Thinking about Knowing

OXFORD UNIVERSITY PRESS

Great Clarendon Street, Oxford ox2 6DP Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Auckland Bangkok Buenos Aires Cape Town Chennai Dar es Salaam Delhi Hong I(ong Istanbul Karachi Kolkata Kuala Lumpur Madrid Melbourne Ivlexico City Mumbai Nairobi Sao Paulo Shanghai Singapore Taipei Tokyo Toronto Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York © Jay F. Rosenberg, 2002

The nloral rights of the author have been asserted Database right Oxford University Press (maker) First published 2002 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, or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover and you must impose this same condition on any acquirer British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data Data available. ISBN 0-19-925133-9

1 3 5 7 9 10 8 6 4 2 Typeset in Calisto MT by Cambrian Typesetters, Frimley, Surrey Printed in Great Britain on acid-free paper by T.J. International Ltd Padstow, Cornwall

Introduction

This book was originally conceived as a series of loosely connected studies addressing topics, both central and peripheral, in contemporary epistemology. What gradually became clear in the writing of it, however, was that in the course of examining various episten10logical questions I inevitably found myself at the same time taking a stand on many of those questions, and-as the publisher's anonymous referees of the draft manuscript both insistedthis collection of stands taken constitutes at least the rudiments of a coherent and fairly comprehensive positive philosophical theory of knowledge. And both anonymous referees also suggested that I should stop being so coy about it and tell the reader straight out what that theory is. The point is well taken, and these introductory remarks are consequently intended to give a brief synoptic overview, first, of the central theses of that theory and, then, of the course of reflections by which I undertake both to develop and to defend them. As is always the case with introductory remarks, what is really in1portant, namely the detailed arguments, will be missing. In the balance of this introduction, then, epistemological doctrines will be mentioned, assorted names will be dropped, and various theses will be claimed to be effectively criticizable or defensible-all very dry and unsatisfying. The full text here summarily described, however, where the promised critiques and defences are actually offered, will, I hope, prove to be both less dry and more satisfying, and perhaps (although I recognize that this is highly unlikely) even convincing.

The positive theory The leading thesis developed in this book, to formulate it most dramatically, is that knowledge is simply adequately justified belief. To put it less dramatically but more accurately, we correctly judge that S knows that p whenever, from our de facto epistemic perspective, we judge S able adequately to justify

Introduction

his belief that p. Since from anyone epistemic perspective the judgements that S has done everything requisite to be entitled confidently to believe that p and that S has done everything requisite to establish the truth of p stand or fall together, a further "truth requirement" is vacuous and idle. Truth may arguably be an outcome of enquiry, but it can function neither as enquiry's goal nor as a constraining condition on any defacto epistemic policy or procedure. This book thus both implicitly and explicitly calls into question in a fundamental way many received understandings regarding the relationships among the concepts of knowledge, justification, and truth. The "perspectivalist" view here advocated, whose origins lie in Peircean pragmatism, is predicated upon explicit recognition of our inescapable epistemic situatedness. Unqualified attributions of knowledge are always made only from our own epistemic perspective, with reference to a particular context of enquiry, and enquiry itself is correlatively understood as always addressed to determinate questions, properly raised only within the framework of a set of background beliefs, themselves not then and there in question, which render appropriate the application of particular investigative procedures, techniques, and norms of epistemic conduct. The corresponding conception of justification is consequently proceduralist, holding that, in the first instance, it is conducts ofpersons that are justified or unjustified. A person's belief is justified, therefore, only if she is justified in believing it, and that will be so just to the extent that she is in a position to justify her believing it, an internalist consequence which I argue applies equally to inferentially derived theoretical and "immediate" spontaneous perceptual beliefs. The Cartesian theses that matter-of-factual knowledge both needs and has available incorrigible foundations are thus rejected in favour of a resolute anti-scepticism coupled to a thoroughgoingfallibilism. Those, then, in short compass, are the chief claims and contentions advanced and defended in the course of this book. Thus sensitized, an attentive reader should be able to detect most of them at work throughout the book, but with various degrees of implicitness. As explicitly formulated t~e­ ses, however, they surface at different times and in a different order. The reason is that, as is my unrepentant custom, I undertake to develop and expound my own philosophical views dialectically, through a careful critical confrontation with and among alternative views and the arguments offered by their advocates. As one of the publisher's referees remarked, the result is not an easy read: "A typical chapter begins by launching the reader into some detailed arguments, with little initial sense of the shape of the chapter as a whole, and relying upon the reader's understanding of the problems that 2

Introd uction

concern him and the assllmptions on which he relies." ("For those equipped to hang 011," he charitably added, "it can be an exhilarating ride" .) The point is again well taken, and I have made some effort, probably inadequate, to address it in the present version. Some additional help, however, may also be provided by the following:

Outline of the chapters As did contemporary epistenl010gy itself, I begin with Rene Descartes. Descartes is often said to have presented, in his first Meditation, a compelling intuitive argu.ment for scepticism regarding our knowledge of the natural world of causally interacting objects in space and time; that is, an argument in support of the claim that it is possible to doubt that any such world exists. Chapter 1 opens with an assessment of this imputation. I argue that a close reading of the original text fails to support the received view. Even when granted a variety of charitable exegetical concessions, Descartes' reasoning turns out to rest upon a series of problematic and highly conte11tious presuppositions quite sufficient to vitiate its claim to be either intuitive or compelling. In the balance of the chapter, I argue that the same holds true for a variety of sceptical reasonings in the general Cartesian style; that is, arguments which appeal to "closure principles" and to our supposed inability to rule out sceptical counter possibilities incompatible with either the truth of or our knowledge of what is supposedly known. The central constructive aim of the chapter, then, is to underscore and support the positive theory's steadfast anti-scepticism; in particular, the thesis that knowledge cannot be undermined by mere possibilities, but only by possibilities which there is some positive reason to suppose actually obtain. Complementary to this resolute anti-scepticism is a thoroughgoingjallibilism with respect to matter-of-factual convictions, which is the central constructive theme of Chapter 2. More specifically, the chapter is addressed to the notion of certainty which Descartes' own epistemological picture sets in opposition to such ostensible global fallibilism. In particular, I consider and find wanting the characteristically Cartesian contention that there is a unique mental operation-a species of epoche, paradigmatically illustrated by the cogitowhich not only suspends judgement and retreats from any "thick" claim to objective truth, but also yields a "thin" truth-claim regarding subjective matters with respect to which we are epistemically infallible. There is, in short, supposed to be an essential connection between certainty and subjectivity. 3

Introduction

The relevant conception of subjectivity turns out to be more difficult to locate than one would initially suspect. The search takes me, inter alia, into the difficult territory of Kant's distinction between "original apperception" and "inner sense", for which I attempt to provide some suitable exegetical enlightenment. My overall conclusion is that neither the deliberate suspension of objective commitments in an act of epoche nor the modes of causal self-affection that Kant thematized under the rubric "inner sense" issues in a class of incorrigible Cartesian certainties. A pure epoche wou.1d yield something that is in a trivial and vacuous sense "infallible"-where no truth-claim is made, no error is possible-but which, for just that reason, is not a truthsusceptible judgement at all, subjective or otherwise. An exercise of "inner sense" indeed issues in a judgement regarding something properly deemed subjective-e.g. a report of the occurrence of a perceptual experience-but it does not exclude the possibility of error. The illusion of subjective certainties results from the fact that an ordinary claim of perceptual appearing-a 'looks'-, 'sounds'-, 'feels'-, 'smells'-, 'tastes'-, and so, generically, 'seems'judgement-combines a trivially "infallible" objective epoche with such an error-susceptible subjective report. Such judgements of perceptual appearing have traditionally been cast in ostensibly foundational epistemic roles. Struck by the large number of falsehoods that he had accepted in his youth and by the highly doubtful nature of the "edifice" that he had subsequently based on them, Descartes proposed "to demolish everything completely and start again right from the foundations". Correlatively, he regarded his method of doubt and the incorrigibly certain beliefs to which its successful application was to lead as two moments in the process of securing such epistemic foundations, upon which he could then"establish [something] in the sciences that was stable and likely to last" . Chapter 3 engages with the question of whether we can expect to find such "foundational" beliefs, beliefs which would constitute "direct" or "immedi.., ate" independent knowledge of a family of facts capable of serving as the ultimate court of epistemic appeal for all factual claims about the world, both particular and general. My particular stalking-horse in Chapter 3 is consequently the confrontation between Wilfrid Sellars' well-known critique of the "Myth of the Given" and, to begin with, William Alston's notion of "immediate knowledge". The latter, however, is ab initio ambiguous between the notions of knowledge that is underived, i.e. non-inferential or spontaneous knowledge, and knowledge that is epistemically independent, i.e. knowledge that does not depend for its epistemic status on other knowledge. Sellars finds no fault with the conclusion that 4

Introduction

some knowledge must be immediate in the former sense-not all knowledge can be derived from other knowledge-but argues that the idea of knowledge that is immediate in the second sense rests on blurring the distinction between non-propositional causes and propositional contents. The only normative paradigm of a belief being "based on" particular "grounds" is the paradigm of inferential support; where the ostensible "grounds" do not have the logical shape of a proposition, a belief can be "based on" them only in the sense of being a response to them. Sellars' rejection of epistemically independent knowledge is thus reflected in a strong epistemic internalism, according to which one's non-inferential (e.g. perceptual) belief counts as knowledge only if one knows that its spontaneous occurrence is, in normal circumstances, a reliable indicator of the fact believed. Since the positive theory of knowledge here advocated in fact endorses such a comprehensive strong internalism, a goodly portion of Chapter 3 is dedicated to exploring the integrated normative account of epistemic justification, language-mastery, concept-possession, and perceptual experience which gives rise to and supports those internalist convictions. Central to that account, and also to the theory of knowledge developed here, is an epistemicproceduralism, according to which the activity of justifying is prior in the order of understanding to the state of beingjustified, which consequently, whether predicated of beliefs or of believers, must be elucidated in terms of the notion of actual and potential justificatory practices. The chapter closes with an argument for the dialectical instability of "concept externalism" , the thesis, articulated by Franz von Kutschera and defended by Crispin Sartwell, that the notion of justificatory reasoning has no essential role to play in an account of the concept of knowledge; that is, that what is essential to knowledge is only true belief. The critical argument highlights the crucial internalist interplay between objective warrant and subjective confidence: A responsible and reflective epistemic agent proportions the strength of his subjective convictions to his objective epistemic entitlements. What secure those entitlements are the justificatory arguments available to him. It follows that spontaneous thoughts and judgements that are "immediate" or "foundational", in the sense of being in fact underived or non-inferential, nevertheless owe their normative epistemic character as expressions of knowledge to their potential embedding in justificatory reasoning practices. The notion of justification having assumed its traditional place of prominence, Chapter 4 turns to the classical "justified-true-belief" account of knowledge. Here is where the positive theory's perspectivalism comes explicitly and articulately into view. The centrepiece of the chapter is a critical 5

Introduction

exploration of Robert Fogelin's reinterpretation of the classical account in terms of a conjunction of two readings of the traditional justification clause, one "internalist" and "deontological", according to which it endorses the propriety of a subject's actual epistemic conduct, and one "externalist" and "evaluative", according to which it asserts the adequacy of the subject's de facto grounds to establish the truth of what is believed. Fogelin's analysis yields an insightful diagnosis of the challenge offered by familiar "Gettier problems" as essentially resting on a contrast of informational states. What is crucial to a Gettier scenario is that someone, given the information ex hypothesi available to him, has responsibly come to believe some claim to be true on grounds that we, who are stipulated to be privy to additional information, recognize do not establish its truth. I offer a revision of Fogelin's diagnosis which replaces his contrast between assessments of epistemic propriety and assessments of truth-determinativeness with one between two wholly "deontological" assessments of propriety from distinct epistemicperspectives. In the balance of the chapter, the resulting "perspectivalist" account is distinguished from such widespread "contextualist" views as that advocated by Keith DeRose, and then deployed critically against both Fogelin's own "Pyrrhonian" sceptical conclusions and the more moderate "elusiveness" and "instability" theses defended by David Lewis and Michael Williams, according to which epistemological enquiry per se generates a unique epistemic context within which matter of factual knowledge becomes impossible. Chapter 5 offers an (admittedly somewhat revisionist) "internalist-perspectivalist" interpretation, based on the positive conclusions developed and defended in preceding chapters, of G. E. Moore's epistemological views and arguments. In particular, I challenge Barry Stroud's diagnostic interpretations of Moore's anti-sceptical strategies and defend Moore's arguments against Stroud's well-known criticisms. The central constructive achievement of the chapter is a more detailed working-out of the perspectivalistpragmatist conception of enquiry as necessarily addressed to determinate questions and conducted within a setting of (defeasible) agreements regarding pertinent investigative methods, applicable epistemic norms, and matters of fact not then and there in question. In this context, the point of Moore's "defence of common sense" is to call our attention to the vast collection of matter-of-factual propositions that, in Wittgenstein's words, "we affirm without special testing"; propositions which supply the normative epistemic standard for closing off a sceptical dialogue of challenge and response with the observation that it is more reasonable to consent to a disputed conclusion than 6

Introduction

to pursue the further enquiries suggested by the latest challenges-an initial non-inferential, non-derivative epistemic authority that is also shared by the crucial premise of Moore's "proof of an external world", 'Here is one hand, and here is another' . Finally, Chapter 6 is devoted to the explicit exploration of the relationships between the notion of truth and such central episten1ic concepts as objectivity, justification, and, of course, knowledge. It begins by revisiting the "concept-externalist" thesis that what is essential to knowledge is only true belief, here freshly defended (by Ansgar Beckermann) as the only acceptable answer to the first question properly addressed by any systematic epistemological theory, "What is the goal of our cognitive efforts?" Appeals to reliable beliefforming mechanisms or to justificatory reasoning, in contrast, are best understood as addressed to subsidiary questions: "How can we reach that goal?" and "How can we ascertain whether and to what extent we have reached it?" On this view, the traditional "justified-true-belief" account of knowledge confusedly conjoins the answers to these distinct questions. Truth is the goal of enquiry; justification only a means or a criterion. In contrast, I argue that, despite almost universal agreement to the contrary, objective matter-of-factual truth simply cannot be the goal of our epistemic activities. Since nothing could count as a reason to conclude that we had finally arrived at such truths and thereby realized our putative goal, believing such truths cannot function in our practical reasoning as a goal. That is, there is no way to transform an ostensible commitment to the goal of objective truth into concrete epistemic policies and practices. This negative conclusion is characteristically pragmatist and, indeed, has already been drawn by C. S. Peirce, although his own "ideal-limit" epistemic conception of truth as "the opinion which is fated to be ultimately agreed to by all who investigate" fares no better on this front. Peirce's own positive suggestion is that enquiry aims at "the settlement of opinion" orfixedbelief, and while this proposal also clearly admits of various unsatisfactory interpretations, I offer one that both is arguably defensible and harmonizes reasonably well with the internalist-perspectivalist account of knowledge developed in earlier chapters. Correlatively, I suggest that Peirce's appeal to the method of science as the fundamental means of fixing belief-parsed "abductively" as in the first instance concerned with the explanatory accommodation of those perceptual experiences which we in fact find ourselves having-can fruitfully be understood as advancing a practice-based constitutive analysis of objectivity as such. For, the epistemic marks of objective matter-of-factual truth are fallibility and 7

Introduction

intersubjectivity. The former is guaranteed, on Peirce's account, by the essential open-endedness of scientific enquiry; the latter by the impersonality and repeatability of scientific observations and experiments. Peirce thus offers us the rudiments of a systematic epistemological theory significantly different from the received. tradition; a theory that explicitly acknowledges our inescapable epistemic situatedness, and so consistently remains both procedural and perspectival in a way that, as Peirce himself insists, precludes the possibility of global sceptical doubt in the Cartesian tradition. None of the theses that I defend in these six chapters is dramatically new, althop.gh some of the arguments are perhaps innovative, and some of the exegetical moves are, let us say, bold. Readers familiar with my previous philosophical work, however (if indeed there are any), will find little to surprise them. Like much of that work, the epistemological views developed here con1bine Kantian and Peircean insights in characteristically Sellarsian ways. I remain, that is, true to my philosophical heritage. This time around, those trusted and familiar conceptual and historical resources have provided me with useful tools both for critically engaging a nu.mber of contemporary epistemological positions and for argumentatively defending a relatively systematic constructive alternative. That their insights, in this way, continue to hold up as well as they have and do is the best reason I know of for concluding that Kant and Peirce and Sellars were generally very much on the right track.

8

The Myth of Cartesian Scepticism: Dreaming, Doubts, and Epistemic Closure

1

Doubt is a kind offear . .. and 'tis the same kind of Madness for a Man to doubt of anything, as to hopefor orfear it, upon a mere possibility John Wilkins 1

Scepticism is said in many ways. Its characteristic classical mark, however, is doubt, and so traditional divisions typically classify sceptics according to what they profess to doubt. Ontological sceptics, for instance, profess to doubt that items of one sort or another exist; epistemological sceptics that we blOW this or that to be the case; for example, that items of one sort or another exist. When talk is of Cartesian scepticism-so called because Descartes is usually thought to have given, in his first Meditation, an intuitive and compelling argument in support of its possibility-the items in question are the contents of a natural world; typically, a system of discrete interacting spatio-temporal objects. In particular, what Descartes is supposed to have shown is that it is possible to doubt that (we know that) any such natural world exists. Cartesian scepticism, that is, is a matter of global existential doubt regarding objects in space and time. The central question that animates this chapter is whether Descartes' putative demonstration of the possibility of such scepticism is as intuitive and compelling as it is custonlarily taken to be. Descartes himself, of course, was neither an ontological nor an epistemological Cartesian sceptic. He doubted neither that the natural world existed 1 Of the Principles andDuties ofNatural Religion, 7thedn. (London: R. Bonwicke, 1675), pp. 25-6. Cited by Christopher Hookway in Scepticism (London and New York: Routledge, 1990), p. 39.

Cartesian Scepticism

nor that we can know not only that the natural world exists but also a good deal about its particular contents and characteristics. Descartes' appeal to the notion of doubt was explicitly methodological, and what was methodologically relevant was never anyone's de facto doubts, but always the possibility of doubting this or that proposition, since that possibility was sufficient to exclude the proposition in question from a particular class of propositions that Descartes is concerned to locate and identify in the opening stages of his Meditations. 2 The opening remark of the Meditations adverts to the propositions at issue. They are what he here refers to as "the foundations": Some years ago I was struck by the large number of falsehoods that I had accepted as true in my childhood, and by the highly doubtful nature of the whole edifice that I had subsequently based on them. I realized that it was necessary, once in the course of my life, to demolish everything completely and start again right from the foundations if I wanted to establish anything at all in the sciences that was stable and likely to last. (AT 17 = CS 12)3

What we should particularly notice about this remark is Descartes' unquestioned confidence not only that the collection of propositions that he accepts or believes-of his opinions, as he usually puts it-has an implicit structure, but that it is a hierarchical structure. That the collection of one's opinions necessarily has some structure or other is not controversial. Propositions, after all, are precisely the sort of items that can and do stand in logical and evidential relationships, in virtue of which some of them can and do serve as reasons for others. But this much interrelatedness is compatible with a variety of structures. (The contemporary image of "the web of belief" comes to mind.) Descartes, however, evidently takes the correlative architectural metaphors of "grounds" 2 The use of 'propositions' here and subsequently is intended to be ontologically innocent and theoretically non-committal. I need a common noun to stand proxy for that-clauses, and 'proposition' will do as well as any. Propositions are thus what can be truly or falsely believed or disbelieved, since, typically, what one truly or falsely believes or disbelieves is that S, where' S' is replaced by some grammatical declarative sentence-but the reader is encouraged to refrain from drawing any provocative ontological or theoretical conclusions from such remarks. I shall also sometin1es say, for instance, that propositions are "expressed" by sentences and "given" by that-clauses, but these remarks, too, are lueant to be read innocently and non-committally. I deliberately avoid 'denoted' here, since I have learned that even charitable readers find it difficult to avoid thinking of denotation or reference as a relation implying ontological status for its relata. My book Beyond Formalism: Naming and Necessity jOr Human Beings (Philadelphia, Pa.: Temple University Press, 1994) offers a variety of contrary considerations which may help the charitable reader to resist that temptation. 3 From the translation of Meditations on First Philosophy by John Cottingham, Robert Stoothoff, and Dugald Murdoch in The Philosophical Writings of Descartes, ii (Cambridge and New York: Cambridge University Press, 1984). Citations here marked "CS" will be from that version; those marked "AT" give the page numbers of the Adam and Tannery edition of the Latin text, given marginally by Cottingham et al.

10

Cartesian Scepticism

and "support" entirely literally. Some of his opinions are "based on" others, and the entire "edifice" admits of being "demolished" and cleared away to reveal its "foundations". Descartes' confidence here, I want to suggest, reflects his tacit acceptance of what we might call the formalist paradigm in epistemology, according to which, explicitly or implicitly, mathen1atical knowledge is treated as a model for knowledge in general. Thereby hangs a classical dialectic. Mathematical truths are traditiol1ally conceived of and characterized as universal, impersonal, a-perspectival, necessary, and apodictically certain. Fundamental to any understanding of our epistemic status vis-a-vis the natural order, however, is the fact that we are situated beings. That is, the natural world is epistemically available to us, if at all, only through perceptual takings, sensuous episodes of perceiving something as something which functionally embody aspects of demonstration and classification. The world that is sensorily present to us is always present only in perspective; that is, from a determinate spatio-temporal point of view. On the face of it, then, empirical knowledge would have to be fundamentally particular, personal, perspectival, contingent, and inden10nstrable. The formalist paradigm thus immediately generates a tension between the conviction that knowledge is fundamentally manifested in the exercise of our capacity to know "eternal" mathematical truths and the commonsensical idea that we are also able to come to know matter-of-factual truths about the natural order. The classical, Socratic-Platonic, response to this tension was to deny the very possibility of empirical knowledge. Allthentic knowledge (episteme) must be of reality, and what is real are the eternal and independent Forms, which cannot be sensed but only grasped by the intellect. ,The provenance of knowledge proper, then, is reason alone, and the very idea of authentic empirical, i.e. experiential, knowledge is intrinsically incoherent. Sensory experience informs us only about appearances, and so, with regard to the world available to us through the senses, the world of such appearances, we can aspire only to opinion (doxa). The emergence of modern science dramatically challenged this classical contrast between episteme and doxa by asserting the mathematical lawfulness of .the experienced world. Galileo and Kepler and later, most strikingly, Newton professed epistemic access to natural necessities; that is, to mathematically representable universal, impersonal, and a-perspectival truths about the natural order which are suited to be objects of knowledge properly so called. The epistemological task which Descartes sets himself at the outset of his Meditations is precisely to vindicate an entitlement to such scientific 11

Cartesian Scepticism knowledge of the sensorily encountered world, an epistemic entitlement to something "in the sciences ... stable and likely to last" . What may have suggested to Descartes that such a project made sense within the formalist framework was the stunning intellectual accomplishment represented by his analytic geometry; that is, the reduction of the classical "science of space" to a series of deductions within the algebraic "science of number" . Why should not something analogous be equally feasible for the emerging "science of nature"? At this point in the dialectic, in other words, a commitment to the formalist paradigm manifests itself in a presumptive demand for a regressive structure of justification terminating in "self-evident truths". Just as one's epistemic entitlement to a theorem of (analytic) geometry is secured by exhibiting it as the terminus of a chain of logical derivations originating in "self-evident" (algebraic) axioms and postulates, the "original foundations" of our empirical knowledge would similarly need to be ostensibly "self-disclosing" truths in which any such regressive justificatory reconstruction within a mathematical natural science would necessarily terminate. It is this dialectical demand of "reason" specifically in service of the formalist paradigm, I think, that persuades Descartes to adopt a methodological principle and an epistemic practice both of which, from any commonsensical or practical perspective, would be utterly unreasonable; namely: that I should hold back my assent from opinions which are not completely certain and indubitable just as carefully as I do from those which are patently false. So, for the purpose of rejecting all my opinions, it will be enough if I find in each of them at least some reason for doubt. (AT 18 = CS 12)

There's quite a bit going on in this short passage. Setting aside for a moment the question of just what counts as a "reason for doubt" in ordinary, practical contexts, we can nevertheless observe that when one is confronted with such an ordinary reason for doubting some mundane propositionp which one has hitherto accepted, the appropriate practical epistemic reaction is typically to modify one's degree of confidence in the truth of p. In the first instance, that is, 'doubt' adverts to a family of epistemic attitudes, correlative to and contrasting with certitude. To be in doubt is characteristically to be epistemically irresolute, to be to some extent unsettled in one's convictions. The ordinary notion of doubt thus belongs to a picture which includes a gradated spectrum of epistemic attitudes toward a given proposition falling between two distinct extremes of certitude: complete confidence that the proposition is true and complete confidence that it is false. This is the notion that C. S. Peirce had in mind when he thematized the removal of doubt as the fixation of belief. 12

Cartesian Scepticism

But it is apparently not quite the notion that Descartes has in mind. Notice that, in Descartes' idiom, it is in the first instance opinions, not persons, that are or are not "completely certain", al1d reasons for doubt are not, so to speak, adventitious to these opinions but rather in some sense to be found "in" them. The Cartesian notion of doubtfulness, in short, is evidently the notion of an intrinsic episten1ic feature of the proposition believed, correlative to and contrasting with indubitability or certainty. In contemporary discussions the attitude of certitude is sometimes introduced under the rubric "psychological certainty" and contrasted with the ostensible propositional feature of "logical certainty" , misleadingly suggesting that we are dealing with two species belongil1g to a single epistemic genus. The supposed propositional epistemic features of doubtfulness and certainty nevertheless plainly have something to do with the personal epistemic attitudes of doubt and certitude, and it is consequently reasonable to ask how, on Descartes' view, the two families of concepts are related. His occasional textual equations of doubtfulness with dubitability and certainty with indubitability obviously suggest a straightforward modal analysis: (dl)

A proposition, p, is doubtful, i.e. not certain, whenever it is possible (for someone) to doubt, i.e. to be epistemically irresolute about, whether p.

The san1e text, however, also offers considerations which, at least prima facie, directly confute (dl). For instance, we shortly find Descartes speaking of "beliefs about which doubt is quite impossible, ... for example, that I am here, sitting by the fire, wearing a winter dressing-gown, holding this piece of paper in my hands, and so on" (AT 18 = CS 13). Nor is it clear whether Descartes would concede that it is genuinely possible to be epistemically irresolute or unsettled regarding what he calls "the simplest and most general things"; for example, that "two and three added together are five; and a square has no more than four sides" (AT 20= CS 14).4 This much at least is clear: Once Descartes reaches the conclusion that some proposition is doubtful, although he begins by speaking of withholding assent from it, his self-directed methodological injunction is ultin1ately not, in the manner of a Pyrrhonist epoche, to adopt an attitude of doubt regarding that proposition by suspending belief. Descartes' goal is not ataraxia but 4 An exquisitely guarded statement! Surely it is equally obvious that a square does not haveftwer than four sides. Or is Descartes perhaps worried that he has inadvertently counted one or more of the square's sides several times?

13

Cartesian Scepticism

something "in the sciences ... stable and likely to last". To this end, he instead reduces the quotidian spectrum of epistemic attitudes to just two, its normal end-points, methodologically adopting a stark epistemic dichotomy of opposed certitudes, unconditional affirmation and unconditional denial. Concretely, he proposes not just to modify his degree of confidence in and "hold back ... assent" from propositions for which he can identify any "reason for doubt", but rather definitively to "reject" any such proposition. His methodological counsel is not the suspension of belief but rather active disbelief. Descartes thus begins his Meditations with an epistemic pretence. He will treat propositions for which he can locate a "reason for doubt"-doubtful propositions-as if he were completely confident that they were false (alternatively, to sound another note, as if he knew them to be false). 5 This is one of the ways in which his method of doubt is exaggerated or "hyperbolic" . Peirce notoriously denies that such a radical pretence is genuinely possible: We cannot begin with complete doubt. We must begin with all the prejudices which we actually have. . . These prejudices are not to be dispelled by a maxim, for they are things which it does not occur to us can be questioned. Hence this initial scepticism will be a mere self-deception, and not real doubt ... Let us not pretend to doubt in philosophy what we do not doubt in our hearts. ("SCFI" v. 265)6

And elsewhere he calls the methodological stance urged by Descartes a "n1ake-believe", a "state of mind in which no man ... actually is", and insists that all enquiry, including philosophical enquiry, unavoidably begins in a state of mind "in which [one is] laden with an immense mass of cognition already formed, of which [one] cannot divest [oneself, even] if one would". "[If] pedantry has not eaten all the reality out of you," he writes, then you must recognize that "there is much that you do not doubt, in the least", and "that which you do not at all doubt, you must and do regard as . . . absolutely true" ("WPI" v. 416). 7 The extent to which Peirce's reservations are plausible and relevant to Descartes' project crucially depends on how we are to understand the notion 5 The spectrum of epistemic attitudes reappears only in the Fourth Meditation, with the introduction of indifference (a literal "suspension of judgement", neither affirming nor denying) as the proper response to "every case where the intellect does not have sufficiently clear knowledge at the time when the will deliberates" (AT 59 =CS 41). 6 C. S. Peirce, "Some Consequences of Four Incapacities" ("SCFI"), v. 208-317, in Charles Hartshorne and Paul Weiss (eds.), Collected Papers of Charles Sanders Peirce (CPCSP) , 6 vols. (Cambridge, Mass.: Harvard University Press, 1931-5). Citation of CPCSPis by volume and paragraph numbers. (All abbreviations used in the text can be found in the Bibliography.) 7 C. S. Peirce, "What Pragmatism Is" ("WPI"), The Monist (1905), in Hartshorne and Weiss (eds.), CPCSP, v. 411-34.

14

Cartesian Scepticism

of a "reason for doubt" . I have already suggested that the contexts which ultimately interest Descartes are neither the apodictic precincts of pure mathematics nor the practical precincts of mundane affairs, but rather fall within the new territory of mathematical natural science being opened by Ga1ileo and Kepler, a region which appears to share epistemic features with both the demonstrable and the practical. We might well expect Descartes to ask explicitly, then, what a "reason for doubt" looks like in connection with, for example, Kepler's conviction that the planetary orbits are not circles but ellipses having the sun as one focus, or Galileo's claim that the distance traversed by a freely falling object is directly proportional to the square of the elapsed time of fall. Instead, he proposes a more radical course, reflecting his hierarchical and foundationalist commitments: "I will go straight for the basic principles on which all my former beliefs rested." Whatever I have up till now accepted as most true I have acquired either from the senses or through the senses. But from time to time I have found that the senses deceive, and it is prudent never to trust completely those who have deceived us even once. (AT 18 =CS 12)

What should we make of the contrast between what is acquired "from the senses" and what is acquired "through the senses"? What, for that matter, are "the senses"? At this stage of Descartes' reflections there is no reason not to adopt the classical catalogue ill its traditional naive acceptation; namely, sight, hearing, smell, taste, and touch regarded as organically based faculties mediating one's perceptual awareness of n1aterial things. 8 On the face of it, then, what is acquired· "from the senses" will be information about (oneself in relation to) items in one's perceptual environment, information expressible in the form of perceptual judgements; for example, again to anticipate, "that I am here, sitting by the fire, wearing a winter dressing-gown, holding this piece of paper in my hands, and so on" (AT 18 = CS 13). But what, on this reading, could then be acquired through, but not from, the senses? As it turns out, what Descartes has in mind is the familiar everyday distinction between first-hand ("eye-witness") knowledge, e.g. perception, and second-hand ("hearsay") knowledge, e.g. testimony. Thus, one knows about, for instance, the circumstances of one's birth "througll" the senses-by seeing 8 At the beginning of the Second Meditation sensing is still characterized as something that "surely does not occur without a body" (AT 27 = CS 18), thus as organically based. Descartes' first recharacterization in turn-on his way to reconceptualizing sensory perception "strictly" as "simply thinking"-brackets the question of an organic basis, but still casts having sensory perceptions as being "aware of bodily things" (as it were through the senses) (AT 29 =CS 19), thus as mediating perceptual awareness of material things.

15

Cartesian Scepticism

photographs, hearing one's parents' accounts, etc.-but not "from" the senses. This intended contrast could clearly use some further elucidation, but, here, I propose instead to treat the notion of a distinction between what is acquired "from" and "through" the senses as an occasion to begin to explore some interesting adjoining regions of the conceptual space of epistemology that is the overall object of the present study. To this end, I will suggest two alternative parsings of Descartes' distinction. To begin with, let us remind ourselves of the Socratic/Platonic doctrine of amnemnesis or recollection. Observe Socrates in the process of questioning Meno's young slave: Does not this line from one corner to the other cut each of these figures in two?- Yes. So these are four equal lines which enclose this figure?-They are.... Each of these lines cuts off half of each of the four figures inside it, does it not?-Yes. How many of this size are there in this figure?-Four. How many in this?-Two. What is the relation of four to two?-Double. How n1any feet in this?-Eight. Based on what line?-This one. That is, on the line that stretches fron1 corner to corner of the four-foot figure?-Yes. Clever men call this the diagonal, so that if diagonal is its name, you say that the double figure would be that based on the diagonal?-Most certainly, Socrates. 9

The conversation here plainly concerns a series of diagrams, figures scratched in the dirt or sand. Using demonstrative expressions, Socrates directs the lad's attention to various parts and features of the diagrams: "these figures", "these lines" . Finally, the boy points out a line-"this one"-suitable to serve as the side of the sought square of double area, the eight-(square-)foot square. The senses are unquestionably implicated in this process. The boy sees what Socrates shows him and hears what Socrates asks him; if he did not, no "recollecting" could occur. But, although it constantly adverts to the diagrams scratched in the dirt, the conversation is not about those diagrams as ephemeral items in its participants' perceptual environment. The conversation is about lines and squares and numbers, mathematical items not sensorily available in any perceptual environment. The diagrams function as "visual aids" for Socrates' "midwifery", helping the young slave to "recollect" a relatively elementary theorem of Euclidean geometry, to find the knowledge not within the diagrams, but "within himself" (84d). 9 As translated by G. M. A. Grube, in Plato: Five Dialogues (Indianapolis, Ind.: Hackett, 1981), 85a-85c.

16

Cartesian Scepticism

In contrast to information about (oneself in relation to) items in one's perceptual environment acquired "from the senses", then, mathen1atical knowledge is here attained only "through the senses". Indeed, the pattern illustrated in the Meno is a standard motif within the rationalists' epistemological repertoire: Sensory experience is not a source of (demonstrative) n1athematical knowledge, but it can serve as an indispensable occasion for such knowledge, a "trigger", required to activate and bring to explicit consciousness an innate knowledge which is otherwise only tacitly present. The Socratic model of midwifery and recollection strikingly reinforces what we might call the "epistemological individualism" of the formalist paradigm. On the one hand, the paradigm takes it for granted that mathematical enquirythe archetype for enquiry per se-is characteristically solitary and a-contextual. Analogously, midwives in general, however helpful, are in principle entirely dispensable. Correlatively, the aim of such enquiry on the Socratic maieutic model is not to bring its subject into a relationship to actualities independent of his epistemic state but to actualize a determinate innate but tacit epistemic state of the individual knower. The initially paradoxical idea of scientific, i.e. mathematical, knowledge of the sensorily encountered world at least begins to make formalist Cartesian sense once it is thought of on the Socratic model, as knowledge attained through, but not acquired from, the senses. The notion of sensory deceptiveness can serve as a point of entry into a further contrast suggested by the distinction between being acquired from and through the senses. We have so far been contrasting (necessary, mathematical) scientific knowledge of objects in general, Socratically attained "through the senses", with (contingent, matter-of-factual) information about the objects in one's immediate perceptual environment, received "from the senses". That what the senses characteristically deceive us about is "objects which are very small or in the distance" , however, suggests that epistemic virtue or vice (also) relevantly attaches to those contingent judgements which depend, inter alia, precisely on the character of one's perceptual environment. On this reading, what is received through the senses will be matter-of-factual information or misinformation about objects. But what, then, is received from the senses? It is tempting to answer, "sensations", but, although that is surely in some sense correct, it is important in this connection to be clear that when Descartes speaks about sensing in epistemic contexts he characteristically has in mind something thoroughly cognitive, and indeed, propositional. To properly appreciate this, however, we need to take a brief look at certain features of the dialectic in the second Meditation. In accordance with the formalist paradigm, Descartes notoriously undertakes to identify having a mind per se with 17

Cartesian Scepticism

having the capacity to reason. 10 His implicit prototype of "thinking" is the reflective deliberation of a mathematician in his study, proving a theorem of analytic geometry. Clearly, however, the sensory presentations and singular indexical cognitions prima facie intrinsic to perceptual experiences ill harmonize with this model. Descartes consequently engages in some remarkable conceptual gymnastics to reconcile the fact and phenomenology of such experiences with his prototypical cogitationes. At the beginning of the second Meditation the senses are explicitly affiliated with the body. Thus, reviewing the results of his n1ethodological doubt, Descartes writes: "[I] have just said that I have no senses and no body ... [But am] I not so bound up with a body and with senses that I cannot exist without them?" (AT 24-5 = CS 16). Shortly thereafter he groups sensing together with such plainly corporeal capacities as eating and walking: Sense-perception? This surely does not occur without a body, and besides, when asleep I have appeared to perceive through the senses many things which I afterwards realized I did not perceive through the senses at all. (AT 27 =CS 18)

The received notion of sensory perception that forms Descartes' point of departure thus implicates the existence of corporeal items in two different ways, through its in1plicit commitment to the existence of sense organs andin its role as a "success verb" contrasting with "appearing" to perceive through the senses (e.g. dreaming)-to the existence of perceived objects. A mere two pages later, however, Descartes has managed to bracket both of these existential commitments: Lastly, it is also the same "I" who has sensory perceptions, or is aware of bodily things as it were through the senses. For exan1ple, I am now seeing light, hearing a noise, feeling heat. But [it is possible that] I am asleep, so [I can suppose] 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' (sentire) is just this, and in this restricted sense of the term it is simply thinking. (AT 29 =CS 19; first emphasis mine)

This is a pregnant passage, and we will have both reason and occasion to return to it later. At this point, however, we can extract from it the conclusion that when Descartes speaks of "sensory perception" one thing that he evidently has in mind is a certain sort of belief, in particular, a belief regarding how matters stand with respect to the objects in his immediate environn1ent and their sensory qualities. Such a belief is precisely the sort of thing that can 10 Hence, for instance, his conviction that non-reasoning animals ("brutes") are simply sophisticated organic automata.

18

Cartesian Scepticism

be correct or incorrect. The senses will be "deceptive" whenever such a sensory belief does not fit one's actual surroundings. !fone is not being deceived, one thereby receives-"through the senses"; that is, through the medium of one's sensings-information about the nature and disposition of objects in one's perceptual environment; if one is being deceived, only misinformation. But what, on this reading, would be received "from the senses"? The answer implicit in the passage from the second Meditation is: whatever one seems to see, hear, and feel. But what is that? One answer perhaps also suggested by the passage is proper sensibles-"light", "a noise", "heat"-but, while there is surely something correct about that, at this point we must tread very carefully. In particular, although we may later uncover good reasons for doing so, we must here resist the temptation straightway to identify such proper sensibles with aspects of the perceptual state of the perceiving subject. As far as the present text is concerned, the light, noise, and heat that one seems to see, hear, or feel are ostensible features of one's perceptual environment-e.g. the light, noise, and heat ofafire-and they are perceived as ostensible features of that environment. In the first instance, that is, proper sensibles are prima facie contents of regions of space. I I Thus, for instance, there looks, sounds, and feels to be a bright yelloworange, crackling, warm fire (over there). On this account, propositional form is already implicated in that which one receives from the senses. The Cartesian way to think of what is received "from the senses", in other words, also seems to be as a series of spontaneous beliefs; for example, that there is a bright yellow-orange, crackling, warm fire situated thus and so (in relation to me) in space. Call these "sensory beliefs". The senses are "deceptive" whenever the spontaneous sensory beliefs received "from" then1 are false. 12 11

Compare Kant's notion of intensive magnitudes as "the real" in space (and time) (cf. A 166-76

= B207-18). Reflection on the role of space in such experiences of seeming to see, hear, and feel reveals that the light, noise, and heat in question here not only aren't aspects of the subject's sensory state, but indeed couldn't be. Not themselves being in space, aspects of an individual's sensory state couldn't (literally) instantiate the shapes and locations that proper sensibles seem to have. These observations form one useful point of entry into Wilfrid Sellars' celebrated critique of the "Myth of the Given" (see e.g. "The Lever of Archimedes", The Monist, 64 (1981), pp. 3-36). 12 Descartes does, of course, recognize what he calls the "second grade of sensory response", comprising "all the immediate effects produced in the mind as a result of its being united with a bodily organ which is affected ... Such effects include the perceptions of pain, pleasure, thirst, hunger, colours, sound, taste, smell, heat, cold, and the like", as well as a "first grade" which is "limited to the immediate stimulation of the bodily organs by external objects". But he makes it clear that when he speaks of the reliability or deceptiveness of the senses what he has in n1ind is the "third grade" of sensory response; that is, "all the judgements about things outside us which we have been accustomed to make from our earliest years": "It is clear that we are not here dealing with the first and second grades of sensory response, because no falsity can occur in them." (See the sixth set of Replies (AT 437-8 = CS 294-6).

19

Ca rtesia n Scepticism

On our first alternative parsing, what is received "from" the senses is distinguished from what is received "through" the senses as contingency is distinguished from necessity; that is, as contingent information about the objects in one's immediate perceptual environment is distinguished from necessary scientific knowledge of objects in general. Only in the context of a SocraticPlatonic identification of the real with what is necessary, eternal, and unchangeable can this distinction be identified with the further distinction between appearance and reality. On our second alternative parsing, in contrast, what is received "from" the senses is distinguished from what is received "through" the senses in essence as a hypothesis is distinguished from a conclusion, both, however, still concerned with the nature and disposition of objects in one's immediate perceptual environment. This distinction, in turn, can be straightforwardly identified with a distinction between how things seem and how they are, and so with a distinction between appearance and reality, but only in a non-Socratic-Platonic sense which allows that how things seem can turn out to be how things are. So far, that is, there is no reason not to understand a belief regarding how things seem as simply a guarded, cautious, or non-con1mittal belief about how things are. Descartes notoriously adopts the Scholastic transposition of this prima facie distinction between two modes of belief, tentative and settled, into two modes of being, "objective" and "formal", but, properly u.nderstood, a thoroughly cognitive model of sensory experience is arguably incompatible with the notion that "appearances" constitute a distinctive category of entities. It is thus interesting to notice that Descartes' epistemic terminology at this point in fact does not prejudge such ontological questions. What has proved son1etimes deceptive are "the senses", not "the (sensory) appearances". Despite its being "prudent never to trust completely those who have deceived us even once" (AT 18 = CS 12), Descartes does not proceed to con, clude that the fact that "the senses occasionally deceive us" is itselfa reason for doubting the truth of any specific sensory belief. On the contrary, as he himself goes on to point out, in the absence of special considerations, it would be literally mad to call into question most of one's ordinary sensory beliefs. Yet although the senses occasionally deceive us with respect to objects which are very small or in the distance, there are many other beliefs about which doubt is quite impossible, even though they are derived from the senses-for example, that I am here, sitting by the fire, wearing a winter dressing-gown, holding this piece of paper in my hands, and so on ... Unless perhaps I were to liken myself to madmen, whose brains are so damaged by the persistent vapours of melancholia that they firmly maintain they are kings when they are paupers, or say they are dressed in purple when they 20

Cartesian Scepticism are naked, or that their heads are made of earthenware, or that they are pumpkins, or made of glass. But such people are insane, and I would be thought equally mad if I took anything from them as a model for myself. (AT 18-19 =CS 12-13)13

On the epistemic picture implicit in these remarks, we continuously find ourselves with spontaneous sensory beliefs regarding the contents of our perceptual environment, most of which we are also quite properly inclined unreflectively to accept, endorse, and act upon. Such sensory beliefs, that is, are not per se intrinsically dubious or uncertain. Subsequent reflection, however, also reveals that these spontaneous sensory beliefs fall into two rough groups: Some of them, e.g. those concerning "objects which are very small or in the distance", are sometimes "deceptive"; others, Descartes says, concern matters "about which doubt is quite impossible". One natural interpretation of these remarks is to see the dialectic here as moving within the confines of a tacit theory regarding factors conducing to perceptual error, although Descartes does not explicitly put it that way. His wording suggests that any sensory belief about "objects which are very small or in the distance", for instance, is n10re likely to be "deceptive" than a belief regarding one's more immediate perceptual environment, "for example, that I am here, sitting by the fire, wearing a winter dressing-gown", but there is no reason to suppose that this is so, for example, with respect to the spontaneous sensory belief that some object is very small or far away. If we take Descartes to be adverting to some such implicit theory of factors conducing to perceptual error, then the basic sort of "reason for doubt" attaching to a given sensory belief will be the fact that it has been formed in unfavourable perceptual circumstances. Since sensory beliefs about, for instance, the shapes of distant objects are less reliable than the corresponding beliefs about objects in one's immediate perceptual environment, the fact that one's current spontaneous sensory belief concerns, for instance, the shape of a distant tower-is a belief, for instance, to the effect that the tower is round 14_gives one some reason 13 We should, I think, pause to appreciate that Descartes' epoch was populated by a much more impressive group of lunatics than our own. I venture to suppose that not one contemporary psychiatrist or psychologist has ever been approached by a prospective client whose complaint is that his head is made of earthenware or that he is a pumpkin. Nowadays it is all diffuse anxieties and sexual neuroses. It is not entirely clear how to explain this striking historical degradation of lunacy, but my own inclination is to blame the dramatic decline in occurrences of the persistent vapours of melancholia. What kind of insanity can one expect, after all, fronl such milksop causes as "stress" and "tension"? 14 "Sometimes towers which had looked round from a distance appeared square from close up; and enormous statues standing on their pediments did not seem large when observed from the ground." (AT 76 = CS 53).

21

Cartesian Scepticism

for suspecting that it may be false. It is indeed at least prudent "never to place our complete trust" in such sensory beliefs formed in such circumstances. Extremes of size or distance will be only two among indefinitely many conditions which might (be discovered to) negatively affect one's perceptual reliability vis-a-vis particular sensible features of objects. In general, however, it will not be a sensory belief's specific content as such which supplies a "reason for doubt", i.e. a reason for accepting the (modal) epistemic conclusion that the belief in question may be false, but rather that content considered in relation to an ancillary epistemic fact; for example, that the circumstances in which the belief spontaneously occurred are (known to be) unfavourable for reliable perception of the relevant sensible characteristics. Contrary to Descartes' programmatic methodological formulations, in other words, the reason for doubt in such cases is not to be found "in" the belief doubted. No such "reason for doubt", in contrast, will attach to sensory beliefs formed in theoretically optimal perceptual circumstances. On the present interpretation, these will be the beliefs about which, Descartes says, "doubt is quite impossible". It is important to see that this remark makes sense only if acceptance or endorsement is the legitimate" default" epistemic stance vis-avis such spontaneous sensory beliefs; that is, only if such beliefs are not intrinsically dubious or uncertain per se. The specific inference from content to doubtfulness that we have just been examining is mediated by a general premise regarding content, circumstances, and reliability, and so presupposes, inter alia, that the falsehood of specific beliefs formed in problematic circumstances, e.g. sensory beliefs concerning the shapes or colours of "objects which are very small or in the distance", can legitimately be discovered and corrected by appeal to spontaneous sensory beliefs formed in unproblematic circumstances; for example, the sensory beliefs with which one finds on~self on closer approach to the originally distant objects. Without such a specific history of known "deceptiveness" attaching to the circumstances of origin of particular spontaneous sensory beliefs, however, "rejecting" those beliefs is, as Descartes himself correctly observes, quite literally mad. Earlier, I identified one respect in which Descartes' methodological doubt is exaggerated or "hyperbolic"; namely, in treating what is shown to be doubtful as if it were known to be false (Descartes' "epistemic pretence"). At this stage of his argument, doubtfulness is established for individual sensory beliefs only through the citing of positive reasons for doubt attaching to determinate families of beliefs formed in determinate kil1ds of circumstances. The specific reason for doubting the truth of a particular spontaneous sensory belief regarding the shape of a given tower, for instance, will be that sensory 22

Cartesian Scepticism

beliefs about the shapes of distant objects have generally (regularly or frequently) proved to be unreliable (mistaken or "deceptive"). Only Descartes' hyperbolic epistemic pretence, however, leads from this observation to the conclusion that the belief in question should be rejected; that is, treated as false. Normally, all that is called for is a more modest modification of one's level of confidence. How likely it is that a particular spontaneous sensory belief is mistaken, that is, characteristically depends upon just how unfavourable the epistem.ic circu.mstances are in which it occurs. What it is reasonable to believe all things considered about, for example, the shape of a distant tower will depend upon one's overall track record vis-it-vis such beliefs, taking into account such disparate matters as the general acuity of one's vision, the actual distances involved, the character of the ambient lighting, the apparent material con1position of the tower, what is known about regional tower-building customs, and so on. The fact that the sensory belief concerns a distant object is indeed relevant to its epistemic appraisal, but, unless one has explicitly adopted Descartes' methodological pretence, it is never in and of itself epistemically determinative. Far from "undermining the foundations" of opinions "acquired either from ... or through the senses", then, the" deceptiveness" of particular spontaneous sensory beliefs is available to fund a "reason for doubt" only if the majority of such beliefs are epistemically unproblematic; that is, are "beliefs about which doubt is qu.ite impossible". At this stage of the argumel1t, doubtfulness is still tied to particular epistemic circumstances which one can have positive reason to suppose obtain. Such dubitable sensory beliefs necessarily form a proper subclass of spontaneous sensory beliefs in general. Plainly, Descartes will need to appeal to different sorts of considerations if he hopes to "undermine" the whole class of sensory beliefs per se without lapsing into prima facie madness. And so, notoriously, he does: As if I were not a man who sleeps at night, and regularly has all the same experiences while asleep as madmen do when awake-indeed sometimes even more improbable ones. How often, asleep at night, am I convinced of just such familiar events-that I am here in my dressing-gown, sitting by the fire-when in fact I am lying undressed in bed! Yet at the moment my eyes are certainly wide awake when I look at this piece of paper; I shake my head and it is not asleep; as I stretch out and feel my hand I do so deliberately, and I know what I am doing. All this would not happen with such distinctness to someone asleep. Indeed! As if I did not remember other occasions when I have been tricked by exactly similar thoughts while asleep! As I think about this more carefully, I see plainly that there are never any sure signs by means of which being 23

Cartesian Scepticism awake can be distinguished from being asleep. The result is that I begin to feel dazed, and this very feeling only reinforces the notion that I may be asleep. (AT 19 = CS 13)

Much of what is going on in this paragraph we can simply ignore. Descartes begins by remarking that he regularly has, while asleep, "all the same experiences ... indeed sometimes even more improbable ones" as his company of exemplary lunatics do when awake. IS The suggestion that someone who insists that his head is made of earthenware or that he is a pumpkin or made of glass is reporting, so to speak, a corresponding characteristic experience is as intriguing as it is implausible, but it is also not at this point centrally apposite to Descartes' project. His chief concern in the text at hand is surely rather with precisely those earlier "beliefs about which doubt is quite impossible". What is important is not that one can have strange and striking dreams but that one can have n1undane and boringly familiar ones. Similarly, we can safely ignore the paragraph's peculiar last sentence. Whether we interpret Descartes' talk of feeling "dazed" as adverting to a literal dulling or paralysis of his mental capacities or, more figuratively, to the less dramatic cognitive malaise of being confused and bewildered, the implicit suggestion that the experience is somehow (epistemically) relevant to the question of whether he is or is not asleep ren1ains utterly obscure. Descartes adduces no considerations at all that would serve to connect his dazed feeling with his beliefs about sleeping and dreaming. It is not immediately clear whether the balance of the paragraph actually presents an argument, but if it does, the intended proximate conclusion appears to be that "there are never any sure signs by means of which being awake can be distinguished from being asleep". Assuming this is the case, then, how is the conclusion to be understood, and what reasons, if any, does Descartes give us for supposing that it is true? What distinction between "being awake" and "being asleep", for instance, does Descartes have in mind here? Just what is it for which we are supposed to lack "sure signs"? This much seems clear: On Descartes' view "being asleep" (i.e. dreaming) resembles "being awake" in that, like a wakeful slJbject (a perceiver), a sleeping subject (a dreamer) also continuously finds himself with a series of spontaneous sensory beliefs. Now, we recall from our earlier anticipatory discussion of the second Meditation that, construed "strictly", having such spontaneous sensory beliefs ("sensing") is not perceiving-seeing light, hearing a noise, 15 I personally cannot imagine what experiences could be "even more improbable" than having a head made of earthenware or being a pumpkin, but if we take Descartes at his word here, many of his dreams ("regularly", he says) must have been real humdingers!

24

Cartesian Scepticism feeling heat, etc.-but only seeming to perceive (to see, hear, feel, etc.). It is therefore worth noting that although Descartes is evidently confident that such dream episodes are not actual perceptions, he seems never to question the view that they are actually beliefs. Alternatively, however, one might argue that if there is a difference between perceiving an X and dreaming that one is perceiving an X, then there is an analogous difference between actually having some sensory belief and dreaming that one is having that sensory belief; that is, that dreaming that one believes that p is not an instance of believing that p.16 On this view, the context "I am dreaming that ..." suspends all epistemic predicates. In particular, dreaming that (one believes that) one is or is not dreaming will not be an instance of believing that one is or is not dreaming. A genuine belief that one is not dreaming, in consequence, could never be mistaken-but, "vacuously" as it were, neither could a dream to the effect that (one believed that) one was not dreaming, since dream contents would no more be subject to such epistemic evaluations than, for example, poems are. 17 On Descartes' view, however, some dream episodes are indeed sensory beliefs, and so epistemically assessable as correct or mistaken. The essential difference between being awake and being asleep concerns the way in which the spontaneous sensory beliefs occurring in the two states are occasioned. In perception, they directly result from the action of "bodily things" on the senses; in dreaming, they do not. 18 It is important to be clear that the distinction between wakeful perceiving and dreaming is independent of that between true and mistaken or "deceptive" sensory beliefs. A distant square tower may (deceptively) look round, but a wakeful perceiver's spontaneous sensory belief to the effect that the tower is round is occasioned by the action of the object itself on his visual sensory apparatus no less than his non-deceptive sensory beliefs regarding the shape of a tower situated immediately before him in broad daylight. Correlatively, I can dream that I am "here in my dressing-gown, sitting by the fire-when in fact I am lying undressed in bed", in which 16 This is the view adopted by Norman Malcolm in "Dreaming and Skepticism", Philosophical Review, 65 (1956), pp. 14-37. 17 It is instructive to consider in this connection Rene Magritte's painting N'est ce pas une pipe, which depicts a (tobacco) pipe and the inscription "N'est ce pas une pipe" ("This is not a pipe"). Treating the inscription as subject to epistemic evaluation leads to a wonderful series of conceptual tangles. 18 The qualifier 'directly' acknowledges the possibility, which Descartes himself evidently accepts, that "the visions which come in sleep are like paintings, which must have been fashioned in the likeness of things that are real" (AT 19 =CS 13). If this is so, a dreamer's spontaneous sensory beliefs might also all ultimately be occasioned by the action of bodily things on his senses, although only remotely.

25

Cartesian Scepticism case the spontaneous sensory beliefs constituting my dream will be "deceptive" . But I can surely also dream that I am lying undressed in bed "when in fact I am lying undressed in bed", in which case my spontaneous sensory beliefs will be non-deceptively correct. 19 To begin with, then, being asleep apparently functions, analogously to extremes of distance and size, as a circumstance that has been found negatively to affect the reliability of families of spontaneous sensory beliefs. Indeed, Descartes in essence reports having discovered the general unreliability of dreams: "As if I did not remember other occasions when I have been tricked by exactly similar thoughts while asleep!" On the face of it, this remark makes sense only if the deceptiveness of a dreamer's spontaneous sensory belief that he is clothed in a dressing-gown and seated next to a fireplace can be discovered and legitimately corrected by appeal to the beliefs with which he subsequently finds himself upon awakening, much as the deceptiveness of a perceiver's spontaneous sensory belief that a distant tower is round can be discovered and corrected by appeal to the sensory beliefs with which he subsequently finds himself in more favourable perceptual circumstances; for example, on closer approach to the object. Furthermore, in the absence of specific positive reasons for supposing so, to believe that one is dreaming is prima facie no less mad than to believe that one is a pumpkin or has a head made of earthenware. What, then, does Descartes' argument gain by shifting our attention from deceptive perceptions to deceptive dreams? Descartes' proximate conclusion, let us recall, is t11at there are "never any sure signs by means of which being awake can be distinguished fronl being asleep". Given our interpretive progress to this point this amounts to the claim that there are no sure signs to distinguish those (generally reliable) sensory beliefs directly occasioned by the actions of objects on the sense organs (perceptions) from those (generally unreliable) sensory beliefs otherwise occasioned; for example, by electrochemical processes occurring in the brain during sleep (dreams). Yet Descartes also evidently suggests that it is possible 19 In contrast to various other interpreters, Barry Stroud-in his influential work, The Significance of Philosophical Scepticism (SPS), (Oxford: Oxford University Press, 1983)-is properly sensitive to this point. His reconstruction of a generally Cartesian scepticism thus invokes"counter possibilities" to p having the generic form of a non-contradictory proposition q-paradigmatically, the proposition that one is dreaming that p-that is (known to be) incompatible, not with p, but rather with one's knowing that p: "[But] if it is obvious ... that Descartes must know that he is not dreaming if he is to know that he is sitting by the fire, it cannot be simply because the possibility in question is known to be incompatible with what he claims to know. It is not" (SPS 29). (As we shall see when we turn explicitly to Stroud's arguments, whether such a corresponding "dream proposition" in fact is incompatible with one's knowing that p is a surprisingly complicated question.)

26

Cartesian Scepticism

to discover that some dreams contain so to speak deceptive counterfeits of wakeful perceptions, and, if this is so, then it must be possible justifiably and legitimately to classify at least some spontaneous sensory beliefs under that category (i. e. to recognize them as dreams). It is not too surprising, then, to find Descartes himself, at the end of the sixth Meditation, surveying the guiding epistemic principles for doing just that: [Dreams] are never linked by memory with all the other actions of life as waking experiences are. If, while I am awake, anyone were suddenly to appear to me and then disappear immediately, as happens in sleep, so that I could not see where he had come from or where he had gone to, it would not be unreasonable for me to judge that he was a ghost, or a vision created in my brain, rather than a real man. But when I distinctly see where things come from and where and when they come to me, and when I can connect my perceptions of them with the whole of the rest of my life without a break, then I am quite certain that when I encounter these things I am not asleep but awake. (AT 89-90 =CS 61-2)

On this account, the "perceptions" (i.e. sensory beliefs) constituting one's waking experience form a mutually spatio-temporally and causally coherent system. A dream, in contrast, consists of spontaneous sensory beliefs which are not "linked by memory" with the other beliefs belonging to this system. Just as a perceiver can legitimately correct some of his "deceptive" spontaneous sensory beliefs by appealing to others occurring in more favourable perceptual circun1stances, so too a dreamer can evidently legitimately reclassify some of his spontaneous sensory beliefs as dreams by appealing to their lack of coherence with the systematic perceptions which he recalls. But, if this is right, then does not such a lack of coherence serve precisely as a "sure sign" by which to distinguish being asleep from being awake? Well, not exactly. Descartes' reference to memory here sensitizes us to the fact that, in both instances, the episten1ic evaluations in question are retrospective. A perceiver corrects a "deceptive" sensory belief by appeal to later perceptions; a dreamer reclassifies some spontaneous sensory beliefs from perceptions to dreams on the basis of a subsequent appeal to memories of an otherwise coherent and systematic sensory experience. Strictly speaking, then, a lack of coherence is, at best, a "sure sign" by which to distinguish having been asleep from having been awake. 2o The continuous-present tense in 20 Why "at best"? Well, so far, I have been treating Descartes' expression "sure signs" as roughly equivalent to "good and sufficient evidence (or: indications)". There is some reason to suppose, however, that Descartes intends 'sure' quite literally as adverting to conclusive evidence or even some species of certainty, and, if so, coherence or its lack will then fail to provide a sure sign of the waking/sleeping distinction in any tense.

27

Cartesian Scepticism

Descartes' proximate conclusion, in other words, is not adventitious but essential. Furthermore, we may note that the sixth Meditation's coherence test is arguably impotent to resolve even retrospective questions regarding the epistemic status, as dream episodes or waking perception, of those mundane sensory beliefs which are Descartes' particular concern in the first Meditation. For, unlike his later hypothetical "perceptions" of a man appearing and disappearing ex and in nihilo, such spontaneous sensory beliefs as that he is seated beside the fireplace in his study, clothed in his dressing-gown, and confronting a sheet of paper are just the sorts of boringly ordinary beliefs which can coherently be "linked by memory" with the spatio-temporally and causally systematic collection of perceptions that constitute his waking life. 21 These last two observations highlight an important difference between dreaming and various other circumstances that can be discovered to affect the reliability of spontaneous sensory beliefs, such as extremes of size or distance. That an object is small or distant is itself in the first instance the content of a sensory belief. Size and distance, that is, are themselves sensible qualities, and so can be epistemically available to a perceiver, qua perceptual circumstances, in precisely the same way that the sensible qualities whose reliable ascription such circumstances can (be known to) negatively influence, e.g. shape and colour, are epistemically available to her, as concurrent aspects or elements of a single (synchronic) complex perceptual episode. Our most recent considerations, however, highlight the fact that this is not true of the perceptual circumstance occurring in a dream. Although being asleep is an epistemic circumstance which can (retrospectively be discovered to) negatively affect the reliability of spontaneous sensory beliefs, occurring in a dream is not itself a sensible quality, and so the fact that a given sensory belief is occurring in that unfavourable circumstance need not be, and characteristically is not, reflected in or indicated by any feature of the dream's content. Descartes' spontaneous sensory belief that he is clothed in his dressing-gown and seated beside the fireplace in his study concerns only familiar and highly salient "middle-sized" features of his immediate perceptual environment. There is, in other words, nothing about its content to suggest that the circumstances in which it occurs are among those (discovered to be) characteristic 21 Parenthetically, we should note that Descartes' hypothetical appearing and disappearing man is a commonplace of the world of Star Trek, whose inhabitants are regularly "beamed" from location to location. Although I will not take time to pursue them here, this observation can serve as the point of entry to a number of interesting branch dialectics, including the notion of systematic dreams and central memory-related aspects of Wittgenstein's "Private-Language Argument".

28

Cartesian Scepticism

for unreliable wakil1g perception. And, crucially, that content is also entirely non-committal, so to speak, regarding the question of whether the given spontaneous sensory belief is a (generally reliable) wakeful perceiving or a (generally unreliable) constituent of a dream. Descartes' claim that there are "never any sure signs by n1eans of which being awake can be distinguished from being asleep" thus amounts to the conclusion that nothing about the content of any spontaneous sensory belief indicates whether it is a wakeful perceiving, occasioned by the action of "bodily things" on the senses, or an episode occurring within a dream, occasioned in some other way during sleep (assuming that such episodes are properly characterized as beliefs in the first place). Nor should this come as a surprise; for, it is arguably the case that, irrespective of whether the beliefs in question are sensory or, for instance, purely mathematical, whatever I can believe, I can also dream. 22 One thing Descartes gains by shifting his attention from perceptions to drean1s, then, is a source of potential "deceptiveness" which is in a significant way content-independent. But does this move advance his project of establishing the dubitability of sensory beliefs per se; that is, of "undermining the foundations" of all those opinions "acquired either from ... or through the senses"? Descartes seems to think so; for, his next sentence apparently invokes precisely the "epistemic pretence" of falsehood that, according to his adopted method, is licensed by having found some reason for doubt attaching to a given belief: Suppose then that I am dreaming, and that these particulars-that my eyes are open, that I am moving my head and stretching out my hands-are not true. Perhaps, indeed, I do not even have such hands or such a body at all. (AT 19 = CS 13)

Evidently, then, Descartes takes himself to have established the existence of "reasons for doubting" that his eyes are open, that he is moving his head, that he is extending his hands, and perhaps even that he has hands, eyes, and a head of the familiar sorts. But what reasons for doubting? Of course, if he is dreaming, then that he is dreaming is indeed a reason for doubting each of these beliefs. Ex hypothesi, the spol1taneous sensory beliefs that occur during 22 Indeed, the sentential matrix "I dreamed that - - " is apparently completely catholic and non-restrictive with regard to its substituends, on the face of it admitting even falsehoods which could not plausibly be believed-e.g. "I dreamed that 2 + 3 equalled 17" and perhaps even "I dreamed that I did not exist" . (Long ago, when we were graduate students together, Michael Dunn reported having dreamed that he was a limit point on the real nutnber line.) The corresponding matrix in the continuous present, "I am dreaming that --", is less accommodating, but, as we have seen, raises interesting questions of its own.

29

Ca rtesia n Scepticism sleep as constituents of one's dreams are generally unreliable and, indeed, can be known to be so. Like waking perceptions occurring in unfavourable perceptual conditions, sensory beliefs occurring in dreams are legitimately subject to doubt. A reasonfor believing that he was (only) dreaming that his eyes are open, etc. would thus also be a reason for doubting that his eyes are open, etc. But does Descartes have any reason at all to believe that he is dreaming? Clearly, he does not. Indeed, the point of the preceding argumentation was precisely to establish that nothing about the content of his mundane occurrent beliefs, sensory or otherwise, could give him any reason at all to believe that he is dreaming, and it is difficult to see what other reasons Descartes might have for reaching that conclusion. 23 Nor does he here suggest any additional relevant considerations. Instead, he incorporates the belief that he is dreaming itself within his "epistemic pretence". He supposes it. On the face of it, this represents a clear departure from Descartes' announced method. Instead of finding a reason for doubting (in the ordinary, attitudinal, sense) that he is seated with open eyes beside his fireplace, he now assumes orposits such a reason. No considerations at all have been adduced, in other words, to show that the sensory beliefs in question are doubtful persein the only sense of 'doubtful' we have so far encountered. During the previous stages of the argument the doubtfulness of a belief was tied to its occurrence in particular epistemically problematic circumstances which one hadpositive reason to suppose obtained. Dubitable sensory beliefs thus necessarily formed a proper subclass of sensory beliefs in general. Since Descartes has no positive reason to suppose that he is dreaming, however, his spontaneous sensory belief that he is seated with open eyes beside his fireplace is not in that sense doubtful. At this point in his reasoning, then, Descartes can salvage his methodological commitments only by implicitly modifying the notion of a "reason for doubt" and, correlatively, what it means for a belief to be "doubtful" . His "reason for doubt" here cannot be that he is dreaming (for he may not be), nor that he has some reason to believe that he is dreaming (for he has no such reason). Indeed, in the light of the coherence considerations later adduced in the sixth Meditation, the very ordinariness of his spontaneous sensory beliefs arguably gives Descartes some reason to believe that he is not dreaming. If 23 This is not to say that stretches of experience cannot, on Descartes' view, warrantedly or justifiably be judged to be constituted by series of dream episodes at all. They indeed can be-but only retrospectively. What Descartes suggests cannot be justified or warranted is a first-person judgement to the effect that one is there and then dreaming; that is, the judgement one would express by asserting "I am dreaming here and now" .

30

Cartesian Scepticism

Descartes has a reason for doubting that he is seated with open eyes beside his fireplace, then, it can only be that he could be dreaming. The mere possibility that his spontaneous sensory beliefs could be episodes in a dream, and so occurring in an epistemic context known to be generally unreliable, now suffices to render those beliefs"doubtful" . What sort of modality is at issue here? What sort of possibility is the "mere possibility" that one is dreaming? Given the epistemic context within which Descartes invokes such possibilities, it is clear that logical contingency alone will not do. It is logically possible to dream that 2 + 3 =5 or even that one oneself thinks and/or exists, but that possibility does not suffice to render those beliefs doubtful. "[Whether] I am awake or asleep, two and three added together are five" (AT 20 = CS 14). A popular answer is "epistemic possibility", meaning thereby logical consistency with what one knows, but, while an argument in the Cartesian style could perhaps be formulated in terms of such a notion, it is highly questionable as a bit of exegesis. On the one hand, at this point in the Meditations, Descartes has not advanced any positive theses regarding what he does know. For all he has so far said, it might turn out that he knows nothing at all, and, should that be the case, the supposed epistemic modality would collapse into its alethic counterpart. On the other hand, it would be premature to conclude, on the basis of the arguments offered up to this point, that Descartes does not know that he's not dreaming. One would evidently need at least an additional pren1ise to the effect that such knowledge requires certainty and indubitability or, at least, the availability of "sure signs", and even if Descartes is in fact implicitly committed to such a premise, it is by no means obvious that he should be. Indeed, interpreting the notion of knowledge in terms of certainty and ind-ubitability at this stage of the dialectic renders the broadening of reasons for doubt to include "mere epistemic possibilities" question begging. The present proposal is that Descartes knows that he is not then and there dreaming that he is seated by the fire in his dressing-gown (briefly: knows that he is not dreaming that s) only if the proposition that he is not dreaming that s is certain and indubitable. According to the n1ethodological strategy operative throughout the first Meditation, however, a proposition will be certain and indubitable only when there can be found in it no "reason for doubt". Conversely, if Descartes can find a reason for doubting that he is not dreaming that s, then the proposition that he is not dreaming that s will not be certain and indubitable, and hence not something that he knows. But can Descartes find a reason for doubting that he is not then and there dreaming that s? What could that reason be? 31

Cartesian Scepticism

Well, if Descartes had a reason for believing that he was dreaming that s, then that would be a reason for doubting that he was not dreaming that s, but, as we have seen, one clear implication of his argument is that he does not and cannot, then and there, have any such reason for believing that he is then and there dreaming that s. But then, as far as I can see, Descartes' only other possible reason for doubting that he is not then and there dreaming that 5 is the alleged fact that he could be dreaming that s. But could he be? Well, on the proposed interpretation, he could be dreaming that 5 only if it is epistemically possible that he is dreaming that 5; that is, only if he does not know anything inconsistent with the proposition that he is dreaming that 5, and, in particular, only if he does not know that he is not then and there dreaming that s. And now the circularity should be apparent; for, what reason is there to suppose that Descartes does not know that he is not dreaming that s? Ex hypothesi, only that it is not certain and indubitable that he is not dreaming that 5; that is, that there is a reason for doubting that he is not then and there dreaming that s. But if we interpret the modality relevant to reasons for doubt in terms of compatibility with what is known, knowledge in terms of certainty and indubitability, and certainty and indubitability in terms of reasons for doubt, then in the absence of any independent non-modal reasons for doubting that he is not dreaming that 5, there is no way for Descartes to argue for the availability of his alleged modal reasons for doubting that. 24 It will help us to see what has gone wrong here, I think, if we consider how Descartes might know that he is not then and there dreaming that he is seated by the fire in his dressing-gown. How, for instance, do Iknow that I am not here and now (i.e. as I write these words) merely drean1ing that I am seated by my computer in my jeans and T-shirt? The most straightforward answer is that I know lots of other things which imply that I am not dreaming at all. I know that I woke up at about 7.30 this morning and haven't gone back to sleep since then, that it's now about three hours later, that the cold cereal I ate for breakfast a few hours ago didn't satisfy me, that I deliberated about what to wear today and decided that jeans and T-shirt would do since I was planning to spend the day at home working on this chapter, that I sat down at my desk shortly after breakfast, and that (apart from attending to some biological 24 An alternative formulation avoids explicit appeal to the notion of knowledge by identifying the possibility that one is (merely) drean1ing with the inability to rule out that hypothesis. The gain in clarity is only apparent, however, since asking what it means to be "unable to rule out" a given hypothesis immediately reintroduces the original exegetical difficulties. The most natural proposal, of course, is that one can rule out a hypothesis whenever one knows something inconsistent with it, in which case the new formulation becomes a mere notational variant of the "epistemicpossibility" reading which we have just found wanting.

32

Cartesian Scepticism

imperatives) I've been sitting here ever since. All of these things that I know, and many other things besides, collectively imply that I an1 not asleep and dreaming at all, and so, ipso facto, that I am not merely dreaming that I am seated by n1Y computer in my jeans and T-shirt. The point, to put it briskly, is that I know much more than simply what is presented in my immediately current spontaneous sensory beliefs, and that this additional body of knowledge can suffice to rule out the epistemic possibility that those sensory beliefs are (merely) episodes in a drean1. N ow, a Cartesian sympathizer will insist, of course, that I don't know any of these other things, but I submit that we have so far seen no sufficient reason to conclude that this is so. Indeed, what was supposed to show that I cannot know such things was precisely to be an argument for the conclusion that I cannot even know that I am seated at my computer in my jeans and T-shirt. And I am in fact quite willing to grant that if I can't even know that I am seated at my computer, then I also can't know the indefinitely many other things which imply that I am not asleep and dreaming. I'm even willing to grant that if all I know is what is presented in my immediately current spontaneous sensory beliefs, then I do not kl10W that I am seated at my computer wearingjeans and a T-shirt. But it is plainly question begging simply to assume that I cannot know all those other things in an argument intended to den1onstrate, by establishing that I cannot know that I am not asleep and dreaming, that I cannot even know that I am seated at my computer. Similarly, nothing in the first-Meditation reasoning that we've so far had occasion to survey implies that Descartes cannot know much more than what is presented in his immediately current spontaneous sensory beliefs, and, if this is so, then interpreting the modality relevant to the broadened sense of "reasons for doubt", i.e. the modality in "could be dreaming", as epistemic possibility is either gratuitous or question begging. What is needed here is rather a different relative modality, one which relativizes not to what is known (in some unexplicated and problematic sense of "known"), but rather to specific belief contents: Let us say that one could, in the relevant sense, be dreaming that p (alternatively, that it is possible that one is "merely dreaming" that p) whenever the proposition that one is dreaming that p is logically consistent with the falsity of p (i.e. with not-p). In this sense, Descartes indeed could be (merely) dreaming that he is seated by the fire it1 his dressing-gown, but could not be (merely) dreaming that 2 + 3 =5 or that he thinks and/or exists-although it is, of course, logically possible that he be (in fact) dreaming those latter things as well. It is the possibility that he is, in this sense, merely dreaming that Descartes now admits as 33

Cartesian Scepticism

a reason for doubting what is presented in his immediately current spontaneous sensory beliefs. This step decisively severs the connection between Descartes' hypothesized epistemic features ofpropositions, doubtfulness and certainty, and the personal epistemic attitudes toward propositions, doubt and certitude. The doubtfulness of a claim in this sense is no longer a reason actually to doubt it, nor is the absence of certainty a ground for lack of certitude. Descartes' "method of doubt" thereby becomes exaggerated or "hyperbolic" in a second way. The first methodological hyperbole, we recall, consisted in the epistemic pretence of "rejecting" (i.e. treating as if they were known to be false) all "doubtful propositions"; that is, to begin with, those sensory beliefs which there was some positive reason to suppose false by virtue of their occurring in epistemically problematic circumstances. The present step expands the class of "doubtful propositions" to include, in essence, all those sensory beliefs not known to occur in epistemically unproblematic circumstances. That is the second hyperbole. Since any spontaneous sensory belief could be occurring in epistemically problematic circumstances, i.e. as an episode in a dream, all such beliefs are in this sense doubtful. Having in this way first suitably broadened his understanding of "dubitability", Descartes is then indeed in a position to invoke his methodological pretence and assume "that these particulars .. . are not true". At this point, then, Descartes proposes to part company with "the foundations ... acquired ... from the senses". The subsequent stretch of Descartes' first-Meditation text is devoted to spreading the contagion of dubitability, henceforth construed broadly in terms of such (mere) possibilities, from "these particulars" to "corporeal nature in general". The proceedings are informed by the apparent conviction that, in some sense, "the visions which come in sleep are like paintings, which must have been fashioned in the likeness of things that are real" (AT 19 = CS 13). To put this framing hypothesis in another way, Descartes is evidently still cOl1vinced that although specific sensory beliefs are perse dubitable, their contents must bear some resemblance or analogy to what is real. However distorted the impression, sensory contents are related to "things that are real" as ectype to archetype. Thus, although what is specifically received from the senses is always doubtful, such spontaneous sensory beliefs might nevertheless serve as a source of general convictions not similarly subject to doubt; general truths thereby received not from but through the senses. The aim of the argument at this point appears to be to fix and limit the scope of the relevant analogy: 34

Cartesian Scepticism Perhaps ... at least these general kinds of things-eyes, head, hands and the body as a whole-are things which are not imaginary but are real and exist. For even when painters try to create sirens and satyrs ... they simply jumble up the lirnbs of various animals. Or if perhaps they manage to think up something so new that nothing remotely sinlilar has ever been seen before-something which is therefore completely fictitious and unreal--at least the colours used in the composition must be real. By similar reasoning, although these general kinds of things-eyes, head, hands and so on-could be imaginary, it must at least be admitted that certain other even simpler and more universal things are real. These are as it were the real colours from which we fornl all the images of things, whether true or false, that occur in our thought. This class appears to include corporeal nature in general, and its extension; the shape of extended things; their quantity, or size and number of these things; the place in which they may exist, the time through which they may endure, and so on. (AT 19-20 =CS 14)

Despite Descartes' repeated references to the pictorial indispensability of colours, the so-called "secondary qualities" (the only genuinely sensible contents of both perceptions and dreams) are strikingly absent from this summation. By "and so on" here, it quickly becomes clear, we are to understand those aspects of "corporeal nature in general" which, like shape, size, number, place, and time, admit of mathematization. As the sixth Meditation confirms, it is precisely here that, epistemically speaking, Descartes intends to come to rest. There he finally concludes that "corporeal things exist", and he continues: They may not all exist in a way that exactly corresponds with my sensory grasp of them, for in many cases the grasp of the senses is very obscure and confused. But at least they possess all the properties which I clearly and distinctly understand, that is, all those which, viewed in general terms, are comprised within the subject-matter of pure mathematics. (AT 80 = CS 55)

And "the fact that I perceive by my senses a great variety of colours, sounds, smells and tastes, as well as differences in heat, hardness and the like" shows only that the bodies which are the source of these various sensory perceptions possess differences corresponding to them, though perhaps not resembling them. (AT 81 =CS 56)

The absence, at either point, of any argunlent to distinguish the epistemic status of colours etc. from that of shapes etc. in point of dubitability merely reflects the fact that, as Berkeley was later to make painfully clear, there is no such distinction to be had: 35

Cartesian Scepticism Phil.

Hyl.

Phil.

Hyl.

Can you ... separate the ideas of extension and motion from the ideas of all those qualities which they who make the distinction term "secondary?" What! Is it not an easy matter to consider extension and motion by themselves, abstracted from all other sensible qualities? Pray how do the mathematicians treat of them? I acknowledge, Hylas, it is not difficult to form general propositions and reasonings about those qualities without mentioning any other, and, in this sense, to consider or treat of them abstractedly. But does it follow that, because I can pronounce the word "motion" by itself, I can form the idea of it in n1Y mind exclusive of body? ... Mathematicians treat of quantity without regarding what other sensible qualities it is attended with ... But when, laying aside the words, they contemplate the bare ideas, I believe you will find they are not the pure abstracted ideas of extension ... Since therefore it is impossible even for the mind to disunite the ideas of extension and motion from all other sensible qualities, does it not follow that where the one exist there necessarily the other exist likewise? It should seem so.25

Nor do the second Meditation's reflections on the piece of wax supply the missing argument. That the persisting wax cannot consist of any configuration of determinate sensible qualities simply does not entail eitl1er the ontological or episten1ic separability of "secondary" sensible contents from mathematicizable "primary" sensibleform. Compare Hume on the "distinction of reason": It is certain that the mind would never have dreamed of distinguishing a figure from the body figured, as being in reality neither distinguishable, nor different, nor separable, did it not observe, that even in this simplicity there might be contained many different resemblances and relations. Thus, when a globe of white marble is presented, we receive only the impression of a white colour disposed in a certain form, nor are we able to separate and distinguish the colour from the form. But observing afterwards a globe of black marble and a cube of white, and comparing then1 with our former object, we find two separate resemblances, in what formerly seemed, and really is, perfectly inseparable. After a little more practice of this kind, we begin to distinguish the figure from the colour by a distinction of reason; that is, we consider the figure and colour together, since they are, in effect, the same and undistinguishable; but still view them in different aspects, according to the resemblances of which they are susceptible. (THNr. i. 7)26 25 George Berkeley, Three Dialogues Between Hylas and Philonous, ed. C. M. Turbayne (Indianapolis, Ind., and New York: Babbs-Merrill, 1954), pp. 34-5. 26 Treatise on Human Nature, ed. L. A. Selby-Bigge (Oxford: Oxford University Press, 1888). Citation by book, chapter, and section.

36

Cartesian Scepticism

What Descartes' reasoning perhaps does establish is that it is possible to doubt the existence of what he calls "composite things"; that is, corporeal things answering to a given determinate description characteristic for menlbership in a specific natural kind: So a reasonable conclusion from this might be that physics, astronomy, medicine, and all other disciplines which depend upon the study of composite things, are doubtful (AT 20 = CS 14)

In the light of Descartes' broader scientific aims, it may seem odd to find physics 011 this list, but we should remember that the discipline includes an inventory of basic natural kinds (atoms, subatomic particles, quantum wavicles, quarks, or whatever) alongside the purely mathematical kinematics and dynamics of stationary and moving "bodies". Given his prima-facie endorsement of at least a determinable al1alogy between the contents of his sensory beliefs, whether perceptions or dreams, and "things that are real", we would expect Descartes to conclude at this point that the existence of "corporeal nature in general", at least, is epistemically secure, falling outside the scope of even his extended and hyperbolic sense of what is doubtful. What we in fact find, however, points to another lacu.na in the argumentation, for Descartes notoriously concludes instead that arithmetic, geometry and other such subjects of this kind, which deal only with the simplest and most general things, regardless of whether they really exist in nature or not, contain something certain and indubitable. For whether I am awake or asleep, two and three added together are five, and a square has no more than four sides. It seems impossible that such transparent truths should incur any suspicion of being false. (AT 20 = CS 14)

As long as the presupposition of sensory ectype and natural archetype remains in place, however, a generic existential doubtfulness regarding "corporeal nature" remains unfounded. The hyperbolic extension of "reasons for doubt" to nlere possibilities will license the corresponding global negativeexistential pretence only if Descartes can consiste11tly specify the requisite possibility; that is, consistently describe an epistemic circumstance compatible with both his defacto sensory beliefs and the non-existence of any natural analogues to their contents, however remote. The final lllovenlent of the First Meditation is addressed to this task: [Firmly] rooted in my mind is the long-standing opinion that there is an oml~ipotent God who has made me the kind of creature that I am. How do I know that he has not

37

Ca rtesia n Scepticism brought it about that there is no earth, no sky, no extended thing, no shape, no size, no place, while at the same time ensuring that all these things appear to me to exist just as they do now? ... [Some] would prefer to deny the existence of so powerful a God rather than believe that everything else is uncertain. Let us not argue with them . . . According to their supposition, then, I have arrived at illy present state by fate or chance or a continuous chain of events, or by some other means; yet since deception and error seem to be imperfections, the less powerful they make my original cause, the more likely it is that I am so imperfect as to be deceived all the time. (AT 21 =CS 14)

The seas of metaphysics are running especially high in these paragraphs. Whereas Descartes' earlier reasoning focused on specific epistemically problematic circumstances ex hypothesi known sometimes to obtain, his point of departure here is a series of purely speculative hypotheses regarding the possible "original cause" of his "present state". The diverse possibilities, he suggests, fall into two n1utually exclusive and jointly exhaustive groups: Either the "original cause" is all-powerful, "an omnipotent God", or less-than-allpowerful, "fate", "chance", "a continuous chain of events" , or whatever. The argument then proceeds as a constructive dilemma. In either case, Descartes argues, his epistemic circumstances will be compatible with his being "deceived all the time", and so, in particular, with his being mistaken regarding the existence of corporeal nature per see The competences of an all-powerful "original cause" prima facie include bringing it about that Descartes is "deceived all the tin1e"; the com_petences of a less-than-all-powerful "original cause" perhaps do not suffice to preclude such all-encompassing error. The originative hypotheses whose epistemic import Descartes undertakes to assess concern his own existence as, as he puts it, "the kind of creature that I am". They are hypotheses, he writes, regarding how "I have arrived at my present state". But what is his "present state"? What kind of creature is he? The balance of the Meditations, as we know, contains an answer, developed in two stages: In the second Meditation Descartes concludes that he is at least a thinking thing; that is, "a mind, or intelligence, or intellect, or reason" (AT 27 = CS 18). And in the sixth Meditation he argues27 that he is only a thinking thing; that (as he puts it) "nothing else be~ongs to my nature or essence except that I am a thinking thing" (AT 78 = CS 54). But how much of this picture is Descartes entitled to presuppose at the present stage of the dialectic? What is it, so to speak, that a hypothetical"original cause" of his "present state" is supposed to have effected?

27 Invalidly, as it happens-but that is a story I have told elsewhere. (See ch. 3 of The Thinking Self (Philadelphia, Pa.: Temple University Press, 1986).

38

Cartesian Scepticism

Well, this much is clear: Whatever else Descartes may be, he is at least a believer. He is the (believing) subject both of diverse particular spontaneous sensory beliefs, all of which have proved to be (in the broad sense) doubtful, and of the general belief whose epistemic status is here in question, a belief in the existence of "bodies" or "corporeal nature in general". That Descartes is at least a thinking thing, in other words, appears to be a non-controversial presupposition of all the proceedings in his first Meditation. But if this is so, then Descartes will have succeeded in describing a "reason for doubting" the existence of "bodies", i.e. in specifying epistemic circumstances compatible both with his de facto sensory beliefs and the non-existence of any natural analogues to their contents, only if his having such sensory beliefs is itselfconsistent with the non-existence of corporeal nature. 28 And that, in turn, will be true only if Descartes' own existence as a thinking thing is independent of the existence of corporeal things. The doubtfulness of Descartes' belief in the existence of "corporeal nature in general" will be a consequence of the (ostensible) possibility that he is "deceived all the time" 011ly if the existence of corporeal nature is something abollt which he could be deceived; that is, only if it is possible for him to believe mistakenly that there are "bodies". But this will be possible only if the existence of bodies is not a necessary condition of the existence of believers. The central claim of the ontological dualism ultimately defended only in the sixth Meditation, in short, surprisingly appears to be a presupposition of the concluding stage of Descartes'hyperbolic doubt in the first. Various objections to this conclusion are likely to be lodged on Descartes' behalf. It will surely be suggested, for instance, that the cogency of Descartes' reasoning does not presuppose the actual ontological independence of mind and body. The mere possibility that a thinking thing could exist without any corporeal things existing suffices to make room for the req"uisite global existential doubt. But how should we understand this? The concept of ontologically independent substances is already a concept in the mode of possibility: mind and body are actually independent only if it is possible for either to exist without the other existing. 29 Now, by Descartes' lights, truths regarding such possibilities can be known by reason alone, and 28 I here deliberately ignore the peculiar, and arguably incoherent, Cartesian conviction that the competences of an all-powerful God transcend the strictures of logical consistency. 29 Hence, Descartes' Meditations establish no eschatological conclusions. That minds can exist apart from bodies does not imply that they ever do. It is consistent with everything that Descartes claims to have demonstrated that a thinking substance which is not present in its body"as a sailor is present in a ship", but rather "very closely joined and, as it were, intermingled with it" (AT 81 = CS 56), cannot survive the dissolution of so intimate a relationship.

39

Cartesian Scepticism

it is surely plausible to hold that what can be known in this way, by the "true light of reason" , is itself necessary. It follows that if the existence of bodies is a necessary condition of the existence of believers then it is necessarily so. Conversely, then, if it is not necessary that believers exist only if bodies doi.e. if it is possible that the existence of bodies is not a necessary condition of the existence of believers-then the existence of bodies isn't a necessary condition of the existence of believers. In the context of Descartes' modal understandings, in short, there is no difference between an argument's presupposing the actual ontological indepel1dence of mind and body and its "merely" presupposing its possibility. 30 Again, it will be objected that even if the existence of bodies is a necessary condition of the existence of believers, it suffices for Descartes' purposes that he does not know with certainty that it is. Surely it is possible for him to doubt that believers exist only if corporeal things do. But this response mislocates the role of presupposition in question. What is at issue is whether the existence of corporeal nature is something about which Descartes could infact be "deceived"; that is, whether it is possible for him to mistakenly believe that there are bodies. Let us suppose that Descartes himself can doubt that the existence of bodies is a necessary condition of the existence of believers. That is not something I have denied. Indeed, I have not proposed that Descartes himself must have any explicit convictions about the ontological independence of mind and body in the context of the first Meditation. What I have contended is that, in accordance with his own methodological precepts, the doubtfulness of Descartes' belief in the existence of "corporeal nature in general" depends on the possibility of his mistakenly believing that there are bodies, and that this will in fact be possible only if the existence of bodies is not in fact a necessary condition of the existence of believers. Quite independently of what Descartes is entitled to believe regarding the ontological independence of mind and body, in other words, the cogency of his argument at this point in the first Meditation-in particu.1ar, Descartes' ability to satisfy his own methodological imperatives by identifying a "reason for doubting" the proposition that there exists a world of corporeal thil1gs-presupposes such independence. Descartes' (modal) reason for doubting that he is seated by the fire in his dressing-gown is that he could be (merely) dreaming that he is, and that will be so 30 The foregoing argument in a nutshell: Let 'T-)B' represent "Thinkers exist only if Bodies exist". I propose that Descartes is committed to: If T-)B then o(T-)B). Contraposing, we obtain: If "'o(T-)B) then "'(T-)B). By modal equivalences, then, it follows that: If o"'(T-)B) then '"'J(T-)B). But ''''(T-)B)' already expresses the ontological independence of mind and bodies, since it (at least) implies'o(T&'"B)' .

40

Cartesian Scepticism

only if his dreaming that he is seated by the fire in his dressing-gown is consistent with his not being seated by the fire in his dressing-gown. Analogously, Descartes' ostensible (modal) reason for doubting that there exists a world of corporeal things is that he could be deceived in believing that bodies exist, and that will be so only if his believing that bodies exist is consistent with there not being any such corporeal things. But just as his dreaming that he is seated by the fire will be consistent with his not being seated by the fire only if his being seated by the fire is not a necessary condition of his dreaming (that, or anything else)-which is unproblematic-so Descartes' believing that bodies exist will be consistent with the non-existence of corporeal things only if the existence of corporeal things is not a necessary condition of his believing (that, or anything else)-which is ab initio contentious and, indeed, by his own lights, stands in need of a separate demonstration not forthcoming until the sixth Meditation. By the same token, Descartes' famous final heuristic pretence, the demoniac deceiver, also presupposes the ontological il1dependence of mind and body: [I] will suppose . . . some malicious demon of utmost power and cunning has employed all his energies in order to deceive me. I shall think that the sky, the air, the earth, colours, shapes, sounds and all external things are . . . delusions ... which he has devised to ensnare my judgment. I shall consider myself as not having hands or eyes, or flesh, or blood or senses, but as falsely believing that I have all these things. (AT 22-3 = CS 15)

Here, too, the expressed SUpposItIon will describe a possibility only if Descartes' believing anything whatsoever, correctly or mistakenly, is consistent with the non-existence of "all external things" . It is surely an understatement to observe that that is hardly obvious. I began this chapter by recalling that Descartes is widely supposed to have given, in his first Meditation, intuitive and compelling arguments in support of the sceptical view that it is possible to doubt that (we know that) there exists a natural world of objects in space and time. The investigations just concluded, however, hardly support this received vie~ In reconstructing such argumentation as the first Meditation does contain, I have, in fact, charitably tried to avoid some of the most implausible Cartesian commitments. I have, for instance, eschewed the Scholastic residue of "modes of being" and so refrained from reading Descartes as problematically ontologizing appearances, and I have taken a relaxed stance toward such notions as "sure signs" and so refrained from reading Descartes as problematically identifying epistemic warraI1t with 41

Cartesian Scepticism

logical entailment. Even granted such exegetical concessions, however, Descartes' reasoning turned out to rest upon a series of controversial and disputable commitments-the presumptive architectural foundationalism and epistemic individualism of the formalist paradigm, a broad notion of "reasons for doubt" which subsumes mere possibilities, and, perhaps most contentiously of all, the ontological independence of mind and body. It would, I submit, be grotesque to call an argument resting on such contentio·us presuppositions"compelling and intuitive" . It should be noted, however, that this negative conclusion concerns only the claimed possibility of global existential doubt regarding a natural world of objects in space and time. The weaker thesis that the actual nature of those items whose perse existence and constitution account for the specific character of our sensory experience might be radically different from what, in that experience, it appears to be, i.e. the thesis that the natural world n1ight (turn out to) be very different from what we initially and pre-reflectively take it to be, rests on no such contentious presuppositions. But that is not as such a scepticalthesis. It formulates, rather, one species of realism, and, unlike global existential scepticism, that the natural world is, in that sense, real is a Cartesian thesis which, for instance, Kantian transcendental idealists, Peircean pragmatists, and contemporary "scientific realists" can all find perfectly congenial. At this point it may well be objected that, whatever the ultimate verdict regarding the particular arguments advanced in Descartes' works, Cartesian sceptical reasoning perse has an intuitive force that still needs to be addressed. Robert Fogelin, for instance, offers the following account of the typical consequences of making salient a variety of remote defeating possibilities: 31 Normally we ignore these possibilities, but if we dwell on them ... we will find ourselves unwilling to claim to know many things that we usually accept as items of knowledge. Do I, for example, know my own name? This seems to me to be as sure a piece of knowledge as I possess. But perhaps, through a mix-up at the hospital, I an1 a changeling. I'm really Herbert Ortcutt, and the person who is called "Ortcutt" is actually RJF. These things, after all, do happen. Given this possibility, do I know my own name? I'm inclined to say that I do not. Not only that, philosophical natfs, namely those who do not see that such an admission may lead to forlorn skepticism, tend to agree. When pressed in this way, people are likely to become impatient, even angry, yet, under pressure, most will acknowledge that strictly speaking-if you are going to be picky-they do not know their own names. (PR 93-4)

31 In Pyrrhonian Reflections on Knowledge and Justification (PR) (Oxford and New York: Oxford University Press, 1994).

42

Cartesian Scepticism

Now, although there is a great deal that can be said about this passage,32 it cannot be disputed that consideration of nlere possibilities can and does evoke this sort of response. The philosophical issue, however, is whether it properly should do so, and we have already encountered some reasons for calling that conclusion into question. One can instead consistently and resolutely hold to the principle-independently advocated by C. S. Peirce and G. E. Moore-that knowledge can be undermined only by possibilities which there is some positive reason to suppose actually obtain, in which case the merepos. sibility that I am a "changeling" will carry no epistemological consequences at all. Pressing the sceptical case will then lead only to dialogues like this: Do I know my own name? Sure, it's'Jay Rosenberg'. But have I ever investigated the possibility that I am a changeling? Well, no, but what sort of possibility is "the possibility that I am a changeling" supposed to be? Of course, it's logically possible that I am a changeling. That is, "Jay Rosenberg is a changeling" is not self-contradictory. But so what? I have no reason to suspect that I am a changeling, and, in fact, I'm not one. But do I know that? Well, for one thing, I've got a certified copy of my birth certificate. Ah, but documents can be forged. Indeed they can, but, again, so what? I have no reason to suppose that this document is a forgery. And so on.

A process of making salient mere possibilities can successfuJly undermine knowledge, in short, only if it can establish that some necessary condition for knowledge has not been met, and that is something that still needs to be cogently and convincingly argued. Barry Stroud, for instance, undertakes to secure by an appropriate argument the genuine sceptical import of Descartes' appeal to dreams and dreaming. The role of that appeal, Stroud agrees, is indeed to call attention to a possibility, but the fact that we are explicitly cOllcerned with judgements of the form "S knows that p" confers relevance on the "dream-possibility". The concept of knowledge which we invoke in everyday life, Stroud argues, is such that, for allY case of claimed knowledge, there exist possibilities "such that if they obtained I did not know what I claimed to know, and they had to be known not to obtain in order for the original knowledgeclaim to be true", and Descartes' dream-possibility, he suggests, fulfils both these conditions:

32 For instance, that Fogelin is confused about proper names. On the" changeling" scenario that he envisions, it might be true that his name once was 'Herbert Ortcutt', and that Herbert Ortcutt's name once was 'Robert J. Fogelin', but their subsequent histories have arguably brought it about that, inter alia, the proper names that were assigned to them at birth are not their present proper nan1es. For the relevant arguments see ch. 4 of my Beyond Formalism.

43

Cartesian Scepticism [It] seems undeniable that it fulfils the first. If he were dreaming Descartes would not know what he claims to know ... But does it fulfil the second condition? Is it a possibility which must be known not to obtain if Descartes is to know that he is sitting by the fire with a piece of paper in his hand? I think it is difficult simply to deny that it is. . . . The idea is that the two conditions ... are not really separate after alL As soon as we see that a certain possibility is incompatible with our knowing such-and-such, it is suggested, we immediately recognize that it is a possibility that must be known not to obtain if we are to know the such-and-such in question. (SPS 26-7)

Thus, even though it is not plausible to require that we know the falsehood of everything incompatible with our knowing what we claim to know, Stroud concludes, it seems plausible to hold the weaker requirement that "if somebody knows something, p, he must know the falsity of all those things incompatible with his knowing that p (or perhaps all those things he knows to be incompatible with his knowing that p)" (SPS 28-9). As they stand, this conclusion and the reasoning leading up to it need to be tidied, stabilized, and refined in a number of ways. One potential confusion-one that Stroud himself is later at pains to highlight-is between the conditions in which it is in order to claim to know something (to assert that one knows) and the conditions in which it is true that one knows that thing. Stroud concedes that "if we insist on ... a realistic account of how we actually behave there seems little doubt that we do not in fact impose that general condition on knowledge-claims" (SPS 49; my emphasis). 33 But it does not follow from this concession that being able to eliminate the dream-possibility is not among the conditions of knovv'ledge per se. And in fact Stroud evidently thinks that it is easy to explain why a principle that correctly formulates one of the conditions for the truth of a knowledge-claim is neglected in mundane contexts of assertion: [We] are ... strongly inclined to reject the sceptical reasoning because what it would require of us deviates so radically from what we require ... in everyday life. The sceptical philosopher has an explanation of that difference. There is a single conception of knowledge at work both in everyday life and in the philosophical investigation of human knowledge, but that conception operates in everyday life under the constraints of social practice and the exigencies of action, co-operation and communication. The

33 Fogelin concurs: "In our common use of knowledge claims-including serious uses of knowledge claims-our assertions are not responsive to every possibility that might refute them. We are willing to make knowledge claims while recognizing that sometimes odd things come up that lead us to withdraw them" (PR 79). Or again: "When people claim to know things, they do not do so in the belief that they have eliminated all eliminable refuting possibilities, nor do their auditors suppose that they believe this" (PR 95).

44

Cartesian Scepticism practical social purposes served by our assertions and claims to know things in everyday life explain why we are normally satisfied with less than what, with detachment, we can be brought to acknowledge are the full conditions of knowledge. (SPS 71)

This is a suggestive passage and one which gives rise to a variety of interesting issues. Just what "social practice", for instance, constrains our mundane knowledge-claims, and just how do such constraints operate? What does Stroud mean by 'detachment', and what are "the full conditions of la.10wledge"? These are all excellent questions, but, for the time being, at least, I shall simply note them and pass on. Second, we must distinguish possibilities incompatible with what is ostensibly known from possibilities incompatible with the knowing of it. In a recent essay entitled "The Structure of the Skeptical Argument,,34 Anthony Brueckner undertakes to determine what "epistemic principles" are required by "what has become the canonical Cartesian skeptical argument" ("SSA" 827). The linchpin of the family of sceptical reasonings considered by Brueckner is an epistemic closure principle, the principle that knowledge is closed under known entailn1ents; that is (letting the arrow represent entailment): (CP)

For all S, , \If, if S knows that