automaton, however ingenious, for a man. Why not? Because, says Descartes, an automaton could never talk: it could 'never arrange its speech in various ways in order to reply appropriately to everything that could be said in its presence'. 15 This for Descartes indicates the crucial difference between animals and man-they do not think. Animals do not penser or cogitate; they are not endowed with a mind (mens, esprit); they lack reason (raison); they do not have a rational soul (,me raisonnable). 16 Descartes is thus explicitly committed to thesis (3), and holds, moreover, that it is entailed by (or at least strongly evidenced by 17 ) thesis (4). Descartes was of course aware that parrots can be made to 'talk', and that dogs make noises which might be analogous to speech; but he has strong and, since Chomsky's updating of them, widely admired arguments against ____________________ 13 Traité de I'homme ( A T XI130-2: C S M1100-1). Cf. Gilson, Rena Descartes, Discours de la méthode, 420 ff. 14 Descartes compares the plants in this connection, 'que [la nature] remplit d'une infinité de petits conduits imperceptibles à la vue': letter to Reneri of Apr. 1638 ( A T1140: C S M K99-100). 15 Discourse ( A T VI57: C S M1140). 16 Ibid. Cf. letter to More, 5 Feb. 1649 : 'loquela unicum est cogitationis in corpore latentis signum certum' ( A T V278: C S M K366). 17 Descartes at one point observes that 'quamvis . . . pro demonstrato habeam, probari non posse aliquam esse in brutis cogitationem, non ideo puto posse demonstrari nullam esse, quia mens humana illorum corda non pervadit' ( A T V276-7: C S M K365). -228construing such utterances as genuine speech. The talking of parrots is dismissed because it is not 'relevant to the topic'; 18 but the most important point Descartes has to make is

that the utterances of dogs, cats, etc. are never, to use the Chomskian phrase, 'stimulusfree'; they are always, says Descartes, geared to and elicited by a particular 'natural impulse'. 19 I shall come back to these arguments, but first an obvious objection must be faced. In admitting that Descartes held thesis (3) (that animals do not think), have I not thereby conceded that he must have held the 'monstrous thesis' (7) (that animals do not feel)? For does not Descartes' special sense of 'think' (cogitare, penser) include feelings and sensations? Well, it is certainly true that Descartes deliberately extended the normal use of cogitatio or pensée. In answer to a misunderstanding of Mersenne (that if man was purely 'res cogitans' he must lack will), Descartes stated that willing was a façon de penser; he further explains that 'la pensée' includes 'non seulement les méditations et les volontés' but 'toutes les opérations de l'âme'. 20 This is generally taken to include sensations and feelings-indeed, seeing and hearing are explicitly included by Descartes in the list of 'opérations de l'âme' just mentioned. Further analysis, however, makes it clear that the matter is not as straightforward as this, and that translators who render cogitatio or pensée as simply 'experience' are moving much too swiftly. 21 When discussing whether 'video ergo sum' might not do as well as 'cogito ergo sum', Descartes says that 'I see' is ambiguous. If understood 'de visione' it is not a good premiss for inferring one's existence, but if understood 'concerning the actual sense or awareness of seeing' (de ipso sensu sive conscientia videndi), it is quite certain, since it is in this case referred to the mind which alone feels or thinks it sees (quae sola sentit sive cogitat se videre). 22 From this we can see that it is misleading to say, tout court, that cogitatio 'includes' sensations and feelings. The only sense in which a sensation like seeing is a true cogitatio is the sense in which it may involve the reflective mental awareness which Descartes calls conscientia--the self-conscious apprehension of the mind that it is aware of seeing. 23 ____________________ 18 To Newcastle, 23 Nov. 1646 ( A T IV574: C S M K303). 19 To More, 5 Feb. 1649; to Newcastle, 23 Nov. 1646; and N. Chomsky, Language and Mind ( New York: Harcourt, Brace & World, 1968), ch. .I 20 To Mersenne, May 1637 ( A T I366: C S M K56); and to Reneri, Apr. 1638 ( A T II36: C S M K97). Cf. Kenny, Descartes, 68 ff. 21 Cf. E. Anscombe and P. Geach, Descartes' Philosophical Writings ( London: Nelson, 1969), pp.xlviif., and my "'Descartes on Thought'", Philosophical Quarterly ( July 1978). 22 Principles, part i, art. 9 ( A T VIII7-8: C S M I195). 23 Conscientia is defined in the Conversation with Burman: 'conscium esse est . . . cogitare et reflectere supra suam cogitationem' ( A T V149: Cottingham (ed.), Conversation with Burman, 7 and 61 [ C S M K335]). -229-

The upshot is that Descartes' assertion of proposition (3) (that animals do not think) need not commit him to denying any feeling or sensation to animals--for example, a level of feeling or sensation that falls short of reflective mental awareness. Notice, moreover, how the language argument fits into all this. In pointing out that animals have no genuine language, Descartes clearly thinks that he has a powerful case for concluding that they do not think. Yet, for Descartes to regard this argument ('non loquitur ergo non cogitat' 24 ) as having such evident force, 'think' (cogitat) here must evidently be used in the fairly restrictive sense described above. If Descartes were using cogitat in the alleged very wide sense, he would be offering us an argument of the form 'non loquitur ergo non sentit' (he does not speak, therefore does not feel). It is inconceivable that Descartes could have proudly produced this argument to his correspondents as selfevidently clinching. IV Our conclusion so far is that neither in calling animals machines or automata, nor in denying they have thought or language, does Descartes commit himself to the monstrous thesis that they have no feelings or sensations. It is now time to look at some positive evidence that he actually regarded the monstrous thesis as false. The strongest evidence, which those who credit Descartes with the monstrous thesis seem strangely blind to, comes from the famous letters already cited, where Descartes denies speech to the animals. Writing to More, Descartes says that the sounds made by horses, dogs, etc. are not genuine language, but are ways of 'communicating to us . . . their natural impulses of anger, fear, hunger and so on'. 25 Similarly, Descartes wrote to Newcastle that: If you teach a magpie to say good-day to its mistress when it sees her coming, all you can possibly have done is to make the emitting of this word the expression of one of its feelings. For instance it will be an expression of hope of eating, if you have habitually given it a titbit when it says the word. Similarly, all the things which dogs, horses and monkeys are made to do are merely expressions of their ____________________ 24 'He does not speak, therefore he does not think.' Strictly, the argument must be of the form 'he does not speak and has no capacity for language acquisi tion, therefore he does not think'; for Descartes says that infants think ( A T VII246: C S M II171)-though only after a fashion ( A T V149-50: C S M K355). 25 'impetus suos naturales, ut iras, metus, famem, et similia . . . significent' ( A T V278: C S M K 366). -230-

fear, their hope, or their joy; and consequently, they can do these things without any thought. 26 'Impulses of anger, fear, hunger', 'expression of one of its feelings', 'expressions of their fear, their hope or their joy'. These are quite extraordinary phrases to use for a man who is supposed to believe animals are 'without feeling or awareness of any kind'. Is it possible that Descartes is here speaking loosely or metaphorically? This seems strange in letters which are explicitly and painstakingly devoted to clarifying the Cartesian position on animals. If this were not enough, in the letter to More, Descartes specifically separates cogitatio (thought) from sensus (sensation), and states that he denies the former, but not the latter, to animals: 'I should like to stress that I am talking of thought, not of . . . sensation; for . . . I deny sensation to no animal, in so far as it depends on a bodily organ.' 27

V The last quotation might make a pleasing and neat vindication of Descartes' kindly fellowship with the beasts: he denied that animals think, but not that they feel. But philosophy is seldom as tidy as this, and we must conclude by discussing a major difficulty which has been put off until now. The difficulty, in a nutshell, is that the monstrous thesis fits in with, and the pleasing vindication clashes with, Descartes' dualism. If substance is divided exclusively and exhaustively into res cogitans and res extensa, what room is there for animal sensations? Since an animal is not a res cogitans, has no mind or soul, it follows that it must belong wholly in the extended divisible world of jostling Cartesian shapes. And this means that what we call (and évidemment, Descartes himself called) 'animal hunger' cannot be anything more than a set of internal muscle contractions leading to the jerking of certain limbs, or whatever. This then must be the authentic Cartesian position--a position summed up when Descartes ____________________ 26 Si on apprend à une pie à dire bonjour à sa maîtresse, lorsque'elle la voit arriver, ce ne peut être qu'en faisant que la prolation de cette parole devienne le mouvement de quelqu'une de ses passions; à savoir, ce sera un mouvement de l'espérance qu'elle a de manger, si l'on a toujours accoutumé de lui doner quelque friandise lorsqu'elle l'a dit; ainsi toutes les choses qu'on fait faire aux chiens, aux chevaux et aux singes ne sont que des mouvements de leur crainte, de leur espérance ou de leur joie, en sorte qu'ils les peuvent faire sans aucune pensée' ( A T IV574: C S M K303). 27 'Velim...notari me loqui de cogitatione, non de vita vel sensu; vitam enim nulli animali denego, utpote quam in solo cordis calore consistere statou; nec denego etiam sensum quatenus ab organo corporeo dependet' ( A T V278: C S M K366). -231-

quotes with approval the passage in Deuteronomy which says that the soul of animals is simply their blood; or when he says that animal life is no more than 'the heat of the heart'. 28

No doubt this is where a pure Cartesian, a consistent Cartesian, would stop. But we have seen that Descartes, dualist or no, undoubtedly and explicitly attributes such feelings as anger, hope, and joy to animals. I think the only explanation of this is that Descartes, either inadvertently or wilfully, failed to eradicate a certain fuzziness from his thinking about consciousness and self-consciousness. To say that X is in pain (angry, joyful) is certainly to attribute a conscious state to X, but this need not amount to a full-blooded reflective awareness of pain that is involved in the term cogitatio. To be dogmatic for a moment, I should certainly say that cats feel pain, but not that they have the kind of full mental awareness of pain that is needed for it to count as a cogitatio (i.e. the sort needed to support the premiss of a cogito-type argument 'patior ergo sum'--'I am in pain, therefore I am'). Descartes is certainly committed to thesis (5)--that animals do not have self-consciousness; but when, as a result, he consigns animals to the realm of res extensa, he simply does not seem to bother that terms like pain, anger, etc., which he uses of animals, clearly imply some degree of conscious (though perhaps not 'self-conscious') awareness. VI It is important to notice, in conclusion, that this strange fuzziness is not simply the result of a blind spot which Descartes had when dealing with animals, but connects with a fundamental and unresolved difficulty in Cartesian metaphysics. There is a fascinating chapter in part iv of the Principles dealing with human sensations (sensus) and feelings (affectus). When we hear a piece of good news, says Descartes, we feel 'spiritual joy' (this is the sort of pura cogitatio that, presumably, God and the angels experience). But when the news is grasped by the imagination, the 'animal spirits' flow from the brain to the heart muscles, which in turn transmit more 'movements' to the brain, with the result that we experience a feeling of 'laetitia animalis'. 29 It is evident that Descartes is in a philosophical mess here. One might expect that joy would be regarded as a purely mental state, and thus confined firmly to the realm of res cogitans. But here is ____________________ 28 To Plempius, 3 Oct. 1637 ( A T1415: C S M K62); and to More (see n. 5). 29 Principles, part iv, art. 190 ( A T VIII317: C S M I280-1). -232Descartes distinguishing between the pure intellectual apprehension of joyful news, on the one hand, and, on the other, a feeling of joy. This latter is the bizarre entity called 'animal joy', which is somehow bound up with heart muscles and brain commotions. The choice of the phrase 'laetitia animalis' here is no accident. Descartes clearly wants to say that the joy of dogs and cats is analysable into just such physiological events. But what he seems to forget is that, as a strict dualist, he should not be using the word 'laetitia' at all in

this case. For a true dualist, if something is laetitia (an inescapably 'mental' predicate) it cannot be animalis (part of res extensa); and conversely, if it is animalis, it cannot be laetitia. The truth, perhaps, is that Descartes was never completely comfortable with strict dualism, however emphatically he affirmed it. As the contortions in the Sixth Meditation show, feelings, or sensations (like those of hunger or thirst), are an insoluble worry for him. We do not merely 'notice' that we are in pain, as a pilot observes that his ship is damaged, we actually feel it; and this shows that there is a 'conjunctio et quasi permixtio' 30 between mind and body--a mysterious 'intermingling' of what are, remember, logically distinct and incompatible substances. This 'substantial union' is the uncuttable knot in the centre of Cartesian metaphysics. Descartes once wrote to a correspondent that if an angel (a pure res cogitans) were in a human body, he would not feel like us; he would merely observe the changes in his nervous system. This shows, Descartes observed, that feelings like that of pain are not the purae cogitationes of a mind distinct from body, but rather are the 'confused perceptions which result from a real union with the body'. 31 Feelings, in other words, are an inexplicable result of the animal side of our nature, our mysterious intermingling with res extensa. If this is what Descartes says about human feelings, it is not surprising that he never got animal feelings properly sorted out. Strict dualism makes nonsense of Descartes' common-sense attribution of feelings like hunger to the animals; but then Descartes is unable to extract from dualism any clear account of the awkwardly undeniable experience of human hunger. At the end of the day, Descartes may not have been completely consistent, but at least he was not altogether beastly to the beasts. 32 ____________________ 30 A T VII81: C S M1156. 31 'sensus doloris, aliosque omnes, non esse puras cogitationes mentis a corpore distinctae, sed confusas illius realiter unitae perceptiones' ( to Regius, Jan. 1642 ( A T III493: C S M K206)). 32 I am indebted to Professor A. G. N. Flew, whose questions about Descartes' position stimulated me to pursue this line of inquiry. -233XI DESCARTES' METHOD AND THE ROLE OF EXPERIMENT DANIEL GARBER It is generally recognized that knowledge for Descartes is the clear and distinct perception of propositions by the intellect; knowledge in the strictest sense is certain, indeed indubitable, and grounded in the purely rational apprehension of truth. But it is also generally recognized that Descartes was a serious experimenter, at least in his biology and his optics, and that in these areas, at least, he seemed to hold that knowledge requires an appeal to experience and experiment. Writing, for example, in Part VI of the Discourse on Method, Descartes laments the fact that he has neither the time nor the

resources to perform all of the experiments (expériences) necessary to complete his system, and calls upon his readers to 'communicate to me those that they have already made, and to help me in performing those which remain to be done' ( A T VI65 [ C S M I144]). (One can see in the Discourse a clear anticipation of an important later literary form, the grant application.) To the twentieth-century philosopher this looks a bit puzzling: how can Descartes be both a rationalist, who sees knowledge as deriving from the intellect, and an experimentalist, who sees experiment and observation as essential to the enterprise of knowledge? This is the puzzle I would like to address in this essay. I shall argue that not only is there no contradiction here, but that the appeal to experience is an essential part of the method for constructing a deductive science. We shall begin with a brief account of Descartes' procedure for constructing his science, his method. While Descartes' method is discussed at great length in any number of books and papers, there is hardly a clear account of what it is in practice in any of the literature. Once we have a clear picture of what Descartes' method is, ____________________ Daniel Garber, ' Descartes and Experiment in The Discourse and Essay' published here under the title ' Descartes' Method and the Role of Experiment' from Stephan Voss (ed.): Essays on the Philosophy and Science of Rena Descartes ( 1993), 288-310. Copyright © 1993 by Oxford University Press, by permission of Oxford University Press. -234and the precise deductive structure of the body of knowledge that he is building, we can then turn directly to the question of experiment, and see how it fits into the programme. METHOD I hold the view that Descartes, in an important sense, gave up his famous method sometime in the late 1630s or early 1640s, and so I do not want to identify the question of Descartes' scientific procedure with that of his method. 1 But to understand Descartes' procedure in science it will be helpful to begin with a brief account of the method as it is in itself and as it is in application, and work from there. In discussing the method, I shall concentrate on the account Descartes gives in the early Rules for the Direction of the Mind, which Descartes worked on intermittently from 1618 or so until 1628 or thereabouts; though never finished. and never published, it is by far the most thorough account of method in the Cartesian corpus, far more intelligible than the brief and enigmatic account of the method Descartes gives in part iv of the Discourse. In order to understand the method, we must understand the goal of inquiry in the Rules, for the method of the Rules is precisely a method of attaining that goal. The goal of inquiry is the subject of the first two rules:

The goal [finis] of studies ought to be the direction of one's native abilities [ingenium] toward having solid and true judgements about everything which comes before it. . . . We should concern ourselves only with those objects of which our native abilities seem capable of certain and indubitable cognition. ( A T X359, 362 [ C S M 19, 10]) By 'certain and evident cognition' here, Descartes seems to mean knowledge grounded in what he calls intuition and deduction. In Rule III Descartes defines intuition: By intuition I understand not the fluctuating faith in the senses, nor the deceitful judgement of a poorly composed imagination, but a conception of a pure and attentive mind, so easy and distinct that concerning that which we understand no further doubt remains; or, what is the same, the undoubted conception of a pure and attentive mind, which arises from the light of reason alone. ( A T X368 [ C S M114]) Deduction is defined in terms of intuition; it is a chain of intuitions, the intuitive grasping of a connection between one proposition and another ____________________ 1 For a full defence of this View, see D. Garber, "'Descartes et la méthode en 1637'", in N. Grimaldi and J.-L. Marion (eds.), Le Discours et sa méthode ( Paris: Presses Universitaires de France, 1987), 65-87, and Garber, Descartes' Metaphysical Physics, ch. 2. -235( A T X369-70, 407 [ C S M I14-15, 37]). This, Descartes argues in the Rules, is the only way to knowledge ( A T X366 [ C S M I12-131). Method is what, in the Rules, is supposed to lead us to such knowledge. But what is this method? Descartes writes in Rule IV: By method I understand certain and easy rules which are such that whoever follows them exactly will never take that which is false to be true, and without consuming any mental effort uselessly . . . will arrive at the true knowledge [vera cognitio] of everything of which he is capable. ( A T X371-2 [ C S M I16]) Descartes summarizes these 'certain and easy rules' in Rule V: The whole of method consists in the order and disposition of those things toward which the mental insight [mentis acies] is to be directed so that we discover some truth. And this rule is observed exactly if we reduce involved and obscure propositions step by step to simpler ones, and thus

from an intuition of the simplest we try to ascend by those same steps to a knowledge of all the rest. ( A T X379 [ C S M I20]) Descartes' rule of method has two steps, a reductive step, in which 'involved and obscure propositions' are reduced to simpler ones, and a constructive step, in which we proceed from simpler propositions back to the more complex. 2 But the rule makes little sense; nor does it connect very clearly with the account of knowledge and certainty in terms of intuition and deduction, unless we know what Descartes means here by the reduction to simples and the construction of the complex from the simples. The precise method Descartes has in mind is nicely illustrated by an example he gives of methodical investigation in Rule VIII (see Table 11.1 ). The problem Descartes poses for himself is that of finding the anaclastic line: that is, the shape of a surface 'in which parallel rays are refracted in such a way that they all intersect in a single point after refraction' ( A T X 394 [ C S M I29]). Now, Descartes notices--and this seems to be the first step in the reduction--that 'the determination of this [anaclastic] line depends on the relation between the angle of incidence and the angle of ____________________ 2 To avoid confusion, I should point out that I am breaking with most commentators, who refer to these as the analytic and synthetic steps, following the distinction Descartes draws in the Second Replies: A T VII155-6 or A T IXA 121-2 [ C S M II110-11]. See e.g. C. Serrus, La Méthode de Descartes et son application à la métaphysique ( Paris: Librarie Fdlix Alcan, 1933), ch. 1; Beck, The Method of Descartes, ch. 11, etc. This is a distinction that has little direct relevance to the stages of the method of the Rules. In the Rules we are dealing with a distinction between two parts of a single method; though they are distinct, both are necessary for a true application of the method. But the distinction between analysis and synthesis in the Second Replies is completely different. There we are dealing with different ways of setting out a single line of argumentation, and we must choose one or the other. On analysis and synthesis see also D. Garber and L. Cohen, 'A Point of Order: Analysis, Synthesis, and Descartes Principles', Archiv für Geschichte der Philosophie, 64 ( 1982), 136-47. -236TABLE 11.1 Anaclastic line example (Rules for the Direction of the Mind, Rule VIII) What is the shape of a line (lens) that focuses Q1. parallel rays of light to the same point? What is the relation between angle of incidence Q2. and angle of refraction (i.e. the law of refraction)? How is refraction caused by light passing from Q3. one medium into another?

Q4. Q5. Q6. Intuition: A natural power is. . . . Construction: The construction consists in traversing the series of questions from Q5 to Q1, deducing the answer to each question from that of the preceding question.

How does a ray of light penetrate a transparent body? What is light? What is natural power?

refraction' ( A T X394 [ C S M129]). But, Descartes notes, this question is still 'composite and relative': that is, not sufficiently simple, and we must proceed further in the reduction. Rejecting an empirical investigation of the relation in question, Descartes suggests that we must next ask how the relation between the angles of incidence and refraction is caused by the difference between the two media, for example, air and glass, which in turn raises the question as to 'how the ray penetrates the whole transparent thing, and the knowledge of this penetration presupposes that the nature of the illumination is also known' ( A T X394-5 [ C S M I29]). But, Descartes claims, in order to understand what illumination is, we must know what a natural power (potentia naturalis) is. This is where the reductive step ends. At this point, Descartes seems to think that we can 'clearly see through an intuition of the mind' what a natural power is ( A T X395 [ C S M I29]). Other passages suggest that this intuition is intimately connected with motion. 3 Once we have such an intuition, we can begin the constructive step, and follow, in order, through the questions raised until we have answered the original question, that of the shape of the anaclastic line. This would involve understanding the nature of illumination from the nature of a natural power, 4 understanding the ways rays penetrate ____________________ 3 Rule IX tells us that in order to understand the notion of a natural power, 'I will reflect on the local motions of bodies' ( A T X402 [ C S M134]). What this suggests is that the understanding of illumination is, somehow, an intuitive judgement about the simple nature, motion, though it is not clear how exactly he thought this would work. 4 Descartes writes: 'If, at the second step, he is unable to discern at once what the nature of light's action is . . . he will make an enumeration of all the other natural powers, in the hope that a knowledge of some other natural power will help him understand this one, if only by analogy' -237transparent bodies from the nature of illumination, and the relation between angle of incidence and angle of refraction from all that precedes. And finally, once we know how angle of incidence and angle of refraction are related, we can solve the problem of the anaclastic line. 5

This example develops the programmatic statement of the method as given in. Rule V in a fairly concrete way. If we take the anaclastic line example as our guide, then methodical investigation begins with a question, a question which, in turn, is reduced to questions whose answers are presupposed for the resolution of the original question posed (i.e. Q1 is reduced to Q2 if and only if we must answer Q2 before we can answer Q1). The reductive step of the method thus involves, as Descartes suggests in Rule VI, ordering things 'in so far as some can be known from others, so that whenever some difficulty arises, we will immediately be able to perceive whether it will be helpful to examine some other [question], and what, and. in what order' ( A T X381 [ C S M I21]). And so, in a sense, the reduction leads us to more basic and fundamental questions, from the anaclastic line, to the law of refraction, and back eventually to the nature of a natural power and to the motion of bodies. Ultimately, Descartes thinks, when we follow out this series of questions, from the one that first interests us, to the 'simpler' and more basic questions on which it depends, we will eventually reach an intuition. When the reductive stage is taken to this point, then we can begin the constructive stage. Having intuited the answer to the last question in the reductive series, we can turn the procedure on its head, and begin answering the questions that we have successively raised, in an order the reverse of the order in which we have raised them. What this should involve is starting with the intuition that we have attained through the reductive step, and deducing down from there, until we have answered the question originally raised. Should everything work out as Descartes hopes it will, when we are finished, it is evident that we will have certain knowledge as Descartes understands it in the earliest portions of the Rules; an answer arrived at in this way will constitute a conclusion deduced ultimately from an initial intuition. ____________________ ( A T X395 [ C S M I29]). In personal correspondence, John Nicholas has emphasized to me the importance (and complexity) of this step in the construction. He suggests, plausibly, I think, that 'human limitations are such that in practice we commonly cannot carry out the downward deduction, and have to fall back on the surrogate step of analogizing and comparing with other natural agencies than the targeted one'. In so far as this analogizing may depend on our experience with the phenomenon in question, as well as with other phenomena, this suggests to him that there may be a use of experience in Descartes other than the one that I emphasize in the following sections. He might well be right. 5 See Pierre Costabel, Démarches originates de Descartes savant ( Paris: Vrin, 1982), 53-8, for an account of the historical background to this example. -238Descartes' strategy here is extremely ingenious. The stated goal of the method is certain knowledge, a science deduced from intuitively known premisses. What the method gives us is a workable procedure for discovering an appropriate intuition, one from which the answer to the question posed can be deduced, and it shows us the path that deduction must follow. This workable procedure is the reduction of a question to more and more basic questions, questions that we can identify as questions whose answers are

presupposed for answering the question originally posed; this reduction both leads us to an intuition, Descartes thinks, and shows how we can go from that intuition back to the question originally posed. This is the story as of 1628 or so, when Descartes abandoned the composition of the Rules. As noted earlier in this section, I think that Descartes' thinking about method changes in his later years. Put briefly, while Descartes always maintains the view that knowledge is to be grounded in intuition, in the immediate apprehension of truths, he changes his mind about which truths lie at the bottom, and about how it is that we are to find them. In the Rules he seems to take the view that our knowledge of the physical world is grounded in certain truths, immediately grasped, about the nature of bodies, natural powers, and so forth. 6 But in the later writings, the grounding is ultimately in metaphysics, our knowledge of ourselves and God, and in God's role as the guarantor of our clear and distinct perceptions; in the later writings, the intuitions he takes for granted in the Rules must be grounded in God, our creator, and in us, God's creation. And furthermore, in the latter writings, the reductive step of the method, a step that can lead us only as far as the unjustified intuitions, is abandoned in favour of a direct attack on the foundations. 7 Despite these changes, though, it will be helpful to begin attacking the question of experiment in Descartes by examining the role it plays in his method. METHOD AND EXPERIMENT In the previous section of this essay I emphasized what might be called the deductive structure of Descartes' project, the view of a completed science as a deduction from initial intuitions. In calling the structure 'deductive', I ____________________ 6 See especially the development in Rule XII ( A T X419 [ C S M I44]), where Descartes discusses the so-called simple natures on which all of our knowledge is supposed to be grounded. The simple natures divide into three classes: intellectual, material, and common. The intellectual simple natures include knowledge, doubt, ignorance, and volition. The material simple natures include shape, extension, and motion. The common simple natures include existence, unity, and duration. 7 For a fuller account of the changes, see n. 1 above. -239do not mean to say that it is deductive in precisely the modern sense, or that it is deductive in any precise sense at all. It must be remembered that when Descartes introduces the notion of deduction in the Rules, it is in explicit contrast to the formal logic of the schools--indeed, in explicit contrast to any formal procedures at all. For Descartes, intuition and deduction are the immediate grasping of the truth of propositions and the inferential connections between propositions, and so there is no reason in principle why a deduction cannot be an ampliative inference in the modern sense of the term, as, for example, the Cogito seems to be. 8 But despite Descartes' refusal to pin down the notion of a deduction in any formal way, a completed science is supposed to be

deductive for him in a rather strict sense; derivative and more complex propositions are supposed to be deduced in his sense from propositions more basic and simpler, and grounded ultimately in intuition. However, Descartes is clear, his natural philosophy is definitely not supposed to be a priori in the modern sense of the term, knowledge obtained without the help of experience. Although Descartes seems to want to proceed deductively, experience and experiment have a significant role to play in this business. It is, of course, well known by now that Descartes was a dedicated experimenter, observer, and dissector, and that the empirical investigation of nature is given significant attention in the Rules, the Discourse, and other writings where he discusses his natural philosophy. Of course, this raises an important problem: how is the appeal to experience consistent with the apparently deductive structure of Descartes' project? There is a considerable literature on this basic question, and answers range from denying (or better, ignoring) the interest in experiment, to denying that Descartes' science was ever intended to be deductive, to claiming that Descartes was simply inconsistent-deductive in theory, and empirical in practice. 9 This is the problem I would like to address in this section. I shall try to show something that may sound a bit paradoxical: that for Descartes experiment functions as an important and, in fact, indispensable tool for discovery in his deductive science, and it is to experience that we must turn to help us sort out the details of the deductive hierarchy of knowledge. A reasonable place to begin is with a passage from part vi of the Discourse, where Descartes attempts to explain to the reader the use of ____________________ 8 See Garber, "'Science and Certainty in Descartes'", esp. 116-23. Desmond Clarke argues that the term 'deduction' is so broad for Descartes that even hypothetical arguments count as deductions for him. See Clarke, Descartes' Philosophy of Science, 63-70, 201-2, 207-10. 9 For a survey of the various views taken in the literature, see Clarke, Descartes' Philosophy of Science, 9-10. -240experiment in his thought. The passage begins with a lengthy account of where experiment is not really necessary. Descartes reports that he began his investigations with 'the first principles or first causes' of everything, which can be discovered from 'certain seeds of truth which are naturally in our souls'. From this Descartes derived 'the first and most ordinary effects that one can deduce from these causes', the heavens, stars, the earth, water, air, fire, and so on. The passage then continues as follows: Then when I wanted to descend to those which were more particular, I was presented with so many different kinds of things that I did not think that it was possible for the human mind to distinguish the forms or kinds of bodies which are on the earth from an infinity of others that could have been there, if God had wanted to put them there, nor, consequently, to

make them useful to us, unless one proceeded to the causes through their effects, and attended to many particular experiments. Afterward, reviewing in my mind all of the objects which have ever been presented to my senses, I venture to say that I have never noticed any thing that I could not easily enough explain by the principles that I have found. But I must also admit that the power of nature is so ample and so vast, and that these principles are so simple and so general, that I have found hardly any particular effect which from the first I did not know could be deduced in many ways, and [I admit] that my greatest difficulty is ordinarily to find in which of these ways it depends on these principles. ( A T VI64-5 [ C S M I144]) Experiment seems not to be at issue in the early stages of investigation. Where experiment becomes important, Descartes indicates, is when we move from the most general features of the world, and, as he puts it, descend to particulars. There, he says, the direct deduction from first principles must stop, and we must 'proceed to the causes through their effects, and attend to many particular experiments'. This has suggested to many, and not implausibly, that at this stage science must become a posteriori, arguing from effect to cause by a kind of hypothetico-deductive method of the kind practised in the Essays and defended in the correspondence of 1637 and 1638. 10 While this may describe Descartes' views later, in certain pessimistic sections of the Principles, this is not, I think, what Descartes had in mind in the Discourse. 11 In the passage in question, ____________________ 10 Charles Larmore suggests such a view in ' Descartes' Empirical Epistemology', in Gaukroger (ed.), Descartes: Philosophy, Mathematics and Physics, 6-22, esp. 9, 12. I presented a similar view in 'Science and Certainty in Descartes', though I no longer think that it is correct. 11 One might point here to the obvious use of hypotheses in the Dioptrics and Meteorology, well before the Principles of Philosophy; see A T VI83 ff., 233 ff.: Olscamp 66 ff., 264 ff. But the Essays constitute an attempt to give the results of inquiry without revealing the full system, and they are not intended to replace proper argument in natural philosophy, which proceeds from cause to effect. By arguing from hypotheses, he thought that he could show some of his results without having to divulge the first principles of his physics, for which, he believed, the public was not ready. But, while pleased with his Essays, he was clear that they represent not the definitive -241Descartes seems clear that he is still interested in deduction, even after he has descended to particulars. The problem is that in any given case, there are many possible ways in which one can deduce from the general principles, 'so simple and so general', to the particular effects we observe. Experiment is somehow supposed to help us find the right deductions, the ones that pertain to our world and to the phenomena that concern us. In

this way, experiments seem not to replace deductions, but to aid us in making the proper deductions. 12 The view is initially quite paradoxical. How can some deductions be right and others wrong? How can it be that experiment is essential for a deductive explanation of a phenomenon? And how could Descartes possibly have maintained a deductive structure in his science, if he admits that there are circumstances in which we must 'proceed to causes through their effects'? To see how this might work, let us turn to some examples. As discussed previously, the anaclastic line problem from Rule VIII involves finding the shape of a surface that refracts all parallel rays into a single point. Descartes' solution to the problem requires us to follow a certain series of steps: first a reduction of the problem to a series of simpler ones, then a constructive step, where the reductive series is traversed backwards, resulting in a deductive solution to the problem, if all works well. Descartes never tells us here where we can or must appeal to experience; experience comes up only in a negative way, where Descartes asserts that we should not try to discover the relation between the angle of incidence and the angle of refraction through experiment, for that would violate Rule III, which tells us that only intuition and deduction are sources of real knowledge ( A T X368 [ C S M I14]). But there is at least one place in the reduction where an appeal to experience would seem to be helpful, if not altogether obligatory. In the very next step of the reduction, Descartes says that the investigator must notice that the relation between the angles of incidence and refraction itself depends on the changes in these angles due to the differences in the media through which the ray is passing (e.g. from air into glass, or water into air), and that these changes, ____________________ treatment of his thought, in accordance with his method of inquiry, but, rather, interesting experiments in exposition. There is an extended discussion of this in part vi of the Discourse: A T VI76-7 [ C S M I150]. This theme also runs through Descartes' correspondence in the period; see A T I562-4; A T II141-4, 199-200: C S M K87-8, 103-4, 107. See also Garber, "'Science and Certainty in Descartes'" and Descartes' Metaphysical Physics, ch. 2. 12 See also Descartes' remarks in Principles, part iii, art. 4. There he talks about having to turn to the phenomena at that stage in his exposition, 'not to deduce an account of causes from their effects', but 'to direct our mind to a consideration of some effects rather than others from among the countless effects which we take to be producible from the selfsame causes'. -242in turn, depend on the way in which the ray penetrates the transparent body ( A T X394 [ C S M I29]). Descartes does not mention experiment or experience in this context. But it is difficult to imagine that this is a step that we can make on the basis of the 'seeds of truth' alone. While it may not require sophisticated optical experiments, it seems that we at least require some minimal experience with light rays and lenses, or other actual

instances of refraction, in order to see that light is typically bent by passing from one medium into another, and to come to the realization that, in order to discover the law which refraction obeys, we must first understand how light passes through media of different sorts. In this way experience would seem to help us to see how we might proceed in our investigation by suggesting what further questions it might be useful for us to look into. Experiment comes up at best only implicitly in the anaclastic line example. But it is quite a visible feature of another example Descartes gives of his method. The example I have in mind is the account Descartes gives of the rainbow in the Eighth Discourse of his Meteorology. This passage contains the only explicit mention of the method in all of the Essays, and it is singled out in a letter from 1638 as an exemplary use of the method in practice (see A T VI325: Olscamp332, and Descartes to Vatier, 22 Feb. 1638 ( A T I559: C S M K85)). The example is a very complicated one, one of Descartes' best but most complex scientific arguments. I shall begin by summarizing the argument, and then try to show how the mass of experimental detail and complex argument sorts itself out into a methodical framework (see Fig. 11.1 ). 13 The problem is to explain how it is that rainbows come about. The account begins with the observation that rainbows appear when and only when there are water droplets in the air. Descartes then turns to the study of large spherical flasks of water, which, he claims, duplicate the effects seen in individual droplets of water that appear to cause the rainbow. Observations of the flask allow Descartes to measure the angles at which colours are observed, and allow him to determine that there are two regions of colour whose red portions are about 42 and 52 degrees from the ____________________ 13 My own interest in the rainbow case here is largely as an illustration of the method of the Rules. For discussions of Descartes' account of the rainbow that emphasize its place in the history of such discussions and in the history of optics more generally, see Carl B. Boyer, The Rainbow ( Princeton: Princeton University Press, 1987), ch. 8, and Jean-Robert Armogathe, 'L'Arc-en-ciel dans les Météores', in Le Discours et sa méthode, 145 -62. Considering Descartes' account in its historical perspective makes it quite clear that, despite the impression he gives in the Meteorology of having discovered everything himself, he owes a great deal to previous investigators. Interesting and important as these historical considerations are, I will focus instead on Descartes' presentation of his theory in an attempt to untangle the methodological underpinnings of his argument. -243FIGURE 11.1 angle at which they are hit by the rays of the sun (see Fig. 11.1 ). These experiments also allow Descartes to determine that these two regions of colour derive from two different combinations of reflection and refraction within the water flask; the brighter colour

region (which corresponds to what is now called the primary bow) at 42 degrees results from two refractions and one internal reflection, while the dimmer colour region (the secondary bow) at 52 degrees results from two refractions and two internal reflections. (The two paths can be discerned within the flask represented in Fig. 11.1 ). There investigations lead Descartes to two further questions: why there is colour at all in these cases, and why it is that the colours appear at two specific angles. The first question, that of colour, is explored experimentally, through a prism, in which, like the flask, colours are produced through the reflection and refraction of light (see Fig. 11.2 ). Observations made with the prism show that a curved surface, like that of the raindrop or the flask, is not needed to produce colour; nor is a -244FIGURE 11.2 reflection necessary, Descartes discovers through experiment. What seems to be necessary, Descartes finds, is at least one refraction, and a restricted stream of light. But in order to understand how the refraction of a restricted beam of light can produce colour, we must press deeper into the nature of light and the way it passes through a transparent body, the very questions that we were pressed back into in the anaclastic line case. The nature of light we know from the Dioptrics: '[The nature of light is] the action of movement of a certain very fine material whose particles must be pictured as small balls rolling in the pores of earthly bodies' ( A T VI331: Olscamp336). 14 And, Descartes argues, what happens when a restricted beam of light passes from one medium into another in refraction is that the balls are given differential tendencies to rotate, depending on where they are in the stream (see Figs. 11.2 and 11.3 ). Since, refraction aside, that is the only mechanical effect that passing from one medium into another has on the light, Descartes argues that colour just must be caused by the differential tendency to rotation. Those balls with a greater tendency to ____________________ 14 This is the paraphrase Descartes gives in the Meteorology; the passage he is referring to in the Dioptrics can be found at A T VI89-93 [ C S M I155-6]. -245FIGURE 11.3 rotate produce the colour red in us, Descartes claims, while those that have a lesser tendency to rotate produce the colour blue/violet in us. (Remember, of course, Descartes held that in the strictest sense, colour is only in the mind, and not in bodies.) And so, from the nature of light and the way it passes through media, we have shown how colours are produced, Descartes thinks. But it still remains to show why the colours are produced in two discrete regions, at characteristic angles from that of sunlight. To solve that problem, Descartes turns back to the flask. Appealing to the law of refraction, which

Descartes alludes to in the anaclastic line example, and derives (after a fashion) in the Second Discourse of the Dioptrics, he demonstrates that after two refractions and one reflection, the vast majority of a bundle of parallel rays hitting the flask, wherever they may hit, will emerge from the flask between 41 and 42 degrees with respect to the angle of the incident light, and after two refractions and two reflections, the majority will emerge at between 51 and 52 degrees ( A T VI 336 ff.: Olscamp339 ff.). 15 From this it follows that at those two regions ____________________ 15 Descartes does the calculation by considering a spherical droplet of water hit by parallel Tays over one hemisphere, and calculating where various of the rays would emerge after an appropriate number of reflections and refractions. His conclusion, carefully stated, reads: I found that after one reflection and two refractions, very many more of [the rays] can be seen under the angle of 41 to 42 degrees than under any lesser one; and that none of them can be seen under a larger angle. Then I also found that after two reflections and two refractions, very many more of them come toward the eye under a 51 to 52 degree angle, than under any larger one; and no such rays come under a lesser. ( A T VI336: Olscamp339) -246on the surface of the sphere, there will be two discrete streams of light that emerge from the flask, moving from one medium into another. And from the previous argument, this will result in two regions of colour at the two angles earlier observed. And so, from the nature of light, the way it passes through media, and the law of refraction, if follows that the rays of sunlight hitting the flask will result in two regions of colour at two characteristic angles. When we have a multitude of such drops, we have a rainbow. It is by no means obvious how this somewhat confused mass of experiment and reasoning can be fitted into the rather rigid mould of Descartes' method. The schematic representation of the argument given in Table 11.2 indicates one plausible way in which the argument might fit. In the schematic representation of the argument, Q1 through Q5 represent the reduction, which leads us from the question originally posed, 'What is the cause of the rainbow?', back to the intuitions which are the starting-point of the Cartesian deduction, intuitions about the nature of light and how it passes through media. But the important thing is, of course, the specific path that Descartes follows to go from the initial question to the intuition, for it is that path that will determine the path followed in the deduction. In this case Descartes proceeds by splitting the question into two questions: one about colour and one about the two regions. Included in square brackets are the empirical results derived from experiment at the point in the argument in which Descartes appeals to them. The path followed after the intuition is relatively straightforward. Here we are dealing with the same steps followed in the reduction, only in the reverse order, as we pass from intuition to the final answer to the question originally posed. But unlike the reduction, experiment and its results seem to play no role

in this part of the argument. The example is certainly much more complex than the anaclastic line example, but it seems to have much in common with it in structure. Before turning back to my main theme, the use of experiment in these arguments, I would like to comment on the kind of deduction that is involved in this case. In the anaclastic line case, we had a definite question, the shape of a lens with such-and-such properties, and at the conclusion of the procedure we can expect a deductive answer to the question, a ____________________ While the conclusion is arrived at by calculation, that calculation must make explicit appeal to the index of refraction for water. When the question comes up in the Second Discourse of the Dioptrics, he notes that we must appeal to experience in order to determine the value of this constant for various sorts of materials ( A T VI101-2 [ C S M I161-2]). This would seem to be another place in which experiment would enter into the method. However, one presumes that Descartes believed that the index of refraction could itself be arrived at by calculation, were we to know the size, shape, and motion of the corpuscles that make up water. -247TABLE 11.2 Descartes' account of the rainbow (Meteorology, Eighth Discourse) -248deduction from basic principles (ultimately, the nature of a natural power) that a lens with this-or-that shape will have such-and-such characteristics. But the situation here is a bit different. What we are seeking is the cause of the rainbow. The answer to this question is, in a sense, not deduced; rather, it is revealed in the deduction itself. The deduction shows us how we can go from the nature of light to the phenomenon of the rainbow; what is deduced, strictly speaking, is just the phenomenon itself, the patches of colour in the sky. But the path followed in deducing the phenomenon shows us that the cause is the passing of light from one medium to another, the differential tendency to rotate this passage gives the particles of light, and the way that the law of refraction causes light rays to converge into two discrete streams at two characteristic angles. This is a deduction, but a deduction of a very different sort from the one in the anaclastic line example. One can quite plausibly ask if Descartes can really be sure that he has given the true sequence of causes that produce the rainbow, as opposed to a possible sequence that produced the same appearances. Descartes himself will later come to see that as a problem. 16 But in the Meteorology it is not; he seems confident that the methodical procedure of investigation he is following assures him that he has captured the real causes. To return to my main thread, a number of interesting things emerge from these two examples. First of all, it would appear that experiment functions strictly at the reductive stage of method, the stage in which we are trying to go from a question posed to the intuition from which the answer is to be derived; experiment seems not to be involved in

the actual deduction. And in that initial stage of inquiry, it seems to function in two not altogether separable roles. First of all, it helps better to define the phenomenon to be deduced or the problem to be solved. This is not at issue in the anaclastic line example, where the problem is set with sufficient precision. But it is an important function of experiment in the rainbow example, where Descartes appeals to experiment to fix what the rainbow is, that it consists of two separate bows, and that the two bows are always at suchand-such an angle with respect to the rays of the sun; in this way, experiment clarifies the question that is to be answered. 17 But just as importantly, experiment aids the reduction by suggesting how things depend on one another, and, in that way, suggesting at a given juncture what question we might turn to next. It is because we know from experiment that refraction ____________________ 16 See e.g. Descartes' remarks in Principles, part iv, arts. 204-6; see the discussion of these passages in Garber, 'Science and Certainty in Descartes'. 17 See the discussion in Rule XIII ( A T X430-1 [ C S M I51-2]), where Descartes discusses the importance of specifying in exact terms what is being sought in an investigation. -249depends on a light ray passing from one medium to another that we know that we must investigate light rays, media, and how light passes through a medium in order to determine the law of refraction. Similarly, it is because of experiments with the prism that we know that reflection is irrelevant to colour, but refraction is not, and it is because we know that colours can arise from the refraction of light that we know that the nature of colour is to be sought in an examination of what light is, and how it is altered by refraction. Once we understand Descartes' method and the roles that experiment does (and does not) play in it, it should come as no surprise that Descartes might suggest that 'it would be very useful if some . . . person were to write the history of celestial phenomena in accordance with the Baconian method . . . without any arguments or hypotheses' (letter to Mersenne, 10 May 1632 ( A T I251: C S M K38)). The sorts of tables that Bacon recommends to the investigator in book ii of his Novum Organum can tell us, for instance, that factor A (colour, say) is always accompanied by factor B (refraction, say), but that factor C (say, reflection) is present in some cases but absent in others. In an investigation of A, this could lead us to questions about B, and prevent us from raising irrelevant questions about C, as when in the rainbow example we learn that refraction is relevant to colour, but reflection is not. Such tables of phenomena and their correlations with one another, independent of any theory, are precisely what Descartes needs to define problems and to determine the relations of dependence of one phenomenon on another necessary to perform the reductive step of the method. In this way, it seems that experiment is not a replacement for deduction, but part of the step preliminary to making a deduction. Science remains deductive for Descartes; in the end, our knowledge of the cause of the rainbow depends on our performing a deduction of the phenomena from an initial intuition. But experiment seems to play its role in

preparing the deduction. In so far as it helps perform the reductive part of the method, the sequence of steps that leads from a question to an intuition, it helps determine the deduction, the same steps followed in reverse order that lead from intuition to the answer to the question posed. The deductive chain that the Cartesian scientist seeks in reason, the chain that goes from more basic to less, is exemplified in the connections one finds in nature itself. In so far as these latter connections are open to experimental determination, we can use experiment to sketch out the chain of connections in nature and find out what depends on what, and thus we can use the connections we find in nature as a guide to the connections we seek in reason. It may not be obvious to us at first just how we can go deductively from the -250nature of light to the rainbow, but poking about with water droplets, flasks, and prisms may suggest a path our deduction might follow. This understanding of how experiment and observation may be useful in a deductive science of the sort that Descartes was attempting to construct allows us to make some sense of some of the more puzzling aspects of Descartes' remarks. On this understanding, we do find causes through their effects, in a sense; experiment is quite necessary in solving problems and helping us to discover the real causes of phenomena in our world. But in no sense are we replacing deductive with a posteriori reasoning. Though we must appeal to experiment, experiment only prepares the deduction that will establish the cause. Furthermore, we can now see how experiment can point the way to the 'correct' deduction, and eliminate the 'incorrect' deductions. There can be alternative derivations of a given phenomenon in the sense that the same bare effect may be produced by different chains of causes. For example, a distribution of colours in a pair of bows in the sky (a bare effect) may be produced by the reflection and refraction of light through raindrops (as it actually is in our world), or by a distribution of tiny Coloured balls suspended in the air, or by colours projected by a slide projector on a cloud of dust, or by any number of other perverse means. But experiment helps us find the correct deduction: that is, the correct chain of causes, by making the phenomenon more precise, and suggesting how it is that the phenomenon is actually produced in this world. In this way experiment can lead us to the correct derivation--correct in the sense that it represents the way the phenomena are caused in our part of the universe. Alternative deductions are not wrong, strictly speaking; one might be able to produce something that looks to us very much like a rainbow in any number of ways. But it's just that it is not the way things are done here, at least not the way it is done in nature. So far I have talked about experiment in the context of Descartes' official method. But, as I pointed out at the very beginning of this essay, I think that Descartes later came to set his method aside. In his later writings, those that follow the Discourse, I would argue that Descartes abandoned the reductive stage of his method in favour of a direct attack on the tree of knowledge, starting from intuition (or, rather, first principles, first philosophy) and deducing on down from there. But I think that much of what I said about experiment in the method also holds good for the systembuilding orientation of later works like the

Principles of Philosophy. Though in the later writings an explicit reductive step is not in evidence, Descartes must find some way of constituting his deductive chain, and here experiment will be useful for the same reason it is in the method. It is, I think, no accident that at the moment that Descartes was working on -251extending the system of the Principles from the inanimate world, derived by the laws of nature from an initial chaos, to the world of plants and animals, he was also doing experiments on the formation of the foetus. 18 I am certain that Descartes thought that in sexual reproduction, the development of a living body from mechanical causes, he might find clues about how living bodies originally arose on this earth through mechanical causes, and that such clues would help him extend the deduction of terrestrial phenomena begun in the Principles to living things. EXPERIMENT AND THE PRIORITY OF REASON In the previous section I tried to show how experiment plays a role in Descartes' scientific procedure, how experiment is needed in at least certain circumstances to aid in the deduction that leads us to genuine knowledge through deduction. But this raises an interesting question. Descartes is usually identified, and rightly so, as the philosopher of reason, the philosopher who rejected the dependence on the senses that characterizes the Aristotelian philosophy that he was eager to replace, in favour of dependence on clear and distinct perception, the immediate dictates of the light of reason. I have tried to show how Descartes' deductive science is not compromised by the way in which he appeals to experiment, how the particular conception Descartes has of the deductive structure of knowledge is fully consistent with the use of reason as an auxiliary to the reductive step of his method. But a deeper question still remains: how any use of experiment at all is consistent with his strictures against the appeal to experience. Descartes certainly does oppose naïve dependence on the senses in passages too numerous to cite; he warns us that things are not at all as our senses tell us they are, that they are not red and green, sweet or salty, that our naïve belief that all of our knowledge derives ultimately from our senses is a prejudice of sense- and body-bound youth, a prejudice that must be rejected before we will be able to penetrate to the true nature of things. In his Meditations, he begins with a series of sceptical arguments that are directed in large part, if not entirely, against our naïve trust in the senses, and in the Fourth Meditation, he appears to recommend that we must limit ourselves to knowledge derived from the light of reason; he appears to argue that only by limiting ourselves to clear and distinct perception can we guarantee that we do not stray into intellectual sin: that ____________________ 18 See Descartes, La Description du corps humain ( A T XI252 ff. [ C S M1321 ff.]). -252-

is, error. And if we are to limit ourselves to clear and distinct perceptions, then there would seem to be no room for any appeal to experience at all, it would seem, even the sort of appeal that I outlined in the previous section. 19 But, I think, the situation is a bit more complex than this textbook summary of Descartes' epistemology might suggest. Descartes does certainly favour reason over the senses, but he certainly does not recommend rejecting the senses altogether. The fullest account of Descartes' views on the senses and the role that they play in the acquisition of knowledge occurs in the Sixth Meditation. The reconsideration of the senses, rejected earlier in the First Meditation, begins early in the Sixth Meditation. Earlier, unsuccessful attempts to prove the existence of bodies led the meditator to consider more carefully the faculty of imagination and the closely related faculty of sensation ( A T VII74 [ C S M II51]). And so the meditator goes back over the considerations that led him first to trust the senses, ending with a review of the considerations that originally led him to question the senses ( A T VII 74-7 [ C S M II514]). At this point, the meditator notes: But now, after I have begun to know myself and my author a bit better, I do not think that everything that I seem to get from my senses should simply be accepted, but then I don't think that everything should be rendered doubtful either. ( A T VII 77-8 [ C S M II54]) The senses loom large in the rest of the Meditation. The meditator first distinguishes between the mind and the body. Then the question turns to the external world, and it is here that the senses make their first positive contribution to the enterprise. The meditator begins: 'Now there is in me a certain passive faculty of sensing, that is, of receiving and knowing the ideas of sensible things' ( A T VII79 [ C S M II55]). We have a passive faculty of sensation. But this would be of use only if there were, somewhere, an active faculty for producing these ideas, a cause. This, Descartes argues, could not be in me, for it seems to involve neither my understanding nor my will, the two faculties I have. So, the meditator reasons, the ideas of sensation he has must come from outside of him, either from God or from bodies (i.e. bodies as understood in the Fifth Meditation, things extended and extended alone) or from something else. The meditator reasons that it must be from bodies themselves that our ideas derive; God has given me a 'great propensity for believing that they come to me from ____________________ 19 For a development of some of these themes in Descartes, see Garber, 'Semel in Vita: The Scientific Background to Descartes' Meditations', in Rorty (ed.), Essays on Descartes' Meditations, 81-116. -253corporeal things', while he has given me 'no faculty at all' for learning that this propensity might be mistaken ( A T VII79-80 [ C S M II55]). So, the meditator argues to himself,

God would be a deceiver if it turned out that our ideas of bodies come from anywhere else but from bodies themselves. And so, he concludes, bodies exist. 20 The argument is a very interesting one. A conclusion is established not because we have a clear and distinct perception that bodies exist, exactly, but because the meditator has a 'great propensity' for believing something, and God has given him no way of correcting that propensity. 21 Descartes admits here that there are at least some circumstances in which a belief that we seem to get from sensation, the inclination to believe that seems to come to us with the sensation, is worthy of our trust. It may not be as worthy of our trust as a genuine clear and distinct perception, as he implies in the Synopsis of the Meditations ( A T VII16 [ C S M II11]), and it may not always be true, as a clear and distinct perception is. But when sensation leads us to a belief, as it does in this case, and when that belief is not overridden, as it were, by a reason for rejecting it, as is the case with our beliefs about colours actually being in things, say, then we can trust the senses. 22 This is the strategy that Descartes pursues in the remainder of the Sixth Meditation in his discussion of the senses. He argues that what he calls the 'teachings of nature', which include the beliefs that appear to arise spontaneously with sensations, can be trusted as being for the most part true when corroborated by reason: that is, when reason does not give us better grounds for rejecting a judgement from the senses, or when reason is in accord with that judgement, or when reason is silent on the question. As with clear and distinct perceptions, Descartes is here dealing with something that God gave us, beliefs that are, in a certain sense, innate: 'I am dealing only with those things that God gave me as a composite of mind and body' ( A T VII82 [ C S M II57]). As such, Descartes argues, they must be in some sense true: 'it is doubtless true that everything that ____________________ 20 For a fuller presentation of this argument, see ibid. 21 In the version of the argument given in Principles, part ii, art. 1, Descartes does seem to argue from the fact that 'we seem clearly to see' that sensation proceeds to us from the object of our idea of body to the real existence of body, and does not appeal to the 'great propensity' that is the nub of the argument in the Meditations. It is not clear why the later text differs from the earlier one on this point. It may represent a genuine change in Descartes' epistemology. But then it may simply reflect Descartes' desire not to enter into his full account of the senses in the Principles. For the relation between the Meditations and the Principles, see Garber and Cohen, 'A Point of Order'. 22 That is, we can trust at least some of the judgements that characteristically accompany our sense perceptions. What seems to be at issue here is the third of Descartes' three grades of sensation; see A T VII436-7 [ C S M II294-5]. -254nature teaches me has some truth in it' ( A T VII80 [ C S M II56]). When it is truth about the nature of things that we are interested in, it is the light of reason, clear and distinct perceptions, that we must turn to first. Descartes writes:

And so, my nature teaches me to flee what gives me pain and to seek what gives me pleasure, and the like. But it does not appear that it teaches us to conclude anything about things outside of us from the perceptions of the senses without a prior examination of the intellect, since knowing the truth about things seems to pertain to the mind alone, and not to the composite [of mind and body]. ( A T VII82-3 [ C S M II57]; emphasis added) And so, while some of the teachings of nature will turn out to be true, it is only the intellectual examination of them that will establish this. In this way Descartes restores the senses and rejects the hyperbolic rejection of the senses that begins the Meditations; indeed, he goes on to reject even the dream argument that is so prominent in the First Meditation ( A T VII 89-90 [ C S M II61-2]). But though the teachings of nature, what we learn from our senses, are restored, they are subordinate to reason; they may be trusted to some extent and in some circumstances, but only after they have been given a clean bill of health by reason. It is with this in mind that we should return to the use of experiment in the rainbow case discussed earlier. One can say that in so far as Descartes does allow the appeal to the senses in at least a general way, there is no inconsistency in Cartesian epistemology; as long as what Descartes takes from the experiments to which he appeals falls within the bounds of proper caution, there is no special problem here. But there is something more interesting to be said in this case about the way in which experience is subordinate to reason. In the previous section, I showed that while experiment might function as an auxiliary to a deduction, it is the deduction itself, and not the experiment, that yields the knowledge. So, for example, in the anaclastic line case, while experience might suggest to us that there is some lawlike relation between angle of incidence and angle of refraction, it is only through deduction that the actual law can be established (see Rule VIII ( A T X394 [ C S M I29])). But the point goes deeper still. In the rainbow case, Descartes begins by observing that on his flask, the stand-in for the raindrop, there are two regions of colour, at roughly 42 and 52 degrees from the ray of sunlight, which angles are then deduced in the end from his. theory. After giving his account, Descartes notes that an earlier observer, the sixteenth-century mathematician Franciscus Maurolycus, set the angles incorrectly at 45 and 56 degrees, on the basis of faulty observations. Descartes notes that 'this shows how little faith one ought to have in -255observations which are not accompanied by the true reason' ( A T VI340: Olscamp342). 23 It is only because we can calculate the angles of the primary and secondary bows from the account we have of the rainbow that we can be sure of what they are, despite the fact that the investigation began with an experimental determination of those angles. 24 Though it is an observation that starts the ball rolling, it is only through a Cartesian deduction that the phenomena and causal dependencies observed can actually enter the body of scientific knowledge, strictly speaking. Similarly, it is only because a deduction

can, indeed, be made in the reverse order of the causal dependencies that experiment has found, that those dependencies ought to be trusted. Descartes is, of course, aware that colour can arise not only from refraction of light, but from the reflection of light off of a surface whose texture is appropriate to cause the changes in the light necessary to produce the colour seen. At one point in his discussion of the rainbow Descartes seems prepared to consider such an account of colour in the rainbow, because, at first glance, the restriction on the beam of light necessary to produce colour through refraction seems to be absent ( A T VI 335: Olscamp338-9). And so, it seems, the causal dependence of the colours of the rainbow on refraction and reflection suggested by experiment is only provisional; while the experimental determination of the path the light follows through the droplet may suggest to us a deductive path that we might be able to follow, it is the actual success of the deduction from intuition to phenomena that actually establishes the causal connections that produce the phenomena. Experiment is important in helping to find the deduction, but it is the deduction that, in an important sense, fixes both the causal path and the phenomena. Experience is important, but only under the control of reason, as Descartes took great pains to emphasize in the Sixth Meditation. This feature of Descartes' position connects in an interesting way with an often discussed problem in the philosophy of science, the question of the theory-ladenness of observation. Whether or not one can have an observation that is not in an important way dependent on some theory or other is a question too often discussed in the abstract. Descartes' appeal to ____________________ 23 For a discussion of Maurolycus's theory of the rainbow, see Boyer, The Rainbow, 15663. The implication of Descartes' remarks is that Maurolycus's values for the angles derive from observation alone. This is not entirely fair. Maurolycus had his reasons for setting the angles as he did, reasons based on his (incorrect) analysis of the path the light follows within the raindrop; indeed, he knew that his calculated value differs from what was known through observation, something for which he attempted to offer an explanation (pp. 159 -60). 24 We must, of course, remember that the calculation does appeal to an experimentally determined value for the index of refraction; however, as I pointed out earlier, Descartes would surely have thought that a 'reason' could be given for that too. -256experiment in the rainbow case shows an interesting complexity in the whole dispute. Descartes does use observation to motivate the theory that he is proposing, or, perhaps, to guide us to that theory. In this sense, observation would seem to be atheoretical for Descartes. But at the same time it is extremely important to realize that the observations Descartes presents as motivating his account of the rainbow, or at least guiding it, are not to be trusted fully until we have an account of the matter, until we can derive those observations from more basic principles. There is such a thing as pre-theoretical observation for Descartes, and this does seem to have a role to play in his procedure. But, at the same time, there is an important sense in which observation does not attain the

status of fact until it becomes integrated with theory-indeed, until it becomes subordinated to theory. In this way, for Descartes, experiment by itself can establish no facts; while experiment can lead us to facts, it is only the final deduction of a phenomenon from intuited first principles that establishes the credentials of a fact, even if first 'discovered' through experiment. In his recent writings, Ian Hacking argues that experiment must be viewed as in an important sense independent of theorization in science: 'experiment has a life of its own,' he insists. 25 By this he means to point out, among other things, that experiment does not function exclusively in the service of theoretical argument, furnishing premisses for theoretical arguments, testing theories proposed, allowing us to eliminate one of a pair of competing theories and accept another, and so forth. This may be true enough for a wide variety of figures. But it is not true for Descartes. For Descartes, at least in the context of the rainbow, experiment plays a carefully regimented role in what is from the start a theoretical project. But, at the same time, neither do experimental phenomena have a role assigned to them in standard hypothetico-deductive conceptions of scientific method, as the touchstone of theory, the atheoretical facts to which we can appeal to adjudicate between alternative theories. If my account of experiment is correct, then however much experiment might help us to find the correct account, it is ultimately reason, not experiment, that is the touchstone of reality, for theory as well as for the experimental facts that help us construct theories. 26 ____________________ 25 Ian Hacking, Representing and Intervening ( Cambridge: Cambridge University Press, 1983), 150. 26 Descartes does say some things that would appear to go against my conclusion. For example, immediately following the long passage from part vi of the Discourse I quoted earlier, Descartes writes: 'I know of no other means to discover this [i.e. how a particular effect depends on the general principles of nature] than by seeking further experiments [expériences] whose outcomes vary according to which of these ways provides the correct explanation' ( A T VI65 -257On the standard view of things, widely shared since the late eighteenth century or so, there are two sorts of philosophers: rationalists and empiricists. Descartes is traditionally viewed as a rationalist--in fact, the founder of the school, in modern times at least. When the extent of Descartes' dependence on experiment and observation is recognized, there is a temptation simply to think that Descartes must have been placed in the wrong slot, and conclude that he must really be some sort of empiricist. 27 I would resist that temptation. It seems to me that what the case of Descartes shows is how crude the scheme of classification really is. For Descartes both reason and experience are important, though in different ways. His genius was in seeing how experience and experiment might play a role in acquiring knowledge without undermining the commitment to a picture of knowledge that had motivated him since his youth, a picture of a grand system of certain knowledge, grounded in the intuitive apprehension of first principles. 28

____________________ [ C S M I144]). But, I think, this must be understood in the context of the interpretation I have offered earlier. The experiments in question must be viewed as leading us down one deductive path rather than down another, and not as a theoryneutral means of choosing between independently constructed theories; for, as Descartes elsewhere insists, we cannot really be sure of an experimental fact until after we have already determined what the correct deduction is. 27 See e.g. Clarke, Descartes' Philosophy of Science, 205. 28 In addition to its presentation at the San José Descartes Conference, earlier versions of this paper were given at the University of Rochester, University of Colorado, Columbia University, Catholic University of America, University of Notre Dame, Georgia Philosophical Association, University of California at San Diego, Virginia Polytechnic Institute and State University, University of Illinois, and the University of Ohio. I would like to thank the audiences there, as well as Peter Dear, Ernan McMullin, John Nicholas, Gary Hatfield, and Beverly Whelton for their very helpful suggestions. Parts of the text also appear in ch. 2 of my book Descartes' Metaphysical Physics. -258XII DESCARTES' CONCEPT OF SCIENTIFIC EXPLANATION DESMOND M. CLARKE If Descartes can be excused of the most characteristic theses of the paradigm rationalist, then two further questions arise with almost equal importunity. One of these concerns the Cartesian understanding of a scientific explanation: what, for Descartes, would satisfy his requirements that a purported explanation is truly scientific? And secondly, in what way would the plausibility of such explanations be evaluated? The two questions are interconnected, for if one imagines that Cartesian explanations are hypothetical, this has corresponding implications for any discussion of their evaluation. And if Descartes is thought to reject experiential evidence in confirming or disconfirming explanations, then this likewise has obvious implications for his understanding of explanation. The first question about the Cartesian concept of explanation--is the subject of this essay. I. CAUSAL EXPLANATIONS To explain a physical phenomenon, for Descartes, was equivalent to (i) specifying its efficient causes, and (ii) describing the mechanism by which the phenomenon results in some 'necessary' way from the assumed causes. If this rather general concept of explanation is combined with various other metaphysical and methodological assumptions of Cartesian science, a more specific understanding of scientific explanation quickly emerges. For example, Descartes assumes that all physical phenomena are explicable in terms of the motions and interactions of small, imperceptible particles of matter. 1 Since nothing less than a causal explanation is

____________________ Desmond Clarke, 'Discartes' Concept of Scientific Explanation' from D. M. Clarke, Descartes' Philosophy of Science ( Manchester University Press, 1982), 108-32, by permission of the author. 1 Principles, part iv, art. 203 ( A T VIIIA 326 [ C S M I289]): 'I have tried to discover, from the observable (sensilibus) effects and parts of natural bodies, the character of their unobservable (insensiles) causes and constituent parts.' -259satisfactory, it follows that any account which Descartes might be willing to endorse must be hypothetical. On the other hand, Descartes was obviously unwilling to settle for mere conjectures, guesswork, or unproved assumptions; so that even though he used the word 'hypothesis' to imply that we have no direct, experiential access to the causes or mechanisms which our scientific theories must describe, at the same time he imagined he could eliminate the uncertainty of hypotheses by a variety of different strategies. Descartes' approach to constructing hypotheses, to choosing between alternative explanations, and to confirming or disconfirming suggested scientific explanations will not be examined in detail here. 1a For the moment it is enough to notice that the imperceptibility of the kinds of cause which are involved and of the mechanisms by which they operate implies that a Cartesian scientific explanation must be hypothetical. 2 Scientific explanations are also a priori. 3 This is a misleading expression for the contemporary reader, whose understanding of 'a priori' is so much influenced by Kant. For while 'a priori' now means something like 'independent of experience or empirical evidence', for Descartes it meant something more like 'causal'. This is what Descartes apparently means when he refers to his account of optical refraction in the Dioptrics as 'a priori'; he wrote to Mersenne in 1638: 'You should know that I demonstrated the refractions geometrically and a priori in my Dioptrics, and I am amazed that you still doubt it' ( A T II31). Likewise for reflection, in a letter to Mersenne in 1640: '1 believe that, in the second discourse of my Dioptrics, I have provided an a priori explanation of why reflection takes place with equal angles' ( A T III82). In each case, an a priori explanation of the optical phenomena in question is an explanation in terms of the motions of 'light' particles and of their interactions with various kinds of media or surfaces. There are two reasons for understanding these texts as references to causal explanations rather than to explanations which are independent of empirical evidence. One reason is that Descartes openly admits that the general claims about the nature of matter and light which are found at the beginning of the Dioptrics are assumptions or hypotheses. Thus he wrote to Plempius, in December 1637: ____________________ 1a These topics are dealt with by Clarke in Descartes' Philosophy of Science, ch. 6. 2 For a much clearer statement of this position, see Pierre-Sylvain Régis, Cours entier de philosophie ou système générale selon les Principes de M. Descartes, i.274: 'in the

case of physical bodies, because their parts are unobservable one cannot perceive either the order or the arrangement of the parts, and the most one could hope for would be to guess at these from the known effects'. 3 Cf. Beck, Method of Descartes, 249 n. 1: ' Descartes admits that the existence of any particular cannot be proved a priori; his contention is that all scientific explanation is a priori.' It appears that Beck does not understand 'a priori' as I do here. -260The things which I propose in the first chapters about the nature of light, and about the shape of particles of salt and fresh water, are not my principles as you seem to object, but rather the conclusions which are demonstrated by everything which follows. ( A T 1476 [ C S M K77]) The following year he wrote in a similar vein to Vatier, concerning the general assumptions which initiate the Meteorology: As regards what I assumed at the beginning of my Meteorology, I could not demonstrate it a priori without giving the whole of my physics; but the expériences which I deduced necessarily from it--which cannot be deduced in the same way from other principles--seem to me to demonstrate it adequately a posteriori. ( A T I563 [ C S M K87]) 4 This suggests that the introductory assumptions of the two essays, the Dioptrics and the Meteorology, are not provided with an a priori demonstration, whatever that might mean; and at the same time, that they can provide an a priori explanation of refraction or reflection. The other reason which supports the causal interpretation of the phrase 'a priori' is Descartes' use of the term in cases where empirical evidence is used as the only source of warrant for an a priori claim. For example, Descartes wrote to Mersenne in 1632 about the possibility of constructing an a priori explanation of all the terrestrial bodies: For, although they [the fixed stars] seem to be very irregulary scattered here and there in the heavens, nevertheless I do not doubt that there is a natural order among them which is regular and determinate. And the knowledge of this order is the key and foundation of the highest and most perfect science . . . For by means of this science one could know a priori all the diverse forms and essences of terrestrial bodies whereas without it we have to content ourselves with guessing them a posteriori, and through their effects. I cannot think of anything which would help me more in coming to know this order than the observation of many comets. ( A T 1250-1 [ C S M K38]; my emphasis). Descartes continues his letter by urging that someone should compile a complete inventory of anything known about the heavens, 'in accordance with the Baconian

method . . . and without introducing any explanations or hypotheses' ( ibid. 251 ). If an a priori explanation of terrestrial phenomena should ideally begin with a Baconian listing of the known facts, it must be obvious that Descartes means something else by 'a priori' than ____________________ 4 Cf. also Descartes (to P. Fournet?), 3 Oct. 1637 ( A T I455 [ C S M K75]); Descartes (to Boswell?), ( 1646?) ( A T IV689), which refers to the discussion of matter on page 42 of the Discourse: 'As regards the subtle matter, it is true that I did not prove it a priori; since I had no intention of dealing with the whole of my philosophy in such a book, I had to begin somewhere and therefore I wrote that I assumed it. But I contend that there are more than five hundred reasons in the Meteorology and the Dioptrics which prove it a posteriori; that is, that I explain more than five hundred difficulties which could not be explained without it.' See also Descartes to Morin, 13 July 1638 ( A T II201 [ C S M K108]), where a similar point is made about the assumptions 'on page 42 of the Discourse on Method' [ A T VI41: C S M I131]. -261'non-empirical'. The natural reading of the text is: empirical information is a startingpoint for a breakthrough to understanding the order in heavenly bodies which underlies their apparent irregularity. And once such a theory is discovered, one could then explain a priori why terrestrial bodies are the way they appear to be. In other words, we would know the causes which explain the observed effects. 5 In more general terms, a Cartesian account of any physical phenomenon involves locating an appropriate description of the explicandum within a broader framework in such a way that the description is deducible (in a rather loose, Cartesian sense) from a description of parts of matter, their motions, and their interactions. The relevant description of the parts of matter may in turn be located within another more general framework, so that one views the logical structure of an explanation as something like the following: . . . Q 1. . . Q 2. . . Q 3. . . , where each Q represents a group of propositions, and the connections between them are Cartesian deductions. In this model, Q 3 can be explained a priori by reference to Q 2 , while Q 2 can be confirmed, a posteriori, by Q 3 . However, Q 2 may be merely postulated as an assumption, because one has not yet regressed as far as Q 1 in one's series of explanations, and therefore Q 2 cannot be said to be explained or demonstrated a priori. Despite its non-demonstrated status, Q 2 can still function to provide an a priori explanation of Q 3 . Although Q 3 is thus provided with an a priori explanation, there is no suggestion at all that it is known a priori in Kant's sense of the term. Descartes clarified this point in the letter to Plempius:

from the oblong and inflexible shape of the salt particles I deduced . . . as many other things as possible which are obvious to the senses. I wished to explain the latter by the former as effects by their causes, but not to prove them since they were already well known. ( A T I476 [ C S M K77]) Nor is there any implication that Q 2 (the explanans) is known a priori, either in Kant's sense or in Descartes'; as in the example quoted above about terrestrial bodies, the explanans may be discovered empirically and, in the case of the Dioptrics and Meteorology, it may not itself be provided with an (a priori) explanation. So that for Descartes, an a priori explanation is an account of a physical phenomenon in terms of more ____________________ 5 See Descartes to an unknown correspondent ( 1648?): 'I have by no means described all the movements of each planet in detail, in my Principles, but I assumed in general all those [movements] which observers have noticed in the planets, and I attempted to explain their causes' ( A T V259). -262simple and primitive concepts: namely, the concepts of the size, shape, and motions of the particles of matter. 6 The postulated Cartesian causes of physical phenomena must furthermore explain, not only that physical events are as they are observed to be, but also that they could not have been otherwise than they are. This feature was underlined in a letter to Mersenne in 1645, and it is close enough to the classical Aristotelian view in the Posterior Analytics to seem almost like a paraphrase: 'As far as Physics is concerned, I believed that I knew nothing at all if I could only say how things may be, without being able to prove that they could not be otherwise' ( 11 Mar. 1640 ( A T III39 [ C S M K145])). This point is linked to an objection from Morin to the effect that it is very easy to construct ad hoc hypothetical causes to explain any conceivable effect. Descartes responded to the objection as follows: Although there are truly many effects to which it is easy to match different causes, with one cause for each effect, it is nevertheless not so easy to fit one single cause to many different effects, if it is not in fact the true cause which produces them. ( Descartes to Morin, 13 July 1638 ( A T II199 [ C S M K107])) This raises the question of how one might come to know which hypothesis is the correct one; 6a however this is done, Descartes clearly intended that an acceptable scientific explanation involves specifying the unique or correct cause of any given phenomenon, and therefore an explanation of how the effect was completely determined by the hypothetical cause. Apart from latent connotations of a strictly deductive logic of explanation, this also suggests that an adequate scientific explanation must be able to

specify the actual mechanism by which the effect results from a cause, rather than merely to suggest a variety of possible ways in which this might happen. So far in this section I have ignored questions about the relationship between physical hypotheses and metaphysics, and have avoided the troublesome grey area where an explanans is no longer unambiguously a physical hypothesis. If we can continue this blinkered attitude for the present, it seems that a scientific explanation of something described by Q n involves Q n -1 , and likewise an a priori or causal explanation of Q n -1 ____________________ 6 In a letter to Mersenne, 15 Nov. 1638, Descartes explains what he means by an a priori knowledge of the lifting power of a machine: 'There is no other way . . . of knowing a priori the magnitude of this effect, namely which weight and what size weight can be lifted with such a machine, except by measuring the quantity of the action which causes the effect, i.e. of the force which must be employed for this purpose' ( A T II433 [ C S M K128]). Again, in this case, an a priori knowledge is a causal account of it. 6a These topics are dealt with by Clarke in Descartes' Philosophy of Science, ch. 6. -263involves reference to Q n -2 . The logic of specifying Q n -1 , or Q n -2 , etc. is, at least in part, hypothetico-deductive. 7 And besides describing the hypothetical cause of a given phenomenon, a physical explanation must also describe the mechanism by which the observed effect inevitably follows from the proposed cause. To the extent that the cause is unobservable, the mechanism by which it operates will also be unobservable, and therefore the description of this mechanism is unavoidably hypothetical. II. CLOCKS, CODES, AND HYPOTHESES The scientific essays of 1637, the Dioptrics and Meteorology, are unambiguously hypothetical. Descartes initiates each essay with various assumptions about matter in motion, which are then put to use in constructing explanations of specific phenomena such as reflection, refraction, or the rainbow. The logic of this approach is clarified by Descartes in part vi of the Discourse: 'since experience makes the majority of these effects very certain, the causes from which I deduce them are used not so much to prove them as to explain them; but, quite the contrary, it is the latter [the causes] which are proved by the former' ( A T VI76 [ C S M I 150]). 8 This led to further inquiries from readers, especially from Morin and Mersenne. 9 In reply to Mersenne, Descartes penned one of the clearest and best known of his attempts to analyse scientific explanations: You ask if I believe that what I wrote about refraction is a demonstration. And I think it is, at least in so far as it is possible to give a demonstration in this kind of study, without having first demonstrated the principles of

physics by metaphysics-a project which I hope to realise some day, but I have not yet done so--and also to the extent that any other question of mechanics, or of optics, or of astronomy, or any other question which is not purely geometrical or arithmetical, has ever been demonstrated. But to demand geometrical demonstrations from me in something which presupposes physics is to wish that I do the impossible. And if one wishes to call only the proofs of geometers 'demonstrations', then one must say that Archimedes never demonstrated anything in mechanics, nor Witelo in optics, nor Ptolemy in astronomy, etc.; but this is not what is said. For in these matters one is content if the ____________________ 7 Only in part, because there is a significant input from various empirical, methodological, and metaphysical assumptions already made by Descartes to what might count as an acceptable explanans. 8 See Descartes to Morin, 13 July 1638 (AT II 198 [CSMK 106-7]); Principles ( A T VIIIA 81 [ C S M I248-9]), where Descartes lists some of the phenomena he hopes to explain and then adds: 'not indeed so that we might use these phenomena as explanations to prove anything; for we wish to deduce the explanations of the effects from causes and not, on the contrary, the explanation of the causes from the effects'. 9 See Descartes to Morin, 22 Feb. 1638 ( A T I536). -264authors presuppose certain things which are not manifestly contrary to experience, and if the rest of the discussion is coherent and free from logical errors, even if their assumptions are not exactly true . . . If people say that they do not accept what I have written because I have deduced it from assumptions which are not proved, then they do not understand what they are asking, nor what they ought to ask for. ( A T II141-2, 143-4 [ C S M K103-4]) 10 Apart from the reference to providing a metaphysical demonstration for physics, this text clearly indicates that Descartes understands his procedure as a hypothetico-deductive one in which the acceptability of the hypotheses is at least partly determined by their explanatory success vis-àvis relevant explananda. This point was developed in a letter to Plempius, for Fromondus, in 1637. Fromondus had apparently objected to the logic of Descartes' arguments in the Meteorology, and in reply Descartes explained that each of his various hypothetical explanations could be rewritten in the form of a syllogism. 11 If O 1 (observed phenomenon), then probably E (likely explanation). But O 1 . Therefore, probably E.

Descartes gives a number of examples of this form of argument, and then adds: 'Although considered separately these only convince us with a certain probability, when they are all taken together they amount to a demonstration' ( A T I423 [ C S M K66]). Unfortunately, Descartes does not make the logic of his arguments entirely clear, because he fails to indicate the logical relationship within the major premiss of each syllogism between the antecedent and the consequent. In each example he claims that the observed effects are a sign (indicium est) of his proposed hypothetical cause. It is not clear, for example, how the relative viscosity of water and oil 'is a sign' that the parts of water are like eels and the parts of oil are branch-shaped. This way of expressing it makes it seem as if the experiential evidence implies the hypothesis; in fact, it is quite the reverse. His argument should read: assuming that the parts of water are like eels and that the parts of oil are branched, one would expect to find observed effects O 1 , O 2 , etc.--the antecedents of the major premiss of each syllogism. Since all of these observational statements are true, the success of a single hypothesis in ____________________ 10 Cf. Descartes to Plempius, 3 Oct. 1637, where he discusses his explanation of why a ray of light passes more easily through water than air, and then adds: 'That is all I wrote, and if I understand what a demonstration is, I have demonstrated it' ( A T I419). 11 'If he would be good enough to read with sufficient attention everything I wrote in the Meteorology and Dioptrics, he would find six hundred explanations from which six hundred syllogisms could be constructed to prove what I say' ( A T I422 [ C S M K65]). -265explaining a diversity of empirically known phenomena amounts to a demonstration of the hypothesis in question.Since explanations in physical science cannot aspire to providing more than hypotheses which conform to experience, it follows that the only way in which a proposed explanation can be faulted is: a. by showing that the initial assumptions are false; 12 b. by finding a logical error in the explanation; or c. by proving that the implications of the explanation are incompatible with experience. 13

An alternative model for the logic of explanations is found in the discussion of clocks and codes in the Principles. These analogies were introduced towards the conclusion of the Principles, partly to explain the way in which hypothetical causes must be postulated in scientific explanations, but also to clarify the kind of certainty which Descartes claims to have realized in his physics. Only the first of these considerations is at issue here. Article 204 of part iv of the Principles begins: 'It suffices in the case of imperceptible entities if one explains how they may be, even if they are not in fact as described.' The French edition adds: 'and that is all that Aristotle tried to do' ( A T VIIIA 327: A T IXB 322 [ C S M I289]). Just as a watchmaker could construct two watches which were

externally similar and equally accurate in keeping time, but with very different internal mechanisms, so, likewise, God could have chosen a variety of causal mechanisms to produce the observable effects which we can observe in nature. The implication here is that the scientist cannot unlock the causal mechanisms at work in nature by direct inspection, and he must be content with postulating the existence of some hidden causes which adequately explain observable physical phenomena. The French text makes the implication explicitly: It is likewise certain that God has an infinity of different ways by each of which he could have arranged that everything in this world appears as it does at present, without it being possible for the human mind to know which of these many ways he chose to follow. ( A T IXB 322) ____________________ 12 Cf. Gilson, Études sur le rôle de la pensée médiévale, 128-37. 13 The three conditions are given in a slightly different form in the third part of the Principles. 'And certainly, if we use no principles except those which are most evident, if we deduce nothing from them unless by mathematical reasoning, and if those things we deduce from them in this way agree with all the phenomena of Nature, we would seem to insult God if we suspect that the causes of things which we discovered in this way were false' ( A T VIIIA 99 [ C S M I255]). Cf. ibid. 101, where Descartes claims that he may assume any hypotheses which are suitable for explaining nature 'on condition that all those things which are deducible from them agree with experience'. -266Conceding this point, Descartes continues: And I would believe that I have done enough if the causes which I have explained are such that all the effects which they could produce are similar to those we observe in fact in the world, without inquiring further if it was by these causes or by some others that they were in fact produced. . . . For medicine, mechanics, and in general all those arts for which physical science is useful have no other objective apart from applying some perceptible bodies to others so that, by a series of natural causes, certain observable effects follow. (Ibid.) This text at least acknowledges that physical science depends on postulating causes which could explain observable effects; it also contains strong suggestions of a crude instrumentalism which Descartes borrows from the saving-the-phenomena tradition. In fact, however, this is not Descartes' own position; he is merely using references to this tradition to excuse his apparent failure to achieve the kind of certainty he promised at the outset of his career. 14 In the subsequent article ( 205 ), Descartes appeals to the analogy of decoding a message to illustrate both the method and the kind of certainty one can expect in physics. If

someone systematically substitutes some of the letters of the alphabet for others in a code written in ordinary language, and if the substitution makes sense of the original, he might justifiably claim to have cracked the code; likewise, If one considers how many different properties of magnets, of fire, and of all the other things in the world have been deduced very clearly from a very small number of causes which I proposed at the beginning of this treatise, even if one imagines that I assumed them by chance and without any reason . . . he will have as much reason to believe that they are the true causes . . . For the number of letters in the alphabet is much greater than the number of first causes which I assumed. ( A T IXB 323-4 [ C S M I290]) There is a more obvious concern here with establishing the certainty of Cartesian hypotheses than with acknowledging the hypothetical procedure which introduces them in the first place; but the analogy speaks for itself with sufficient clarity to suggest that the physicist is comparable to the code-breaker in postulating whatever is necessary-hypothetical causes for the scientist--to explain a given phenomenon successfully. Although hypotheses are unavoidable in science, conjectures or ____________________ 14 Cf. A T IXB 185: 'As regards these particular effects, of which we have insufficient empirical evidence to determine the true causes which produce them, we should be content to know some causes by means of which they could have been produced'; a letter (to Huygens?), 5 Oct. 1646: 'in such matters in which one has not done many experiments, it suffices to imagine some cause which could produce the effect in question, even if it could also be produced by another cause and one does not know which is the true cause' ( A T IV516). [These topics are dealt with by Clarke in Descartes' Philosophy of Science, ch. 6.] -267unfounded guesses are unacceptable. In the Fifth Replies to Objections the attempt to guess at God's purpose in creating the universe is called a conjecture: And although in ethics, where it is often acceptable to use conjectures, it is sometimes a mark of piety to guess what objective God has in mind in ruling the universe, it is certainly inappropriate in physics where everything should be based on most secure reasons. ( A T VII375 [ C S M II258]) It is not clear whether the conjectural status of an unconfirmed hypothesis depends on either of these two factors: a. The hypothesis has no initial plausibility, prior to further examination or testing. This initial plausibility might derive from its intrinsic reasonableness, its simplicity, its

b.

coherence with a structure of scientific explanations, or even from some empirical evidence in its favour. The hypothesis is such that, once adopted for examination as a possible explanation, there is no further evidence available in favour of its truth, or perhaps there could never be any such evidence which could decide for or against it.

Whereas in the discussion of God's purpose in creation the second factor above seems to be uppermost in Descartes' mind, the passage at the conclusion of the Principles quoted above suggests that (a) is sufficient to reduce a hypothesis to a mere conjecture: 'even if one allows that I assumed them by chance and without any reason'. This implies that before any evidence is proposed in favour of a hypothesis, one is dealing with nothing more than an unfounded guess or mere conjecture, and this has no permanent place in physical science. Therefore Descartes can both exclude guesswork from physical science and at the same time endorse hypotheses if either of the following conditions is satisfied: an assumption has some initial plausibility, either because of favourable empirical evidence, its simplicity, etc.; or an assumption is corroborated by experience or by reason, after its initial adoption for further examination. In this way it is possible for a mere conjecture to become a plausible hypothesis as a result of corroborating evidence, and this is the minimal claim being made for the general assumptions which were adopted at the beginning of the Principles. Even if they began life as mere conjectures, the evidence introduced later in the text redeemed them as acceptable physical hypotheses. III. CONSTRAINTS ON HYPOTHESES One could hardly conceive of Descartes approaching some problem in physical science with a completely open mind as to what may be legiti-268mately postulated to provide an acceptable explanation. Some theoretical entities are excluded a priori; more importantly, there is an implicit but significant complicity on Descartes' part with a kind of crude empiricism which avoids the theoretical in favour of scaled-down pictures of the objects of our ordinary perception. So that parallel to the acknowledgement that a scientific explanation must inevitably be hypothetical, Descartes relies on a number of metaphysical and methodological assumptions to specify further the range of entities which may figure in hypotheses. It is at this point that metaphysics has a decisive influence on the content of Cartesian science. As is well known, Descartes rejected scholastic forms as nonexplanatory. Le Monde, chapter 5, rejects such explanatory qualities as heat, cold, etc., because 'these qualities seem to me to need explanation themselves' ( A T XI25-6 [ C S M I123]). The Discourse more generally 'expressly assumes' none of the 'forms or qualities which are disputed about in the schools' ( A T VI42-3 [ C S M I89]). 15 It is consistent with this rejection of scholastic explanations that Descartes is not concerned to provide any explanation of the substances of physical phenomena, in so far as substances are distinct from physical

properties: 'this substantial form of the sun, in so far as it is distinct from those qualities which are found in its physical nature (matière), is once again a philosophical entity which is unknown to me' ( Descartes to Morin, 12 Sept. 1638, A T II 367 [ C S M K122]; cf. ibid. 364, for a similar remark about the 'form' of motion). The suggestion that scholastic forms are themselves in need of explanation perhaps implies that Descartes understood scholastic explanations as redescriptions--in an esoteric language--of explananda rather than as genuine explanations. 16 Whether or not he would have accepted this interpretation of his comments, it is unquestionable that his proposed alternative was to look for the efficient and material causes of physical phenomena, and that formal and teleological causes had no place in his science. The question then arises as to how he could so easily specify, apparently a priori, what could count as a cause in the kinds of explanation which he was willing to accept. The suggestion which is most often made at this juncture is that Descartes was attempting to conflate distinctions between physical and mathematical sciences, and that he therefore hoped to explain all physical phenomena in terms of the geometrically specifiable ____________________ 15 See also A T VI61-2 [ C S M I142-3]. 16 In a letter to Huygens, Mar. or Apr. 1638, Descartes gave as an example of a nonexplanation: 'lux est medium proportionale inter substantiam et accidens' [light is a proportional medium between substance and accident'] ( A T II51). It would be difficult to resist agreeing with him! -269features of bodies in motion. I think this suggestion has some validity; but paradoxically, a more influential factor is Descartes' fundamental empiricism. Descartes' most comprehensive discussions of the relevance of mathematics as a paradigm of scientific methodology are found in the Regulae and the Discourse. Descartes was evidently enamoured of the certainty of mathematics and its relative progress as a science, compared to philosophical confusion, and there is no doubt that he recommended some adjustments in philosophical and scientific method to exploit what he considered to be the source of the mathematicians' success. On the other hand, there is ample evidence, especially in the correspondence after 1630, that Descartes had lost interest in pure mathematics; so that one needs to be especially cautious in interpreting references to 'mathematical' method in Descartes' mature physical science. For example, he wrote to Mersenne, 15 April, 1630: As far as the [mathematical] problems are concerned, I will send you a million of them to set for others if you wish; but I am so tired of mathematics and take so little account of them now that I would hardly take the trouble to solve the problems myself. ( A T I139)

The same sentiment is clearly expressed in writing to Mersenne, in October or November 1631; to Stampioen in 1633; to Mersenne in March and September 1638: 'Please do not expect anything else from me in Geometry; for you know that for a long time I've protested that I do not wish to work at it, and I honestly think that I can give it up completely' ( A T II 361-2). 17 Descartes is obviously reluctant to spend time with purely formal, or abstract, mathematical problems. 'To tell you the truth, I am so tired of abstract mathematics that I can no longer work at it' (to Mersenne, A T II 507). The disillusionment with abstract mathematics is reflected in the Discourse in terms of a distinction between pure and applied mathematics ( A T VI17-18 [ C S M I119-20]). When his correspondents expressed surprise at the change of heart (for example, M. Desargues), Descartes replies: I have decided to abandon only abstract geometry . . . and the reason for this is because I will have so much more time to cultivate a different type of geometry, which is concerned with explaining the phenomena of Nature. For if he cares to consider what I wrote about salt, the snow, and the rainbow, he will quickly realise that all my physics is nothing else but geometry. ( A T II268 [ C S M K118-19]) ____________________ 17 Descartes to Mersenne, Oct. or Nov. 1631 ( A T I229-30); to Stampioen, 1633 ( A T I275); to Mersenne, 31 Mar. 1638 ( A T II95); to Mersenne, 1 Apr. 1640 (?), ( A T III50); to Dozem, 25 Mar. 1642 ( A T III534); the passage quoted is from 12 Sept. 1638 ( A T II361-2). -270The last line here is partly an attempt to placate an importunate correspondent. It is just as much a reflection of the standard distinction between pure and applied mathematics, the latter of which included music, astronomy, and optics. 18 Descartes speaks of two kinds of geometry, (rather than of mathematics), where 'geometry' is understood as 'a science which, in general, teaches one to know the measures of all bodies' ( A T VI 389; cf. ibid. 392). In other words, any exact science is part of applied mathematics. The correspondence after 1630 thus shows a constant disillusionment with purely formal mathematics. This is consistent with an established and predominant interest in the physical sciences and, more confusingly, with Descartes' use of the term 'mathematics' to refer to any scientific knowledge which is subject to mathematical treatment. This wider use of the term survives in a number of places where a physical explanation is called mathematical. And surely, if we use only those principles which seem to be most evident, if we deduce nothing from them unless by mathematical arguments, and if we find that those things we have thus deduced from them correspond accurately with all the phenomena of nature, then we would seem to insult

God if we suspect that the causes which we have thereby discovered are false. ( A T VIIIA 99 [ C S M I255]) Evidently Descartes was impressed with the rigour and certainty of mathematics, and he wished to emulate these features of the discipline in all scientific knowledge. But rather than reduce physics to abstract mathematics, he planned to minimize uncertainty by rather stringent limitations on the conceptual resources of physical science. The assumption of a viable distinction between primary and secondary qualities partly determined this move; those qualities of physical phenomena which were assumed to be a function of the perceiver's response could be reduced, within a comprehensive account of both the perceiver and what is perceived, to qualities which were not explicable in the same way. In the language of Le Monde, such qualities could not be used to explain anything, because they were in need of explanation themselves. When Descartes is eventually faced with specifying what kind of qualities could count as explanatory, there is a noticeable absence of arguments for his choice. There is an obvious advantage in small parts ____________________ 18 Cf. A T VI19-20 [ C S M I120-1] and Gilson, Discours, 216-17. For Descartes' distinction of geometry and mechanics, cf. A T VI389 and 392. Pierre Boutroux discusses Descartes' attitude towards abstract mathematics in L'Imagination et les mathématiques selon Descartes, Université de Paris, Bibliothèque de la Faculté des Lettres, no. 10 ( Paris: Alcan, 1910), 34; see also P. Golliet, "'Le Problème de la méthode chez Descartes'", Revue des sciences humaines, 61 ( 1951), 56-73, at 62. -271of matter exchanging motion through contact action--for this kind of explanation is subject to mathematical control. But even here the promise of a rigorous presentation of the explanans is more than can be realized; for Descartes recognizes that he has no way of dealing with the multiplicity of factors which determine the results of collisions between moving particles. In default of such a mathematically exact description of the parts of matter in motion, Descartes falls back on ordinary experience as a source of explanatory concepts. This is what is meant in the passage from Le Monde, where matter is said 'to contain nothing which is not so perfectly known to the reader that you could not even pretend not to know it' ( A T XI35 [ C S M I91]). The Discourse summary of the procedure in Le Monde accurately reflects the same point: Thus, I first described this matter and tried to represent it in such a way that there is nothing in the world, it seems to me, which is clearer or more intelligible, except what was earlier written about God and the soul; for I explicitly assumed that it had none of those forms or qualities which are disputed in the schools nor, in general, anything the knowledge of which is

not so natural to our minds, that one could not even pretend not to know it. ( A T VI42-3 [ C S M I132]) And the kinds of qualities which are such that one could hardly pretend not to be acquainted with them are precisely those observable characteristics of bodies such as motion, size, etc. which are known through experience. The limitations imposed by this kind of conceptual empiricism only become apparent when the explanation of specific phenomena is examined. Thus, the concept of attraction at a distance would seem to be ruled out in advance by a good Cartesian; gravity must be explicable by reference to impact action. Magnetism is explained by reference to eel-like particles; muscular action is the result of contact action; in fact, no other explanatory concept is considered at a later stage of the scientific project apart from those which are initially admitted. This feature of Cartesian science is too often absorbed by the presumed dominance of mathematics and mechanics as an explanatory model. But when it is considered in conjunction with the Cartesian reliance on models, the very rough sense in which the conceivability of these models can be thought to confirm a hypothesis, and the emphasis in both the Regulae and the Discourse on the reduction of complex problems to the simple and familiar, it may be seen just as easily as pointing towards a kind of crude empiricism in Descartes' conceptual scheme. Physical reality is much too complex to be amenable to mathematical description; there are too many factors in every physical situation to be brought within the scope of a mathematical formula. On the other hand, the theoretical -272concepts introduced by the scholastics were considered to be suspect from the point of view of a viable explanation. For want of anything better, Descartes decided to restrict the conceptual tolerance of theories to those concepts which are known from our ordinary experience of physical bodies, and to structure theories along the lines dictated by crude models of colliding pieces of matter which are described in non-theoretical language. IV. MODELS There is little doubt that Descartes relies extensively on models and analogies in articulating scientific theories; he also makes the much stronger claim that the possibility of constructing a model is a sine qua non condition for explanations in physics. I claim that they [i.e. models and analogies] are the most appropriate way available to the human mind for explaining the truth about questions in physics; to such an extent that, if one assumes something about nature which cannot be explained by any analogy (comparaison), I think that I have conclusively shown that it is false. ( To Morin, 12 Sept. 1638 ( AT II368 [ CSMK122])) A similar sentiment is expressed in writing to Plempius, 3 October, 1637:

There is nothing more in keeping with reason than that we judge about those things which we do not perceive, because of their small size, by comparison and contrast with those which we see (ad exemplum et similitudinem eorum quae videmus). ( AT II421 [ CSMK64]) These reflections on method coincide with Cartesian practice. As early as the Regulae ( AT X395 [ CSM I29]) the nature of light was to be explained, if necessary, by analogy with other natural powers which were already known. And the construction of a model universe in Le Monde is introduced as a 'fable' ( AT XI31 [ CSM I90]). Nor is it plausible to argue that the switch to a model was a move on Descartes' part to avoid the kinds of difficulties encountered by Galileo--for Descartes talks about a 'fable' long before Galileo's Inquisition problems ( 25 Nov. 1630 ( AT I179 [ CSMK28])). In the second part of Le Monde, the 'Treatise on Man', there is a similar move to construct a model of how the body operates and to understand the body's functioning by analogy with a machine. So that there are frequent references to the machine of the body ( AT XI120, 141 [ CSM199, 101]), and the relationship between model and reality is one of 'representation' ( AT XI173 [ CSM I 104 ]). -273The most widely cited examples of models are the three 'comparaisons' which are introduced at the beginning of the Dioptrics. As in the more general discussion in Le Monde, the author refuses to speculate about the true nature of light and substitutes a hypothetical approach which is comparable to the procedure of astronomers: But, since I have no need to talk about light in this context except in order to explain how rays of light enter the eye and how they may be turned by the various bodies which they encounter, it is not necessary that I undertake to say truly what its nature is; and I believe that it will be enough if I use two or three analogies (comparaisons) which help us understand it in a way which is most convenient to explain all its properties which experience acquaints us with, and to subsequently deduce all those other properties which cannot be so easily noticed; in this approach [we are] imitating astronomers who, although their assumptions are almost entirely false or uncertain, nevertheless, because they agree with various observations which they have made, do not fail to draw many very true and very certain conclusions from them. ( AT VI83 [ CSM I153]) Three models 19 are then introduced, in turn, before the sine law is discussed: a. a blind person's use of a walking-stick as an analogy for the propagation of light by a 'force' or 'pressure' which eventually impinges on our eyes; b. the example of wine leaking from a vat, as an analogue for the linear transmission of light; and c. a tennis-ball moving towards a thin sheet of material which it punctures on impact, and then continues its motion with reduced speed. 20

The models are recognized to have some negative analogies also; 21 for example, after introducing the comparison between our perception of light and the blind person's perception of physical objects, Descartes adds: 'But, because there is a great difference between this blind man's stick and the air or other transparent bodies, through the medium of which we see, I must use another analogy at this stage' ( AT VI86 [ CSM I 154 ]). Likewise, in the case of the tennis-ball analogy, the weight, size, and shape of the ball are excluded as irrelevant to the model, so that the change in direction ____________________ 19 The word 'modèle' is often used by Descartes, as in 'modèle de bois'-- Descartes to Huygens, 25 Jan. 1638 ( AT I505), and 8 Feb. 1638 ( AT I520)--and in 'modèle de cuivre', Descartes to Ferrier, 13 Nov. 1629 ( AT I55). The word is also used to mean a diagram, as in AT VI215, 217, 224. 20 For earlier uses of the same analogy, see Sabra, Theories of Light from Descartes to Newton, 93-9. 21 For the terminology of negative analogies, see Hesse, Models and Analogies in Science. -274resulting from impact with a permeable surface is highlighted as of primary concern ( AT VI94, 99 [ CSM I157, 160]). 22 The model-building feature of Cartesian science is much more pervasive than these few examples suggest; it would hardly be an exaggeration to say that Descartes' whole scientific project is one of imaginatively constructing descriptions of the motions of particles which might explain natural phenomena and our experience of them. The 'real world' is abandoned with unusual haste in favour of the imaginary world in almost every attempt to explain. The ease with which this is done and the value of the resulting discussions were challenged by Morin, in a letter of 12 August 1638: Problems in physics can rarely be resolved by analogies (comparaisons); there is almost always some difference [between model and reality], or ambiguity, or some element of the obscure being explained by the more obscure (obscurum per obscurius). ( AT II291) Descartes' reply to this was not only impenitent, but boldly steadfast in defending the analogical method in explaining the motions of particles in the transmission of light: And I was right in using these observable spheres to explain their turning motion, rather than the parts of subtle matter which are invisible, so as to submit my explanations (raisons) to the test of the senses, just as I always try to do. ( AT II 366) Despite this methodological claim, the proposed explanation was not in fact subjected to empirical test, and Morin rightly objected on this precise point in a later reply; 23 what

was done was that the invisible boules of subtle matter were compared with the wooden models which could be imagined more easily, without any experience of the characteristic motions of such model boules, and without any attempt on Descartes' part to do appropriate experiments to test his theory. ____________________ 22 A similar recognition of disanalogy is found in Descartes to Mersenne, 27 July 1638. Descartes had tried to explain the colour of light in terms of the rotation of small spheres; he then adds the qualification: 'the spheres which are painted in the figure on this page serve only as an example, and should be taken to be spheres of wood or some other visible material, and not parts of subtle matter' ( AT II269). Cf. Descartes to Mersenne, 28 Oct. 1640: 'it is for this reason that I assumed, in my Dioptrics, that the surface and the ball are perfectly hard, and that the ball has neither weight, nor size (grosseur), etc. in order to make my demonstration mathematical [i.e. exact]. For I am well aware that the reflection of an ordinary ball never occurs at exactly equal angles' ( AT III208). 23 Morin had objected to this defence in a letter of Oct. 1638, to which Descartes seems not to have responded. 'There is no one on earth who could perform the experiment you mention on wooden spheres . . . why have you not explained the motions which are characteristic of the spheres of subtle matter . . . neither one nor the other of these spheres can be experienced' ( AT II418-19). -275The defence of this analogical approach is interesting. A viable explanation of any natural phenomenon can be constructed in terms of the size and motions of its parts. Therefore one can always construct a large-scale model of whatever is happening at the micro-level, because the only disanalogy involved is one of size; in fact, the concepts used to describe the properties of the small particles are themselves parasitic on our everyday language for talking about the medium-sized physical objects of ordinary experience: in the analogies I use, I only compare some movements with others, or some shapes with others, etc.; that is to say, I compare those things which because of their small size are not accessible to our senses with those which are, and which do not differ from the former more than a large circle differs from a small one. ( AT II367-8 [ CSMK122]; my italics) The central place of models, as scaled-up versions of micro-events without any corresponding change in the concepts which are used to describe both model and explanandum, is consistent both with Descartes' peculiar procedure for confirming explanations which are formulated by reference to a model, and with the even more surprising thesis that any model--even if not the 'true' one--is preferable to none at all in explaining physical phenomena. One usually thinks of scientific confirmation or corroboration as establishing the truth of propositions which have a direct bearing on the truthvalue of a hypothesis. What one

normally expects to find is evidence which implies the truth of a hypothesis, or evidence which is consistent with it and inconsistent with its likely alternatives. Descartes stretches the already tenuous link between hypothesis and confirming evidence by substituting the confirmation of claims about models for the confirmation of claims about the explanandum being modelled. This is an understandable move if one believes, as he apparently does, that there is no significant difference between the model and the reality apart from obvious disanalogies of size which are already allowed for. Thus, for example, one finds the case in which heating snow in a container confirms a theory about clouds ('cecy est aysé a experimenter en la neige' ( AT VI292)); burning earth and other chemicals together confirms a hypothetical explanation of lightning ('ainsi qu'on peut voir, par experience' ( AT VI320)); shaking beads on a plate tests a theory about ice ('ainsi que vous pourres voir par experience' ( AT VI288)); or the opacity of sunspots corroborates a theory about the origin of the earth's surface ( AT VIIIA 205). When asked how the subtle matter surrounding the earth can push heavy bodies towards the centre, he replies: get a vessel with heavy and light particles mixed together, and spin it. -276The heavy ones displace the light ones towards the centre. ( Descartes to Mersenne, 16 Oct. 1639 ( AT II593-4).In each of these cases, and in many others which could be quoted, Descartes corroborates a description of a model rather than introduce evidence which bears directly on the hypothesis at issue. The logic of the argument seems to be: 1. To explain phenomenon ϕ, under the description D, construct hypothesis(es) H. 2. ϕ is similar to ϕ 1. 3. To explain ϕ1, under description D1, construct hypothesis H1. 4. H1 implies evidence or effects E. 5. E is true. 6. Therefore H1 is probable. 7. Therefore H is also probable. The plausibility of the move from (6) to (7) is a function of the similarity or otherwise of ϕ and ϕ 1, or, in other words, of the appropriateness of the model adopted. One is initially surprised at the logic of this type of argument--for it provides little warrant for the conclusion. The simplest explanation for the poor logic is to say that Descartes is a sloppy thinker in matters physical-and this is unquestionably true. A second reason is the one already mentioned: that prior methodological assumptions about what counts as an explanation have underestimated the significance of possible disanalogies between particles of subtle matter and the everyday models which Descartes prefers to discuss. There is one further reason, however, which helps reinforce an otherwise out-of-character reference on Descartes' part to constructing likely stories: where an ideal, demonstrative, and fully guaranteed explanation is not available, the Cartesian is apparently willing to settle for a plausible reconstruction which may not accurately reflect how matters actually stand in the real world.

This represents a further weakening of the concession already made by Descartes that, since explanation in the physical sciences is explanation of the observable in terms of non-observable causes, the construction of viable explanations is unavoidably hypothetical; that concession still held out the hope that the resulting uncertainty could be effectively minimized by discovering a small number of hypothetical causes which would explain a wide variety of disparate effects. Whether or not this ideal can be realized in practice is something I discuss elsewhere, 23a in connection with ____________________ 23a Clarke, Descartes' Philosophy of Science, ch. 6. -277Descartes' views on confirmation. Even if it cannot be realized, there is a persistent suggestion that when nothing better is available, any plausible hypothesis at all will do in physics. The distinction between 'true' hypotheses and those which are acceptable despite their presumed falsehood emerges most explicitly in the Cartesian account of the evolution of the Earth. Descartes seems to have believed--or at least claims to have believed--that the Earth was created by God, in an instant, in a fully developed state. On the other hand, he was, proposing an account of a model world which might have gradually evolved from chaos to the structured natural universe we observe today: Nevertheless I did not wish to conclude from all these things that the world was created in the way which I proposed; for it is much more likely that God made it as it should be, from the beginning. But it is certain--and this is an opinion which is commonly accepted among theologians--that the action whereby he now conserves the universe is exactly the same as that by which he created it. So that even if he had given it, from the beginning, no other form apart from chaos, provided that he established the laws of nature and gave it his concurrence to act in the way it usually does, one could believe, without denigrating the miracle of creation, that by that act alone all purely material things could have developed with time into the condition we see them in today. And their nature is much easier to understand if one sees them develop gradually in this way rather than when one thinks of them in their complete state. ( AT VI45 [ CSM I1334]) 24 One could construe this passage as a phony compromise between genuinely held philosophical beliefs and apparently accepted religious dogma. It seems more likely that Descartes had a genuine problem here, and that he was in fact modifying his concept of explanation in the light of irreconcilable difficulties with theology. Since the world was created in an instant by God, and one cannot understand God's creative action, the only alternative is to imagine a way in which the instantaneously implemented creative act could have been accomplished by a more gradual and autonomous 25 structuring of the

physical universe under the control of the laws of nature. So that an explanation of a physical event is not necessarily an account of how it happened historically; it is rather an account of how it could have happened which is consistent with one's prior methodological and physical principles. This point is taken up again at greater length in the Principles: There is no doubt that the world was created with all of its perfection from the very beginning . . . This is what the Christian religion teaches us, and natural reason ____________________ 24 Cf. Principles, part iii, arts. 43, 44 ( AT VIIIA 99). 25 The term 'autonomous' means in this context: that God does not need to intervene directly in the evolution of the universe once he established the laws of nature and imparted an initial measure of motion to the parts of matter. -278likewise convinces us of the same. For once we take account of the omnipotence of God we must conclude that whatever he created would have had its entire perfection from the beginning. Nevertheless, to understand the nature of plants or of man, it is much better to consider how they can gradually develop from seeds, than to consider how they were created by God at the beginning of the universe. Thus, if we can think of a few very simple and easily known principles from which we can show that the stars and the earth, and everything else we can observe on earth, could have developed as if from seeds--although we know they did not in fact develop in this way--we could explain their nature much better in this way than if we simply described them as they are now, [added in French] or how we believe they were created. ( AT VIIIA 99-100: IXB 123-4 [ CSM I256]). In other words, a scientific explanation of a physical phenomenon is equivalent to an account of how it could have happened as a result of the normal influence of the laws of nature. Once this distinction between the historically 'true' and the scientifically viable is adopted--perhaps initially because of the need to compromise for religious reasons--it is difficult to contain the latent instrumentalism which begins to colour much less obvious cases. For example, astronomical hypotheses are canvassed in the Principles not with a view to their truth, but with an eye on their explanatory value: 'Three different hypotheses, that is suggestions, have been discovered by astronomers, which are considered not as if they were true, but merely as suitable for explaining the phenomena' ( AT VIIIA 85 [ CSM I250]). The alternative Cartesian theory was proposed 'merely as an hypothesis and not as the truth of the matter' ( ibid. 86 ). In fact, Descartes advises the

reader to understand all the explanations which are to follow the discussion in part ii of the Principles as hypotheses: I wish everything I write from this point forward to be regarded as an hypothesis. Even if they are thought to be false, I think it will have been worth while if everything I deduce from these hypotheses agrees with experience. For we can see that they are as useful for life as the knowledge of the truth itself, [added in French] for one can use them in the same way to arrange natural causes to produce whatever effects one wishes. ( AT VIIIA 99: IXB 123 [ CSM I255]) There is no suggestion at this stage that the general principles themselves are merely hypotheses in this sense; nor that the facts to be explained are in some sense uncertain because they are empirically known. The status of the general principles is already agreed; and the empirical basis of our knowledge of explananda in no way compromises its certainty. What is at issue is the status of the intermediate claims which relate general principles of physics to the descriptions of specific phenomena: P-principles . . . observational judgements -279The Cartesian view of the dotted line is that, at least initially, it cannot hope to be anything but a plausible story which is consistent both with the P-principles and with the description of observed explananda. It is precisely at this juncture that the Cartesian reliance on models is most evident. For in physical explanations where one has insufficient experimental evidence, 'it suffices to imagine a cause which could produce the effect in question, even if it could have been produced by other causes and we do not know which is the true cause' (to an unknown correspondent, AT IV516). The Principles reflects this moderate position: As far as particular effects are concerned, whenever we lack sufficient expériences to determine their true causes, we should be content to know some causes by which they could have been produced. ( AT IXB 185) I believe that I have done enough if the causes which I have explained are such that all the effects which they could produce are found to be similar to those we see in the world, without inquiring whether they were in fact produced by those or by some other causes. ( AT IXB 322) One could not automatically conclude from these concessions that Descartes is willing to settle for a pervasive uncertainty in science. There are hints even here that the remaining uncertainty could perhaps be diminished by further empirical evidence. Whatever one decides on that point, it remains that a plausible or hypothetical account can constitute an explanation, in physical science, as long as it is 'deduced' from Cartesian P-principles and

is consistent with our experience. One of the primary functions of models is to facilitate the construction of such plausible accounts, and thereby complete the otherwise unfinished story of how very general and simple principles can explain the natural occurrence of specific physical phenomena. -280XIII FORCE (GOD) IN DESCARTES' PHYSICS GARY C. HATFIELD It is difficult to evaluate the role of activity--of force or of that which has causal efficacy-in Descartes' natural philosophy. On the one hand, Descartes claims to include in his natural philosophy only that which can be described geometrically, which amounts to matter (extended substance) in motion (where this motion is described kinematically). 1 Yet, on the other hand, rigorous adherence to a purely geometrical description of matter in motion would make it difficult to account for the interactions among the particles that constitute Descartes' universe 2 since the notions of extension and kinematical motion do not in themselves imply any causal agency. There is, after all, no reason to expect that a particle whose single essence is extension, even if we suppose it to be moving, should impart motion to another particle, while conversely, there is no reason to expect a resting particle to hinder the motion of an impinging one. It is not difficult to locate the source of this problem. Descartes desired to divest the material world (or rather our explanation of it) of the 'real qualities' of the scholastics as well as the active, vital spirits and occult qualities of Renaissance natural philosophy. 3 All natural phenomena were ____________________ Gary Hatfield, "'Force (God) in Descartes' Philosophy'" from Studies in the History and Philosophy of Science, Vol. 10 ( 1979), 113-40, by permission of Elsevier Science Ltd, Oxford, England. 1 That it was Descartes' intention to include nothing in his physics that could not be described geometrically is well known. In his letter to Mersenne of 27 July 1638. Descartes wrote: 'toute ma Physique n'est autre chose que Géometrie' ( A T II268 [ C S M K228]. Cf. Principles of Philosophy, part ii, art. 64. 2 After having asserted the particulate nature of matter and the laws governing its motion in part ii of the Principles, Descartes says, in article 1 of part iii, that we must now consider whether the established principles concerning material objects 'suffice to explain all natural phenomena' ( A G222). While the desired explanations are to be given in terms of size, shape, and motion (geometry and kinematics), the principles or laws of nature that govern motion involve concepts transcending geometry, as we shall see. 3 Descartes' scorn for the 'real qualities' of the scholastics is well known (see e.g. Descartes to Morin, 13 July 1638, A T III199-200). For his exclusion of vital

principles from physiology, -281to be accounted for in terms of an austere ontology of matter in motion. Yet such austerity is in danger of excluding causal agency. Perhaps, in overturning the conception of a living world so prevalent among his predecessors and contemporaries, Descartes forced himself into a position from which he could not allow any causal agency whatsoever into his explanation of the world. 4 To the modern reader it may seem obvious that Descartes did not get himself into such a fix, since matter in motion so readily reminds us of kinetic energy or of some more primitive notion of force. Yet by no means is it obvious that Descartes attributed causal efficacy to matter in motion per se. Serious scholars have held opposite positions on this issue. Westfall, Gabbey, and others 5 have argued that although on the metaphysical plane Descartes attempted to eliminate force from his mechanical universe, none the less 'force is a real feature of the mechanical world which Descartes presents to us'. 6 The thrust of this view is that Descartes conceived of force as 'the capacity of a body in motion to act', 7 by means of impact, upon other bodies. An opposing interpretation is that Descartes did in fact deny causal agency to moving matter per se, restricting agency to immaterial substances such as the human mind, angels, and God. The intention of this interpretation is not to de-emphasize the role of matter in motion in Descartes' explanation of nature, but rather to stress the fact that ____________________ see Traité de l'Homme (AT XI 202 [CSM I 108]); Descartes to Regius, May 1641 ( K102 [ C S M K 182]); and Hall, introduction to his translation of L'Homme, Treatise of Man, pp. xxvi-xxx. For his denial of occult qualities, see Principles, part iv, art. 187. Cf. R. Lenoble, Mersenne, ou La Naissance du Mécanisme ( Paris: Vrin, 1943), 6-13. 4 This seeming contradiction in Descartes' thought has been marked by commentators past and present: N. Kemp Smith, Studies in the Cartesian Philosophy ( New York: Russell & Russell, 1962, 1st pub. 1902), 75-9; Leon Pearl, Descartes ( Boston: Twayne, 1977), 206-9. The problem of accounting for the activity of matter was pervasive in the seventeenth century; see e.g. Robert Boyle , Of the Excellency and Ground of the Corpuscular or Mechanical Philosophy ( 1694), in M. B. Hall, Robert Boyle on Natural Philosophy ( Bloomington, Ind.: Indiana University Press, 1966), 187-209; Locke, Essay Concerning Human Understanding, II. 21. 2; J. E. McGuire, "'Force, Active Principles, and Newton's Invisible Realm", Ambix, 15 ( 1968), 145208; and the remarks of Leibniz discussed below. 5 Richard S. Westfall, Force in Newton's Physics: The Science of Dynamics in the Seventeenth Century ( New York: American Elsevier, 1971), ch. 2; Gabbey, 'Force and Inertia in Seventeenth-Century Dynamics'; Desmond M. Clarke, "'The Impact Rules of Descartes' Physics'", Isis, 68 ( 1977), 55-66; and Thomas L. Prendergast, "'Motion, Action and Agency in Descartes' Physics'", Journal of the History of

Philosophy, 13 ( 1975), 453-62, take positions similar to that of Westfall and Gabbey. Gabbey, "'Force and Inertia'", 8. Cf. ibid. 9, where Gabbey claims that for Descartes 'forces, whether of motion or of rest, are causes in their own right, distinct from motion and rest, and are in effect real properties belonging to body'. One wonders whether a philosopher such as Descartes, who worried so much over his metaphysics, would allow real properties into his ontology 'in effect'. 7 Westfall, Force in Newton's Physics, 64. 6

-282Descartes did not conceive of this moving matter dynamically. This second view has heretofore been proffered in summary accounts of Descartes' position, with the benefit of little or no supporting evidence. 8 As I shall show, the required evidence is available. The position I will defend can be better understood once the usage of the term 'dynamics' is clarified. The word itself was coined by Leibniz, who used it to name his new 'science of force'. 9 Leibniz denied the Cartesian view that extension is the single essence of material substance. If extension were truly the essence of body, he argued, then bodies should be completely indifferent to motion or rest, so that a moving body should be able to carry along all the resting bodies that it encounters without the slightest diminution in its own velocity. But bodies exhibit natural inertia, i.e. a resistance to being put into motion. Whence it follows: 'There is in matter something else than the purely Geometrical, that is, than just extension and bare change. And in considering the matter closely we perceive that we must add to them some higher or metaphysical notion, namely, that of substance, action, and force'. 10 While extension may be an attribute of matter, it is not its essence; the notion of force 'is really prior to it'. 11 Motion, too, is the expression of this more fundamental entity: 'there is ____________________ 10 Leibniz, "'Whether the Essence of a Body Consists in Extension'", Journal des Sçavans, 18 June 1691, trans. P. Wiener, in Leibniz: Selections ( New York: Scribner's, 1951.), 101. 11 Leibniz, Specimen Dynamicum, trans. Wiener, Leibniz, 120-1. Cf. Leibniz, "'On the Correction of Metaphysics and the Concept of Substance'", trans. L. Loemker, Philosophical Papers and Letters ( Dordrecht: Reidel, 1969), 433: 'The concept of forces or powers, which the Germans call Kraft and the French force, and for whose explanation I have set up a distinct science of dynamics, brings the strongest light to bear upon our understanding of the true concept of substance.' 8 M. Jammer, Concepts of Force: A Study in the Foundations of Dynamics ( Cambridge, Mass.: Harvard University Press, 1957), asserts that Descartes denied ontological status to force and conceived it as a 'fictitious appearance' (p. 103 ). As we shall see below (Sect. III and V), Richard J. Blackwell, "'Descartes' Laws of Motion'", Isis, 57 ( 1966), 220-34, and Peter Machamer, "'Causality and Explanation in Descartes' Natural Philosophy'", in P. Machamer and Robert G. Turnbull (eds.), Motion and Time, Space and Matter ( Columbus, Oh.: Ohio State University Press, 1976), ch. 7, have taken positions that imply or are very similar to my own. These writers each had some goal

other than analysis of the textual evidence bearing on Descartes' concept of force. Two articles, however, have been directed specifically toward this problem: H. Carteron, "'L'Idée de force mécanique clans le système de Descartes'", Revue Philosophique, 94 ( 1922), 243-77,483-511; and Martial Gueroult, "'Métaphysique et physique de la force chez Descartes et chez Malebranche'", Revue de Métaphysique et de Morale, 59 ( 1954), 1-37, 113-34 [trans. as "'The Physics and Metaphysics of Force in Descartes'"]. According to Carteron, Descartes was led 'de rejeter l'idée de force, au nom du principe des idées claires'; yet this rejection became 'une source d'obscurités et d'erreurs' ( "'L'Idée de force'", 490), because it led Descartes to introduce the obscure notions of 'impenetrability' and 'force of repose' (qualités occultes!) into-his physics to do the work of 'force'. Gueroult contends that Descartes left motion in the lurch (as it were), 'distinguée des corps' and 'distinguée de Dieu'; motion takes on life of its own, being passed from body to body by means of direct contact during collisions ( Métaphysique', 32-3). 9 P. Costabel, "'Newton's and Leibniz's Dynamics'", trans. J. Briggs, Texas Quarterly, 10 ( 1967), 119-26; Gabbey, "'Force and Inertia'", 1-6. -283nothing real in motion itself apart from the reality of momentary transition which is determined by means of force and a nisus for change'. 12 Since force is fundamental to both material substance and motion, dynamics, the science of force, is the fundamental science of nature. It is in the Leibnizian sense that Descartes lacked a science of 'dynamics', and did not conceive of motion 'dynamically'. That is, he did not attribute force to matter as an active causal agent; he therefore did not rest his analysis of the cause of motion upon force in matter. Descartes consciously denied matter any true causal agency, i.e. any force in the strict sense of the word. (In this he was followed by Malebranche and other Cartesians, cf. below.) This is not to say that Descartes took no interest in the reality behind motion, but simply to deny that he conceived force as a real property of matter. II A closer examination of the view that Descartes did attribute force to matter will bring into relief my contention that he did not. I will take as representative of this view the cogent version presented by Westfall in his book on the concept of force in the seventeenth century. 13 According to Westfall, Descartes' principal contribution to the development of the science of mechanics was to provide clarity to the concept of motion as a state of being equivalent to other states (such as the state of rest). Descartes saw motion--rectilinear motion--as a state in which a body would remain so long as nothing changed it. It so doing, he focused attention upon 'the central question to which the principle of inertia has directed dynamics', 14 the cause of changes of motion.

Beyond this, Westfall continues, Descartes reduced most changes of motion to a single causal agency. Because matter as conceived by Descartes is wholly inert, it can be moved only by something external to it. This might be a human soul, or God. Yet God's action 'was confined to the origin of motion . . . ; since then He has limited Himself to maintaining the original quantity'. 15 Outside God and the human soul, in the cases of all purely physical phenomena, only the impact of another body can operate to change a body's motion. . . . The quantity of motion in the universe remains constant because of the immutability of God, but God has constructed the ____________________ 12 Specimen Dynamicum, trans. Wiener, 120. 13 Westfall, Force in Newton's Physics, ch. 2. 14 Ibid. 59. 15 Ibid. 60. -284universe such that motion can pass from one body to another in the process of impact. 16 Westfall has Descartes treat motion as something God put into the universe in the beginning, and which then took on a life of its own, being passed from body to body through impact, which serves as the primary causal agency in all physical interactions. Even though Descartes made matter in motion the primary causal agency of his universe, this did not lead him, so Westfall laments, to develop a science of dynamics. Yet, while Westfall finds no fully formulated science of dynamics in the work of Descartes, he does find several terms that were used with 'dynamical signification', of which the two principal ones are 'agir' and 'force'. It is from the usage of these terms that Westfall constructs his account of what a Cartesian science of dynamics would have looked like. Descartes used 'agir', as one might expect, in the context of describing interactions among bodies. He also used it to describe the jiggling of the particles of bodies that were 'agités', or in 'agitation': in this manner he described heat. Finally, he attributed 'action' both to bodies actually in motion, and to those that merely had an inclination to move, as in the case of light. From these usages Westfall draws the following conclusion: Descartes' use of 'agir' and its cognates illustrates how the main thrust of his philosophy directed him away from the formulation of a dynamics such as we know. The attention of his mechanical philosophy of nature was focused on the causes of phenomena, and his principal dynamic concepts were directed, not toward the changes of motion that a body undergoes, but toward bodies that 'agi[ssen]t' on others to change their

motions. What 'action' signified to him was the operation of one body on another to change its state of motion or rest. 17 While the usage of the term was not precise, it none the less betrayed to Westfall the essence of Descartes' dynamics--the action of one body upon another. Descartes employed the term 'force' in several contexts, the most common of which was in referring to the force of a body's motion, or its force to continue to move. 18 A body's size and velocity determine its force ____________________ 16 Ibid. 17 Ibid. 63. 18 Descartes also used the term 'force' in a manner analogous to our word 'work'. This usage was restricted to several letters on simple machines ( Descartes to Constantijn Huygens, 5 Oct. 1637 ( A T I431-47); to Mersenne, 13 July 1638 ( A T II222-45 [ C S M K111 ff.])). As Westfall says regarding these letters, Descartes simply employed the term 'force' according to accepted usage ( Force in Newton's Physics, 72). Carteron, in 'L'Idée de force', says that 'la statique de Descartes est isolée clans son of, œuvre; il l'écrivit à la demande de Huygens, poussé par des considérations pratiques, elle n'est pas rattachée à une science plus générale, elle est dénuée de toute préoccupation encyclopédlque' (p. 257 ). -285of motion, a concept 'roughly equivalent to our concept of momentum'. 19 This usage again betrays Descartes' emphasis on impact as the model for force, in which force was conceived as something possessed by the moving body. Such a conception of force could only be a hindrance to progress: Although he formulated a conception of motion that made changes in motion the objects of explanation, Descartes' quantitative dynamical concept focused, not on changes, but on the steady motion which, by his own words, requires no explanation. As such, the idea of the force of a body in motion became, more of an obstacle to a quantitative mechanics than an aid. 20 Although Descartes' conception of motion as a natural state was a step in the right direction, his conception of force as something bodies possess, rather than something that acts upon bodies, was a stumblingblock (as it were) on the road to a well-conceived quantitative mechanics ( Newton). Beyond the problem just mentioned, Westfall finds two difficulties in the application of the concept of the force of a body's motion. The first is with Descartes' distinction between the motion or force of a body and its 'determination' in any particular direction. Descartes maintained that the motion of a body is completely independent of the

direction in which the body is moving: a change can occur in one without any change in the other. According to Westfall, the primary consequence of this distinction 'was to remove changes of direction from the realm of dynamics'. 21 Descartes' position denied the vectorial nature of motion. This, says Westfall, was 'disastrous' for Descartes' dynamics: His first two laws of motion, taken together, asserted the vectorial nature of inertial motion. A body will continue to move in a straight line until something external acts upon it. As far as motion is concerned, a change of direction is as significant as a change of speed; and once the idea of rectilinear inertia is assumed, a workable dynamics will require that both changes be reduced to the same quantitative measurement of action. Descartes' distinction denied that the two changes have anything in common, and left changes of direction in the anomalous status of changes that are not changes, changes that involve no act. 22 This contradictory state of affairs threatened to take from Descartes one of his most prized feathers: his emancipation from the concept of natural circular motion. If no action is required to change a body's direction, the question of the dynamics of circular motion does not arise at all. 23 The second difficulty concerns whether matter is indeed inert, that is, ____________________ 19 Westfall, Force in Newton's Physics, 63. 20 Ibid. 21 Ibid. 66. 22 Ibid. 67. 23 Ibid. 82. -286wholly indifferent to motion. In league with Galileo, Descartes 'held that matter is wholly inert, lacking in any internal force and dominated in its motion by external actions upon it'. 24 But if matter is truly inert, how can it be (this is the same question asked by Leibniz) that the size of a body is opposed to its motion? Descartes solved this dilemma by application of the concept of quantity of motion, which was defined kinematically in terms of size and velocity. If quantity of motion equals the product of size and velocity, then the larger a body the greater the quantity of motion needed in order to produce the same velocity. But, says Westfall, in the application of this ostensibly kinematical concept to cases of impact, dynamical elements once more entered the fray. Descartes spoke of a 'resistance to motion'; larger bodies have more of this resistance than smaller bodies, and from this Descartes concluded that a larger body at rest could not be put into motion upon impact by a smaller body in motion, no matter what its velocity. But this is problematic: 'From the point of view of his own philosophy, the resistance to motion that he admitted could not be reconciled to inertness which was a consequence of the essential

nature of matter.' 25 This move, was an unfortunate one for Descartes, since he had attempted 'to remove [the problem of impact] from dynamics and to convert it to a kinematic problem based on the principle that God in His immutability maintains eternally the same quantity of motion in the universes'. 26 Despite Descartes' efforts, 'dynamic considerations slipped in unawares and with them a resistance to motion that cannot be reconciled with the inertness of matter'. 27 Descartes was revealed as having a dynamics in spite of himself. The picture that Westfall paints of Descartes' writings on motion is that they are fraught with inconsistency. While on the one hand Descartes asserts that each piece of matter is inert, on the other hand he attributes causal agency to particulate matter. While he asserts principles that ought to imply the vectorial nature of motion, he denies that force must act in order to change the direction of a body in motion. Finally, he attributes to his inert matter the ability to resist motion once it is at rest. III Westfall's picture of Descartes contrasts with a view gained from paying closer attention to the metaphysical foundations of Descartes' ____________________ 24 Ibid. 69. 25 Ibid. 72. 26 Ibid. 82. 27 Ibid. 70. -287physics. 28 In his analysis of Descartes' views on causality, Peter Machamer has sketched this alternative interpretation: Since matter is completely passive and characterised by extension, the efficient cause of the motion of matter must come from elsewhere. In Descartes' ontology, there are only two active principles: God and mind. In the case of matter, God must be the active principle and thus the efficient cause of motion. 29 Matter is truly inert. All causal agency must be attributed to God or to created minds. Such a view of Descartes' theory of motion follows not from a systematic analysis of the various contexts in which the terms 'force' or 'action' are employed, but from close inspection of those passages in which Descartes himself elucidates the manner in which 'the force of a body to act' ought to be conceived. Thus, in arguing counter to Westfall and others, I will not simply offer more of the same kind of evidence, i.e. more contextual analysis, 30 but rather I will examine those passages (primarily from the Principles 31 ) in which Descartes sought to clarify the concept of force; I will seek not to distil his analysis

of this concept from various sentential contexts, but rather to discern the manner in which 'force' fits into his metaphysics and system of nature. The upshot of my view is that when Descartes used terms such as 'force' he did so according to customary usage. However, as a good mechanical philosopher, his aim was not simply to make rigorous the common conception of force, but rather to reveal the reality behind that conception. ____________________ 28 Descartes himself says that his physics is grounded on his metaphysics, in a letter to Mersenne of 28 Jan. 1641: 'I may tell you, between ourselves, that these six Meditations contain all of the foundations of my Physics' ( A T III297-8: K94 [ C S M K173]). Cf. to Mersenne, 17 May 1638 ( A T II141-2: K55 [ C S M K103]). Of course, Descartes' stated intention to ground his physics upon his metaphysics is not sufficient evidence that he in fact did so, but it suggests that we should give close scrutiny to this matter, in order to decide for ourselves. I take it that the foundation of his physics is found in those portions of the metaphysics in which the essence of material substance is identified with extension and it is concluded that the properties of material substance are all 'comprehended in the object of pure mathematics' ([ A T VII80: C S M II55]; the argument runs throughout the Fifth and Sixth Meditations; see esp. H R I179 and 185 [ A T VII62-3, 71: C S M II43, 49]), and the passage in which the doctrine of continual creation is established (Third Meditation, H R I168 [ A T VII49: C S M II33]). 29 Machamer, "'Causality and Explanation'", 178-9. A view similar to the one stated by Machamer was considered by Westfall, only to be rejected in favour of an interpretation based on contextual analysis of terms having 'dynamical signification' ( Force in Newton's Physics, 61). 30 Gabbey, like Westfall, relies upon sentential context in his analysis of Descartes' concept of force ( "'Force and Inertia'", 20-31), as does Clarke ( "'Impact Rules'"). 31 Of the two works by Descartes that extensively treat of the problem of motion, Le Monde and the Principles of Philosophy, the latter warrants primary consideration both because it reflects Descartes' mature thought (it was written in 1640-4, whereas Le Monde was written in 1629-33) and because it was the first to be published (in 1644, as opposed to 1664) and has enjoyed a wider reputation. -288It is well known that Descartes considered the essence of matter to be extension, 32 a concept that does not in itself include the notion of causal agency. But while the essence of matter is extension, matter can be in motion. Perhaps moving matter has (or accrues) agency. In part ii of the Principles Descartes defines the concept of motion. Building upon the rhetorical stance he had adopted in part i, where he emphasized the need for correcting the errors that we have fallen into because of the uncritical tutelage taken from our senses since childhood, 33 Descartes first describes the common, or 'vulgar', conception of motion: 'motion . . . in the vulgar sense is nothing more than the action [actio, action] by which any body passes from one place to another.' 34 The most important thing to notice about the common notion is that it refers to action. Descartes then contrasts the common view with the correct definition: that motion is the

transference [translationem, transport] of one part of matter or one body from the vicinity of those bodies that are in immediate contact with it, and which we regard as in repose, into the vicinity of others. . . . And I say that it is the transportation [translationem, transport] and not either the force or the action which transports, in order to show that the motion is always in the mobile thing [mobili, mobile], not in that which moves [the mover: movente, celuy qui meut]; for these two do not seem to me to be accurately enough distinguished. 35 While it is vulgarly held that motion entails action, Descartes takes pains to distinguish these, and to treat motion solely in terms of changes in spatial relation. Motion is defined purely kinematically as the translation of matter relative to other pieces of matter; beyond this, it should be made clear that it has no existence apart from that which is moved. Motion 'is a mode of the mobile thing and not a substance, just as figure is a mode of the figured thing, and repose of that which is at rest'. 36 The significance of saying that motion is a mode of matter is that motion has no reality outside the moving thing, just as shape has no reality outside the thing that has shape. Descartes gives a more vivid illustration of this distinction in a letter to Mersenne: ____________________ 32 Principles, part i, art. 53. 'There is always one principal property of substance on which all others depend. Thus extension in length, breadth and depth constitutes the nature of corporeal substance' ( H R I240). 33 Ibid. part i, art. 71-6 ( H R I249-53). 34 Ibid. part ii, art. 24 ( H R I265). The words in brackets are from the original Latin and Picot's French translation (which Descartes is said to have read and approved, and to which he may have contributed himself ( A T IXB pp. vii-xx), respectively. 35 Principles, part ii, art. 25 ( H R I266). 36 ibid. Descartes discusses his usage of the word 'mode' at Principles, part i, art. 56 ( H R I 241-2), where he also discusses the closely related terms 'attribute' and 'quality'. -289I do not believe there are in nature any real qualities, attached to substances and separable from them by divine power like so many little souls in their bodies. Motion, and all the other modifications of substance which are called qualities, have no greater reality, on my view, than is commonly attributed by philosophers to shape, which they call only a mode and not a real quality. 37 As long as the discussion is limited to matter and motion, it need be concerned only with things that are definable geometrically. But what about the cause of motion? Descartes had no intention of leaving this topic unmentioned, even though he believed that motion could best be understood 'if we enquire only about locomotion [locali, d'un lieu en autre],

without taking into account that force [vi, force] that produces it'. 38 In article 36 of part ii of the Principles, Descartes begins his treatment of the cause of motion as follows: After considering the nature of motion, we must treat of its cause. First, the universal and primary cause--the general cause of all the motions in the universe; secondly the particular cause that makes any given piece of matter assume a motion that it had not before. 39 We shall take up these two causes in turn. The first or primary cause of motion is God himself, who acts not only at the creation to impart motion to matter at the first instant, but who continues to act at each instant: As regards the general cause, it seems clear to me that it can be none other than God himself. He created matter along with motion and rest in the beginning; and now, merely by his ordinary cooperation, he preserves just the quantity of motion and rest in the material world that he put there in the beginning. 40 Though motion is but a mode of bodies, it can still exhibit a definite quality, 'and it is readily conceived that this quantity may be constant in the universe as a whole, while varying in any given part'. 41 The quantity of ____________________ 37 Descartes to Mersenne, 26 Apr. 1643 ( K135 [ A T III648: C S M K216]). While Descartes' distinction of motion as a mode from the substance of which it is a mode is clear enough, one wonders how the attribution of this mode to single particles can be reconciled with Descartes' avowed relativity ( Principles, part ii, art. 29 ( A G, 21011)). There is a fundamental tension between Descartes' treatment of motion and his supposed relativity, and it has been suggested (e.g. by Blackwell, 'Descartes' Laws of Motion', 227) that Descartes' relativity was an artificially contrived means of avoiding difficulties similar to those of Galileo. This view was put forth by More in 1662: 'his [ Galileo's] Imprisonment frighted Des-Cartes into such a distorted description of Motion, that no mans Reason could make good sense of it, nor Modesty permit him to phansy any thing Non-sense in so excellent an Author' ( A Collection of Several Philosophical Writings, 2nd edn. ( London), p. xi). 38 Principles, part i, art. 65 ( H R I246-7). 39 Ibid. part ii, art. 36 ( A G215). 40 Ibid. 41 Ibid. It is not, however, readily conceived how this quantity can be constant (if that means determinate) in an indefinitely large universe ( Principles, part ii, art. 21 ( H R I264-5)). -290-

motion is reckoned as the product of the size and velocity of a piece of matter; an instance of its conservation goes as follows: 'if the motion of one piece of matter is retarded, we must assume an equal acceleration of some other body of the same size'. 42 This wording brings to mind God conserving motion in the case of colliding bodies, and we shall examine such cases below. In any event, the fact that the quantity of motion is conserved follows directly from God's immutability: since he put a definite quantity of motion into the universe in the beginning, and since 'we conceive it as belonging to God's perfection, not only that he should in himself be unchangeable, but also that his operation should occur in a supremely constant and unchangeable manner', it follows that 'from the mere fact that God gave pieces of matter various movements at their first creation, and that he now preserves all this matter in being in the same way as he first created it, he must likewise always preserve in it the same quantity of motion'. 43 While maintenance of a constant quantity of motion in the universe might have been attributed solely to the properties God gave matter together with the initial divine impulse, Descartes attributes it to God's continued operation. Indeed, it seems that so long as Descartes grounded the conservation of the quantity of motion upon the immutability of God, he could do nothing else. For if God put a certain quantity of motion into the universe at the beginning and were no longer actively involved in the production of that movement, so that the passage of motion from one particle to another took place independently of God's operation, there seems to be no reason for God's immutability to imply conservation. In such a case Descartes would have to base his argument for conservation on the immutability of nature itself, not on that of God. 44 God is the cause of all the motions that occur in the universe, but the particular motions that each body assumes, that is, the changes in a body's motion, must still be accounted for. To these Descartes turns in article 37: 'From God's immutability we can also know certain rules or natural laws which are the secondary, particular causes of the various motions we see in different bodies.' 45 Perhaps the most striking aspect of this passage is that the secondary causes of motion, the causes of change in ____________________ 42 Ibid. That the intent of this example is that some other body of the same size must receive equal acceleration or a body of different size receive an appropriately proportional acceleration, is suggested by the French version, which reads: 'toutesfois & quantes que le mouuement d'vne partie diminuë, celuy de quelque autre partie augmente à proportion' ( A T IXB 84). 43 Principles, part ii, art. 36 ( A G215-16). 44 Nor does it seem that the conservation of the quantity of motion could rest solely upon the fact that at each instant God preserves the entire universe ( Principles, part ii, art 21), for surely his act of preservation could be immutable, while the thing preserved (the universe) could exhibit changing attributes (including the quantity of motion). 45 Principles, part ii, art. 37 ( A G216). -291-

motion, are said to be a set of rules or natural laws. While it may seem odd that rules or laws could be causes, it should be noticed that these laws follow from God's immutability, and have to do with the way in which he preserves the motion in matter. The laws are therefore manifestations of the ways of God; they 'cause' changes in motion in that they are the reason for God's alteration of the motion in any given piece of matter. In order to understand this more fully, let us look at the laws, of which there are three.The first two laws of nature are: (1) Every reality, in so far as it is simple and undivided, always remains in the same condition so far as it can, and never changes except through external causes. . . . So our conclusion must be: a moving body, so far as it can, goes on moving. (2) Any given piece of matter considered by itself tends to go on moving, not in any oblique path, but only in straight lines.46 The first law is a general rule that applies to all states of being, including such states as being of a particular shape, being at rest, or being in motion. The first and second laws taken together are, at least descriptively, 47 equivalent to the principle of inertia, a fact that has justifiably earned Descartes acclaim by historians of mechanics. Unlike Newton's principle of inertia, however, these two laws do not follow from a notion of vis inertiae; Descartes says that the reason (causa) for these laws 'is the immutability and simplicity of the operation by which God preserves motion in matter. For he preserves the motion in the precise form in which it occurs at the moment when he preserves it, without regard to what it was a little while before.' 48 This immutability is responsible for the fact that projectiles remain in motion after having left the hand of the thrower (and until brought to rest, for example, by air resistance), which is an instance of ____________________ 46 Principles part ii, arts. 37, 39. ( A G216, 217). 47 For an exposition of the descriptive equivalence but theoretical non-equivalence of Descartes' laws and the principle of inertia, see Blackwell, 'Descartes' Laws of Motion. 48 Principles, part ii, art. 39 ( A G217). It has been contended on the basis of this passage that God conserves matter along a rectilinear path because a straight line is simpler ( Machamer, "'Causality and Explanation'", 189-90) or more perfect ( Blackwell, "'Descartes' Laws of Motion'", 224) than a curved line. These readings, however, conflict with the next sentence in article 39: 'In the instant, of course, no motion takes place; but obviously the motion of any moving body is determined at any assigned instant of its duration as capable of being continued in a given direction; continued, that is, in a straight line, not in some curve.' We might express Descartes, argument by saying, roughly, that the only path assignable to the motion of an object in infinitely small time, and therefore along an infinitely small line segment, is a straight line. Notice that this is not an appeal to simplicity or perfection. Descartes' usage of 'simplicity' in the passage quoted above (in the text) should be read in the etymologically pure sense of the word, as signifying a onefold action, such as would be an action performed in a single instant. -292-

the first law. Moreover, since God conserves the motion as it is in an instant, without regard for the previous instant, i.e. since he conserves it along a straight line, the tendency of a rock in sling (at each moment) to move along the tangent is accounted for. This is an instance of the second law. In the two laws considered so far the word 'force' has not been used; but then neither are these two laws sufficient for explaining changes in motion, since they imply eternal motion along a straight line. Indeed, if the 'secondary causes' of motion, i.e. the three laws of nature, were really intended to explain 'the particular cause that makes any given piece of matter assume a motion that it had not before', 49 i.e. changes in motion, one might at this point wonder why Descartes bothered even to consider the first two of these laws. Leaving this question aside for the moment, let us turn to the third of Descartes' laws, in which he does mention force, and from which he generates a set of impact rules describing the manner in which colliding bodies undergo a change in their motion. 50 (3) When a moving body collides with another, then if its own force [vim, force] of going on in a straight line is less than the resistance of the other body, it is reflected in another direction and retains the same amount of motion, with only a change in its direction; but if its force of going on is greater than the resistance, it carries the other body along with it, and loses a quantity of motion equal to what it imparts to the other body. 51 There can be no doubt that Descartes intended this rule to do the primary work in accounting for changes in motion, for he explicitly says that 'this third law covers all the particular causes of corporeal change--so far as they are themselves corporeal'. 52 ____________________ 49 Principles, part ii, art. 36 ( A G215). 50 These are the notorious seven rules of impact ( Principles, part ii, arts. 46-52). Of these rules, I make only passing reference to rule 4 (below). For a detailed discussion of all seven rules, see Blackwell, "'Descartes' Laws of Motion'". 51 Principles, part ii, art. 40 ( A G218). I have altered this and following translations where necessary in order to translate vis uniformly as 'force'. Westfall's and Gabbey's analysis of Descartes' concept of force focuses upon the usage of 'vis' and upon Descartes' statement of the third law. 52 Ibid. Descartes goes on to say that he is 'not now considering whether, or how, human or angelic minds [mentes] have the power to move bodies'. It is perhaps significant that of the three non-extended beings that can affect matter--God, human minds, and angelic minds--Descartes here specifically excludes human and angelic action, but not divine; this is as would be expected if Descartes believed that divine power is the cause behind corporeal change in general. (For two of the many passages in which Descartes lists God, human minds, and angelic minds as the three forces or powers able to act upon corporeal substance, see " Descartes to More, 5 Feb. 1649" ( K239 [ C S M K 3611]) and 15 Apr. 1649 ( K249 [ C S M K373]).

-293The third law comes in two parts, covering two classes of effect arising from the impact of a pair of bodies. The first part concerns cases in which the only effect is a change in determination or direction of one of the bodies, with no change in the quantity of motion. As a proof of this first part of the law, Descartes offers the following: There is a difference between a motion as such and its determinate direction; it is thus possible for the direction to change while the motion remains unaltered. Now, as I said, any given reality which, like motion, is not complex but simple, persists in being so long as it is not destroyed by any external cause. In a collision with a hard body, there is an obvious reason why the motion of the other body that collides with it should not continue in the same direction; but there is no obvious reason why this motion should be stopped or lessened, for one motion is not the opposite of another motion; so the motion ought not to be diminished. 53 A moving body collides with a second (larger) body, which is at rest. Descartes declares that there is an 'obvious reason' for the first body to change course. This obvious reason follows from the fact that extension is the essence of matter, and that extension (in Descartes' view) implies impenetrability. 54 If the bodies are not to pass through or destroy one another (they are assumed to be perfectly inelastic), either the first body must change course or the second must move (or there must be a compromise). Since the first part of the third law is restricted to cases in which the moving body cannot move the stationary body, the moving body must change course. It loses none of its velocity, and so the quantity of motion is conserved, consistent with Descartes' strictures. The second part of the third law covers those cases in which motion apparently is transferred from the moving body to the body with which it collides. It is here, especially, that one would expect Descartes' dynamics (if he had one) to be revealed. Descartes' proof of this part of the third law, however, is disappointing to one seeking to find causal agency in the motion of matter: The second part is proved from the immutability of the divine operation; God preserving the world by the same activity by which he once created it. For all places are filled with body, and at the same time the motion of every body is rectilinear in tendency; so clearly, when God first created the world, he must not only have assigned various motions to its various parts, but also have caused their mutual impulses and the transference of motion from one to another; and since he now ____________________ 53 Principles, part ii, art. 41 ( A G218). 54 Descartes to More, 15 Apr. 1649 ( K249 [ C S M K372]): 'It is impossible to conceive of one part of extended substance penetrating another equal part without eo ipso thinking that half the total extension is taken away or annihilated; but what is

annihilated does not penetrate anything else; and so, in my opinion, impenetrability can be shown to belong to the essence of extension and not to that of anything else.' -294preserves motion by the same activity and according to the same laws, as when he created it, he does not preserve it as a constant inherent property of given pieces of matter, but as something passing from one piece to another as they collide. Thus the very fact that creatures are thus continually changing argues the immutability of God." 55 God caused motion in the beginning, as well as the transference of motion from one body to another. Does he cause one particle to transfer its motion to another by endowing the particles with force, or does he, by his action, do the transferring? The fact that he conserves motion 'as something passing from one piece to another' suggests the latter. But this interpretation ignores the fact that Descartes speaks of 'force' in his statement of the third law. What is that notion of 'force'? As if having anticipated the question, Descartes provides an answer in the article immediately following the one just quoted. He says: It must be carefully observed what it is that constitutes the force [vim, force] of a body to act on another body or resist its actions; it is simply the tendency of everything to persist in its present state so far as it can (according to the first law). Thus what is joined to another thing has some force of resisting separation from it; what is separate has some force of remaining at rest, and consequently resisting everything that might change its state of rest; what is moving has some force of persisting in its motion-in a motion constant as regards velocity and direction. 56 Here Descartes elucidates both the force (vis, force) of a body to act on another and the force of one body to resist the action of another. These 'forces' are nothing but the tendency of bodies expressed in the first law and manifested in the second: the tendency of each body to persist in its own state. Here, then, we find the reason for including the first two laws in the section that accounts for changes in motion: the tendency that a body manifests at each instant to move in a given direction or to remain at rest explains not only its own behaviour, but also, together with the same tendency in other bodies, its interactions with those bodies. The notion that each body has at each instant a certain tendency to move (or to remain at rest) is thus fundamental to Descartes' physics. But notice that Descartes gives no reason why bodies should tend to persist in their own state other than the immutable nature of divine action that preserves bodies at each moment. A fortiori, in explaining the tendency to persevere, he does not appeal to the force that a body has by virtue of its motion. Indeed, the relation is just the reverse: 'the force of a body to act . . . is simply the tendency of everything to persist in its present state', a tendency

____________________ 55 Principles, part ii, art. 42 ( A G219). 56 Ibid. part ii, art. 43 ( A G219). -295that does not follow from any property of matter, but rather from an attribute of God. Just as with the preservation of the quantity of motion, it is difficult to see how a tendency to move could be grounded upon the immutability of God, unless God were directly responsible for the tendency itself. The picture that results is one of God imparting motion to matter at the creation and continuing to impart the same quantity of motion at each instant in the duration of the universe. But not always the same quantity of motion to the same individual particles: the universe is full, and in such heavy traffic each particle is not free to realize its tendency to move forever in a straight line. Thus in order to preserve the same quantity of motion in the face of constant contact among the various particles, God is eternally giving one particle more motion and another particle less, according to their size and speed, in the manner described by the three laws of motion (and the attendant rules of impact). The laws of nature, then, are the 'secondary, particular causes' of various motions inasmuch as it is according to them that God finds reason to change the direction and speed of any given moving particle as it comes in contact (one might say 'upon the occasion of its contact' 57 ) with one or more other particles. 58 Or, if we focus upon the fact that God creates the universe anew at each instant, 59 these laws provide the reason for God to create the particle at successive instants along one path rather than along another. This reading of the Principles is reinforced by consideration of Descartes' earlier written work on natural philosophy, Le Monde. 60 While it is true that in this work Descartes speaks of bodies 'impelling' one ____________________ 57 The implication of the interpretation of Descartes that I am developing is that his view of causal interaction in the material realm was implicitly occasionalistic (see Sect. IV, esp. n. 87 below). 58 As James Collins puts it ( Descartes' Philosophy of Nature), 'Laws of nature are quite pregnantly laws for nature, in the sense of being dependent and instrumental expressions of the manner in which the divine power conveys movement to the whole field of material particles. The laws or rules themselves are intellectual formulations of the pattern followed by God in communicating motions to bodies' (p. 27 ). One might think that Descartes need not have made the relation between God's establishment of the rules or laws and their continuing manifestation quite so direct; perhaps it would be enough for God to decree the laws at creation, so that matter would forever after follow the decree. But unless matter forever after remembered the decrees and understood them well enough to carry them out, this would require that God's decrees act at a temporal distance, which is as unacceptable as having the matter follow the laws intentionally. A third possibility, that the predilection to follow the laws is

inherent in the essence of the matter (cf. Leibniz), is not available to Descartes, since for him the sole essence of matter is extension. 59 Principles, part ii, art. 21, and part ii, art. 39; also Third Meditation ( H R I168) [ A T VII 49-50: C S M II34]). Cf. Jean Wahl, Du Rôle de l'idée de l'instant dans la philosophie de Descartes, 2nd edn. ( Paris: Vrin, 1953). 60 Since the argument for a Cartesian dynamics draws upon evidence taken from this early work of Descartes (written in 1629-33, first published 1664), it is in order to consider some of -296another by contact, and that he uses the word 'force' in speaking of moving particles (as he does in the Principles as well), attention to those passages in which Descartes directly considers the cause of motion suggests that in Le Monde he is even more explicit about God's direct activity in the universe. Thus he says that God put matter into motion in the beginning, and that he 'always acts in the same way, and consequently always produces the same effect in substance'. 61 He then presents two rules (one and three from the Principles) that are meant to explain the changing motions that bodies exhibit. 62 Since God is immutable, the particular changes in motion that a body undergoes cannot be explained by changes in God's mode of acting; but that which underlies these changes-the tendency of each particle to remain in the same state (of rest or motion) at successive moments--is attributable to God and follows from his immutability. 63 This becomes clearer in light of the conclusion that Descartes draws from the third rule he introduces, which concerns the tendency toward rectilinear motion: 64 It must be said, then, that God alone is the author of all the movements in the universe, in so far as they exist, and in so far as they are in straight lines; but that there are various arrangements of matter which render them irregular or curved; just as the theologians teach us that God is the author of all our actions, in so far as they have any goodness in them, but that it is the various dispositions of our wills which make them bad. 65 While the tendency to move in a straight line is owing to God alone, Descartes does not give him responsibility for the irregular paths that ____________________ this evidence. Much weight is placed upon a passage in which Descartes says that 'the virtue or power of moving itself which is found in a body is indeed able to pass wholly or in part into another [body]' ( A T XI11; Gueroult, "'Métaphysique'", 32, and Gabbey, "'Force and Inertia'", 20, put particularly heavy weight upon this passage), and indeed it appears that Descartes is here analysing the cause of motion in terms of a power or virtue that passes from one body to another during collisions. Yet at the beginning of the paragraph in which this passage occurs, Descartes says he is not inquiring into the cause of movement, and in the final sentence he says that if the reader wishes, these movements may be assigned to a primuni mobile revolving about the world, all of which suggets that this passage ought not to be taken as Descartes' final word. Indeed, this passage is found in ch. 3 of Le Monde, on solids and fluids. I

will restrict my attention to ch. 7, on the laws of nature, as it is here that we should expect to find Descartes' serious attempts to analyse motion. 61 A T XI37 (my emphasis), trans. H. Torrey, The Philosophy of Descartes ( New York: Holt, 1892), 239 (1 have used but not always followed Torrey) [ C S M I93]. 62 A T XI38-42: Torrey, Philosophy, 239-43. 63 Descartes repeats the phrase about God 'always producing the same effect' ( A T XI43: Torrey, [ C S M K96] Philosophy, 243). 64 This rule also depends upon the fact that God conserves each thing 'by one continuous activity' ( A T XI44: Torrey, Philosophy, 244). 65 A T XI46-7 [ C S M I97]: Torrey, Philosophy, 245. Cf. Malebranche, Méditations Chrétienties, Med. V, arts. 14-15, in ( Œuvres de Malebranche, x). -297bodies take. In what sense are these paths not assignable to God? There are two factors that determine these paths; the laws of nature, which determine the paths that particles take after any given encounter, and the 'various arrangements of matter' throughout the plenum, that is, the pattern of particles of diverse sizes and shapes. Since the laws of nature are direct manifestations of God's activity, and therefore must be attributed to him, it must be that he is not responsible for the second factor determining the paths, which is the pattern of particles. But since each pattern proceeds from a previous pattern according to the laws of nature, it must be that God is not responsible for the sizes and speeds of the particles at the creation--that he created a chaotic universe, which would sort itself out (into the universe we now see) by the ordinary course of nature. Indeed, that is the story of the formation of the universe that Descartes tells in Le Monde and again in the Principles. 66 So it remains that God is the cause behind each motion in the universe, but that he is not responsible for the fact that some one pattern of motions has occurred from the creation rather than another. A consideration of Descartes' major writings on natural philosophy leads to the conclusion that he indeed banned force from inclusion among the properties of matter, and that, by reducing the notion of the force of a moving body to a tendency to persist that depends entirely upon God, he put the causal agency behind motion into divine hands. If doubts remain about this conclusion, Cambridge Platonist Henry More had none. On 11 December 1648, More began a correspondence with Descartes that primarily concerned the Principles. There were several exchanges before Descartes' death in 1650. Of the questions that More posed to Descartes in these letters, those concerning God and motion are of interest here. In his first letter, More questioned Descartes' reason for denying extension to God: 'For how does he communicate movement to matter, as he has done at some time, and does even now according to you, if he does not so to speak touch the material universe, or at least if he has not touched it at some time?' 67 Taking it as given that God (in Descartes' view) still ____________________

66

Le Monde ( A T XI34-6: Torrey, Philosophy, 235-8 [ C S M K I91-2]; Principles, part iii, arts. 45-7 ( A G224-6). It should be noted that even though the formation of the universe proceeds from chaos toward order through God's active production of movement, God acts immutably in this production, which implies that he does not form the various features of the universe by particular (providential) acts of will. Indeed, since Descartes stresses that the laws of motion follow from God's immutability, one wonders to what extent God is free to choose the laws of nature, and hence the kind of world that will emerge from the initial chaos. 67 More to Descartes, 11 Dec. 1648 ( A M VIII94). Cf. ibid. 96: 'if God truly impresses movement into matter, as you have asserted. . . .' -298communicates motion to matter, More asked how God can do so without being extended. Descartes answered that immaterial substances such as God, the human mind, and angels can act on matter even though they are not extended, and therefore may be considered to have 'an extension of power'. 68 These passages are interesting because Descartes tacitly admitted that he believed God to be continually imparting motion to matter, and explicitly proclaimed God's power to do so. In a letter of 5 March 1649, More asked whether rest is a privative or positive mode of body. 69 Descartes answered, in a roundabout fashion, by referring to an example in which he concluded that there is nothing more positive in movement than in rest. 70 More rejoined that from his point of view motion is a force or action, and must be so defined if it is not to be considered merely an extrinsic relation, consisting wholly of changes in the distance between bodies. He asked Descartes how, if motion is merely a mode, it can pass between bodies. 71 Descartes replied: You observe correctly that a motion, being a mode of a body, cannot pass from one body to another. But that is not what I wrote: [though] indeed, I think that motion, considered as such a mode, continually changes. . . . But when I said that the same amount of motion always remained in matter. I meant this about the force which impels its parts, which is applied at different times to different parts of matter in accordance with the laws set out in article 45 and following of the Second Part. So there is no need for you to worry about the transmigration of rest from one possessor to another, since not even motion, considered as a mode which is the contrary of rest, transmigrates in that fashion. 72 So there is no passage of motion from particle to particle. But what is this 'force which impels' the parts of matter and is distinct from motion considered as a mode of substance? Descartes answered this question in the same letter to More: The translation which I call motion, is a thing of no less entity than shape: it is a mode in a body. The force [vis] causing motion may be the force of

God himself conserving the same amount of translation in matter as he put in it in the ____________________ 68 Descartes to More, 15 Apr. 1649 ( K249 [ C S M K372]). This was Descartes' second attempt to answer More's query. In his first answer, Descartes had said that he conceived of God, angels, and human minds, 'as powers [virtutes] or forces [vires] which though they can act upon extended substances, are not themselves extended' ( 5 Feb. 1649 ( A M VIII125: K239 [ C S M K361]) ). Included in Descartes' very conception of God was the notion of force. 69 A M VIII183. 70 To More, 15 Apr. 1649 ( A M VIII216, 212-13 [ C S M K372 ff.]). 71 More to Descartes, 23 July 1649 ( A M VIII247-52). 72 To More, Aug. 1649 ( K258 [ C S M K382 ]). Descartes began his reply before his departure to Sweden at the end of August, but did not put it in final form or post it ( A T V401-2). The letter was discovered by Clerselier and published in the first volume of the Lettres de Mr Descartes ( Paris, 1657; numerous editions followed). The full Descartes-More correspondence, including this letter, was published in More Collection of Several Philosophical Writings. -299first moment of creation; or it may be the force of a created substance, like our soul, or of any other thing to which he gave the force to move a body. This force is a mode in a creature, but not in God; but because this is not easy for everyone to understand, I did not want to discuss it in my writings. I was afraid of seeming inclined to favour the view of those who consider God as a world-soul united to matter. 73 Descartes continued this line of thought into the next paragraph, in which he responded to More's query about the tendency of matter to remain at rest: 'I agree that if matter is left to itself and receives no impulse from anywhere it will remain entirely still. But it receives an impulse from God who preserves the same amount of motion or translation in it as he placed in it at the beginning.' 74 God continually imparts motion to matter so that he might conserve the total quantity. This is the reality, the force or power, behind that mode of body which is called motion. IV In the Principles, and more explicitly in his correspondence with More, Descartes provides abundant evidence that he conceived matter as completely passive, totally lacking the property of, or the ability to communicate, force. Therefore any attempt to attribute a true dynamics to Descartes is doomed. But of what significance is this conclusion? It would be of small consequence if its only implication were that Westfall's analysis failed to follow the letter of Descartes' writings: perhaps, after all, the dynamics that Westfall teased out of Descartes is there despite Descartes' stated intentions to the

contrary. In order to explore this possibility, let us examine Westfall's analysis of Descartes in the light of the conclusions that I have reached thus far. Westfall's first criticism is that Descartes (along with others) put ____________________ 73 A M VIII264: K287 [ C S M K381]. Descartes' statement that the force causing motion may be God, our soul, or 'any other thing [alterius rei] to which he gave the power to move a body' led Prendergast ( "'Descartes' Physics'", 98) to suppose that the 'other things' include material bodies. The text certainly does not force that conclusion, and the grouping of the two qualifications after the semicolon, as well as the wording, suggests that the 'other thing' is another intelligence, such as an angel (cf. nn. 52, 68). 74 K258 [ C S M K381). In these passages from the letters to More, Descartes writes as if God is continually pushing matter about. This conflicts with the implications of the view that God creates the world anew at discrete instants, and thus simply creates each particle at various points along its path of motion. The first view implies continuous preservation of the universe together with continuous pushing, the second view a series of static creations, with each particle having a tendency to be created in either the same or some other position at the next instant (cf. n. 62 above). I shall not attempt to resolve this issue here, though it is of some interest, as the second view is bound up with one of Leibniz's criticisms of Descartes ( Critical Remarks Concerning Descartes' Principles, part ii, arts. 45-52, in Loemker, Philosophical Papers). -300mechanics on the wrong path when he focused attention upon the moving particle as the causal agent in dynamics, rather than taking seriously his own assertion that motion is a natural state and making changes in motion the paradigm case of the action of force. In reply to Westfall, it is false, strictly speaking, that Descartes shifted attention away from changes in motion. We have seen that he introduced his three laws of nature as 'secondary causes' in order to explain changes of motion. His celebrated seven rules of impact all dealt with cases in which changes occur in the motion of the bodies involved. In these cases, in so far as one body provides God with reason to change the state of motion of another body, it may be viewed as 'an external action that changes a body's inertial state', the model of force that Westfall approvingly attributes to Newton. 75 The phrase that Westfall makes so much of in his treatment of Descartes, 'the force of a body to act', implies the force of a body to act on something. Nevertheless, in saying that Descartes viewed matters from the standpoint of the moving particle rather than that of the particle whose motion gets changed, Westfall has captured an aspect of Descartes' analysis of motion. While it is not true that Descartes attributed to the moving particle itself a force enabling it to act on other particles, it is true that God conserves in the individual moving particle its motion in a straight line. It is this divine act of conservation (extending to the rectilinearity of the motion) that causes bodies to persevere in their state of motion and that represents the reality behind the apparent action of one moving body upon another. Thus Westfall, while misunderstanding

Descartes' full intent, rightly draws attention to the importance of rectilinear perseverance in Descartes' analysis of motion. The concept of perseverance which lies behind the apparent force of a moving body also applies to bodies that are at rest. Thus if one were (mistakenly) to attribute to Descartes the opinion that 'force of movement' is a real property of matter, one would be expected also (mistakenly) to attribute to him a similar view with respect to 'force at rest'. Such an attribution is the basis of Westfall's assertion that Descartes falls into inconsistency when he attributes to resting particles a 'resistance to motion'. 76 The problem of resistance to motion arises primarily with Descartes' ____________________ 75 Westfall, Force in Newton's Physics, 63-4. 76 Principles, part ii, art. 49 ( A T IXB 91). Descartes' mention of this 'force to remain at rest' led Carteron ( 'L'Idée de force', 490-7) to assert that Descartes here allows an occult quality into his system of nature. The arguments I will give against Westfall apply equally to Carteron. -301fourth rule of impact, in which he states that a smaller body can never, no matter what its velocity, put a larger resting body into motion. Descartes attributes this behaviour to the larger body's 'force to remain at rest'. 77 Westfall charges that this resistance to motion cannot be reconciled with Descartes assertion that matter is inert, and that Descartes is here revealed as allowing dynamical elements into his universe. But surely this is mistaken: just as Descartes reduced the force of a body's motion to divine preservation, so too he reduced its 'resistance to motion' to the same divine act. Thus he said that God 'preserves just the quantity of motion and rest in the material. world that he jut there in the beginning'. 78 We have seen above that Descartes attributed to matter both the tendency to remain in motion and the tendency to stay at rest: 'what is at rest has some force of remaining at rest, and consequently resisting everything that might change its state of rest'. 79 Both of these tendencies (or 'forces') are referred, in the first law of nature, to the immutability of God as it is expressed in his preservation of states of being such as motion and rest. Even if perseverance in a state of rest can be attributed to divine action, why should it be that a small body can never move a larger one? 80 We have seen that the 'force of resistance' is founded upon divine preservation; but mere preservation of rest does not in itself rule out a compromise result for cases in which a rapidly moving body strikes a resting body that is only slightly larger. To rule out such compromises, one needs a further premiss, one which Descartes did not make explicit in the Principles, but which he revealed in a letter to Clerselier in which he sought to justify the fourth rule. After going through an example in which he concluded that the smaller body cannot move the larger, Descartes continued:

Nevertheless, I am glad that the first and foremost difficulty that you have found in my Principles concerns the rules according to which the movement of colliding bodies is changed; for from that I judge that you have not found difficulty in what precedes the rules, and also that you will not find much difficulty in the rest, nor in the rules either, when you have noticed that they depend only on a single principle, ____________________ 77 Principles, part ii, art. 49 ( A T IXB 91). 78 Ibid. part ii, art. 36 ( A G215); my emphasis. 79 Ibid. part ii, art. 43 ( A G219) . 80 While Westfall could find only inconsistency here, he none the less could praise the fourth rule as 'the first attempt to bring the fact of experience, that more effort is required to give a large velocity to a body than a small one, within the domain of rational mechanics' ( Force in Newton's Physics, 72). Yet, however true it may be that Descartes' conclusion concerning resistance to motion was intended to take experience into account, his justification of the tendency to remain at rest was neither in violation of the inertness of matter (it depended upon God) nor based upon empirical considerations to the exclusion of metaphysics (as we shall see). -302which is that when two bodies in incompatible modes collide, some change in these modes must truly occur, so as to render them compatible, but that this change is always the least possible; that is to say, if, some quantity of these modes being changed, they are able to become compatible, a greater quantity of change will not occur. 81 Descartes grounds his fourth rule on a principle of minimal change; since the smaller body must undergo some change, and since the situation can be resolved if it changes direction without changing speed, there is no reason for a greater quantity of change to occur. The other problem that Westfall finds is with Descartes' distinction between motion and its determination. Westfall's claim amounts to the assertion that Descartes held that changes in direction require no cause. Yet surely Descartes believed that changes in direction are caused: a cursory examination of the third law of nature and the rules of impact reveals that for a body to change its direction, it must at least encounter another body. What Westfall really objects to is that Descartes did not analyse changes in direction by means of a vectorial treatment of motion-that he did not see changes in direction as manifestations of the action of force. 82 However, Descartes in fact reduced both persistence in a straight line and changes of direction to a single causal agency: both reflect the way in which God conserves matter in motion. In the same letter to Clerselier from which I have just quoted, Descartes states explicitly that changes in velocity and determination are, as changes in motion, equivalent: 'Two separate modes must be considered in [the case of] movement: the one is

____________________ 81 Descartes to Clerselier, 17 Feb. 1645 ( A T IV185 [ C S M K247]). 1 do not mean to affirm that Descartes had this principle in mind when he formulated the fourth rule of impact; but this passage does exemplify the kind of justification that Descartes chose to give the fourth rule. 82 Westfall based this criticism (discussed more fully in Sect. II above) upon the contention that Descartes' notion of rectilinear perseverance implies the vectorial nature of motion, in which case changes in direction require force (à la Newton). This contention might be true if Descartes had based this rectilinear perseverance upon a property of matter. But since he based it upon an attribute of God, Westfall's (Newtonian) intuitions about the ideal properties of matter and the implications of rectilinear perseverance are not likely to be of much use. In Descartes' system, the cause of motion (and indeed the reason for regularity or order among natural phenomena) is a rational (the most rational) being. So the factors that enter into the laws can depend upon this being's rationality and, perhaps, his purposefulness. Thus if, in his discussion of impact, Descartes invokes a principle of minimal change (as above), while this might be counter to our intuitions about the nature of matter, there seems to be no reason that God might not be conceived to act in this manner. And if Descartes answers the current difficulty by referring to the reasons that God needs for conserving motion in one direction rather than another, such an answer is hardly open to objections based upon intuitions about material substance simpliciter. In general, it would be interesting to evaluate the role of rationality and purposiveness (final causes) in Descartes' physics from the point of view that the laws of nature are descriptions of God's conserving activity. Cf. Machamer, 'Causality and explanation', and Collins, Descartes' Philosophy of Nature (esp. ch. 4), who have proceeded some way toward such an analysis. -303motion alone or velocity, and the other is the determination of that motion in some direction. Each mode is changed with equal difficulty.' 83 The notion of 'equal difficulty' employed here is not, of course, to be read in dynamical terms as 'equal force'. The implication of the phrase is that each change is as much in need of a cause (a reason for it to take place) as the other. Changes or velocity occur in those cases in which a moving particle encounters another and 'transfers' part of its motion to the other; a change of direction occurs when such an encounter takes place but without any 'transferal' of motion. All of which is to say that in cases of change in velocity, God conserves one body with less velocity than it previously had and adds this velocity to the second body, whereas in cases of change in direction, God simply conserves the first body in a different direction than previously. In each case a cause or reason for (God to make the) change is required. And in each case the cause falls outside the realm of (Leibnizian) dynamics, for in neither is an internal force attributed to one body to explain its behaviour or its effect upon another. Our examination of Westfall's various criticisms of Descartes has shown that Descartes lacked not only an explicit dynamics, but an implicit one as well. While Descartes did

emphasize the rectilinear perseverance of the moving particle, he attributed this to divine action. Similarly, his concept of 'resistance to motion' did not reveal an implicit dynamics, but rather was simply another example of the manner in which God's immutability is expressed in his conservation of the world. Finally, Descartes put changes in direction on the same plane as changes in velocity, and he analysed neither of these changes dynamically. In the end, then, the view that assigns to Descartes a wholly geometrical treatment of matter and motion--a treatment of matter as inert extended substance and a kinematical treatment of motion--stands; Descartes removed causal agency from the material world, and placed it in the hands of God and created minds, that is, in the hands of immaterial substances endowed with the power to act upon matter.83 The implication is that ____________________ 83 A T IV185 [ C S M K247]. 84 In the Principles (part ii, art. 40), Descartes merely poses the question of the ability of human minds to act upon matter; in his correspondence with Elizabeth ( 21 May 1643 ( K137 ff. [ C S M K217 ff.]), 28 June 1643 ( K141-3 [ C S M K226 ff.])) and More ( 15 Apr. 1649 ( K249 [ C S M K372]), he affirms that the mind can act upon body (though he qualifies this by saying that this ability is a mode of the mind, as opposed to part of its essence (to More, Aug. 1649 ( K257 [ C S M K381])). This raises an interesting question: if the mind can affect matter, perhaps it can impel matter to motion, increasing the quantity of motion and thereby violating the principle of the conservation of that quantity. Leibniz answered this question for Descartes, by saying that Descartes 'believed that we ought not to ascribe to the soul the power of increasing or diminishing the force of bodies, but simply that of changing their direction, by changing the course of the animal spirits' ( 'Considerations on the Principles of Life, and on Plastic Natures' ( 1705), in -304Descartes' physics is purely kinematical--it is built upon descriptions of the sizes, shapes, and motions of particles of matter. The motions of these particles exhibit regularity, which can be described by rules referring only to the size and speed of the particles. Within this physics mechanisms can be posited--involving only the description of patterns of moving particles--to account for various phenomena encountered in nature, as when gravity is explained by vortices, magnetism by corkscrew-shaped effluvia and threaded pores in bodies, or prismatic colour by differential velocities of rotation of spinning particles. To go beyond this, to understand the foundations of the principle that particles tend to persist in a state of motion or rest, or of the lawful manner in which particles change course upon collision, or of the conservation of the quantity of motion, it is necessary to enter the realm of metaphysics and to consider the activity of the Deity. And not the activity of God who intervenes providentially, working a miracle now and then, 85 but the constant activity of an immutable preserver, whose very immutability is responsible for the regularity found in the sea of matter.

Returning to the problem raised in the introduction, concerning whether Descartes was constrained by his geometrical treatment of matter in motion to exclude all causal agency from matter itself, we can answer in the affirmative. Motion is fundamental to Descartes' system of nature, but it is not by itself causally fundamental. God is the cause of motion. 86 Descartes was able to divest the material world of the active principles ____________________ 84 Wiener, Leibniz, 91). While I know of no direct statement to this effect by Descartes, it is suggestive to note that in his Treatise on Man the mechanism by which animal spirits are able to induce various motions in the body relies upon a change in the direction of flow (down one tube rather than another). The initial motion (and, presumably, quantity of motion) of the spirits is derived from their origin in the blood ( A T XI179-82: Hall, L'Homme, 91-4). 85 It is possible that Descartes would not have allowed any miraculous intervention into his universe. For example, he says of his Le Monde that we may assume 'God will never work any miracle here' ( A T XI48: Torrey, Philosophy, 246). And the few words about miracles found in his correspondence have a sceptical tone ( A T II558; III214 [ C S M K155]). 86 In light of the conclusion that God is intimately involved with and has a fundamental role in Descartes' universe at each instant in time, it would be interesting to re-evaluate the common conception that ' Descartes, by his teaching, leads to materialism and by his exclusion of God from the world, to atheism' ( A. Koyré, From the Closed World to the Infinite Universe ( Baltimore: Johns Hopkins University Press, 1957), 138; see also ibid. 213, 225, and 237 ; cf. Richard S. Westfall , Science and Religion in Seventeenth Century England ( Ann Arbor: University of Michigan Press, 1973; orig. pub. 1958), 80, 110, 205; compare Newton's views to those I have attributed to Descartes). There is no doubt that this view had currency among some of Descartes' contemporaries. For example, Pascal is said to have remarked that Descartes'aurait blen voulu, dans toute sa philosophie, pouvoir se passer Dieu; mais il n'a pu s'empêcher de lui faire donner une chiquenaude, pour mettre le monde en mouvement; après cela, il n'a plus que faire de Dieu' ( Pensées, [no. 1001]). But on the other side, P. André could write in his Vie du R. P. Malebranche (first written before 1718) that Descartes established his laws of nature and rules of impact upon the following two principles: '10Que la volonté de Dieu étant la force, qui -305ascribed to it by his (as he saw them) less rationally inclined contemporaries only by calling upon the grandest active principle of them all. Thus there is not only a 'ghost in the machine' of the human body, but one in the machine of the universe as well. And if this ghost is not viewed in such a crass way as to make him the 'world-soul united to matter', he is none the less there, extended in his power, accounting for whatever action is found in the world of geometrically conceived matter and motion. Indeed, he is necessary to that world, if it is to contain any activity at all.

Which brings us to a final point. Beyond the implications of this paper for our reading of Descartes, a further implication is that when Cordemoy (d. 1685), De La Forge (d. 1666), Geulincx (d. 1669), Malebranche, Regis (d. 1707), and other Cartesians developed their systems of occasional causes, they were not perverting Descartes, but were extending what was already implicit in his treatment of motion. 87 The central features of ____________________ 86 tient les corps en repos, aussi bien que celle qui les met en mouvement, ces deux manières d'être sont de leur nature aussi positives. . . . 20Que Dieu étant aussi constant dans sa conduite qu'immuable dans son essence, ii conserve toujours dans l'univers une égale quantité absolue de mouvement. D'où il conclut, que le mouvement ne se perd point, et qu'il n'augmente ni ne diminue dans tout I'assemblage des corps; mais qu'il ne fait, pour ainsi dire, que passer de l'un à l'autre par une communication perpétuelle. La raison, qu'il en rapporte, est que la force mouvante, ou la cause motrice de la nature, qui est la volonté divine, demeurant tousjours la mesme, doit tousjours produire le mesme effet' (printed in Œuvres de Malebraiiche, xvii 1. 19). Furthermore, Leibniz reprimanded 'the Cartesians' for leaving God in the universe by referring motion 'miraculously' to divine power ( "'New System of Nature'", Journal des Sçavans, 27 June 1695, trans. Wiener, Leibniz, 113-14; see also below). More recognized Descartes' God as communicating motion to matter, but none the less he later attacked Descartes as a 'nullibist' whose views lead to atheism. This attack may have been prompted by Descartes' denial of More's view of God as a world-soul ( True Notion of a Spirit ( 1681, Latin edn. in Enchiridion Metaphysicum ( 1671)), sects. 2, 5, and 6). One suspects that More was also concerned about the question of particular providence (More, Antidote against Atheism (1653), Bk. II ch. 1.). This issue is indeed complex. It is false to claim that Descartes excluded God from the world after the creation. However, despite the Cartesian God's continual commerce with the material world, his character is closer to the impersonal deistic First Cause than the Christian God the Father. 87 Gerould de Cordemoy, Le Discernement du corps et de l'Âme ( 1666), fourth discourse, in P. Claire and F. Girbal (eds.), (Euvres philosophiques ( Paris: Presses Universitaires de France, 1968), 135-44; Louis de La Forge, Traité de l'esprit de I'homme ( 1666), ch. 14, in P. Claire (ed.), ( Euvres philosophiques ( Paris: Presses Universitaires de France, 1974), 235-46; Arnold Geulincx , Metaphysica vera, pt. 2, arts. 11-13, pt. 3, arts. 4-5, in J. Land (ed.), Opera Philosophica. 3 vols., ( The Hague: Nijhoff, 1892), ii. 176-9, 188-91; Nicolas Malebranche, Recherche de la Verité (16745), Bk. 6, pt. 2, ch. 3, in ( Elivres, ii; Pierre S. Regis, La Physique, Bk. 1, pt. 2, chs. 46, in Système de Philosophie ( 7 vols., Lyon, 1691), ii. As T. M. Lennon has recently argued in his "'Occasionalism and the Cartesian Metaphysic of Motion'" (in T. Penelhum and R. A. Shiner (eds.), New Essays in the History of Philosophy, Canadian Journal of Philosophy, supp. 1, pt. 1 (Guelph, Ontario: Canadian Association for Publishing in Philosophy, 1975), ( 29 40 ), Cartesian occasionalism extended to all worldly change, material and mental, so that mindbody occasionalism was but a special case of this more general analysis of causality. For a recent treatment of occasionalism (focusing on mind-body) in the Cartesians mentioned above and others, see Rainer Specht, Commercium mentis et corporis: Über Kausalvorstellungen im

Cartesianismus (Stuttgart-Bad Canstatt: Fromman, 1996). Even though Descartes says souls have, as a mode, the power of acting upon matter (cf. n. 84), since souls as well as particles of -306Descartes' treatment of motion are that God moves individual particles, and that the encounter of any one particle with others of specific size, speed, and direction gives God reason (or cause) to move the particle one way rather than another, according to the laws of motion. Such a view is, essentially, the doctrine of occasionalism applied to the material realm. This leads to the conclusion that the 'Cartesians' mentioned above were, at least on this one fundamental point, truly Cartesian; a conclusion that is prima facie trivial, historiographically surprising (to those who view occasionalism as a perversion of Descartes), but, upon reconsideration, not so surprising at all (so many Frenchmen couldn't have been wrong). V Those who have found an implicit dynamics in Descartes' physics have commonly erred by relying primarily upon contextual analysis at the expense of fitting Descartes' remarks about force into the broader scheme of his physics and metaphysics. Contextual analysis, of course, can be an important tool in the examination of a concept, especially when no explicit discussion of the concept is provided by the author in question (or when his usage contradicts this discussion). With such writers as Descartes, however, care must be taken in its application--the more so if one ignores the writer's own analysis. Descartes is a special case, because he quite often employed words according to the common usage throughout a work, only to indicate, when explicitly dealing with the concept in question, that some other reality lies behind the common usage, so that things are not as they seem at all. Colour is the best-known example. Descartes denied that colour is a real quality of matter--what is commonly called colour (the cause in objects of our colour sensations) is really nothing more than an arrangement of matter that imparts a certain spin to particles of matter so as to cause sensations of light and colour. 88 An example closer to our ____________________ 87 matter depend upon God's conserving power for their continued existence ( Principles, part i, art. 21; Third Meditation ([ A T VII49: C S M K II33]), one wonders to what extent the successive states of our soul, not to mention the effects of our will upon matter, must ultimately depend upon God. Finally, it would be interesting to analyse Cartesian occasionalism by comparison with that of St Augustine and Islam; cf. Majid Fakhry, Islaniic Occasionalisin and its Critique by Averroës and Aquinas ( London: Allen & Unwin, 1958), introduction. For one of the many statements that occasionalism is a perversion of Descartes, see Gueroult, 'M6taphysique', pt. 2, Malebranche. For a Cartesian who apparently did not follow the trend toward occasionalism, see Jacques Rohault, Traité de Physique ( Paris, 1671), pt. 1, ch. 10, arts. 12-13; English trans. by Samuel Clark, Rohault's System of Natural Philosophy ( London, 1723; repr. New York: Johnson, 1969), 45-6.

88

Optics, part i: 'colours are nothing else, in the bodies that we call coloured, than the diverse ways in which these bodies receive light and reflect it against our eyes' ( Olscamp, 67: A T VI85). -307-

subject is that of gravity. It is well known that Descartes did not regard gravity as a real quality of matter or as resulting from attraction; he put it into the same category as colour. Yet in a letter to Cavendish, written after the public ation of his account of gravity in the Principles, 89 Descartes could treat gravity as a cause: 'The causes that move a body are [1] the gravity of those of its parts that descend and [2] agitation, as much of those that descend as those that rise. The causes that hinder a body are the gravity of those that rise and the resistance of the air.' 90 If one were to isolate such a passage and analyse the usage of 'pesanteur', one might conclude that Descartes, in spite of his own protestations, had let gravity slip in as a real quality. Such a conclusion would surely be unwarranted, and would not recognize the mammoth task that faced Descartes and his contemporaries as they sought to reveal, by means of the mechanical philosophy, the reality lying behind the obscure and confused notions associated with such terms as colour, gravity, and (I might add) force, 91 as these terms were employed in common usage according to the prejudices of the senses, that Descartes so vehemently inveighed against. 92 Yet, as Descartes wrote to More, in his published work he did not embrace his views about God's direct role in motion as fully as he might have. So it would perhaps not be difficult (either for us or for Descartes' ____________________ 88 So surely when Descartes speaks of coloured objects, we are not to suppose that he thought colour was a real property of matter. 89 Principles, part iv, arts. 20-7. The postil to art. 20 reads: 'Explanation of the second action [of vortices] which is called gravity [gravitas, pesanteur]' ( H R1281 [ C S M K1268]). 90 Descartes to Cavendish, 2 Nov. 1646 ( A T IV559 ). Cf. Descartes to Huygens, 18 or 19 Feb. 1643 ( A T III 61.9): 'I consider also, concerning gravity, that it augments the velocity of bodies that it makes descend, nearly in the same proportion as are the times during which they descend.' 91 While colour, gravity, and force share in common (for Descartes) the fact that each has some other reality behind it than is apparent to the untutored senses, force ought to be assimilated to such qualities as colour and gravity (felt heaviness) only when one speaks of the sensation of force (as, e.g., the sensation that one is exerting force). As Gabbey ( 'Force and Inertia', 8-9) points out, it would be an error to assimilate force as the cause of motion to the category of subjective qualities. However, Gabbey is mistaken in concluding that force should therefore be given ontological status as a property of matter. According to Descartes, colour and other qualities are reducible to matter in motion (and the interaction of moving matter with sentient beings); the reality behind this matter in motion is not to be found in some more basic property of

the material world (such as the Leibnizian concept of force), but rather In God. It is according to the manner suggested in this paragraph that I would account for Descartes' usage (in various contexts) of the word 'force' as if force were an attribute of matter (e.g. Descartes to Mersenne, 11 June 1640 ( A T II75); to Mersenne for Bourdin, 29 July 1640 ( A T III113); to the Marquis of Newcastle [?] ( A T V135); Optics, part ii ( Olscamp, 75 -6, A T VI 94); Le Monde ( A T XI430)). To attribute ontological beliefs about force to Descartes on the basis of the fact that he uses the word in such contexts is something like claiming an astronomer is a closet Ptolemaicist after hearing him extoll the beauty of the sunrise.

92

-308more immediate successors 93 ) to mistake his view on that account, to detach the tendency toward rectilinearity from its direct dependence upon God, and to thereby find an implicit dynamics in Descartes' works. Such detachment would, on the surface, do little else than transfer the cause of the tendency toward rectilinear perseverance from dependence upon God to dependence upon the properties of matter. But, as Westfall's analysis makes clear, such a transference would leave Descartes' geometrically conceived matter loaded down with a pair of inexplicable 'forces'. Even beyond whatever obscurity is found in Descartes' treatment of 'force', the proclivity of recent commentators to 'dynamicize' Cartesian material substance is not difficult to understand. It might seem today that there is something odd about bringing divine immutability into physics as a constant. Indeed, when Richard Blackwell, in his comparison of Descartes' laws of motion with Newton's principle of inertia, expressed Descartes' view that God is responsible for rectilinear perseverance by saying that the reason for such perseverance is 'external . . . to the material world', whereas for Newton the reason is 'internal', 94 at least one historian found Blackwell's usage of 'external' (and, by implication, 'Internal') to be incomprehensible. 95 If such language is difficult for us to comprehend today, at a time when science and metaphysics are considered to be distinct, even incompatible, realms, it was well understood by at least one of Descartes' near contemporaries. Leibniz, after trying to make do with the notion that matter is inert (in the Cartesian sense) and finding that he could not, was led to attribute an inherent force to matter. He emphasized that this force is internal to matter, to differentiate his position from those who referred all force to something external to matter, that is, to God himself: We have elsewhere explained that there was something contained in material things which has to be added to mere extension, and is really prior to it, viz., a natural force implanted by the creator in all things. . . . As it is not our concern now to refer this force miraculously to God himself, we must assume that it has been placed by him in bodies themselves, verily to constitute their innermost nature. 96 ____________________ 93 Some of his successors read Descartes as asserting that the machine of the universe runs independently of God, while others did not (see n. 87 above). Further study of

these two views among Descartes' followers and detractors is needed. Blackwell's entire statement is: 'A body perseveres in its state of motion of rest for Descartes because this is demanded by the immutability of God. The same holds true for Newton because of the body's mass. In the first case the reason is external, in the second case internal, to the material world' ( 'Descartes' Laws of Motion', 234). 95 Gabbey, 'Force and Inertia', 52 n. 125. 96 Leibniz, Specinieti Dynamicutii, trans. Wiener, Leibniz, 119-20. 94

-309Leibniz asserted, against 'those few exceptional men . . . [who] deny any internal force and activity in things', that the reason 'for any natural truth is not to be sought immediately in the activity or will of God, but rather in the fact that God has enclosed in things themselves properties and determinations from which all their predicates can be explained'. 97 Rather than viewing laws of nature in the manner of Descartes, as continuing manifestations of divine activity and thus as laws of God's activity, Leibniz viewed the laws as following from the properties that God has instilled in matter itself. This last point concerning Leibniz and Descartes--and, for that matter, my central point concerning Descartes' notion of force--might seem to lie more in the realm of metaphysics than of empirical science. If so, there is good reason: Descartes' physics is unintelligible without recourse to his metaphysics. Since our present modes of thought are in part the product of recent science, we may tend to brush aside the metaphysical component of Descartes' physics in order to get at the 'real science'. If any lesson is to be learned from an investigation of Descartes' notion of force, it is that such a move is unwise, and is liable to lead one to attribute to Descartes concepts that he would not claim, and to discover inconsistencies in his work that vanish in view of the metaphysics behind the physics. ____________________ 97 Leibniz, Specimen Dynamicum, 130-1. -310BIBLIOGRAPHY 1. EDITIONS OF DESCARTES Standard Editions ADAM C., and TANNERY P, (eds.), Euvres de Descartes ( 12 vols., rev. edn., Paris: Vrin/CNRS, 1964-76). The complete works in the original Latin or French. Referred to as 'AM'. COTTINGHAM J., STOOTHOFF R., and MURDOCH D. (eds.), The Philosophical Writings of Descartes, Vols. I and II ( Cambridge: Cambridge University Press, 1985). Referred to as 'CSM'.

COTTINGHAM J., STOOTHOFF R., MURDOCH D., and KENNY A. (eds.), The Philosophical Writings of Descartes, Vol. III, The Correspondence ( Cambridge: Cambridge University Press, 1991). Referred to as 'CSMK'. Other Editions ADAM C., and MILHAUD G. (eds.), Correspondence de Descartes ( Paris: Presses Universitaires de France, 1936-63). Referred to as 'AM'. ALQUIÉ F. (ed.), Descartes, Œuvres Philosophiques ( 3 vols., Paris: Garnier, 1963). ANSCOMBE E., and GEACH P. T. (eds.), Descartes, Philosophical Writings ( London: Nelson, 1969). Referred to as 'AG'. COTTINGHAM J. G. (ed.), Descartes' Conversation with Burman ( Oxford: Clarendon Press, 1976). Referred to as 'CB'. GILSON E. (ed.), Rena Descartes, Discours de la méthode, texte et commentaire ( Paris: Vrin, 1925; 4th edn., 1967). HALDANE E. S., and Ross G. R. T. (eds.), The Philosophical Works of Descartes ( 2 vols., Cambridge: Cambridge University Press, 1911). Referred to as 'HR'. HALL T. S. (ed.), Descartes, Treatise on Man ( Cambridge, Mass.: Harvard University Press, 1972). Referred to as 'Hall'. KENNY A., (ed.), Descartes: Philosophical Letters ( Oxford: Oxford University Press, 1970). Referred to as 'K'. MAHONEY M. S. (ed.), Descartes, The World ( New York: Abaris, 1979). MARION J.-L. (ed. and trans.), Règles utiles et claires pour la direction de l'esprit ( The Hague: Nijhoff, 1977). MILLER V. R., and MILLER R. P. (eds.), Descartes, Principles of Philosophy ( Dordrecht: Reidel, 1983). OLSCAMP P. J. (ed.), Discourse on Method, Optics, Geometry and Meteorology ( Indianapolis: Bobbs-Merrill, 1965). Referred to as Olscamp. VOSS S. (ed.), The Passions of the Soul ( Indianapolis: Hackett, 1989). WILSON M. D. (ed.), The Essential Descartes ( New York: New American Library, 1969). -313-

II. OTHER PRE-TWENTIETH-CENTURY WRITERS BAILLET ADRIEN, La Vie de Monsieur Des-Cartes ( Paris: Horthemels, 1961; photographic reprint, Hildesheim: Olms, 1972). HUME DAVID, An Enquiry concerning Human Understanding [ 1748], ed. L. A. SelbyBigge , rev. P. H. Nidditch ( Oxford: Clarendon Press, 1975). KANT IMMANUEL, Kritik der reinen Vernunft (Critique of Pure Reason), 1st edn. ('A'), 1781; 2nd edn. ('B'), 1787, in Kant's gesammelte Schriften ( Berlin: Reimer/ de Gruyter, 1900-), iv. 1-152. LEIBNIZ GOTTFRIED WILHELM, Die philosophischen Schriften von G. W. Leibniz, ed. C. I. Gerhardt ( Berlin: Weidmann, 1875-90). ----- Nouveaux Essais sur l'entendement humain [c. 1704], trans. New Essays Concerning Human Understanding, ed. P. Remnant and J. Bennett ( Cambridge: Cambridge University Press, 1981). LOCKE JOHN, An Essay concerning Human Understanding [ 1690], ed. P. H. Nidditch ( Oxford: Clarendon Press, 1984). MALEBRANCHE NICOLAS, Méditations chrétiennes et métaphysiques [ 1683], and Recherche de la Vérité [ 1674-5], in A. Robinet (ed.), ( Euvres completes de Malebranche ( Paris: Vrin, 1958-67). PASCAL BLAISE, Pensées [c. 1660], ed. L. Lafuma ( Paris: Éditions du Seuil, 1962). RÉGIS PIERRE-SILVAIN, Cours entier de philosophie ou système générale selon les Principes de M. Descartes (final edn., Amsterdam: Huguetan, 1691). ----- Système de philosophie ( Lyon: Thierry, 1690; facsimile edn., New York: Johnson Reprint Corporation, 1970). ROHAULT JACQUES, Entretiens sur la philosophie ( Paris: Michel le Petit, 1671). SEXTUS EMPIRICUS, The Outlines of Pyrrhonism [c. AD 1 90)], trans. R. G. Bury, Loeb Classical Library ( London: Heinemann; Cambridge, Mass.: Harvard University Press, 1933; repr. 1967). III. COLLECTIONS OF SECONDARY SOURCES COTTINGHAM J. (ed.), The Cambridge Companion to Descartes ( Cambridge: Cambridge University Press, 1992).

----- Reason, Will, and Sensation: Studies in Descartes's Metaphysics ( Oxford: Clarendon Press, 1994). DONEY W. (ed.), Descartes: A Collection of Critical Essays ( Garden City, NY: Doubleday, 1967). HOOKER M., (ed.), Descartes: Critical and Interpretive Essays ( Baltimore: Johns Hopkins University Press, 1978). MOYAL G., Descartes: Critical Assessments ( 4 vols., New York: Routledge, 1991). RORTY A. O. (ed.), Evsays on Descartes' Meditations ( Berkeley: University of California Press, 1986). Voss S. (ed.), Essays on the Philosophy and Science of René Descartes ( Oxford: Oxford University Press, 1992). IV. GENERAL BOOKS ON DESCARTES' PHILOSOPHY BECK L. J., The Method of Descartes ( Oxford: Clarendon Press, 1952). BEYSSADE J.-M., La Philosophie premiere de Descartes ( Paris: Flammarion, 1979). -314COTTINGHAM J., A Descartes Dictionary ( Oxford: Blackwell, 1993). ----- Descartes ( Oxford: Blackwell, 1986). ----- The Rationalists ( Oxford: Oxford University Press, 1988). GAUKROGER S., Descartes: An Intellectual Biography ( Oxford: Clarendon Press, 1995). GILSON E., Études stir le rôle de la pensée médiévale dans la formation du système cartésien ( Paris: Vrin, 1951). ----- Index scolastico-cartésien ( Paris: Alcan, 1913). GUEROULT H., Descartes selon l'ordre des raisons ( Paris: Montaigne, 1952; 2nd edn. 1968), trans. R. Ariew as Descartes' Philosophy Interpreted According to the Order of Reasons ( Minneapolis: University of Minnesota Press, 1983). KENNY A., Descartes ( New York: Random House, 1968).

LOEB L., From Descartes to Hume: Continental Metaphysics and the Development of Modern Philosophy ( Ithaca, NY: Cornell University Press, 1981). RODIS-LEWIS G., Descartes ( Paris: Libraire Générale Franqalse, 1984). SMITH N. KEMP, New Studies in the Philosophy of Descartes ( London: Macmillan, 1966). WILLIAMS B., Descartes: The Project of Pure Enquiry ( Harmondsworth: Penguin, 1978). WILSON M. D., Descartes ( London: Routledge and Kegan Paul, 1976). V. DOUBT BLUMENFELD D. and BLUMENFELD J. B., "'Can I Know I Am Not Dreaming?'", in M. Hooker (ed.), Descartes (see section III above), 234-55. BENNETT J., "'Truth and Stability in Descartes's Treatment of Scepticism'", Canadian Journal of Philosophy, supp. vol. 16 ( 1990), 75-108. CURLEY E., Descartes Against the Skeptics ( Oxford: Blackwell, 1978). FRANKFURT H. G., Demons, Dreamers, and Madmen ( Indianapolis: Bobbs-Merrill, 1970). MALCOLM N., "'Dreaming and Skepticism'", Philosophical Review, 65 ( 1956), 14-37-7 repr. in Doney (ed.), Descartes (see section III above), 54-79. WILLIAMS B., "'Descartes's Use of Scepticism'", in Myles Burnyeat (ed.), The Skeptical Tradition ( Berkeley: University of California Press, 1983), 18-39. VI. THE COGITO AYER A. J., "'I Think Therefore I Am'", in Doney (ed.), Descartes (see section III above), 80-7. FELDMAN F., "'On the Performatory Interpretation of the Cogito'", Philosophical Review, 82 ( 1973), 345-63. FRANKFURT H. G., "'Descartes' Discussion of his Existence in the Second Meditation'", Philosophical Review, 75 ( 1966), 329-56. HINTIKKA J., "'Cogito ergo sum: Inference or Performance'", in Doney (ed.), Descartes (see section III above), 108-39.

----- "'Cogito ergo sum as an Inference and a Performance'", Philosophical Review, 72 ( 1963), 487-96. MARKIE P., Descartes's Gambit ( Ithaca, NY: Cornell University Press, 1986). -315MARKIE P., "'The Cogito Puzzle'", Philosophy and Phenomenological Research, 43 ( 1982). SIEVERT D., "'Descartes' Self-Doubt'", Philosophical Review, 84 ( 1975), 51-69. -----, "Sellars and Descartes on the Fundamental Form of the Mental'", Philosophical Studies, 37 ( 1980). ZEMACH E., "'De Se and Descartes: A New Semantics for Indexicals'", Nous, 19 ( 1985), 181-204. VII. CLEAR AND DISTINCT IDEAS AND THE CARTESIAN CIRCLE DONEY W., "'The Cartesian Circle'", Journal of the History of Ideas, 16 ( 1955), 324-38. -----, "Descartes's Conception of Perfect Knowledge'", Journal of the History of Philosophy, 8 ( 1970), 387-403. ----- (ed.), Eternal Truths and the Cartesian Circle ( New York: Garland, 1987). FELDMAN F., "'Epistemic Appraisal and the Cartesian Circle'", Philosophical Studies, 27 ( 1975), 37-55. FRANKFURT H. G., "'Memory and the Cartesian Circle'", Philosophical Review, 71 ( 1962), 504-11. GEWIRTH A., "'The Cartesian Circle'", Philosophical Review, 50 ( 1941), 368-95. ----- "'The Cartesian Circle Reconsidered'", Journal of Philosophy, 67 ( 1970), 685-700. ----- "'Descartes: Two Disputed Questions'", Journal of Philosophy, 68 ( 1971), 288-96. LOEB L. E., "'The Cartesian Circle'", in Cottingham (ed.) The Cambridge Companion to Descartes (see section II, above). VIII. JUDGEMENT AND THE WILL, AND ETERNAL TRUTHS ALQUIÉ F., La Découverte métaphysique de l'homme chez Descartes ( Paris: Presses Universitaires de France, 1950; 2nd edn., 1987).

COTTINGHAM J., "'The Cartesian Legacy'", Proceedings of the Aristotelian Society, supp. vol. 66 ( 1992), 1-21. CURLEY E., "'Descartes on the Creation of the Eternal Truths'", Philosophical Review, 93 ( 1984), 569-97. DONEY W. (ed.), Eternal Truths and the Cartesian Circle ( New York: Garland, 1987). ETCHEMENDY J., "'The Cartesian Circle: Circulus ex tempore'", Studia Cartesiana, 2 ( 1981), 5-42. FRANKFURT H. G., "'Descartes on the Creation of the Eternal Truths'", Philosophical Review, 86 ( 1977), 36-57. GAUKROGER S., Cartesian Logic ( Oxford: Clarendon Press, 1989), ch. 2, pp. 60 ff. JOLLEY N., The Light of the Soul: Theories of Ideas in Leibniz, Malebranche and Descartes ( Oxford: Oxford University Press, 1990). KENNY A., "'The Cartesian Circle and the Eternal Truths'", Journal of Philosophy, 67 ( 1970), 685-700. -316IX. DUALISM AND THE MIND-BODY UNION BEDAU M., "'Cartesian Interaction'", Midwest Studies in Philosophy, 10 ( 1986), 483502. COTTINGHAM J., "'Cartesian Trialism'", Mind, 94 ( 1985), 218-30. GARBER D., "'Understanding Interaction: What Descartes Should Have Told Elizabeth'", Southern Journal of Philosophy, 21 ( 1983), 15-32. MALCOLM N., "'Descartes's Proof that his Essence is Thinking'", Philosophical Review, 74 ( 1965), 315-38; repr. in Doney (ed.), Descartes (see section III above), 313. RADNER D., "'Descartes's Notion of the Union of Mind and Body'", Journal of the History of Philosophy, 9 ( 1971), 159-70. RICHARDSON R., "'The Scandal of Cartesian Interaction'", Mind, 91 ( 1982), 20-37. ROZEMOND M., "'The Role of the Intellect in Descartes'" Case for the Incorporeity of the Mind', in S. Voss (ed.), Essays on the Philosophy and Science of René Descartes (see section III above), 97-114.

VAN J. CLEVE, "'Conceivability and the Cartesian Argument for Dualism'", Pacific Philosophical Quarterly, 64 ( 1983). X. THE PASSIONS, HUMAN BEINGS, AND ANIMALS COTTINGHAM J., "'Cartesian Ethics: Reason and the Passions'", Revue internationale de philosophie, 195 ( 1996), 193-216. HARRISON P., "'Descartes on Animals'", Philosophical Quarterly, 42 ( 1992), 219-27. HATFIELD G., "'Descartes' Psychology and its Relation to his Physiology'", in Cottingham (ed.), Cambridge Companion to Descartes (see section III above), 335-70. HOFFMAN P., "'Cartesian Passions and Cartesian Dualism'", Pacific Philosophical Quarterly, 71 ( 1990), 310-33. ----- "'The Unity of Descartes' Man'", Philosophical Review, 95 ( 1986), 339-70. RORTY A., "'Cartesian Passions and the Union of Mind and Body'", in Rorty (ed.), Essays on Descartes' Meditations (see section III above), 513-34. XI. CARTESIAN SCIENCE CLARKE D., Descartes' Philosophy of Science ( Manchester: Manchester University Press, 1982). COLLINS J., Descartes' Philosophy of Nature, American Philosophical Quarterly Monograph Series, 5 ( Oxford: Blackwell, 1971). GABBEY A., "'Force and Inertia in the Seventeenth Century: Descartes and Newton'", Studies in the History and Philosophy of Science, 2 ( 1971), 1-67; repr. in Gaukroger (ed.), below. GARBER D., Descartes' Metaphysical Physics ( Chicago: University of Chicago Press, 1992). ----- "'Science and Certainty in Descartes'", in Hooker (ed.), Descartes (see section III above), 114-51. GAUKROGER S. (ed.), Descartes, Philosophy, Mathematics and Physics ( Brighton: Harvester, 1980). -317GEWIRTH A., "'Experience and the Non-Mathematical in the Cartesian Method'", Journal of the History of Ideas, 2 ( 1941), 183-210.

GUEROULT M., "'The Physics and Metaphysics of Force in Descartes'", in Gaukroger (ed.), above. HESSE M. B., Models and Analogies in Science ( Edinburgh: Nelson, 1961). SABRA A. I., Theories of Light from Descartes to Newton ( London: Oldbourne, 1967). SHEA W. M., The Magic of Numbers in Motion: The Scientific Career of René Descartes ( Canton, Mass.: Science History Publications, 1991). -318-