New Issues in Epistemological Disjunctivism [1 [edition].] 9781351603560, 1351603566

Table of contents :
1. IntroductionCasey Doyle, Joseph Milburn, and Duncan PritchardPart I: Situating Disjunctivism2. Perceptual Experience and Empirical RationalityJohn McDowell3. Epistemological Disjunctivism and Anti-Luck Virtue EpistemologyDuncan PritchardPart II: Historical Antecedents4. Ancient Philosophy and Disjunctivism: The Case of the StoicsIakovos Vasiliou 5. The Kantian Roots of Epistemological DisjunctivismThomas Lockhart6. Was Wittgenstein a Disjunctivist avant la lettre? Genia Schoenbaumsfield7. Settling a Question: Austin and DisjunctivismGuy LongworthPart III: Epistemological Disjunctivism: Prospects and Problems8. Disjunctivism and Realism - not Naive but ConceptualSonia Sedivy9. Epistemological Disjunctivism and its Representational CommitmentsCraig French10. Either Epistemological or Metaphysical DisjunctivismVeli Mitova11. Neither/NorClayton Littlejohn12. Disjunctivism and CredenceRam Neta13. Disjunctivism, Skepticism, and the First PersonAdrian Haddock Part IV: Disjunctivism in Other Domains14. Two Forms of Memory Knowledge and Epistemological DisjunctivismJoseph Milburn and Andrew Moon15. Testimonial DisjunctivismStephen Wright16. Epistemological Disjunctivism: Perception, Expression, and Self-KnowledgeDorit Bar-On and Dustin Johnson17. Ringers for BeliefCasey Doyle18. Disjunctivism and Other MindsAnita Avramides

Citation preview

New Issues in Epistemological Disjunctivism

This is the first volume dedicated solely to the topic of epistemological disjunctivism. The original essays in this volume, written by leading and up-and-coming scholars on the topic, are divided into four thematic sections. The first section presents work from two of the leading defenders of epistemological disjunctivism. The second set of chapters addresses the historical background of epistemological disjunctivism. It features essays on ancient epistemology, Immanuel Kant, J.L. Austin, and Ludwig Wittgenstein. The third section tackles a number of contemporary issues related to epistemological disjunctivism, including its relationship with perceptual disjunctivism, radical skepticism, and reasons for belief. Finally, the fourth group of essays extends the framework of epistemological disjunctivism to other forms of knowledge, such as testimonial knowledge, knowledge of other minds, and self-knowledge. New Issues in Epistemological Disjunctivism is a timely collection that engages with an increasingly important topic in philosophy. It will appeal to researchers and graduate students working in epistemology, philosophy of mind, and philosophy of perception. Casey Doyle is Junior Researcher at the University of Hradec Kralove in the Czech Republic. He was formerly Junior Research Fellow in Philosophy at St Hilda’s College, University of Oxford. His major research interests are in philosophy of mind and epistemology, especially self-knowledge. He also works on certain topics in ethics and history of philosophy. Joe Milburn is a Visiting Professor at the University of Navarra in Pamplona. His research centers on epistemology and philosophy of religion. He has recent publications dealing with epistemic luck and with the philosophy of John Henry Newman. Duncan Pritchard is Distinguished Professor of Philosophy at the University of California, Irvine, and Professor of Philosophy at the University of Edinburgh. His monographs include Epistemic Luck (2005), The Nature and Value of Knowledge (co-authored, 2010), Epistemological Disjunctivism (2012), and Epistemic Angst: Radical Skepticism and the Groundlessness of Our Believing (2015).

Routledge Studies in Epistemology Edited by Kevin McCain University of Alabama at Birmingham, USA Scott Stapleford St. Thomas University, Canada

Pragmatic Encroachment in Epistemology Edited by Brian Kim and Matthew McGrath New Issues in Epistemological Disjunctivism Edited by Casey Doyle, Joe Milburn, and Duncan Pritchard For more information about this series, please visit: www.routledge.com/ Routledge-Studies-in-Epistemology/book-series/RSIE

New Issues in Epistemological Disjunctivism

Edited by Casey Doyle, Joe Milburn, and Duncan Pritchard

First published 2019 by Routledge 52 Vanderbilt Avenue, New York, NY 10017 and by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2019 Taylor & Francis The right of the editors to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data A catalog record for this book has been requested ISBN: 978-1-138-09409-3 (hbk) ISBN: 978-1-315-10624-3 (ebk) Typeset in Sabon by Apex CoVantage, LLC

Contents

1 Introduction

1

CASEY DOYLE, JOE MILBURN, AND DUNCAN PRITCHARD

PART I

Situating Disjunctivism 2 Perceptual Experience and Empirical Rationality

29 31

JOHN McDOWELL

3 Epistemological Disjunctivism and Anti-Luck Virtue Epistemology

41

DUNCAN PRITCHARD

PART II

Historical Antecedents 4 Ancient Philosophy and Disjunctivism: The Case of the Stoics

59

61

IAKOVOS VASILIOU

5 The Kantian Roots of Epistemological Disjunctivism

89

THOMAS LOCKHART

6 Was Wittgenstein a Disjunctivist Avant la Lettre?

113

GENIA SCHÖNBAUMSFELD

7 Settling a Question: Austin and Disjunctivism GUY LONGWORTH

131

vi

Contents

PART III

Epistemological Disjunctivism: Prospects and Problems

151

8 Disjunctivism and Realism: Not Naïve but Conceptual

153

SONIA SEDIVY

9 Epistemological Disjunctivism and Its Representational Commitments

169

CRAIG FRENCH

10 Either Epistemological or Metaphysical Disjunctivism

194

VELI MITOVA

11 Neither/Nor

215

CLAYTON LITTLEJOHN

12 Disjunctivism and Credence

241

RAM NETA

13 Disjunctivism, Skepticism, and the First Person

259

ADRIAN HADDOCK

PART IV

Disjunctivism in Other Domains

279

14 Two Forms of Memory Knowledge and Epistemological Disjunctivism

281

JOE MILBURN AND ANDREW MOON

15 Testimonial Disjunctivism

298

STEPHEN WRIGHT

16 Epistemological Disjunctivism: Perception, Expression, and Self-Knowledge

317

DORIT BAR-ON AND DREW JOHNSON

17 Ringers for Belief CASEY DOYLE

345

Contents 18 Disjunctivism and Other Minds

vii 366

ANITA AVRAMIDES

List of Contributors Author Index Subject Index

388 392 395

1

Introduction Casey Doyle, Joe Milburn, and Duncan Pritchard

Tim Crane pointed out that disjunctivism is “a philosophical doctrine with a name like an eye disease” (2009). Despite its unbeautiful name, debates about disjunctivism are at the forefront of contemporary analytic philosophy. Varieties of disjunctivism have been advanced and contested in the theory of knowledge, philosophy of perception, philosophy of action, and philosophy of reasons and rationality. Although it was perhaps once common to refer to disjunctivism as a single position, whose main exponents included J. M. Hinton, Paul Snowdon, John McDowell, M.G.F. Martin, and John Campbell, it is now more common to distinguish at least two species of disjunctivism: metaphysical and epistemological disjunctivism. Metaphysical disjunctivism is a thesis about the nature of perceptual experience. Many different formulations of metaphysical disjunctivism can be found in the literature (see Soteriou 2009 for a helpful overview). The weakest formulation holds that veridical perceptual experiences and hallucinations differ mentally in some important respect. Naïve Realists are committed to metaphysical disjunctivism (Martin 2004; Snowdon 2005) and much discussion has focused on evaluating the prospects of naïve realism by considering the truth of metaphysical disjunctivism. By contrast, epistemological disjunctivism is a thesis about the nature of perceptual knowledge, and in particular the kind of support that perceptual experience provides. In its weakest formulation, epistemological disjunctivism denies that veridical perceptual experiences and hallucinations provide the same support for perceptual beliefs. It holds that, in the right circumstances, veridical perceptual experience can provide better support for beliefs than hallucinations. The two most prominent defenders of epistemological disjunctivism—John McDowell and Duncan Pritchard—hold that perceptual experience provides factive, indefeasible, or falsity-inconsistent reasons for belief, as it is variously put, and these reasons are reflectively accessible to the subject. In the terminology of the Oxford Realists, veridical perceptual experience provides “proof” rather than evidence. This is the first collection of essays dedicated to epistemological disjunctivism. A central aim of this volume is to broaden the discussion by

2

Casey Doyle et al.

considering epistemological disjunctivism in the context of other issues in epistemology and philosophy of mind. It considers historical precedents of the view, presents the state of the art in discussions of epistemological disjunctivism as an account of perceptual knowledge, and explores disjunctivist treatments of other domains of knowledge, such as memory, testimony, introspection, and knowledge of other minds. In this introduction we provide some background on the view and the relation between epistemological disjunctivism and metaphysical disjunctivism before offering brief summaries of each contribution.

1. Epistemological Disjunctivism A range of views might rightfully receive the label epistemological disjunctivism. As mentioned earlier, any view claiming that veridical experience gives us better rational support than subjectively indistinguishable illusory or hallucinatory experience can be considered a weak form of epistemological disjunctivism. In what follows, however, we focus on epistemological disjunctivism as it figures in the work of John McDowell and Duncan Pritchard, as these are its two most prominent defenders in the literature. According to Pritchard, Epistemological disjunctivism is the view that in paradigm cases of perceptual knowledge the knowledge in question enjoys a rational support that is both factive and reflectively accessible. In particular, it is the view that when one has perceptual knowledge in such cases, the reflectively accessible rational support one has for one’s knowledge that p is that one sees that p. (2016: 124, emphasis in original) To say that one has factive rational support in paradigm cases of perceptual knowledge is to say that the rational support one possesses is truthguaranteeing. One cannot have factive rational support for believing that p, unless p is true. To say that one’s rational support is reflectively accessible is to say that one can know by reflection alone that one possesses this rational support. The epistemological disjunctivist, then, holds that in paradigm cases of perceptual knowledge one has rational support for one’s perceptual knowledge that is both factive and reflectively accessible. Pritchard’s version of epistemological disjunctivism takes this rational support to be the fact that one sees that p. In this case, the rational support provided by our perceptual experience is of two kinds. Either it is factive and reflectively accessible, as in paradigmatic cases of knowledge, or it is not. In motivating epistemological disjunctivism, Pritchard focuses on the benefits of accepting the view, as opposed to the disadvantages of accepting an alternative. According to Pritchard, epistemological disjunctivism

Introduction

3

provides a new approach to perceptual knowledge. Normally, views of perceptual knowledge can be classified either as internalist or externalist. Internalist theories require that knowledge is grounded on reflectively accessible rational support; externalist theories, on the other hand, require that knowledge be grounded on truth-conducive processes of belief formation. By making perceptual knowledge depend on reflectively accessible rational support, epistemological disjunctivism respects the internalist insight that knowledge is an achievement for which we are in part responsible. By making knowledge depend on factive support, at the same time it respects the externalist insight that our grounds for knowledge must make our knowledge reliable. By integrating internalist and externalist components, epistemological disjunctivism represents “a holy grail” for theories of perceptual knowledge (Pritchard 2012: 1–4). Furthermore, in Pritchard’s view, epistemological disjunctivism respects common sense and gives us the means to provide a satisfactory response to skepticism (see Pritchard 2012: 17–18, 131–135, 2016: 132–136). Given all this, if epistemological disjunctivism is an available position, we should accept it. Showing that it is an available position is a matter of removing standing objections to the view. So Pritchard’s argument for epistemological disjunctivism consists in large part in arguing that there is no good reason to reject it (Pritchard 2012: 18). Pritchard presents three standing problems for epistemological disjunctivism and the solution to these problems. First, there is the basis problem. Many philosophers hold that seeing that p is just a way of knowing that p. But if seeing that p is a way of knowing that p, it does not seem we could have perceptual knowledge in virtue of the rational support that seeing that p affords us. So epistemological disjunctivism fails. Pritchard’s response to this problem is to challenge the idea that seeing that p is simply a way of knowing that p. Pointing to cases in which we possess misleading mental state defeaters, Pritchard argues that we can see that something is the case, even when we fail to know that it is the case (see Pritchard 2012: 21, 26–28; Pritchard 2016: 127–128). Second, there is the access problem. Epistemological disjunctivists hold that we can know that we see that p by reflection alone. But we can also know by reflection alone that if we see that p, then p. So it seems that if epistemological disjunctivism is true, then we should be able to know, by reflection alone, various facts about our environment, which is absurd. Pritchard’s response to this problem, roughly, is to deny that one can know by reflection alone that one sees that p, if one does not already have perceptual knowledge that p. The epistemological disjunctivist is committed only to the idea that in paradigmatic cases of perceptual knowledge, one can know by reflection alone that one sees that p (see Pritchard 2012: 20, 50–52; Pritchard 2016: 129–130).

4

Casey Doyle et al.

Finally, there is the distinguishability problem. According to epistemological disjunctivism, if one has paradigmatic perceptual knowledge that p, then one can know by reflection alone that one sees that p. But this seems to imply that in paradigmatic cases of perceptual knowledge, one is able by reflection alone to distinguish between good cases, in which one has perceptual knowledge, and corresponding bad cases, in which appearances are the same but misleading. But this seems wrong. After all, we are liable to fall into error when we are in the bad case. Pritchard’s response to the distinguishability problem is to argue that in paradigmatic cases of perceptual knowledge that p, while we know reflectively that we see that p, and so in some way are able to distinguish good cases from bad cases, this does not imply that we are able to discriminate between good and bad cases. Thus, while it is a commitment of epistemological disjunctivism that when one is in the good case, they can know by reflection that they are in the good case and not in the bad case, epistemological disjunctivism is in no way committed to the idea that we could somehow know in the bad case that we are in the bad case. Pritchard puts this by saying that in paradigmatic cases of perceptual knowledge, we have favoring grounds for believing that we see that p instead of merely seeming to see that p, as opposed to discriminating grounds. So while we can know reflectively that we are in the good case when we are in the good case, and so in some sense distinguish the good case from the bad, we cannot discriminate between good and bad cases (see Pritchard 2012: 21, 96–97; Pritchard 2016: 132). Whereas Pritchard motivates epistemological disjunctivism by appeal to its theoretical benefits, John McDowell presents his disjunctivism as an alternative to what he takes to be a seductive but ultimately incoherent view of empirical knowledge. To understand McDowell’s epistemological disjunctivism, it is helpful to focus on two closely related features of his thought. The first is McDowell’s disjunctive account of appearances or seemings. The second is his criticism of “hybrid conceptions of knowledge”. McDowell gives voice to a disjunctive account of appearances or seemings in various papers. For instance, in “Criteria, Defeasibility, and Knowledge”, he makes the claim that “an appearance that such-andsuch is the case can be either a mere appearance or the fact that suchand-such is the case making itself perceptually manifest to someone” (1998a: 386–387, emphasis in original). Elsewhere, McDowell states, “Of facts to the effect that things seem thus and so to one, we might say, some are cases of things being thus and so within the reach of one’s subjective access to the external world, whereas others are mere appearances” (1998d: 241). And again, “an appearance is either a case of things being thus and so in a way that is manifest to the subject or a case of its merely seeming to the subject that that is how things are” (2010: 244, emphasis in original).

Introduction

5

McDowell’s disjunctive account of seemings or appearances is supposed to present an alternative to a view on which appearances always “fall short of the fact” (cf. McDowell 1998a: 387). If appearances always fall short of the fact, then whenever one is in a mental state in which it seems to them that things are thus and so, it is possible that the world is otherwise. He writes: The idea of a fact [making itself perceptually manifest to someone] is in itself purely negative; it rejects the thesis that what is accessible to experience falls short of the fact in the sense I explained, namely, that of being consistent with there being no such fact. (1998a: 387) On McDowell’s disjunctive account of appearances, at least some mental states of it seeming to one that things are thus and so are such that being in them is a sufficient condition for things actually being thus and so; some appearances are factive. The disjunctive account of appearances serves as an alternative to what McDowell calls “a fully Cartesian view of subjectivity” (McDowell 1998d: 236). On this fully Cartesian view, “there are no facts about the inner realm besides what is infallibly accessible” for the subject (Ibid.: 241). If the fully Cartesian view is correct, a subject runs no risk of making a mistake about how things stand in the subjective realm; she cannot mistake herself to be in one subjective state when she is in another. McDowell holds that the Cartesian view of our subjectivity obscures not only how we could have perceptual knowledge of the objective world, but how we could even have appearances that purport to be about our objective environment (McDowell 1998d: 243). By endorsing the disjunctive conception of appearances, McDowell rejects the Cartesian idea that transparency is a mark of the subjective. No subject is infallible as to whether a fact concerning her objective environment is being made perceptually manifest to her. We are liable to perceptual error. Nevertheless, McDowell holds that when facts concerning one’s environment make themselves manifest to one’s subjectivity, this is itself a fact about one’s subjectivity. We need not, and should not, understand such an event as factoring into an independent subjective component (how things appear to the subject) and an independent objective component (how things stand in the subject’s environment). Rather, we should allow that certain facts about one’s environment are implied by certain facts about one’s subjectivity. Intuitively, there is a close connection between a disjunctive account of appearances and epistemological disjunctivism. However, there is good reason for suspecting that McDowell thinks that adopting a disjunctive account of appearances is insufficient for adopting the latter. On McDowell’s telling, ancient skepticism, as opposed to modern

6

Casey Doyle et al.

Cartesian skepticism, accepted a disjunctive account of appearances but called into question the idea that possessing either kind of appearance could allow one to have knowledge (McDowell 1998d: 241). But the ancient skeptic is clearly not an epistemological disjunctivist. So, while adopting a disjunctive account of appearances might be a necessary condition for accepting epistemological disjunctivism, it certainly is not a sufficient one. To understand how McDowell fits disjunctivism about appearances into a version of epistemological disjunctivism, it is useful to focus on his criticisms of “hybrid conceptions of knowledge”. Hybrid conceptions of knowledge hold that “knowledge is a standing in the logical space of reasons”. That is, they hold that knowing that p involves being able to justify one’s claim that p (McDowell 1998c: 395, fn). However, they deny that being able to justify one’s claim that p is sufficient for knowing that p. They hold that a standing in the space of reasons can never guarantee the truth of one’s belief. So according to the hybrid account, knowing that p involves both a standing in the space of reasons and something beyond that standing, namely, that one’s belief is true (McDowell 1998c: 400). Putting aside McDowell’s Sellarsian terminology of the space of reasons, we can say that hybrid accounts of knowledge involve distinct internalist and externalist components. According to hybrid accounts, knowledge minimally involves having doxastic justification for one’s belief, where this involves awareness or knowledge of the grounds of one’s belief. This awareness must be such that one could, if prompted, cite these grounds to justify one’s believing as one does. But, on the hybrid account, meeting this internalist condition is insufficient for knowledge, since one can meet the condition and still fail to believe truly. So there is an externalist condition for knowledge as well, namely, that one’s belief be not only justified, but true. In other words, hybrid conceptions of knowledge are “fallibilist”. While they hold that having reflectively accessible rational support is necessary for knowing, they also hold that this reflectively accessible rational support cannot guarantee the truth of the relevant belief. McDowell thinks it is necessary to reject a hybrid conception of knowledge. He states: I think . . . that if we cannot see our way to accepting the Sellarsian idea in full, we should reject it, as in full-blown externalist accounts. It is not a good idea to suppose a satisfactory standing in the space of reasons might be part but not the whole of what knowledge is. (1998c: 404) Why is this? McDowell offers a battery of arguments. First, he argues that hybrid conceptions of knowledge allow for accidentally true beliefs to count as knowledge (McDowell 1998b: 403–404). Second, McDowell

Introduction

7

argues that hybrid conceptions of knowledge cannot coherently understand knowledge to be a standing in the space of reasons (Ibid.: 402–403). Third, McDowell argues that hybrid conceptions of knowledge fail to explain how knowledge is a cognitive achievement of the knower (1998a: 373–374). Fourth, McDowell argues that hybrid accounts of knowledge are skeptical. He writes: I do not think anyone has ever given a satisfactory answer to this question: if one acknowledges that one’s warrant for believing something leaves open a possibility that things are not as one believes them to be, how does that differ from acknowledging that for all one knows things are not as one believes them to be, so that one does not know them to be that way? I doubt that anyone would go on supposing that a belief based on less than conclusive warrant can be knowledgeable, refusing to be embarrassed by that question, were it not for being unable to envisage any alternative—apart, that is, from explicitly conceding that we ought to be sceptical about the possibility of experience-based knowledge. (2013: 148) Finally, McDowell appeals to what might be called “transcendental concerns” in arguing against the hybrid conception of knowledge. He argues that hybrid conceptions of knowledge, like the fully Cartesian view of mind, put into question not only the possibility of having knowledge of the world, but also the possibility of thought being intentionally directed at the world (McDowell 1998b: 408–409). While McDowell holds that we must reject a hybrid conception of knowledge, he acknowledges that it can seem that it is the only alternative to skepticism or a version of externalism on which knowledge does not involve a standing in the space of reasons. (This comes out in the previous block quotation.) According to McDowell, if we combine the fact that we are incorrigibly fallible believers with a certain conception of the space of reasons—that is, with a certain understanding of how doxastic justification must work—then it will seem that our standing in the space of reasons will never be able to guarantee the truth of our beliefs about the “external world”. What is this conception of the space of reasons? McDowell calls it “the interiorized conception of the space of reasons” (1998c: 404). It holds: [Reason has] a proper province in which it can be immune to the effects of luck; not in the sense of sheer chance, but in the sense of factors that reason cannot control, or control for. The idea is that reason can ensure that we have only acceptable standings in the space of reasons, without being indebted to the world for favours received; if we exercise reason properly, we cannot arrive at defective standings

8

Casey Doyle et al. in the space of reasons, in a way that could only be explained in terms of the world’s unkindness. (Ibid.: 404–405)

This conception of the space of reasons is the epistemological analogue to the fully Cartesian view of the mind. According to it, the only facts that bear on our justification are those that we not liable to fall into error about, given that we are sufficiently conscientious in making our judgments. The facts relevant to our standing in the space of reasons must be open to our view, though we might need to look closely to see them. Otherwise, we could come to have a defective standing in the space of reasons that could only be explained in terms of the world’s unkindness. Accepting the interiorized conception of the space of reasons leads one to accept what McDowell calls the Highest Common Factor conception of experience (cf. McDowell 1998a: 386–394). On this conception of experience, the justifying grounds given to us in cases of veridical experience are the same as those given to us in cases of illusory or hallucinatory experience. Note that the interiorized view of the space of reasons does not allow for the fact that one has a veridical experience to bear on one’s justificatory status. We are liable to fall into error about whether our experiences are veridical or not, no matter how conscientious we are in forming our judgments. (It is exactly for this reason that we are liable to fall into error when making perceptual judgments, no matter how conscientious we are.) Thus, on the interiorized conception of the space of reasons, having a veridical experience cannot bear on our standing in the space of reasons. Nor, for similar reasons, can the fact that we are having an illusory or hallucinatory experience. What facts concerning our experience, then, can bear on our standing in the space of reasons? The most plausible candidate will be something like the fact that it appears to us that such and such is the case; plausibly, if we were sufficiently conscientious we would not fall into error in our judgments about how things appear to us. But facts about how things appear to us hold the same in cases of veridical experience and in the corresponding cases of illusory and hallucinatory experience. Veridical experience and illusory/hallucinatory experience, then, provide the same justifying grounds for belief. Similar reasoning will apply to any domain of knowledge for which it makes sense to talk about being deceived by appearances through no fault of one’s own. So, for instance, once one accepts the interiorized conception of the space of reasons, one is led to a Highest Common Factor conception of memory or testimony (cf. McDowell 1998c: 397). Once one accepts the interiorized conception of the space of reasons, and the Highest Common Factor conception of experience (and memory, and testimony, etc.) that goes with it, hybrid conceptions of knowledge seem to be the only alternative either to skepticism or to a brute

Introduction

9

externalism regarding knowledge. For on a Highest Common Factor conception of experience (or memory, or testimony), the grounds provided by experience (or memory, or testimony) never guarantee the truth of one’s relevant beliefs. In this case, one’s options seem to be either to deny that knowledge requires truth-guaranteeing grounds, or to admit a wideranging skepticism. According to McDowell, however, we should reject both the interiorized conception of the space of reasons and the Highest Common Factor conception of experience that it generates. We should allow that we are liable to mistake what grounds we have for our beliefs, while holding fast to the idea that in the good case, when we have knowledge, our grounds for belief are truth-guaranteeing. Once we reject the suspect view of the space of reasons, the disjunctive account of appearances provides us with an alternative view of the grounds we possess for our knowledge. One’s knowledge that p need not rest on the grounds that it merely seems to one that p. Rather, in cases of knowledge one’s ground for believing that p is the very fact that p making itself manifest to one’s subjectivity. The disjunctive conception of appearances, then, allows for a view of knowledge in which we possess and are aware of possessing justifying grounds for our beliefs that guarantee the truth of these beliefs. McDowell’s view leaves it open that there may be cases in which we mistakenly take ourselves to have an experience in which a fact makes itself manifest to our subjectivity. In this case, neither do we have perceptual knowledge nor does our experience provide us with justifying grounds that guarantee the truth of our beliefs. On McDowell’s view, however, this does not give us reason for calling into question the idea that normally we know that we possess grounds that guarantee the truth of our beliefs. As McDowell states, Defective exercises of a perceptual capacity can be indiscriminable from non-defective exercises. It is a mistake to infer that even on an occasion on which the capacity is working perfectly, the current exercise of it is, for all one knows, defective. (2011: 42) Rather than making this mistaken inference, McDowell would have us consider perceptual knowledge to be the manifestation of a fallible capacity for self-conscious knowledge. When this capacity is fully manifested, one has perceptual knowledge, knowing that one’s experience guarantees the truth of one’s belief. But this capacity is a fallible capacity. Some of the times the manifestations of the capacity will be defective, resulting not in perceptual knowledge but in merely apparent knowledge. Nevertheless, the fallibility of this capacity does not prevent it from normally being exercised (cf. Ibid.: 50–53).

10

Casey Doyle et al.

As should be clear, there are important differences between McDowell and Pritchard’s formulations of epistemological disjunctivism, but there are also a number of shared characteristics. Both McDowell and Pritchard accept that in paradigmatic cases of perceptual knowledge, one is in a position to provide a reason for their belief that guarantees the truth of this belief. This entails that the reflectively accessible rational grounds or reasons we receive in experience differ in good cases, cases in which we have perceptual knowledge, and in bad cases, cases in which we fail to have knowledge, perhaps due to the fact that we are in the grips of illusory experience.

2. Epistemological and Metaphysical Disjunctivism As mentioned, early discussions of disjunctivism tended to run together different views and motivations under a single banner. This was unfortunate, because it obscured the different motivations and aims of different versions of disjunctivism, and so made their assessment difficult (Haddock and MacPherson 2008). Recent work has corrected for this by distinguishing between epistemological and metaphysical disjunctivism. While the core commitments of epistemological disjunctivism have been stated in the previous section, we can understand the core commitment of metaphysical disjunctivism as follows. Metaphysical Disjunctivism, Core Commitment: Veridical perceptual experiences and hallucinations differ mentally in some important respect. Stronger claims are often made by defenders of metaphysical disjunctivism. For example, some hold that veridical experiences and hallucinations are fundamentally different kinds of mental events (Martin 2004, Snowdon 2005). While denying that these events have a common kind, this view is consistent with accepting that veridical experiences and hallucinations share some important properties, such as phenomenal qualities. Others deny even this and insist that veridical perceptual experience and hallucination share no common mental element, sharing only the property of being subjectively indistinguishable (cf. Byrne and Logue 2008 for discussion). What matters for our purposes is the core commitment. When the distinction between epistemological and metaphysical disjunctivism is made, it is often accompanied by two further claims. The first is that defenders of one form of disjunctivism can remain neutral on the truth of the other (Snowdon 2005; Byrne and Logue 2008: 67; Pritchard 2012: 29; Soteriou 2016). Even if there is some significant, perhaps fundamental, mental difference between perception and hallucination, this needn’t be taken to have any immediate epistemological

Introduction

11

consequences (cf. the earlier discussion of how, for McDowell, the disjunctive account of appearances is necessary but not sufficient for the epistemological conclusion he wants). Just because the two mental events are different in their essences, it doesn’t follow that they are different in the support they lend to beliefs. (An analogy: claims about the metaphysical differences between zebras and cleverly disguised mules don’t have any epistemic upshot when it comes to answering questions about the presence of zebras.) So a metaphysical disjunctivist needn’t be an epistemological disjunctivist. Likewise, one might insist that the feature that makes a difference to the kind of support that veridical perceptual experience lends to belief isn’t essential or fundamental to the kind of mental event that it is. Thus, one can be an epistemological disjunctivist without being a metaphysical disjunctivist. The second claim is that McDowell, a prominent disjunctivist, accepts epistemological disjunctivism but not metaphysical disjunctivism (Snowdon 2005; Byrne and Logue 2008; Haddock and MacPherson 2008). A theme in many of our contributions is that we should reevaluate one or both of these claims. Let’s consider them in turn. Even if epistemological and metaphysical disjunctivism are logically independent, it might be thought that they are dialectically related (Neta 2008, Pritchard 2012: 20). One strategy for relating them runs as follows. Suppose that epistemological disjunctivism is true: veridical perceptual experiences and hallucinations differ in the kind of support they lend to beliefs. That doesn’t entail metaphysical disjunctivism. But we can ask: what explains the truth of epistemological disjunctivism? What is it about veridical perceptual experience and hallucination in virtue of which they lend different support to perceptual beliefs? A perfectly natural thought is that the two lend different support because they are different kinds of mental states; they differ epistemically because they differ in some fundamental way. Call this the “Explanatory Connection”. Explanatory Connection: Metaphysical disjunctivism provides the best explanation of the truth of epistemological disjunctivism. Stephen Wright presses the point by insisting that it is “mysterious how experiences that are metaphysically identical could come apart in terms of providing reasons” (2013: 256). Duncan Pritchard describes the Explanatory Connection as follows: In a nutshell, the disjunctivist view of the metaphysics of perceptual experience seems to offer the most natural way of explaining why there is this radical epistemic difference in these pairs of cases—viz., the reflectively accessible rational support is different because the very nature of one’s experiences is different. (2012: 24)

12

Casey Doyle et al.

While Pritchard mentions the connection and suggests that it is plausible, he does not discuss it in detail. That is because, as mentioned, his discussion of epistemological disjunctivism, like others (Neta 2008), is primarily defensive. On the other hand, denying that there are logical and explanatory connections between epistemological disjunctivism and metaphysical disjunctivism has some obvious benefits. Both are controversial claims beset with serious objections. If the Explanatory Connection is true, then defenders of epistemological disjunctivism must respond to the many objections facing metaphysical disjunctivism. For example, Tyler Burge has alleged that metaphysical disjunctivism is inconsistent with what we know about vision from empirical psychology (2005). In particular, he argues that it is inconsistent with the “Proximality Principle”, which roughly states that any two states of a perceptual system with the same proximal stimulation, internal input, and antecedent psychological states are the same type of perceptual state (2005: 22). Burge charges that metaphysical disjunctivism is committed to rejecting this principle, since it holds that veridical perceptual experience and hallucination are importantly or fundamentally different states despite sharing the three properties the principle mentions. As Bar-On and Johnson point out in their contribution (Chapter 16), the conjunction of the Explanatory Connection and the Proximity Principle spells trouble for epistemological disjunctivism. Their conjunction seems to show that epistemological disjunctivism is committed to an empirically untenable conception of perceptual experience. In her contribution (Chapter 10), Veli Mitova considers another reason to be wary of the Explanatory Connection: accepting both epistemological disjunctivism and metaphysical disjunctivism commits one to a particular conception of our motivating reasons for belief, one which she argues is implausible. While previous work has distinguished epistemological disjunctivism and metaphysical disjunctivism and has emphasized their logical independence, contributors here consider other reasons for and against holding both views together. The general consensus in the literature, such as it is, is that epistemological disjunctivism and metaphysical disjunctivism are distinct theses and that “disjunctivism” is ambiguous between them. When reading philosophers who embrace disjunctivism, one must ask: which do they accept? Philosophers interested in the nature of perception might accept metaphysical disjunctivism and remain neutral on epistemological disjunctivism (Martin 2004; Snowdon 2005; Fish 2009). Others embrace epistemological disjunctivism and reject metaphysical disjunctivism (Millar 2007, 2008). A common claim in the literature is that McDowell defends epistemological disjunctivism but not metaphysical disjunctivism (Byrne and Logue 2008; Snowdon 2005; Haddock and MacPherson 2008). The contributions from Lockhart (Chapter 5), Sedivy (Chapter 8), Haddock (Chapter 13), and Avramides (Chapter 18)

Introduction

13

contest this common reading. On the alternative, developed in different ways in these chapters, McDowell should be read as seeking to restore a commonsense picture of our perceptual relation to the world with both metaphysical and epistemological aspects, which cannot easily be pulled apart. As Sedivy puts it, “McDowell’s work has always emphasized that explaining perception’s objectivity and warrant is a single task”. And, as Avramides emphasizes, McDowell’s work has sought to undermine a Cartesian picture of the mind as a self-standing realm, divorced from the “external world”. The disjunctive conception represents an alternative to the Cartesian picture, and the rejection of this picture brings in its train both epistemological and metaphysical consequences. Our brief overview of McDowell’s disjunctivism in this chapter characterized it in terms of two moves: the proposal of a disjunctive conception of appearances (a metaphysical thesis), which undermines the felt need to endorse a Highest Common Factor conception of the support perception provides (an epistemological thesis). Although they differ in their focus, the chapters here that engage with McDowell’s work reveal that, on his version of disjunctivism, it is a mistake to suppose we can factor out the epistemological and metaphysical motivations for advancing views about perceptual experience.

3. Historical Antecedents and New Applications In addition to evaluating epistemological disjunctivism on its own and as it relates to metaphysical disjunctivism, our contributors consider disjunctive approaches to other sources of knowledge as well as the historical antecedents of disjunctivism. It is worth pausing to consider the significance of these topics. As noted earlier, McDowell and Pritchard offer different motivations for epistemological disjunctivism. McDowell’s motivations for epistemological disjunctivism require a disjunctive account of knowledge wherever there is a source of knowledge that depends upon appearances, and these appearances are such that we might be misled by them through no fault of our own. In contrast, Pritchard’s epistemological disjunctivism pertains only to visual perceptual knowledge. One issue addressed in this volume is the plausible scope of disjunctive accounts of knowledge. Are there sources of knowledge that allow us to be deceived by appearances but for which there is no plausible disjunctive account? Can Pritchardean motivations for epistemological disjunctivism be extended beyond perceptual knowledge? These chapters should be of interest to more than those with a craving for generality or a love of disjunctivism. As is the case with perceptual knowledge, the disjunctivist conception can be difficult to locate on the topography of views in the contemporary literature about memory, testimony, other minds, or introspection. (See especially Wright’s contribution in Chapter 15 on this point.) Considering the

14

Casey Doyle et al.

disjunctive conception in these other areas promises to help us better understand the available theoretical options. Discussions of disjunctivism situate it historically with reference to influential papers and a monograph by J.M. Hinton in the late 1960s and 1970s (Hinton 1967, 1973; see also the discussion of Hinton in Snowdon 2008). Others have noted that disjunctivism, in both its forms, has arisen largely out of Oxford and can be taken to descend from the “Oxford Realism” of Cook Wilson and H.A. Prichard (Kalderon and Travis 2013). Our volume considers other historical antecedents of the view, looking at the Stoics, Kant, Wittgenstein, and Austin. Rather than aiming for anything like being historically thorough, we have instead selected historical figures whose work is most likely to be of use to contemporary philosophers with an interest in disjunctivism. There are two reasons behind our selections. First, three of these figures have been influential in McDowell and Pritchard’s presentations of disjunctivism and, in the case of Austin, influential on both epistemological disjunctivism and metaphysical disjunctivism. Second, its defenders have long claimed that resistance to disjunctivism comes from broadly Cartesian conceptions of the relation between mind and world (see especially Schönbaumsfeld in Chapter 6 and Avramides in Chapter 18). If that is correct, then it would surely be useful to reflect on the work of philosophers who either are innocent of that Cartesianism, such as the Stoics, or are actively combatting it, as in the case of Kant, Wittgenstein, and Austin.

4. Summary of Chapters In what follows, we provide summaries of the chapters of the book as they appear. McDowell—“Perceptual Experience and Empirical Rationality” In this chapter, John McDowell defends his conception of perceptual knowledge from a particular objection. In McDowell’s view, perception is a rational capacity for knowledge through impacts upon the senses. In full exercises of the capacity, exercises that yield knowledge, one’s perceiving (as opposed to one’s merely seeming to perceive) that such and such is the case is the ground for one’s knowledgeable judgment. Since we cannot perceive something unless it is the case, these grounds guarantee the truth of the relevant belief. Furthermore, in McDowell’s account, perception contains grounds not only for perceptual knowledge, but for knowledge that one has the grounds of one’s perceptual knowledge. Our capacity for perceptual knowledge, as a rational capacity, is at the same time a capacity for knowing the grounds of our perceptual knowledge. McDowell defends this view from an objection. It is a “familiar fact” that one can, through no fault of their own, judge one’s self to perceive

Introduction

15

something when one does not. This can make it seem that a perception can never contain within itself the potential grounds for knowledge that it is a perception rather than a merely apparent perception. But McDowell argues that the familiar fact would tell against his conception of experience only if it were the case that we needed an independent capacity for knowing what our perceptual judgments are grounded on, and that this capacity could give us knowledge only if it were infallible. But this is exactly what McDowell denies. As such, the familiar fact of misjudging whether we perceive something shows us only that our rational capacity for perceptual knowledge is a fallible capacity. Pritchard—“Epistemological Disjunctivism and Anti-Luck Virtue Epistemology” How does epistemological disjunctivism relate to the wider issue of the nature of knowledge? Duncan Pritchard answers this question by considering how epistemological disjunctivism, construed as a claim only about a particular kind of knowledge, might be embedded within a theory of knowledge he calls anti-luck virtue epistemology. This is a theory of knowledge that, he claims, can accommodate both virtue-theoretic and anti-luck insights about knowledge. He maintains that not only is epistemological disjunctivism (so construed) consistent with anti-luck virtue epistemology, but that it is also a natural fit. Pritchard here distinguishes between anti-luck virtue epistemology as a general schematic proposal that can accommodate several distinct interpretations, and a specific rendering of anti-luck virtue epistemology that he claims is particularly plausible. He argues that the latter, even despite being a form of epistemic externalism about knowledge, is entirely compatible with the distinctively decisive internalist epistemic support provided by epistemological disjunctivism. What they share, he claims, is the idea that knowledge is primarily more a matter of skillful cognitive manifestation than of being the product of ratiocination. What epistemological disjunctivism brings to anti-luck virtue epistemology, however, is the necessary kind of cognitive contact with reality to ensure that knowledge is, contra the radical skeptic, even possible. Vasiliou—“Ancient Philosophy and Disjunctivism: The Case of the Stoics” In his contribution, Iakovos Vasiliou reads the Stoics as assuming epistemological disjunctivism. Central to Stoic epistemology is the doctrine of cataleptic appearances. According to the stoics, cataleptic appearances (a) arise from what is, (b) are stamped and impressed exactly in accordance with what is, and (c) are of such a kind as they could not arise from what is not. Vasiliou notes that contemporary commentators either give an “externalist” interpretation of cataleptic appearances, or they understand the Stoic doctrine to fail, because for any veridical experience there can be phenomenologically indistinguishable illusory experience. Vasiliou

16

Casey Doyle et al.

argues, however, that we can understand the Stoic doctrine concerning cataleptic appearances as a kind of epistemological disjunctivism. On this understanding, cataleptic appearances are factive, and we can know that we have them by reflection alone, since the cataleptic appearances are both the criterion of truth regarding appearances and of whether appearances are cataleptic. While understanding cataleptic appearances in this way aligns the Stoic doctrine with the epistemological disjunctivism of John McDowell and Sebastian Rödl, Vasiliou notes that McDowell diverges from the Stoics in denying the possibility of a Sage who through their wisdom only ever assents to cataleptic appearances. Lockhart—“The Kantian Roots of Epistemological Disjunctivism” The interpretation of Kant has played an important role in the development of John McDowell’s version of epistemological disjunctivism. In his contribution, Thomas Lockhart explains the connection between McDowell’s disjunctivism and his reading of Kant. Disjunctivists hold that, in the good case, perceptual experience provides a factive reason or indefeasible warrant for belief. According to McDowell’s brand of internalism, beliefs based on reasons must be “manifestations of rationality as such”. That is, they must be such that the subject can determine, by reflection alone, that they support the belief, were she to “step back” and assess its credentials. If the support is factive or indefeasible, then reflection upon one’s perceptual experience must present it as “revealing” the world. Thus, the disjunctivist needs a conception of perceptual experience on which it reveals the world to the subject in such a way that beliefs based upon it are manifestations of rationality. According to Lockhart, McDowell finds in Kant’s concept of an intuition the resources for such a conception. While leaving open the question whether Kant himself was an epistemological disjunctivist, Lockhart explains why the idea of an intuition matters for McDowell’s Kant and, in particular, why epistemological disjunctivists might make use of an account of perceptual experience on which conceptual capacities are “drawn on” in it. Lockhart’s chapter explores many of the same themes as Sedivy’s, for both defend McDowell’s claim that perceptual experience “reveals the world” and that our conceptual understanding must inform our perceptual experience. Lockhart shows the relevance of history to contemporary issues in philosophy of mind and epistemology. Schönbaumsfeld—“Was Wittgenstein a Disjunctivist Avant la Lettre?” Genia Schönbaumsfeld considers the question whether Wittgenstein was a disjunctivist. Given that Wittgenstein is widely thought to be a quietist whose approach to philosophy is therapeutic rather than constructive, it might be thought the answer would have to be “no”. But that presumes that disjunctivism must be thought of as a bit of constructive philosophical theorizing. Like Sedivy, Schönbaumsfeld rejects that characterization,

Introduction

17

instead insisting that disjunctivism is merely the articulation of our commonsense conception of perceptual knowledge. Central to that conception is the rejection of what she calls the “Reasons Identity Thesis”, that is, the Highest Common Factor conception of the support that perceptual experience provides belief. As Schönbaumsfeld notes, the Reasons Identity Thesis tends to be motivated by a “fully Cartesian” conception of the mind on which our encounter with objects in perception is always mediated by representational states that “fall short of the facts”. Once in the grips of the Cartesian conception, it seems that one must provide a proof that there is an external world, “without any help from perception”, which could then entitle one to accept one’s perceptual experiences at face value. The Cartesian conception, in turn, seems to be forced on us by the possibility of radical doubt, that is, a doubt that undermines all of our beliefs about the “external world” in one go. In On Certainty, Schönbaumsfeld finds an argument that radical doubt is unintelligible. The consequences of this argument are powerful. As she puts it: If Wittgenstein is right . . . that radical doubt is an illusion, then the notion that we either first need to establish that there is an external world, or else groundlessly take it ‘on trust’ (or ‘assume’) that there is, goes out of the window (p. 120). If radical doubt is an illusion, then we do not stand in need of a proof of an external world, and so we do not face the prospect of constructing claims about the external world out of claims about our sense experience. “[P]hysical object propositions and propositions about experiential seemings can be cognitively on a par”. The rejection of radical doubt restores our commonsense conception of perceptual knowledge by undermining the motivations for the Reasons Identity Theory. In addition to its exploration of Wittgenstein’s relation to disjunctivism, Schönbaumsfeld’s essay offers an interpretation of Wittgenstein’s writings on Moore’s “proof” and a novel reading of Wittgensteinian “hinge propositions”. (For an alternative reading, one which also considers the relation between Wittgenstein’s views and disjunctivism, see Pritchard 2016). Longworth—“Settling a Question: Austin and Disjunctivism” In this contribution, Guy Longworth considers J.L. Austin’s relationship to contemporary disjunctivism and highlights the ways in which epistemic considerations figure into the debates concerning metaphysical disjunctivism. Longworth argues that while Austin sets the stage for contemporary disjunctivism, he failed to go beyond a schematic disjunctivist position, one that espouses commonsense realism regarding perception. According to Longworth, the key move of contemporary disjunctivists is to block the inference from a lack of distinguishable differences between veridical and hallucinatory experience to the claim that there

18

Casey Doyle et al.

is no qualitative difference between veridical and hallucinatory experience. While Austin rejects the idea that all qualitatively identical experiences must be identical in every way, he does not explicitly reject the idea that indistinguishable experiences are qualitatively identical. Nevertheless, Longworth shows how Austin’s thought anticipates the key move of contemporary disjunctivism. By distinguishing between cognitive and sensory powers, and by recognizing introspection as a fallible cognitive power, Austin opens up the idea that we might be unable to distinguish between veridical and hallucinatory perception, not because these kinds of experience are qualitatively identical, but because of cognitive limitations inherent to human introspection. So while Austin fails to make the key move of contemporary disjunctivism, he make this move possible. Sedivy—“Disjunctivism and Realism: Not Naïve but Conceptual” As mentioned, it has become common to label John McDowell an epistemological disjunctivist where this is interpreted to mean that his position is neutral on questions about the metaphysics of perception. Sonia Sedivy’s contribution challenges this reading and outlines what she calls McDowell’s “conceptual realism”. This position is comprised of two central theses, both of which are part of common sense: that our conceptual understanding informs our seeing, and that in perceptual experience features of our environment can be made “manifest” for us. The first claim amounts to the idea that our experience picks out instances of kinds and properties. We recognize these kinds in perception thanks to our conceptual understanding. The second idea is that objects are “presented”, “revealed”, “manifest”, or directly there for us in experience. (See Lockhart for a discussion of both claims as they relate to Kant’s conception of an intuition.) As Sedivy emphasizes, this position should be contrasted with naïve realism, which is the standard view assumed in discussions of metaphysical disjunctivism. Sedivy argues that an advantage of conceptual realism is that it makes available resources for explaining things— such as the commonality between good and bad cases and the epistemic upshot of metaphysical disjunctivism—that naïve realism does not. She also considers how the de se content involved in perceptual experience on the conceptual realist picture might play an explanatory role in aesthetic experience, thus revealing the significance of a disjunctivist conception of experience for areas of philosophy beyond epistemology and philosophy of mind. French—“Epistemological Disjunctivism and Its Representational Commitments” In his contribution, Craig French develops what he calls the thing-seeing approach to epistemological disjunctivism. According to this approach, “[t]he particular kind of rational support that the epistemological

Introduction

19

disjunctivist claims that our beliefs enjoy in paradigm cases of perceptual knowledge is that provided by visual perceptual states or episodes of thing-seeing” (p. 120). The thing-seeing approach is non-committal about whether perception involves propositional or representational content. Because of this, French’s thing-seeing approach allows one to be an epistemological disjunctivist without taking on controversial positions in the philosophy of perception. Much of French’s chapter is dedicated to defending the neutrality of the thing-seeing approach from two objections. The first objection holds that thing-seeing cannot provide factive rational support unless it involves propositional content, since factive rational support can only be provided by factive propositional attitudes. French’s response is that while states of thing-seeing might not offer factive rational support in this technical sense, given that thing-seeing is relational, they offer factive rational support in the sense that they guarantee the truth of the relevant perceptual beliefs. The second objection is that if perceptual states were non-representational, then they could not provide rational support for our beliefs. Engaging with the work of Hannah Ginsborg, French suggests that the conscious character of states of thing-seeing could allow states of thing-seeing to rationalize our beliefs by providing us with the relevant sorts of reasons, even if perceptual states were non-representational. As a result, the thing-seeing approach allows one to accept epistemological disjunctivism while being agnostic about the nature of perception. Mitova—“Either Epistemological or Metaphysical Disjunctivism” Veli Mitova demonstrates that philosophers interested in disjunctivism need to keep in mind debates concerning the nature of reasons. She does this by arguing against the conjunction of epistemological and metaphysical disjunctivism in a novel way. According to Mitova, the conjunction of epistemological and metaphysical disjunctivism implies disjunctivism about motivating reasons for perceptual belief. According to this view, motivating reasons for perceptual beliefs are either factive states partly external to the agent, or non-factive states purely internal to the agent. But disjunctivism about motivating reasons for belief is an implausible view. Mitova has us consider a case in which an individual contemplates a flowering tree for an hour; during this time, an evil demon intervenes at three different points, interposing a hologram of the flowering tree. If disjunctivism about motivating reasons is correct, then there are seven different motivating reasons for why the individual believes there is a flowering tree in front of her. In Mitova’s view, this makes motivating reasons for belief too independent from the believer’s perspective, calls into question the idea that motivating reasons are causes, and on the whole makes motivating reasons too protean—it simply isn’t plausible that our motivating reasons change so fluidly. Thus, we should not hold the conjunction of epistemological and metaphysical disjunctivism. This result

20

Casey Doyle et al.

goes against some traditional strategies for arguing for metaphysical disjunctivism in which the truth of epistemological disjunctivism is offered as a reason for accepting metaphysical disjunctivism, and it threatens both epistemological and metaphysical disjunctivism if these mutually imply each other. Littlejohn—“Neither/Nor” Clayton Littlejohn attacks epistemological disjunctivism on the grounds that it mistakenly makes the justification of our non-inferential perceptual beliefs depend upon their being based upon sufficient reasons for belief. In particular Littlejohn rejects the idea that “[i]f S knows that p, S’s being in a position to know supervenes upon the reasons she possesses and her belief is properly based on sufficient reasons” (p. 217). These reasons would be reasons that put one in position to know that something is the case. Littlejohn’s attack is two-pronged. First, Littlejohn attacks the idea that “seeing that p” is a visual relation between a perceiver and a fact that could put us in a position to know that p. Littlejohn has us imagine a case in which one sees a tomato from a very far distance, and one represents what one sees as a tomato. Given that one is so far away from the tomato, one is not in a position to perceptually know that what one sees is a tomato. This creates a dilemma for the disjunctivist. Either she must deny that in this case one sees that there is a tomato in front of them, or she must deny that seeing that there is a tomato in front of oneself always puts one in a position to know. Second, Littlejohn attacks the idea that the normative status of belief supervenes upon the reasons at hand of an individual believer. Littlejohn argues that we are tempted to believe that our rational evaluable responses supervene on the reasons we have at hand because we are tempted to believe that we could always (at least in principle) respond to the reasons that we have. Once we can see how to reject this latter idea, however, we will find ourselves free to reject the former. Neta—“Disjunctivism and Credence” Epistemological disjunctivism implies that in paradigmatic cases of perceptual knowledge the successful perception of a mind-independent object provides a better basis for belief about that object than an indistinguishable illusion or hallucination can. Ram Neta uses this idea to formulate an objection to epistemological disjunctivism. This objection makes two plausible assumptions. The first is that S’s having a better basis for believing that p at time t2 than at time t1, implies that, all else being equal, S has better evidence for p at time t2. The second is that if S has better evidence for p at time t2 than at time t1, then S is rationally obliged to be more confident in p at t2 than at t1. With these assumptions in place, Neta has us suppose that someone gradually transitions from being under the grips of an illusion or hallucination into successfully

Introduction

21

perceiving a mind-independent object and so having perceptual knowledge concerning this object. If epistemological disjunctivism is true, then the individual would have better grounds for their belief. Given the two plausible assumptions, the individual would then be rationally obliged to be more confident in the relevant perceptual belief at the end of the transition. But this seems plainly false. Neta’s response to the objection is to reject the assumption that having better evidence rationally requires us to be more confident. Neta holds that while rational confidence is determined by evidential support, evidential support is not determined simply by what is in one’s evidence set. Rather, evidential support is determined by what is in one’s evidence set and how confident one should be that one has this evidence. If there is an imperceptible change to one’s evidence set, one should be neither more nor less confident that they have a particular set of evidence. Haddock—“Disjunctivism, Skepticism, and the First Person” With some notable exceptions (McDowell 1998a, 1998b), discussions of epistemological disjunctivism have focused on the epistemology of perception. Some chapters of this volume consider disjunctivist proposals for other domains of knowledge, such as testimony (Wright), memory (Milburn and Moon), other minds (Avramides), and self-knowledge (Bar-On and Johnson, Doyle). Adrian Haddock presents disjunctivism as a conception of knowledge generally, one that aims to complete the project of modern epistemology, paradigmatically expressed in Descartes’s Meditations. According to Haddock, that project is essentially first-personal, in that it aspires to provide an account of knowledge in the face of the possibility of error that would yield self-understanding, rather than a merely theoretical analysis from a third-person perspective. Roughly, this means that it aims to articulate an account of knowledge drawing on resources available to the subject herself, from the first-person point of view, that would reveal that she can know despite the possibility of error. As Haddock emphasizes, the project of modern epistemology, so understood, has for the most part been ignored in contemporary epistemology, which favors the third-person perspective. On Haddock’s reading, central to McDowell’s disjunctivism is the thesis that “there is no knowledge without self-knowledge”, which reveals the disjunctivist to respond to the Modern project. How could self-knowledge help in revealing the possibility of knowledge in the face of error? One way would be Cartesian: we happen upon a self-ascription expressible as “I think that p”, which is self-verifying, guaranteeing the truth of both the self-ascription and the proposition contained within it (e.g. “I experience that there is a red square”). But this proposal is toothless if we want to vindicate our knowledge claims when p is not verified by my thinking that I think it, that is, knowledge claims about the external world. The disjunctivist denies that

22

Casey Doyle et al.

the self-knowledge that can secure our claims to know the external world must be self-verifying. But, for just this reason, it seems unable to respond to the skeptical threat. If I defend my claim to know that p by citing the fact that I see that p, we can simply ask how I know this. Since “I see that p” isn’t self-verifying, it would seem we stand in need of some further justification for our self-knowledge. On Haddock’s telling, then, the disjunctivist’s appeal to self-knowledge is of no use. But there is an insight behind the thesis that knowledge requires selfknowledge: self-knowledge is not knowledge of some fact in addition to the fact that p, but is, instead, a “use of ‘I perceive p’ [that] does reveal this thought [that p] to be true, because to use ‘I perceive p’ is to grasp, not merely the thought expressed by ‘p’, but the truth of this thought” (Haddock, p. 264). Standing in the way of this insight is the widespread view, rejected by Anscombe, that “I” is a referring expression. Offering a novel interpretation of her argument, Haddock argues that, if “I” does not refer, then the project of modern epistemology and so the skeptical threat are incoherent. Milburn and Moon—“Two Forms of Memory Knowledge and Epistemological Disjunctivism” In their chapter, Joe Milburn and Andrew Moon consider the prospects for extending Pritchard’s disjunctivism to memory knowledge. Milburn and Moon distinguish between experiential and stored memory knowledge. Experiential memory knowledge is knowledge of the past that is grounded upon memory experience. Stored memory knowledge is simply knowledge that we retain, having acquired it in the past. Milburn and Moon argue that the prospects of a disjunctivist account of experiential memory knowledge are as good as the prospects of such an account for perceptual knowledge. Just as we commonly justify our knowledge claims by appealing to the fact that we see that something is the case or that we see a particular object, so too do we commonly justify our memory knowledge claims by appealing to the fact that we remember that something is the case, or that we remember a particular object or event. And just as seeing that is factive and seeing an object is relational, so too remembering that is factive, and remembering an object or event is relational. So we have reason to believe that in paradigmatic cases of experiential memory knowledge we know in virtue of having factive, reflectively accessible rational support. Things are different, however, in the case of stored memory knowledge. First, there is reason to deny that one possesses stored memory knowledge that p in virtue of possessing any rational support for p. Second, even if we possess stored memory knowledge that p in virtue of possessing rational support for believing that p, there are paradigmatic cases of stored memory knowledge that seem to lack factive rational support, for example stored memory knowledge that rests on inductive grounds.

Introduction

23

Wright—“Testimonial Disjunctivism” Stephen Wright considers the viability of an epistemological disjunctivist treatment of testimonial knowledge. His discussion is structured around three questions: (1) What are the elements of disjunctivism about testimony? (2) How does disjunctivism fit into the existing topography of views on testimony? (3) Is disjunctivism about testimony viable? On the disjunctivist proposal Wright considers, a subject can know that p on the basis of hearing the speaker say that p, where this is a factive and reflectively accessible reason. As Wright explains, this position can be difficult to locate on the familiar map of positions in the epistemology of testimony, divided between Reductionists and Anti-Reductionists. Like Milburn and Moon, Wright observes that some of the objections to epistemological disjunctivism about perception and the responses to them provided by Pritchard and McDowell are not restricted to the perceptual case. Thus, the disjunctivist about testimony can appeal to them in defending her view. Wright then argues that, despite some reasons to be optimistic about disjunctivism about testimony, it is unclear what would count as a paradigmatic instance of testimonial knowledge, on the disjunctivist analysis. Like Milburn and Moon, Wright abstracts from the case of perceptual knowledge and considers epistemological disjunctivism, as formulated by Pritchard (2012), as a schematic view, applied here to the case of testimony. In addition to its insights about testimony, the chapter provides a template for thinking about disjunctivism in other domains of knowledge. Bar-On and Johnson—“Epistemological Disjunctivism: Perception, Expression, and Self-Knowledge” In important work, including her monograph Speaking My Mind (2004), Dorit Bar-On has developed a Neo-Expressivist account of first-person authority. Such an account must explain why self-ascribing a mental state in speech or thought from the first-person perspective, issuing an “avowal”, is immune from most ordinary forms of doubts and scrutiny. The explanation is that avowals express first-order mental states because they are issued from them. In a good case, a subject’s ground for selfascription is the very mental state self-ascribed, whereas, obviously, it cannot be that in the bad case. Thus, the Neo-Expressivist proposal is committed to the disjunctivist claim that good and bad cases differ in the support they lend. In their contribution, Bar-On and Drew Johnson compare disjunctivism about self-knowledge with the more-familiar disjunctivism about perceptual knowledge. Like others, they accept that epistemological disjunctivists should accept a version of metaphysical disjunctivism. Impressed by an argument from Tyler Burge mentioned earlier, Bar-On and Johnson contend that the latter is incompatible with principles about the metaphysics of perception. Disjunctivism about selfknowledge avoids this worry, since it denies that self-knowledge is based

24

Casey Doyle et al.

on introspection or any form of perceptual experience. Next, they compare their favored Expressivist proposal with Constitutivism, another theory of self-knowledge committed to disjunctivism, arguing that the Neo-Expressivist proposal fares better. Finally, they consider ways in which this disjunctivist treatment of self-knowledge might help the more familiar case of perceptual knowledge. In the case of self-knowledge, BarOn and Johnson favor a view on which the subject’s support for selfascribing mental states is a form of entitlement, rather than justification, in Burge’s sense. Their suggestion is that epistemological disjunctivists can avoid difficulties from the metaphysics of perception by thinking of support in terms of entitlement. Doyle—“Ringers for Belief” Like Bar-On and Johnson, Casey Doyle considers a disjunctivist treatment of self-knowledge. Epistemological disjunctivism claims that when all goes well, seeing that p places one in a position to know that p and to know that one sees that p. But if the good and bad cases of perceptual experience are subjectively indistinguishable, how can the subject know, by reflection alone, that she is in the good case when she is? This is the worry from subjective indistinguishability. Doyle considers whether it confronts a disjunctive treatment of doxastic self-knowledge, focusing on the thesis that one can know that one believes that p on the ground that one believes that p. First-person selfascriptions of belief are based on their truthmakers. Defenders of this view (like Bar-On 2004; Zimmerman 2006; Travis 2012) have argued that it does not confront the worry. Doyle provides reasons for thinking that it does. He frames things in terms of the question whether there are, in Charles Travis’s terms, “ringers” for beliefs: cases in which it subjectively seems to one, in the first-personal way, that one believes that p, though one doesn’t. He argues that there are such cases, such as imagination, self-deception, and delusion. In these cases one consciously judges that p, or performs a conscious act indistinguishable from judging, thanks to which it seems to one that one believes that p, though one doesn’t. This argument presupposes a particular, if widely assumed, conception of conscious judgment as an episode in phenomenal consciousness, perhaps realized in inner speech. He concludes by considering an alternative view, on which self-consciousness of judgment. Such a view has been proposed by disjunctivists like McDowell, Rödl, and Kern in response to the subjective indistinguishability worry about perceptual knowledge. (This view is considered in detail in Vasiliou’s contribution, Chapter 4.) While it avoids the objection, it is a radical view about self-knowledge, which makes it difficult to see how our capacity for self-knowledge is fallible. Thus, in addition

Introduction

25

to considering an extension of epistemological disjunctivism to another domain, Doyle shows how reflection on self-knowledge can bring out something important about at least some formulations of epistemological disjunctivism about perceptual knowledge. Avramides—“Disjunctivism and Other Minds” McDowell’s version of disjunctivism about appearances was originally presented in the context of a discussion about knowledge of other minds. In her contribution, Anita Avramides considers the relation between McDowell’s disjunctivism in these two areas. Her focus is “McDowellian Disjunctivism”, meaning a disjunctivism that shares the motivations behind McDowell’s view. Avramides rejects the common picture of McDowell’s disjunctivism as merely an epistemological thesis. Her contention is that in both the perceptual and other minds cases, McDowell offers a diagnosis of the temptation to accept a Highest Common Factor conception of the epistemic import of perceptual experience and another’s behavior. In both cases, that view seems to be forced on one because of a commitment to a broadly Cartesian picture of the mind, conceived as an autonomous or self-standing realm populated by items to which we have infallible access. This picture of the mind threatens our knowledge of both the external world and other minds. If there is a divide between our sense impressions and objects in the world, skepticism looms. And if the minds of others are kept “behind” their behavior in an autonomous realm, then it is unclear how we could gain access to them. Since the same picture prevents us from accepting a disjunctivist view of the two domains of knowledge, it is “dubiously coherent” to be a disjunctivist about perceptual knowledge (for McDowell’s reasons) while resisting it as part of an account of knowledge of other minds. As Avramides puts it, drawing on Davidson, the epistemic barrier from the inside out (to the external world) is also a barrier from the outside in (to the mind of another). The two stand or fall together. Avramides’s chapter brings out why disjunctivists with broadly anti-Cartesian motivations should be more concerned with the problem of other minds, and it indicates how that problem is central to the philosophy of mind and epistemology.

References Burge, Tyler, 2005, Disjunctivism and perceptual psychology, Philosophical Topics 33: 1–75. Byrne, Alex and Logue, Heather, 2008, Either/or, in Adrian Haddock and Fiona Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 57–94. Crane, Tim, 2009, The pigs in my garden, Times Literary Supplement, 31 July 2009.

26

Casey Doyle et al.

Fish, William, 2009, Perception, Hallucination, and Illusion. Oxford: Oxford University Press. Haddock, Adrian and Macpherson, Fiona, 2008, Introduction: Varieties of disjunctivism, in Adrian Haddock and Fiona Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 1–24. Hinton, John Michael, 1967, Visual experiences, Mind 76(April): 217–227. ———, 1973, Experiences: An Inquiry Into Some Ambiguities. Oxford: Clarendon Press. Kalderon, Mark and Travis, Charles, 2013, Oxford Realism, in Michael Beaney, ed. The Oxford Handbook of the History of Analytic Philosophy. Oxford: Oxford University Press, pp. 489–517. Martin, M. F., 2004, The limits of self-awareness, Philosophical Studies 120: 37–89. McDowell, John, 1998a, Criteria, defeasibility and knowledge, in his Meaning, Knowledge, and Reality. Cambridge, MA: Harvard University Press. ———, 1998b, Knowledge and the internal, in his Meaning, Knowledge and Reality. Cambridge, MA: Harvard University Press. ———, 1998c, Knowledge by hearsay, in his Meaning, Knowledge and Reality. Cambridge, MA: Harvard University Press. ———, 1998d, Reply to Wright, in Crispin Wright, Barry Smith and Cynthia Macdonald, eds. Knowing Our Own Minds. Oxford: Oxford University Press. ———, 2010, Tyler Burge on disjunctivism, Philosophical Explorations 13(3): 243–255. ———, 2011, Perception as a Capacity for Knowledge. Milwaukee, WI: Marquette University Press. ———, 2013, Perceptual experience: Both relational and contentful, European Journal of Philosophy 21(1): 144–157. ———, this volume, Perceptual experience and empirical rationality. Millar, Alan, 2007, What the disjunctivist is right about, Philosophy and Phenomenological Research 74(1): 176–198. ———, 2008, Perceptual-recognitional abilities and perceptual knowledge, in Adrian Haddock and Fiona Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 330–347. Neta, Ram, 2008, In defence of disjunctivism, in Fiona Macpherson and Adrian Haddock, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 311–329. Pritchard, Duncan, 2012, Epistemological Disjunctivism. Oxford: Oxford University Press. ———, 2016, Epistemic Angst. Princeton, NJ: Princeton University Press. Snowdon, Paul, 2005, The formulation of disjunctivism: A response to Fish, Proceedings of the Aristotelian Society, New Series 105: 129–141. ———, 2008, Hinton and the origins of disjunctivism, in Fiona Macpherson and Adrian Haddock, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 35–56. Soteriou, Matthew, 2009, The disjunctive theory of perception, in Edward N. Zalta, ed. The Stanford Encyclopedia of Philosophy (Winter 2016 ed.). https:// plato.stanford.edu/archives/win2016/entries/perception-disjunctive/. ———, 2016, Disjunctivism. London: Routledge.

Introduction

27

Travis, Charles, 2012, While under the influence, in his Perception. Oxford: Oxford University Press, pp. 391–412. Wright, Stephen, 2013, Review: Duncan Pritchard: Epistemological disjunctivism, Dialectica 67(2): 252–257. Zimmerman, Aaron, 2006, Basic self-knowledge: Answering Peacocke’s criticism of constitutivism, Philosophical Studies 128: 337–379.

Part I

Situating Disjunctivism

2

Perceptual Experience and Empirical Rationality John McDowell

1. Perception is the capacity for knowledge through impacts of reality on the senses. At least since Aristotle, human beings have been singled out as a kind of living beings differentiated from others by rationality. In human beings, knowledge, including the knowledge yielded by perception, is rationality at work. As Sellars says, identifying something as a case of knowledge— and he means human knowledge—is placing an episode or state in the space of reasons, which he explains as the space of justifying and being able to justify what one says.1 I will say something at the end about how an account of human knowledge on these lines relates to what we might understand perception to be in non-human animals, which are not in the space of making and justifying claims. Perceptual experience, in the sense of experience in which the subject is perceiving, as opposed to merely seeming to be perceiving, affords human beings the grounds for the judgments that are the units of their perceptual knowledge: judgments that are knowledgeable by virtue of being non-defective acts of the capacity for knowledge that perception is. In an experience in which someone is perceiving, ways things are in the world are perceptually manifest to her. There is an obvious rationality in judging that things are some way if the judging subject can ground her judgment on an experience in which it is manifest to her that things are that way. And for a judgment to be perceptually knowledgeable just is for it to have that kind of rationality. To say that a judgment is grounded on an experience in which the subject perceives that things are as she judges them to be is just another way of saying the judgment is a non-defective act of the capacity for knowledge that perception is. One can be perceiving that things are some way only if things are that way. If a judgment that things are some way is grounded on an experience in which the subject perceives that things are that way, it conforms to a necessary and sufficient condition for being knowledgeable: the judgment’s being grounded as it is entails that things are as they are judged to be. Philosophers often think we should not put such a condition on perceptual knowledge; I will come back to this.

32

John McDowell

I have already broached the topic of empirical rationality. In judgments that are knowledgeable by virtue of being grounded on experiences in which the subject perceives that things are as she judges them to be, empirical rationality is at work in its fundamental mode. It is true that empirical rationality is also at work in judgments inferentially grounded on things the judging subject perceives to be the case, and that is relevant to the rational significance of perceptual experience; the idea of a capacity to have it perceptually manifest to one in one’s experience that things are some way makes sense only as the idea of part of a more extensive capacity for empirical knowledge, including a capacity for knowledge inferentially grounded on what one perceives to be the case. But for my present purposes, we can restrict ourselves to empirical rationality in what I said is its fundamental mode. I have spoken of the operation of empirical rationality in perceptual knowledge, and I think that is where we must begin in understanding empirical rationality. It is a familiar fact that, without carelessness or inattention, someone can judge that things are some way on the ground of an experience in which it appears to her that things are that way, though the experience is not one of perceiving that things are that way, and the judgment is not knowledgeable. Perhaps things are not that way at all; or perhaps things are that way, but that they appear to be that way in the experience is accidental in relation to the fact that they are that way. Such judgments are operations of empirical rationality in what I said is its fundamental mode, but they are defective operations of it. If someone makes a judgment of this sort, she takes her judgment to be knowledgeable in the way that figures in non-defective operations of empirical rationality in its fundamental mode, and that is why she is acting rationally in making the judgment. In her own view of herself, her judgment has the rationality that perceptually knowledgeable judgments have. But to give an account of how experience relates rationally to judgment, we should focus in the first instance on the role of experience in grounding judgments of the kind that judgments of this sort merely seem to their subjects to belong to. We should focus on non-defective operations of empirical rationality in the fundamental mode; we should focus on empirical rationality at work in perceptually knowledgeable judgments, judgments the subject makes on the ground that it is manifest to her in her experience that things are as she judges them to be. Compare the fact that to give an account of a kind of organ, for instance kidneys, we should focus on what kidneys do when they work as they should, and not let ourselves be distracted by things that can go wrong in the functioning of kidneys. There is a specific way in which the fact that there can be defective operations of empirical rationality in its fundamental mode can be distracting, and I will come to it soon.

Perceptual Experience

33

2. As Sellars in effect says, knowledgeable judgments are acts of their subject’s rationality. An act of rationality is self-conscious. Not only the act itself but also the subject’s ground for it are within the scope of the self-consciousness with which the subject engages in the act. Someone who makes a judgment knows, at least implicitly, not only what she judges but on what ground she judges it. That is an internalism in one sense: the ground on which someone makes a judgment must be available for her to know it in the act of judging, and that holds in particular for grounds by virtue of which judgments are knowledgeable. In the case of grounds by virtue of which judgments are perceptually knowledgeable, the point comes out in this: if someone who has expressed a bit of perceptual knowledge is asked “What is your ground for that?” or, equivalently, “How do you know that?” she can answer straight off, without further investigation, “I perceive (have perceived, can perceive) that it is so”. There is more to this internalism about the grounds for perceptual knowledge than I have put in place so far. In an experience that is one of perceiving, there are more ways things are perceptually manifest to the subject than she judges to be ways things are. There are things someone who is perceiving could knowledgeably judge to be the case on the ground of her experience but does not judge to be the case. If she did judge that things are some way on the ground that her experience makes it manifest to her that things are that way, then according to the internalism I have put in place, she would know, at least implicitly, that it is on that ground that she makes her judgment; that would be part of her self-consciousness in the act of her capacity for knowledge through perception that her judgment would be. It would be internal to her act of judging that she would know at least implicitly that it is grounded on an experience in which she perceives things to be as she judges them to be. Now, the potential for knowledge that experiences are experiences of perceiving extends beyond a subject’s knowing that experiences of perceiving ground judgments she actually makes. The potential for knowing that an experience is one of perceiving is internal to acts of the capacity for knowledge through perception, even in respect of the experience’s being such that it would provide a ground for knowledgeable judgments that she could but does not make. The rational significance of an experience of perceiving is that it is potentially a ground for perceptually knowledgeable judgments. We can conceive perceptually knowledgeable judgments as complete acts of the capacity for knowledge through perception, and we can conceive experiences of perceiving, in abstraction from whether they actually ground knowledgeable judgments, as in themselves partial acts of the capacity for knowledge through perception. Suppose an experience has as part of its rational significance that it could ground a knowledgeable judgment that things are some way; even if that

34

John McDowell

potential for grounding a judgment is not actualized in knowledge that things are that way, the availability for knowledge in self-consciousness that characterizes an act, even a partial act, of the capacity for knowledge through perception, as the act of rationality it is, includes a potential for knowing that aspect of the experience’s rational significance. So the internalist thought is not just that when someone makes a perceptually knowledgeable judgment, she knows at least implicitly, in the self-consciousness with which she makes the judgment, that her ground for the judgment is that it is manifest to her in her experience that things are as she judges them to be. As I said, an experience of perceiving has a rational significance that consists in its being potentially a ground for knowledgeable judgments. It has that rational significance independently of whether the potential for grounding judgments is actualized. And even if the potential for grounding some knowledgeable judgment that is part of the rational significance of an experience of perceiving is not actualized, the subject of the experience still has it in her power, just in undergoing the experience, to know that it has a rational significance that includes making it manifest to her that things are as she would judge them to be in that judgment. It is true that this potential for knowing aspects of the rational significance of an experience of perceiving cannot be actualized except in what I have described as complete acts of the capacity for knowledge through perception, knowledgeable judgments that things are some way the experience makes it manifest to the subject that things are. Someone cannot actually know she is enjoying an experience whose rational significance includes its being manifest to her in the experience that things are a certain way, except as part of the self- consciousness with which she makes the perceptually knowledgeable judgment that things are that way. Judging (in fact knowledgeably) that things are that way is included in knowing that it is manifest to her in her experience that things are that way. That does not preclude us from holding that a subject whose experience makes knowledge that things are some way available to her need not avail herself of that opportunity for knowledge; she need not judge that things are that way. Even if she does not judge that things are that way, her experience contains the potential for knowing that it is one in which it is manifest to her that things are that way, a potential that would be actualized if, and only if, the potential for grounding a knowledgeable judgment that things are that way, which is part of the experience’s rational significance, were actualized. 3. According to the internalism I have been elaborating, someone who makes a perceptually knowledgeable judgment knows, at least implicitly, that her ground for her judgment is an experience in which it is manifest to her that things are as she judges them to be. And that knowledge of her ground is an act of the same capacity for knowledge that is in act in

Perceptual Experience

35

the perceptually knowledgeable judgment she makes on that ground, the capacity for perceptual knowledge of how things are in the world. Her knowledge of what grounds her judgment is part of the self-consciousness with which she engages in the act of the capacity for perceptual knowledge that her perceptually knowledgeable judgment is. I have proposed an expansion to this internalism: an experience in which the subject perceives that things are a certain way contains a potential for knowledge that the experience has that rational significance, even if the experience’s potential for grounding a judgment that things are that way, which would be knowledgeable, is not actualized. In this expanded internalism, there is still only that one capacity for knowledge in play, the capacity for knowledge through perception of how things are in the world. If the potential for knowledge that the experience is one in which it is manifest to the subject that things are a certain way were actualized, that knowledge about the experience would be an act of that same capacity, the capacity for knowledge through perception of how things are in the world. The potential for knowledge that the experience has that rational significance is internal to the experience, even if it does not ground a judgment that things are that way, by virtue of the fact that the experience is, in itself, a partial act of that same capacity, the capacity for knowledge through perception of how things are in the world. Now, it is important that only one capacity, the capacity for knowledge through perception of how things are in the world, is at work here, because it preempts a line of thought that can seem to rule out the conception of perceptual knowledge I have been describing. I have described an epistemology for perceptual knowledge with two elements: first, a general internalism according to which a subject who makes a judgment knows, at least implicitly, the ground for her judgment—which, if her judgment is knowledgeable, is the ground by virtue of which it is knowledgeable—in her self-consciousness in making the judgment on that ground; and, second, a thesis about perceptual knowledge in particular, that the ground for a judgment that is perceptually knowledgeable is an experience in which the subject perceives, has it perceptually manifest to her, that things are as she judges them to be. Together, these theses imply that someone who makes a perceptually knowledgeable judgment knows, at least implicitly, in her selfconsciousness in making her judgment, that the experience on which she grounds her judgment is one in which she is perceiving things to be as she judges them to be. As I have acknowledged, it is a familiar fact that someone can, without carelessness or inattention, take herself to be perceiving that things are a certain way in her environment when she is not. As I said, that familiar fact can distract us. It does that by seeming to show that the account of perceptual knowledge that I have described cannot be correct.

36

John McDowell

The familiar fact shows that if an experience is not an experience of perceiving that things are some way, that need not be something the subject has it in her power to know as part of her self-consciousness in the act of judging that things are the way it appears to her that they are in the experience. And it can seem to follow that if an experience is an experience of perceiving that things are some way, that cannot be something the subject has it in her power to know as part of her selfconsciousness in the act of judging that things are the way it appears to her that they are in the experience. More generally, it can seem to follow that the potential for knowing that an experience is one of perceiving that things are some way cannot be, as in my expanded internalism, contained in the experience itself, even if the subject does not judge, on the ground of the experience, that things are that way. Perhaps a subject can know, in her self-consciousness in the act of making a judgment on the ground of an experience, that in the experience things appear to her to be as she judges them to be. Perhaps the potential for that knowledge can be contained in the experience itself. But according to this line of thought, a subject cannot know, in her self-consciousness in the act of making a judgment, that the experience that is her ground for the judgment is one in which she perceives things to be that way, as opposed to one in which things merely appear to be that way. The potential for knowledge that the experience is one of perceiving cannot be contained in the experience itself. On this account, knowing that an experience is one of perceiving that things are some way would require information extraneous to whatever knowledge is at the subject’s disposal in grounding judgments on the experience, or just in undergoing the experience. That would rule out the conception of perceptual knowledge that I have described. How does the familiar fact seem to establish that one cannot know, in one’s self-consciousness in making a judgment on the ground of an experience, that the experience is one of perceiving that things are as one judges them to be; and, more generally, that the potential for knowing that an experience is one of perceiving that things are some way cannot be contained in the experience itself? The implicit argument has an extra premise, to this effect: if the potential for knowing that an experience is one of perceiving were contained in the experience itself, the knowledge that actualizes that potential would have to be an act of a capacity that would enable one to know of any of one’s experiences, whether it is an experience of perceiving or not, provided that one exercised the capacity with care and attention. The familiar fact shows that there is no such capacity. If there were, it would not be possible to take an experience to be one of perceiving when it is not, except through lack of care or inattention. But the familiar fact is that that does happen. And now the argument is that, since there is no such capacity, there can be no such knowledge.

Perceptual Experience

37

But this line of thought is, as I said, preempted by the fact that knowledge of one’s ground for a perceptually knowledgeable judgment is an act of the same capacity for knowledge that is in act in the perceptual knowledge constituted by the judgment; more generally, the potential for knowing that an experience has a rational significance by virtue of which it can ground perceptually knowledgeable judgments is contained in the experience itself by virtue of its being, as such, a partial act of that same capacity. A perceptually knowledgeable judgment is an act of the capacity to know through perception how things are in the world. What the familiar fact reflects is that the capacity to know through perception how things are in the world is fallible; it can issue in judgments that are not knowledgeable even if it is exercised with care and attention. In more or less unsurprising ways, the capacity can issue in judgments that are only seemingly knowledgeable, even if its exercise is free from that kind of flaw. Properly understood, this poses no threat to the thought that when the capacity works perfectly, it issues in a perceptual judgment that is knowledgeable in the way I have been describing: a judgment that something is so for which the subject has, and knows she has, a ground consisting in the fact that it is manifest to her in her experience that the relevant thing is so. That seems to be threatened only if we accept that additional premise. The threat comes from supposing that if someone could know, without information extraneous to what is contained in her perceptual experience, that she has such a ground for a judgment, that knowledge would have to be an act of a capacity other than her capacity to make judgments about how things are in the world that are correct by virtue of being perceptually knowledgeable; and that this other capacity, unlike the capacity to make judgments about how things are in the world that are correct by virtue of being perceptually knowledgeable, would have to be infallible: if one exercised it with care and attention on any experience one is under- going, it would be guaranteed to issue in a correct answer to the question whether the experience is one of perceiving or not. As before, the familiar fact shows that there is no such capacity. If we allowed this line of thought to begin, we might wonder why we should accept that the supposedly different capacity, the capacity to know the rational significance of one’s experiences, would have to be infallible. But the question does not arise, because, as I said, the capacity to know that an experience one is having has a rational significance by virtue of which it could ground a knowledgeable judgment that things are some way is not other than the capacity to make judgments about whether things are that way that are correct by virtue of being perceptually knowledgeable. The familiar fact shows that the capacity to make judgments about how things are in the world that are correct by virtue of being knowledgeable is not infallible, even if exercised with care and attention; it can issue in judgments that are not knowledgeable.

38

John McDowell

But that does not prevent us from recognizing that in non-defective acts the capacity issues in judgments that are knowledgeable. And knowing, at least implicitly, that the experience on which one grounds one’s judgment is an experience in which one perceives that things are as one judges them to be is part of how it is with the subject in a non-defective act of that same capacity, the capacity to make judgments about the world that are correct by virtue of being knowledgeable. We have to grant that one can innocently think one is perceiving that things are some way when one is not; one can innocently fail to know that an experience one is having is not an experience of perceiving that things are that way. But that is how it is in some defective acts of the capacity for perceptual knowledge, and there is nothing problematic or surprising about the fact that the capacity can issue in defective acts. It shows nothing about non-defective acts of the capacity. A subject can innocently fail to know that her experience is not one of perceiving, but that leaves unthreatened the claim that if her experience is one of perceiving, the potential for knowing that it is one of perceiving—and so, if you like, for knowing whether it is one of perceiving or not—is contained in the experience itself. 4. The line of thought I have been rejecting purports to show that if an experience is one of perceiving, that is not something the subject has it in her power to know just in undergoing the experience. It would follow that even if an experience on which someone grounds a judgment that things are some way is one in which she is perceiving that things are that way, it cannot be as such that the experience is the ground for her judgment that things are that way, except in a sense that allows the grounds for someone’s judgments to be unknown to her—a sense in which the ground for a judgment need not be what the judging subject, exercising her rationality, grounds her judgment on. This would be an externalism about grounds, in a sense corresponding to the sense in which the conception I have described is internalist. On this account, being an experience in which the subject perceives things to be some way must involve satisfaction of a condition extra to whatever it is about the experience that enables it to be what the subject, exercising her rationality, grounds judgments on; that an experience is one of perceiving cannot be its rational significance in the sense I have been using. There is no shortage of candidates for what the rational significance of an experience that is in fact one of perceiving might be, conformably to this line of thought. What they have in common is the idea that, even if an experience is one in which the subject is in fact perceiving that things are some way, the rational significance of the experience does not guarantee that things are that way, and so does not guarantee that if the subject grounds a judgment that things are that way on the experience, her judgment is true. The belief that some such position is compulsory is what underlies the thought I mentioned and promised to return to: that it cannot be right

Perceptual Experience

39

to impose on perceptually knowledgeable judgments the condition that the subject’s grounds for them, in the sense of what the subject goes on, exercising her rationality, in making them, guarantee that things are as they are judged to be in the judgments. But if we give up that condition, how can we suppose there is such a thing as perceptually knowledgeable judgment? If what someone grounds a judgment on, exercising her rationality, leaves it open that things are not as she judges them to be, how can that not imply that for all she knows things are not as she judges them to be? So how can her judgment be knowledgeable? On a conception of this kind, as I said, an experience’s being one of perceiving is extra to what enables its subject to ground judgments on it, exercising her rationality. No doubt we can make out a sense in which someone who grounds a judgment on an experience is in a better position, with respect to her judgment, if the supposedly extra condition for the experience to be one of perceiving is satisfied than if the extra condition is not satisfied. But according to this conception, even if the experience that the subject grounds her judgment on is one in which she is in fact perceiving that things are as she judges them to be, still whatever it is about the experience that enables her to ground her judgment on it, exercising her rationality, does not entail that things are the way she judges them to be. And, as I suggested, that seems to imply that for all she knows things are not as she judges them to be. If we cannot set aside the implication, the position constituted by satisfaction of the supposedly extra condition for an experience to be one of perceiving, even though that position is better in some sense, cannot amount to knowing that things are as the subject judges them to be. And I have never seen a plausible account of how we might contrive to set aside the implication. Usually philosophers do not address the question; they think we have to content ourselves, somehow, with counting as knowledgeable judgments for which the subject’s grounds, in the sense of what the subject goes on, exercising her rationality, in making the judgments, leave it open that things are not as she judges them to be. They think we have to settle for that, on pain of giving up the idea that there is such a thing as perceptual knowledge, because they think conclusive grounds are not to be had. I think settling for inconclusive grounds is giving up the idea that there is such a thing as perceptual knowledge. The fact is that conclusive grounds are to be had, in experiences of perceiving, conceived in the way I have described and defended. 5. I have so far spoken exclusively about perception as a capacity for knowledge in human beings. When Sellars equates attributing knowledge with placing something in the space of reasons, he does not note that what he says fits only rational subjects. As I said, non-human animals are not in the space of justifying

40

John McDowell

and being able to justify what they say. It might seem that Sellars means to deny knowledge to non-human animals except perhaps in a loose sense. But even if Sellars does mean to reject literal attributions of knowledge to non-human animals, which I doubt, there is no need for us to be so restrictive about the very idea of knowledge, perceptual knowledge in particular. There is a perfectly intelligible generic notion of knowledge, which we can begin to spell out in terms of being informed about things, and a correspondingly generic notion of perception, which we can begin to spell out in terms of a capacity to come to know about things in a way that depends on the functioning of sense organs. Sellars’s topic, to which I have aimed to make a contribution, is best understood as a special form in which the generic notions apply to human beings. It is no objection to the account I have been giving that it does not fit the form in which the generic notions apply to non-human animals.

Note 1. Sellars (1997: §36).

Reference Sellars, Wilfrid, 1997, Empiricism and the Philosophy of Mind with introduction by Richard Rorty and Study Guide by Robert Brandom. Cambridge, MA: Harvard University Press.

3

Epistemological Disjunctivism and Anti-Luck Virtue Epistemology Duncan Pritchard

1. Introductory Remarks One philosophical lesson that I have learnt over the years—learnt the hard way, I might add—is the importance of recognising when philosophical projects can be usefully run together, and when they really need to be kept apart. I’ve discovered more than once that projects that initially seemed to go naturally together are in fact to some significant degree orthogonal to each other. So, for example, in earlier work—see especially Pritchard (2005)—I had a tendency to suppose that one could usefully simultaneously offer an account of (i) knowledge in general, (ii) perceptual knowledge in particular, and (iii) how best to respond to the problem of radical scepticism. In fact, all three projects I’ve discovered—eventually!—need to be kept apart. This point is relevant to our current purposes, since now that I do keep these projects separate, it raises the question of how they all relate. In particular—see, especially Pritchard (2012b)—epistemological disjunctivism is core to how I respond to project (ii). But I am also quite explicit that epistemological disjunctivism is not, thereby, an answer to project (i). Relatedly, although I claim that epistemological disjunctivism is part of the solution to (iii), I am (these days anyway) very clear that it is not the complete solution.1 And yet these three projects are clearly all closely connected, even if one would be mistaken in simply trying to answer them all at once. So how do they all relate to one another? Since I have explored the question of how epistemological disjunctivism fits into a solution to the problem of radical scepticism at length elsewhere, my focus here will largely be on understanding how epistemological disjunctivism fits within an account of knowledge more generally.2 To that end, I will be locating this proposal within an approach to epistemology that I call anti-luck epistemology, and the theory of knowledge that goes along with that approach: anti-luck virtue epistemology.3

2. Anti-Luck Virtue Epistemology It is widely held in epistemology that there cannot be lucky knowledge. This is the so-called anti-luck intuition.4 The methodology behind

42

Duncan Pritchard

anti-luck epistemology is that we should take this claim seriously by offering an account of both luck and of the specific sense in which knowledge is incompatible with luck. The thought is that by putting these two elements together, we will be able to formulate, in a principled way, an anti-luck condition on knowledge. Sparing the reader the details somewhat, such a methodology, in my hands at least, leads, via a modal account of luck, to a defence of a specific formulation of the safety condition on knowledge. Roughly, to have knowledge is to have a true belief formed on a safe basis such that it could not have easily been a false belief, formed on that same basis.5 There is more to knowledge than merely non-lucky (i.e., safe) true belief, however. I have tried to make this point vivid via the following scenario. Imagine that the subject’s true belief is such that in all close possible worlds where the subject continues to form a belief on the same basis as in the actual world, her belief continues to be true. The twist in the tale, however, is that what is underwriting the subject’s safe—and hence non-lucky—true belief has nothing to do with her exercise of relevant cognitive agency, but is rather the result of the epistemically helpful interference from another agent.6 The crux of the matter is that while the subject clearly has a safe true belief, since it isn’t rooted in her exercise of relevant cognitive agency, it cannot count as knowledge. Merely having a non-lucky true belief thus cannot suffice for knowledge. Such cases are meant to remind us that aside from the anti-luck intuition that guides the theory of knowledge, there is also an ability intuition. This holds that a subject’s knowledge must in some substantive way be attributable to the subject’s exercise of relevant cognitive agency. The anti-luck and ability intuitions can at first blush seem like two sides of the same coin, and indeed they are often run together. Isn’t getting to the truth via one’s cognitive ability what excludes epistemic luck? Conversely, aren’t the conditions under which one gets to truth in a non-lucky way precisely those conditions where one has appropriately manifested one’s cognitive agency? The case just described where one has a clearly safe true belief without thereby having knowledge should make us realise that these two intuitions are imposing distinct constraints on a theory of knowledge. In this scenario, after all, while there will be nothing at all lucky about the subject’s cognitive success (i.e., her true belief), the fact that this non-lucky cognitive success is not the result of her manifestation of relevant cognitive agency ensures that it doesn’t count as knowledge. In short, the antiluck intuition is satisfied, but not the ability intuition. Crucially, we can also describe cases in which the subject’s true belief is clearly attributable to her manifestation of relevant cognitive agency but where it is even so just a matter of luck that the belief, given how it was formed, is true (and hence the belief is unsafe). This is because a success, cognitive or otherwise, can be fully attributable to one’s (cognitive)

Anti-Luck Virtue Epistemology

43

agency even when there are features of the modal environment that nonetheless ensure that it is unsafe (i.e., it is a success which, given how it was brought about, could very easily have been a failure). In short, the ability intuition is satisfied, but not the anti-luck intuition. These days I express this point in terms of what I call an epistemic twin earth case (Kallestrup and Pritchard 2014; Pritchard 2016b). Imagine two agents who are microphysical duplicates. They occupy a causally identical ‘local’ environment (i.e., the environment that they are currently causally interacting with), and have identical causal histories. Moreover, not only is their local environment identical, but also their ‘normal’ environment, where this concerns the features of the environment that they ordinarily causally interact with. The point of these stipulations is to keep fixed across the two subjects any possible factor that might be relevant to the attribution of agency. Now imagine that both subjects form a true belief in the very same proposition. The only way that our two agents’ situations differ is that for only one of them there are features of the modal environment that ensure that the belief so formed could have easily been false (for the other agent, the true belief, so formed, could not have easily been false). One subject thus has an unsafe, and thus luckily true, belief, while the other subject does not. And yet their manifestations of agency, and thus cognitive agency, are bound to be identical, and hence the extent to which their cognitive success is attributable to their manifestation of relevant cognitive agency is bound to be identical, too. This means that a pure—or, as I put it, ‘robust’—virtue epistemology, which is a proposal that essentially defines knowledge in terms of virtuous true belief, is unsustainable (just as a pure, or robust, anti-luck epistemology was argued to be unsustainable earlier).7 One’s cognitive skills and intellectual virtues are, after all, manifestations of one’s cognitive agency and hence are ways of capturing the ability intuition. But as epistemic twin earth cases illustrate, success (including cognitive success) that is appropriately attributable to agency (cognitive or otherwise), can nonetheless be modally fragile, in the sense that it is a success that could have easily been a failure (even if, as in the cognitive case, we keep the basis for belief fixed). So unless the proponent of robust virtue epistemology is to reject the anti-luck intuition, and hence allow that there can be lucky knowledge, their position is untenable.8 So the first kind of case demonstrates that there can be non-lucky true beliefs that don’t amount to knowledge because they are not attributable to the subject’s manifestation of cognitive agency. And the second kind of case (the epistemic twin earth case) demonstrates that there can be true beliefs that are fully attributable to the subject’s manifestation of cognitive agency while nonetheless being only luckily true (i.e., unsafe). The moral is that the anti-luck and ability intuitions, far from being two sides of the same coin, are in fact imposing distinct, albeit overlapping, constraints on a theory of knowledge. This means that no anti-luck condition

44

Duncan Pritchard

alone will satisfy the ability intuition, and no ability (i.e., virtue) condition alone will satisfy the anti-luck intuition. It follows that a pure, or robust, version of either anti-luck epistemology or virtue epistemology will inevitably be unsound. I conclude that we need a theory of knowledge that is duly sensitive to the distinct constraints imposed by these two master intuitions. In particular, notice that it would not be enough to simply bolt-on an antiluck condition to a virtue condition. Aside from the fact that it would be easy to ‘Gettierize’ such a view—one just needs both conditions to hold, but for entirely independent reasons—it wouldn’t in any case capture the dialectical situation as just described. After all, that these two intuitions impose distinct conditions on a theory of knowledge does not entail that they impose independent epistemic conditions; indeed, since they are overlapping intuitions, we have every reason to treat them as not being independent. Here is the theory of knowledge that I claim does fit the bill: anti-luck virtue epistemology (ALVE). This holds that one has knowledge when one has a safe (non-lucky) cognitive success (i.e., true belief) that is significantly attributable to one’s manifestation of relevant cognitive agency. A few points of clarification are in order regarding this account of knowledge. The claim that the safe cognitive success needs to be ‘significantly attributable’ to one’s manifestation of relevant cognitive agency means that one’s manifestation of relevant cognitive agency must play a significant part in a causal explanation of one’s safe cognitive success. Note that a significant part is not the same as being the primary or overarching part of the causal explanation, which is what would be required if the safe cognitive success were fully attributable to one’s manifestation of relevant cognitive agency. In addition, notice that the explanatory relationship is between one’s manifestation of relevant cognitive agency and one’s safe cognitive success, rather than one’s cognitive success simpliciter. Putting these two points together, it follows that ALVE is both in a sense stronger (i.e., more demanding) and in a sense weaker (i.e., less demanding) than robust virtue epistemology. It is stronger, in that the explanatory relationship is not merely with one’s cognitive success but rather with one’s safe (non-lucky) cognitive success. But it is also weaker, in that while robust virtue epistemology is interested in whether one’s manifestation of relevant cognitive agency is the primary or overarching component of a causal explanation of one’s cognitive success, ALVE is only interested in whether it is a significant (but not necessarily primary or overarching) component in such a causal explanation (albeit now regarding one’s safe cognitive success). It is the way in which the anti-luck and ability conditions on knowledge intersect in this fashion (i.e., as opposed to being independent) that enables ALVE to avoid the kinds of problems that face other proposals. That we have an anti-luck condition built in ensures that we do not need

Anti-Luck Virtue Epistemology

45

to worry about Gettier-style cases, including those, as in the epistemic twin earth case, where the subject’s lucky true belief is nonetheless fully attributable to her manifestation of relevant cognitive agency. And that we have an ability condition ensures that we can explain why knowledge is lacking in cases where the subject’s safe cognitive success has nothing to do with her manifestation of relevant cognitive agency, since the relevant explanatory demand is not met. But by weakening the nature of this explanatory relationship, we can also account for cases where an epistemically friendly environment ensures that the subject’s manifestation of relevant cognitive agency need not be the primary or overarching element in a causal explanation of her safe cognitive success in order for her to have knowledge, such as cases of easy testimonial knowledge.9 Note that ALVE is a structural proposal about the nature of knowledge, and as such, a range of different theories of knowledge could be plugged into that structure (Pritchard 2012a, 2017a). Consider, for example, how one ends up with a different way of thinking about knowledge if one treats the ‘relevant cognitive agency’ at issue in ALVE in terms of specifically intellectual virtues, rather more generally in terms of a wider class of cognitive traits (e.g., cognitive skills, faculties, and intellectual virtues). Given the essentially reflective nature of the intellectual virtues—not just in their application, but also in terms of their acquisition and maintenance—this would lead to a version of anti-luck virtue epistemology that was much closer in spirit to traditional internalist accounts of knowledge (i.e., those views which incorporate an internalist justification condition) than a competing position cast in terms of a broader notion of cognitive agency. We can thus distinguish between ALVE as a structural proposal, and particular renderings of the ALVE thesis. For what it’s worth, my own preferred rendering of ALVE—let’s call this ALVE*, to differentiate it from the more general, structural thesis—is along epistemic externalist lines. I claim that the relevant sense of cognitive agency in play is one that doesn’t essentially involve the intellectual virtues, or otherwise demand good reflectively accessible reasons on the part of the subject. I maintain that sometimes at least merely manifesting one’s cognitive agency in completely unreflective ways—such as when we appropriately engage our reliable cognitive faculties and in doing so form safe beliefs—can suffice for knowledge. So even though mature human knowledge might well tend to involve subjects having good reflectively accessible reasons in support of their beliefs—and even, one might hope, the manifestation of intellectual virtue—I would claim that this is not required for knowledge.

3. Anti-Luck Virtue Epistemology and Epistemological Disjunctivism While ALVE is a theory of knowledge in general, epistemological disjunctivism, at least as I present the view at any rate, is specifically a

46

Duncan Pritchard

claim about perceptual knowledge in paradigmatic conditions. This proposal holds that one’s perceptual knowledge in such conditions enjoys rational support that is both reflectively accessible and factive. In particular, it maintains that one’s perceptual belief in the target proposition, p, is supported by the factive reason that one sees that p (which entails p), where this is factive rational support that is reflectively accessible to the subject.10 The combination of the rational support being both reflectively accessible and factive makes the view distinctive. According to traditional views in epistemology, ‘internalist’, reflectively accessible rational support is by its nature non-factive. Indeed, internalist epistemic support is meant to not be world-involving at all, in the sense that it is epistemic support that one’s beliefs would enjoy even if those beliefs were radically false.11 Conversely, there are varieties of epistemic support that are world-involving, in the sense of being if not factive, then at least such that they entail empirical claims about the world (think, for example, of a reliability condition on knowledge); but these ‘externalist’ varieties of epistemic support are held not to be reflectively accessible. Epistemological disjunctivism, if coherent, thus offers a kind of epistemic internalism (via the reflective accessibility of the rational support) about the type of knowledge in question that can nonetheless incorporate aspects of epistemic externalism (in virtue of involving epistemic support that is directly world-involving, via the factivity of the rational support in play).12 I’ve defended the view at length elsewhere, so I don’t propose to review the arguments for and against epistemological disjunctivism here. Instead, I want to focus on how such a position should be understood within the wider ALVE theory of knowledge. As I just noted, I understand epistemological disjunctivism only as being about a specific kind of knowledge. In my view, it isn’t even concerned with perceptual knowledge in general, but only with paradigm cases of perceptual knowledge. But it is at least consistent with thinking of epistemological disjunctivism as applying to a particular kind of knowledge that it applies to all knowledge (in the sense that an epistemological disjunctivist thesis about knowledge in general would obviously entail an epistemological disjunctivist thesis about paradigmatic perceptual knowledge). Would such a generalised epistemological disjunctivism about knowledge be compatible with ALVE? Construed as a thesis about knowledge in general, epistemological disjunctivism would be a fairly radical form of (non-classical) epistemic internalism about knowledge. It would demand not just that all knowledge is supported by good, reflectively accessible reasons (i.e., of a kind that would satisfy an internalist justification condition), but that all knowledge is supported by reflectively accessible factive reasons.13 Such a view strikes me as implausible. The idea that one’s reflectively accessible reasons could be factive in cases of paradigm perceptual knowledge trades on the fact that sometimes at least the relevant fact to be known

Anti-Luck Virtue Epistemology

47

is made utterly manifest to one, as when one sees a pig before one in optimal conditions for pig perception. But not all perceptual knowledge is like that. Sometimes what we see is more suggestive, as when we see a curly tail go behind a wall in a pig pen. That might also lead to knowledge that a pig is present (knowledge which is at least significantly perceptual), but it doesn’t seem to be in any obvious way supported by a reflectively accessible, factive reason. Indeed, in the normal case anyway, it seems to be rather supported by a collection of non-factive, reflectively accessible reasons of the usual epistemic internalist kind. Alternatively, of course, one could embrace epistemological disjunctivism about knowledge in general and hence argue that genuine knowledge is not possessed in these scenarios involving mere non-factive, reflectively accessible rational support; of course, that would be a very restrictive account of knowledge to offer. In any case, putting concerns about the cogency of the proposal to one side, it ought to be clear that there is no obvious tension between epistemological disjunctivism construed as a general theory of knowledge and ALVE. Remember that ALVE is a structural proposal, and hence compatible with a range of different interpretations. Could a generalised epistemological disjunctivism not be one of them? The issue rests on whether we can think of knowledge supported by reflectively accessible factive reasons along the lines set out by ALVE. This, in turn, rests on what kinds of conditions epistemological disjunctivism claims need to be in place in order for one’s beliefs to enjoy such rational support. On my presentation of epistemological disjunctivism, at least, the conditions in play are very demanding, as befitting the fact that one only has this kind of rational support in epistemically paradigmatic perceptual conditions. In particular, one’s veridical perception has to be the result of one’s perceptual faculties functioning appropriately in epistemic conditions that are both objectively and subjectively good (this is what I refer to as the ‘good+’ case, to differentiate it from merely ‘good’ scenarios where the epistemic conditions are only objectively good). This means that not only is the subject in epistemic conditions that are in fact ideal for the subject to gain perceptual knowledge, but that these epistemic conditions also seem to the subject to be ideal for the gaining of perceptual knowledge (e.g., no misleading defeaters are in play).14 With the conditions for the obtaining of factive, reflectively accessible reasons characterised in this epistemically demanding fashion, it is fairly plausible that in satisfying the rubric for epistemological disjunctivism, the subject will thereby satisfy the corresponding rubric for ALVE. Let’s begin with the anti-luck element of ALVE. If one’s basis for belief is a factive reason, then the safety condition is straightforwardly met, as clearly one cannot form a belief on that same basis in close possible worlds and yet believe falsely.15 This shouldn’t surprise us, given that one needs to be in epistemic conditions that are both objectively

48

Duncan Pritchard

and subjectively good in order to enjoy the factive reason in play. In particular, objectively good epistemic conditions are explicitly defined as being conditions that put one in a position to know, to the extent that so long as one appropriately forms one’s belief in those conditions, one will acquire knowledge. This is how poor subjective epistemic conditions can prevent knowledge even when the objective epistemic conditions are optimal. A misleading (undefeated) defeater, for example, would prevent knowledge, since it would preclude one from appropriately forming a belief in the target proposition, even if the epistemic conditions are objectively good. Given that objective epistemic conditions are understood in this fashion, it shouldn’t be surprising that the subject has a basis for belief that would ensure truth not just in the actual world but also in the subject’s modal neighbourhood. The point is thus that in order to have reflective access to the factive reason, the subject must have a true belief that is safe. Moreover, notice that the safety of the subject’s belief, based on the reflectively accessible factive reason, is also significantly attributable to her manifestation of relevant cognitive agency. For remember that it is not just that factive reasons are reflectively accessible to the subject only in epistemic conditions that are objectively and subjectively good, but also that the subject herself has to be cognitively functioning in an appropriate manner. This means that having reflective access to the factive reason in play must go hand in hand with one’s appropriate cognitive functioning in this regard. Accordingly, given that these properly functioning cognitive capacities are generating factive reasons, which are in turn ensuring that one’s beliefs so formed are safe, one would expect that one’s safe cognitive success ought to be at least significantly (if not primarily) attributable to one’s exercise of relevant cognitive agency.

4. Localised Epistemological Disjunctivism Within ALVE* Epistemological disjunctivism, construed as a general theory of knowledge, is thus compatible with ALVE. But as I noted earlier, this would be a rather implausible account of knowledge, and thus an unattractive way of fleshing out an instantiation of the ALVE schema. This leads us to the more interesting question of how to think about the relationship between the kind of localised epistemological disjunctivism that I defend—whereby knowledge is supported by factive, reflectively accessible rational support in specific conditions—and ALVE. I noted above that my preferred way of cashing out the ALVE schema—ALVE*—is cast along epistemic externalist lines. This proposal allows that the manifestation of cognitive abilities in the broadest sense—where this includes mere cognitive skills and innate cognitive faculties, in contrast to intellectual virtues, specifically—can generate knowledge. Given that epistemological disjunctivism is meant to capture key epistemic internalist insights, one

Anti-Luck Virtue Epistemology

49

might think that ALVE*, as an epistemic externalist thesis, must be in at least some dialectical tension with it. On the contrary, I want to suggest that ALVE* and a localised epistemological disjunctivism are natural bedfellows. Notice that even though ALVE* is an epistemic externalist thesis, it is still meant to accommodate at least some of the intuitions that drive epistemic internalism, and in particular the idea that knowledge demands cognitive responsibility. In this respect, ALVE* follows the model of virtue-theoretic proposals in the theory of knowledge more generally. It is in the nature of one’s cognitive abilities, and which makes them more than just reliable belief-forming processes, that they are integrated aspects of one’s cognitive character. This is why knowledge that is attributable to one’s manifestation of relevant cognitive ability is thereby attributable to one’s cognitive agency, and thus to one’s cognitive character. In this way, knowledge on this view directly brings with it the notion of cognitive responsibility, in a way that a mere process reliabilism about knowledge does not.16 Where ALVE* departs from a version of ALVE that insists on the intellectual virtues—and which thus insists that knowledge arises out of intellectual character specifically—is that the latter brings in a more beefed-up notion of cognitive responsibility, one that is far more reflective given the nature of the intellectual virtues. Either way, there is a substantive notion of cognitive responsibility in play in all virtue-theoretic accounts of knowledge.17 As an epistemic externalist thesis, ALVE* doesn’t demand that the subject should have any reflectively accessible reasons in support of her true belief in order for that belief to amount to knowledge (though, as noted earlier, it is consistent with this point that most mature human knowledge does tend to be so supported). All that is important is that this true belief is formed in such a way that it is safe, and this safety is significantly attributable to the subject’s manifestation of relevant cognitive ability (and is thus attributable to her cognitive agency). When ALVE* is combined with epistemological disjunctivism, however, then there will be a paradigm kind of perceptual knowledge that is supported by reflectively accessible reasons; indeed, supported by reflectively accessible reasons of a particularly robust kind in virtue of being factive. Now one might think that the robustness of the internalist epistemic support at issue with epistemological disjunctivism in epistemically paradigm cases entails, even in the context of ALVE*, the manifestation of the relevant intellectual virtues. I don’t think that this follows. In general, being sufficiently proficient in adducing reflectively accessible (and usually non-factive) reasons will tend to require some degree of manifestation of intellectual virtue, but I don’t think that this applies to the kind of factive reasons available according to epistemological disjunctivism (even despite their being factive). This is because such factive reasons are meant to be epistemic standings that one is entitled to by default. If there

50

Duncan Pritchard

is nothing to alert one (or that should alert one) to there being anything epistemically amiss, and in fact there is nothing epistemically amiss, then one thereby has the factive reason available to one. It is in this sense that such factive reasons are not ‘earnt’, in that one doesn’t need to actively do anything from a rational point of view in order to acquire them. In particular, they don’t require any specific reasoning on the part of the subject, in the way that other kinds of rational beliefs do. In this sense, then, they are not the result of ratiocination, or otherwise grounded in ratiocination. One has to be inducted into the relevant epistemic practice, of course, and one’s cognitive agency needs to be functioning appropriately and suitably attentive to relevant environmental conditions (which, note, might not require any particular kind of conscious monitoring of that environment; being able to spot when conditions are abnormal does not require a conscious monitoring of when they are normal). Beyond that, nothing from a specifically rational point of view is required of the agent in order to have access to them.18 So while epistemological disjunctivism allows for an elevated (factive) kind of internalist epistemic standing, it doesn’t intellectualise what that epistemic standing involves. Indeed, it is interesting to note how well epistemological disjunctivism and ALVE* sit together in this respect. In being inducted into the epistemic practices of one’s community, and thereby being such that one counts as a believer at all, one has both externalist and a strong kind of internalist knowledge made available to one. In the former case, this is through one’s appropriate exercise of one’s cognitive abilities and faculties. In the latter case, this is specifically through one’s appropriate exercise of those cognitive abilities and faculties leading to one having reflective access to factive reasons. In neither case does one need to be particularly reflective, much less intellectually virtuous. The point is rather that one’s appropriate manifestation of one’s cognitive abilities and faculties can thereby put one in direct epistemic contact with the world. In the best case, this delivers knowledge grounded in factive, reflectively accessible reasons. The acquisition of the intellectual virtues, and thus the kinds of good reflective cognitive traits that go with them, such as good reasoning, is something that comes after one’s acquisition of factive reasons. This signals the fact that these involve cognitive traits that are intellectually more sophisticated, even though they marshal reflectively accessible reasons that are typically non-factive. This should not surprise us. Consider what it takes to make a complex observation of one’s environment (i.e., as opposed to merely passively perceiving one’s environment), of a kind that essentially involves intellectual virtue. To be observant, in the sense relevant to intellectual virtue, is to be able to interrogate one’s surroundings for salient information. This will mean attending to important details of what one sees and drawing inferences appropriately, suitably informed

Anti-Luck Virtue Epistemology

51

by one’s background knowledge. Even if factive reasons play some role in this reasoning, there will inevitably be non-factive reasons also in play. More importantly, the resulting beliefs will be buttressed by a complex piece of reasoning that delivers good, but defeasible, rational support. Even though the rational support that results is non-factive, this kind of justification of one’s belief generates a form of knowledge that is paradigmatically internalist. Nonetheless, it would be a mistake to think that internalist knowledge is always of this intellectualist kind, or even that this kind of internalist knowledge offers the strongest form of epistemic support for one’s beliefs. Knowledge supported by factive reasons enjoys stronger epistemic support but is also a kind of internalist knowledge that is cast along non-intellectualist lines. The combination of ALVE* and epistemological disjunctivism enables one to have a broadly externalist theory of knowledge that can also incorporate core insights from epistemic internalism. Indeed, it is a theory of knowledge that has factive internalist knowledge at its heart, in a non-intellectualist guise. This highlights an important point about epistemological disjunctivism, something that is brought into sharp relief when it is set within ALVE*: although it offers us a particularly robust kind of internalist epistemic support, it does not follow standard forms of epistemic internalism in treating our reflective rational processes as epistemically privileged. Our epistemic position is not one of reasoning from an epistemically secure ‘inner’ reflective realm to derive conclusions about an epistemically insecure ‘outer’ realm. Rather, our cognitive grip on reality is ensured by the development of cognitive skills that have been acquired through induction into a social epistemic practice, and which bring with them factive, reflectively accessible reasons. In this sense, a grip on the ‘outer’ realm precedes a mastery of any ‘inner’ realm involving reflective rational processes, where the former enables the latter rather than being derivative on it. This last point also highlights the importance of combining ALVE* with epistemological disjunctivism. The former by itself, after all, doesn’t secure a decisive reflective cognitive grip on reality at all. Without epistemological disjunctivism, that is, one gets at most defeasible, reflectively accessible rational support of the usual kind. But this is not enough, as the problem of radical scepticism demonstrates. In particular, as I’ve argued elsewhere, epistemological disjunctivism is the antidote to the underdetermination-based formulation of Cartesian scepticism, where anything else is either too weak to be effective, or too revisionary to be intellectually satisfying, in response to the paradoxical nature of this sceptical puzzle. ALVE* may thus be an adequate account of knowledge, but it lacks the resources by itself to show that widespread knowledge is possible, and for that we need the combination of ALVE* and epistemological disjunctivism.19,20

52

Duncan Pritchard

Notes 1. To be fair, this wouldn’t have been clear from Pritchard (2012b), though there are good reasons for this—for example, I wanted that book to stand alone and hence not be dependent on any wider philosophical theses that might be thought contentious. For my complete response to the problem of radical scepticism, which locates epistemological disjunctivism as part of a larger ‘biscopic’ treatment of the problem that incorporates Wittgensteinian insights, see Pritchard (2015b). See also Pritchard (2016a). 2. As indicated in note 1, I offer a detailed account of how epistemological disjunctivism fits within a wider response to radical scepticism in Pritchard (2015b). 3. For the key works developing anti-luck epistemology, see Pritchard (2004, 2005, 2007a, 2008b, 2012c, 2015a). For the key works developing anti-luck virtue epistemology, see Pritchard (2009a, 2009b, 2009c, 2012a, 2016d, 2017), Pritchard, Millar, and Haddock (2010: chs. 1–4). Note that in recent work I have argued that our focus should be not on epistemic luck but rather on the closely related notion of epistemic risk, and hence that we should develop an anti-risk epistemology (and thus an anti-risk virtue epistemology). See Pritchard (2015d, 2016c, 2017b, forthcominga). Since the differences between epistemic luck and epistemic risk do not concern us here, I will set this complication to one side in what follows. See also note 9. 4. Which, of course, is not to say that there isn’t dissent (in philosophy there is nearly always dissent). See, for example, the exchange between Hetherington (2013) and Pritchard (2013) on this question. Note that even those who accept the general thesis about the incompatibility of knowledge and luck might still diverge about how they interpret this claim. For example, as I explain below (see note 8), some proponents of a robust form of virtue epistemology are willing to allow that knowledge can be lucky in certain, quite specific ways. 5. For a list of the key works defending anti-luck epistemology, and the notion of safety that goes with it, see note 3. For specific defence of the modal account of luck, see Pritchard (2014; cf., 2005: Ch. 5). For some of the key defences of safety, see Luper (1984, 2003), Sainsbury (1997), Sosa (1999), Williams (2000, passim). 6. See, for example, the ‘temp’ case that I offer and explore in Pritchard (2009b: chs. 3–4, 2012a, 2016c: chs. 3–4) and in Pritchard, Millar, and Haddock (2010: chs. 2–3). 7. For some of the key defences of robust virtue epistemology, see Sosa (1991, 2007, 2009, 2011, 2015), Zagzebski (1996, 1999), Greco (2003, 2007, 2009a, 2009b, 2012). Note that there are significant differences between these authors (none of which are relevant for our current concerns). Epistemic twin earth cases are essentially refinements of a previous critique that I offered of robust virtue epistemology, which is that while it can offer a principled basis for eliminating a certain form of knowledge-undermining veritic epistemic luck that I call intervening epistemic luck, it lacks the resources to eliminate a second kind of knowledge-undermining veritic epistemic luck that I call environmental epistemic luck. Epistemic twin earth cases essentially concern the kind of problem cases involving environmental epistemic luck. More generally, I claim that successes that are attributable to agency— i.e., achievements, at least according to some proponents of robust virtue epistemology—whether cognitive or otherwise, are compatible with environmental luck. See Pritchard (2009a, 2009b, 2009c, 2012a, 2016d, 2017); Pritchard, Millar, and Haddock (2010: chs. 1–4). Robust virtue epistemology

Anti-Luck Virtue Epistemology

8.

9.

10. 11.

12.

13.

14.

53

is to be contrasted with ‘modest’ (or ‘impure’) virtue epistemology which holds that an ability, or virtue, epistemic condition on knowledge is at most a necessary condition. As we will see, the anti-luck virtue epistemology that I defend is a version of modest virtue epistemology. I present and explore this distinction between robust/pure and modest/impure virtue epistemology in a number of places—see, for example, Pritchard, Millar and Haddock (2010: chs. 1–4), Pritchard (2012a, 2016d). Some robust virtue epistemologists explicitly acknowledge the potential modal fragility of our manifestations of agency, including cognitive agency, most notably Sosa (2007: ch. 5). Rather than taking this to show that there is more to knowledge than virtuous true belief, they instead argue that knowledge can sometimes involve lucky true belief. I discuss Sosa’s view in particular on this front in Pritchard (2009a). See also Pritchard (2012a), where I examine the more general issue of whether robust virtue epistemology should grant that there can be lucky knowledge. One of the advantages that I claim anti-risk epistemology (and thus anti-risk virtue epistemology) has over anti-luck epistemology (and thus anti-luck virtue epistemology)—see note 3 for more on this distinction—is that it is better placed to explain why robust virtue epistemology would be unwise to allow that there can be lucky knowledge, since they would be committed to ascribing knowledge to subjects even when there is a high epistemic risk of error. See Pritchard (2016c, 2017b, forthcominga) for further discussion of this point. For further discussion of such cases in this regard, see Pritchard, Millar and Haddock (2010: chs. 2–3), Pritchard (2012a, 2016d: chs. 3–4). Elsewhere I have characterised these cases in terms of positive epistemic dependence (where a subject knows even despite her cognitive agency playing a relatively low explanatory role in her safe cognitive success), in contrast to the negative epistemic dependence at issue in epistemic twin earth cases (where the subject’s cognitive agency plays an overarching explanatory role in her safe cognitive success and yet she still lacks knowledge). See Pritchard (2016b) for further discussion. For the key text where I defend epistemological disjunctivism, see Pritchard (2012b). See also Neta and Pritchard (2007); Pritchard (2007b, 2008a, 2011a, 2011b, 2015b, 2015c, 2016e, 2018, forthcomingb). This is the upshot of the so-called new evil genius intuition, which is rejected by epistemological disjunctivism. The loci classici in this regard are Lehrer and Cohen (1983) and Cohen (1984). For a helpful general discussion of the new evil demon intuition and its epistemological significance, see Littlejohn (2009). See also note 12. As I explain in Pritchard (2011b), epistemological disjunctivism represents a non-classical form of epistemic internalism, in that while it embraces the characteristic internalist thesis of accessibilism, and is at least consistent with another characteristic internalist thesis (mentalism), it rejects a third characteristic internalist thesis (the new evil demon intuition). See also Neta and Pritchard (2007) and note 11. My presentation of epistemological disjunctivism is inspired by the account of knowledge offered by McDowell (e.g., 1995). Although the matter is not entirely clear, there are passages that suggest that McDowell may well hold that something like epistemological disjunctivism (not his terminology but mine) is the right way to think about knowledge in general; he certainly holds that the view has application beyond perceptual knowledge. See, for example, McDowell (1994). For further discussion of the good+ case, along with a taxonomy of different types of ‘good’ and ‘bad’ cases, see Pritchard (2011a, 2011b: part 1).

54

Duncan Pritchard

15. Some complex issues are raised by the basing relation as it applies to epistemological disjunctivism, though it would take us too far afield to explore them here. For discussion, see Pritchard (forthcomingb). 16. See, for example, Greco’s (1999) seminal presentation of virtue reliabilism, including how it departs from, and thereby improves on, earlier process reliabilist proposals. 17. That all virtue-theoretic accounts of knowledge are devoted to capturing a substantive notion of cognitive responsibility gets obscured by the fact that proponents of epistemic internalist forms of virtue epistemology are often characterised as virtue reponsibilists, in contrast to the virtue reliabilism defended by epistemic externalist virtue epistemologists. But this categorisation in fact just signals that a particularly robust form of cognitive responsibility is being defended. The issue is further complicated in the literature by the fact that many virtue responsibilists explicitly eschew the project of defining knowledge altogether. For some of the key defences of virtue responsibilism, see Code (1987), Montmarquet (1993), Hookway (2003), Zagzebski (1996), and Roberts and Wood (2007). For more on the distinction between virtue reliabilism and virtue responsibilism itself, see Axtell (1997). 18. It is important to note that while factive reasons might be in this sense ‘unearnt’ according to epistemological disjunctivism, this doesn’t mean that the epistemic standing they provide is a form of epistemic entitlement, at least as that notion is usually understood, anyway—e.g., by Burge (e.g., 1993, 2003) and Wright (e.g., 2004). In particular, the factive epistemic reasons are meant to be genuine positive reasons (indeed, entailing reasons) for believing the target proposition. 19. I explore these issues in detail in Pritchard (2015b, passim). In particular, I argue there that we need to treat the contemporary version of Cartesian radical scepticism as in effect two distinct (but overlapping) sceptical problems, one that turns on the closure principle and another that trades on an underdetermination principle. Epistemological disjunctivism provides us with the resources to effectively deal with the latter. (But not the former. For that, one needs to combine epistemological disjunctivism with a Wittgensteinian hinge epistemology, a combination which I claim offers us a complete, ‘biscopic’, treatment of this kind of radical scepticism.) 20. I am grateful to Casey Doyle and Joe Milburn for comments on a previous version of this chapter.

References Axtell, G., 1997, Recent work in virtue epistemology, American Philosophical Quarterly 34: 410–430. Burge, T., 1993, Content preservation, Philosophical Review 102: 457–488. ———, 2003, Perceptual entitlement, Philosophy and Phenomenological Research 67: 503–548. Code, L., 1987, Epistemic Responsibility. Hanover, NH: University Press of New England. Cohen, S., 1984, Justification and truth, Philosophical Studies 46: 279–296. Greco, J., 1999, Agent reliabilism, Philosophical Perspectives 13: 273–296. ———, 2003, Knowledge as credit for true belief, in M. DePaul and L. Zagzebski, eds. Intellectual Virtue: Perspectives From Ethics and Epistemology. Oxford: Oxford University Press, pp. 111–134. ———, 2007, The nature of ability and the purpose of knowledge, Philosophical Issues 17: 57–69.

Anti-Luck Virtue Epistemology

55

———, 2009a, Achieving Knowledge. Cambridge: Cambridge University Press. ———, 2009b, Knowledge and success from ability, Philosophical Studies 142: 17–26. ———, 2012, A (different) virtue epistemology, Philosophy and Phenomenological Research 85: 1–26. Hetherington, S., 2013, There can be lucky knowledge, in M. Steup and J. Turri, eds. Contemporary Debates in Epistemology (2nd ed.). Oxford: Blackwell, §7. Hookway, C., 2003, How to be a virtue epistemologist, in M. DePaul and L. Zagzebski, eds. Intellectual Virtue: Perspectives From Ethics and Epistemology. Oxford: Oxford University Press, pp. 183–202. Kallestrup, J., and Pritchard, D.H., 2014, Virtue epistemology and epistemic twin earth, European Journal of Philosophy 22: 335–357. Lehrer, K., and Cohen, S., 1983, Justification, truth, and coherence, Synthese 55: 191–207. Littlejohn, C., 2009, The new evil demon problem, in B. Dowden and J. Fieser, eds. Internet Encyclopaedia of Philosophy. www.iep.utm.edu/evil-new/. Luper, S., 1984, The epistemic predicament, Australasian Journal of Philosophy 62: 26–50. ———, 2003, Indiscernability skepticism, in S. Luper, ed. The Skeptics: Contemporary Essays. Aldershot: Ashgate, pp. 183–202. McDowell, J., 1994, Knowledge by hearsay, in B.K. Matilal and A. Chakrabarti, eds. Knowing From Words: Western and Indian Philosophical Analysis of Understanding and Testimony. Dordrecht, Holland: Kluwer, pp. 195–224. ———, 1995, Knowledge and the internal, Philosophy and Phenomenological Research 55: 877–893. Montmarquet, J., 1993, Epistemic Virtue and Doxastic Responsibility. Lanham, MD: Rowman & Littlefield. Neta, R. and Pritchard, D.H., 2007, McDowell and the new evil genius, Philosophy and Phenomenological Research 74: 381–396. Pritchard, D.H., 2004, Epistemic luck, Journal of Philosophical Research 29: 193–222. ———, 2005, Epistemic Luck. Oxford: Oxford University Press. ———, 2007a, Anti-luck epistemology, Synthese 158: 277–297. ———, 2007b, How to be a neo-Moorean, in S. Goldberg, ed. Internalism and Externalism in Semantics and Epistemology. Oxford: Oxford University Press, pp. 68–99. ———, 2008a, McDowellian neo-Mooreanism, in Adrian Haddock and Fiona Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 283–310. ———, 2008b, Sensitivity, safety, and anti-luck epistemology, in J. Greco, ed. Oxford Handbook of Scepticism. Oxford: Oxford University Press, pp. 437–455. ———, 2009a, Apt performance and epistemic value, Philosophical Studies 143: 407–416. ———, 2009b, Knowledge. London: Palgrave Macmillan. ———, 2009c, Knowledge, understanding and epistemic value, in A. O’Hear, ed. Epistemology (Royal Institute of Philosophy Lectures). Cambridge: Cambridge University Press, pp. 19–43. ———, 2011a, Epistemological disjunctivism and the basis problem, Philosophical Issues 21: 434–455.

56

Duncan Pritchard

———, 2011b, Evidentialism, internalism, disjunctivism, in T. Dougherty, ed. Evidentialism and its discontents. Oxford: Oxford University Press, pp. 362–392. ———, 2012a, Anti-luck virtue epistemology, Journal of Philosophy 109: 247–279. ———, 2012b, Epistemological Disjunctivism. Oxford: Oxford University Press. ———, 2012c, In defence of modest anti-luck epistemology, in T. Black and K. Becker, eds. The Sensitivity Principle in Epistemology. Cambridge: Cambridge University Press, pp. 173–192. ———, 2013, There cannot be lucky knowledge, in M. Steup, J. Turri, and E. Sosa, eds. Contemporary Debates in Epistemology (2nd ed.). Oxford: Blackwell, pp. 152–164. ———, 2014, The modal account of luck, Metaphilosophy 45: 594–619. ———, 2015a, Anti-luck epistemology and the gettier problem, Philosophical Studies 172: 93–111. ———, 2015b, Epistemic Angst: Radical Skepticism and the Groundlessness of Our Believing. Princeton, NJ: Princeton University Press. ———, 2015c, Epistemological disjunctivism: Responses to my critics, Analysis 75: 627–637. ———, 2015d, Risk, Metaphilosophy 46: 436–461. ———, 2016a, Epistemic Angst, Philosophy and Phenomenological Research (Online First). doi:10.1111/phpr.12280. ———, 2016b, Epistemic dependence, Philosophical Perspectives 30: 1–20. ———, 2016c, Epistemic risk, Journal of Philosophy 113: 550–571. ———, 2016d, Epistemology. London: Palgrave Macmillan. ———, 2016e, Epistemological disjunctivism: Responses to my critics, Journal of Philosophical Research 41: 221–238. ———, 2017a, Knowledge, luck and virtue: Resolving the gettier problem, in C. Almeida, P. Klein, and R. Borges, eds. Explaining Knowledge: New Essays on the Gettier Problem. Oxford: Oxford University Press, pp. 57–73. ———, 2017b, Anti-Risk Epistemology and Negative Epistemic Dependence, Synthese (Online First). doi:10.1007/s11229-017-1586-6 ———, 2018, Epistemological disjunctivism and the biscopic treatment of radical scepticism, in V. Mitova, ed. The Factive Turn in Epistemology. Cambridge: Cambridge University Press, pp. 15–31. ———, forthcominga, Anti-risk virtue epistemology, in J. Greco and C. Kelp, eds. Virtue Epistemology. Cambridge: Cambridge University Press. ———, forthcomingb, Epistemological disjunctivism and factive bases for belief, in P. Bondy and J.A. Carter, eds. Well-Founded Belief: New Essays on the Epistemic Basing Relation. London: Routledge. Pritchard, D.H., Millar, A., and Haddock, A., 2010, The Nature and Value of Knowledge: Three Investigations. Oxford: Oxford University Press. Roberts, R. and Wood, W.J., 2007, Intellectual Virtues: An Essay in Regulative Epistemology. Oxford: Oxford University Press. Sainsbury, R.M., 1997, Easy possibilities, Philosophy and Phenomenological Research 57: 907–919. Sosa, E., 1991, Knowledge in Perspective: Selected Essays in Epistemology. Cambridge: Cambridge University Press. ———, 1999, How to defeat opposition to Moore, Philosophical Perspectives 13: 141–154.

Anti-Luck Virtue Epistemology

57

———, 2007, A Virtue Epistemology: Apt Belief and Reflective Knowledge. Oxford: Oxford University Press. ———, 2009, Reflective Knowledge: Apt Belief and Reflective Knowledge. Oxford: Oxford University Press. ———, 2011, Knowing Full Well. Princeton, NJ: Princeton University Press. ———, 2015, Judgment and Agency. Oxford: Oxford University Press. Williamson, T., 2000, Knowledge and Its Limits. Oxford: Oxford University Press. Wright, C.J.G., 2004, Warrant for nothing (and Foundations for Free)? Proceedings of the Aristotelian Society (supp. vol.) 78: 167–212. Zagzebski, L., 1996, Virtues of the Mind: An Inquiry Into the Nature of Virtue and the Ethical Foundations of Knowledge. Cambridge: Cambridge University Press. ———, 1999, What is knowledge? in J. Greco and E. Sosa, eds. Blackwell Guide to Epistemology. Oxford: Blackwell, pp. 92–116.

Part II

Historical Antecedents

4

Ancient Philosophy and Disjunctivism The Case of the Stoics Iakovos Vasiliou

1. Introduction Disjunctivism and ancient philosophy might seem a rather random pairing: why connect a quite recent cluster of positions in the philosophy of perception with philosophical theories about perceptual experience and the epistemology of perception that are thousands of years old? A couple of big-picture considerations may help to motivate the project. Some notable disjunctivists, such as John McDowell, attribute the formation of the “traditional” philosophical picture against which many of their arguments are aimed to a particular historical time period around Descartes and the rise of modern philosophy. On such a view, “traditional” problems in epistemology, as well as in other areas of philosophy, stem from a particular conception of mind and world and the relationship between them. The mind is an immediately and transparently accessible space within which things stand in rational relations to one another, while the world is accessible to the mind only indirectly, via inference or some such, if at all. So, modern philosophy becomes obsessed with a series of philosophical problems aimed at bridging a “gap” between mind and world. An alternative approach, on this account, would avoid such dualisms by showing that what seemed to be mandatory and compelling philosophical problems are in fact misguided questions asked because of misguided, but fortunately optional, assumptions. The point here is only to foreground the idea that if there is something to this sort of critical philosophical project—to which disparate philosophers such as Wittgenstein, Sellars, Davidson, Rorty, and McDowell have arguably signed on—then Greek philosophy would occupy a particularly interesting philosophical position. Since it obviously takes place before the alleged significant shift in philosophical outlook, one might suppose a priori that it would be a place to look for philosophical positions that were innocent of certain offending modernist assumptions. So, if it were true, for example, that disjunctivism makes a sort of naïve realism and lack of concern with certain skeptical problems philosophically defensible, then this might change our attitude toward Plato’s and

62

Iakovos Vasiliou

Aristotle’s remarks about perceptual knowledge.1 This would be an interesting but relatively weak use of disjunctivism. I mean simply that it is unsurprising that the correct view about how to understand certain questions in epistemology would make different sense of certain historical texts in epistemology. What would be more substantive would be if there were textual evidence that Greek philosophers or schools held disjunctivism itself and not simply views that might be well supported by adherence to it. There might seem to be an immediate problem, however, with attempting a more substantive connection between Greek philosophy and disjunctivism, which is where my brief mention of a sweeping historical picture comes into play. If the Greeks were genuinely innocent of the “traditional” but, it is argued, distinctively modern philosophical framework, against which disjunctivism is a reaction, then their relationship toward disjunctivism must be fundamentally different. For regardless of whether one adopts the overarching historical thesis according to which a fundamental shift of philosophical framework occurs sometime after the ancient period,2 there is no doubt that contemporary disjunctivism, especially epistemological disjunctivism, is a minority position developed explicitly against and in reaction to what the disjunctivist sees as stalemate between two sorts of positions in epistemology, internalist and externalist, neither of which provides a satisfactory, non-skeptical account of perceptual knowledge on its own, but which also arguably cannot be harmoniously coupled. It is this dissatisfaction with “traditional” epistemology and, according to the disjunctivist, its “dreary history,”3 that generates the motivation to seek an alternative. This leaves us in an odd position relative to Greek philosophy, since it is obviously not in a position to be reacting against a framework that had yet to be invented. In order to carry out a more ambitious agenda with respect to the relationship between disjunctivism and Greek philosophy, I will proceed by hypothesis. I will assume that in fact Greek philosophers are innocent of the traditional framework—the key components of which for our purposes I detail in this chapter—and then attempt to interpret the Stoics, in particular, on the assumption that disjunctivism is the default view. This will be difficult in part because for most of us epistemological disjunctivism is a difficult and subtle position and many philosophers are suspicious even of its cogency. The effect of this is that when one reads, “it appears to S that p,” we automatically default to our “traditional” interpretation rather than understanding it as a disjunctivist would. I will read our texts from the perspective of a disjunctivist, but of a disjunctivist, unlike our contemporary exponents, for whom disjunctivism is the natural default position and not a reaction to or rejection of something.4 I hope to show that this approach will generate numerous and substantive parallels between the Greeks, in particular the Stoics, and contemporary proponents of epistemological disjunctivism, in particular John McDowell.

Ancient Philosophy and Disjunctivism

63

2. “Traditional” Epistemology and Its Disjunctivist Critics While people ordinarily take themselves to be knowers in virtue of episodes of visual perception all the time—that is, take themselves to know that p because they see that p—such a position is manifestly not philosophical common sense.5 Philosophical common sense says that any and every time one believes that p on the basis of merely seeing that p, one does not have sufficient entitlement or warrant6 to know that p. The Argument from Illusion teaches us, allegedly, that we are not epistemically entitled to claims about how the world is from the way it looks. The idea that what is experientially available to us is assumed to be the same with respect to epistemic warrant in the veridical and non-veridical cases is the heart of the idea of the “Highest Common Factor (HCF),” which McDowell describes as follows: [T]he argument [from illusion] is that since there can be deceptive cases experientially indistinguishable from non-deceptive cases, one’s experiential intake—what one embraces within the scope of one’s consciousness—must be the same in both kinds of case. In a deceptive case, one’s experiential intake must ex hypothesi fall short of the fact itself, in the sense of being consistent with there being no such fact. So that must be true, according to the argument, in a nondeceptive case too. One’s capacity is a capacity to tell by looking: that is, on the basis of experiential intake. And even when this capacity does yield knowledge, we have to conceive the basis as a highest common factor of what is available to experience in the deceptive and non-deceptive cases alike.7 What is common to all cases of putative seeing, veridical and not veridical, is a seeming to see, which delivers exactly the same epistemic warrant in either case. Thus one is stuck with a mental state, a seeming to see, that is of the same epistemic significance whether or not it happens to be a case of actual seeing. This is supposed to follow from the premise that for any actual veridical episode of seeing, there could be a phenomenologically indistinguishable episode of seeming to see (e.g., a hallucination) instead. So the subject is never in a position to grasp whether or not she is actually seeing. All she can do is hope that the world is doing her a favor and not “playing her false.”8 The upshot—and the conundrum of traditional epistemology—is that since whether or not the world is playing us false is something that can never be reflectively accessible to us, we are forever cut off with how things seem to us, and the best we can do within this “space” is to reflect on these internal goings-on and, so to speak, clean them up rationally by rendering them, for example, coherent with one another. But no matter how coherent we make them, we will never be able to jump the gap and be able to know how the world is merely from how it looks to us.

64

Iakovos Vasiliou

McDowell further describes the effect of the Argument from Illusion as a retreat. One retreats from claims about how the world is to how the world looks or seems: Consider the Argument from Illusion. Seeing, or perhaps having seen, that things are thus and so would be an epistemically satisfactory standing in the space of reasons. But when I see that things are thus and so, I take it that things are thus and so on the basis of having it look to me as if things are thus and so. And it can look to me as if things are thus and so when they are not; appearances do not give me the resources to ensure that I take things to be thus and so, on the basis of appearances, only when things are thus and so. If things are indeed thus and so when they seem to be, the world is doing me a favour. So if I want to restrict myself to standings in the space of reasons whose flawlessness I can ensure without external help, I must go no further than taking it that it looks to me as if things are thus and so. One might hope that this inward retreat is only temporary.9 Once the Argument from Illusion is in place, the world has to do me a favor for p to be the case; that p is the case is not something that is up to me to establish for myself. At this point an externalist-type thought enters the account: there has to be, for example, a causal relationship of the right sort between the fact that p and my seeing that p. But that connection from p’s seeming to be the case to p’s being the case is one that is not accessible to the subject. If one wishes to avoid skepticism, even the most optimistic views (what McDowell calls “hybrid” views) leave one with what is at best a “likely”10 story: Now in the hybrid conception of knowledge, it is admittedly not a complete accident, relative to someone’s standing in the space of reasons, if things are as she takes them to be; the position of her belief in the space of reasons makes it likely to be true. But the reason why the extra stipulation that the belief is true—that is distinctive of the hybrid approach—is needed is that likelihood of the truth is the best that the space of reasons yields, on the interiorized conception of it: the closest we can come to factiveness.11 The “world doing me a favor” language means that, with the Argument from Illusion in place, I can’t ever know simple empirical truths by myself. This is the picture the disjunctivist reacts against. Epistemological disjunctivism claims that our grounds for knowledge can be both reflectively accessible to the subject and also factive.12 It argues that there is a difference in epistemological warrant when it looks to S as though p in the case where p obtains and in the case where it does not. In a veridical case, the epistemological disjunctivist holds that seeing that p is (sufficient,

Ancient Philosophy and Disjunctivism

65

infallible, and conclusive) grounds for knowing that p, despite the fact that seeming to see p, which may be phenomenologically indistinguishable from seeing p and may share the same representational content as seeing p, is not.13 This combination of seemingly incompatible positions underlies the resistance to disjunctivism in contemporary epistemology,14 and, I shall argue here, drives certain Academic objections against the Stoic position.

3. Stoic Epistemology of Perception Gisela Striker describes the shift from the fourth to the third century BC as one in which the central question of epistemology shifted from “What is knowledge?” to “Is there any knowledge?”15 Plato and Aristotle are relatively unconcerned, if perhaps for different reasons, about the veridicality of perception; they share a standing confidence in the veridicality and reliability of perception. When opponents, such as Protagorean relativists, raise puzzles about conflicting appearances, seeking to conclude that the truth about how things are must be relativized correspondingly, Plato and Aristotle respond forcefully but also somewhat dismissively. Things are as they appear to a person in good condition. It is not each individual that is the measure of how things are, as the Protagoreans would have it, but the one in the proper condition. If you want to know what color things are, you don’t ask someone with jaundice. If you want to know whether the wine is off, you don’t ask someone who is ill.16 In a memorable passage, Aristotle remarks that if one wants to know whether snow is white, one needs perception, not argument (Topics I.11, 105a2–7). Conflicting appearances should not destabilize the epistemic value of ordinary perceptual episodes of normal perceivers under normal conditions.17 Such a position fits well with disjunctivism insofar as cases of misperceiving can be taken in stride, without their threatening to cut a perceiver off from being able to experience genuine episodes of perceiving the world in veridical cases. But, as I remarked, there is no endorsement of a disjunctivist thesis as such.18 The Epicureans preserve the absolute veridicality of the senses, encapsulated in their much-quoted and much-mocked slogan, “all sensations are true,” but run the risk of undermining what typically motivates philosophers to defend the reliability of the senses: the desire that the deliverances of sense perception may be able to serve as a foundation for knowledge. While all of the deliverances of sense perception are perfectly accurate, they are accurate only about how things appear. According to Long and Sedley’s analysis, “all sensations are true” is best understood as analogous to the idea that the camera does not lie: it accurately records the light waves that are striking it. Seeing a person from very far away as tiny, indistinct, and unclear is in fact accurate because that is exactly how a person looks from such a distance. Long and Sedley argue, however,

66

Iakovos Vasiliou

that this does not mean sensation is useless for gaining knowledge about the external world, for it can count as reliable evidence for what the world is like, despite obviously not being an infallible guide.19 Sensory evidence, used properly by reason, can help to make true judgments; one would not expect, for example, a square tower to look the same close up as it does at a distance (LS 16G=Lucretius, 4.353–363). As with Plato and Aristotle, Epicurus’s views are consistent with a disjunctivist thesis, but also perhaps with other contemporary views.20 I hope to show that the Stoics, however, are the most interesting case to consider with respect to disjunctivism. Central to Stoic epistemology is the idea of the cataleptic appearance, which they suppose to be the “criterion” of truth.21 An appearance (phantasia), for a mature human being, is a mental state as of something’s being the case; for example, that an apple is on the table.22 When a subject assents to an appearance, then they are in a certain epistemic state. Which epistemic state a person is in depends on who they are (e.g., a Sage or not) and to what sort of appearance they have assented. A cataleptic appearance (katalēptikē phantasia) is an appearance that is, somehow, guaranteed to be correct and is thus supposed to be the criterion of truth.23 Assent to cataleptic appearances yields what in contemporary English could be called “knowledge,” i.e. katalēpsis.24 Typically, however, katalēpsis is translated “cognition” or “apprehension” to distinguish it from epistēmē, which is often translated “knowledge” but is more like systematic scientific knowledge, particularly in Stoicism. The Stoics believe that only the Sage possesses epistēmē, and that Sages rarely exist (perhaps one every couple of centuries). The Stoic Sage is an ideal person (in all respects) who, as we will discuss a bit further, assents exclusively to cataleptic appearances. Nevertheless, the rest of us also assent all the time to cataleptic appearances, although we precipitously assent to non-cataleptic appearances as well.25 We should also note, although it will not be of central importance for us, that assent to appearances occurs very frequently in everyone’s lives. According to the Stoics, when one acts in a certain way, the action itself necessarily implies or perhaps even constitutes an assent to an appearance. So when you reach for your cup of coffee, you are assenting to the appearance that this is a cup of coffee. While I am not interested here in philosophy of action, what is important about this for us is to note the frequency and ubiquity of assent, which we might keep in mind in connection with disjunctivism’s claim to be a vindication of common sense. Appearances are matters of objects and states of affairs impressing themselves in a physical way on our senses, and via our senses to our soul (LS 53G, H). This is one reason why many have translated phantasiai as “impressions.” This translation captures well the empirical quality of phantasiai and, what’s more, perhaps helpfully has the Stoic account resonate with modern empiricist accounts like Hume’s. The problem with “impression,” however, is that it suggests in English at least some level

Ancient Philosophy and Disjunctivism

67

of belief, as in “I have the impression that . . .” But as we have seen, one can have an impression without believing it at all because one refuses to assent to it; so an impression or appearance is a mere seeming to be thus and so, with no commitment as to the appearance’s truth implied on the part of the perceiver. Appearances, when true, are direct, unmediated takings in of states of affairs in the world. They are not images or pictures; we do not perceive images of objects but objects (and states of affairs).26 Appearances, when true, are a direct stamping of a state of affairs onto the soul. So, in more contemporary parlance, the mind reaches right out to the world and does not stop short of the fact. So, as with disjunctivism, we seem to have a direct realist picture of veridical perception. This much alone, however, makes the connection between Stoic epistemology and disjunctivism rather weak, in the sense that I described earlier. The Stoics, like Aristotle, seem to have a naïve realist conception of perception, and in fact have perhaps a more plausible physicalist account than Aristotle, for whom perceiving consists, somehow, in the sensible form being received by the perceiver without the matter (see, e.g., De Anima Ι.12). To this extent, disjunctivism, insofar as it is correct, would fit nicely with such a view. But what I am interested in is whether a more substantive connection between disjunctivism and the Stoic position can be established. For this, we should begin by examining the Stoic account of appearances in greater detail.

4. Stoic Appearances In a couple of passages from doxographical accounts of Stoicism in Diogenes Laertius and Aëtius, we find distinctions that are quite striking, if we consider them with disjunctivism in mind. Aëtius is the most interesting text for our purposes (LS 39B=Aëtius, 4.12.1–6; Inwood and Gerson (2008), trans., modified): [1.] Chrysippus says that these four things [i.e., appearance (phantasia), appearing object (phantaston), the phantastic (phantastikon), phantasm (phantasma)] differ from each other [Χρύσιππος διαφέρειν ἀλλήλων φησὶ τέτταρα ταῦτα]. Appearance [φαντασία], then, is an experience [πάθος] which occurs in the soul and which, in [the experience] itself, also indicates what caused it. For example, when we observe something white by means of vision, there is an experience that has occurred in the soul by means of vision; and this experience we are able to say that there exists something white which stimulates us. And similarly for touch and smell. [2.] “Appearance” [φαντασία] gets its name from “light” [φῶς]; for just as light reveals itself and the other things which are encompassed in it, so too appearance reveals itself and that which caused it. [3.] The appearing

68

Iakovos Vasiliou object [φανταστόν] is that which causes the appearance. For example, the appearing object is the white and the cold and everything that is able to stimulate [κινεῖν] the soul. [4.] The phantastic [φανταστικόν] is a groundless attraction [διάκενος ἑλκυσμός], an experience in the soul [πάθος ἐν τῇ ψυχῇ] that occurs as the result of no appearing object (phantaston), as in the case of people who fight with shadows and punch at thin air. For some appearing object underlies appearance, but not the phantastic. [5.] A phantasm [φάντασμα] is that to which we are attracted in the phantastic groundless attraction. This occurs in the melancholic and in madmen; at any rate when Orestes in the tragedy says . . .27 he says this like a madman, and sees nothing [ὁρᾷ δὲ οὐδέν], but only thinks [that he does] [ἀλλὰ δοκεῖ μόνον].

What we get here, corroborated in certain details by Diogenes/Diocles,28 is a division of two species of mental states, falling under the genus pathos. Each of these mental states has a different cause and object. So appearance, one sort of pathos or affection of the soul, has as its object and cause an “appearing object (phantaston),” while “the phantastic (phantastikon),” is a different kind of pathos that has as its object and cause a phantasm (phantasma). Although it is not explicit, it seems clear that appearances here are restricted to true appearances—or more apt to these descriptions, appearances are of the real, fantasies are of mere phantasms.29 The genus then is pathē of the soul, which include (true) phantasiai and (false) phantastika as species. Chrysippus clearly labels the different states: appearances from real objects are those that yield a distinctive mental state: a phantasia (appearance). Appearances are revelatory like a light, showing what is the case: Turning on a light (lighting a torch) simply reveals (visually speaking) what is there. This passage, then, is straightforwardly readable as a statement of disjunctivism.30 A phantasia generated by a phantaston is a different mental event (pathos in the soul) than a phatastikon generated by phantasma. Correspondingly, it seems plausible to understand Aëtius’s (and Diogenes’s) account(s) as implying that Chrysippus must therefore reject a conception of experience according to which the experience one has in veridical and non-veridical cases are the same, the difference lying only in the presence or absence of the object/state of affairs. A mere appearance, as we would say, is defined as a different pathos of the soul: It has a different name—a phantastikon—and is a “thoroughly empty attraction [διάκενος ἑλκυσμός],” and it is generated by a phantasma, which, in the example from Orestes and the gloss by Aëtius, is a nothing: a mere appearance (conceived of from the perspective of the object). So we have here more substantive evidence for attributing a disjunctivist view to the Stoics. If our overall historical hypothesis is correct, of course, they are not presenting a disjunctivist position as a reaction to an HCF conception that they are consciously rejecting. Rather, we should,

Ancient Philosophy and Disjunctivism

69

as a working hypothesis, suppose that an HCF conception is yet to be formulated. It is not explicit in this text that these different kinds of mental events will have different epistemic significance, but that should be obvious. Since appearance goes with a taking in of an actual fact or object, it will be true. A true appearance, of course, will have a very different epistemic significance from a false one. So the implicit rejection here of an HCF conception of experience (i.e., experiential disjunctivism) will support and go hand in hand with epistemological disjunctivism. It is only in these two passages that the distinction between a true and false “appearance” is labeled with different terms, thereby making a disjunctivist reading quite straightfoward. Other testimony is more ambiguous, although still clearly interpretable in disjunctivist terms. In a wellknown, if problematic, passage on the Stoic conception of appearance in Sextus (M. 7.242–252),31 Sextus proceeds by dividing up appearances according to various differentiae: the persuasive and unpersuasive, the true and false, the cataleptic and non-cataleptic.32 This way of dividing up appearances is more easily read along HCF lines. For in Sextus, the divisions and distinctions all take place within the class of appearances. So, if we divide impressions into the true and false (as done in LS39G), we naturally understand that not as distinguishing between two classes of distinct mental events (or two distinct types of epistemic warrant), but one kind of event—an impression—that is on some occasions true and on other occasions false. Of course, all parties agree that there is a big difference between true and false impressions, but a “traditional” reading might interpret this as though there is one impression, common to true and false cases alike, which either corresponds or does not. On a disjunctivist reading, by contrast, there are some impressions that are direct revelations of reality, and others that are mere figments. While this comes out more clearly in Aëtius and Diocles (Diogenes), if we read the subsequent division in Sextus in light of the immediately preceding passage, the disjunctivist interpretation is more tempting: 241. Since appearance comes about either from external things or from effects in us (and this is more properly [κυριώτερον] called by them “thoroughly empty attraction [διάκενος ἑλκυσμός]”), in the account of appearance there is absolutely implied at the same time the fact that the process of being affected occurs either by way of impact from outside [κατὰ τὴν ἐκτὸς προσβολὴν] or by way of experiences (pathē) in us. . . . Thus it is hard to give an account of appearance (phantasia) as it figures in Stoicism. Now among appearances there are many additional differences; [242], however, the ones about to be mentioned will be sufficient. [Start LS 39G] Of appearances, some are persuasive, some unpersuasive. (M 7.241, Bett (2005) trans.)

70

Iakovos Vasiliou

This text contains the only other mention of διάκενος ἑλκυσμός (“thoroughly empty attraction”). As we have seen, in the Aëtius passage, this is what gets the label phantastikon in explicit distinction from a phantasia. Sextus associates this phrase with what is “properly called by them [the Stoics]”; but this aside perhaps suggests that he is not bothering with most of their “proper” or “strict” terminology, and so none of the associated technical distinctions that we saw in Aëtius or Diogenes, especially between a phantasia and a phantastikon, is maintained. This is not particularly surprising since in the context the view is being presented and simultaneously criticized (cf. 7.232–240). Nevertheless, the important point is that nothing in the Sextus passage rules out its being construed consistently with a disjunctivist reading. The central focus in our comparison between Stoic epistemology and contemporary epistemological disjunctivism, however, will be the cataleptic appearance.

5. The Cataleptic Appearance It is clear that the Stoics want the cataleptic appearance to be able to function as a foundation in their epistemology. The cataleptic appearance is a mental episode that would constitute “the holy grail” in epistemology (as epistemological disjunctivism has recently been described by Duncan Pritchard)33, insofar as it would seem, at first glance, to be a perceptual appearance that is (a) reflectively accessible and (b) factive. As the criterion of truth it would be the “means by which we acquire immediate, non-inferential knowledge of a perceptible state of affairs.”34 The canonical definition of the cataleptic appearance contains three parts:35 (a) that which arises from what is (ἀπὸ ὑπάρχοντος); (b) stamped and impressed exactly in accordance with what is (κατ᾽αὐτὸ τὸ ὑπάρχον ἐναπεσφραγισμένην καὶ ἐναπομεμαγμένην); (c) of such a kind as could not arise from what is not (ὁποία/ὅια οὐκ ἂν γένοιτο ἀπὸ μὴ ὑπάρχοντος). Part (a) demands that the appearance be derived from what exists or, on some interpretations such as Cicero’s (Acad 2.42, 112), from what is true. The latter interpretation has certain advantages, insofar as it allows us to read part (c) more easily as making a distinct claim. Focused as we are on adult human perception, the subsequent discussion will not hinge on whether we consider “I see a green thing” or “I see that this is green,” “I see Socrates” or “I see that this is Socrates.” As I mentioned earlier, for the Stoics, adult human beings’ perceptions have conceptual content. Part (b) claims that it must be stamped precisely and clearly on the sense organs; that is, the appearance must be clear and distinct (cf.

Ancient Philosophy and Disjunctivism

71

LS 40C).36 Part (c) claims, arguably, that the impression must be unique. On Cicero’s account (2.77–78=LS40D), part (c) was added by Zeno in response to a challenge to the cataleptic appearance raised by Arcesilaus: what if there were another appearance that were just like the putative cataleptic appearance, but which was false? Part (c) is meant to rule out this possibility. But how to interpret part (c) and its connection to parts (a) and (b) remains difficult and disputed. What we might call the “traditional” reading, stemming from Michael Frede,37 is an externalist one. If we understand the cataleptic appearance as guaranteeing the truth of its propositional content on an externalist interpretation, then it will be plausibly understood to be describing a certain causal relationship that obtains between p and the appearance that p. As one would expect of an externalist reading, this takes care of the factive element: the fact that S has a cataleptic appearance that p, entails p. But whether or not S is having a cataleptic appearance is not something reflectively accessible to S. As Gisela Striker puts it: It is important to realize, however, that the definition [of the cognitive impression] does not purport to tell us how we can find out whether a given impression is cognitive or not—it tells us only what sort of impressions can lead to cognition in the first place. The Stoic assumption that there must best be such impressions relies on the premise that knowledge is indeed possible, and that it must ultimately come from the senses. . . . How we can tell whether a given impression is cognitive is a different question—and, as it turns out, quite a difficult one.38 The cost of the externalist reading is clear since it seems to deprive us of the idea that the experience of a cataleptic appearance, as such, is within our cognitive access; thus, it deprives us of what one might certainly think was the aim of the cataleptic appearance, namely to provide the perceiver with an indefeasibly conclusive entitlement for claiming to know that p from the fact that she is aware that she is having a cataleptic appearance that p. For the Stoic position to be one in which a perceiver, for all she knows, has at most an appearance that is at best probably true, or plausibly true, is to concede the argument to their Academic skeptical opponents. The externalist interpretation turns to the Stoics’ metaphysical thesis of the identity of indiscernibles, applying it to appearances. According to the Stoics, there could be no two things that have all and only the same properties. This seems to be the role of part (c), at least, on Sextus’s interpretation: “of such a kind as could not arise from what is not” was added by the Stoics, since the Academics did not share their view of the

72

Iakovos Vasiliou impossibility of finding a totally indiscernible [but false] appearance. For the Stoics say that one who has a cataleptic appearance fastens on to the objective difference of things in a craftsmanlike way, since this kind of appearance has a peculiarity which differentiates it from other appearances, just as horned snakes are different from others. (Sextus, M. 7.252=LS40E; slightly modified)

There are craftsmanlike, precise aspects of a cataleptic appearance that distinguish it from other superficially similar, but non-cataleptic appearances.39 Of course, as quickly becomes obvious, this is helpful only if the metaphysical discernibility is itself perceptually discernible, so that one can discern one appearance from another. Nevertheless, part (c) is sometimes understood as suggesting that there is something phenomenologically accessible about the appearance that uniquely marks it as cataleptic and guarantees its truth. Unsurprisingly, this has led in the present century to a number of interpretations that attempt a hybrid interpretation of the cataleptic appearance, according to which the cataleptic appearance not only guarantees the truth of itself, but also can stand in some internal epistemic relation to the perceiver, so it can provide a conclusive reason for her belief that p.40 Perhaps, as in the earlier passage on some interpretations, the Stoics held out hope at certain points that there could be some perceptually discernible difference between cataleptic and non-cataleptic appearances, but that ends up being simply untenable. All of these more recent commentators agree that in the end, the Stoic position falls to the Academic objection that there can be a false appearance that is phenomenologically indistinguishable from a true one.41 At best, with care, attention, and practice, a perceiver might be able to ensure that what she perceives is likely or probable; but this is pretty clearly to give up the game to the Academics. What is needed to defend fully the Stoic position would be an epistemological disjunctivism that allowed that, under certain conditions, a perceiver could have an appearance that was both reflectively accessible and factive. McDowell makes such a case in his Perception as a Capacity for Knowledge, which, we will see, argues in ways that strongly support a Stoic position and parallels certain Stoic claims, while opposing a position that is very similar to that of their Academic opponents.

6. McDowell’s Perception as a Capacity for Knowledge and the Stoic/Academic Debate At this point I hope it is less surprising that scholarship on the cataleptic appearance has reflected in its exegetical options the same options we witness in contemporary epistemology. Settling on what McDowell described as a “hybrid” interpretation (section 2 in this chapter), the consensus has been that the Stoics’ Academic opponents fairly easily win the day. In these final sections, I will show how McDowell defends a position

Ancient Philosophy and Disjunctivism

73

very much like the Stoics. The Stoic/Academic debate is strikingly similar to how McDowell describes his own dispute with Tyler Burge in Perception as a Capacity for Knowledge.42 Burge’s thought is this: to argue that the perceptual state is a seeing—to argue that the warrant it provides is good enough for knowledge—one would need to argue that in the present circumstances the warranting force it has, as the perceptual state it is, is not undermined, though it is warrant of a kind that can be undermined, warrant of a sort that does not guarantee what it warrants. . . . If a subject brings the warrant a perceptual state gives her for believing something about how things are in her environment within the scope of her selfconsciousness, she becomes aware that she has something like evidence, in itself inconclusive, for the truth of the belief.43 As McDowell remarks:44 “Everything turns here on the assumption that the warrant a perceptual state provides for a belief cannot guarantee the truth of the belief.” This is what the Academic skeptics, like Burge, say when they deny the Stoic claim part (c) about the cataleptic appearance, namely that it is “of such a kind as could not arise from what is not” (see LS40H=M. 7.402). Carneades’s rejection of (c) is a rejection of the idea that the warrant a perceptual episode provides for a belief can guarantee its truth. McDowell offers a Sellarsian-inspired alternative: when all goes well in the operation of a perceptual capacity of a sort that belongs to its possessor’s rationality,45 a perceiver enjoys a perceptual state in which some feature of her environment is there for her, perceptually present to her rationally self-conscious awareness. This presence is an actualization of a capacity that belongs to the subject’s reason. . . . And if a perceptual state can consist in a subject’s having a feature of her environment perceptually present to her, that gives the lie to the assumption that a perceptual state cannot warrant a belief in a way that guarantees its truth. If a perceptual state makes a feature of the environment present to a perceiver’s rationally self-conscious awareness, there is no possibility, compatibly with someone’s being in that state, that things are not as the state would warrant her in believing that they are. . . . The warrant for belief that the state provides is indefeasible; it cannot be undermined. . . . When one sees something to be so, one is in a perceptual state in which its being so is visually there for one, so that one has a conclusive warrant for a corresponding belief.46 Here is a statement of a position that would cut through the internalist/ externalist readings of the Stoic cataleptic appearance. The cataleptic appearance, as a perceptual state under excellent conditions, is a taking

74

Iakovos Vasiliou

in of some feature of the environment and, crucially, a taking in that is part of the perceiver’s rational self-conscious awareness (and so satisfies our internalist intuitions) and yet at the same time guarantees the veridicality of the appearance. One’s warrant is conclusive. It is clear, I think, that this is what the Stoics want from the cataleptic appearance. Now, whether the Stoics are entitled to it, or McDowell is entitled to it, of course is another matter. But what should be apparent so far is the striking similarity between the Stoic view and the position McDowell stakes out and, as we will see, attempts to defend. McDowell describes his position as one in which “rational subjects can be, and can know themselves to be, in perceptual states that make features of the environment present to them, and so provide conclusive warrant for corresponding beliefs.”47 What is mistakenly, according to McDowell, understood to block the availability of such a position is the assumption that perceptual states can provide at best defeasible warrant for beliefs, which supposedly follows from the fact that any perceptual capacity is fallible. Since any perceptual warrant is compatible with being fooled, “perceptual warrant is in itself neutral as to whether a belief it warrants is true or not.”48 This is precisely the Academic objection in denying the cataleptic appearance: there is no perceptual appearance such that having that appearance supplies conclusive warrant for one’s belief, and so knowledge. They provide a parade of examples of phenomenologically indistinguishable perceptual states, as illustrations of the capacity’s fallibility: two grains of sand, twins, delusions of Orestes by the Furies, wax pomegranates, and so on (e.g., Cicero, Acad. 2.84–86; Sextus, M. 7.409–410; LS40F). What exactly the Stoic response (or responses) to these examples is, is itself a matter of controversy. The Stoics may well have thought, at least at certain points, that a practiced perceiver could perceptually discern a cataleptic appearance from a non-cataleptic but very similar appearance (e.g., Cicero, Acad. 2.20). After all, the Stoic claim is that some appearances are cataleptic and can be known by the perceiver to be cataleptic. In fact, the “trick” examples of twins, grains of sand, similar eggs, and so on would seem to have a limited affect on undermining the cogency of a cataleptic appearance.49 An inability to tell via perception whether one grain of sand is the same as the grain of sand you were just shown previously hardly undermines the idea that one can tell via perception that, say, a book rather than a rabbit is directly in front of you. Indeed, the fact that these kinds of examples are raised suggests to me that their target is not primarily the idea that no appearance could be cataleptic, but rather the distinct claim that the Sage will never err. There are two distinct positions held by the Stoics, often confusingly discussed simultaneously in our sources. One is similar to the position that McDowell is defending here, namely, that some perceptual episodes (those involving cataleptic appearances) provide conclusive warrant for belief; that is, they constitute what we would call knowledge or a “grasp”

Ancient Philosophy and Disjunctivism

75

for the Stoics. This is central to the epistemological disjunctivist and, most naturally, to the account of the cataleptic appearance: both seek a position in which the perceptual state of the putative knower is both reflectively accessible and factive. The second aspect of the Stoic position, with which McDowell will disagree, is the belief that in theory, someone, namely the Sage, can wield his perceptual capacities in a way such that they never err and are never fooled. Nevertheless, the Stoics agree that our perceptual capacities are fallible: there are “incognitive” or “acataleptic” appearances, whether because they are false or because they are unclear (LS40C=DL 7.46).50 Much of the debate between McDowell and his critics, however, as between the Stoics and Academics, turns on what follows from the possibility of error. So let us turn back to McDowell’s understanding of fallibility. Fallibility, in general, is applied to capacities. There can be exercises of a capacity in which the possessor fails to exercise that capacity successfully; that is, fails to do what the capacity is specified as a capacity to do. The perceptual capacity is a capacity to be in certain perceptual states that have features of the objective environment perceptually present to one’s rational self-conscious awareness.51 This dovetails nicely with (even if it does not require) experiential disjunctivism and the rejection of the HCF conception of experience, which we saw the Stoics accept; there are some perceptual experiences that reach right out to the facts. This is what our perceptual capacity does when there is a non-defective exercise of it.52 In such non-defective instances, there is no possibility, compatibly [sic] with the subject’s being in such a state, that things are not as she would believe them to be in the beliefs that the state would warrant. When we acknowledge that a capacity is fallible, we acknowledge that there can be exercises of it that are defective, in that they fail to be cases of what the capacity is specified as a capacity to do. That does not preclude us from holding that in non-defective exercises of the perceptual capacity subjects get into perceptual states that provide indefeasible warrant for perceptual beliefs. . . . That is just what one’s position is in a non-defective exercise of the capacity.53 Thus in a case of, say, seeing that something is green, one’s perceptual state leaves no possibility that it is not green. What is difficult about McDowell’s position—and what goes to the heart of the disjunctivism he defends—is that he agrees that the possessor of the capacity to perceptually identify green objects can be fooled and so can take something to be green when it is not. But the response to this is not to “retreat” (as he puts it in “Knowledge and the Internal”)54 to a view about what the capacity does that is common to both the successful

76

Iakovos Vasiliou

and defective exercises of it. Here is the heart of the Academic (and “traditional” epistemology’s) objection. To say that our perceptual capacities are fallible is not just to say that they sometimes, for whatever reason, go awry, but that our perceptual capacity does not “ensure that its possessor is always in a position to discriminate defective from non-defective exercises.”55 McDowell’s response is to concede what the Stoics, with their conception of the Sage, may want to deny: If a capacity did ensure that [its possessor is always in a position to discriminate defective from non-defective exercises], it would in principle be infallible; through carelessness or insufficient attention, someone might still mistake defective exercises for non-defective exercises, but the capacity itself would be perfect. But that sort of perfection is surely not within our reach. . . . It is not just that we are prone to carelessness or inattentiveness, but that no perceptual capacity excludes all possibility of defective exercises such that, however careful or attentive one was, one would not recognize them as defective, at least at the time.56 This is stated by McDowell in propria persona. Here he agrees, at least partly, with the Academic skeptics insofar as he, too, rejects the possibility of a Sage, someone who could completely insulate herself from the possibility of error through care and attention. McDowell appreciates that experiential disjunctivism is insufficient to establish the conclusive warrant of perceptual episodes but nevertheless wants to defend the idea of preserving the criterion, as a reflectively accessible guarantor of truth in the face of denying that anyone, no matter how careful, could ensure that the exercise of their conceptual capacity was non-defective. Only epistemological disjunctivism aims to achieve what the Stoics wish to with the cataleptic appearance. The skeptical threat, however, still looms: Even if one is in a perceptual state that is the upshot of a non-defective exercise of a perceptual capacity—and so puts one in genuine contact with the objective world—for all one knows, one’s state is not of that kind; for all one knows, it is defective.57 So one cannot know what one’s perceptual state (apparently) puts one in a position to know. This is the Academic position (see, e.g., Cicero Acad 2.77–78, 83). How to respond? Essentially, McDowell runs the same point about the fallibility of our perceptual capacities with respect to the fallibility of our knowledge of our own perceptual capabilities. In a good case, one’s perceptual state, being a rational self-conscious state of awareness (as the Stoics too believe), is at the same time also a bit of self-knowledge, namely, knowledge that one is in such a state. So, in a good case, one’s seeing that such and such is the case brings with it the knowledge not only that such and such is the case, but also that one knows that such

Ancient Philosophy and Disjunctivism

77

and such is the case by seeing it. Now, this self-knowledge is, like the perceptual capacity itself, fallible. But McDowell argues that we should not let the defective exercise of this capacity on certain occasions—cases in which we only seem to know that we are seeing—derail the non-defective cases in which we genuinely do know that we are seeing.58 Setting the question of the correctness of the view aside for a moment, it is clearly a defense of exactly the position the Stoics wish to maintain: that a subject can acquire perceptual knowledge and know that she has it. The position, although it acknowledges fallibility and even indiscriminability (on which more later in this chapter), explicitly refuses to retreat to a position where one admits merely some sort of “likelihood” or “convincingness” of one’s perceptual appearances, while nevertheless agreeing that “for all one knows,” one could be wrong; positions similar to this were maintained by the Academics (e.g., see Sextus M. 7.169–175; cf. LS 69D).59 To support his view, McDowell appeals to an argument from Sebastian Rödl, which McDowell cites as an “excellent statement of the point.” The opponent agrees that in a veridical case of perceiving, one does have conclusive grounds on which to base one’s belief, but the problem is that one cannot know whether one is in that state or not; thus the conclusive nature of the grounds is out of the reflective reach, as it were, of the subject. Rödl, like McDowell, then wants to repeat the same move about the awareness of one’s own state made in connection with the perceptual state in the first place. Rödl writes: This objection repeats the mistake: from the fact that, when I am fooled, I do not know that I am, it does not follow that, when I am not fooled, I do not know that I am not. When I know that p as I perceive it to be the case, then I know that I perceive that p. Thus I am in a position to distinguish my situation from any possible situation in which I would be fooled, for, in any such situation, I would not perceive that p, while in the given situation I do.60 There is a considerable impression of bootstrapping here. When I am not fooled about p—whether that is seeing that p or knowing that I see that p—then it does not follow, Rödl says, that I do not know that I am not. But does it follow that I know that I am not being fooled? Apparently, according to Rödl (and McDowell), it does. This is supposed to disarm the “for all one knows” objection. Both parties agree, of course, that if one knows one perceives, then one knows one is not being fooled; and both agree that when I am fooled, I do not know that I am. If one rejected the latter, one would believe in the possibility of a Sage: one who could make oneself invulnerable to error. What McDowell and Rödl then insist on, despite this, is that when I am not fooled, I can know that I am not being fooled. To appreciate their position (even if one is not persuaded

78

Iakovos Vasiliou

by it), it is again helpful to focus on the disjunctivist’s interpretation of ordinary episodes of seeing: when I am seeing, and not merely seeming to see (e.g., hallucinating), then I am in a perceptual state that is both reflectively accessible and factive. Part of the reflective accessibility here is captured in Rödl’s line, “When I know that p as I perceive it to be the case, then I know that I perceive that p.” So just as the possibility of cases of mere seeming to see does not alter the nature of cases of actual seeing, so the cases of being fooled into thinking one knows that one is perceiving ought not to undermine the cases of knowing that one is perceiving, when one is not being fooled. Nothing demands that fallibility automatically generalize in a way that affects all of the cases. Fallibility simply means that there will be particular cases in which one’s perceptual capacity fails to work, and in some of those cases, one’s capacity to know whether one is in a certain perceptual state will also fail to work. But it doesn’t follow that in the non-defective cases, one is not perceiving or knowing that one is perceiving. The Stoics may, in their defense of the cataleptic appearance, be insisting on the same point. In Sextus (M. 7.429), the opponent challenges how a Stoic can identify a cataleptic appearance as cataleptic: the criterion for judging that a cataleptic appearance is cataleptic, must itself be either cataleptic or non-cataleptic. Of course, the Stoics reject the latter; but if they adopt the former, the opponent alleges, they will be led to an infinite regress, requiring, essentially, a way of knowing that one knows that one knows, and so on. Sextus then suggests a response, on behalf of the Stoics, that is very much like Rödl’s and McDowell’s: “But perhaps some [Stoic] will say that the cataleptic appearance is the criterion both of the thing that appears [τοῦ φανταστοῦ], that it does truly exist, and of itself, that it is cataleptic.” (7.430) Here is the self-reflective move.61 Unfortunately, as so often, the Stoic positive argument is stated only by a hostile source without details as to how it was developed.62

7. McDowell and the Stoics’ “When It Has No Obstacle” McDowell raises a final argument that provides another striking parallel with an addition by later Stoics in their discussion of the cataleptic appearance and its role as a criterion of truth. McDowell imagines a case where a person, who would otherwise say, “I know a green thing when I see it,” is told that she will be in a situation where the light will be changed undetectably, and when it is, one cannot reliably judge colors. In such a situation, the subject should now suspend judgment about appearances of things being green. So, in this situation, even when the light has not been altered, and so the perceptual state of the subject is the same as it is when perceiving a green thing—since that is in fact what is going on—she would not be entitled to say she knows she is seeing green. For a perceptual state to be one in which “the environment is present” to the

Ancient Philosophy and Disjunctivism

79

subject, it must be one in which the perceptual state is more than simply accurately presenting the environment to the subject; in addition to that, it must also be able to play the role of a conclusive, indefeasible warrant for knowing what is the case. So, in this hypothesized experimental case, even if in fact the light is never deceptive and so, in one, thinner sense, the deliverances of perception are the same (we might say, metaphysically) as they would be in an ordinary case of seeing a green object outside the context of this experiment, they are still not a matter of “accurately presenting the environment” to the subject because they cannot play the relevant epistemic role necessary to make the subject a knower. The traditional Stoic view holds that for an appearance to be cataleptic, it must be able to play the role of an indefeasible, conclusive warrant in a claim of perceptual knowledge. But the later Stoics add a fourth element to their account of the criterion of truth, beyond the appearance being cataleptic:63 Now, the older Stoics say that this cataleptic appearance is the criterion of truth. The later Stoics, on the other hand, added “when it has no obstacle.” For there are times when a cataleptic appearance does strike us, yet it is not trusted because of the external circumstances. For example, when Heracles stood by Admetus, having brought Alcestis up from below the earth, Admetus did catch a cataleptic appearance from Alcestis, but yet did not trust it. And when Menelaus having come back from Troy saw the true Helen at Proteus’ place (when he had left on the ship her phantom, over whom a ten-year war had been fought) he grasped an appearance that was from a real thing and in accordance with just that real thing, and stamped and impressed, but he did not have in it. So that the cataleptic appearance is the criterion when it has no obstacle; these ones were cataleptic, but had obstacles. For Admetus figured that Alcestis was dead and that a dead person does not rise up.  .  . . And Menelaus observed that he had left Helen under guard on the ship and that it was not unlikely that the one found on Pharos was not Helen, but some phantom or spirit. (Sextus, M.7.253–256, Bett (2005) trans., slightly modified) In these examples, a person is confronted with a cataleptic appearance, at least as far as the first three criteria (see section 5) are concerned: the appearance is (a) from what is, (b) stamped and impressed in accordance with what is, and (c) of such a kind as could not come from what is not. Admetus sees his wife Alcestis; Menelaus sees his wife Helen. These are two examples of veridical perception and an accurate taking in of what is the case. The problem, however, is the context. For Admetus actually to be seeing Alcestis, she must have come back from the dead, which he does not take to be possible (although it is actual in this case). Similarly, for

80

Iakovos Vasiliou

Menelaus actually to be seeing Helen, she must be there in front of him and not on the ship where he thinks he left her. To account for such cases, the Stoics add (d) “when there is no obstacle.” The background beliefs that Admetus and Menelaus hold, which are entirely reasonable but in fact false in these instances, block the possibility that what is otherwise a cataleptic appearance can count as knowledge. It is plausible to think that by the time the Stoics have added “when there is no obstacle,” they must have conceded that one could have an appearance that is phenomenologically indistinguishable from a true appearance. For Menelaus’s thought that he is seeing a phantom (even though it is the real Helen) means that he is having a perceptual experience that meets conditions (a) through (c) for a cataleptic appearance. Nevertheless, in McDowell’s language, the environment is still not present to him because the belief that his perception affords him cannot, because of the “obstacle,” count as an episode of knowing. Just as McDowell argues that a mere true appearance of something as green, even if it is phenomenologically and, indeed, metaphysically the same as one that would be had outside the context of the envisaged experiment, is not the same sort of appearance, so the Stoics say the same once they add “when it has no obstacle” to their criterion of truth. For McDowell, the experimental subject need only believe that the experimenter is “authoritative” in his announcement that an experiment with deceptive light is occurring.64 Even if that turns out to be false (just as in the Stoic case Admetus’s belief that people cannot return from the dead turns out, ex hypothesi, to be false), it still changes the sort of appearance that the experimental subject and Admetus have: it is no longer a case of the “environment being present” because it can no longer play the epistemic role it would outside of that context. In short, it severs the typical sufficiency of a cataleptic appearance for having a criterion of truth.

8. Conclusion Within the context of the experiment and of Alcestis’s miraculous resurrection, for all the subjects know, they are in a situation where the light is deceptive and dead people do not come back to life, and so McDowell and the Stoics agree that they are not knowers even though they are having perceptual experiences that in other contexts would conclusively, indefeasibly be entitlements to perceptual knowledge. But we may still wonder, against both McDowell and the Stoics, why this consideration will not generalize. Isn’t it the case that in general, outside the context of any experiment, the subject cannot know whether she is in a deceptive environment or not, given that it has already been conceded that “the look” of the thing as green will be the same in the deceptive and nondeceptive cases?

Ancient Philosophy and Disjunctivism

81

McDowell does confront this: The mere fact that it is possible for lighting conditions to be undetectably unsuitable for knowing the colours of things by looking at them is not a reason to say that for all she knows the present lighting conditions are unsuitable. That would just be a version of that mistaken view about what follows from the sheer fact of fallibility. We go wrong if we generalize from the case of the experimental subject to cases in which no specific possibility that would make knowledge unavailable is in the offing. The experimental subject has a specific reason to think the light may be unsuitable on this occasion, and the fact that she cannot rule out that possibility undermines her capacity to know the colours of things by looking, even on an occasion on which the light is in fact suitable for that capacity. In the general case, in contrast, the sheer fact that lighting conditions can be undetectably unsuitable does not present the subject with a possibility that she would need to rule out before she could claim to know the colours of things by looking. . . . As always, the sheer fact of fallibility does not show that on an occasion when she makes such a claim and that risk does not materialize, she was not giving expression to knowledge, and knowledge that she knew she had. (49–50, his emphasis, my bold)65 What is so special, however, about the specificity of the reason, as I highlighted in bold? The answer has to be, I think, that for McDowell and for the Stoics, the default assumption, rendered reasonable by the rejection of the HCF conception and views about human beings and their relationship to the world, is that we are perceptually in touch with the world in a way that very frequently renders us knowers. When there are specific conditions, specific reasons to think that this situation is undermined, then it undermines us as knowers as well. But without reason to think these specific conditions are operative, there is not the standing, default worry that the skeptic supposes. In obscure cases we can, as the Stoic defenders say (e.g., Cicero, Acad. 2.19), investigate more closely, because this is not simply a vain activity of adding one more mere appearance about how the world seems, but, absent any specific challenge, potentially additional conclusive evidence about how the world is.66 I have tried to show that McDowell’s defense of epistemological disjunctivism against “traditional” epistemology yields an illuminating and attractive defense of what the Stoics call the cataleptic appearance, detachable from the Stoic claims about what is possible for the Sage. McDowell wants to make epistemology allow the regular possibility of cataleptic appearances as episodes of our knowing about the world. While his view is certainly controversial (and I, of course, have not

82

Iakovos Vasiliou

attempted to defend it here but only to present some of its details), it does give us a sense of the possibility of a defense of the Stoic cataleptic appearance that could sustain the back and forth of two hundred years of Academic attack.67

Notes 1. I argue for a similar view regarding Aristotle’s unquestioning confidence in the overall veridicality of perception in Vasiliou (1996). 2. Burnyeat (1982) is a well-known argument attributing a radical break in philosophical framework to Descartes; Burnyeat’s position is explicitly followed and endorsed by McDowell (1998b [1986]: 238–240). For critical doubts about whether the differences between Descartes and the ancients are as numerous or as radical as Burnyeat and others believe, see Fine (2000). Burnyeat himself suggests that Descartes may be anticipated by Augustine in several important claims about the mind and its relation to the world, which is a view presented in Matthews (2002; see esp. ix–xxii); see further bibliography there. Vogt (2015) argues explicitly that Augustine plays the critical role in the development of the conception of mind and world necessary for external world skepticism. 3. McDowell (1998c [1995]: 396). 4. In contrast with how Reed (2002), for example, connects the Stoics and disjunctivism; see further in this chapter. 5. Throughout this chapter I will focus primarily on the epistemology of visual perception, as do both the contemporary and Greek debates with which I am most concerned. 6. I will use “warrant,” “entitlement,” and “justification” synonymously, unless otherwise specified. 7. McDowell (1998a [1983]: 386). 8. McDowell (1998c [1995]: 405–408). 9. McDowell (1998c [1995]: 396). 10. The “likely” or “convincing” or “probable” are hallmarks of Carneadean Academic skepticism; see Long and Sedley (1987), texts 69D, E, and F (henceforth: “LS 69D,” etc.) and 457–460. 11. McDowell, “Knowledge and the Internal,” 403. 12. See Pritchard (2012); also Greco (2014). 13. Haddock and MacPherson (2008b: 7 ff). 14. As Pritchard (2012, 4), remarks: it suggests that we can “have our cake and eat it too.” 15. Striker (1990: 43), see Brunschwig (1999: 230–232), for some qualifications. 16. See, e.g., Plato, Theaetetus, 151d–184a and Aristotle, Metaphysics, 1010b1– 30, 1053a35–b2, 1063a1–6. 17. There is a lot of secondary literature on responses to “conflicting appearances” in pre-Hellenistic philosophy: as a sample, see Lee (2005), Burnyeat (1990); for a general account of pre-Hellenistic views of perception, see Caston (2015). 18. Interpreting Aristotle’s position as inconsistent with the HCF conception of experience, however, makes good sense of his lack of worry about the generalizing effect of considerations of misperception, illusion, and so on; see Vasiliou (1996). 19. See Long and Sedley (1987: 85–86). 20. See, for further details, Everson (1990); also Asmis (1999).

Ancient Philosophy and Disjunctivism

83

21. There are many discussions of the cataleptic appearance (or “cognitive impression,” as it is often translated): see, as a sample, Frede (1987 [1983], 1999), Long and Sedley (1987: 249–253), Annas (1990), Striker (1990), Reed (2002), Sedley (2002), Hankinson (2003), Perin (2005), Nawar (2014), Brittain (2014), Shogry (2018). 22. Most appearances are perceptual, although not all (cf. LS39A). As throughout, I am interested in the perceptual (particularly, visual) experiences of language-using rational human beings. It seems that the Stoics believe that there are cataleptic appearances of a non-perceptual variety as well, but these will not concern us; see Brennan (1996). 23. I have simply left katalēptikē transliterated as cataleptic in the phrase “cataleptic appearance;” it is often translated “cognitive” (thus matching the resulting epistemic state: cognition) or “apprehensive” (thus matching “apprehension”). Each of these English translations has drawbacks. “Cognitive” is too vague and general a term, since in English all of the appearances would seem to be cognitive in a sense at least. “Apprehensive” as an adjective also has the wrong connotations in English. 24. See Annas (1990: 187): “unlike belief, this kind of assent [katalēptikē phantasia] is guaranteed to get things right. We might think that we now had knowledge—and so we would if knowledge is a grasp of particular facts which excludes error.” See also Long and Sedley (1987: 257), who point out that katalēpsis could be translated “knowledge” in many contexts. We should note that assent to a cataleptic appearance does not always yield the epistemic state of katalēpsis; for if one were a Sage, then one’s assent to a cataleptic appearance would amount to a special kind of katalēpsis, namely scientific knowledge (epistēmē). 25. In ancient testimony, as in contemporary scholarship, appearances are sometimes spoken about as true or false, although technically this may not be the case. For the Stoics, there is a corresponding intelligible content, a lekton, that goes with a given appearance (again, for a mature human being). It is lekta (basically, propositions) that are true and false and also the objects to which we assent, thereby generating belief or cognition or knowledge, again depending on who we are and what we assent to. But, given my focus on the epistemology of perception, I will ignore this, as do other scholars as well as some ancient testimony, and speak more casually of certain appearances as true or false and also of people as assenting to appearances. For a discussion of lekta and their relationship to appearances see de Harven (2018). 26. See Frede (1987 [1983], 1999), Long and Sedley (1987: 239), Striker (1990). 27. Electra is in the presence of mad Orestes, but Orestes thinks he is seeing the Furies; see Euripides, Orestes, 255–259. 28. See LS39A=DL 7.49–50, reporting Diocles of Magnesia, Inwood and Gerson (2008) translation, modified: “[50.] Appearance and phantasm are different. For a phantasm is a semblance in the intellect of the sort that occurs in sleep, and appearance is an impression (τύπωσις) in the soul, i.e. an alteration, as Chrysippus supposes in book 3 of his On Soul.” 29. One might speculate that the reason “true” and “false” are not mentioned here explicitly is because in On the Soul, Chrysippus is not going to get into the fact that technically one assents to lekta, and only propositions are true or false, but he does not want to speak incorrectly. 30. As Reed (2002: 173–174), also takes it. Overall, Reed connects Stoicism with disjunctivism, but focuses on whether it is right to interpret the Stoics, at different points, as direct realists. While he does not explicitly distinguish between epistemological disjunctivism and experiential (or metaphysical) disjunctivism (see Haddock and MacPherson 2008a: 1–32; Pritchard 2012:

84

31.

32.

33. 34. 35. 36.

37. 38.

39. 40. 41.

42. 43. 44. 45. 46.

Iakovos Vasiliou 23–24), he ends up concentrating on the latter (see, e.g., 175–176); my focus will be on the former, sections 6 and 7. I should note here that two of our main sources for information about Stoic epistemology, Sextus’s Against the Mathematicians=Against the Logicians and Cicero’s Academica, are written by authors generally hostile to the Stoics. Cicero (first century BC) is sympathetic with Academic skepticism, while the much later Sextus (probably second century AD) is an advocate of Pyrrhonian skepticism. There is some dispute about how reliable this text is as an account of Stoicism because of concerns that it was perhaps influenced by later Academic, Carneadean views about appearances; see Long and Sedley (1987: vol. 2, 242), who note that Sextus’s distinctions like the persuasive and nonpersuasive have no parallel in Diogenes’s account and that Sextus’s description is “wholly different in content and sequence.” They might have added that the same is true in comparison with the passage we just considered from Aëtius, whereas the Diogenes/Diocles and Aëtius passages have considerable overlap. One might keep in mind, too, that while Cicero and Sextus are explicitly hostile to Stoicism (as I mentioned earlier), Diogenes/Diocles and Aëtius are merely claiming to describe it. Pritchard (2012: 1). Perin (2005: 386). See: DL 7.46; Sextus, M., 7.248, 402, 426; Cicero Acad. 2.77–78; several texts collected in LS40. Hankinson (2003: 61), argues that (b) itself makes two claims, adding the following to the issue I described earlier: that it be impressed in accordance with what is, which is meant to rule out the case where someone might hallucinate not, for example, that there is a person in front of them when there is none (ruled out by [a]), but that the person in front of them, who is an ordinary person, is a ghost; cf. LS 40E=Sextus, M. 7.247–252. See Frede (1987 [1983]). Striker (1990: 152); see also, Frede (1987 [1983]: 160) (cf., 167–168): “Evidence is an objective feature of the impressions which is not to be confused with a subjective feeling of conviction or certainty, however strong that feeling may be, just as having a clear view of something is a matter of objective fact and not of subjective feeling. How we know that an impression is evident is a different matter, to which we will turn later; for this, our ‘feeling’ may very well be relevant, but it seems, even in optimal circumstances, to be no more than a symptom of the evidence of an impression” (my italics). See Perin (2005: 397). See Reed (2002), Perin (2005), Nawar (2014). Reed (2002) is the only commentator to discuss disjunctivism and the Stoics, but his focus on experiential rather than epistemological disjunctivism leads him to say that even if the veridical experience is of a different nature from a phenomenologically indistinguishable hallucination, that difference is not accessible to the perceiver and so not available to help with her epistemological entitlement. In what follows, I can only attempt to articulate how McDowell’s argument goes, what he hopes to achieve by it, and trace its close connections with the views of the Stoics. I cannot undertake to defend it in its own right. McDowell (2011: 28–29, emphasis in original). McDowell (2011: 30). As perception does for rational human beings, according to the Stoics. McDowell (2011: 30–31, emphasis in original).

Ancient Philosophy and Disjunctivism 47. 48. 49. 50. 51. 52.

53. 54. 55. 56. 57. 58. 59.

60. 61. 62.

63. 64.

85

McDowell (2011: 34). McDowell (2011: 36). See Vogt (2013) for some discussion. Even for the Sage, it is not that his perceptual capacity is infallible; it is, rather, that he uses it unerringly so that he assents exclusively to cataleptic appearances. McDowell (2011: 37). We should keep in mind that rejection of the HCF conception of experience makes the default position with respect to perceptual episodes to construe them as being in direct contact with the world. That is what the job of perception is: for the Stoics this is part of their overall cosmic teleology, where perception is part of our ability as rational creatures to come to know and understand the rational order of the cosmos; for us today, our having (fallible, but largely functioning) perceptual capacities is understood as part of a Darwinian explanation of why certain capacities were selected for. McDowell (2011: 38–39). See McDowell (1998c [1995]: 396–399). McDowell (2011: 39). Ibid., 40; see the same point in McDowell (1998c [1995]: 399–400); cf. also 407–408. McDowell (2011: 40–41). McDowell (2011: 42). Near the end of the essay (53–54), McDowell asserts that if, contrary to what McDowell himself has argued for, one insists that the best that perception can ever deliver is an inconclusive warrant, then one has only two choices: skepticism (i.e., there is no such thing as perceptual knowledge) or dogmatism, “claiming that an inconclusive warrant can be good enough for a belief it warrants to be knowledgeable.” (McDowell cites Pryor [2000] as holding this position.) But then he adds: “it is hard to see how dogmatism genuinely contrasts with giving in to skepticism. We are to grant that the warrant for a belief leaves open a possibility that the belief is false. I know no convincing explanation of how that can be made out to differ from granting that for all the believer knows her belief may be false.” This collapsing of the contemporary “dogmatic” view into skepticism, in McDowell’s view, matches the ancient philosophical landscape, where the alternative to the Stoic view is the Academic skeptic’s position, which embraces the claim that “for all one knows,” one’s belief may be false. The ancients too would refuse to call a modern, “hybrid” position such as Pryor’s non-skeptical. Rödl (2007: 157–158), quoted by McDowell (2011: 43–44). One should make clear too that not all contemporary epistemological disjunctivists pursue this self-reflective move as part of their defense: See, e.g., Pritchard (2012). In the difficult and somewhat obscure skeptical argument against this selfreflective move (7.430–432), Sextus does not rely on the infinite regress argument but adduces what he takes to be other absurd consequences of claiming that a cataleptic appearance could be a judge of its own status. The flow of the argument here seems to suggest that Sextus takes the self-reflective stance to be sufficient to block the regress, even though of course he did not think it sufficient to vindicate the cataleptic appearance as a conclusive basis for knowledge and as a criterion for truth. See Brittain (2014), esp. 340 ff., for further discussion. McDowell (2011: 47).

86

Iakovos Vasiliou

65. In McDowell’s (2011) final argument, 52–53, he imagines two scenarios. The first is one where the experimenter (truthfully) tells the subject that the light will be inadequate for color identification in one out of a hundred cases. In the second case, we are to imagine that we have scientifically discovered that our perceptual capacity to identify colors (under good conditions) will misfire in one out of a hundred cases. McDowell argues that the mere fact that the statistical chance of misidentifying the color is the same in both cases, in the experimental case the subject can never be a knower, since she is permanently in a “for all one knows” situation; while in the latter case she will be a knower, and know that she is, in the ninety-nine cases in which she identifies the color correctly. 66. This is one reason, so far as I can see, that what is being confronted here is not radical New Evil Genius or Brain-in-a-Vat Skeptical hypotheses, which are not entertained in the Stoic/Academic debate. I think that McDowell must rely on more than I have discussed here to treat this sort of skepticism. See Neta and Pritchard (2007) and Pritchard (2012, part 3) for epistemological disjunctivist responses to radical skepticism. 67. I would like to thank the audience for the Ancient Philosophy Work in Progress Group at New York University for their helpful feedback on an initial presentation of these ideas. I am especially grateful for comments on subsequent drafts by the editors of this volume as well as Vanessa de Harven, Brad Inwood, Whitney Schwab, Rosemary Twomey, Katja Vogt, and Nancy Worman.

References Annas, Julia, 1990, Stoic epistemology, in Stephen Everson, ed. Companions to Ancient Thought 1: Epistemology. Cambridge: Cambridge University Press, pp. 184–203. Asmis, Elizabeth, 1999, Epicurean epistemology, in K. Algra, J. Barnes, J. Mansfeld, and M. Schofield, eds. The Cambridge History of Hellenistic Philosophy. Cambridge: Cambridge University Press, pp. 260–294. Bett, Richard, 2005, Sextus Empiricus: Against the Logicians. Cambridge: Cambridge University Press. Brennan, Tad, 1996, Reasonable impressions in stoicism, Phronesis 41: 318–334. Brittain, Charles, 2006, Cicero: On Academic Scepticism. Indianapolis, IN: Hackett. Brittain, Charles, 2014, The compulsions of stoic assent, in M.-K. Lee, ed. Stategies of Argument: Essays in Ancient Ethics, Epistemology, and Logic. Oxford: Oxford University Press, pp. 332–355. Brunschwig, Jaques, 1999, Introduction: The beginnings of Hellenistic epistemology, in K. Algra, J. Barnes, J. Mansfeld, and M. Schofield, eds. The Cambridge History of Hellenistic Philosophy. Cambridge: Cambridge University Press, pp. 229–259. Burnyeat, Myles, 1982, Idealism and Greek philosophy: What Descartes saw and Berkeley missed, The Philosophical Review 90: 3–40. ———, 1990, The Theaetetus of Plato. Indianapolis, IN: Hackett. Caston, Victor, 2015, Perception in ancient Greek philosophy, in Mohan Matthen, ed. The Oxford Handbook of Philosophy of Perception. Oxford: Oxford University Press, pp. 29–50. de Harven, Vanessa, 2018, Rational impressions and the Stoic philosophy of mind, in John Sisko, ed. Philosophy of Mind in Antiquity. London & New York: Routledge, pp. 215–235.

Ancient Philosophy and Disjunctivism

87

Everson, Stephen, 1990, Epicurus on the truth of the senses, in his Companions to Ancient Thought 1: Epistemology. Cambridge: Cambridge University Press, pp. 161–183. Fine, Gail, 2000, Descartes and ancient skepticism: Reheated cabbage? The Philosophical Review 109: 195–234. Frede, Michael, 1987 [1983], Stoics and skeptics on clear and distinct perceptions, in his Essays in Ancient Philosophy. Minneapolis: University of Minnesota Press, pp. 151–176. ———, 1999, Stoic epistemology, in K. Algra, J. Barnes, J. Mansfeld, and M. Schofield, eds. The Cambridge History of Hellenistic Philosophy. Cambridge: Cambridge University Press, pp. 295–322. Greco, John, 2014, Pritchard’s Epistemological Disjunctivism: How right? How radical? How satisfying? The Philosophical Quarterly 64: 115–122. Haddock, Adrian and Macpherson, Fiona, eds., 2008a, Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press. ———, 2008b, Introduction: Varieties of disjunctivism, in Adrian Haddock and Fiona Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 1–32. Hankinson, Robert J., 2003, Stoic epistemology, in Brad Inwood, ed. The Cambridge Companion to the Stoics. Cambridge: Cambridge University Press, pp. 59–84. Inwood, Brad and Gerson, Lloyd, 2008, The Stoics Reader. Indianapolis, IN: Hackett. Lee, Mi-Kyoung, 2005, Epistemology After Protagoras: Responses to Relativism in Plato, Aristotle, and Democritus. Oxford: Oxford University Press. Long, Anthony A. and Sedley, David N., eds., 1987, The Hellenistic Philosophers (2 vols.). Cambridge: Cambridge University Press. Matthews, Gareth B., ed., 2002, Augustine: On the Trinity Books 8–15. Cambridge: Cambridge University Press. McDowell, John, 1998a [1983], Criteria, defeasibility, and knowledge, in his Meaning, Knowledge, and Reality. Cambridge, MA: Harvard University Press, pp. 369–394. ———, 1998b [1986], Singular thought and the extent of inner space, in his Meaning, Knowledge, and Reality. Cambridge, MA: Harvard University Press, pp. 228–259. ———, 1998c [1995], Knowledge and the internal, in his Meaning, Knowledge, and Reality. Cambridge, MA: Harvard University Press, pp. 395–413. ———, 2011, Perception as a Capacity for Knowledge. Milwaukee, WI: Marquette University Press. Nawar, Tamer, 2014, The Stoic account of apprehension, Philosopher’s Imprint 14(29): 1–21. Neta, Ram and Pritchard, Duncan, 2007, McDowell and the new evil genius, Philosophy and Phenomenological Research 74: 381–396. Perin, Casey, 2005, Stoic epistemology and the limits of externalism, Ancient Philosophy 25: 383–401. Pritchard, Duncan, 2012, Epistemological Disjunctivism. Oxford: Oxford University Press. Pryor, James, 2000, The skeptic and the dogmatist, Nous 34(4): 517–549. Reed, Baron, 2002, The stoics’ account of the cognitive impression, Oxford Studies in Ancient Philosophy 23: 147–180.

88

Iakovos Vasiliou

Rödl, Sebastian, 2007, Self-Consciousness. Cambridge, MA: Harvard University Press. Shogry, Simon, 2018, Creating a mind fit for truth: the role of expertise in the Stoic account of the kataleptic impression, Ancient Philosophy 38: 357–381. Sedley, David, 2002, Zeno’s definition of phantasia kataleptike, in T. Scaltsas and A.S. Mason, eds. Zeno of Citium and His Legacy: The Philosophy of Zeno. Larnaca: Municipality of Larnaca, pp. 133–154. Striker, Gisela, 1990, The problem of the criterion, in Stephen Everson, ed. Companions to Ancient Thought 1: Epistemology. Cambridge: Cambridge University Press, pp. 143–160. Vasiliou, Iakovos, 1996, Perception, knowledge, and the sceptic, Oxford Studies in Ancient Philosophy 14: 83–131. Vogt, Katja, 2013, The Hellenistic academy, in Frisbee Sheffield and James Warren, eds. Routledge Companion to Ancient Philosophy. New York: Routledge, pp. 482–495. ———, 2015, From investigation to doubt: The beginnings of modern skepticism, in Gareth Williams and Katharina Volk, eds. Roman Reflections. Oxford: Oxford University Press, pp. 260–174.

5

The Kantian Roots of Epistemological Disjunctivism Thomas Lockhart

John McDowell is one of the principal proponents of epistemological disjunctivism.1 McDowell has also articulated and defended a distinctive interpretation of Kant’s Critique of Pure Reason.2 These statements do not simply report two independent facts about McDowell’s intellectual biography. McDowell’s interpretation of Kant is closely intertwined with McDowell’s conception of epistemological disjunctivism. The exact nature of the interconnections between McDowell’s version of epistemological disjunctivism and McDowell’s Kant are not, I think, widely appreciated. So this is not an interpretative paper about Kant. I do not directly defend the claim that Kant was an epistemological disjunctivist, although I do argue that McDowell’s Kant was an epistemological disjunctivist. My task in this chapter is simply to show how McDowellian epistemological disjunctivism is intertwined with McDowell’s interpretation of Kant. Connections between McDowell’s epistemological disjunctivism and McDowell’s Kant run in both directions. On the one hand, I argue that the epistemological disjunctivist must articulate a conception of perceptual experience according to which perceptual experience can (if all goes well) ‘reveal’ the world in such a way that beliefs based on those perceptual experiences are entitled by, and thus rational responses to, those perceptual experiences. For McDowell, the best available account of perceptual experience which meets this constraint is Kant’s notion of an empirical intuition, as articulated in the first Critique. On the other hand, we’ll see that the conception of an empirical intuition which McDowell attributes to Kant is rich enough to sustain the claim that McDowell’s Kant was an epistemological disjunctivist.

1. Epistemological Disjunctivism According to the epistemological disjunctivist, in a paradigmatic case of perceptual knowledge, a subject S knows that p in virtue of having a perceptual experience which is, for her, an indefeasible warrant to judge that p. In such cases, S’s warrant for her perceptual knowledge that p is her

90 Thomas Lockhart perceptual experience. S’s perceptual experience in such cases is factive: it follows from the fact that S has such a perceptual experience that p is true. S’s warrant is also, in such cases, reflectively accessible: S can come to know, through reflection alone, that her warrant for her perceptual knowledge is her factive perceptual experience.3 So we can formulate epistemological disjunctivism (ED) thus: (ED) In a paradigmatic case of perceptual knowledge, a subject S knows that p in virtue of having a perceptual experience which is, for her, an indefeasible warrant to judge that p. It will be useful to have a formulation of the core commitments of epistemological disjunctivism, the wording of which will allow us more readily to move to a discussion of Kant. I propose the following McDowellian slogan, which I call the Rationality Manifestation Thesis (RMT): (RMT) Perceptually based beliefs are manifestations of rationality as such.4 In section 2, I explain what McDowell means by RMT, paying specific attention to the phrase ‘rationality as such’. In sections 3 and 4, I show why ED and RMT can be thought of as equivalent. In section 5, I argue that in order to defend RMT (and thus ED), it will be necessary to articulate a conception of perceptual experience which satisfies two constraints. On the one hand, an adequate account of perceptual experience must show how perceptual experiences have content in a manner which allows us to say that perceptually based beliefs are rational responses to those perceptual experiences. On the other hand, an adequate account of perceptual experience must show how it could be that perceptual experience can ‘reveal’ the world in a manner suitable to entitle a perceiver to a knowledge claim. That is, the epistemological disjunctivist must articulate a conception of perceptual experience according to which perceptually based beliefs can be understood as both entitled by, and rational responses to, perceptual experiences. This puts me in a position, in sections 6 and 7, to show why Kant’s account of empirical intuition, as interpreted by McDowell, best satisfies these constraints on an adequate epistemological disjunctivist conception of perceptual experience.

2. Rationality as Such In this section, I explain what McDowell means by saying that rationality ‘as such’ is manifest in perceptually based beliefs.

Kantian Roots of Epistemological Disjunctivism

91

Perceptually based beliefs are, by definition, based on perceptual experience. If such beliefs are to be ‘manifestations of rationality’, we’ll need to explain how perceptually based beliefs can count as rational responses to perceptual experience. Here is McDowell: If someone has a perceptually based belief, she believes something because her experience reveals to her, or at least seems to reveal to her, that things are as she believes them to be. And that ‘because’ introduces an explanation that depends on the idea of rationality in operation. McDowell (2009b: 127) Let us apply this wording to a particular case. Suppose that Gwendolen has a perceptually based belief that there is a chestnut horse in view. According to McDowell, we can offer the following explanation: (1) Gwendolen believes the horse is chestnut because her experience reveals to her, or at least seems to reveal to her, that the horse is chestnut. Explanation (1) is, I think, perfectly ordinary and ought to be philosophically uncontentious. One feature worth noticing is this: explanation (1) deploys the idea that Gwendolen’s experience ‘reveals to her’ that the horse is chestnut. This choice of terminology is intended to be neutral; it is not supposed to smuggle in any particular account of perception before we start to develop one. It is true that McDowell takes ‘reveal’ to be factive: if Gwendolen’s experience reveals to her that the horse is chestnut, then the horse is indeed chestnut. But this does not undermine the innocence of explanation (1), for there is a crucial disjunction: explanation (1) explains Gwendolen’s perceptually based belief that the horse is chestnut in terms of a perceptual experience which reveals to her, or at least seems to reveal to her, that the horse is chestnut. Given this disjunction, I see no reason to be wary of following McDowell and allowing ‘reveals’ to be factive. McDowell takes it that Gwendolen’s rationality is ‘manifested’ in explanation (1). We’ve seen him put what he takes to be the same point this way: the ‘because’ in explanation (1) ‘depends on the idea of rationality in operation’. But in what sense is Gwendolen’s rationality ‘in operation’ or ‘manifested’ in an explanation such as (1)? The kind of rationality at play in an explanation such as (1) is what McDowell calls the “notion of responsiveness to reasons as such” (McDowell 2009b: 128; emphasis in original). To have the ability to respond to reasons ‘as such’ is to be able to “step back and assess whether putative reasons warrant action or belief” (McDowell 2009b: 129). And

92 Thomas Lockhart for this form of rationality to be ‘operative’, for McDowell, the ability need not be exercised: [I]f the capacity is present without being exercised, we have in view someone who can respond to reasons as the reasons they are. And rationality in the sense I am explaining may be actually operative even though the capacity to step back is not being exercised. McDowell (2009b: 129) So when McDowell says that the explanation (1) depends on ‘the idea of rationality in operation’, he means this: for explanation (1) to be admissible, it must be that Gwendolen could assess whether the putative reason for her belief does indeed warrant that belief (although explanation (1) takes no stand on whether Gwendolen has in fact assessed the putative reason for her perceptually based belief). In other words, explanation (1) presupposes that Gwendolen could assess whether her perceptual experience warrants her belief that the horse is chestnut. The ability McDowell attributes to Gwendolen is not extraordinary. Beliefs based on perceptual experiences are (at least prima facie) embedded in a network of rational connections: perceptually based beliefs can serve as reasons for other beliefs, and we can ask for what reason one holds a perceptually based belief. This is exhibited in perfectly quotidian exchanges. Suppose I text Gwendolen the question, ‘Why do you think Henleigh’s horse is chestnut?’ She might give her reason for this belief by texting back, ‘I can see it’. This exchange is enough to exhibit the relevant structural feature that McDowell needs: Gwendolen has the ability to give a reason for her perceptually based belief. McDowell emphasizes the fact that rationality as such also requires that one have some facility with being able to address the question: How good is my reason for a particular belief? But this, too, is routine. Suppose Gwendolen has a perceptual experience which seems, to her, to reveal a chestnut horse in the distance. On this occasion, the object which looks to her like a chestnut horse is rather far away. Accordingly, she decides not to judge, on the basis of her perceptual experience, that there is a chestnut horse in the distance. For McDowell, this is enough to show that Gwendolen has the ability to ‘step back and assess’ whether a perceptual experience does (or does not) warrant a perceptually based belief. To claim that Gwendolen could assess whether her perceptual experience warrants her belief is not to say that Gwendolen can always successfully answer the question whether her perceptual experience warrants whatever perceptually based belief the experience has prompted. It might be that there are particular cases in which Gwendolen’s attempts to determine whether her perceptual experience warrants some belief are doomed to arrive at the wrong answer. A vivid illustration is furnished by skeptical scenarios. Suppose that Gwendolen is confronted with a

Kantian Roots of Epistemological Disjunctivism

93

sculpture of a chestnut horse, which is, to her, visually indistinguishable from a real horse. Suppose that on the basis of a visual experience of this sculpture, she comes to believe that she is looking at a horse. Now she asks herself: ‘Does my visual experience warrant my perceptually based belief?’5 It is quite possible that, even after reflection, she will (wrongly) answer ‘Yes’. But the possibility of such cases does not undermine McDowell’s point: simply that Gwendolen has the ability to step back and assess her putative reasons. Even if Gwendolen is trapped in some skeptical scenario, she can always ask herself whether her perceptual experience gives her a warrant for a particular belief putatively warranted by that experience. I’ve made some effort to defend it, but one might nevertheless reject McDowell’s claim about the manner in which rationality enters into an explanation such as (1). For example, one might simply deny that explanations such as (1) are rationality-invoking (the ‘because’ in explanation (1) might be a merely causal ‘because’). Or one might agree that explanations such as (1) are rationality-invoking but claim that the form of rationality invoked in such explanations is weaker than the form on which McDowell insists. For example, one could be moved by the following thought: animals and infants clearly have perceptually based beliefs which we can explain as rational responses to perceptual experiences, but animals and infants do not possess the ability to ‘step back’ and assess whether their perceptual experiences warrant those perceptually based beliefs. McDowell is perfectly happy to accommodate this point by claiming that perceptually based beliefs in such animals and infants merely exhibit rationality, not rationality as such.6 Instead of further litigating this issue, let us simply record what McDowell means by claiming that rationality as such is manifested in explanations such as (1): (2) For a perceptually based belief to be a manifestation of rationality as such, the believer must have the ability to assess whether the putative perceptual reason for her belief warrants that belief.

3. The Rationality Manifestation Thesis and Epistemological Disjunctivism I turn now to the task of explaining why RMT and ED can helpfully be thought of as equivalent. In section 2, we worked with the following example of an explanation of a perceptually based belief: (1) Gwendolen believes the horse is chestnut because her experience reveals to her, or at least seems to reveal to her, that the horse is chestnut.

94 Thomas Lockhart According to RMT, rationality as such is manifest in this form of explanation. Now if explanation (1) involves rationality as such, then it will follow that rationality as such is also at work in the following explanation: (3) Gwendolen believes that the horse is chestnut because her experience reveals to her that the horse is chestnut. There is, I take it, room for philosophical anxiety concerning the kind of explanation on offer in (3). One source of anxiety might arise from standard skeptical considerations. If Gwendolen is at the mercy of an evil demon, it might never be correct to explain any of her perceptually based beliefs in the manner deployed in explanation (3). To respond to this worry, notice that my claim was simply that rationality as such is at work in an explanation such as (3). My claim was not that we will ever be in the happy situation of being correctly able to deploy such an explanation. RMT does not entail that explanation (3) is ever true; RMT simply requires that explanation (3) is (to use McDowell’s preferred term-of-art) intelligible. Note that for something to be ‘intelligible’, in McDowell’s sense, it need not be correct. For something to be intelligible, is, roughly speaking, for it to be coherent. A claim would be unintelligible if there were some philosophical argument which succeeded in showing that it must lapse into incoherence. We can accommodate this first anxiety about admitting explanation (3) thus: (4) RMT entails that rationality as such is at work in explanation (3) (even if we are never, in fact, in a position correctly to be able to use an explanation such as (3)). A second source of anxiety concerning the explanation on offer in (3) might arise from a philosophical account of experience according to which experience does not ‘reveal’ the world. A philosophical account of experience according to which experience does not ‘reveal’ the world might come in various guises. One form such an account of experience might take would be to claim that the nature of experience is such that at best we can make use of the following kind of explanation: (5) Gwendolen believes that the horse is chestnut because her experience seems to reveal to her that the horse is chestnut. But this kind of worry need not trouble the proponent of RMT. Notice that explanation (5) characterizes Gwendolen’s experience using ‘seems’-talk. And ‘seems’-talk presupposes ‘is’-talk: in explanation (5),

Kantian Roots of Epistemological Disjunctivism

95

Gwendolen’s perceptual experience seems just like the experience she would be having if we were able to use explanation (3).7 But because ‘seems’-talk presupposes ‘is’-talk, if explanation (5) is coherent, then explanation (3) must also be coherent. So even if some philosopher is committed to an account of the nature of perceptual experience according to which experience does not ‘reveal’ the world, if she makes use of an explanation such as (5), then, because ‘seems’-talk presupposes ‘is’-talk, she will be committed to the intelligibility of explanation (3) (even though she will insist that, in virtue of the nature of perceptual experience itself, we can never in fact make use of an explanation such as (3)). And the mere intelligibility of explanation (3) is sufficient for my current purposes. A philosophical account of experience according to which perceptual experience does not ‘reveal’ the world might take on a more radical guise and insist that perceptual experience is not in the business of even seeming to reveal the way the world is. On such an account, perceptual experience would have no representational content. But this would simply amount to a denial of RMT. For according to RMT, perceptually based beliefs are rational responses to perceptual experience. If Gwendolen’s perceptual experience has no representational content, then a belief based on that perceptual experience cannot be a rational response to that perceptual experience. To think otherwise is (as I’ll explain in a little more detail in section 4), simply to fall into the ‘Myth of the Given’. So far I’ve argued that if RMT is correct, then rationality as such is at work in an explanation such as: (3) Gwendolen believes that the horse is chestnut because her experience reveals to her that the horse is chestnut. We saw (in section 2) that what it means to say that rationality as such is at work in explanation (3) is this: Gwendolen is able to assess whether the reason she takes herself to have for her perceptually based belief does, or does not, entitle her to that belief. In explanation (3), Gwendolen’s experience does indeed reveal to her that the horse is chestnut. Suppose, in this case, she attempts to determine whether her perceptual experience entitles her to her perceptually based belief; and suppose furthermore she is successful in this attempt. Since her perceptual experience reveals that the horse is chestnut, then (when she successfully exercises that ability) Gwendolen will determine that her perceptual experience entitles her to her perceptually based belief because that experience reveals to her a chestnut horse. One might object to the argument I’ve just rehearsed as follows: Why think that Gwendolen could ever successfully determine that her perceptual experience entitles her to her perceptually based belief?

96 Thomas Lockhart The response (again) is that we don’t need (yet) the claim that Gwendolen ever could successfully determine that her perceptual experience entitles her to her perceptually based belief. If she is at the mercy of a Cartesian demon, then she will never successfully determine that her perceptual experience entitles her to her perceptually based belief. All we need at this stage is the mere coherence of the idea that she could successfully determine that her perceptual experience entitles her to her perceptually based belief. And the coherence of this idea is guaranteed by RMT. For RMT claims that rationality as such is at work in an explanation such as (3). This means, according to gloss (2), that Gwendolen has the ability to assess whether the reason she takes herself to have for her perceptually based belief entitles her to that belief. Since she has this ability, it must at least be intelligible that she could successfully exercise it (even if, as a matter of fact, she never can), for that is precisely what it is an ability to do. So if RMT is correct, then rationality as such is at work in an explanation such as (3) and, accordingly, it must be coherent to think of perceptual experience, when all goes well, as rationally (in the sense of rationality as such) entitling a perceiver to a perceptually based belief by revealing the world to her in a suitable way. To summarize: according to RMT, Gwendolen has an ability to assess whether her perceptual experience entitles her to her perceptually based belief. If she is in a situation in which her experience reveals to her that the horse is chestnut, and she successful exercises that ability, then she will determine that her perceptual experience does indeed so entitle her. But this is just the position of the epistemological disjunctivist: in a paradigmatic perceptual experience, Gwendolen knows that the horse is chestnut in virtue of enjoying a perceptual experience which reveals to her that the horse is chestnut; furthermore, she can, on reflection, come to tell that her perceptual experience entitles her to judge that the horse is chestnut because the experience reveals, to her, the chestnut horse. So RMT entails ED. It is straightforward to see that ED entails RMT. For the epistemological disjunctivist is committed to the claim that in a paradigmatic perceptual experience, Gwendolen can know that the horse is chestnut in virtue of having a perceptual experience which provides her with an indefeasible warrant to judge that the horse is chestnut. I’ve stipulated that since the warrant is, for the epistemological disjunctivist, indefeasible, this means that Gwendolen can, through reflection, determine that her perceptual experience entitles her to judge that the horse is chestnut. But since, for the epistemological disjunctivist, Gwendolen has the ability to reflect on the question whether her perceptual experience gives her a warrant for her perceptually based belief, then the epistemological disjunctivist must be committed to the thought that Gwendolen’s perceptually based belief is a rational response to her perceptual experience, in the sense of rationality as such. And this is simply RMT.

Kantian Roots of Epistemological Disjunctivism

97

4. Responding to the Cartesian Skeptic One might think that epistemological disjunctivism, as I have articulated it here, would not be enormously useful in responding to the skeptic. For epistemological disjunctivism, as I have characterized it, merely enters a claim about the nature of paradigmatic perceptual experiences; it makes no claim about whether anybody ever enjoys any paradigmatic perceptual experiences. Thus far I’ve merely insisted that certain conceptions of perceptually based belief are intelligible, and I’ve said nothing about their actuality. So do we ever enjoy any paradigmatic epistemological disjunctivist perceptual experiences? Common sense, according to McDowell, tells us that (6) We can know about the world around us on the basis of our perceptual experience. The epistemological disjunctivist claims that we can make philosophical sense of the idea that perceivers can know about the world around them on the basis of perceptual experience. So (if the epistemological disjunctivist conception of perceptually based belief is intelligible), there is nothing to stand in the way of our adopting the common-sense position of stipulation (6). Thus, in response to the question, ‘Is it the case that Gwendolen ever has a paradigmatic perceptual experience?’ McDowell answers: ‘Of course, for that is nothing more than common sense’.8 But what of the skeptic? For McDowell, we should abandon the position of common sense only if we are rationally forced to do so. A position which ruled the epistemological disjunctivist conception of perceptually based belief unintelligible would force us to abandon explanation (6). Can the skeptic show that such a conception is unintelligible? It depends on which form of skepticism we have in mind. According to McDowell, the standard forms of external world skepticism (call them Cartesian skepticism) cannot show that it is unintelligible that we can have perceptually based knowledge of the external world. Indeed, since the Cartesian skeptic makes extensive use of ‘seems’-talk, and ‘seems’-talk presupposes ‘is’-talk, the Cartesian skeptic must be committed to the intelligibility of the idea that perceptual experience can reveal how the world is. On McDowell’s interpretation, the Cartesian skeptic does nothing more than attempt to provoke anxiety by drawing attention to the existence of (what are often called) ‘bad’ cases. But the mere existence of ‘bad’ cases cannot rationally force the epistemological disjunctivist to abandon stipulation (6), for (McDowell argues) the epistemological disjunctivist can straightforwardly account for the existence of ‘bad’ cases without abandoning her claim that (what are typically called) ‘good’ cases are possible.

98 Thomas Lockhart How does the epistemological disjunctivist accommodate the possibility of ‘bad’ cases? The argument turns on the nature of fallibility. Gwendolen is human, and human capacities are fallible. Thus on any occasion on which Gwendolen attempts to exercise her capacity to assess whether her perceptually based belief is warranted, her attempt might go awry. There are various ways her attempt to exercise her capacity might backfire. She might come to believe that her perceptual experience does indeed warrant her belief, even though this is not so. Or she might judge that her perceptual experience does not warrant her perceptually based belief, even though it does. But if Gwendolen’s perceptual experience does indeed reveal, to her, that the horse is chestnut, and Gwendolen successfully exercises that capacity, then, as a result of that exercise of her capacity, she will determine correctly that her perceptual experience warrants her perceptually based belief. Call such a case a ‘good’ case. Cartesian skepticism makes much of ‘bad’ cases. These are situations in which Gwendolen’s perceptual experience is, to her, indistinguishable from the perceptual experience she enjoys in the good case. But the epistemological disjunctivist has no difficulty accounting for bad cases. As we have seen, because Gwendolen’s capacity is fallible, her attempts to exercise it will sometimes go awry. One way her attempts to exercise her capacity might go awry is in the manner beloved by the skeptic. But this possibility has no power to show that she can never successfully exercise that capacity. And it is only that conclusion which would force us to abandon stipulation (6). The mere possibility (and indeed reality of) ‘bad’ cases cannot force us to abandon stipulation (6). The possibility of ‘bad’ cases should inspire in us exactly as much anxiety as the familiar fact that human cognitive capacities are fallible.9 Since the epistemological disjunctivist can, by acknowledging the familiar fact that our capacities are fallible, accommodate the ‘bad’ cases deployed by the Cartesian external world skeptic without needing to abandon the position of common sense, the epistemological disjunctivist need not bother with such a form of skepticism.10 Instead, we are free to maintain the position of common sense and hold that Gwendolen often does come to know about the world on the basis of her perceptual experience. The point of this summary of McDowell’s manner of treating the Cartesian skeptic is to show that McDowell thinks the epistemological disjunctivist need not fear skeptical attack unless the attack can force us to abandon the position of common sense by revealing that the epistemological disjunctivist is committed to an incoherent conception of perceptual experience. McDowell thinks that standard varieties of external world skepticism cannot force us to abandon common sense. However, there is a form of skepticism which does claim to show that the epistemological disjunctivist conception of perceptual experience is incoherent: the Kantian skeptical argument.11 In the next section, I introduce that problem.

Kantian Roots of Epistemological Disjunctivism

99

5. Kantian Skepticism and the Myth of the Given Epistemological disjunctivism requires a conception of perceptual experience according to which perceptually based beliefs are both entitled by and rational responses to perceptual experience. The Kantian skeptical dialectic claims that no conception of perceptual experience can meet these demands, because there is no available conception of perceptual experience according to which perceptually based beliefs can be understood as rational responses to perceptual experience. The Kantian skeptical argument is perhaps most familiar as that argument to which the so-called ‘Myth of the Given’ is an inept (at least according to McDowell) response. Here is one version of the argument and the Mythical response. On the one hand, any perceptual experience stands in a causal relationship to various events in the world and for this reason can, presumably, be described and explained in terms appropriate to the natural sciences. On the other hand, perceptually based beliefs are embedded in a network of rational relationships with other beliefs. According to RMT, we must be able to think of these perceptually based beliefs as rational responses to perceptual experiences. But if perceptual experiences are merely the causal upshots of certain natural processes, then such experiences cannot stand in rational relationships to perceptually based belief. At best, perceptual experiences so understood would be causal triggers for perceptually based beliefs, not rational bases for such beliefs. This line of thought is at the heart of the Kantian skeptical argument. For our purposes, the main upshot of the Kantian skeptical argument will be the claim that perceptual experience is not the kind of thing which can stand in a rational relationship to anything. According to McDowell, there is a significant temptation in the history of the philosophy of perception to simply ignore the Kantian skeptic and insist that perceptual experiences, conceived merely as items in a naturalcausal chain, can, so understood, give rational guidance to perceptually based beliefs. But, McDowell claims, such an insistence is hopeless. It is the myth of a ‘Given’, which simultaneously lies outside the rational order but is nevertheless capable of providing justificatory constraints on transactions within that rational order.12 One of McDowell’s attempts to formulate the Myth in a manner appropriate to capture its various guises is this: “Givenness, in the sense of the Myth, would be an availability for cognition to subjects whose getting what is supposedly Given to them does not draw on capacities required for the sort of cognition in question” (McDowell 2009a: 256). As McDowell points out, it is straightforward to see that Givenness, in this sense, must be mythical. One falls foul of the Myth when one fails to “realize that knowledge of some kind requires certain capacities” (Ibid.). This is precisely the problem with (what he calls) ‘traditional empiricism’:

100 Thomas Lockhart it fails to realize that experience cannot make things available for our cognition “without any involvement of capacities that belong to our rationality” (McDowell 2009a: 257). And it is only through embracing this idea, ignored by the traditional empiricist, that one can avoid the Myth. I cannot here defend McDowell’s claim that it is compulsory to avoid the Myth of the Given. Perhaps it is not a Myth. What I have shown is this: the epistemological disjunctivist must articulate a conception of perceptual experience which shows how perceptually based beliefs can count as rational responses to perceptual experiences. Kantian skepticism constitutes a fundamental challenge to such an account of perceptual experience. In order to respond to this form of skepticism and avoid falling into the Myth, McDowell insists that (7) Perceptual experience brings into play capacities that ‘belong to our rationality’. In his Mind and World (1994), McDowell defended the idea that in order for perceptual experience to bring ‘into play’ capacities that ‘belong to our rationality’, perceptual experiences would have to be understood as containing or making propositional claims. But McDowell has come to think that the content of propositional experience is not correctly thought of as propositional (McDowell 2009a). We’ll also follow McDowell in adopting this requirement.13 My discussion of the Kantian skeptical argument generates a constraint on an adequate account of how perceptually based beliefs can be understood as rational responses to perceptual experiences: (8a) It will have to be the case that ‘capacities that belong to our rationality’ are ‘drawn into play’ by perceptual experiences, although the manner in which perceptual experience draws on rational capacities must not be by way of having propositional content. We also developed, in section 3, a second constraint on an account of perceptual experience adequate to the task of the epistemological disjunctivist: (8b) In order for a perceptual experience to be capable of rationally entitling a perceiver to a perceptually based belief, it must be coherent to understand perceptual experience (at least when all goes well) as capable of ‘revealing’ the world. In a sense, there is no great difficulty in avoiding the Myth of the Given and satisfying explanation (8a). It is, at least currently, a relatively commonplace thought that perceptual experience is, in some way or another,

Kantian Roots of Epistemological Disjunctivism

101

contentful. The core problem lies in simultaneously satisfying both explanations (8a) and (8b). Why? Once we admit the thought that perceptual experience has content, it is plausible to think that the content of Gwendolen’s experience when she is, in fact, looking at the chestnut horse is identical to the content of the experience she would have if it merely seemed to her that she were looking at the chestnut horse. In other words, surely the content of Gwendolen’s perceptual experience in the ‘good’ case is the same as in the ‘bad’ case. But if this is so, it is hard to see how Gwendolen’s perceptual experience in the ‘good’ case can be said to ‘reveal’ the world to her. It can seem as if the representational content is acting as an intermediary, standing between Gwendolen and a direct ‘revelation’ of the world. This worry might move us in the direction of insisting that in order to ‘reveal’ the world, perceptual experience must be understood as non-representational. But this (at least according to McDowell) is simply to fall back into the Myth of the Given. In effect, the task of satisfying the twin constraints (8a) and (8b) is the task of showing how it can be that perceptual experience is both immediate enough to allow us to say that Gwendolen can (in the best kind of case) enjoy a perceptual experience which entitles her to a belief in virtue of ‘revealing’ the world to her, and, simultaneously, contentful enough for Gwendolen’s perceptually based beliefs to be rational responses to those perceptual experiences in a way which avoids the Myth. Put otherwise, the challenge is this: How can it be that the content of Gwendolen’s perceptual experience in the ‘good’ case is, in some sense, the same as the content of her perceptual experience in the ‘bad’ case, and yet her perceptual experience in the ‘good’ case entitles her to her perceptually based belief by ‘revealing’ the world to her, whereas her perceptual experience in the ‘bad’ case does no such thing? I’ve been making the case that, as McDowell understands matters, the epistemological disjunctivist must develop a conception of perceptual experience which meets the twin demands of (8a) and (8b). Can such an account of perceptual experience be found? According to McDowell, we can find an appropriate conception of experience in Kant’s notion of an empirical intuition. The proposal to turn to Kant at this point might seem surprising. For what we need is a form of empiricism which avoids the Myth of the Given. And, according to McDowell, It can be tempting to cast the empiricistic version of the Myth of the Given that is Sellars’s primary target as an interpretation of the familiar Kantian duality of understanding and sensibility, spontaneity and receptivity. In this interpretation, sensory receptivity yields immediately given cognitions. Conceptual capacities, which belong to the spontaneous understanding, come into play only subsequently, in basic empirical judgments, conceived as directly warranted by

102 Thomas Lockhart those immediately given cognitions and in turn warranting the further reaches of a world view. And now it is a way of putting Sellars’s point to say that this picture of how sensibility and understanding cooperate is hopeless. . . . When we conceive the operations of sensory receptivity as prior to and independent of any involvement of conceptual capacities, we debar them from intelligibly standing in rational relations to cases of conceptual activity. We ensure that they could at best be triggers or promptings to bits of conceptual activity, not justifications for them. McDowell (2009g: 92–93) In fact, McDowell claims, “it is a form of the Myth to think sensibility by itself, without any involvement of capacities that belong to our rationality, can make things available for cognition. That coincides with a basic doctrine of Kant” (McDowell 2009g: 93).14

6. Non-Propositional Intuitional Content How then does McDowell’s Kant teach us to think about the content of an empirical intuition? McDowell urges us to “forget much of the philosophical resonance of the English word [‘intuition’]”. An Anschauung (intuition) is “a having in view” (McDowell 2009a: 260). About intuitions, Kant writes: In whatever manner and by whatever means a mode of knowledge may relate to objects, intuition is that through which it is in immediate relation to them, and to which all thought as a means is directed. But intuition takes place only in so far as the object is given to us. This again is only possible, to man at least, in so far as the mind is affected in a certain way. The capacity (receptivity) for receiving representations through the mode in which we are affected by objects, is entitled sensibility. Kant (1929: A19/B33) Drawing on this passage, McDowell’s Kant begins with the idea that in an (empirical) intuition, we have an object sensibly in view. But although it is correct to say that, for McDowell’s Kant, in an empirical intuition an object is sensibly in view, that does not yet tell us how an empirical intuition has content. McDowell leans heavily15 on a ‘Clue’ from Kant: The same function which gives unity to the various representations in a judgment also gives unity to the mere synthesis of various representations in an intuition; and this unity, in its most general expression, we entitle the pure concept of the understanding. Kant (1929: A79/B104–5)

Kantian Roots of Epistemological Disjunctivism

103

Since judgments can be thought of as representations unified by a particular function, we can say that judgments have logical structures. That function which accounts for the unity of a judgment is also responsible for the manner in which representations are combined in an intuition (this different manner of combination Kant calls ‘mere synthesis’).16 Suppose we can figure out how the manner in which representations are unified in a judgment is also responsible for the manner in which representations are synthesized in an intuition. That might help see how capacities that belong to our rationality (capacities involved in judgments) are ‘drawn on’ in experience. So let us start by asking: How are representations unified in a judgment? McDowell switches to the linguistic idiom. Following suit, let us start with the case of assertion. Suppose that Gwendolen has a visual intuition of a chestnut horse and, on the basis of this intuition, asserts (correctly) that the horse is chestnut. Assertion is a means of making “something discursively explicit” (McDowell 2009a: 262). And according to McDowell, “the idea of making things explicit extends without strain to judging. We can say that one makes what one judges explicit to oneself” (Ibid.). One thing McDowell means by this is: suppose Gwendolen were merely to judge, without asserting it, that the horse is chestnut. The content of her judgment would be the same as the content of the corresponding assertion. But by connecting assertion to judgment, McDowell is also drawing our attention to the fact that the content of an assertion and the content of a judgment are articulated. The idea that assertions, being linguistic, are articulated is familiar: Gwendolen’s assertion deploys, for example, a particular predicative expression the ‘significance’ of which is, in some sense, ‘part of’ the content which Gwendolen makes explicit by her assertion. Since judgments have the same content as their corresponding assertions, we can also think of judgments as articulated. Gwendolen’s judgment can, much like her assertion, helpfully be thought of as having ‘parts’ (representations) which are combined (unified) in such a way as to constitute the content of that judgment. For McDowell, this kind of discursive articulation is characteristic of propositional content. We can understand the articulated parts of the content of Gwendolen’s judgment as being associated with conceptual capacities. For example, when Gwendolen judges that the horse is chestnut, she exercises a conceptual capacity associated with the predicate ‘chestnut’. Of course, this is not the only conceptual capacity Gwendolen exercises in making such a judgment. She also exercises, for example, the conceptual capacity associated with ‘horse’. Furthermore, she exercises these conceptual capacities with a very specific kind of “logical togetherness” (McDowell 2009f: 30): she’d exercise the same capacities at the same time if she were to judge that there is a chestnut woodpecker and a bay horse in front of her, but they would not be exercised with the same logical togetherness as in her judgment that the horse is chestnut. In the language of the

104 Thomas Lockhart ‘Clue’, the representations are unified in accordance with a particular function. In order to avoid the Myth, we need it to be the case that capacities which belong to our rationality are ‘drawn on’ in an empirical intuition. The ‘Clue’ suggested that the kinds of capacities which are drawn on in an empirical intuition are those which are unified in a judgment. And I’ve just been arguing that we can think of the kinds of capacities that are unified in a judgment as conceptual capacities. So the natural suggestion is that we can avoid the Myth by articulating a conception of empirical intuition according to which conceptual capacities are drawn on in empirical intuition. The idea that conceptual capacities are ‘drawn on’ in a perceptual experience is perfectly familiar to any reader of McDowell. But we must proceed very carefully at this point to isolate the rather surprising way in which conceptual capacities are involved in perceptual experience. Following McDowell, I said that in order to avoid the Myth, we must find a conception of perceptual experience according to which capacities that belong to our rationality are drawn on in perceptual experience. If conceptual capacities are to meet this constraint, we’ll need to understand conceptual capacities as ‘belonging to our rationality’. But remember that for McDowell, ‘rationality’ for us humans means ‘rationality as such’. That is to say: it is conceptual capacities understood as belonging to our rationality as such that must be drawn on in perceptual experience. McDowell is clear about this point. He writes: the notion of a ‘conceptual capacity’ is governed by a “stipulation: conceptual capacities belong essentially to their possessor’s rationality in the sense I am working with, responsiveness to reasons as such” (McDowell 2009b: 129). In this sense, capacities count as conceptual “only in subjects who can exercise those capacities in reasoning” (McDowell 2009b: 130). One way to draw attention to the strength of McDowell’s view is this. McDowell’s view leaves (logical) room for the possibility of at least three kinds of perceptual experience. First, there might be living beings to whom we wish to attribute no conceptual or rational apparatus whatsoever. If such beings have perceptual experiences, those perceptual experiences will not draw on any conceptual capacities at all. Second, there might be beings (for example, dogs and infants) to whom we are inclined to ascribe rationality, but not rationality as such. To these beings we are free to attribute conceptual capacities, but not conceptual capacities which meet McDowell’s ‘stipulation’. The perceptual experiences of such beings might draw on their conceptual capacities, but they would do so in a different way than perceptual experience draws on the conceptual capacities of those who possess rationality as such. The third kind of perceptual experience is reserved for those beings who possess rationality as such. But how are conceptual capacities drawn on in the perceptual experience of those who possess rationality as such?

Kantian Roots of Epistemological Disjunctivism

105

McDowell’s suggestion is this: a conceptual capacity (in the rich sense governed by his stipulation) is ‘drawn on’ in an intuition when the intuition constitutes a warrant for the deployment, in a judgment, of the content associated with that capacity. In the case of Gwendolen, her empirical intuition ‘draws on’ the conceptual capacity associated with the term ‘chestnut’ in the sense that the intuition entitles Gwendolen to deploy, in a judgment, the content associated with that term.17 But how is intuitional content not propositional? Recall that Kant insists that in an intuition, an object is in view. McDowell expands on this point by saying that the object is “visually present to a subject” “through the presence” of “those of its features that are visible to the subject from her vantage point” (McDowell 2009a: 265). That is, the various representations in an intuition—corresponding to the properties of the object— are “synthesized” with a kind of unity appropriate to an object. To make a judgment simply is to make explicit some of the content united in the intuition (“one makes what one judges explicit to oneself”; McDowell 2009a: 262). Since the horse Gwendolen is looking at is chestnut, then content associated with this concept is part of her visual intuition. To make that content explicit is just to represent, to herself, that she herself is confronted with a chestnut something; which is simply to judge that something chestnut is in view. Kant writes, “It must be possible for the ‘I think’ to accompany all my representations” (B131). In particular, the ‘I think’ can accompany any of the representations in an intuition. To add the ‘I think’ to one of the representations united in an intuition is what McDowell means by making that content explicit (to oneself), and that is just what a judgment consists in: “For the ‘I think’ to accompany some of the content of an intuition, say a visual intuition, of mine is for me to judge that I am visually confronted by an object with such-and-such features” (McDowell 2009a: 266). In making the content explicit in this way, Gwendolen gives the content discursive articulation. In the intuition, it is not so articulated: in order to ‘exploit’ “some content given in an intuition . . . one needs to carve out that content from the intuition’s unarticulated content before one can put it together with other bits of content in discursive activity” (McDowell 2009a: 263–264).18 Some examples might help to clarify this conception of non-propositional intuitional content and its relationship to judgment. According to McDowell, Gwendolen’s visual intuition of a chestnut horse constitutes, for her, a warrant to judge that there is a chestnut horse in view. But she need not make that judgment. Suppose, for example, she has a visual intuition of the chestnut horse but, deep in conversation with Henleigh, does not make any judgment about the horse. The content of her perceptual experience is the same, whether or not she makes an explicit judgment on the basis of that experience. And we can capture that natural thought thus: Gwendolen’s visual intuition constitutes, for her, a warrant to judge that the horse is chestnut whether or not she makes such a judgment. When

106 Thomas Lockhart she makes the judgment, she makes explicit something to which her experience entitles her, but the experience is the same whether she makes the judgment or not. Suppose Gwendolen now turns her attention to a bay horse, but she does not happen to have the discursive capacity required to judge that the horse is bay (for example, she doesn’t know what the word ‘bay’ means). It is not unreasonable, I think, to say that her perceptual experience constitutes a rational entitlement to the judgment that the horse is bay. Here is one consideration in favor of this view: suppose she learns later that day what the word ‘bay’ means. She might then recall: ‘I saw a bay horse this morning’. In such a case, it seems to me correct to say that her belief is rationally based on her perceptual experience of a bay horse. If we take it that the content of a perceptual experience is propositional, it is not obvious how to account for this kind of case. Was the proposition ‘The horse is bay’ part of the content of Gwendolen’s initial perceptual experience even though she lacked, at that time, the relevant discursive conceptual capacity? If not, what propositional content did the original experience have which could rationalize the later judgment? McDowell’s non-propositional account of the content of an empirical intuition better accommodates the case of the bay horse. For McDowell’s Kant, the relevant content is actualized in her intuition, even if she does not have the relevant discursive capacity. For, even if she does not currently possess the relevant discursive capacity, nevertheless the intuition (would) warrant the judgment that the horse is bay. She can acquire the relevant discursive capacity readily enough, and, in so doing, she would simply be exploiting a “potential for discursive activity that is already there in the capacities actualized in having an intuition with that content” (McDowell 2009a: 265).19 There are various other issues it would be desirable to address at this point. For example, is every conceptual capacity which I could acquire and which would accurately report my perceptual experience actualized in my perceptual experience? McDowell’s intriguing answer is that we can draw the line (for visual intuitions) at “proper sensibles of sight and common sensibles accessible to sight” (McDowell 2009a: 260). But we’ll have to leave such questions aside for now.

7. McDowell’s Kant and Epistemological Disjunctivism In section 5, I made much of a particular challenge facing any attempt to articulate a conception of perceptual experience adequate to the task of the epistemological disjunctivist. The challenge was to explain how both of the following can be true: on the one hand, when all goes well, a perceptual experience can entitle Gwendolen to a perceptually based belief in virtue of revealing the world to her; and on the other hand, the content of Gwendolen’s perceptual experience in the ‘good’ case is, in some sense,

Kantian Roots of Epistemological Disjunctivism

107

the same as the content of her perceptual experience in a ‘bad’ case. I want now to show how the account of empirical intuition developed in section 6 can meet that challenge. Suppose first that Gwendolen is in a ‘bad’ case: she is looking at a realistic sculpture of a bay horse illuminated in such a way as to appear chestnut. And suppose further she is unable to tell that she is not confronted by a real chestnut horse. In this bad case, we can say that Gwendolen’s perceptual experience draws on the same conceptual capacities that her perceptual experience would draw on in the good case. The difference between her perceptual experience in the good case and the bad case is this: in the good case, the relevant conceptual capacities are drawn on in such a way that they constitute a warrant for Gwendolen to deploy those capacities discursively in a judgment. In the bad case, the conceptual capacities are drawn on in such a way that it merely seems to Gwendolen that she has a warrant for the discursive deployment of those conceptual capacities. Her perceptual experience ‘draws on’ the same conceptual capacities in both cases, but (to stretch the terminology) defectively so in the bad case.20 We can (and should) reserve the term ‘intuition’ for those cases where Gwendolen’s perceptual experience does indeed constitute, for her, a warrant for a discursive deployment of the capacities drawn on in that intuition. We can apply this point to other skeptical scenarios. Suppose, unbeknownst to Gwendolen, she finds herself in an area which is populated with zebras cunningly disguised to look like chestnut horses. And suppose that Gwendolen cannot tell the difference between a real chestnut horse and a disguised zebra. Suppose now she is in fact looking at a chestnut horse. In such a case, the relevant conceptual capacities are exercised with the same kind of logical togetherness as in the good case. But in this case, the conceptual capacities are not exercised in a way that constitutes, for Gwendolen, an intuition of a chestnut horse. Why not? In a visual intuition of a chestnut horse, the relevant conceptual capacities are drawn upon in such a way as to constitute a warrant for Gwendolen to judge that she is confronted by a chestnut horse. But when she is looking at a (real) chestnut horse in a terrain rife with undetectable disguised zebras, she would not be so warranted. For this reason, the capacity corresponding to the predicate ‘chestnut’ is not drawn upon in the manner characteristic of an intuition. To put the point another way: in this case, she is (through no fault of her own) unable successfully to exercise her capacity to tell, by looking, that there is a chestnut horse in front of her.21 Consider again the case where Gwendolen is confronted with a sculpture of a bay horse illuminated so as to appear chestnut. On this occasion, I’ve claimed that her visual experience ‘draws on’ the conceptual capacities associated with ‘chestnut’ and ‘animal’, but not in the manner in which those capacities would be ‘drawn on’ in an intuition of a chestnut horse. Now Kant says that the ‘I think’ can accompany all of

108 Thomas Lockhart my representations. And we saw that McDowell, applying this claim, writes: “For the ‘I think’ to accompany some of the content of an intuition, say a visual intuition, of mine is for me to judge that I am visually confronted by an object with such-and-such features” (McDowell 2009a: 256; emphasis in original). But couldn’t the ‘I think’ accompany some of the content of the perceptual experience Gwendolen has when she examines the cunningly illuminated bay horse sculpture? And wouldn’t it be natural to say that in such a case she judges (mistakenly, although through no fault of her own) that there is a chestnut horse in view? I think we can answer yes to both of these questions. In the case of the cunningly illuminated sculpture, Gwendolen might well judge that there is a chestnut horse in view. But what, then, is the difference between the two cases? In the good case, Gwendolen has a visual intuition of the chestnut horse. And a visual intuition of a chestnut horse entitles Gwendolen to judge that there is a chestnut horse in view. In the bad case, Gwendolen has no such entitlement, so her judgment will not be warranted (although she may well take it that her judgment is warranted in that case). We can go further. In the ‘good’ case, Gwendolen can know that she is warranted in making her judgment. Why? Since her perceptually based belief is (as McDowell has insisted from the start) a manifestation of her rationality as such, Gwendolen has the capacity, should she wish to exercise it, to consider the question whether her perceptual experience warrants her perceptually based belief. And if she successfully exercises that capacity, she will determine that her perceptual experience does indeed warrant her perceptually based belief. I claimed that McDowell’s Kant is an epistemological disjunctivist. It should now be clear why this is so. We’ve seen that an empirical intuition, for McDowell’s Kant, ‘draws into play’ conceptual capacities. Crucially, the conceptual capacities ‘drawn into play’ by an empirical intuition “belong essentially to their possessor’s rationality . . . as such” (2009b: 129). The conceptual capacity associated with ‘chestnut’ is ‘drawn into play’ by Gwendolen’s visual intuition of the chestnut horse because the intuition gives her a warrant for an explicit deployment of that capacity in a judgment. So a Kantian empirical intuition is, precisely in virtue of the manner in which it ‘draws into play’ conceptual capacities, a perceptual experience such that a belief based on that intuition is a manifestation of rationality as such (to use the language of RMT). Put otherwise, precisely because the conceptual capacities which are drawn into play in an empirical intuition are drawn on in such a way as to constitute an indefeasible warrant for a judgment based on that intuition, McDowell’s Kant is committed to ED. So much for McDowell’s Kant. What of Kant himself? To defend McDowell’s interpretation of Kant (and fully vindicate the claim that Kant was an epistemological disjunctivist), we’d have to find resources in Kant for an especially strong notion of an empirical intuition. It would

Kantian Roots of Epistemological Disjunctivism

109

not be enough to argue that the capacity to have intuitions depends on the capacity to make judgments. One would have to argue first that conceptual capacities can be ‘actualized’ in empirical intuition in a manner which is non-propositional (see, for example, Land 2015a). One would then have to argue that the conceptual capacities non-propositionally actualized in the intuition are to be understood in the sense proper to conceptual capacities which belong to their possessors’ rationality as such. Specifically, the conceptual capacities would have to be actualized in the intuition in a manner which constituted a warrant for an explicit, articulated deployment of the associated concept in a judgment. I do not know whether one could defend this claim about Kant, but it should be obvious that it would be no small feat to do so.

8. Conclusion The central difficulty facing the epistemological disjunctivist, according to McDowell, is to articulate a conception of perceptual experience according to which, when all goes well, perceptual experience rationally (in the sense of ‘rationality as such’) entitles a perceiver to a perceptually based belief by revealing the world to her in a suitable way. In order to articulate a suitable conception of perceptual experience, McDowell turns to the notion of a Kantian empirical intuition. On this account, when Gwendolen enjoys a visual intuition of a chestnut horse, the intuition ‘draws into play’ conceptual capacities by constituting a warrant for the explicit deployment of those capacities in a judgment. This interpretation of empirical intuition is enough to make McDowell’s Kant an epistemological disjunctivist. I leave open the question of whether Kant himself was an epistemological disjunctivist, although I hope to have brought this question into sharper focus.22

Notes 1. See, for example, McDowell (1998, 2009c, 2010, 2011). Another principal proponent of epistemological disjunctivism is Duncan Pritchard (2012, 2015). 2. See, for example, McDowell (2009a, 2009d, 2009e, 2009f, 2009h, 2009i). 3. Pritchard’s characterization of epistemological disjunctivism employs the terms ‘reflectively accessible’ and ‘factive’ (2012: 13). McDowell tends to deploy the term ‘indefeasible’ (2010: 45). For McDowell, an indefeasible warrant is both factive and reflectively accessible (2010: 46). I propose to follow his lead here; although it would make no difference to my argument if we substituted ‘factive and reflectively accessible’ for ‘indefeasible’ in the definition of ED. 4. For this wording, see McDowell (2009b: 127–128). 5. Unless Gwendolen is a philosopher, she is unlikely to ask herself this question. But it is nevertheless a perfectly ordinary question: a non-philosopher might ask the same question by saying, ‘Is that really a horse?’

110 Thomas Lockhart 6. McDowell (2009b). He argues that we are under no obligation to maintain that rationality amounts to the same thing in adult humans, on the one hand, and infants and animals, on the other. Adult humans manifestly do have the ability to respond to reasons as such, and it is (he claims) perfectly plausible to think the capacity to step back and assess is an essential component of the form of rationality they possess. 7. McDowell makes this point (2009c: 230). He refers to Sellars’s discussion of ‘looks’-talk at 1997: sec. III. 8. This section is an attempt to rehearse the central argument of McDowell (2009c). 9. I leave open the question: ‘How much anxiety should the fallibility of human cognitive capacities inspire in us?’ 10. I recognize that it is likely to seem unsatisfactory, but this is not the place to take this argument further. For more discussion, see Pritchard (2012, 2015), Schönbaumsfeld (2013). 11. The terminology here is that of Conant (2012). 12. For an articulation of the Myth and its relationship to ‘Kantian skepticism’, see Conant (2012). For a different, but illuminating, approach to these issues, see Byrne (2016). 13. For now I’ll have to be stipulative about this. I’ll be in a position to say something more about the advantages of the non-propositional account in section 6. 14. The question of whether this is so, that is, the question of how—if at all— concepts are involved in Kantian intuitions is a lively topic in recent Kant scholarship. For non-conceptualist accounts, see, for example, Hanna (2008, 2011) and McLear (2014, 2016). For conceptualist accounts, see, for example, Bowman (2011), Bauer (2012), and Land (2011, 2015a, 2015b). 15. For example, at McDowell (2009f: 30–31; 2009d: 70; 2009g: 94). 16. See especially McDowell (2009g: 94). 17. We could say that Gwendolen has the capacity to know, by looking, that something chestnut is in front of her. 18. McDowell claims that adding the ‘I think’ to the content of an intuition (and thereby making a judgment) is what Kant means by ‘bringing the content to the unity of apperception’: “I find that a judgment is nothing but the manner in which given [cognitions] are brought to the objective unity of apperception. This is what is intended by the copula ‘is’. It is employed to distinguish the objective unity of given representations from the subjective.” (B141). In quoting this passage, McDowell (2009d: 70) substitutes ‘cognitions’ for ‘modes of knowledge’. 19. There are other considerations which speak in favor of a non-propositional account of the content of perceptual experience. Suppose the content of Gwendolen’s perceptual experience of the chestnut horse is the proposition that the horse is chestnut. What is the relationship between this experience and her belief that the horse is chestnut? Standardly, the relationship between two items both of which have propositional content is inferential. But this cannot be acceptable, for one reason: there is a tight connection between seeing that the horse is chestnut and knowing that the horse is chestnut. If the relationship between the experience and the belief is merely that of inference, there is a risk of severing that tight connection. For another reason, McDowell is perfectly clear that perceptually based knowledge must be arrived at, for the empiricist, non-inferentially (McDowell 2009j: 223). 20. McDowell does not quite put the point this way. His initial instinct was to adopt the view that there is an ‘illusion of content’ in ‘bad’ cases (McDowell

Kantian Roots of Epistemological Disjunctivism

111

2009e: 48–49). He has since explicitly rejected this (McDowell 2013: 156). The content, he observes, can be the same in the good case and the bad case. We might put the alternative I have described by saying that, in the bad case, Gwendolen is having an illusion of an intuition. 21. This case is structurally parallel to barn façade cases. McDowell sometimes seems to write as if having the object visually present to one is sufficient to secure an indefeasible warrant. But, plausibly, in the case at hand Gwendolen has the chestnut horse visually present to her but lacks an indefeasible warrant. We need something more than mere visual presence. McDowell is aware of this point. He comes closest to handling the issue in the manner of this paragraph in McDowell (2013: 152). 22. Thanks to Jennifer Lockhart, Arata Hamawaki, and the editors of this volume for helpful discussions of this material.

References Bauer, Nathan, 2012, A peculiar intuition: Kant’s conceptualist account of perception, Inquiry 55(3): 215–237. Bowman, Brady, 2011, A conceptualist reply to Hanna’s Kantian non-conceptualism, International Journal of Philosophical Studies 19(3): 417–446. Byrne, Alex, 2016, The epistemic significance of experience, Philosophical Studies 173: 947–967. Conant, James, 2012, Two varieties of skepticism, in Günter Abel and James Conant, eds. Rethinking Epistemology (vol. 2). Berlin: Walter De Gruyter, pp. 1–73. Hanna, Robert, 2008, Kantian non-conceptualism, Philosophical Studies 137(1): 41–64. ———, 2011, Beyond the myth of the myth: A Kantian theory of non-conceptual content, International Journal of Philosophical Studies 19(3): 323–398. Kant, Immanuel, 1929, Critique of Pure Reason. Trans. N. Kemp Smith. Basingstoke: Palgrave. Land, Thomas, 2012, Kantian conceptualism, in Günter Abel and James Conant, eds. Rethinking Epistemology (vol. 1). Berlin: Walter de Gruyter, pp. 197–239. ———, 2015a, No other use than in judgment?: Kant on concepts and sensible synthesis, Journal of the History of Philosophy 53(3): 461–484. ———, 2015b, Nonconceptualist readings of Kant and the transcendental deduction, Kantian Review 20(1): 25–51. McDowell, John, 1994, Mind and World. Cambridge, MA: Harvard University Press. ———, 1998, Knowledge and the internal, in his Meaning, Knowledge and Reality. Cambridge, MA: Harvard University Press. ———, 2009a, Avoiding the myth of the given, in his Having the World in View. Cambridge, MA: Harvard University Press, pp. 256–272. ———, 2009b, Conceptual capacities in perception, in his Having the World in View. Cambridge, MA: Harvard University Press, pp. 127–144. ———, 2009c, The disjunctive conception of experience as material for a transcendental argument, in his The Engaged Intellect. Cambridge, MA: Harvard University Press. ———, 2009d, Hegel’s idealism as radicalization of Kant, in his Having the World in View. Cambridge, MA: Harvard University Press, pp. 69–89.

112 Thomas Lockhart ———, 2009e, Intentionality as a relation, in his Having the World in View. Cambridge, MA: Harvard University Press, pp. 44–65. ———, 2009f, The logical form of an intuition, in his Having the World in View. Cambridge, MA: Harvard University Press, pp. 23–43. ———, 2009g, Self-determining subjectivity and external constraint, in his Having the World in View. Cambridge, MA: Harvard University Press, pp. 90–107. ———, 2009h, Sellars on perceptual experience, in his Having the World in View. Cambridge, MA: Harvard University Press, pp. 3–22. ———, 2009i, Sensory consciousness in Kant and Sellars, in his Having the World in View. Cambridge, MA: Harvard University Press, pp. 108–126. ———, 2009j, Why is Sellars’s essay called “Empiricism and the philosophy of mind”? in his Having the World in View. Cambridge, MA: Harvard University Press, pp. 221–238. ———, 2010, Tyler Burge on disjunctivism, Philosophical Explorations 13: 243–255. ———, 2011, Perception as a Capacity for Knowledge. Milwaukee, WI: Marquette University Press. ———, 2013, Perceptual experience: Both relational and contentful, European Journal of Philosophy 21(1): 144–157. McLear, Colin, 2014, The Kantian (non)-conceptualism debate, Philosophy Compass 9(11): 769–790. ———, 2016, Kant on perceptual content, Mind 125: 95–144. Pritchard, Duncan, 2012, Epistemological Disjunctivism. Oxford: Oxford University Press. ———, 2015, Epistemic Angst: Radical Skepticism and the Groundlessness of Our Believing. Princeton, NJ: Princeton University Press. Schönbaumsfeld, Genia, 2013, McDowellian neo-Mooreanism? International Journal for the Study of Skepticism 3: 202–217. Sellars, Wilfred, 1997, Empiricism and the Philosophy of Mind. Cambridge, MA: Harvard University Press.

6

Was Wittgenstein a Disjunctivist Avant la Lettre? Genia Schönbaumsfeld

1. Introduction For the purposes of this chapter, I take disjunctivism to be the view that perceptual knowledge of the world is possible—that is to say, that if we are in the good case, then perception will apprise us of the facts. Disjunctivists, in other words, reject what in The Illusion of Doubt I call the Reasons Identity Thesis (RIT): the thought that in both the good case and the bad case, our perceptual reasons are the same.1 It is this thesis, widely accepted in contemporary epistemology,2 that pushes us towards scepticism and the thought that knowledge of the ‘external world’ is never possible, even in the good case. Now, on the face of it, the denial of such a thesis, disjunctivism, looks like a substantial epistemological thesis of its own. But this appearance is deceptive, as disjunctivism is really only the philosophical name for our ordinary conception of what perception can achieve—namely, yield knowledge of the world in the good case. It only looks like a substantial philosophical theory, because a bad picture—the Cartesian picture of our evidential situation—has turned our ordinary, pre-theoretical conception into something that seems dubious and epistemologically inflated. To someone who rejects this picture and the Reasons Identity Thesis, on the other hand, disjunctivism is just the name philosophers give to our ordinary way of thinking about the world. When I raise the question, therefore, whether Wittgenstein was a disjunctivist (or would espouse disjunctivism if he were still alive today), I am not thereby asking whether Wittgenstein would accept a substantial theoretical position in epistemology. Rather, I am asking whether Wittgenstein would accept our ordinary, pre-theoretical notion that we have perceptual access to the world, which, in the good case, yields knowledge of ‘externally’ existing physical objects. It would be odd, given that Wittgenstein thinks that philosophy ‘leaves everything as it is’ (Philosophical Investigations §124), if he didn’t. Wittgenstein once said, ‘not empiricism and yet realism in philosophy, that is the hardest thing’ (Remarks on the Foundations of Mathematics,

114

Genia Schönbaumsfeld

Part VI, §23). So, one might call the conception that I believe one can find in Wittgenstein’s writings ‘realism without empiricism’,3 and this, I will show in this chapter, is a form of disjunctivism. If my reading is correct, Wittgenstein not only was one of the first proponents of a disjunctivist view; he also gives us good reasons to resist its converse, the Reasons Identity Thesis, which one might call the fourth or last dogma of empiricism.4

2. The Reasons Identity Thesis and the Cartesian Picture In order to get a better grip on what ‘realism without empiricism’ might be, let us start by asking the question, ‘what is realism with empiricism’? Here is a possible answer: any view that says that our contact with the world is either representationally or inferentially mediated, and that the world might therefore be both metaphysically and epistemically inaccessible to us. Such a conception is, of course, part and parcel of the Cartesian picture according to which knowledge of the so-called ‘external world’5 is always ‘indirect’ in the sense that it depends on an inference from how things subjectively seem to one (i.e. on an inference from one’s mental states or ‘sense-experiences’) to how things actually are ‘out there’ in the ‘external world’.6 For, as Michael Williams perceptively notes, on this view, ‘external’ does not mean ‘in one’s surroundings’; it rather means (in Berkeley’s phrase) ‘without the mind’: ‘And what is within the mind is what is “immediately known” by virtue of its presence to consciousness’ (Williams 2011: 58). Once this contrast between the immediately known and the inferentially known is firmly in place, ‘we are in a position to imagine that, however disorderly when viewed thematically, beliefs about external things share an underlying epistemology, in virtue of their uniform dependence on “evidence of the senses”’ (Ibid.). Such a picture leaves one in the most dire epistemological straits, of course, for it thrusts, among many other unwelcome things, the Reasons Identity Thesis upon one. For if one’s epistemological starting-point cannot be anything other than ‘purely sensory knowledge’7 which always falls short of providing one with knowledge of an ‘externally existing’ world, then it is only ever possible to come to know something about the world ‘outside of one’s own mind’, if a proof can first be given that one is not confronted by mere appearance. In other words, if we espouse the Cartesian picture, then we get the view that a proof of the ‘external world’s’ existence is necessary before any of our perceptual claims can be warranted (see, for example, Wright 2008; Coliva 2014, 2015), as we can only ever have knowledge of how things appear to us, never of how they are (even in the best possible case). On such a conception, that is, my epistemic standing, even in the good case, can never be better than what I have in the bad case, and hence, whatever case I happen to find myself

Was Wittgenstein a Disjunctivist?

115

in, my perceptual reasons are always, subjectively speaking, the same (the Reasons Identity Thesis (RIT)).8 In what follows, I will show that Wittgenstein rejects such a view,9 and why.

3. Knowledge and Radical Doubt On Certainty, Wittgenstein’s final collection of remarks, is most relevant to the question of whether Wittgenstein would accept disjunctivism. In this work, in which Wittgenstein is primarily in dialogue with G. E. Moore (1959), Wittgenstein examines the concepts of knowledge, certainty and doubt, and raises the question whether radical scepticism—the hypothesis that there might be no ‘external world’—is, ultimately, fully intelligible. To this end, Wittgenstein seeks, inter alia, to establish three things: 1. The idealist’s or sceptic’s doubts are not practical doubts. 2. A reasonable person will not doubt that I know that I have hands. 3. Radical scepticism is an illusion. All three ideas can already be found in the following two passages occurring towards the beginning of On Certainty (OC): ‘I know’ often means: I have the proper grounds for my statement. So if the other person is acquainted with the language-game, he would admit that I know. The other, if he is acquainted with the languagegame, must be able to imagine how one may know something of the kind. The statement ‘I know that here is a hand’ may then be continued: ‘for it’s my hand that I’m looking at’. Then a reasonable man will not doubt that I know.—Nor will the idealist; rather he will say that he was not dealing with the practical doubt that is being dismissed, but that there is a further doubt behind that one.—That this is an illusion has to be shown in a different way. (On Certainty §18–19; emphasis in original) Radical scepticism is the thought that all our perceptual beliefs taken together could be false—i.e. could fail to ‘match’ the way the world is (if, indeed, there is one). That is to say, even in paradigmatic cases of perceptual knowledge, such as when I’m looking at my hands and nothing out of the ordinary is occurring, I might be wrong. One way of countering the sceptic’s proposal would be to insist, with Moore, that in such cases I do know that I’ve got hands. But Wittgenstein thinks that this is actually not a good tactic: ‘Moore’s mistake lies in this— countering the assertion that one cannot know that, by saying “I do know it”’ (OC §521).

116

Genia Schönbaumsfeld

Why is it not a good tactic, though (apart from perhaps being dialectically ineffective)? Presumably because the sceptic (or idealist)10 would say that he is not dealing with the practical doubt that Moore seems to be dismissing—for what else could he be doing by holding up his hands?— but that there is a further doubt behind the practical one. That this is an illusion, Wittgenstein believes, has to be shown in a different way: one cannot refute or deny an illusion, only dissolve or remove it, for only a coherent claim can be refuted, and an illusion is only apparently intelligible. In other words, it’s not that Wittgenstein thinks that Moore is wrong about knowing the things he claims to know; it’s rather that insisting on one’s knowledge in this context is not going to remove the illusion under which the sceptic is labouring. Hence, doubt first needs to be cast on the sceptic’s doubt by asking what doubting that one’s got hands, in ordinary circumstances where normal grounds for doubt are absent, could so much as mean. Wittgenstein says: If a blind man were to ask me ‘Have you got two hands?’ I should not make sure by looking. If I were to have any doubt of it, then I don’t know why I should trust my eyes. For why shouldn’t I test my eyes by looking to find out whether I see my two hands? What is to be tested by what? (OC §125) If the existence of my hands is genuinely in doubt, in other words, then I cannot rely on any ordinary way of checking that a physical object is in front of me, say, by looking more closely, as I cannot, in such radical circumstances, trust my eyes, either. Nothing that we would ordinarily call ‘evidence’ can help us here.11 For this reason, the sceptic’s doubt is quite unlike the following, ‘ordinary’ doubt: I visit someone in hospital who has been in an accident and his whole body is covered in bandages. In such a case, I might not know whether this person still has hands, for they might have been amputated. If the poor soul reassures me by saying ‘I know I have hands’, then I will take this to imply that he has been able to check (say by having had the bandages removed by a doctor earlier in the day).12 Here the claim ‘I know I have hands’ is perfectly intelligible, since we also have a clear idea of what it would be like not to know that one has hands (say, by having failed to check). None of these conditions are met, however, in the case of what one might call ‘radical’ doubt. If I assert in ordinary circumstances where nothing unusual (such as accidents, etc.) has occurred, that I know that I have hands (in order to counter a radical sceptical claim, say), it is unclear what sort of justification I could offer for this. That is to say, if, in such a context, I could nevertheless be ‘wrong’ about having hands, then this would be no ordinary ‘mistake’, but would rather constitute what Wittgenstein calls ‘an annihilation of all yardsticks’:

Was Wittgenstein a Disjunctivist?

117

What if it seemed to turn out that what until now has seemed immune to doubt was a false assumption? Would I react as I do when a belief has proved to be false? Or would it seem to knock from under my feet the ground on which I stand in making any judgments at all? . . . Would I simply say ‘I should never have thought it!’—or would I (have to) refuse to revise my judgment—because such a ‘revision’ would amount to an annihilation of all yardsticks? (OC §492) What Wittgenstein seems to be saying here is that while a mistake does not call into question the intelligibility of the entire practice in which one is participating—but only one’s own competence—the notion that I could be ‘wrong’ about the most fundamental things undermines the coherence of the very practices on which my expression of doubt at the same time depends (since if there were no such practice, there would be no such doubt). Consequently, there is something very odd about the radical sceptic’s doubt. Not only is it not a practical doubt, it is not a particular one, either. Rather, it is a ‘global’ doubt, because if I doubt here, then everything else I have ever believed or hitherto taken for granted is plunged into chaos as well. So, one might begin to suspect that the radical sceptic is really after not a demonstration that my hands (or any other particular physical objects) exist, but rather a proof that the ‘external world’ as such exists—that all of our epistemic practices taken together manage to ‘latch on’ to something ‘outside’ of us. Contrary to what Moore is trying to do, however—namely, to play the same game as the sceptic, but to come down on the opposite side—Wittgenstein thinks that what we need to show is that this kind of ‘global validation’ demand is not a coherent idea. There is no such thing, nor could there be such a thing, as a global validation of all our epistemic practices taken together, and, consequently, its absence is neither lamentable nor an epistemic limitation.13 For, as Wittgenstein goes on to show, the sceptic’s ‘radical’ doubt is not a doubt, and, hence, the validation demand is not just incoherent, but obsolete, too: ‘if you tried to doubt everything you would not get as far as doubting anything. The game of doubting itself presupposes certainty’ (OC §115). Since if we were to suppose, with the radical sceptic, that all of our ‘external world’ propositions taken together could be false, then the words we use to formulate this ‘doubt’, would lose their meaning, too: ‘If you are not certain of any fact, you cannot be certain of the meaning of your words either’ (OC §114).14 That your words mean what they do is also a fact about them. So, if everything is uncertain, as the sceptic supposes, then ‘meaning’—if indeed there could still be such a thing— would no longer be stable, either. Consequently, it’s not just that scepticism is unstateable—but might, for all that, be true15—it is rather that,

118

Genia Schönbaumsfeld

if Wittgenstein is right, it is no longer clear what the sceptic’s ‘thought’ can even amount to (i.e. this is not merely a ‘semantic’ point). For, if, for example, it were conceivable that I were falsely ascribing pain to myself all the time, then it is just no longer clear what ‘pain’ really means.16 But if the meaning of ‘pain’ is under threat, what, then, could it mean to say, with the sceptic, that ‘I doubt that there is such a thing as pain’? So, it seems that the kind of justification that the sceptic is after—a fully general evaluation of our entire rational or epistemic system—is impossible to provide. But this is not because human powers are unequal to the task. Rather, Wittgenstein thinks, the very idea is misguided: ‘So is the hypothesis possible that all the things around us don’t exist? Would that not be like the hypothesis of our having miscalculated in all our calculations?’ (OC §55). For a ‘miscalculation’ in all our calculations is not, as it were, just an ‘aggregated’ mistake; rather, it implies that we have never calculated at all—that nothing that we have ever done counts as an instance of calculating. But, if so, the sceptic owes us an explanation of what ‘calculation’ really means, and of what it is, that we allegedly cannot do (or know). In the absence of such an explanation, we need not accept the sceptic’s challenge that unless we can show in advance that we are not radically mistaken about everything, we aren’t entitled to our ordinary knowledge claims. Doubt, in other words, needs a context, a common human background that stands fast and against which specific epistemic claims can be assessed. A ‘global’ doubt is a paper doubt, as it tries to doubt everything at the same time and in this way fails to doubt at all. As Wittgenstein says, I did not get my picture of the world by satisfying myself of its correctness; nor do I have it because I am satisfied of its correctness. No: it is the inherited background against which I distinguish between [what is] true and [what is] false. (OC §94) Hence, it is not that I should have satisfied myself of its correctness, and did not do so because I was either too rash, too lazy or too gullible. Rather, with respect to our whole rational system, there is no such thing as satisfying oneself of its correctness (or not doing so), since, as Williams (2011) has shown, so-called ‘knowledge of the external world’ is not a natural kind. Instead, what we mean by this term is an amalgamation of all sorts of different beliefs, methods and practices. Consequently, there is no unitary ‘entity’ here that one could so much as stand in an epistemic relation to.17 In this much, Wittgenstein’s own claim, ‘the difficulty is to realize the groundlessness of our believing’ (OC §166), is not felicitously expressed. For ‘groundlessness’ is itself an epistemic notion which implies that there could—or ought to—be grounds, although in fact there are none; whereas Wittgenstein is really trying to say that it is logically

Was Wittgenstein a Disjunctivist?

119

impossible for there to be any, and that the absence of the possibility of ‘global validation’ is therefore not to be viewed as a lack or as an epistemic shortcoming. So, it’s not that we groundlessly need to accept that we know that we have hands, for example; it is rather that we have no clear idea of what it might mean not to ‘accept’ it. This is why Wittgenstein says that if, in ordinary circumstances, Moore were to pronounce the opposite of those propositions which he declares certain, ‘we should not just not share his opinion: we should regard him as demented’ (OC §155).18 If this were an ‘ordinary’ case of not knowing, where one could cite evidence, then one could just not share Moore’s opinion. If this is correct, then we don’t stand in an epistemic relation— groundless or otherwise—to what has come to be known as ‘hinge propositions’: those propositions that are exempt from doubt and that constitute the ‘hinges’ around which our enquiries turn (OC §341–343). Rather, ‘their role is like that of rules of a game; and the game can be learned purely practically, without learning any explicit rules’ (OC §95). That is to say, propositions expressing our hinge commitments may look like ordinary empirical claims (and so one might try to assert or refute them like Moore or the sceptic), but that appearance is deceptive, for their role is ‘descriptive of the language-game’ (OC §56), and in this much ‘logical’ or ‘technique-constituting’. So one might say that ‘hinge propositions’ are an attempt to articulate the logical enabling conditions that allow our epistemic practices to operate, and without which even our words could not mean anything.19 But if this is so, then ‘hinge propositions’ cannot be in the market for knowledge, and hence it neither makes sense to assert nor to deny that one ‘knows’ them. At best, and as Wittgenstein says, such propositions can express a ‘logical’ (or ‘grammatical’) insight: ‘I know I have two hands’, asserted in ordinary circumstances, really means ‘there is no such thing as a doubt in this case’—it does not make sense as a Moorean antisceptical claim. In order not to get confused between what Wittgenstein calls a ‘grammatical’ or ‘logical’ sense of ‘to know’ and ordinary uses of these terms, I therefore propose to distinguish between a ‘logical’ and an ‘epistemic’ sense of ‘to know’. For example, an unproblematic—i.e. straightforwardly ‘epistemic’—employment of ‘I know I have two hands’ would be in the context of the accident scenario discussed earlier. Here, the claim ‘I know I have hands’ makes sense, since we also have a clear idea of what it would be like not to know that one has hands. Furthermore, if one does know, one is able to offer justification for one’s knowledge claim. But as we have already seen, neither of these two conditions are satisfied in the radical sceptical context (i.e. in ordinary circumstances where radical sceptical error-possibilities have been raised), and so one cannot use ‘I know I have two hands’ in order to counter such a scenario. At best,

120

Genia Schönbaumsfeld

‘I know’ can have a ‘logical’ use here, informing one’s interlocutor that a doubt makes no sense in this context; that it is ‘logically’ excluded (OC §194).20 This is the important insight that Wittgenstein believes Moore misses and that constitutes the mistake that we quoted Wittgenstein as identifying earlier. Moore is, in effect, conflating the ‘logical’ and ‘epistemic’ senses of ‘to know’ in the misguided belief that this verb can tolerate metaphysical emphasis (OC §482) and that one can use ‘I know I have two hands’ to refute the sceptic. In spite of this criticism, however, Wittgenstein believes that there is something right that Moore is perhaps inchoately gesturing towards: ‘So one might grant that Moore was right, if he is interpreted like this: a proposition saying that here is a physical object may have the same logical status as one saying that here is a red patch’ (OC §53). In other words, Wittgenstein thinks that Moore’s ‘proof’ is ineptly trying to express the thought that physical object propositions and propositions about experiential seemings can be cognitively on a par: we can have the same kind of ‘direct access’ to physical objects as is commonly thought, on the Cartesian (or empiricist) picture, to accrue only to one’s experiential seemings (or ‘sense data’). In normal circumstances (i.e. in the good case where nothing is interfering with my perceptual and cognitive capacities), there is no epistemic ‘gap’ between perceptual awareness of my hands and ‘knowledge’ of them that could be ‘bridged’ by evidential grounds. I am directly (i.e. non-inferentially) aware of my hand rather than of ‘handlike’ seemings, appearances or sense-data, from which I would have to infer the presence of my hands.21 If this constitutes the neo-Moorean insight that Wittgenstein is endorsing, it implies that Wittgenstein rejects the Reasons Identity Thesis, since this takes our epistemic starting-point as always consisting of an experience ‘as of something’s being the case’, never of a factive experience (i.e. an experience that entails the truth of what is perceived). Given that most proposals in the contemporary epistemological literature—be they ‘liberal’,22 ‘conservative’23 or ‘moderate’24— grant this starting-point, however, all these offerings concede to the sceptic that an independent warrant for the proposition ‘there is an external world’ is in principle required before one is able to take one’s good-case perceptual experiences at face value.25 But the idea that we need independently—and without any help from perception—to establish that ‘there is an external world’ only makes sense if there is an ‘internal’ world to contrast it with; that is to say, if we believe that claims about our mental states or ‘sensory experiences’ are somehow epistemologically prior to claims about ordinary physical objects. If Wittgenstein is right, however, that radical doubt is an illusion, then the notion that we either first need to establish that there is an external world, or else groundlessly take it ‘on trust’ (or ‘assume’) that there is, goes out of the window, too.

Was Wittgenstein a Disjunctivist?

121

For this reason, Wittgenstein says that ‘there are physical objects’ is nonsense (OC §35). For if we cannot ultimately make sense of the notion that there could be no physical objects—just ‘sensory experiences’—then we cannot coherently ‘prove’ (or ‘assert’) that there are, either.26 Rather, and as Wittgenstein says, ‘A is a physical object’ is a piece of instruction which we give only to someone who doesn’t yet understand either what ‘A’ means, or what ‘physical object’ means. Thus it is instruction about the use of words, and ‘physical object’ is a logical concept. (OC §36) That is to say, the concept ‘physical object’ belongs to our epistemic (or logico-grammatical) scaffolding. It is not an empirical concept that could be up for grabs,27 as there are no empirical means of finding out that the world consists only of experiential seemings, say. Consequently, one cannot assert that ‘there are physical objects’, as this would constitute an attempt to ‘ground’ the ‘background’ (an attempt to assert a logical or grammatical truth as if it were an ordinary empirical claim), and any such attempt will itself presuppose the very background in question:28 ‘If the true is what is grounded, then the ground is not true, nor yet false’ (OC §205). This will seem threatening only if we continue to believe, as foundationalists, for example, do, that our relation to the world ‘as a whole’ must be descriptive and epistemic all the way down. But if Wittgenstein is right, this is wrong-headed. For just as a game, in order to be a game, must have rules distinct from the moves that they make possible, so our practices require an unquestioned background in order to function. It is for this reason that Wittgenstein uses the famous ‘riverbed’ metaphor: The mythology[29] may change back into a state of flux, the river-bed of thoughts may shift. But I distinguish between the movement of the waters on the river-bed and the shift of the bed itself; though there is not a sharp division of the one from the other. (OC §97) The problem with radical scepticism and traditional attempts to refute it is that they don’t take on board this fundamental distinction, and so the temptation to think that one could, or should, have the movement of the waters on its own continues to persist. If Wittgenstein is right, however, that a ‘global’ doubt is not a doubt, then there is no radical sceptical problem to answer. Furthermore, it is a prior endorsement of the Cartesian picture itself which makes radical scepticism seem unavoidable, as this grants to the sceptic that one can know the content of one’s sense-experiences even though these

122

Genia Schönbaumsfeld

experiences as a class may never be experiences of an ‘external’ world at all.30 So, once the plausibility of this picture is undermined, we can stop conceiving of our epistemic situation in terms of Wright’s I-II-III31 argument and instead start with the notion that our good-case perceptual experiences can be taken at face value. But, if so, then these (factive) experiences will, in the good case, themselves provide conclusive perceptual warrant for the propositions in question,32 and, hence, no prior proof of the ‘external world’s’ existence is required.

4. Implications for Closure At this point in the discussion, problems about ‘closure’ start to loom. In the words of Michael Williams: To say that our relation to framework judgements [hinge propositions] is non-epistemic is to say that they neither amount nor fail to amount to knowledge. But surely they can be entailed by judgements to which we do have an epistemic relation, in which case knowledge fails to close under known entailment. (1996: 187) Most philosophers who have considered Wittgenstein’s On Certainty think that Wittgenstein must deny closure, for the reasons Williams gives.33 Since most epistemologists don’t exactly welcome this implication,34 however, some form of damage limitation seems to be required. Williams himself, for example, argues for the instability of knowledge under the conditions of reflection, while Wright (2004) has famously defended a non-evidential warrant—or ‘entitlement’—strategy for what he calls ‘cornerstones’.35 Pritchard (2014, 2015), on the other hand, believes that the way out of this predicament is to recognise that commitment to hinges is non-optional, and specifically not to be construed as ordinary belief (2014: 208). Given that a hinge commitment is necessarily immune to rational considerations, it is best not thought of as a ‘belief’ at all, even though it does share some salient characteristics with belief, notably, full commitment to the target proposition (Ibid.: 202). This means that if one cannot acquire a belief in a hinge proposition on the basis of a belief-forming process like competent deduction, let alone base one’s beliefs on the evidential outcome of such a process, one can fail to know the propositions that express one’s hinge commitments and still not contravene the closure principle36 (Ibid.: 209). Of the proposals currently on offer, this strikes me as the most plausible. But we can do even better if we bear in mind the distinction I have drawn between ‘logical’ and ‘epistemic’ senses of ‘to know’. For if I am right and ‘I know I have two hands’ does not, in ‘global’ sceptical contexts, have an epistemic sense, then ‘closure’-based sceptical arguments

Was Wittgenstein a Disjunctivist?

123

are impossible to construct altogether. Since if Moore’s insistence against the sceptic that he knows that he has two hands is really used to mean something like ‘I know I’m not radically mistaken’ (or ‘I know I can’t be wrong all the time’), and we substitute this into a ‘closure’-based sceptical argument, we get the following: (BIV1) If I know I have two hands (read ‘if I know that I’m not radically mistaken’), then I know I’m not a BIV. (BIV2) I don’t know I’m not a BIV. (BIV3) I don’t know I have two hands (read ‘I know that I’m not radically mistaken’). But since, as everyone would agree, the BIV scenario is just colourful shorthand for the thought that I am radically mistaken, the closure-based sceptical argument now reads: (BIV1*) If I know I’m not radically mistaken, then I know I’m not radically mistaken. (BIV2*) I don’t know I’m not radically mistaken. (BIV3*) I don’t know I’m not radically mistaken. This, of course, isn’t an argument, but a tautology followed by a (two) bald assertion(s). So, it seems that, on my interpretation of Wittgenstein’s conception, there can be no such thing as a closure-based sceptical argument: the appearance that there might be one is generated purely by the confusion that ‘I know I have two hands’ functions like an ordinary empirical proposition in radical sceptical contexts (i.e. has an epistemic sense).37 If, for the reasons Wittgenstein gives, it doesn’t, however, then Pritchard’s attempt to preserve the closure principle by arguing that hinge commitments cannot be acquired through competent deduction is unnecessary (even if true). For without a closure-based sceptical argument, the closure principle simply has no application in radical (‘global’) sceptical contexts.38,39 Given that Wittgenstein thinks that radical scepticism is an illusion, this conclusion should not surprise us.

5. Conclusion I have tried to show in this chapter that Wittgenstein would accept the ordinary notion that we perceive physical objects and can thereby come to know about them, but that this is not a theoretical view requiring substantial epistemological commitments. Rather, what makes it seem that way is an illusion of doubt which suggests that we can never know anything about the world ‘outside our own minds’ unless we can first demonstrate—without any help from perception—that the ‘external world’ exists. If my interpretation of Wittgenstein is correct, however,

124

Genia Schönbaumsfeld

he regards such an idea as fundamentally misguided. We cannot, for the reasons given in this chapter, make sense of the thought that everything around us might not exist or that all propositions about physical objects are false. Since Wittgenstein’s views have often, in the past, been associated with forms of anti-realism (Dummett 1978; Wright 1980; Boghossian 2007), it is worth pointing out, by way of concluding, that the conception that I claim one can find in Wittgenstein’s work—realism without empiricism—is a modest conception and not to be confused with what is sometimes called ‘philosophical realism’ or ‘metaphysical realism’. But neither is it to be conflated with idealism or, pace Coliva (2014, 2015), with a form of ‘internal rationalism’ that holds that assuming the existence of the external world is constitutive of rationality. For, if I am right, Wittgenstein would regard the idea that one could assume the external world’s existence as equally wrong-headed as trying to demonstrate its existence. If doubting that anything around us doesn’t exist isn’t, as Wittgenstein tries to show, fully intelligible, then neither is assuming that it exists. For what could this so much as mean? That I believe that tables and chairs exist, for example, but that I can’t be sure? How would this absence of certainty manifest itself? It seems that either it wouldn’t or it would be a form of madness. That is to say, the ‘difference’ between someone who is sure and someone who only ‘assumes’ would either be one of ‘battle cry’ (see Zettel §414) only and not otherwise discernible in behaviour, or, if it did manifest in behaviour, for example, if the person in question kept touching objects in order to assure himself of their existence, then we would have to conclude that he’s lost his marbles (OC §155). If this is right, then Wittgenstein has provided us with powerful reasons to resist the Reasons Identity Thesis and the Cartesian picture, which stand in the way of accepting the disjunctivist thought that good case perception is factive and gives us knowledge of the world around us (i.e. of our environment). For, if a global doubt is no longer intelligible, then the most that can threaten us is a ‘local’ sceptical scenario—such as Dretske’s zebra example—but local sceptical cases can, in principle, be answered by appeal to background knowledge or by making further, special checks.40 That is to say, without the notion that there might be no ‘external world’ hovering in the background, we can say to the person who wants to know how we can know that we see a zebra in front of us and not a cleverly disguised mule, that zoos don’t generally go to the trouble of producing credible fakes, that people who commit fraud go to jail, etc. Furthermore, and if need be, one could call in a zoologist who could perform special checks in order to ascertain that we are indeed confronted by a real, live zebra and not a fastidiously painted mule. In a nutshell, if Wittgenstein is right, ordinary doubts (even outlandish ones) can be answered, whereas ‘extraordinary’ doubts make no sense.

Was Wittgenstein a Disjunctivist?

125

Notes 1. McDowell (1994, 1998a, 1998b) calls this the ‘highest common factor conception’ (HCF). 2. See, for example, Bennett (1971), Pollock (1974), McGinn (1984), Stroud (1984), Nagel (1986), Pryor (2000), Wright (2002, 2014), Burge (2003), Conee (2007), Coliva (2012), White (2014). 3. See chapter 5 of The Illusion of Doubt. See also Diamond (1991). 4. See chapter 5 of The Illusion of Doubt. 5. I have put quotation marks around ‘external world’, as I am challenging the Cartesian idea that we could be locked into an ‘internal’ world of appearances, from the perspective of which the existence of an ‘external’ world could be doubtful. 6. These inferences need not be conscious, but the relevant point is that one is starting from an ‘inner mental realm’ of experience from which one must try to work out ‘what is going on out there in the external world’ (White 2014: 299). 7. See Stroud (2009: 7): ‘We should question the whole idea of purely sensory knowledge understood as knowledge that someone could have without knowing or having the resources for knowing anything else—in particular, without knowing anything about ‘things located outside’ us. That is the assumption about perception that appears to have completely general sceptical consequences. That is the assumption that must be rejected’. 8. So Coliva (2014: 326) is wrong to think that one can accept this picture without falling into McDowell’s ‘highest common factor conception’, for the latter is the same as accepting RIT. 9. For more on the Cartesian picture and how to resist it, see The Illusion of Doubt. 10. There are debates in the literature whether Moore is addressing the sceptic or the idealist. Coliva thinks it’s the idealist, Maddy that it is the sceptic (see their recent exchange in 2018). Nothing much hangs on this for my purposes, as both the sceptic and the idealist doubt that there is an external, physical world. 11. This is why Maddy, in her unfairly underrated recent book (2017), says the sceptic really wants ‘extraordinary’ evidence. But, of course, we have no clear idea what this could be. 12. Cf. On Certainty §23. 13. Pace what Pritchard (2014) seemed to argue. Pritchard’s current view is that this is merely an appearance, however—i.e. he thinks that there is no genuine limitation at work here; it is more that we cannot be returned to our original state of ‘epistemic innocence’ once we have engaged with the radical sceptical problem (see the final chapter of Epistemic Angst). On my diagnosis, we have never really been in a state of epistemic innocence (even if things seemed that way), as the radical sceptical problem is an inbuilt by-product of endorsing the Cartesian picture, which blocks us off from the ‘external world’ from the get-go. 14. Compare: ‘If my name is not L.W., how can I rely on what is meant by “true” and “false”?’(OC §515). 15. A view that Stroud seems to defend (in conversation). 16. For a similar point see Thomas (2014). 17. It only seems that way because on the Cartesian picture, all claims about things ‘outside the mind’ are necessarily doubtful. 18. This means that, on pain of irrationality, the sceptic cannot just bite the bullet here and ask why her demand would be reasonable only if our inability

126

19. 20.

21. 22. 23. 24. 25. 26. 27.

Genia Schönbaumsfeld to meet it is a contingency. For one cannot coherently ‘want’ what is logically impossible (it only makes sense to want what one could, at least in principle, have). If, for example, there is no such thing as ‘round squares’, then I cannot coherently ‘want’ to find such things. If it makes no sense for people ‘to be’ prime numbers, then I cannot coherently want to ‘find’ a prime called ‘Caesar’. Hence, by parity of reasoning, if I grant that ‘global validation’ is logically impossible, then I cannot coherently continue to demand that there nevertheless be such a thing. Since there would be no practices in which they could play a role. At this point, a sceptically minded philosopher might perhaps object that, contrary to what I have just argued, we do have a clear idea of what it would be like not to know that one has hands, for, surely, that’s just the brain-ina-vat (BIV) (or radical sceptical) scenario! That is to say, if I were a BIV, I could not know that I have hands, since, firstly, I could not know anything, and, secondly, I would not even have hands. Although the latter two claims are true, the objection does not constitute a genuine counter-example to the foregoing argument. Taking the second point first: if I don’t have hands (say, because I’m a BIV), and I could (miraculously) come to know this fact, then my claim ‘I know I don’t have hands’ (or ‘I doubt that I have hands’) would be perfectly meaningful and, indeed, not relevantly different from the accident scenario discussed earlier. If, on the other hand, the BIV scenario is only a fancy way of fleshing out the thought that one might be globally wrong about everything (as it tends to be), then ‘I know I have two hands’ no longer functions as an ordinary proposition about hands (or other body parts) but rather means something like: ‘I know I cannot be globally wrong’ or ‘I know I’m not the victim of a “global” sceptical hypothesis’. But, if so, we have moved from an ‘epistemic’ use of ‘to know’ to what I have called a ‘logical’ or ‘grammatical’ sense. So, we are not concerned with an ordinary factual (empirical) claim, but rather with a logical enabling condition: if the obtaining of this condition is called into question, then the notions of ‘truth’ and ‘falsity’ that the sceptic needs in order to be able to formulate his doubt lose their meaning, too. It is this problem that the radical sceptic overlooks. One might, in the spirit of Descartes’s ‘pure enquiry’, believe that one can call everything into doubt at the same time without realising that the attempt to do so simultaneously deprives one’s expression of doubt of a sense: ‘A doubt that doubted everything would not be a doubt’ (OC §450). Consequently, ‘I know I have two hands’ can look like an ordinary epistemic (empirical) claim but, in the radical sceptical context (i.e. in ordinary circumstances where radical sceptical error-possibilities have been raised), either has no clear meaning or else constitutes an articulation of the ‘logical truth’ that ‘there is no such thing as a doubt in this case’ (OC §58). This implies that Wittgenstein rejects the priority of sensory knowledge. See e.g. Pryor (2000). See e.g. Wright (2004). See Coliva (2012, 2014, 2015). Pryor (2000) heroically defends the notion that an experience as of P is sufficient, if one has no reason to doubt Q (that there is an external world). For more on this, see The Illusion of Doubt, chapter 3. I cannot do full justice to this point here. Of course, we could decide to use another word instead of ‘physical object’, say, ‘sense-datum’. But if this word behaves in exactly the same way as ‘physical object’ does, then all we have done is change the sign, not the symbol (i.e. we are still operating with the same concept). See chapter 3 of The Illusion of Doubt for more on this.

Was Wittgenstein a Disjunctivist?

127

28. This is why Moore’s attempt seems question-begging. 29. Wittgenstein uses the word ‘mythology’ here not to indicate that the background is ‘made up’ or ‘arbitrary’ (for then it might be false), but merely to get us away from thinking that the background must itself be a set of true or false propositions (i.e. something ‘factual’). 30. In this respect, my diagnosis of what creates the radical sceptical problem is quite different from that offered by Pritchard (2015). On my account, radical scepticism is an illusion generated by prior acceptance of the Cartesian picture (and hence of RIT), whereas for Pritchard the problem (or problems) arise from two paradoxes—one generated by the underdetermination principle, the other by the closure (RK) principle. According to Pritchard, disjunctivism can answer only the first problem; for the second, we need Wittgenstein’s insight that all epistemic assessment is essentially local. I discuss Pritchard’s proposal at length in The Illusion of Doubt, chapter 2. 31. I. My current experience is in all respects as if P. II. P. III. There is an external world.

32. 33. 34. 35. 36. 37.

38.

With this schema in place, one can only know (II) if one already knows (III), which leads to acceptance of the Reasons Identity Thesis. And as Pritchard (2015) points out, this is far more convincing than ‘dogmatism’ that takes the ‘liberal’ view that non-factive perceptual experiences can achieve this. Also compare Wright (2004), Coliva (2014). Some, of course, such as Dretske (1970, 1971, 2005a, 2005b) and Nozick (1981), do. For a discussion of Dretske’s strategy, see Chapter 1 of The Illusion of Doubt. Coliva (2014, 2015) defends an ‘internal rationalism’ where assuming the truth of hinge propositions is thought to be constitutive of rationality. The closure principle states that if I know that P and I competently deduce Q from P, thereby coming to know that Q on this basis while retaining my knowledge that P, then I know that Q. In this respect it is irrelevant what phrase you substitute after ‘I know’. I.e. if, in global sceptical contexts, ‘I know I have two hands’ does not function like an ordinary claim about hands, then neither would ‘I know my cat’s name’ or ‘I know that London is the capital of England’ function like ordinary claims about cats’ names or London. For the sceptic is not calling into question whether I know my cat’s name or can see my hands, which would be ordinary questions that one could respond to by saying ‘I know my cat’s name’, etc. If I say this, however, in response to a ‘global’ sceptical challenge, then this really constitutes, if it means anything, a roundabout way of insisting that I cannot be wrong about such things. One might wonder, however, whether in contexts where it does make sense to claim that one knows that one has hands, such as in the accident scenarios described earlier, where one would, if one accepted disjunctivism, be in possession of a factive reason (such as seeing that one has hands) that would entail the truth of this proposition, one could not use this knowledge to refute radical scepticism. But this would be a mistake. One could not turn the fact that one knows that one has hands in such scenarios into an anti-sceptical claim à la Moore, for the attempt to do so would result in a radical change of context: one would be moving from an ordinary context, where mundane knowledge claims can be made, to a radical sceptical context, where we are trying to assert that we are not radically mistaken. If what I have argued in this chapter is right, however, then this is something we cannot do. For it

128

Genia Schönbaumsfeld

implies that we have moved from an ‘epistemic’ use of ‘I know’ to a ‘logical’ or grammatical one, which means that Moore would not be informing us about a true state of affairs (as he wants), but would rather be articulating a condition for the meaningfulness of our epistemic practices taken as a whole. Hence, it is not that closure fails, but rather that this principle has no application in this ‘global’ sceptical context. 39. This is not a revisionary proposal—it does not attempt to reinterpret the ordinary sense of ‘to know’. Rather, it is because ‘to know’ means what it does that Wittgenstein believes that knowledge claims (or denials) are not fully intelligible in the radical sceptical context. In such a context, the only sense that can be made of them is therefore ‘logical’, not ‘epistemic’. I believe that this is also the significance of Wittgenstein’s notorious remark at PI §402: ‘For this is what disputes between idealists, solipsists and realists look like. The one party attacks the normal form of expression as if they were attacking an assertion; the others defend it, as if they were stating facts recognized by every reasonable human being’. On my reading, it is clear why Wittgenstein thinks one should do neither. 40. For more on the distinction between ‘local’ and ‘global’ sceptical scenarios, see The Illusion of Doubt, chapter 2.

References Bennett, Jonathan, 1971, Locke, Berkeley, Hume: Central Themes. Oxford: Oxford University Press. Boghossian, Paul, 2007, Fear of Knowledge. Oxford: Oxford University Press. Burge, Tyler, 2003, Perceptual entitlement, Philosophy and Phenomenological Research 67: 503–548. Coliva, Annalisa, 2012, Moore’s proof, liberals and conservatives—Is there a (Wittgensteinian) third way? in Annalisa Coliva, ed. Mind, Meaning and Knowledge. Oxford: Oxford University Press, pp. 323–351. ———, 2014, Moderatism, transmission failures, closure and humean scepticism, in D. Dodd and E. Zardini, eds. Scepticism and Perceptual Justification. Oxford: Oxford University Press, pp. 248–276. ———, 2015, Extended Rationality: A Hinge Epistemology. London: Palgrave Macmillan. ———, 2018, What do philosophers do? Maddy, Moore and Wittgenstein, International Journal for the Study of Skepticism 8(3): 198–207. Conee, Earl, 2007, Disjunctivism and anti-skepticism, Philosophical Issues 17: 16–36. Descartes, Rene, 1986 [1641], Meditations on First Philosophy. Trans. John Cottingham. Cambridge: Cambridge University Press. Diamond, Cora, 1991, The Realistic Spirit. Cambridge, MA: MIT Press. Dretske, Fred, 1970, Epistemic operators, Journal of Philosophy 60: 1007–1023. ———, 1971, Conclusive reasons, Australasian Journal of Philosophy 49: 1–22. ———, 2005a, The case against closure, in Ernest Sosa and Matthias Steup, eds. Contemporary Debates in Epistemology. Oxford: Blackwell, pp. 13–25. ———, 2005b, Reply to Hawthorne, in Ernest Sosa and Matthias Steup, eds. Contemporary Debates in Epistemology. Oxford: Blackwell, pp. 43–46. Dummett, Michael, 1978, Truth and Other Enigmas. London: Duckworth. Maddy, Penelope, 2017, What Do Philosophers Do? Scepticism and the Practice of Philosophy. New York: Oxford University Press.

Was Wittgenstein a Disjunctivist?

129

———, 2018, Replies to Coliva, Leite and Stroud, International Journal for the Study of Skepticism 8(3): 231–244. McDowell, John, 1994, Mind and World. Cambridge, MA: Harvard University Press. ———, 1998a, Criteria, defeasibility and knowledge, in his Meaning, Knowledge and Reality. Cambridge, MA: Harvard University Press, pp. 369–394. ———, 1998b, Knowledge and the internal, in his Meaning, Knowledge and Reality. Cambridge, MA: Harvard University Press, 395–413. McGinn, Colin, 1984, The concept of knowledge, Midwest Studies in Philosophy 9: 529–544. Moore, George Edward, 1959, Philosophical Papers. London: George Allen and Unwin. Nagel, Thomas, 1986, The View From Nowhere. Oxford: Oxford University Press. Nozick, Robert, 1981, Philosophical Explanations. Oxford: Oxford University Press. Pollock, John, 1974, Knowledge and Justification. Princeton, NJ: Princeton University Press. Pritchard, Duncan, 2014, Entitlement and the groundlessness of our believing, in D. Dodd and E. Zardini, eds. Scepticism and Perceptual Justification. Oxford: Oxford University Press, pp. 190–212. ———, 2015, Epistemic Angst. Princeton, NJ: Princeton University Press. Pryor, James, 2000, The skeptic and the dogmatist, Noûs 43: 517–549. Schönbaumsfeld, Genia, 2016, The Illusion of Doubt. Oxford: Oxford University Press. Stroud, Barry, 1984, The Significance of Philosophical Scepticism. Oxford: Oxford University Press. ———, 2009, Scepticism and the senses, European Journal of Philosophy 17(4): 559–570. Thomas, Alan, 2014, McDowell on transcendental arguments, scepticism and error theory, International Journal for the Study of Skepticism 4(2): 109–124. White, Roger, 2014, What is my evidence that here is a hand? in D. Dodd and E. Zardini, eds. Scepticism and Perceptual Justification. Oxford: Oxford University Press, pp. 298–321. Williams, Michael, 1996, Unnatural Doubts. Princeton, NJ: Princeton University Press. ———, 2011, External world scepticism and the structure of epistemic entitlement, in Jason Bridges, Nico Kolodny and Wai-Hung Wong, eds. The Possibility of Philosophical Understanding. Oxford: Oxford University Press, pp. 43–61. Wittgenstein, Ludwig, 1969, On Certainty. Ed. G.E.M. Anscombe and G.H. von Wright, trans. Anscombe and D. Paul. Oxford: Blackwell. ———, 1975, Zettel. Ed. G.E.M. Anscombe and G.H. von Wright, trans. G.E.M. Anscombe. Oxford: Blackwell. ———, 1989, Remarks on the Foundations of Mathematics. Ed. G.H. von Wright, R. Rhees and G.E.M. Anscombe, trans. G.E.M. Anscombe. Blackwell: Oxford. ———, 2009, Philosophical Investigations. Ed. P.M.S. Hacker and J. Schulte, trans. G.E.M. Anscombe, P.M.S. Hacker and J. Schulte. Oxford: Blackwell. Wright, Crispin, 1980, Wittgenstein on the Foundations of Mathematics. London: Duckworth.

130

Genia Schönbaumsfeld

———, 2002, Anti-sceptics simple and subtle: G. E. Moore and John McDowell, Philosophy and Phenomenological Research 65: 330–348. ———, 2004, On epistemic entitlement: Warrant for nothing (and foundations for free?), Aristotelian Society (supp. vol.) 78(1): 167–212. ———, 2008, Comment on John McDowell, in Adrian Haddock and Fiona Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 390–404. ———, 2014, On epistemic entitlement (II): Welfare state epistemology, in D. Dodd and E. Zardini, eds. Scepticism and Perceptual Justification. Oxford: Oxford University Press, pp. 213–247.

7

Settling a Question Austin and Disjunctivism Guy Longworth1

And it’s the differences that matter. —J. L. Austin

1. Introduction Was J. L. Austin a disjunctivist? The question is made difficult to answer both by Austin’s unwillingness to propound doctrine and by a lack of consensus as to the core commitments of disjunctivism. I shall attempt to make some progress in settling it by articulating some central aspects of some contemporary forms of disjunctivism and indicating some ways in which those aspects are foreshadowed in Austin’s work. I shall suggest that, although Austin failed clearly to articulate or defend a version of this contemporary form of disjunctivism, his work was distinctively hospitable to its articulation and defence. I’ll begin, in the following section, by sketching some central themes in Austin’s discussions of perception and indicating some ways in which it is not entirely straightforward to classify Austin’s position as an interesting (or non-schematic) form of disjunctivism. In section 3, I discuss the way in which Austin sought to distinguish sensory perception from knowledge. In section 4, I discuss some ways in which that distinction figured in Austin’s treatment(s) of delusions, including sensory hallucinations. Austin’s discussions of sensory hallucinations are then compared, in section 5, with a central contemporary specification of disjunctivism, initially proposed by J. M. Hinton and developed in different ways by John McDowell, Paul Snowdon and M. G. F. Martin (see Hinton 1967, 1973; McDowell 1982; Snowdon 1980–1981; Martin 1997). That comparison reveals that Austin’s position was closer to these contemporary forms of disjunctivism than it is sometimes credited with being, but is nonetheless distinguishable from them.

2. Problems of Classification Three central themes in Austin’s approach to philosophical discussion of perception and knowledge can be discerned in the opening pages of Sense

132

Guy Longworth

and Sensibilia: (Theme (1)) that what is being offered is not so much a set of philosophical claims as a technique for dealing critically with others’ attempts to make such claims; (Theme (2)) that one main component in that technique is the development of an appropriate sensitivity to differences amongst cases; (Theme (3)) that a second main component is to foster an attitude of common sense realism (Putnam 1994; Martin 2007: 15–29; Longworth forthcoming), or trust in one’s naïve judgements— both one’s first-order judgements about one’s environment and one’s second-order assessments of those first-order judgements. [Theme (1):] I am not, then—and this is a point to be clear about from the beginning—going to maintain that we ought to be “realists”, to embrace, that is, the doctrine that we do perceive material things (or objects). . . . [Theme (2):] There is no one kind of thing that we “perceive” but many different kinds, the number being reducible if at all by scientific investigation and not by philosophy: pens are in many ways though not in all ways unlike rainbows, which are in many ways though not in all ways unlike after-images, which in turn are in many ways but not in all ways unlike pictures on the cinemascreen—and so on, without assignable limit. [Theme (1):] So we are not to look for an answer to the question, what kind of thing we perceive. What we have above all to do is, negatively, to rid ourselves of such illusions as “the argument from illusion”. . . . [Themes (1) and (3):] It is a matter of unpicking, one by one, a mass of seductive (mainly verbal) fallacies, of exposing a wide variety of concealed motives—an operation which leaves us, in a sense, just where we began. [Themes (1), (2), and (3):] In a sense—but actually we may hope to learn something positive in the way of a technique for dissolving philosophical worries (some kinds of philosophical worry, not the whole of philosophy). (Austin 1962: 4–6) Insofar as disjunctivism amounts to a doctrine rather than a technique, Theme (1) tends to count against finding its traces in Austin. However, Themes (2) and (3) are more hospitable. Theme (3) seems naturally supportive of naïve approaches to perceptual experience. Such approaches seek to treat perceptual experience as at the same time subjective and environmental. Perceptual experience is treated as subjective in that it is viewed as fully determined by how things seem to its subjects and, so, as accessible to its subjects by introspection. And it is treated as environmental in that it is taken to depend constitutively on mind-independent elements, potentially including objects, activities, or conditions. Thus, we would ordinarily take ourselves to be able to tell, by introspection, whether we are seeing aspects of our environments. Naïveté is threatened by arguments to the sophisticated conclusion that

Settling a Question

133

subjective and environmental conditions are independent. These arguments aim to show that it is possible to hold fixed how things appear to subjects whilst arbitrarily varying their environments; and that it is possible to hold fixed their environments whilst arbitrarily varying how things appear to them. The function of disjunctivism is defensive: it is designed to block those arguments by explaining away the appearance that it is possible for all of the ways things appear to subjects to vary independently of all of the aspects of their environments. To a first approximation, the disjunctivist explanation appeals to differences amongst the ways things appear. According to the disjunctivist, the ways things appear fall into disjoint classes depending on whether or not they are constitutively dependent on aspects of the environment. The misleading appearance of independence is fostered only by slipping between the unthreatening claim that there are some ways things appear which are environment-independent; and the threatening, but undefended, claim that, since the ways things appear form a unified kind, there are no ways things appear which are environment-dependent. Thus, Theme (2) figures in conditioning the disjunctivist to discern differences amongst the ways things appear to which the sophisticated theorist is insensitive. Not only are Themes (2) and (3) indicative of a propensity towards disjunctivism, there is some reason to think that when the issue is considered at this level of generality, our finding that Austin endorsed a naïve view of perception would settle the question in favour of his also endorsing a form of disjunctivism. As Paul Snowdon puts the thought, disjunctivism is, in a nutshell, the idea that apparent perceptual experiences should be thought of as belonging to two alternate, disjoint classes. They belong, that is, to one or the other case, where one case is the kind of experience that constitutes genuine perception of an item or feature in the percipient’s environment, and the other is an experience which seems to be that way but which is actually not that way, and is, rather, an hallucination. We might, therefore, using a somewhat resonant term, describe disjunctivism as experiential dualism. Now, the conclusion of most reflection about perception in our philosophical tradition has been that all such experiences have a single nature, usually taken to be the apprehending, or directly perceiving, of some inner item. Consequently it can be said that the standard philosophical approach has been in favour of experiential monism. There has, however, always been a tradition of thought in which naïve realism about perception is defended. It seems, though, that this naïve realist tradition must accept the claim that there are, besides the perceptual experiences which can be described in accordance with the naïve realist conception (whatever it is), other sorts of experiences, say hallucinations, which seem to the subjects of them to

134

Guy Longworth be of the genuinely perceptual sort, but which are certainly not of the same sort. This seems to mean that they must be experiential dualists, and so disjunctivists, of some sort. (Snowdon 2008: 36–37; emphasis in original)

Although I’ve suggested that Theme (3) is naturally supportive of naïve approaches to perception, I haven’t yet provided any evidence that Austin would have agreed. So, even if Snowdon’s sketch showed that a commitment to a naïve view of perception brought in train a commitment to disjunctivism, more work is required in order to decide whether Austin was so committed. Part of the evidence in favour of attributing to Austin a commitment to a naïve approach to perception stems from his resistance to arguments against naïve treatments, as will emerge more fully in the remainder: that resistance is, I think, best explained by his being committed to a naïve treatment. Further evidence is supplied by Austin’s conception of the role of perceptual experience in settling questions and, so, in sustaining knowledge about one’s environment: One might add that Warnock subtly intensifies this air of chanciness by taking his examples from the sphere of hearing. It is, as a matter of fact, quite often true that, just going by the sound, we do make some sort of inference in saying what we hear, and it is quite often easy to see how we might go wrong. But then seeing is not, as Warnock quietly takes for granted, exactly like this; for it is, characteristically, by seeing the thing that the question is settled. (Austin 1962: 138–139) The idea that one might exploit one’s seeing a thing in order to settle a question seems to depend on the idea that it is possible, in propitious circumstances, to tell that one is seeing the thing. Austin treats similarly the role of plainly viewing a pig: But if the animal then emerges and stands there plainly in view, there is no longer any question of collecting evidence; its coming into view doesn’t provide me with more evidence that it’s a pig, I can now just see that it is, the question is settled. (Ibid.: 115) Plainly viewing, or seeing, a pig can put one in a position to see that it is a pig and, so, to know—in a sight-specific way—that it is a pig. And Austin’s appeal to plainness plausibly suggests, once again, a role for the availability of what one is seeing in enabling one to use perception to settle questions. (For further discussion of plainness, see Hinton 1973:

Settling a Question

135

29–34, 41–43; Snowdon 2008: 41–43. For further discussion of Austin’s views about knowledge, and about the role of perception in sustaining knowledge, see Putnam 1994; Martin 2007; Longworth 2018.) However, even granting reason to think that Austin would have endorsed a naïve view of perception, at least three sorts of reasons might be offered for resisting the further attribution, made solely on that ground, of a commitment to disjunctivism. The first reason for resisting the easy transition to an attribution of disjunctivism is that, even if a naïve approach to perception or knowledge were to entail disjunctivism, Austin might not himself have recognised the entailment. And we shouldn’t assume that if he were to have been brought to acknowledge the entailment, he would have treated it as a reason for accepting disjunctivism, rather than as a reason for rejecting any approach to perception or knowledge that entailed it. The second reason is that we might hold that the attribution of sheer experiential dualism, in Snowdon’s sense, is insufficient for an interesting attribution of disjunctivism. That is, we might hold that disjunctivism, as understood by its contemporary proponents, is a more specific doctrine than the simple denial that the ways things seem form a unified kind—a doctrine that incorporates, for example, a more or less detailed explanation for how the ways things seem can merely seem to form, without in fact forming, a unified kind. Put another way, we might be willing to allow that Austin instances the schematic form of disjunctivism that figures in Snowdon’s sketch whilst at the same time seeing the ways in which his position differs from contemporary forms of disjunctivism to be more important than the schematic similarities. (I don’t mean to suggest that Snowdon is under any illusion about this.) The third reason is more specific to Austin and derives from Theme (2). As Snowdon presents schematic disjunctivism, it embodies both negative and positive components. The negative component is the rejection of experiential monism. Here, I think that Austin—and, indeed, any other proponent of a naïve approach to perceptual experience—would be inclined to agree. The positive component is the acceptance of experiential dualism: the claim that sensory experiences fall neatly into two classes, those in which experience is partly constituted by mind-independent aspects of the environment, and those in which it isn’t. Here, I think that Austin’s attitude is likely to have been more equivocal. For, on one hand, the appeal to mind-dependence will determine exactly two classes only if the relevant notion of mind-dependence does so, and Austin might well have resisted the idea that the notions of mind, and so mind-dependence, are sufficiently sharp to sustain the required classification. (Here, one might consider, for example, what Austin’s attitude would have been to the question whether after-images are mind-dependent. See Austin 1962: 27.) On the other hand, even if Austin were willing to accept that sensory experiences can be sorted into these two classes, he might at the same

136

Guy Longworth

time have held that differences within each of the two classes, or similarities across the class boundary, are of greater importance. Thus, for example, I am not denying that cases in which things go wrong could be lumped together under some single name. A single name might in itself be innocent enough, provided its use was not taken to imply either (a) that the cases were all alike, or (b) that they were all in certain ways alike. What matters is that the facts should not be prejudged and (therefore) neglected. (Austin 1962: 14, fn. 1; emphasis in original) For these reasons, it would be safer to think of Austin’s rejection of experiential monism as supporting not experiential dualism, but rather experiential pluralism (see also Snowdon 2008: 38, fn. 6; Kalderon 2011). With those warnings about seemingly easy resolutions in hand, the question before us divides into two. (i) Did Austin’s commitment to a naïve treatment of perception translate into a commitment to schematic disjunctivism—the rejection of experiential or cognitive monism? (ii) Insofar as Austin was committed to schematic disjunctivism, did his position also manifest more specific similarities with, or differences from, contemporary forms of disjunctivism?

3. Perception and Cognition Underpinning Austin’s various discussions of knowledge and perception is a distinction between these two elements. On one side of the distinction is propositional knowledge, an outcome of the operation of a power of propositional cognition or judgement. Exercises of that power can be susceptible to assessment as correct or incorrect, in light of the fit, or lack of fit, between the propositions they engage and the ways things are. On the other side of the distinction is sensory or perceptual experience, the outcome of a power to enjoy awareness of, or acquaintance with, concrete environmental elements—for example, objects, events, conditions, stuffs, processes, or other environmental ephemera. Unsuccessful exercises of that power can fail to sustain perceptual awareness of environmental elements, but not in a way which sustains an easy distinction between correct and incorrect such exercises (in addition to Austin 1946, 1962, see Austin 1950, 1954). The distinction plays two central parts in Austin’s drama, as well as taking on a number of important subsidiary roles. The first central role played by the distinction is in undermining treatments of propositional knowledge, and in particular propositional knowledge about one’s own mind, as comprising a form of acquaintance

Settling a Question

137

(Austin 1946: 96–97, 115–116). Austin’s concerns here are twofold: first, if introspective self-knowledge is treated as a mode of acquaintance, then it is hard to see how other people could enjoy the same mode of knowledge about us that we can enjoy about ourselves; second, if selfknowledge is treated as a mode of acquaintance, then it is hard to make room for errors and uncertainties—again marking a principled distinction between one’s knowledge of one’s own mind and one’s knowledge of others’. In response, Austin emphasises the essential role of predication, or description, in propositional cognition, both as something that cannot be delivered up through simple acquaintance, and as what makes room for the possibility of error: Any description of a taste or sound or smell (or colour) or of a feeling, involves (is) saying that it is like one or some that we have experienced before: any descriptive word is classificatory, involves recognition and in that sense memory, and only when we use such words (or names or descriptions, which come down to the same) are we knowing anything, or believing anything. But memory and recognition are often uncertain and unreliable. (Austin 1946: 92) Similarly, if “on the basis of being in pain” only means “when I am (what would be correctly described as) in pain”, then something more than merely saying “I’m in pain” is necessary for knowing I’m in pain: and this something more, as it involves recognition, may be hesitant and/or mistaken. (Austin 1946: 95; for discussion, see Pears 1979; emphasis in original) The second central role played by the distinction between propositional cognition and sensory experience is in undermining the converse mistake: treating sensory experience as if it were a form of propositional cognition. The treatment at issue would be one on which what was given in sensory experiences determined how things would have to be in order for the sensory experiences to be correct or incorrect, and so determined the sorts of propositional contents which could also figure in the operations of cognition or its expression in speech: Very clearly detailed, this is the view that, at least and only in a certain favoured type of case, I can “say what I see (or otherwise sense)” almost quite literally. (Austin 1946: 90) So, on this treatment, sensory experience already involves classification which can then simply be replicated in sense-based cognition.

138

Guy Longworth

Much of Austin’s 1962 discussion of perception is taken up with trying to undermine this assimilation of sensory experience with propositional cognition. His main lines of argument here appeal to the idea that sensebased cognition is the upshot of the interaction of sensory experience with propositional cognition. Because of this, there is no backward path from errors in sense-based cognition to specifically sensory dysfunction. In particular, there is no such path that is suitable to sponsor a distinction between correct and incorrect (veridical and non-veridical) sensory experiences. For example, when the plain man sees on the stage the Headless Woman, what he sees (and this is what he sees, whether he knows it or not) is not something “unreal” or “immaterial”, but a woman against a dark background with her head in a black bag. If the trick is well done, he doesn’t (because it’s deliberately made difficult for him) properly size up what he sees, or see what it is; but to say this is far from concluding that he sees something else. (Austin 1962: 14; emphasis in original) Seeing what it is that is before one is a matter of seeing that something is so, and so is a form of propositional cognition. Indeed, plausibly, it is a form of propositional knowledge. Failures, or errors, in propositional cognition needn’t trace to failures, or dysfunctions, in sensory experience. Furthermore, seeking so to explain such errors can lead to definite mistakes, as when one tries to explain why someone failed to see that there was a woman against a dark background with her head in a black bag by appeal to the idea that they saw something else—so that, for example, their sensory experience would have been correct just in case there had been a headless woman there. More generally, with responsibility for errors in sense-based cognition distributed over the capacity for propositional knowledge, the capacity for perception, and the interactions of those capacities, one can better account for the variety of ways in which things can go wrong without seeking to trace all those ways to simple incorrectness in sensory experience: That is to say, once again there is no neat and simple dichotomy between things going right and things going wrong; things may go wrong, as we really all know quite well, in lots of different ways— which don’t have to be, and must not be assumed to be, classifiable in any general fashion. (Austin 1962: 13; emphasis in original) Austin summarises his response to the assimilation of sensory experience to propositional cognition in the following way:

Settling a Question

139

though the phrase “deceived by our senses” is a common metaphor, it is a metaphor; and this is worth noting, for in what follows the same metaphor is frequently taken up by the expression “veridical” and taken very seriously. In fact, of course, our senses are dumb— though Descartes and others speak of the “testimony of the senses”, our senses do not tell us anything, true or false. (Austin 1962: 11) Austin’s diagnosis is, again, that this treatment of sensory experience is fostered by insufficient sensitivity to the differences between perception and propositional knowledge: Uncritical use of the direct object after know seems to be one thing that leads to the view that (or to talking as though) sensa, that is things, colours, noises, and the rest, speak or are labelled by nature, so that I can literally say what (that which) I see: it pipes up, or I read it off. It is as if sensa were literally to “announce themselves” or to “identify themselves”. . . . But surely this is only a manner of speaking . . . : sensa are dumb, and only previous experience enables us to identify them. If we choose to say that they “identify themselves” (and certainly “recognizing” is not a highly voluntary activity of ours), then it must be admitted that they share the birthright of all speakers, that of speaking unclearly and untruly. (Austin 1946: 97) Austin’s distinction between sensory experience and propositional cognition plays important subsidiary roles in underwriting lifelike explanations of a variety of ways in which sense-based cognition can go wrong, or in which subjects can be tempted towards making sense-based mistakes. Strictly, the distinction makes available three general forms of error-explanation, which may operate in combination: (a) explanations which appeal to cognitive failure, perhaps through inattention or carelessness, or in more extreme cases, through some form of cognitive disorder; (b) explanations which appeal to more general failures of the system as a whole, including the interactions between its sensory and cognitive elements, as in some forms of visual illusion, wherein one sees how things look and responds cognitively in ways typically appropriate to things’ looking that way, and yet the outcome cognition fails to fit how things are; (c) explanations which appeal to specifically sensory malfunction. However, Austin’s view is that cases of type (c) are uncommon, and certainly less common than philosophers have typically thought. And insofar as errors, or inclinations towards error, can be given type-(a) or type-(b) explanations, that fact can help to insulate naïve accounts of perception from objections which might otherwise be thought to arise

140

Guy Longworth

simply from the obtaining of tendencies towards sense-based error. In order to be effective, such objections would have to be supported by specific reasons to accept two claims about cases of sense-based errors: first, that some such errors must be explained by appeal to forms of specifically sensory malfunction that are insusceptible to naïve treatment; and, second, that the required explanations generalise, so as to preclude naïve treatments of any case of perceptual experience. In the following section, we’ll consider in greater detail how Austin proposes to treat some specific cases. (For further discussion and development of Austin’s distinction, building especially on his 1962: 33–43 discussion of “looks”, “appears”, and “seems”, see Travis 2004; see also Price 1952.)

4. Perception and Hallucination A type of case which might be thought to make trouble for a naïve treatment of perception would be a case in which one had an experience in which it seemed to one as if one were perceiving environmental elements but in which no such environmental elements were present, a type which Austin characterises as delusion. For according to the naïve approach, the experience one has when genuinely perceiving has a nature partly constituted by environmental elements. Hence, one couldn’t be having a sensory experience of the same nature as a genuine perception of particular environmental elements if no such environmental elements were present. Part of Austin’s response to the thought that such cases would make trouble for a naïve treatment of perception is to appeal to the distinction between perception and cognition and to try to explain such cases by appeal to cognitive, rather than sensory, disorder. That is, Austin appeals to the idea that it might seem to one as if one were perceiving environmental elements not because of the kind of sensory experience one was undergoing, but rather because of disorder in the power responsible for sizing up what one’s experience reveals about one’s environment: Delusions, on the other hand, are something altogether different from this [i.e., typical cases of illusions]. Typical cases would be delusions of persecution, delusions of grandeur. These are primarily a matter of grossly disordered beliefs (and so, probably, behaviour) and may well have nothing in particular to do with perception. (Austin 1962: 23) However, Austin also seems to allow that there might be delusions which cannot be explained merely by appeal to cognitive disorder, but must be explained partly by appeal to features of the sensory experiences they involve:

Settling a Question

141

But I think we might also want to say that the patient who sees pink rats has (suffers from) delusions—particularly, no doubt, if, as would probably be the case, he is not clearly aware that his pink rats aren’t real rats. (Austin1962: 23) Here, Austin appears to leave open that someone might have a sensory experience that was qualitatively similar to that involved in a genuine perception of pink rats, but undergone in the absence either of pink rats or of any other suitable environmental elements. That is, Austin appears to allow here for the possibility of hallucinatory sensory experiences which are qualitatively similar to, and perhaps qualitatively indistinguishable from, perceptual experiences. (Austin would certainly have disallowed the transition from the former to the latter, on the grounds that things can be alike without being perfectly alike; see Austin 1962: 41–42). If Austin allows for the possibility of such cases, and at the same time seeks to uphold a naïve treatment of the experiences involved in genuine perception, then we would expect him to appeal to a difference in nature between the genuine and the hallucinatory cases. That is something we do find: If I am told that a lemon is generically different [that is, different in nature] from a piece of soap, do I expect that no piece of soap could look just like a lemon? Why should I? (Ibid.: 50) Austin’s thought seems to be that, similarly, it does not follow from the supposition that pink rats are generically different from hallucinated pink rats, or pink rat sense data, that pink rats and pink rat sense data cannot appear similar or, indeed, exactly similar. And if pink rats and pink rat sense data can appear exactly similar, then perhaps nothing precludes an experience of one from being generically different to, and yet qualitatively indistinguishable from, an experience of the other. Thus, Austin develops the thought: But if we are prepared to admit that there may be, even that there are, some cases in which “delusive and veridical perceptions” really are indistinguishable, does this admission require us to drag in, or even to let in, sense-data? No. For even if we were to make the prior admission (which we have so far found no reason to make) that in the “abnormal” cases we perceive sense-data, we should not be obliged to extend this admission to the “normal” cases too. For why on earth should it not be the case that, in some few instances, perceiving one sort of thing is exactly like perceiving another? (Ibid.: 52)

142

Guy Longworth

Now, at first sight, it’s natural to feel sympathy with Snowdon’s comment about this passage: Austin here looks to be precisely making room for what I have been calling experiential dualism. (Snowdon 2008: 38) However, a second glance is apt to be less comforting. For one thing, Austin seems to concede a potential role for sense data only for the sake of (further) argument. For another, the proposal that a naïve treatment of perception entails experiential dualism seemed to rely on two ideas: first, that on a naïve treatment of the experience involved in perceiving some environmental elements, enjoying such an experience requires perceiving such elements, and so that kind of experience would not be possible in the absence of suitable such elements; and, second, that it is possible for there to be sensory experiences in the absence of any such environmental elements, and so in the absence of perception. By contrast, Austin doesn’t say anything here about the possibility of sensory experiences in the absence of perception. Rather, the two sorts of experiences which he discusses both involve perception; it is just that they involve perceiving elements of different sorts, since only some of them are sense data. An attempt could be made to press Snowdon’s suggestion by appeal to the idea that sense data are not environmental elements, and so are not potential objects of perception (see e.g. Snowdon 2008: 39). However, since Austin doesn’t commit to a view about the nature of sense data, doing so would involve going beyond the available evidence. Whatever its bearing on Austin’s classification as a schematic disjunctivist, his thought that qualitative sameness and generic sameness might cohabit is anyway apt to be found unsatisfying as a response to the challenge to naïve treatments of perception sustained by the possibility of certain sorts of sensory hallucinations (for articulation of that source of dissatisfaction, see e.g. Martin 1997, 2007; Soteriou 2016: 16–26). Furthermore, it is precisely in order to provide a more adequate response to that type of challenge that contemporary articulations of disjunctivism have developed beyond the schematic form. In what follows, I seek to explain both concerns via a stepwise development of the challenge. The aim is to establish at which step this proposal of Austin’s stumbles. (The challenge sketched here, as well as the contemporary disjunctivist response, derives from, and is developed in greater detail in, Martin 1997, 2004, 2006. See also Soteriou 2016: 39–52, 158–184). Central to naïve treatments of perception is the claim that one couldn’t undergo the very same kind of experience that is involved in perceiving aspects of one’s environment in the absence of such perceived aspects. Now consider a subject who is perceiving aspects of their environment. It is plausible that their sensory experience is enabled by specific physical

Settling a Question

143

aspects of what is going on in their brains and nervous systems. Furthermore, it is plausible that by perfectly replicating, in the subject, those specific physical aspects of what is going on in the perceiving subject’s brain and nervous system, it is possible to induce a subjectively similar hallucination. It seems reasonable to allow that it is possible to replicate those aspects of what is going on in the perceiving subject’s brain and nervous system without replicating the aspects of environment that they are perceiving. It follows that any sensory experience enjoyed by the subject of replication can be undergone in the absence of relevant environmental aspects and, so, is not susceptible to naïve treatment. That much is consistent with Austin’s proposal, since all we are bound to accept to this point is that the experiences of perceiving and induced hallucination are subjectively similar. In particular, we are not yet forced to accept that any experience undergone by the replicating subject has the same qualities as, or is of the same kind as, the experience undergone by the perceiving subject. However, on further natural assumptions, that result seems to be compelled. It is very natural to assume that the subjective similarity between the target hallucination and its matching perceptual experience is due to their sharing an array of qualities and that its having those qualities explains why the hallucination seems to its subject to be of the same kind as the matching perceptual experience. Thus, it is because the hallucination possesses those qualities that it has the phenomenal character it does. If that assumption is accepted, then it follows that it is possible to explain why an experience seems to its subject to be a case of perceiving aspects of the environment without affording perceived aspects of the environment a constitutive role in determining the nature of that experience. However, it doesn’t yet follow that it is possible to explain why experiences of perceiving aspects of the environment seem the way they do without affording perceived aspects of the environment a role as constituents of those experiences. For the considerations to this point leave open that, although the qualities of the hallucination explain why it seems to its subject to be a case of perceiving, and although the experiences one has when perceiving also have those qualities, nonetheless the experiences one has when perceiving have additional qualities, and those additional qualities play an essential role in explaining its phenomenal character. Thus, further assumptions are required in order to undermine Austin’s proposed treatment. A key further assumption concerns the nature of the target hallucination. We’ve already assumed that it is possible to bring about that hallucination in the absence of perceived aspects of the environment. The new assumption is that it is possible for an experience with exactly the same qualities as the target hallucination to occur in the presence of perceived aspects of the environment and, indeed, for such an experience to occur consistently with the occurrence of all the other conditions that enable

144

Guy Longworth

the subject to perceive aspects of their environment. It follows from the assumptions made to this point that an experience with the same qualities as the target hallucination occurs whenever a perceptual experience occurs. As J. M. Hinton characterises the former kind of experience, it would be a quasi-hallucination, for which one must no more supply an actual object than for a genuine hallucination, but which occurs whenever one has a veridical experience. (1967: 227) But it seems to follow from the assumptions made to this point, first, that the qualities of the quasi-hallucination would suffice to explain the way that the experience of perceiving seems to its subject, and second, that the explanation furnished by appeal to those qualities would be more general than the explanation offered by the naïve realist, since it would explain not only the experience of perceiving but also the experience of undergoing a matching hallucination. The challenge facing the naïve realist at this point is to defend the claim that perceived aspects of the environment figure in the constitution of the experience of perceiving those aspects given that the qualities of the coincident quasi-hallucination would appear to screen off any qualities peculiar to the experience of perceiving and, so, to prevent those qualities from playing a role in explaining how the experience of perceiving seems to its subject. Even at this stage, it might still be possible to defend the letter of Austin’s proposal that the natures of perceptual experiences and quasihallucinations differ. And it might be possible to defend the claim that perceptual experiences have qualities which quasi-hallucinations lack (for discussion, see Martin 1997, 2002). However, it seems reasonable to allow that, since they are screened off by the qualities of the quasihallucination, those differences will play no role in explaining the ways perceptual experiences seem to their subjects. It therefore seems reasonable to accept that perceptual experiences and quasi-hallucinations fall into the same experiential kind. And to accept that would be to reject the naïve treatment of perceptual experiences.

5. Disjunctivism The path from Austin’s proposal to the failure of the naïve treatment seems to proceed straightforwardly from his opening concession that there can be cases of hallucinatory experiences which are qualitatively indistinguishable from—that is, which have the same experience determining qualities as—perceptual experiences. As was mentioned earlier, contemporary forms of disjunctivism have developed beyond the schematic form precisely in order to block that path. The central development

Settling a Question

145

is aimed at explaining why a hallucination might seem to be qualitatively indistinguishable from a matching perceptual experience even though it doesn’t in fact share qualities with the perceptual experience. As Hinton observes, we would ordinarily allow that the things we perceive might merely seem to us to share qualities without really doing so. The disjunctivist seeks to extend that permission to experiences: If things had to have a common property for you to take one for the other then a dagger, or a flash of light (such as may occur unobserved) would have to have properties in common with “dagger of the mind” or a “phosphene”: a flash you see when an electric current is passed through your brain. Or else it would have to be, strictly speaking, a sense-datum of the one that you took for a sense-datum of the other. Why, if we don’t think that, should events have to have properties in common in order to be mistaken for one another? Why should it not just seem as if they had properties in common? Seeing a flash of light and having that illusion seem, but only seem, to have in common the property “when x occurs a flash of light occurs”. (Hinton 1967: 225) The view is one on which the ways our experiences seem, and, in particular, the qualities that they seem to us to instantiate, are determined, not only by the qualities they have, but also by our abilities to tell which qualities they have and, so, to classify them as similar or different. On this view, our awareness of the qualities of our experiences is not a matter of simple acquaintance with those experiences or their qualities. Rather, it involves the operation of the power of introspective cognition. It is therefore possible to explain why it might seem to us that two experiences shared a quality despite their not in fact sharing that quality by appealing to limitations on the operations of that power. On the assumption that operations of the power of introspective cognition are so limited that we might fail to discriminate amongst experiences that fail to share relevant qualities, we can explain why such experiences might seem to us to share such qualities without doing so. If such a view could be made out, then it would provide the resources to block the crucial assumption, according to which a perceptual experience and an internally physically matching hallucination may share all of the qualities that are relevant to determining the way the experience seems to the subject. The assumption is made to seem plausible in two steps. In the first step, a case is made for thinking that the subject of such a hallucination would be unable to discriminate it on the basis of introspection from its internally physically matching perceptual experience. In the second step, a case is made for thinking that the best explanation of the subject’s inability to discriminate the hallucination from its matching perceptual experience is that the two experiences share qualities. The

146

Guy Longworth

disjunctivist proposal that we are considering seeks to block the move from the first step to the second. According to the proposal, a superior explanation of the subject’s inability to discriminate amongst their experiences is provided by appeal only to limitations on their power of introspective cognition. As Hinton puts it, It can indeed be the same experience, but this only means that it can “be the same” experientially or subjectively or “qualitatively”, i.e. that you can be quite unable to tell the difference. It is no more allowable to twist subjectively indistinguishable events into indistinguishable subjective events than to twist subjectively indistinguishable girls into indistinguishable subjective girls. (Hinton 1967: 226) When so conceived, this contemporary specification of disjunctivism seems to differ from Austin’s proposal in blocking the transition—which in our target passage Austin seemed to accept—from introspective or subjective indiscriminability to qualitative indistinguishability (at least, where the latter is understood as entailing qualitative sameness). Thus, although the target passage might sustain the attribution to Austin of a commitment to schematic disjunctivism, it doesn’t support the attribution of a commitment to the contemporary specification. Albeit that the contemporary specification of disjunctivism is rooted in the distribution of labour availed by Austin’s clarification of the distinction between perceptual and cognitive powers, is that the most we can say about Austin’s role in its development? Hinton himself suggests that there may be more to say here. Characterising his own stand against the slide from introspective indiscriminability to qualitative indistinguishability, he writes: The stand is of course like Austin’s. As just intimated, it is not confined to sensa—or, one may add, sensibilia. (Hinton 1980: 39) To what extent was Austin’s stand like Hinton’s? We’ve already noted a further point of likeness, in the form of Austin’s use of the distinction between acquaintance and propositional knowledge in order to make room for introspective error and uncertainty. Furthermore, in the passage leading up to the proposal considered above, Austin exploits the distinction in order to block transitions like that from indiscriminability to qualitative indistinguishability: I do not, of course, wish to deny that there may be cases in which “delusive and veridical experiences” really are “qualitatively indistinguishable”; but I certainly do wish to deny (a) that such cases

Settling a Question

147

are anything like as common as both Ayer and Price seem to suppose, and (b) that there have to be such cases to accommodate the undoubted fact that we are sometimes “deceived by our senses”. We are, after all, not quasi-infallible beings, who can be taken in only where the avoidance of mistake is completely impossible. (Austin 1962: 52; emphasis in original) Again, just before the target passage, Austin says: From the fact that I am sometimes “deluded”, mistaken, taken in through failing to distinguish A from B, it does not follow at all that A and B must be indistinguishable. Perhaps I should have noticed the difference if I had been more careful or attentive; perhaps I am just bad at distinguishing things of this sort (e.g. vintages); perhaps, again, I have never learned to discriminate between them, or haven’t had much practice at it. Austin (1962: 51; emphasis in original) Now, one difference between what Austin says here and Hinton’s proposal is that to which Hinton alludes: Austin’s eyes are on the objects of perceptual experience, rather than the experiential events. Another is that Austin focuses here on cases in which discrimination by sight is possible in principle, but not achieved, or achievable, by particular subjects. That is, he appeals to contingent, rather than principled, limitations on subjects’ powers of introspective cognition. However, although Austin’s attention is on the objects of experience, his claim about what does or doesn’t follow from discriminatory error is entirely general. Furthermore, Austin seems to drop the restriction to contingent failures of discrimination in the following passage: Again, when “the quickness of the hand deceives the eye”, it is not that what the hand is really doing is exactly like what we are tricked into thinking it is doing, but simply that it is impossible to tell what it is really doing. In this case it may be true that we can’t distinguish, and not merely the case that we don’t; but even this doesn’t mean that the two cases are exactly alike. Austin (1962: 52; emphasis in original) Here again, Austin’s focus is on the objects of perceptual experience. However, in that context, he gives the appearance of being prepared to disallow the transition from (a) its being impossible to discriminate through sight what the hand is doing from what it might be doing but isn’t to (b) the event of its doing the one thing and the event of its doing the other being exactly alike. However, since, as we’ve seen, Austin was prepared to disallow the transition from qualitative indistinguishability to exact likeness—at least,

148

Guy Longworth

where the latter amounts to sameness of nature—we still lack decisive evidence that Austin would have disallowed the earlier transition. So, Austin has to hand some of the materials that are required in order to construct the contemporary disjunctivist treatment of internally physically matching hallucinations. And there is some evidence that Austin was prepared to exploit those materials in order to block the problematic transition from indiscriminability to qualitative indistinguishability. However, as far as we can now tell, he didn’t make use of those materials in the way that Hinton did in order to provide an account of the ways matching hallucinations seem based on principled limits to introspective cognition. It is natural to wonder why. Insofar as Austin didn’t see clearly the possibility of adopting the contemporary specification of disjunctivism, that is plausibly due to the operation of three main factors. First, Austin seems not to have considered in detail the type of challenge to schematic disjunctivism, that arises from internal physical replication (see Martin 2007). As we saw, it was reflection on that challenge which motivated the contemporary specification. Second, although Austin considered contingent failures of introspective discrimination—failures due, for example, to inexperience or inattention— he didn’t have clearly in view the possibility of there being principled limits to the power of introspective cognition as such. Third, in those cases with respect to which Austin did apparently consider principled limits of discrimination—for example, in considering the upshot of principled limits to what it is possible to discriminate by sight—the nature of the target power would have made it plausible that indiscriminability in such cases corresponds with the sharing of qualities—albeit superficial qualities, like looks (see Austin 1962: 33–43; Martin 2007, 2010). Thus, exploiting the notional distinction between introspective indiscriminability and qualitative indistinguishability in order to respond to the challenge to schematic disjunctivism required attaining a view of introspective cognition on which seeming sameness needn’t correspond with sameness of superficial qualities (‘seemings’), and Austin didn’t clearly attain such a view (for developments of the required view, see especially Martin 2008; Soteriou 2016: 169–184). It is worth observing, however, that Austin’s rejection of an acquaintancebased model of introspection (Austin 1946) (as well as his differential treatment of “looks”, “appears”, and “seems”; Austin 1962: 33–43, as developed by Travis 2004) is distinctively congenial to the development of the required view (see again section 3).

6. Conclusion My aim has been to defend two suggestions about Austin’s relation to contemporary forms of disjunctivism. The first suggestion is that Austin

Settling a Question

149

failed to articulate or defend a central element in those contemporary forms of disjunctivism, the proposal that principled limits to the power of introspection can be exploited in order to block the transition from introspective indiscriminability to qualitative indistinguishability. The second suggestion is that Austin’s discussions of perception provided an intellectual environment that was distinctively hospitable to the later development of more specific forms of disjunctivism. In summary form, the three components of Austin’s position which shape that environment are the following: (i) his distinction between cognitive and sensory powers; (ii) his treatment of introspection as a cognitive, rather than sensory, power; (iii) his treatment of the cognitive power of introspection as fallible and, to that extent, limited. With those components of Austin’s position in view, it is possible to appreciate better his accomplishments whilst at the same time recognising the precise extent of the challenges that he bequeathed to his successors.

Note 1. I’m grateful to Thomas Crowther, Naomi Eilan, Mark Kalderon, Hemdat Lerman, Clayton Littlejohn, Matthew Soteriou, and Charles Travis for discussion and comments.

References Austin, John Langshaw, 1946, Other minds, Proceedings of the Aristotelian Society (supp. vol.) 20: 148–187. (Page references to the reprint in his 1979: 76–116.) ———, 1950, Truth, Proceedings of the Aristotelian Society (supp. vol.) 24: 111– 128. (Page references to the reprint in his 1979: 117–133.) ———, 1954, Unfair to facts, in J.O. Urmson and G.J. Warnock, eds. Philosophical Papers (3rd ed.). Oxford: Clarendon Press, pp. 154–174. ———, 1962, Sense and Sensibilia. Reconstructed from the manuscript notes by G.J. Warnock. Oxford: Oxford University Press. ———, 1979, Philosophical Papers (3rd ed.). Eds. J.O. Urmson and G.J. Warnock. Oxford: Clarendon Press. Hinton, John Michael, 1967, Visual experiences, Mind 76(302): 217–227. ———, 1973, Experiences. Oxford: Clarendon Press. ———, 1980, Phenomenological specimenism, Analysis 40(1): 37–41. Kalderon, Mark Eli, 2011, Before the law, Philosophical Issues 21: 219–244. Longworth, Guy, 2018, Enough is enough: Austin on knowing, in S.L. Tsohatzidis, ed. Interpreting J. L. Austin. Cambridge: Cambridge University Press, pp. 186–205. ———, 2018, The ordinary and the experimental: Cook Wilson and Austin on method in philosophy, British Journal of the History of Philosophy (Online First). doi:10.1080/09608788.2017.1413539 Martin, Michael G. F., 1997, The reality of appearances, in M. Sainsbury, ed. Thought and Ontology. Milan: FrancoAngeli, pp. 81–106.

150

Guy Longworth

———, 2002, Particular thoughts & singular thought, Royal Institute of Philosophy Supplements 51: 173–214. ———, 2004, Limits of self-awareness, Philosophical Studies 120: 37–89. ———, 2006, On being alienated, in T.S. Gendler and J. Hawthorne, eds. Perceptual Experience. Oxford: Clarendon Press, pp. 354–410. ———, 2007, Sense and Sensibilia. http://sas-space.sas.ac.uk/631/ ———, 2010, What’s in a look? in B. Nanay, ed. Perceiving the World. Oxford: Oxford University Press, pp. 160–225. McDowell, John, 1982, Criteria, defeasibility, and knowledge, Proceedings of the British Academy 68: 455–479. Pears, David, 1979, A comparison between Ayer’s views about the privileges of sense-datum statements and the views of Russell and Austin, in G.F. Macdonald, ed. Perception and Identity: Essays Presented to A. J. Ayer With His Replies to Them. London: The Macmillan Press, pp. 61–83. Price, Henry Habberley, 1952, Seeming, Proceedings of the Aristotelian Society (supp. vol.) 26: 215–234. Putnam, Hilary, 1994, Sense, nonsense and the senses, Journal of Philosophy 91: 445–517. Snowdon, Paul, 1980–1981, Perception, vision, and causation, Proceedings of the Aristotelian Society 80: 175–192. ———, 2008, Hinton and the origins of disjunctivism, in A. Haddock and F. Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 35–56. Soteriou, Matthew, 2016, Disjunctivism. London: Routledge. Travis, Charles, 2004, The silence of the senses, Mind 113(449): 57–94.

Part III

Epistemological Disjunctivism Prospects and Problems

8

Disjunctivism and Realism Not Naïve but Conceptual Sonia Sedivy

This chapter argues that John McDowell’s account of perception as both contentful and relational tends to go unnoted when the options for disjunctive theories are laid out. McDowell’s approach is important because it comes up the middle between ‘intentional’ and ‘relational’ views of perception. In doing so, it offers theoretical resources for explaining perceptual experience and its epistemic standing that purely relational views associated with naïve realism do not have. This challenge claims that naïve realist or purely relational views are a natural fit for disjunctive theories of perceptual experience. Though McDowell’s work has always emphasized that explaining perception’s objectivity and warrant is a single task, his account of both perceptual experience and its epistemic standing are reconstructed as epistemically motivated. I will focus on detailed work in papers dating between 2006 and 2013 to show how they open a unified approach to perception and its epistemic potential that turns on the claim that perception is both contentful and relational: that contents, and the broader context of capacities in which such contents figure, secure the perceiver’s relation to what she sees (cf. McDowell 2009, 2010, 2013a, 2013b). I will call views of this kind conceptual realism, though commonsense realism might be more apt. It is part of commonsense that understanding informs seeing since our phenomenological and intuitive sense of our experience is that it is always something of a kind that one sees. As Peter Strawson noted, if one is asked to describe one’s experience, one would say something like: “I see the red light of the setting sun filtering through the black and thickly clustered branches of the elms, I see the dappled deer grazing in groups on the vivid green grass” (Strawson 1988: 94). Ordinary intuitive descriptions convey that one sees individual members of kinds and instances of properties. Conceptual realism takes up this part of commonsense with a philosophically technical notion of conceptual capacities. Second, it is also “natural and intuitive” that perceptual experience “at its best makes aspects of objective reality present to us” (McDowell 2010: 245). Commonsense or conceptual realism explains this intuitive conviction—that

154

Sonia Sedivy

“some aspect of objective reality is there for a subject, perceptually present to her”—with the proposal that our conceptual capacities play a role in securing a subject’s relation to her surroundings (Ibid.: 245). Conceptual realism puts both commonsense commitments together, aiming to explain how understanding helps secure relatedness to objective reality. Here is one way to think about the situation that this chapter addresses. As the topic of disjunctive theories has evolved, a distinction has been drawn between epistemic and metaphysical types of disjunctivism. Epistemological disjunctivism is concerned with the epistemic standing of perceptual experiences as opposed to illusions or hallucinations. Metaphysical disjunctivism is concerned with the essential nature of perception and disjoins perceptual experiences from illusions or hallucinations in terms of their content or phenomenal character or something that gets at the very nature of perception. Most discussions state clearly that epistemological disjunctivism does not entail (some version of) metaphysical disjunctivism—in other words, one may put forward an epistemological account that disjoins the epistemic standing or warrant of perceptual experiences from other visual experiences without being committed to a specific theory of perception. But as Duncan Pritchard, for example, notes, a theory of perception that is itself disjunctive “seems to offer the most natural way of explaining” the epistemological thesis that one’s “reflectively accessible rational support is different because the very nature of one’s experiences is different” (Pritchard 2012: 24). The problem is that if we turn our attention to metaphysical disjunctivism, what tends to be cited is naïve realism; it seems to be the main player in the field. Naïve realism claims that the good case of a genuine perception involves the individuals and properties in the world, disjoining perceptions as a type of mental state from illusory or hallucinatory states. While naïve realists such as Mike Martin and William Fish recognize that “disjunctivism does not entail naïve realism” though “naïve realism entails disjunctivism,” (Fish 2009: 37) they also discount other variants of disjunctivism and focus on clarifying what they take to be a natural alignment. As Mike Martin puts it, “disjunctivism seeks to defend Naïve Realism” (Martin 2006: 361). Or, as William Fish writes, “the major motivation for endorsing disjunctivism would be to sustain naïve realism” (Fish 2009: 38). Naïve realism is joined by other relational views that hold that perception is purely relational: objects figure in perceptual experience as constituents by virtue of a primitive or ‘acquaintance-like’ (or subpersonal) relation that does not draw on the perceiver’s understanding of her situation or of herself as a perceiver (Brewer 2011; Campbell 2002; Travis 2004, 2013). This increasingly common alignment—between purely relational, naïve realist and disjunctive accounts of perception—distorts our understanding of the explanatory options insofar as we do not recognize the alternative provided by conceptual realism. It overlooks the following option:

Disjunctivism and Realism

155

“Perception makes knowledge about things available by placing them in view for us. But it is precisely by virtue of having content as they do that perceptual experiences put us in such relations to things” (McDowell 2013a: 144). McDowell’s point in part is to distinguish between content that is not itself relational but that figures in relational perceptual states of a subject—de se content—and content that is itself relational—de re content of singular thoughts made possible by a perceiver’s relational experiences. The chapter will proceed in the following four steps. The first part offers a reconstruction of McDowell’s account of perceptual experience; the second part shows the explanatory resources this provides for understanding the epistemic standing of perceptual experiences.1 The third section explores how conceptual and purely relational, naïve realism adjudicate a range of cases differently. I will explain how commonsense or conceptual realism offers theoretical resources for explaining the epistemic standing of perceptual experience in various situations that purely relational, naïve realist theories cannot. The fourth section offers an example of how conceptual realist theories might disagree or diverge from the detail of McDowell’s approach.

1. Perceptual Presence: Contentful and Relational McDowell suggests that perception makes individuals and their properties immediately present or “there for a subject,” and this is possible through a triad of explanatory conditions: (i) de se content; (ii) a context of capacities; (iii) a world context of a specific configuration. The three conditions specify that seeing is a relation between a perceiving person and their surroundings, a relation that is secured through a specific kind of content in a context of capacities. The claim that seeing is relational is stronger than saying that it has content that is veridical. It is possible that truth conditions might obtain and yet a perceiver might not be perceptually related to her surroundings so that her surroundings would not be present to her. To understand just what immediate presence is, we need to examine the three conditions and how they work together. First, the account explains the distinctively visual contribution to perceptual experience: de se content that places objects and their visible properties in relation to oneself. Such content is highly specific in locating an object and its properties. Nevertheless, it can be indicated through descriptions—that lose some detail—either in propositional or objectual form, ‘that there is something red and rectangular in front of me’ or ‘something red and rectangular in front of me.’ For such content to be part of a seeing that makes a particular object immediately present, the second and third conditions need to obtain. McDowell argues that such visual content needs to be integrally connected to capacities of understanding that have the specific character

156

Sonia Sedivy

that they allow for self-determination. In other words, the second condition strives to capture what is distinctive of human mature perception in terms of the Kantian idea that we have the capacity for self-determination whereby we can choose to act or to refrain from acting, and that this capacity gives human perception its objectivity. In the case of perception, our capacity for self-determination does not enable us to choose what we see (except in the trivial sense that we can choose where we look). We cannot change what objects or properties we see. But we can recognize that our situation is not such that an object would be present to us. The potential to recognize that the circumstances may be inadequate is due not to a single capacity, but to a broader context of capacities that we might call a context of understanding, which makes it possible to question or examine our perceptual circumstances either on occasion or after the fact. In view of the extensive debate over concepts and their role in perception over the past several decades, the point of departure for McDowell’s approach is that there “is not . . . a universally shared idea of conceptual capacities, which determines a subject matter about whose properties people disagree. The notion of the conceptual can be used in a variety of ways, for a variety of purposes” (McDowell 2009: 32). The philosophical notion of concepts is often linked to generality and to a ‘fine-ness of grain’ that goes with attributing contents that have truth-value. This helps explain the character of thoughts or language uses. Alternatively, the notion of conceptual capacities might be reserved for capturing the way a wide range of animals respond to kinds that are their reasons for acting. If we wish to take the latter course, McDowell’s approach would need to adopt a different term for identifying the capacities at issue. The point stands that such capacities need to be identified, however we wish to designate them. McDowell uses the notion of the conceptual to highlight the Kantian connection between rationality and understanding, the capacity for generating and applying concepts. As noted earlier, the over-arching idea is that humans have the potential for rational self-determination, whether it be in action, thought or perception. One way that McDowell has tried to put this point is that capacities are conceptual in that when we think or act, what we reason about or act on are not simply reasons but reasons as such—which means reasons that we can consider as such so that they do not compel an action or thought, but allow for self-determination in whether we think or act on that reason.2 To be sure, we do not need to explicitly entertain reasons as such to arrive at further beliefs or actions in most cases. But capacities to do so—which may or may not be exercised—must be available. When it comes to perceptual experience, “one does not choose to accept that things are the way one’s experience plainly reveals that they are” (Ibid.: 139). Rather, “recognizing reasons as compelling is itself an

Disjunctivism and Realism

157

exercise of one’s capacities for rational self-determination” insofar as the potential is there to question one’s experience or its circumstances. It is in this sense that perceptual experience is informed by capacities for self-determination that are internal to it. This is what gives perceptual experience its objectivity, its purport to be of objects in the world that do not induce response from us, but to which we stand in a relation that we understand and that we can examine should need arise. This proposal emphasizes points of analogy between action and perception even as it recognizes disanalogies. First, both action and perception must be amenable to rational control, though how rational control figures in perception and action differs. As noted, we can assess experiential states but not choose them as we can choose our actions. Second, action and perception are analogous in their largely unreflective nature; we carry on perceiving and we carry on with our activities to a large measure unreflectively. Rational self-control requires that we can entertain an action or a perception by articulating what we have done or what we have perceived—or perhaps what we are doing or what we are seeing. In neither case does it mean that the action or perception is itself propositional or needs to involve a state with propositional content. This is not considered especially problematic in discussions of skillful knowhow where the idea that knowledge can be manifest in action seems au currant (for example, see Wiggins 2009; Stanley 2011a, 2011b; Kremer 2016). The analogous view should also be unproblematic for perception. It counters the charge that conceptualism renders perceptual content language-like or hyper-intellectual. What is being proposed is that de se contents are such that they can be expressed in a form requisite for entertaining a possibility of error. This is what it means to say that conceptual capacities are internal or integral to experiential contents: a de se content that places an object in relation to me is such that it can be expressed in either objectual or propositional form, which means that it figures for me as a reason as such. Last but not least, the third condition turns to the other relatum—not the subject and her contents and capacities, but the perceptible scene or configuration. The object must be how and where de se content places it for the perceiver to stand in perceptual relation to it. This third condition underscores the relational nature of perceptual experience: if one of the relata—an object and its properties—is missing, the relational state cannot obtain. These three conditions show how the conceptual realist notion of relational perceptual presence is stronger than the claim that perceptual experiences have truth conditions. Consider seeing a bird flying. On the truth or veridicality condition view, one sees a bird flying if and only if the conditions are such that there is a flying bird. McDowell argues that this is not sufficient for a relation to the bird whereby the bird is objectively present to one, which also means that it can figure as a reason as such.

158

Sonia Sedivy

It is crucial that the relation is neither independent of content nor just a matter of content. Rather, the relation is secured by the capacities in the context of which the content obtains, and the content is such that a context of understanding can secure a relation because the content places the object in relation to oneself. A slightly different way to make the point is that seeing or perceptual experience is multi-dimensional—it involves a de se content, which places objects and properties uniquely in relation to me, together with a context of capacities whereby I can understand that the circumstances are such that the object is there, is present to me or not. This complex state is one whereby it appears to the subject that ‘things are a certain way.’ The analysis shows that how things appear to a subject in a case of seeing involves all three dimensions working together. If things appear to a subject to be a certain way without all three of the dimensions working together, such states of appearing would be defective cases of seeing or belong to a different kind altogether. McDowell writes: It is part of the point of my disjunctive conception of experience that having an aspect of objective reality present to one entails having it appear to one that things are a certain way. But that is not to say that having an aspect of objective reality perceptually present to one can be factored into some non-mental conditions and an appearance conceived as being the mental state it is independently of the non-mental conditions. The factoring fails; the state is the appearance it is only because it is a state of having something perceptually present to one. (2010: 251) Though this approach disjoins perceptual experiences from states where things appear the same to the subject, it is not a metaphysical disjunctivism: it does not deny that there is a type of state—an appearance— common to experiences that are perceptual, hallucinatory and illusory. McDowell argues that a description that characterizes how things appear to a subject captures a genuine commonality, so that the description identifies a type of state that may be common to both defective and nondefective perceptual experiences, as well as illusory states and hallucinatory states (Ibid.: 244). The disjunctive phrasing makes clear that when it appears to a subject that an object is in front of her, the state she is in is either one of “having an aspect of objective reality perceptually present to [her]” or one where it seems that an aspect of objective reality is perceptually present to her.

2. Perceptual Presence and Perceptual Warrant Insofar as perception is a capacity whereby individual objects and their properties are present to a person in McDowell’s sense, it is a capacity

Disjunctivism and Realism

159

that provides self-consciously “indefeasible, and so knowledge constituting, warrant for belief about the environment” (Ibid.: 247). The detailed explanation of perceptual presence offers the following account of the epistemic standing of experiences. First, “the epistemic significance of an experience consists in its having content in the way that it does” (McDowell 2013a: 147; emphasis in original). That is, the epistemic significance of perceptual experience needs to be understood in terms of the way in which its content together with the other conditions yield perceptual presence. This provides resources for explaining epistemic significance not only of cases where everything works, but of cases where something is amiss. I will return to a more detailed analysis in the next section. Second, the epistemic significance of a particular experience is a matter of the capacity of which it is an exercise. From an epistemic point of view, perception is “a capacity to be in positions in which one knowingly has . . . environmental realities present to one” (Ibid.: 151). This follows from approaching mature human perception as integrated within capacities that allow for self-determination. Insofar as agents have capacities not just for acting on reasons but on reasons as such—which involves capacities to reflect on those reasons—then the objects and properties that are thereby perceptibly present provide warrant. One’s knowledge that there is something red and rectangular in front of one includes knowledge of its own credentials as knowledge. And it is the knowledge it is because it is a non-defective act of a capacity to know such things through perception. (McDowell 2013a, p. 151) The account of perceptual experiences as exercises of a capacity whereby objects are immediately present sets up McDowell’s discussion of fallibility. He argues that most discussions treat particular perceptual experiences as fallible, but it is capacities that are fallible; exercises are either defective or non-defective. In part, the point is that the fact that a capacity may be fallible does not change what the capacity is or what it is for. McDowell introduces the example of a basketball player’s capacity to shoot free throws, shots that yield one point in a specific context (McDowell 2010: 245–246). This capacity is fallible, with many defective or flawed exercises even by star players. But the fact that a particular free throw misses the basket does not make it something other than a free throw—it is a flawed free throw. Similarly, if the correct analysis of the capacity of perception is that it makes objects and their properties present to the perceiver, the capacity is such that particular perceptual experiences fail to have conclusive warrant when all does not go well. Defective perceptual experiences do

160

Sonia Sedivy

not provide conclusive warrant just as defective free throws do not yield points—but non-defective ones do. This analogy may seem to run afoul of a significant disanalogy between free throws and perceptual experiences: experiences have subjective character and on many accounts, including McDowell’s, their warrant must be part of their subjective character (McDowell 2013a: 150). “Someone who has knowledge of this [perceptual or experiential] sort must be in a position to know the warrant by virtue of which her state counts as knowledge” (Ibid.: 148). Since some experiences are defective in a way that one might not discern, it may seem that the requirement cannot be met. Given that “one can take an experience to make knowledge available when it does not” (Ibid.: 151), the analogy might seem to fail. But according to McDowell, this line of reasoning makes a mistaken inference from the fallibility of a capacity: it is a mistake “to infer that when one is not being fooled, one’s experience does not put one in a position to know that one is not being fooled” (McDowell 2010: 246). The reply relies on his analysis of the perceptual capacity and on the relationship between a capacity and its exercises. If it is the nature of perceptual capacity to make objects present, then one is in a position to know that one is not being fooled when objects are perceptually present to one. If perceptual presence lies in the way that one has de se content— in a context of capacities that allow for assessment of oneself and one’s conditions—then perceptual warrant is available in good cases. One may be mistaken. But if one is not mistaken, one is exercising a capacity whereby an object is present to one, and one has indefeasible warrant. At this point, the argument that fallibility attaches to a capacity, while exercises either are flawed or do not kick in. The capacity to get into such positions [where one knows that is in a position to have indefeasible warrant for believing that things are a certain way] is fallible. It does not follow that that cannot really be what it is a capacity to do. (Ibid.: 246) To return to the analogy with free throws, the mistaken inference from fallibility would have one hold that because one makes mistakes in shooting free throws, when one makes no mistakes, one nevertheless does not shoot a free throw. McDowell’s counter is that the subjectivity of experience does not cancel out the basic point but is part of the point: when an exercise of a perceptual capacity is not defective, one is in a perceptual relation to objects with conclusive warrant for perceptually based belief. My disjunctive approach to experience is a way of expressing this rejection of a faulty inference from fallibility. The disjunctive formulation states the point positively: of experiences that seem to reveal to

Disjunctivism and Realism

161

one some aspect of how things are in one’s environment, some make that aspect of reality perceptually present to one, whereas the other only seem to do that. (Ibid.: 246) In sum, McDowell’s account of our perceptual capacity gives the following epistemic picture: On the content conception, the epistemic significance of an experience consists in its having content in the way that it does. An experience that is a seeing can be like an experience that merely appears to put its subject in touch with a corresponding environmental reality in respect of what content it has. But a seeing is unlike a mere appearing in how it has its content. Seeings . . . put the subject in a position that leaves open no possibility of things not being as they would be believed to be in suitably related beliefs. (2013a: 147)

3. Explanatory Potential: Conceptual vs Naïve Realism Consider the following scenarios from “Perceptual Experience: Contentful and Relational” (McDowell 2013a: 152–153). First, consider an experimental situation where (i) the lighting is such that in many cases, objects do not look to have their colours, while in some they do; and (ii) this is not detectable. In such a scenario, according to McDowell’s approach, all of the subject’s perceptual experiences are defective even though some are veridical and normally caused. The subject exercises her perceptual capacity, but the conditions are such that the subject is not “in a position to know the thing’s colour by looking.” The subject has experiential de se content, whereby it appears to her that an object has a specific colour. But the subject is in contexts where it is not possible to exercise her understanding of perceptual situations, so she is not perceptually related to the object’s colour even in the specific cases where the object has the colour that it appears to her to have. Second, consider the same experimental scenario—where some lighting conditions are undetectably unsuitable while others are suitable—but the subject has been told that she is in this situation. This means that even in cases where the lighting “is suitable for knowing colours by looking,” the subject understands that she cannot tell when she is in such a case. Because of her knowledge of her situation, she cannot exercise her capacities to reflect on the conditions in which she finds herself. As a result, “she is not entitled to take its colours to be visually present to her in her experience” (McDowell 2013a: 153). As in the first case, she will have de se experiential content whereby it appears to her that objects have certain colours.

162

Sonia Sedivy

In some cases, these contents are veridical and normally caused; in other cases, they are not. But even where the contents are veridical and normally caused, the subject is not perceptually related to an object’s colour because she knows that her capacity to reflect on her situation has been undercut. Compare purely relational, naïve realist accounts of such cases. The following comparisons will be at a level of generality aimed at the claim that subjects stand in a perceptual relation to objects and properties that is not integrated with conceptual capacities and does not involve contents. The proposed perceptual relation might be a matter of ‘bare acquaintance’ or of a primitive personal level capacity or of subpersonal visual processes not subject to cognitive effects. Across both scenarios, purely relational theories would hold that the subject is related to her surroundings in those cases where the lighting is suitable. This is because the relational state obtains when the conditions are suitable. The perceiver’s understanding of her circumstances—her conceptual capacities—are not integral to but subsequent to the relational capacity. Given that the coloured object figures in her experience, it would be hard for purely relational theories to hold that the subject does not see the object. And insofar as such theories hold that the subject sees the colour of the object, it is hard to deny epistemic standing that would belong to seeing. These scenarios show that purely relational theories push the epistemic account into an externalism according to which the subject sees the object’s colour regardless of her reflective access to warrant. Though purely relational theories may argue that the subject’s understanding of her situation may be such that she does not take her perceptual experience to constitute knowledge or provide warrant for further empirical belief, it is explanatorily complex to do so given that the theory holds that the experience is relational and factive independently of and antecedently to any exercise of the subject’s understanding. One might go further and argue that given that the first scenario is analogous to ‘barn façade’ cases, naïve realism does not deliver the epistemic result many believe to be needed: to distinguish cases where the subject has perceptual warrant from cases where she does not. In the second scenario, the fact that the perceiver knows that she is in an experimental situation—in which she cannot detect suitable from unsuitable conditions—makes no difference to whether she is related to the objects and their colours which figure in her experiential states when conditions are suitable. In sum, the two scenarios bring out the following points. 1. Purely relational and conceptual realist approaches yield different results in all these cases, and the two approaches pull in different epistemic directions.

Disjunctivism and Realism

163

2. Conceptual realism and naïve realism explain the good and bad cases differently. Naïve realism distinguishes good and bad cases in both scenarios: where the lighting is suitable, coloured objects figure in the subject’s perceptual experience; where the lighting is not suitable, coloured objects do not figure in her perceptual experience. Conceptual realism holds that all of the cases across both scenarios are bad for perception—in none of the cases is the subject related to coloured objects. That is, conceptual realism denies that the colours are present to her. 3. Conceptual realism explains that the two scenarios impact differently on the perceiver’s experience. Though both scenarios undermine a subject’s capacities to understand her situation, they do so differently: in one case because she does not know that there are undetectable differences in lighting, in the other case because she does know this. Though naïve realism can countenance that a subject’s understanding of her situation is different in the two scenarios, this does not make a difference as to whether coloured objects figure in her perceptual experience. Even though the subject knows that she is in circumstances where she cannot distinguish suitable conditions from unsuitable ones, objects and properties figure in her experiences nonetheless, so that the experiences are factive. 4. Conceptual realism explains that across both the first and second scenarios, the subject is in a type of state that is common to defective cases of seeing and non-defective cases of seeing—it appears to the subject that objects have certain colours. In all of these cases, though it appears to the subject that objects have certain colours, she cannot tell that they do. 5. Conceptual realism provides an explanation of how it is that things may appear the same to the perceiver across veridical and nonveridical cases. The subject is in a type of state where de se contents may be the same. 6. Purely relational accounts deny a commonality in the good and bad cases in both scenarios because objects figure in experiences in a way that does not draw on a subject’s capacities and does not involve contents. Insofar as such accounts deny that contents are involved in a relational state, they deny a key resource for explaining commonality in a subject’s experience in good and bad cases. 7. These differences in explanation—of whether a subject is exercising her perceptual capacity successfully and whether she has perceptual warrant—stem in part from the fact that the subject’s understanding of her circumstances can make a difference on the naïve realist account only after the relation is secured. According to purely relational accounts, the subject’s state is factive independently of her understanding of the situation.

164

Sonia Sedivy

4. Conceptual Realism: Contents and Relations McDowell’s explanation of perceptual experience as contentful and relational opens up the possibility of diverging accounts of both dimensions. Specific theories might diverge in their accounts of the content of perceptual experiences or of the conceptual capacities in the context of which such contents figure. For example, I would disagree with McDowell’s characterization of visual de se contents as obtaining independently of the subject’s conceptual capacities with respect to the kind to which an object belongs. McDowell suggests that what an experience has in view is something red and rectangular, for example, but not a book or a book cover or a copy of the Tractatus. So even though cognitive capacities are involved in having something red and rectangular in view, these other things that I know by virtue of having the experience are “acts of cognitive capacities that would not need to be in operation for me to have the thing in view at all” (McDowell 2013a: 156, my italics). This makes a distinction between the cognitive capacities drawn into having something in view at all, and knowing the sort of thing it is. I suggest that taking the aesthetic dimension of perceptual experience into account indicates that this concession to arguments against the ‘pervasiveness of the conceptual’ is mistaken. Having an object in view at all may involve aesthetic impact, and such impact draws on our understanding of what we are looking at. To telescope extensive discussion in aesthetics for the purposes of this chapter, let’s focus the issue in terms of two key claims.3 First, aesthetic response depends at least on the high-level kind to which an object belongs—natural, artefactual or artistic—and arguably on more specific kinds as well. (I will leave more specific kinds out for this discussion.)4 Second, at least some aesthetic responsiveness is experiential and a dimension of perceptual experience (see Walton 1970). If the first contention from aesthetics is correct—that aesthetic response depends at least on higher-level kinds to which the object belongs—then it follows that insofar as aesthetic responsiveness can be perceptual, the requisite understanding needs to be able to inform perceptual experience. Since aesthetic responsiveness may be integral to any perceptual experience, understanding of the kind to which the object belongs needs to be in operation. But what about appearances, and what it is to have a thing in view at all, which are of importance to disjunctive theories? Can aesthetic responsiveness be part not only of having a thing perceptually present to one, but also of having a thing in view at all? This is the key question with respect to McDowell’s view that we can have a coloured and shaped object in view without any understanding of the kind to which it belongs. Let’s take this in two steps that offer illustrative examples to argue, first, that perceptual experience of colour and shape may include aesthetic responsiveness, which draws on understanding; and second, that having

Disjunctivism and Realism

165

a thing in view at all involves aesthetic responsiveness (so that it draws on understanding). First, consider two monochromatic canvases with the same paint and colour, where one is a paint sample and the other an artwork. A large stretched canvas that is painted all over with a certain shade of grey colour would have one range of aesthetic properties if it is a decorative sample in a fancy paint shop—for example, drab, gloomy or elegant. But a canvas of the same size, painted all over with the same shade of grey colour, would have a different range of aesthetic properties if it is a painting made by a specific artist at a particular time, such as Gerhard Richter’s Grau 1970 (247–10). A work’s aesthetic properties would be connected to ‘what’ it conveys or ‘what’ it is about—a feeling, a mood, a content. One might find the same grey colour gloomy for example, but in the case of the artwork it would be gloominess intentionally conveyed, rather than the gloomy effect of a certain colour. Perhaps an all-over-grey painting might seem not to convey anything—it might seem strangely neutral or ambiguous, lacking in a forceful effect or content. But then this would be precisely its content and aesthetic impact—ambivalence or absence, or a withholding of message. This example illustrates that aesthetic response depends on the high-level kind to which an object belongs, here artefact versus artwork; and it suggests that the response at issue may be perceptual, since it is a response to the object’s colour (and shape, size, texture). One might counter that aesthetic properties are higher-level properties that depend on basic visual properties such as colours, contours and perhaps shapes to draw a distinction in kind between aesthetic and basic visual non-aesthetic properties. But monochromatic paintings put pressure on distinguishing the basic property of colour and the aesthetic property that is that same colour. Return to a monochromatic work that is an expanse of a single colour and has aesthetic effect, such as Gerhard Richter’s all-over-grey painting, Grau 1970 (247–10). There is only one colour, and that colour is aesthetically present to one. A distinction in kind between aesthetic properties and basic visual properties would entail that the aesthetic impact of the single colour depends on the nonaesthetic property which is also that single colour. The one colour is both an aesthetic and a non-aesthetic property. At this step we are at the beginning of more extended argumentation that lies beyond the scope of this chapter. But this should suffice to indicate that we can avoid undue complexity by maintaining the intuitive view that the colour of the monochrome has aesthetic impact; it is an aesthetic property. Undercutting a distinction in kind between aesthetic and basic nonaesthetic properties such as colours and contours suggests that (i) perception of colours and contours can include aesthetic impact, and (ii) involves understanding of the kind to which the particular coloured object belongs.

166

Sonia Sedivy

But what about the claim that having the object in view at all does not need to involve understanding of the kinds to which the object belongs? Let’s return to our grey monochromes, but in the following scenario. I enter a fancy paint store and consider the grey colour of a large stretched canvas, displaying the latest ‘in’ shade. I find it too dark and gloomy to be surrounded by it in a living space. Now, suppose I learn that I have actually entered a gallery done up like a paint store, so that I am in an immersive installation space that probes our differential responsiveness to artefacts and artworks, to a grey monochrome paint sample and painting. It is safe to say that my experience will change. What I am seeing—the grey rectangular object, including its aesthetic impact—would change as a whole. The experiential change in perceiving an installation in place of a paint store—and, more specifically, a paint sample in place of an artwork—involves a change in how the rectangular object in front of me appears to me, and that change involves aesthetic impact. Some change in aesthetic property or impact is integral to the whole in the immediate experiential shift from sample to work. This change is explained by my understanding of the object I am looking at, incorrect or correct as it may be. Insofar as I was wrong in apprehending a paint sample, the content that explains how the object appeared to me includes its aesthetic impact or aesthetic property. When I learn that I am looking at a painting, the content that explains how the painting appears to me also includes its aesthetic impact or aesthetic property. Though these considerations touch only the tip of the iceberg that aesthetics provides, they should suffice to suggest the following. (i) In any particular case, perceptual experience may involve aesthetic responsiveness, and such responsiveness necessarily draws on understanding. (ii) In any particular case, the way an object appears, and hence the content that explains such appearance, may include aesthetic responsiveness; such aesthetic appearance and content necessarily involves understanding. (iii) In any particular case, having an object in view at all may involve aesthetic responsiveness, and hence having the object in view at all would necessarily involve understanding. In short, since any particular experience and appearance may be aesthetic, our understanding is internal to having an object in view at all. This discussion illustrates how McDowell’s approach opens up a range of topics—for example, concerning the nature of the contents, perhaps de se, that need to figure in a context of conceptual capacities to secure perceptual relatedness whereby an object and its properties are perceptually present to one.

5. Conclusion This chapter has focused on McDowell’s account to illustrate the explanatory resources that become available if we explain perceptual relatedness

Disjunctivism and Realism

167

in terms of contents and capacities of understanding. I have not argued directly that such an approach is correct—that is beyond the scope of this chapter—though showing its explanatory power is part of such an argument. The main aim has been to show that commonsense or conceptual realism makes good sense of our perceptual capacity—and that it is important to recognize the position when we think about disjunctive accounts of perceptual experience and its epistemic standing. In particular, conceptual realism offers an important alternative to naïve realist, purely relational approaches with which ‘disjunctivism’ has come to be readily associated.

Notes 1. In presenting the account of perceptual experience to show the explanatory resources it offers for epistemic issues, I am not going against McDowell’s clear statement that he has argued for two disjunctivisms—one concerning singular thought and one “about the epistemology of knowledge warranted by perceptual experience”—and that the latter is not based on or does not “exploit considerations about the former.” See McDowell (2013b). McDowell’s point is that singular thought is made possible by perceptual experience and that the nature of its content—de re—content is different from the de se nature of perceptual content, as I will explain. 2. McDowell makes this connection by contrasting the capacity to act for reasons (which we share with animals) with the capacity to act for reasons as such, which is distinctive to us. 3. The following discussion draws on a more extended discussion in Sedivy (2018). 4. I also leave out of the discussion here Immanuel Kant’s seeming proposal in the Critique of Judgement that pure judgements of beauty do not involve any conception of the object, even a high-order one.

References Brewer, Bill, 2011, Perception and Its Objects. Oxford: Oxford University Press. Campbell, John, 2002, Reference and Consciousness. Oxford: Oxford University Press. Fish, William, 2009, Perception, Hallucination, and Illusion. Oxford: Oxford University Press. Kremer, Michael, 2016, A capacity to get things right: Gilbert Ryle on knowledge, European Journal of Philosophy 25: 25–46. Martin, Michael, 2006, On being alienated, in T.S. Gendler and J. Hawthorne, eds. Perceptual Experience. Oxford: Oxford University Press, pp. 354–410. McDowell, John, 2009, Conceptual capacities in perception, in his Having the World in View, Essays on Kant, Hegel, and Sellars. Cambridge, MA: Harvard University Press, pp. 127–144. ———, 2010, Tyler Burge on disjunctivism, Philosophical Explorations 13: 243–255. ———, 2013a, Perceptual experience: Both relational and contentful, European Journal of Philosophy 21: 144–157.

168

Sonia Sedivy

———, 2013b, Tyler Burge on disjunctivism (II), Philosophical Explorations 16: 259–279. Pritchard, Duncan, 2012, Epistemological Disjunctivism. Oxford: Oxford University Press. Sedivy, Sonia, 2018, Aesthetic properties, history and perception, The British Journal of Aesthetics 58(4): 345–362. doi:10.1093/aesthj/ayy039 Stanley, Jason, 2011a, Know how. Oxford: Oxford University Press. ———, 2011b, Knowing (how), Noûs 45: 207–238. Strawson, Peter F., 1988, Perception and its objects, in Jonathan Dancy, ed. Perceptual Knowledge. Oxford: Oxford University Press, pp. 92–112. Travis, Charles, 2004, The silence of the senses, Mind 111: 57–94. ———, 2013, Perception: Essays After Frege. Oxford: Oxford University Press. Walton, Kendall, 1970, Categories of art, Philosophical Review 79: 334–367. Wiggins, David, 2009, Knowing how to and knowing that, in H.J. Glock and J. Hyman, eds. Wittgenstein and Analytic Philosophy: Essays for P.M.S. Hacker. Oxford: Oxford University Press.

9

Epistemological Disjunctivism and Its Representational Commitments1 Craig French

1. Introduction As it is standardly presented, epistemological disjunctivism involves the idea that paradigm cases of visual perceptual knowledge are based on visual perceptual states which are propositional—states of seeing that p (McDowell 1982, 1995, 1998, 2008; Haddock and Macpherson 2008b; Pritchard 2012, 2016). I look at the crow perched in the tree, in excellent perceptual conditions, with fully functioning perceptual and cognitive capacities, and come to know that the crow is black. I know this on the basis of visual perception. And the epistemological disjunctivist spells this out as follows: I have this knowledge in virtue of the fact that I can see that the crow is black. Given this, the orthodox version of epistemological disjunctivism takes on controversial commitments in the philosophy of perception. Specifically, it is committed to the claim that some of our perceptual states are propositional attitudes, and hence have propositional representational contents. And, in virtue of this, it is committed to the more general claim that some of our perceptual states are representational, where here this is understood merely as the idea that some of our perceptual states have representational contents. These are the representational commitments of orthodox epistemological disjunctivism. Must epistemological disjunctivism involve these commitments? Must the view be hostage to the fortunes of representational views in the philosophy of perception? I don’t think so. Here I will argue that we can take epistemological disjunctivism in a new direction and develop a version of the view free of these representational commitments. I will develop the non-orthodox version of epistemological disjunctivism I set out in French (2016)—what I will call the thing-seeing approach. The basic idea is that instead of conceiving of knowledge grounding perceptions as states in which one sees that such-and-such is the case (e.g., that the crow is black), we should instead conceive of them as states or episodes in which one sees a thing (e.g., a black crow), we should conceive of them as thing-seeings.

170

Craig French

I’ll suggest that we can cast such seeings in this knowledge grounding role without conceiving of them as representational. But this is because we can put thing-seeings to epistemological work, in the framework of epistemological disjunctivism, whilst remaining neutral on whether or not they are propositional, or representational at all—call this the neutrality claim. The point, then, is not to replace epistemological disjunctivism’s controversial representational commitments with controversial non-representational commitments. The point is, rather, that epistemological disjunctivism can be developed with fewer commitments in the philosophy of perception than is usually appreciated. I’ll proceed as follows. First, I’ll outline the orthodox form of epistemological disjunctivism and its controversial representational commitments a little more fully. I’ll then set out the thing-seeing approach and some of what motivates it. Finally, I’ll consider and respond to two challenges to the neutrality claim. The first challenge is that unless states or episodes of thing-seeing are representational, they cannot provide factive rational support. The other challenge—to which I devote most of the discussion in this chapter—is that unless states or episodes of thing-seeing are representational, they cannot rationalize perceptual beliefs.

2. Epistemological Disjunctivism What is epistemological disjunctivism? I understand it as a view about specifically visual perceptual knowledge. The view involves a negative claim underpinned by a positive claim. Thus, Duncan Pritchard states that it is the rejection of the idea that the . . . rational support one possesses in favour of one’s perceptual belief is the same regardless of whether one is having a normal veridical perceptual experience as opposed to being the victim of an introspectively indistinguishable experience which is in fact deceptive or untrustworthy in some way (e.g., a hallucination). (2011: 434) [The] two rational standings are radically different in kind (this is what makes this epistemological proposal disjunctivist). (2012: 16) We can thus put the negative strand of epistemological disjunctivism in this way. The Negative Claim The rational support in the Good and Bad Cases is not of the same nature or fundamental kind.

Representational Commitments

171

Epistemological disjunctivism also involves a positive thesis which underlies the negative claim. Pritchard puts the positive thesis like this: The Core Thesis In paradigmatic cases of perceptual knowledge an agent, S, has perceptual knowledge that ϕ in virtue of being in possession of rational support, R, for her belief that ϕ which is both factive (i.e. R’s obtaining entails ϕ), and reflectively accessible to S. (2012: 13; emphasis in original)

A paradigmatic case of visual perceptual knowledge is one where that knowledge is grounded in a perfectly veridical case of visual perception. The nature of the rational support provided by these cases of perception, the Core Thesis tells us, is such that it is reflectively accessible, and factive. How does this positive claim underpin the Negative Claim? Here, factivity is doing the work. To say that a kind of rational support is factive is to say that if rational support of that specific kind obtains, then that guarantees the truth of the belief it supports. So the Core Thesis tells us that in a perfectly veridical experience of a black crow before me, say, the rational support to believe that there is a black crow there, which perception provides, is, in its nature, factive. But now consider the rational support I have in a subjectively indistinguishable but non-veridical hallucinatory experience. Such rational support is not, in its nature, factive. Thus, given the Core Thesis, Bad Case rational support differs in its nature to Good Case rational support. Epistemological disjunctivism, understood along these lines, is standardly underpinned by a more specific commitment. Once again, we can draw on Pritchard’s presentation, and we can put the further specification of the Core Thesis, like this: The Specification The particular kind of rational support that the epistemological disjunctivist claims that our beliefs enjoy in paradigm cases of perceptual knowledge is that provided by seeing that the target proposition obtains. So when one has paradigmatic perceptual knowledge of a proposition, p, one’s reflectively accessible rational support for believing that p is that one sees that p. (Ibid.: 14)

The orthodox epistemological disjunctivist claims that states of seeing that p themselves constitute the rational support for belief had in paradigm cases of perceptual knowledge.2 Such states are factive, so if we assume that they are reflectively accessible, then this works just fine as a specification of the Core Thesis and gives us a more complete version of epistemological disjunctivism.

172

Craig French

3. Representational Commitments The orthodox epistemological disjunctivist holds that the visual states which underlie paradigmatic cases of visual perceptual knowledge are states of seeing that p. The orthodox epistemological disjunctivist is therefore committed to the claim that some visual states are propositional attitudes with propositional representational contents, and hence that some visual states are representational. As noted in the introduction, these are the representational commitments of the orthodox form of epistemological disjunctivism. These commitments are controversial in the philosophy of perception. It is not obvious that any states of perception are propositionally structured, or that any states of perception are representational at all.3 Recent philosophical literature on perception bears this out. Some think that perception is representational but deny that it is propositional (e.g., Burge 2010; Crane 2011). Others claim that perception isn’t even representational (e.g., Travis 2004; Weir 2004; Hutto and Myin 2012).4 Now, I don’t propose to get into a discussion here about whether any visual states are propositional or representational (for recent discussion, see the articles in Brogaard 2014). And I am certainly not objecting to the orthodox epistemological disjunctivist for having controversial commitments in the philosophy of perception. I’m merely highlighting that the orthodox epistemological disjunctivist has these controversial commitments. But why is this worth highlighting? Because it reveals an underexplored way in which epistemological disjunctivism is vulnerable. For the view is controversial not only from within epistemology, but also from within the philosophy of perception. Those who deny that perception is ever representational or propositional must reject orthodox epistemological disjunctivism. Arguments against the view that perception is ever representational or propositional are arguments against orthodox epistemological disjunctivism. In light of this, it is natural to ask whether the epistemological disjunctivist needs to take on controversial representational commitments. Are these at the heart of epistemological disjunctivism? Or is there a way of upholding epistemological disjunctivism without these controversial commitments in the philosophy of perception? If there is, then it may turn out that non-propositional or non-representational manoeuvring in the philosophy of perception will not necessarily move us away from epistemological disjunctivism, and may well be consistent with upholding the epistemological insights that the epistemological disjunctivist offers. So, is there a more neutral form of epistemological disjunctivism available? I want to suggest that there is. I will develop a form of epistemological disjunctivism which is consistent with representational commitments but also with the denial of these commitments.

Representational Commitments

173

4. The Thing-Seeing Approach Consider states or episodes of thing-seeing, that is, states or episodes in which we see particular things in our environments. I will suggest that we can develop an epistemology of visual perception which fits the epistemological disjunctivist mould, with thing-seeings at its heart—that is, where we take thing-seeings to be the visual states or episodes which ground paradigm cases of visual perceptual knowledge.5 Before proceeding, let me note two limitations of the discussion to follow. First, I won’t attempt to defend or motivate the framework of epistemological disjunctivism itself. I’ll just assume that it is in good order. Second, I’ll be restricting attention to cases of visual knowledge concerning a thing in one’s environment (e.g., there is a black crow there, the package is brown, etc.). I leave discussion of other cases of visual knowledge (e.g., it’s foggy) for another occasion. The idea at the heart of the thing-seeing approach is that Good Case rational support comes from states or episodes of thing-seeing. Thus, we can replace the aforementioned specification of epistemological disjunctivism (see ‘The Specification’ in section 2) with the following: The Thing-Seeing Specification The particular kind of rational support that the epistemological disjunctivist claims that our beliefs enjoy in paradigm cases of perceptual knowledge is that provided by visual perceptual states or episodes of thing-seeing.

Here, ‘thing’ is a dummy term, which, as Dretske notes, is ‘intended to cover such disparate items as tables, houses, cats, people, games, sunsets, signals, tracks, shadows, movements, flashes, and specks’ (1979: 98). Thus, ‘thing’ covers the heterogenous class of particular things or entities in our environments that we can see. There are three motivations for developing epistemological disjunctivism in this new direction. First, it captures an aspect of common sense. Consider the example where there is a crow before me and I can see it. It seems to be part of common sense to suppose that rational support for beliefs about the crow can come from my seeing it. Suppose I know that the crow on the tree before me is black. How do I know that the crow there is black? Because I can see it. I draw here on Cassam (2007: 347–348), who makes the similar claim that we readily accept explanations of knowledge in terms of seeing objects (e.g., ‘How do you know that the cigarette lighter is under the desk?’ ‘Because I can see it’.). Second, the approach doesn’t face the basis problem facing the orthodox form of epistemological disjunctivism (on which see Pritchard 2011).

174

Craig French

Briefly: the basis problem for orthodox epistemological disjunctivism arises once we add the standard view of seeing that p into the mix. Standardly, seeing that p is thought to be a way of knowing in the sense that to see that p just is to know that p by visual means (Dretske 1969; Williamson 2000; Cassam 2009). If the standard view is accepted, then it is unclear how states of seeing that p can be, as the orthodox epistemological disjunctivist claims, the basis of perceptual knowledge: for being in such states already amounts to having perceptual knowledge. But this problem doesn’t arise on the thing-seeing approach, since the knowledge grounding visual states or episodes appealed to—thingseeings—clearly do not entail knowledge (see here Dretske 1969: ch. 2). Thus, if we re-work epistemological disjunctivism so that the rational support for paradigm instances of perceptual knowledge comes not from states of seeing that p but from states or episodes of thing-seeing, then there is no basis problem (for fuller discussion of this, see French 2016).6 A third motivation for developing epistemological disjunctivism along the lines of the thing-seeing approach is that the approach is, per se, less committal in the philosophy of perception when it comes to the representational commitments outlined earlier. The idea here is not that the approach explicitly conceives of thingseeings as non-propositional or non-representational. Though typical linguistic expressions associated with thing-seeings, e.g., sentences of the form ‘S can see an F x’, ‘S can see x’ (e.g., ‘Jaya can see a black crow’, ‘Barry can see Lola’), aren’t themselves propositional constructions (unlike sentences of the form ‘S sees that p’), it might be that thing-seeings are, in their nature, propositional or representational in some way. It might be that the correct theory of seeing a thing—linguistic expression aside—is one on which seeing a thing constitutively involves representation, even propositional representation, of some sort. We shouldn’t take the thing-seeing approach to deny this. But we shouldn’t take the approach to affirm it, either. And that’s the point. Even though there is, presumably, a fact of the matter about whether thing-seeings are representational, the thing-seeing approach—that is, the construal of epistemological disjunctivism in terms of the thing-seeing specification—is, per se, non-committal on that issue. It is metaphysically neutral when it comes to representational commitments. Thus, if the thing-seeing approach is to work, then we can put states or episodes of thing-seeing to epistemological work, in the framework of epistemological disjunctivism, whilst remaining neutral on whether or not they are propositional, or representational at all. It is this claim—the neutrality claim—that I want to defend in the remainder.

5. Neutrality: The Factivity Challenge Do thing-seeings constitute factive rational support for the beliefs they support? If not, then the thing-seeing approach cannot be a way of

Representational Commitments

175

developing epistemological disjunctivism, since the idea that the rational support perception provides in Good Cases is factive is at the heart of epistemological disjunctivism. Reflecting on this gives rise to the factivity challenge. Unless thingseeings have propositional structure, it is unclear how they can constitute or provide rational support which is factive. For on one understanding of a factive mental state, it is a propositional attitude that one can have only to truths. (Think of paradigms: e.g., knowing that p.) Seeing that p is such a mental state. It is a propositional attitude one can have only to a true content. But if seeing a black crow, for instance, is not a propositional attitude, then it can’t be a factive mental state in the technical sense just outlined. How, then, can thing-seeings constitute or provide factive rational support for beliefs? The key to answering this is to highlight how a state or episode of thing-seeing can constitute factive rational support for p without being factive in the technical sense. A way to see how this is possible is to note that seeing an F thing is similar to seeing that a (the, that) thing is F, in this sense: S cannot see an F thing if there isn’t an F thing there to be seen. Seeing a black crow requires that there is a black crow there to see. Whether or not seeing a black crow is a factive mental state in the technical–propositional sense, if I see a black crow, then there must be a crow which I see, and it must be black. Thus, a formulation of epistemological disjunctivism in terms of thing-seeing can capture the idea that Good Cases involve factive– truth-guaranteeing–rational support. Seeing a black crow guarantees the truth of the belief (or beliefs) it rationally supports, e.g., that the crow is black. And this doesn’t require states or episodes of thing-seeing to be propositional. This draws on the following idea we find in McDowell: [I]f a perceptual state can consist in a subject’s having a feature of her environment perceptually present to her, that gives lie to the assumption that a perceptual state cannot warrant a belief in a way that guarantees its truth. If a perceptual state makes a feature of the environment present to a perceiver’s rationally self-conscious awareness, there is no possibility, compatibly with someone’s being in that state, that things are not as the state would warrant her in believing that they are, in a belief that would simply register the presence of that feature of the environment. (2011: 31) Seeing a crow there is a way for a crow to be perceptually present to one. If it merely looks to one as if there is a crow there, then a crow is not perceptually present to one; but if one genuinely sees a crow there, then a crow is present to one. Thus that perceptual state or episode can serve

176

Craig French

as a truth-guaranteeing warrant for a belief with the content: there is a crow there. What I suggested previously is an expansion of McDowell’s idea in this passage, and we can put it alternatively like this: if one sees a black crow, then, in the McDowellian idiom, there is no possibility, compatibly with someone’s being in that perceptual state, that things are not as the state would warrant her in believing that they are in a belief that attributes the feature in question to the object in question (i.e., a belief that the crow is black). So, states or episodes of thing-seeing can provide factive rational support for beliefs which simply register the presence or existence of the things seen, but also for beliefs which attribute features to the things seen. What is doing the work here is that seeing a thing requires the existence of the thing, that seeing is subject to what Dretske calls an ‘existence condition’ (1969: 43–54), or, to put it another way, that seeing is relational.7 In this respect, seeing is just like standing next to a thing. If I stand next to a black crow, then there must be a black crow there which I stand next to. This is not a special theoretical point about seeing; it is just a basic point about the logic or structure of seeing, as Burge notes, it holds ‘as a matter of the most elementary and superficial understanding’ (2010: 62, fn. 1). Whatever the deep nature of states or episodes of thing-seeing— propositional, representational or otherwise—they have a relational structure and are thus factive in the non-technical sense.8 Note, the claim that thing-seeing is relational is not the same as the controversial claim that the visual experiences we have when we see mindindependent things are relations between perceivers and those things, or, as it is sometimes put, that such visual experiences involve their mindindependent objects as constituents (see Campbell 2002; Martin 2004; Brewer 2011; Logue 2012; Soteriou 2013). To spell out why these claims are different, suppose that one denies the relational theory of experience and holds instead that a visual experience as of a black crow doesn’t involve the crow as a constituent, but instead is to be understood in non-relational, representational terms: in terms of the visual representation of a black crow (or crow-shaped thing). If one holds this, must one deny the claim that seeing a black crow requires that there is a black crow which one sees? No. What one will deny is merely that this relational structure is part of the nature of the experience one has when one sees. That’s not to deny that seeing has a relational structure. It is perfectly consistent with the idea that seeing is relational and that this relational structure is, so to speak, located elsewhere (e.g., in the perceiver’s being causally related to the thing seen). I’ve argued that we can meet the factivity challenge by appealing to the claim that thing-seeing is relational. This is to appeal to a theory-neutral

Representational Commitments

177

feature of thing-seeing. Thus, the epistemological disjunctivist can appeal to states or episodes of thing-seeing as the visual states or episodes which provide factive rational support in Good Cases, and they can do so without committing to (or denying) a propositional or representational view of such states or episodes.9

6. Neutrality: The Rationalizing Role Challenge The second challenge to neutrality that I want to consider has the following form: the visual states or episodes which constitute factive and reflectively accessible rational support in Good Cases must play a rationalizing role with respect to the perceptual beliefs they support. But we can’t capture how thing-seeings play this role unless we admit that they are representational. We can call this the rationalizing role challenge. I am adapting this challenge from work by Hannah Ginsborg (2006, 2011). It’s an adaptation because Ginsborg is not explicitly concerned with the epistemological disjunctivist framework. But in Ginsborg’s discussion we find (a) a conception of the rationalizing role of perception, and then (b) the suggestion that to capture how visual perceptions can play this role, we must admit that they are representational. 6.1 The Rationalizing Role So, first, what is the relevant conception of the rationalizing role of perception? Ginsborg expresses it as follows: As a subject assessing my own beliefs, what I need to determine is what the facts are independently of those beliefs: if the issue is whether I am justified in believing that it has rained, I need to determine whether the streets are wet, not whether I believe that the streets are wet. But if I am assessing someone else’s beliefs, then I need to determine how things present themselves as being from her point of view. As William Alston puts it, ‘when we ask whether S is justified in believing that p . . . we are . . . asking a question from the standpoint of an aim at truth; but we are not asking whether things are in fact as S believes. We are getting at something more “internal” to S ’s “perspective on the world”. . . . We are asking whether the truth of p is strongly indicated by what S has to go on’ (1985: 71). To give someone’s reason for a belief by way of answering the kind of question Alston describes, is to specify a psychological state, typically another belief, in the light of which her original belief can be recognized, from a third- person perspective, as rational. Reasons in this sense might be referred to as ‘third-person’ reasons for belief. (2006: 290)

178

Craig French

Ginsborg is drawing a distinction between two senses of ‘reason’ for belief. Take Riya, who believes that it has rained. On the one hand, she has an excellent reason for this belief: the fact that the streets are wet. If asked why she believes that it has rained, she can reply: ‘because the streets are wet’. She thus gives her reason in one sense of ‘reason’: she offers up the fact that the streets are wet as the reason for her belief. On the other hand—and this is what is immediately relevant to us— there is, Ginsborg suggests, a second sense of ‘reason’ on which reasons are psychological states of subjects which rationalize their beliefs. Riya believes that it has rained. She is rational in holding this belief because her perspective includes the other belief that the streets are wet: the belief that the streets are wet rationalizes (together with other background beliefs, no doubt) her further belief that it has rained. To put it in Alston’s terms, the belief that the streets are wet is an aspect of her perspective on the world which ‘strongly indicates’ that it has rained. It thus plays a rationalizing role with respect to the belief that it has rained. 6.2 Ginsborg’s Argument: Outline Having highlighted this rationalizing role, Ginsborg argues that it is hard to see how thing-seeings can be rationalizing if they are not representational. Thus she says that we need to make sense of how “our beliefs and judgements can be rationally intelligible in the light of our perceptions” (2011: 135). But she adds that this seems to require that we ascribe some kind of representational content to perceptual experience. . . . For on a certain natural line of thought, perception of an object cannot rationalize a belief, that is, make it rationally intelligible, unless it presents the object as being a certain way, that is, as having a certain general property or feature. (Ibid.: 135) There are two related waves to Ginsborg’s argument which I’ll discuss in turn. First, Ginsborg argues that for a subject’s perception to rationalize their belief that p, their perception must provide them with a ‘perspective on the world’ which ‘indicates’ that p is true, but we can’t capture this unless we admit that perception is representational. Second, Ginsborg argues that for a subject’s perception to rationalize their belief that p, their perception must make available to them a reason to believe that p, but we can’t capture this unless we admit that perception is representational.

Representational Commitments

179

6.3 Ginsborg’s Argument: First Wave Ginsborg considers ‘how seeing a package in front of you might provide you with rational grounds for believing that there is a package in front of you’ (2011: 137). One idea about this, which Ginsborg (Ibid.: 147) draws from Brewer (2004), is that one sees the package, and in seeing it, one has experiential access to some of the features which are relevant to the judgement that it is a package (e.g., its color and shape, its being a package, etc.), and that’s why seeing the package rationalizes the belief that there’s a package there. But Ginsborg suggests that it is not entirely clear how to satisfactorily unpack the idea of the relevant features being ‘experientially accessible’, and a natural way to do so invites appeal to representation. Consider Asha, who sees the package before her. Suppose we unpack the idea that she has ‘experiential access’ to the relevant features, as follows: she sees the package as brown, as rectangular, and as a package. Now, if Asha’s perception is understood in this way, we can understand how it rationalizes her belief that there is a package there: her perception is such as it gives her a perspective on her environment which ‘indicates’ to her how things are. But in construing her visual perception in these terms, we are, Ginsborg suggests, taking it to be representational (2011: 138). Is there an alternative construal of the experiential accessibility of the relevant features which doesn’t invite appeal to representation? Ginsborg considers some of Brewer’s discussion of the Müller-Lyer Illusion. In this illusion, two lines in fact equal in length look unequal in length to the subject. Brewer claims that, despite the illusion, the equality in length of the lines is experientially accessible to the subject (2004: 70). What does this mean? Ginsborg spells it out as follows: what this seems to amount to in the Müller-Lyer case is just that we are capable, under appropriate circumstances, of coming to represent the lines as having the property of identity in length: for example, we can come to represent them as identical in length if the misleading arrow-heads and -tails are removed. (Ibid.: 138) So the sense of ‘experiential accessibility’ in play is that the features are experientially accessible if they are instantiated by what is perceived, and we are capable of representing them or coming to see that they are instantiated in what is perceived, in appropriate circumstances. So if we apply this to the case of the package, the idea is the following: ‘the color, shape, and functional kind of the package are experientially accessible in this weaker sense if the package has them, and if we are capable of coming to see that it has them’ (Ibid.: 38).

180

Craig French

But Ginsborg notes that this doesn’t help since saying that [the relevant features] are experientially accessible in this sense does not seem to add anything to the claim that the package presented to us is in fact brown, rectangular, and a package, since these are features of a kind which we can, typically, come to see things as having. So it does not help with the question of how having a particular brown, rectangular package perceptually presented to us can rationalize the belief that it has those properties of being brown, being rectangular, and being a package. (Ibid.: 138) The weak notion of experiential accessibility doesn’t help us to see how a subject who sees the package is rational in judging that it’s a package, since though the relevant features are experientially accessible, they are not accessible in such a manner that the perception ‘indicates’ that the object has those features. It is, Ginsborg thinks, not sufficient for one’s perception of an F thing to rationalize the belief that the thing is F that one simply sees the thing and it is in fact F. There has to be, Ginsborg supposes, something about the perception which makes F rationally applicable to the thing in light of the perception. Saying that F is experientially accessible in the weak sense under consideration doesn’t capture this any more than the fact that one sees the thing and it in fact is F. Consistently with the presence in experience of a package which in fact has the relevant features (and which are in the weak sense accessible), Asha’s experience might nonetheless take such a form that it doesn’t ‘indicate’ to her that there’s a package there, or it might take such a form that it ‘indicates’ something quite different (e.g., there’s a patch of light, or that’s a large brick). Similarly, in the Müller-Lyer case, even though identity in length is a feature of the lines, and experientially accessible, the perception of those lines does not itself rationalize the belief that those lines are identical in length. So instead, drawing on Brewer (2008), Ginsborg considers the idea that ‘in the straightforward case where you take the package to be a package, the visual similarity of the package to paradigm packages makes [rationally] intelligible your taking it to be a package’ (2011: 138). But Ginsborg worries about this, too: Here again, though, it is not clear how the similarity is supposed to figure in your perceptual experience. It cannot be that your perceptual experience proper represents the package as similar to paradigm cases of packages . . . [because at this stage Ginsborg is considering how someone can capture the rationalizing role of perception without

Representational Commitments

181

appeal to representational content.] But the only alternative would seem to be that the package is in fact similar to paradigm packages, where that similarity might or might not become salient to a given perceiver in a given context. And while that might make it possible to understand why, as a matter of psychological fact, your perception of the package leads you to believe that the package is a package, it does not help with the question of how your perception can make your belief ‘rationally intelligible’ in the sense at issue. (2011: 139) The fact that the package is visually similar to a paradigm package might be a reason for Asha to judge that it’s a package. But how does her perception ‘indicate’ to her that this is how things are? This is intelligible if we are allowed to invoke seeing the object as visually similar to a paradigm package, where this is construed representationally. But not if all we appeal to is Asha’s simply seeing a package which in fact is visually similar to a paradigm package. Consistently with the presence in perception of a package which is in fact similar to a paradigm package, Asha’s perception might nonetheless take such a form that it doesn’t ‘indicate’ to her that there’s a package there, or it might take such a form that it ‘indicates’ something quite different (e.g., there’s a patch of light, or that’s a large brick). Considerations such as these lead Ginsborg to suggest that ‘we cannot make sense of a perception as rationalizing a belief unless we take the perception to represent its object as having some general feature or other’ (Ibid.: 140). I think there is something to Ginsborg’s argument here. That is, if all we know about one’s perception is that they see an F thing, we are not thereby entitled to suppose that they are rational in judging the thing to be F on the basis of their perception. Since if all we know is that they see an F thing, then we don’t know much at all about the perspective on the world they have in so seeing. We just don’t know anything about what their perception ‘indicates’ to them about their environment. For all we know, their perception may be illusory, lacking in detail or focus, confused, and so on. But it is just not obvious that the only way to correct for this is to conceive of the rationalizing states or episodes of seeing as representational. Instead, we can appeal to their conscious characters. That is, if we know that Asha has judged that there is a brown rectangular package there on the basis of seeing a brown rectangular package, before we ascribe rationality to her, we can ask: what kind of experience does she have in seeing the brown rectangular package, that is, what is it like for her to so perceive on this occasion? (The answer to that question certainly cannot be read off from the fact that she sees a brown rectangular package).

182

Craig French

A familiar way to express or describe the kind of experience a perceiver has or the conscious character of their experience, is in terms of how things appear to the perceiver in perceiving as they do. Suppose, then, that it turns out that in seeing the package Asha has an experience the character of which can be captured by saying that the package appears brown, rectangular, and package-like to Asha. Given this, we can see how it is rational for her to judge that there is a brown rectangular package before her on the basis of her visual perception. It is rational in light of her seeing the package. If all we could admit about the kind of experience she has is that it is a seeing of a brown rectangular package, we wouldn’t be admitting enough to make her subsequent judgement rationally intelligible. But by appeal to the conscious character of her experience, we can admit more, and thus regard her subsequent judgement as rational in the light of her visual perception. One might have the following worry about the suggestion we’ve just considered. One might think that we are just agreeing with Ginsborg that states of seeing must be representational if they are to be rationalizing, though we are using different terms to express it. But then the suggestion would have to be that in talking of the conscious character of the subject’s experience, of what it is like for the subject to experience as she does, we are just referring to how the subject perceptually represents the world, in different terms. However, this would need strong argument, since it is tantamount to supposing that there is a certain incoherence in the idea of a theory of experience which explicitly rejects appeal to representation in its account of conscious character. One kind of constitutive account has it that the kind of experience a subject has when they see something is a matter of the representational nature of the experience. For instance, one view is that when Asha sees a brown rectangular package, and has an experience with a certain character, e.g., which can be captured by saying that the package looks rectangular, brown, and package-like to her, this is a matter of her perceptually representing the package as being brown, rectangular, and package-like. But another account has it that the character of her experience is explicable in terms of a primitive non-representational perceptual relation to the package itself. On this view the package itself (and some of its features), and not a perceptual representation of the package, is constitutive of the character of her experience.10 Such a view may, of course, turn out to be false. But there is no evident incoherence in it. What we want to make sense of is how Asha’s visual perception in which she sees a brown rectangular package can rationalize her subsequent judgement that there is a brown rectangular package there. I am suggesting that such a judgement is rational in the light of her visual perception given that her perception is or involves an experience of a certain kind, with a certain character. What the nature of the experience is, given

Representational Commitments

183

that it is of that kind, or given that it has that character, is a further question that doesn’t need to be settled in order to capture what we want to capture. Asha’s visual perception rationalizes her judgement because it is (or involves) an experience of certain a kind, whatever the correct constitutive account of that turns out to be. To return to Alston’s terminology, we can acknowledge that for Asha’s visual perception to rationalize her belief that there is a brown rectangular package there, it has to give her a ‘perspective on the world’ which ‘strongly indicates’ the truth of what she judges. (For future reference, let’s call this the perspective claim.) But that’s what having a visual perception of an object with a certain conscious character, which we can describe in terms of things appearing a certain way to one, does for Asha. Having a certain conscious character is part of what it is for her visual perception to give her a perspective on the world which strongly indicates that there is a package there. So here again, as with factivity, it looks as though the explanatory work is being done by a theory-neutral feature of states or episodes of thing-seeing: that such states or episodes have conscious characters. We can intelligibly talk about the character of an experience without appealing to, smuggling in, or suggesting any constitutive account of what that amounts to. To capture what Ginsborg wants us to capture, I have suggested, does require us to consider the character of the experience the subject has when they perceive. But to capture what Ginsborg wants to capture doesn’t require appeal to any particular constitutive account of that. Ginsborg goes straight for a constitutive account in appealing to perceptual representation, and thus neglects the idea that we can capture what she wants to capture in a more neutral way. 6.4 Ginsborg’s Argument: Second Wave The second wave of Ginsborg’s argument brings in an account of why psychological states are rationalizing. Consider, then, the following: On the approach we are considering, the perception of the package has to play the same kind of role in rationalizing the belief that there is a package present that the belief that the streets are wet plays in rationalizing the belief that it rained. In our paradigm case it is plausible to suppose that the belief plays that rationalizing role because it, so to speak, makes available to you a reason for your belief, namely the consideration that the streets are wet. When you form your belief that it rained, you have ‘in view’ the fact or proposition that the streets are wet, and you are in a position to cite that fact or proposition as a reason for your belief that it rained. So if your perception of the package is to play the same kind of role

184

Craig French with respect to your belief that there is a package in front of you, then it is plausible to suppose that it must also bring into view a consideration which supports the belief that there is a package in front of you. (Ginsborg 2011: 144–145)

There is a lot to unpack here. First is the assumption that we are to understand the rationalizing role of perception as Ginsborg understands the rationalizing role of belief. This is questionable. One might disagree with Ginsborg about the rationalizing role of belief. Or one might disagree with Ginsborg that the rationalizing role of perception should be modelled closely on that of belief. And one might question both things. But here I’ll accept the way Ginsborg understands the rationalizing role of belief, and I’ll grant that the rationalizing role of perception should be understood similarly. Ginsborg’s argument has to do with the idea that for perceptions to be rationalizing, they must bring reasons into view. The idea seems to be that we can’t capture how perceptions bring reasons into view (and hence how perceptions are rationalizing), unless we admit that perceptions are representational. This is the second wave of Ginsborg’s argument. Now, one line of thought is that seeing the package brings the package into view, and the package itself can be a reason for beliefs about it. Perhaps seeing the package in front of one can rationalize the belief that there’s a package there, similarly to how beliefs rationalize other beliefs, that is, by bringing an appropriate reason into view which supports the belief in question. Nothing in this line of thought requires us to endorse a representational model of thing-seeing, though the line of thought is consistent with such a model. But in response to this, Ginsborg says that the package itself cannot count in favor, either of the belief that it is a package, or of the belief that there is a package in front of you. Not being a fact or proposition, it is simply not the right kind of thing to serve as a reason for a belief. (Ginsborg 2011: 145) This suggests that Ginsborg is operating with a factualist or propositionalist assumption about reasons such that the reasons which rationalizing psychological states or episodes bring into view must be either facts or propositions, considerations of the form that suchand-such is the case. As we’ll see shortly, another aspect of Ginsborg’s thinking is more relaxed on this matter, but for now, let’s proceed as if factualism about reasons is true (I drop discussion of propositionalism from now on).11

Representational Commitments

185

We are assuming that states or episodes of thing-seeing play a rationalizing role. I now want to suggest that even if we grant factualism, and so hold that states or episodes of thing-seeing bring facts into view, that doesn’t force upon us a representational model of states or episodes of thing-seeing. Again, we can remain neutral on whether such a model holds. And the flaw in Ginsborg’s argument, I want to suggest, is similar to what we encountered earlier: neglect of the fact that we can capture what she wants us to capture—something about the rationalizing role of visual perception—in terms of a theory-neutral feature of visual perception–conscious character. Now one might wonder how merely seeing a package can make the fact that there’s a package there available to a subject if it is not propositionally structured. However, given what we said earlier, the question is not whether merely seeing a package can make the relevant fact available to a subject, but rather whether seeing a package where that involves having a certain kind of experience, an experience with a certain character, can make the relevant fact available to one. Suppose that Asha sees the package and in seeing it has an experience the character of which can be captured by saying that it appears to her as if there’s a package there. Part of what Asha’s seeing the package, and what having an experience with such a conscious character, does for Asha is make relevant facts available to her—such as the fact that there’s a package there. What is explanatorily relevant to Asha’s having in view the fact that there’s a package there, is that she sees the package and has an experience with a certain character, which can be captured by saying that it appears to her as if there’s a package there. That’s not the same as the idea that she sees the package and perceptually represents that there is a package there—not unless we have already smuggled in a representational account of seeing, experience, and conscious character. Consistently with this, it might be that Asha’s having in view the fact that the package is there is in part a function of what she can make of her visual perception. That is, it might be that it is in part a function of her being able to tell that she is seeing a package. And this, we might think, is partly enabled by the fact that she is seeing a package, but partly enabled by the fact that she has an experience with the aforementioned character. For in virtue of the experience being of this kind (and not, say, of such a kind as would be captured by saying that it appears to her as if there is a patch of light there), it can reveal itself to her to be a seeing of a package. Now, it might be that this requires Asha to have certain capacities, the exercise of which involves representation—presumably, it requires Asha to be a rational, self-conscious subject, with appropriate discriminative and recognitional capacities. But none of this demands a representational model of states or episodes of thing-seeing, or of experience and conscious character.

186

Craig French

So even if we assume that factualism is true, and that states or episodes of thing-seeing rationalize beliefs by making facts available to their subjects, this doesn’t force upon us a representational model of thing-seeing (or of experience and conscious character). We can remain neutral on such a model. What is explanatorily relevant is, again, theory-neutral: that rationalizing states or episodes of thing-seeing are or involve experiences with conscious characters. Asha’s visual perception of the package makes the fact that there’s a package there available to her because it is (or involves) an experience with a certain character. A similar point applies when it comes to what Ginsborg has to say about views which don’t assume factualism. Consider, for instance, her remarks on the view she finds in Johnston (2006). As Ginsborg notes, on Johnston’s view: We can perceive such things as the snubnosedness of Socrates, or the astringency of the calvados, where these are conceived of as states or conditions of Socrates or the calvados, and as being on a par with events such as a particular chiding of Socrates by Xanthippe. (2011: 146) Ginsborg notes that for Johnston, such things are not facts or propositions (Ibid.: 146). And she notes that the states and events which we perceive might be thought to serve as reasons for belief. Perhaps the ‘brownness of the package’ on Johnston’s construal, that is the state or condition of the package constituted by its being brown, can count in favor of believing the package to be a package. . . . And in that case your perception of the package’s brownness—of its state of being brown—can rationalize the belief that it is a package. Alternatively, your perception can rationalize, as a limiting case, your believing the package to be brown, where that belief in turn can rationalize your believing it to be a package.12 (Ibid.: 146–147) One might develop this so as to reject Ginsborg’s suggestion that in order to capture how rationalizing perceptions bring reasons into view, we must admit that such perceptions are representational. For one might suppose that Asha’s seeing the brownness of the package brings a reason to believe that the package is brown into view, namely, the package’s brownness. And here we are appealing to the idea that she sees the package’s brownness, not any particular account of what such seeing amounts to (representational or otherwise).13 What, then, is wrong with such an account, according to Ginsborg?

Representational Commitments

187

First, to switch examples, Ginsborg suggests that [W]hatever plausibility there is to the idea that perceiving Socrates’s snubnosedness can rationalize the belief that he is snubnosed, seems to rely on the assumption that, in perceiving Socrates’s snubnosedness, one perceives that he is snubnosed, and hence is presented with a fact or proposition counting in favor (here, as a limiting case) of the belief that he is snubnosed. (Ibid.: 147) But why does Ginsborg think this? She explains with reference to another example: The point can be brought out most readily in connection with events. It is possible to perceive the event of a chiding of Socrates by Xanthippe without realizing that Socrates is being chided by Xanthippe: one might at the time be capable of describing what one is hearing only as ‘a muffled voice coming from the next room,’ and find out only later, if at all, that one had heard Xanthippe chiding Socrates. If that is the way in which one hears Xanthippe chiding Socrates, then, even if one’s perception causes one to form the belief that Xanthippe is chiding Socrates, the belief is not rationalized by the perception. It is the fact or proposition that Xanthippe is chiding Socrates which can serve as a reason for, or count in favor of, this or that belief, not the event of Xanthippe’s chiding Socrates. (2011: 147) The point seems to be that one can perceive the event of a chiding of Socrates by Xanthippe and still not be rational in judging on that basis that Xanthippe is chiding Socrates. For one might hear the event in such a way that it would not be rational to judge, on the basis of one’s auditory perception, that Xanthippe is chiding Socrates (for instance, one has nothing but an impression of muffled voices in the next room). Applying this back to the previous claims, presumably the argument is as follows: one can perceive Socrates’s snubnosedness and still not be rational in judging on that basis that Socrates is snubnosed. For one might perceive the snubnosedness in such a way that it would not be rational to judge, on the basis of one’s perception, that Socrates is snubnosed (for instance, if one’s perception is confused or illusory). Finally, Ginsborg considers the case of the package: [I]f Asha sees the brownness of the package, the package’s brownness is there to be seen. But if the package’s brownness is there to be seen, then the package must be brown. So, Asha’s seeing the package’s

188

Craig French brownness, just like her seeing the brown package, guarantees the truth of her perceptual belief that the package is brown. If we follow Johnston in assuming that states and conditions are to be understood on the same model as events, the same must hold true for them. Seeing the package’s brownness, or the package’s being brown, can rationalize the belief that the package is brown, or a package, only if it involves seeing the package to be brown. (2011: 147)

To apply the argument: Asha can perceive the package’s brownness and still not be rational in judging on that basis that the package is brown. For she might perceive the brownness in such a way that it would not be rational to judge, on the basis of her perception, that the package is brown (for instance, if her perception is confused or illusory). What Ginsborg concludes from these considerations is that insofar as one’s perception of the snubnosedness of Socrates rationalizes one’s belief that Socrates is snubnosed, it is only because one perceives that Socrates is snubnosed, and only because this perception makes available to one the fact that Socrates is snubnosed. And insofar as one’s perception of the chiding of Socrates by Xanthippe rationalizes one’s belief that Xanthippe is chiding Socrates, it is only because one perceives that Xanthippe is chiding Socrates, and only because this perception makes available to one the fact that Xanthippe is chiding Socrates. And insofar as Asha’s perception of the brownness of the package rationalizes her belief that the package is brown, it is only because she perceives that the package is brown, and only because her perception makes available to her the fact that the package is brown. Now, Ginsborg is right that when Asha sees the brownness of the package, then whether or not she is rational in judging that the package is brown, on the basis of her perception, depends upon the way in which she sees brownness of the package. After all, for her perception to be rationalizing, it has to give her a perspective on the world which ‘indicates’ the truth of what she judges (the perspective claim). And this will be fixed, in part, by the way in which she sees the brownness of the package. The problem is that Ginsborg assumes that once we acknowledge this point, we must appeal to propositional representation. Yet, as I have argued, it’s enough to acknowledge that Asha’s visual perception has an appropriate conscious character (for instance, that in seeing the brownness of the package she has an experience with a character that we can capture by saying that the package appears brown to her). Again, Ginsborg goes straight for something theoretical to capture what can be captured with a theory-neutral notion. Similar points apply to Ginsborg’s other examples. Thus, despite what Ginsborg says, one might hold that Asha’s seeing the brownness of the package rationalizes her belief that the package is brown, and does so by bringing a reason into view: the package’s brownness. We can add that seeing the package’s brownness plays this

Representational Commitments

189

rationalizing role only when it has an appropriate conscious character. But none of this requires us to take a stance on whether Asha’s seeing of the brownness of the package (and the experience and conscious character this involves) is representational. Thus, the second wave of Ginsborg’s argument, like the first, doesn’t show that we cannot capture the rationalizing role of states or episodes of thing-seeing unless we admit that such states are representational. 6.5 Summary In the first wave, Ginsborg argues that for a subject’s visual perception to rationalize their belief that p, their perception must provide them with a ‘perspective on the world’ which ‘indicates’ that p is true; but to capture this, we have to admit that visual perceptions are representational. In response, I argued that we can capture this instead with the theoryneutral idea that our visual perceptions have conscious characters. This is what explains how our perceptions afford us a perspective on the world which ‘indicates’ the truth of the beliefs they support. In the second wave, Ginsborg argues that for a subject’s visual perception to rationalize their belief that p, their perception must make available to them a reason to believe that p; but to capture this, we have to admit that visual perceptions are representational. In response, I argued that we can capture this instead with the theory-neutral idea that our visual perceptions have conscious characters. This is what explains how our perceptions make reasons available to us—whether we conceive of such reasons as objects, facts, states, conditions, or events.

7. Conclusion I’ve explored a non-orthodox form of epistemological disjunctivism—the thing-seeing approach. On this approach, states of episodes of thing-seeing— such as seeing a black crow and seeing a brown package—are the states or episodes of visual perception which ground paradigmatic visual knowledge. A central motivation for developing epistemological disjunctivism in this new direction is that it is a form of the view which, unlike the orthodox form, is neutral on whether the states or episodes which ground visual knowledge are propositional or representational. It is thus less vulnerable in the philosophy of perception than orthodox epistemological disjunctivism. It is not hostage to the fortunes of propositional or representational views in the philosophy of perception. I have considered whether we really can put states or episodes of thing-seeing to epistemological work, in the framework of epistemological disjunctivism, whilst remaining neutral on whether or not they are propositional, or representational. According to the factivity challenge, unless we admit that these states or episodes are representational, we

190

Craig French

cannot capture how they provide factive rational support. In response, I argued that the fact that thing-seeings are relational explains how they can constitute factive rational support, and that this doesn’t require a commitment to, or denial of, any propositional or representational model of thing-seeing or its relationality. According to the rationalizing role challenge, unless we admit that knowledge grounding visual states or episodes are representational, we cannot capture how they rationalize beliefs. In response, I argued that we can capture the rationalizing role of states or episodes of thing-seeing by appeal to the fact that they have conscious characters, and that this doesn’t require a commitment to or denial of any propositional or representational model of thing-seeing or of conscious character. I have argued, then, that with respect to both challenges, what is relevant to explaining how states or episodes of thing-seeing can be put to epistemological work, in the framework of epistemological disjunctivism, are theory-neutral features.

Notes 1. Some of this material was presented at the Perception and Evidence Workshop, organized by Frank Hofmann, at the University of Luxembourg in June 2016. I am grateful to the workshop participants on that occasion for helpful discussion, especially to Frank Hofmann, Clayton Littlejohn, and Chris Tucker. For other helpful discussions relating to this material, thanks to Anil Gomes, Mark Kalderon, Mike Martin, Aidan McGlynn, Ian Phillips, Duncan Pritchard, Christopher Ranalli, and Paul Snowdon. 2. See also McDowell’s claim that ‘When one knows something to be so by virtue of seeing it to be so, one’s warrant for believing it to be so is that one sees it to be so . . .’ (2011: 33). 3. If one denies that there are any propositionally structured states of perception, that doesn’t commit one to denying that there are states of the form: S sees that p. For one might deny that such states are states of visual perception. An alternative is to hold that for S to see that p (at least in one sense ‘see’) is for S to know that p, on the basis of vision. On which see French (2012a, 2013, 2016). 4. Though these are relatively recent examples, note that non-representational theorizing about perception is not new, and that it can take many different forms. There are, for instance, versions of sense-datum views, adverbial views, and sensationalist/raw feel views, which reject the idea that perception has representational content. 5. Here I develop some of the ideas in French (2016). 6. For further discussion of the basis problem see also Millar (2016: 56, fn. 27). 7. For further discussion of this point about seeing, and potential counterexamples, see French (2012b: §1.3.1) 8. The fuller passage from Burge is as follows: ‘The standard specification of such states [seeing, knowing, remembering, etc.] entails—as a matter of the most elementary and superficial understanding—truth, veridicality, or some relation, such as perceptual reference, to the environment. Knowing something entails that it is true. Seeing something entails perceptually referring to it and being causally related to it’ (2010: 62, fn. 1). Note here how swiftly Burge

Representational Commitments

9.

10.

11. 12. 13.

191

moves from the theory-neutral point that seeing something is relational to more theoretically loaded ideas such as the idea that seeing something requires a causal relation, and reference. It is not these ideas, but the theory-neutral point that I am suggesting the epistemological disjunctivist avail themselves of. And though Burge appeals to the idea of successful perceptual reference to what is seen, this isn’t built into the very idea of seeing or its relational structure (it is not a matter of the most elementary and superficial understanding that seeing involves reference). For one might attempt to explain the relational structure of seeing by invoking not representational notions such as reference, but rather non-representational notions such as the non-representational relation of acquaintance appealed to by authors such as Brewer (2011). This may turn out not to be the correct approach, but it is certainly not at odds with the very idea of seeing and its relational structure. An epistemological disjunctivist might deny that knowledge grounding visual states or episodes have to play the rationalizing role that Ginsborg highlights. To satisfactorily deny this, they would have to spell out their notion of reflectively accessible rational support such that it doesn’t entail that visual perceptions must have this rationalizing role. Though this is an option, here I’ll just accept that the disjunctivist’s knowledge grounding visual perceptions must have this rationalizing role. I don’t mean to suggest that such relational views have to eschew a role for perceptual representation in accounting for conscious character, but they may do so, and some proponents of such views do seem to develop their views in this way (see, for instance, Campbell 2002; Brewer 2011). Note also that such views don’t have to say that the character of Asha’s experience is entirely constituted by the package (and its features). For views of this sort can allow that there are other aspects of the relational nature of experience, the specification of which is relevant to what shapes conscious character, such as facts about the subject of experience (Logue 2012), the manner in which one is acquainted with the objects of experience (Soteriou 2013), and the ‘third relatum’ or standpoint from which one perceives (Campbell 2009; Brewer 2011). For further discussion, see French (2018). For helpful discussion of the ontology of reasons, see Sylvan (2016). See also Kalderon (2011). This also fits the mould of the thing-seeing approach, on the assumption that Asha’s seeing the brownness of the package provides factive rational support for some of her perceptual beliefs.

References Alston, William, 1985, Concepts of epistemic justification, The Monist 68(2): 57–89. Brewer, Bill, 2004, Realism and the nature of perceptual experience, Philosophical Issues 14(1): 61–77. ———, 2008, How to account for illusion, in Adrian Haddock and Fiona Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 168–180. ———, 2011, Perception and Its Objects. Oxford: Oxford University Press. Brogaard, Berit, ed. 2014, Does Perception Have Content? New York: Oxford University Press. Burge, Tyler, 2010, The Origins of Objectivity. Oxford: Oxford University Press. Campbell, John, 2002, Reference and Consciousness. Oxford: Oxford University Press.

192

Craig French

———, 2009, Consciousness and reference, in Brian McLaughlin, Ansgar Beckermann and Sven Walter, eds. The Oxford Handbook of Philosophy of Mind. Oxford: Oxford University Press, pp. 648–662. Cassam, Quassim, 2007, Ways of knowing, Proceedings of the Aristotelian Society 107(3): 339–358. ———, 2009, Knowing and seeing: Responding to Stroud’s dilemma, European Journal of Philosophy 17(4): 571–589. Crane, Tim, 2011, Is perception a propositional attitude? in Katherine Hawley and Fiona Macpherson, eds. The Admissible Contents of Experience. Oxford: Blackwell, pp. 83–100. Dretske, Fred, 1969, Seeing and Knowing. London: Routledge. ———, 1979, Simple seeing, in his Perception, Knowledge and Belief: Selected Essays. Cambridge: Cambridge University Press, pp. 97–112. French, Craig, 2012a, Does propositional seeing entail propositional knowledge? Theoria 78(2): 115–127. ———, 2012b, Visual Perception as a Means of Knowing, PhD thesis. London: University College London. ———, 2013, Perceptual experience and seeing that p, Synthese 190(10): 1735–1751. ———, 2016, The formulation of epistemological disjunctivism, Philosophy and Phenomenological Research 92(1): 86–104. ———, 2018, Näive realism and diaphaneity, Proceedings of the Aristotelian Society 118(2): 149–175. Ginsborg, Hannah, 2006, Reasons for belief, Philosophy and Phenomenological Research 72(2): 286–318. ———, 2011, Perception, generality, and reasons, in Andrew Reisner and Asbjørn Steglich-Petersen, eds. Reasons for Belief. Cambridge: Cambridge University Press. Haddock, Adrian and Macpherson, Fiona, eds., 2008a, Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press. ———, 2008b, Introduction: Varieties of disjunctivism, in their Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 1–24. Hutto, Daniel D. and Myin, Erik, 2012, Radicalizing Enactivism: Basic Minds Without Content. Cambridge, MA: MIT Press. Johnston, Mark, 2006, Better than mere knowledge? The function of sensory awareness, in Tamar Gendler and John Hawthorne, eds. Perceptual Experience. Oxford: Oxford University Press, pp. 260–290. Kalderon, Mark Eli, 2011, Before the law, Philosophical Issues 21(1): 219–244. Logue, Heather, 2012, Why naive realism? Proceedings of the Aristotelian Society 112(2pt2): 211–237. Martin, Michael, 2004, The limits of self-awareness, Philosophical Studies 120: 37–89. McDowell, John, 1982, Criteria, defeasibility, and knowledge, Proceedings of the British Academy 68: 455–479. ———, 1995, Knowledge and the internal, Philosophy and Phenomenological Research 55: 877–893. Reprinted in McDowell (1998); page references in the main body to that version. ———, 1998, Meaning, Knowledge, and Reality. Cambridge, MA: Harvard University Press.

Representational Commitments

193

———, 2008, The disjunctive conception of experience as material for a transcendental argument, in Adrian Haddock and Fiona Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 376–389. ———, 2011, Perception as a Capacity for Knowledge. Milwaukee, WI: Marquette University Press. Millar, Alan, 2016, Perceptual knowledge and well-founded belief, Episteme 13(1): 43–59. Pritchard, Duncan, 2011, Epistemological disjunctivism and the basis problem, Philosophical Issues 21(1): 434–455. ———, 2012, Epistemological Disjunctivism. Oxford: Oxford University Press. ———, 2016, Epistemic Angst: Radical Skepticism and the Groundlessness of Our Believing. Princeton, NJ: Princeton University Press. Soteriou, Matthew, 2013, The Mind’s Construction: The Ontology of Mind and Mental Action. Oxford: Oxford University Press. Sylvan, Kurt, 2016, Epistemic reasons I: Normativity, Philosophy Compass 11(7): 364–376. Travis, Charles, 2004, The silence of the senses, Mind 113(449): 57–94. Weir, Alan, 2004, An ultrarealist theory of perception, International Journal of Philosophical Studies 12(2): 105–128. Williamson, Timothy, 2000, Knowledge and Its Limits. Oxford: Oxford University Press.

10 Either Epistemological or Metaphysical Disjunctivism Veli Mitova

Suppose I see a fever tree outside my window. According to the epistemological disjunctivist, this experience makes my beliefs about the tree rational in a way that they wouldn’t be had I merely hallucinated the tree. Seeing is believing, as the old adage goes. Epistemological disjunctivism is a view about the normative significance of our perceptual experiences. Although itself neutral on the nature of these experiences, it is nonetheless often taken to be more congenial to certain ontologies than to others. In particular, it is commonly thought to smooth the way for so-called metaphysical disjunctivism, the view that veridical perception is one kind of ontological entity while merely apparent perception is a radically different kind. Thus, when I see the fever tree, I am in one kind of mental state; when I am hallucinating it, I am in another.1 John McDowell, the mascot of both epistemological and metaphysical disjunctivism, is often taken to have argued from the former to the latter.2 The main aim of this chapter is to show that it isn’t a good idea to hold both forms of disjunctivism together, and a fortiori to move from the one to the other. I first set the scene by introducing the two forms of disjunctivism (section 1) and a distinction that my argument will need—the one between normative and motivating reasons (section 2). The argument, then, is this: Premise 1: The conjunction of epistemological and metaphysical disjunctivism entails disjunctivism about motivating reasons for belief. Premise 2: Disjunctivism about motivating reasons for belief is implausible. Conclusion: So, the conjunction of epistemological and metaphysical disjunctivism is implausible. In sections 3 and 4, I argue for these premises. In section 5, I defend each against the most pressing objection. The upshot is that friends of both forms of disjunctivism are in trouble: if they think of the two views as logically independent, they shouldn’t hold both together; if they think of them as dependent, they shouldn’t hold either.

Either Epistemological or Metaphysical Disjunctivism

195

This conclusion should be at the very least surprising, given the popularity of holding both views together. But I also hope that it will be constructive. Disjunctivists of various stripes take themselves to be accountable to considerations from philosophy of mind and epistemology. This chapter suggests that they are also accountable to broader debates on the nature of reasons. If I show nothing else in this chapter, I hope to at least show that.

1. Disjunctivisms Let me adopt the increasingly common practice of reserving the labels ‘perceptual experience’ and ‘perception’ for veridical experiences alone, and dubbing non-veridical experiences, such as hallucination and illusion, ‘apparent perceptual experiences’, or ‘appearances’ for short.3 In most general terms, epistemological disjunctivism is a view about the differing normative significance of perceptual experiences and appearances, while metaphysical disjunctivism is a view about their differing ontologies. In this section, I contrast the two views, and introduce the move some philosophers like to make from the former to the latter. 1.1 Epistemological Disjunctivism The basic idea of epistemological disjunctivism is that a perceptual experience endows beliefs based on it with a different kind of epistemic support from the one we get from mere appearances. McDowell puts it like this: [P]erceptual appearances are either objective states of affairs making themselves manifest to subjects, or situations in which it is as if an objective state of affairs is making itself manifest to a subject, although that is not how things are. Experiences of the first kind have an epistemic significance that experiences of the second kind do not have. They afford opportunities for knowledge of objective states of affairs. (2008: 380–381, my italics; see also McDowell 1998a: 390–391) And here is Duncan Pritchard, a more recent champion of the view:4 [H]owever we understand the rational support [that the experience provides] in the second case, it is not merely a ‘stunted’ version of the rational support in the first case, as if the latter were just the former supplemented in some way with additional support. Rather, the two rational standings are radically different in kind (this is what makes this epistemological proposal disjunctivist). (2012: 16, his italics)

196 Veli Mitova Perceptual experiences and mere appearances, in other words, put us in different positions in relation to knowledge, rationality, justification, and other such epistemically good things. Craig French (2016: 87) calls this the ‘negative thesis’ of epistemological disjunctivism, and helpfully distinguishes it from a positive one which specifies the kind of support perceptual experiences provide. According to Pritchard (2012: 13), this support is reflectively accessible and factive.5 Other authors add other positive bits. According to McDowell (1998a: 390), for instance, the support is also non-inferential and indefeasible. My concern here is not with the positive theses which various disjunctivists espouse. (Though, for obvious reasons, the factivity of perception must be assumed throughout.) I will just focus on the negative one. First, I take it to be common ground for all epistemological disjunctivists. And, second, it is the thesis that I will argue makes trouble for the combination of epistemological and metaphysical disjunctivism. Put it this way: however else one specifies epistemological disjunctivism, its central negative commitment makes it a bad idea to also endorse metaphysical disjunctivism. 1.2 From Epistemological to Metaphysical Metaphysical disjunctivism is the view that perceptions and appearances are two fundamentally different kinds of mental state. Thus, when I see a fever tree out of my window, I not only come into possession of different rational support for my beliefs about the tree, as epistemological disjunctivism has it. I am also in a fundamentally different mental state from the one I am in when hallucinating the tree. The factive state is often thought to include the worldly object itself, but there is considerable room for variation and subtlety here.6 Suffice it to say that the perceptual and apparently perceptual experiences are taken to have radically different ontologies: the former is partly external, while the latter is purely internal. Epistemological and metaphysical disjunctivism are generally understood to be logically independent (Byrne and Logue 2008: 67; Pritchard 2012: 29). Nevertheless, some have claimed that the former naturally leads to the latter (arguably, starting with McDowell 1998a).7 Ram Neta (2008) offers the clearest idea of this move, to my mind. Very roughly, it goes like this: The Move 1. The only satisfactory explanation of the differing normative significance of perceptions and appearances is that each provides a different kind of reason for belief. 2. The only satisfactory explanation of the difference in kind of reasons is that there is a metaphysical difference in kind between perceptions and appearances.

Either Epistemological or Metaphysical Disjunctivism

197

3. So, metaphysical disjunctivism supplies the only satisfactory explanation of the negative epistemological disjunctivist thesis.8 I find this a really appealing way of spelling out the move. Doing so in abductive—rather than deductive—terms makes perspicuous the natural affinities between the two forms of disjunctivism, while also preserving their logical independence. But I won’t insist either on the superiority of The Move, or on its fidelity to McDowell’s own argument. My beef, recall, is with the combination of the two forms of disjunctivism, not with the move from the one to the other. The point of sketching it here is just to give a taste of how natural it seems. Now, some have argued that it is a little bit too natural. So natural, in fact, that epistemological disjunctivism entails metaphysical. Stephen Wright, for instance, makes the sensible point that it is ‘mysterious how experiences that are metaphysically identical could come apart in terms of providing reasons’ (2013: 256). Whether this point constitutes an argument for the logical dependence of the two views will turn on how exactly metaphysical disjunctivism is understood. I find the abductive alternative offered here a persuasive way out of logical dependence. But in case you don’t, let’s just say that if the two are logically interdependent in any way, then my argument that we shouldn’t hold both together amounts to an argument that we should hold neither.

2. Reasons So far, I have talked, as typically do disjunctivists, in perfectly general terms about epistemic standings, normative significance, and so on. I now wish to unify these claims as well as make them a bit more precise, by switching to talk of reasons. The way has already been paved by The Move, which introduced a difference in kind of reason as the best explanation of the differing normative standings of perceptions and appearances. But we need one more tool to complete the transition—the distinction between normative and motivating reasons. The former will allow us to articulate epistemological disjunctivism in terms of reasons. The latter will help us see what is wrong with combining it with metaphysical disjunctivism. According to a venerable tradition in the philosophy of action, there are (at least) two kinds of reasons—normative and motivating. A normative reason for an action is a consideration that favours the action.9 A motivating reason is the consideration for which the agent acted as she did. The normative reason is a good reason, something we cite to at least partly justify your action, to show what there is to be said in favour of it; the motivating reason is something we cite to explain your action, to show what you thought there was to be said in favour of it, and hence what moved you to act as you did.10

198 Veli Mitova For instance, the fact that she has been supposedly working late for the last two months, yet keeps coming home smelling like a smoky pub, is a normative reason for you to get a marriage councillor or a divorce attorney. When things go well, this will also be your reason, the motivating reason, for getting an attorney, say. But it may be that your reasons don’t track the normative reasons out there. Thus, suppose that you haven’t even noticed the late nights, but get an attorney, instead, in order to spite her. In this case, the motivating reason is that doing so will spite her, while the normative one remains her late nights. (Your not noticing the late nights and your spite are, presumably, further normative reasons to end the marriage.) There are various ontologies one might adopt for each kind of reason. So, for instance, we could think of the motivating reason as a fact (that this will spite her) or a psychological state (your desire to spite her, or your belief that this will). Although I have my own views on the subject, I will keep them to myself here. All we need from this picture are the concepts: a normative reason to φ is something that favours φ-ing; my reason for φ-ing is what I saw in favour of φ-ing and hence what I acted out of.11 Although not many epistemologists use the language of normative and motivating reasons, they do in fact assume a similar distinction in discussions of the so-called basing requirement.12 This is the idea that in order for a belief to constitute knowledge, it needs to be based on the good reasons available to the believer. Suppose I believe a friend to be honest. Suppose further that there are good reasons for this belief at my disposal: I have often seen him return lost valuable objects to their rightful owner, tell the truth even when it is to his disadvantage, and so on. But suppose I believe that he is honest, because, say, I am in love with him and so am really bent on thinking well of him. In this case, my belief is not based on the good reasons available to me. I then only have ‘propositional’ (or ‘ex ante’) justification, which is insufficient for knowledge. If I had based my belief on the good reasons I have for it, by contrast, I would also have ‘doxastic’ (or ‘ex post’) justification.13 This contrast in justification mirrors the story I gave for action a paragraph ago, suggesting the tacit play of a distinction between normative and motivating reasons for belief. When I base my belief on the reasons I have, then these are the reasons for which I believe as I do. A motivating reason for believing that p, then, is what I saw in favour of believing that p; a normative reason to believe that p is something that favours believing that p.14 The only disadvantage of introducing the normative-motivating distinction through the propositional-doxastic one is that propositional justification is a matter of good reasons available to me, whereas normative reasons aren’t necessarily restricted to considerations available to me.15 Be that as it may, the reason-concepts that tacitly feature in propositional

Either Epistemological or Metaphysical Disjunctivism

199

and doxastic justification are clearly the ones of a normative and a motivating reason, respectively.

3. Disjunctivisms and Reasons With this distinction in hand, we can now get more precise about the negative thesis of epistemological disjunctivism. I do this in subsection 3.1 and then defend the first premise of my core argument, the claim that combining epistemological with metaphysical disjunctivism commits us to disjunctivism about motivating reasons (subsection 3.2). 3.1 Epistemological Disjunctivism and Reasons The negative thesis of epistemological disjunctivism, remember, was that perceptual experiences and mere appearances provide differing support for the beliefs based on them. In reformulating the thesis in the language of reasons, I take my cue from The Move (subsection 1.2), which featured reasons prominently. But reasons were also implicit in McDowell’s formulation of disjunctivism in the passage cited earlier (subsection 1.1). The idea was that perceptual experiences ‘afford opportunities for knowledge’, while apparent perceptual experiences don’t. The most natural way of hearing this is that perceptions are good reasons for belief, while mere appearances aren’t. This was how we started: seeing is believing. Perceptual experiences are—while mere appearances aren’t— things that genuinely favour belief, things that can justify belief, make it rational, and qualify it for knowledge. They are, that is, normative reasons. The negative thesis, then, can be formulated like this: Epistemological Disjunctivism If E is S’s perceptual experience that p then E provides a normative reason for S to believe that p; if E is only an apparently perceptual experience, then it doesn’t provide such a reason.16 Three comments. First, in characterising the view in this way, I don’t take myself to be doing anything controversial. As we have seen, epistemological disjunctivists draw the distinction between perceptions and appearances in terms such as ‘rational support’, ‘epistemic standing’, ‘warrant’, ‘justification’, ‘being in a good position to gain knowledge’, and so on. One needn’t even commit to the (rather trendy) view that a reason is our most basic normative notion, in order to see that these terms are all easily translatable into reasons-speak.17 Indeed, the translation may be helpful for making perspicuous what is common to epistemological disjunctivist views which describe themselves in these different terms.

200 Veli Mitova Second, Epistemological Disjunctivism is obviously not a particularly precise thesis. I leave it at that, nonetheless, to avoid incurring substantive commitments. Such commitments would rob it of its value—to capture the common (negative) tune to which different epistemological disjunctivists dance. Finally, Epistemological Disjunctivism makes the move from epistemological to metaphysical disjunctivism even more natural. In particular, we needn’t go through step (1) of The Move, the claim that the best explanation for the negative thesis of epistemological disjunctivism is a difference in the kinds of reasons perceptions and appearances provide. It seems to me, rather, that the thesis is itself directly formulable in terms of reasons in the way I have just suggested. But I hasten to add that, whether you agree or not, skipping (1) is, at any rate, dialectically unimpeachable in the context of my argument: by representing the relationship between epistemological and metaphysical disjunctivism as even more intimate than earlier, the conclusion that we shouldn’t hold onto both together becomes both harder-earned and more surprising. 3.2 Disjunctivism About Motivating Reasons I now argue that the combination of Epistemological Disjunctivism and metaphysical disjunctivism entails disjunctivism about motivating reasons. Epistemological Disjunctivism is fairly unequivocal about normative reasons for perceptual belief: they are all of one kind—perceptual states (or facts about such states, or the objects of such states, depending on your ontology of normative reasons). But recall the basing requirement from subsection 2.2. It meant, effectively, that evaluations of belief aren’t solely based on normative reasons. We also need to consider motivating reasons, otherwise our evaluation would be silent on something vital for knowledge—doxastic justification. What account of motivating reasons can the epistemological disjunctivist give? Clearly, our perceptual beliefs can be based on either perceptual or apparently perceptual experiences. So, we get: Motivating Reasons A motivating reason for a perceptual belief that p can be (provided by) either a perceptual experience that p or an apparent perceptual experience that p. I give the option of this being a thesis about what provides reasons— rather than what constitutes them—in order to keep neutral on the actual ontology of motivating reasons.18 The thesis is, thus, compatible with the idea that such reasons are facts (the fact that p), or mental states

Either Epistemological or Metaphysical Disjunctivism

201

(the experience that p), or propositions (p).19 It is similarly compatible with the idea that the reasons provided or constituted by perceptual and apparent experiences are of the same ontological kind. So, Motivating Reasons is ontologically neutral on two levels: less abstractly, on what the exact ontology of these reasons is; and more abstractly, on the question of whether all motivating reasons for belief have the same ontology. But now suppose that the epistemological disjunctivist also endorsed metaphysical disjunctivism. Call this ‘double disjunctivism’. Motivating Reasons would immediately lose its ontological neutrality at least at the second, more abstract, level. For, according to metaphysical disjunctivism, as explained in subsection 1.2, the two types of experience belong to two fundamentally different ontological kinds: the perceptual experience contains a bit of world, so to speak, while the appearance is a purely mental state. Thus, the double disjunctivist is committed to a form of disjunctivism about motivating reasons. Let me formulate this third kind of disjunctivism in such a way as to preserve at least the ontological neutrality at the first level, so we don’t fight on too many fronts: Motivating Reasons Disjunctivism Motivating reasons for perceptual belief are of two different ontological kinds: they are (provided by) either factive states partly external to the agent, or non-factive states, purely internal to the agent. So, when things go well, my belief that there is a fever tree outside my window is based on my seeing it. When things don’t go well, my belief is based on an altogether different ontological entity.

4. Disjunctivism About Motivating Reasons Is Implausible If this is right, we have established the first premise of the core argument: the conjunction of epistemological and metaphysical disjunctivism entails disjunctivism about motivating reasons for perceptual beliefs. I now argue for the second premise: this view is implausible. The reason is that it implausibly implies that motivating reasons are shapeshifters. A thought-experiment should drive the point home. Manish lives in a house with a beautiful banana tree just outside his top veranda. The tree is in flower, and Manish cannot get enough of the flowers. In fact, he is so fond of them that he sets aside an hour every day between 1 and 2 p.m. to do nothing but contemplate them, banishing his whole family from the veranda at those times. Suppose that the Evil Demon’s first cousin, the Perverse Demon, decides to play a trick on Manish. Every day she waits for him to make himself comfortable on the veranda, and then starts switching back and forth between allowing the

202 Veli Mitova banana tree to come into full view and interposing a perfectly indistinguishable hologram of it. She follows this particular timeline: • • • • • • •

1:00–1:15 Manish sees the tree; 1:15–1:20 he is looking at the hologram; 1:20–1:30 he is looking at the tree; 1:30–1:35 the Perverse Demon interposes the hologram; 1:35–1:45 she graciously allows Manish to see the real thing; 1:45–1:59 back to the hologram; 1:59–2:00 he sees the tree.

During this hour Manish continuously believes that there is a flowering banana tree in front of him. What kinds of reasons are there for his belief, according to the double disjunctivist? Normative reasons first. Since epistemological disjunctivism says that only veridical experiences provide normative reasons, Manish may be said to have a normative reason to believe as he does four times (1:00–1:15, 1:20–1:30, 1:35–1:45, and 1:59–2:00). The reason goes away thrice, as the Perverse Demon introduces the hologram. I say ‘may be said’, because we might wonder whether Manish has a normative reason at any point, given the Gettier quality of the situation.20 The present argument is silent on normative reasons, so it can go with either reading. Our business is with motivating reasons. So, what is the reason for which Manish believes that there is a flowering banana tree in front of him? If disjunctivism about motivating reasons is correct, there are seven reasons. Between 1:00 and 1:15, it is the fact that he is seeing that there is a flowering banana tree; between 1:15 and 1:20, it is the non-veridical experience, then back to perception, and so on.21 How plausible is it that motivating reasons are this protean? Not terribly, it seems to me, for at least three reasons. First, a view that allows such protean changes would be theoretically infelicitous. The whole point of positing motivating reasons in the first place is to give one’s account of reasons the resources to represent the agent’s perspective on what she does and believes. The concept of motivating reason, in other words, is tied to what it is that the agent saw in favour of the action or belief, something that from her point of view makes acting and believing as she does intelligible. Although we are not, of course, infallible about our motivating reasons, to have them float completely free of our perspective would seem to undermine the whole theoretical point of positing such reasons. But this is just what the present proposal does at both the metaphysical and epistemic levels. At the metaphysical, according to the story, Manish’s reasons have changed seven times without his perspective having changed once. Thus, our reasons can’t depend on our perspective. At the epistemic level, the view implies that the reasons for which we believe

Either Epistemological or Metaphysical Disjunctivism

203

can change without our having the slightest inkling of this change. But if our motivating reasons were unmoored from our perspective both metaphysically and epistemically, it would be difficult to see how they are something that reveals what the agent saw in favour of believing as she does and what she takes to make her belief intelligible. It would be thus difficult to see what we gain theoretically by positing an entity of this kind over and above normative reasons.22 Notice that this point does not presuppose a particular ontology of motivating reasons, just the idea that they are dependent on our perspective. This dependence is granted on all sides. Even the factualist—who puts the greatest distance between our reasons and our perspective— thinks that such reasons are facts as the agent takes them to be. Hence, a factualist doesn’t face the problem presented here. She would say that the motivating reason remains the same thing—a fact as Manish takes it to be—even though he is mistaken three times that the fact obtains. The second reason why it is implausible to think that motivating reasons are shapeshifters is that such reasons are most naturally thought of as causes. The understanding is natural partly because the basing requirement is typically understood in causal terms. For a belief to be based on the good reason for it, it needs to be caused by that reason.23 But also, the notion of a motivating reason for action, on which the present epistemic one is modelled, is generally taken to be a causal one, since such reasons, as their name suggests, are what moves the agent to action.24 Yet it doesn’t seem plausible to suppose that there are seven causes sustaining Manish’s belief in the space of an hour without any change in the belief itself. After all, one common way of thinking of causes is as those things without which the effect wouldn’t have occurred.25 When Manish’s perceptual state is replaced with an appearance, the original cause of the belief goes away, so, presumably, there should be at least some change in the belief. I don’t mean to make too much of this point, as causation is a vexed issue by all accounts. So is the claim that a motivating reason is a cause. Nonetheless, it seems to me that the point does provide some suggestive support for the implausibility of attributing seven motivating reasons to Manish. So far, I have only discussed an example of motivating reasons for a perceptual belief. The final reason for thinking that such reasons are not shapeshifters is that it is implausible to think that any motivating reasons—either for belief or for action—are shapeshifters. First, take other, perception-unrelated, kinds of motivating reasons for belief. Suppose that at 1 p.m. I believe that the meeting will take place at 2 p.m., based on your testimony.26 We could imagine cases in which you keep losing and regaining the justification for your belief that the meeting will start at 2 p.m., following the timeline of the Manish example. This may, in turn, affect the justificatory status of my belief (again, depending on whether your account of reasons requires just truth or more robust

204 Veli Mitova normative elements). It may affect, that is, the normative reasons for believing as I do. But it seems implausible that the reason for which I believe that the meeting will start at 2 p.m. should change in the course of the hour. Next, take motivating reasons for action.27 Suppose I am driving to your place where I am meant to meet you at 1 p.m. And suppose that I am stuck in traffic for an hour without being able to let you know. Suppose you kept popping in and out of your place for various reasons in the space of this hour (first you went to get some coffee, then some smokes, etc.). It is plausible to think that my motivating reason for driving to your place remains constant throughout—it is the fact, or my belief, that you are waiting for me (depending on your ontology of reasons).28 Yet, clearly, the truth value of the proposition that you are waiting for me can change as often as it did in the Manish case. How does this support the claim that epistemic motivating reasons aren’t protean? Such reasons, recall, were modelled on motivating reasons for action. So, it would be strange if the one bunch of reasons were shapeshifters while the other remained constant through thick and thin. This point is enforced by the rising popularity of what we might call the Uniformity of Reasons Thesis, the idea that at a sufficient level of abstraction, your stance on reasons should be consistent across the practical and epistemic normative domains.29 If you thought that motivating reasons for action can’t shift shape while motivating reason for belief can, you’d run afoul of this thesis. Of course, it may be that you don’t like this thesis, or you think that reasons for action can be protean. If so, you wouldn’t buy this third argument for why motivating reasons for belief can’t be shapeshifters. Nonetheless, you will still have the first two considerations to reckon with. But, in any case, I would be happy to have at least achieved the minor triumph of demonstrating the interrelatedness of these issues. This, recall, was the more modest aim of the chapter: to show that disjunctivists are accountable to broader debates about reasons.

5. Objections If the thinking so far is on the right track, one shouldn’t endorse both epistemological and metaphysical disjunctivism. And if the two views are logically dependent, one shouldn’t endorse either. The core argument, recall, was this: Premise 1: The conjunction of epistemological and metaphysical disjunctivism entails disjunctivism about motivating reasons for belief. Premise 2: Disjunctivism about motivating reasons for belief is implausible. Conclusion: So, the conjunction of epistemological and metaphysical disjunctivism is implausible.

Either Epistemological or Metaphysical Disjunctivism

205

In this section I consider an objection to each of the argument’s premises. By so doing, I hope to place the argument on firmer footing. But if I don’t quite succeed, I will also be happy to simply shed light on the commitments one incurs by rejecting each premise. 5.1 No Disjunctivism About Motivating Reasons? The most obvious way for the double disjunctivist to resist the core argument is to deny that she is committed to disjunctivism about motivating reasons. Recall Motivating Reasons Disjunctivism: Motivating reasons for perceptual belief are of two different ontological kinds: they are (provided by) either factive states partly external to the agent, or non-factive states, purely internal to the agent. Earlier, I added the parentheses in order to preserve the neutrality of the thesis at least at the level of what the exact ontology of perceptions and appearances is. But suppose we now took them out and argued that the two kinds of reasons being provided by two different ontological entities doesn’t amount to the two reasons themselves being different ontological entities.30 Compare: a quarter of Joe’s money is provided by his university job, the other three quarters by his partner. A job and a partner are two different ontological entities (unless, perhaps, you are a rabid workaholic), but no one would suppose that the two portions of money have different ontologies. Likewise, it could be that all perceptual motivating reasons are of the same ontological kind, even though they have their source in two different ontological entities, perceptions and appearances. The most plausible way of spelling out this suggestion (that I know of) is to go ‘representationalist’.31 A motivating reason, according to representationalism, is one’s appreciating that a certain consideration counts in favour of a belief or action. Jay Wallace articulates the idea neatly: The motivating reason—the consideration we cite in the perspective of explanation, to make sense of what the agent has already done—is the agent’s seeing that normative consideration as counting in favor of the action performed, a normative consideration that, as it happens, was also a good reason for so acting. (2003: 435; emphasis added) The disjunctivist could adopt this view and argue that a motivating reason for belief is not dependent on whether one’s experience is a perception or an appearance. It is always the same ontological entity—the agent’s seeing or appreciating that a certain experience favours a certain belief. The trouble with this suggestion comes into view when we start wondering how exactly we should understand this single entity of seeing or appreciating? In a word: if we think of the seeing in factive terms, our

206 Veli Mitova account of motivating reasons will be incomplete; if we think of it in experiential terms, we are back with disjunctivism about motivating reasons; but if we think of it in doxastic terms, we end up with a picture of justification which is both ill-motivated for a disjunctivist and ill-advised in general. Suppose we thought that the relevant state of seeing or appreciating is factive. Then clearly, our account of motivating reasons will be incomplete. People sometimes think they see a consideration favouring a belief, but it doesn’t in fact favour the belief. An account which thought of motivating reasons as factive would not have the resources to account for such cases, and would hence be incomplete. Suppose next we understood in perceptual terms this seeing or appreciating that a certain experience favours a certain belief. Then clearly, our motivating reasons, which are (facts about) such seeings, will be of two ontological kinds according to the double disjunctivist. Epistemological disjunctivists think that the good rational support that perceptions provide is precisely provided by (the fact that one is) seeing that p.32 Once you are a double disjunctivist, you also think that perceiving is a different ontological entity from merely apparently perceiving. If you didn’t think this, you couldn’t think that the former state is a normative reason while the latter isn’t. But if you did think this, you would think that seeing that you have a reason is different in kind from merely appearing to see that you have a reason. So, motivating reasons would be of two different ontological kinds. Hence, thinking of motivating reasons as either facts about experiences or the experiences themselves would land the double disjunctivist precisely back into the trouble that she was trying to avoid by being a representationalist about motivating reasons. Perhaps, then, the motivating reason, according to the present representationalist proposal, is a belief, or some weaker doxastic state, that my perceptual experience favours believing that p. But this suggestion won’t fare much better. For it would make whatever justification my experiences give me inferential, and this is something that disjunctivists would resist. Now, I promised at the beginning to steer clear of positive disjunctivist commitments, but there is a very good reason for this one, both for a disjunctivist and in general. The reason people are keen on thinking that perceptual beliefs are justified by the perceptual experiences themselves is that this is the only way of forestalling an infinite regress of justification.33 If my belief that I am seeing that p is what justifies my belief that p, we can always ask what justifies my belief that I am seeing that p. Whatever belief we cite in answer to this question will in turn need to be justified, and so on ad infinitum. Experiential justifiers are seen as the most promising way of stopping this regress. But if we now claim that our reasons are further beliefs, then we have just resuscitated the regress in full force. So, we need the motivating reason to consist of the experience itself, not of beliefs

Either Epistemological or Metaphysical Disjunctivism

207

about it or facts about such beliefs. But in that case, if we are also epistemological and metaphysical disjunctivists, we need to be disjunctivists about motivating reasons. Premise 1 of the core argument stands. 5.2 No Protean Changes? I now turn to the second premise of the core argument, the claim that disjunctivism about motivating reasons for belief is implausible. The subargument I gave for this premise was that such disjunctivism commits its friend to the implausible view that motivating reasons are protean. One could block this argument by denying either the commitment or its implausibility. Since I already spent some time arguing that protean changes are implausible (section 4), the most promising way to go at this point would be to deny the commitment. This wouldn’t show that disjunctivism about motivating reasons is plausible, but it would certainly undermine my argument for the claim that it isn’t. How can the disjunctivist about motivating reasons maintain that Manish’s reason stays the same even though his experiences keep changing from perception to appearance and back again between 1 and 2 p.m.? The most natural way would be to follow up on the second argument I gave for the implausibility of protean changes. Look, the disjunctivist could say, you claim that a motivating reason is a cause, and that it would be hence difficult to see how the motivating reason can change while the belief stays the same. But this point, the disjunctivist can continue, cuts both ways. Far from showing that it is implausible to be committed to protean changes, it shows that we are not committed to changes at all. The question then becomes what Manish’s motivating reason for believing as he does is, such as it remains the same throughout the episode. The most obvious candidate is the fact that it seems to him that there is a flowering tree in front of him.34 But this candidate is dialectically ineligible at this point, for two reasons. First, the present objector is trying to show that disjunctivism about motivating reasons does not entail that such reasons are protean. To do this, she needs to stay disjunctivist about motivating reasons. But the present suggestion abandons this view. For if the reason is the seeming itself, then it would be so regardless of whether the seeming is veridical or illusory. If it made an ontological difference to the seeming whether it is veridical or illusory, then we would be back with protean changes. So, the seeming must be of one kind to avoid such changes. But then all our motivating reasons (for perceptual belief) are of one kind. Second, and more tentatively, the present suggestion also seems to tacitly abandon epistemological disjunctivism. For usually people who think of our reasons as seemings also think that such seemings give us prima facie warrant for believing as it seems to us (e.g., Huemer 2007). But that is just to say that the seeming is a good reason for belief. This, of course,

208 Veli Mitova goes against the grain of epistemological disjunctivism, whose central negative thesis is that only genuine perceptions provide good reasons. So, the single state that is Manish’s motivating reason for believing that there is a flowering banana tree before him between 1 and 2 p.m. can’t be its seeming so to him. The only plausible candidate, then, for the single motivating reason which will avoid protean changes is Manish’s first perception of the tree. It is presumably this that caused his belief. And the present objection works from the causal character of motivating reasons to their constancy in the face of changing experiences such as Manish’s. This suggestion at least has the virtue of not abandoning disjunctivism about motivating reasons. For it is quite possible that, in general, motivating reasons are of two kinds—perceptions and appearances—but in Manish’s case, only one kind is operant throughout. Notwithstanding this virtue, however, the suggestion founders in a different way: it seems to render motivating reasons normatively useless. A motivating reason, if you recall, is the reason on which one bases one’s belief. It is, hence, the reason we look at in order to evaluate the belief’s (doxastic) justificatory status. But this wouldn’t make sense if the motivating reason was just the initial cause of the belief. Two variations on the Manish case will drive the point home. First, suppose that instead of starting with the real tree, the Perverse Demon started by mistake with the hologram, left it there for two seconds, and then followed the timeline of the original thought-experiment. On the current proposal, this means that Manish’s motivating reason is now the appearance rather than the perception, since it was the appearance that caused the belief. But this would mean, in turn, that in this scenario, Manish’s belief that there is a flowering banana tree before him would be based on a worse reason (an appearance) than in the original scenario (where it is based on a perception). Hence, its justificatory status would be inferior in the present scenario. But this seems implausible. The belief is the same; its causal trajectory is exactly the same with the slight two-second difference; Manish’s perspective is the same; so intuitively, the belief should enjoy the same justificatory status in the two cases. Suppose alternatively (and this is the second case) that after briefly allowing Manish to see the real tree, the Perverse Demon got side-tracked and forgot the hologram there for a whole week. According to the present story, the motivating reason for Manish’s belief over that week would remain the same—it would be the perception. Worse, since Manish based his belief on the perception, the justificatory status of the belief would be positive (or at least more positive than if he’d started with the hologram). This, I hope you agree, is counterintuitive. The brief presence of a perception in a sea of illusion shouldn’t be granted such hefty normative significance. Conversely, the week of illusion that follows should play a more prominent role in fixing the justificatory status of the belief during that week.

Either Epistemological or Metaphysical Disjunctivism

209

We can continue imagining alternative stories, but I will spare the reader. Although the considerations presented here aren’t conclusive, I think they do suggest that the present proposal gives counterintuitive verdicts in both directions: it creates normative differences where there are none, and doesn’t register normative differences when it should. Hence my claim that it renders the notion of motivating reasons normatively useless. Perhaps it is open to the disjunctivist about motivating reasons to deny these results by arguing that the perception and illusion in each case do not do the positive normative work that I am envisaging, because of the Gettier quality of the situation. I am not sure how she can do that without acknowledging that what happens after the initial bringing about of the belief matters normatively, and hence edging towards the protean commitment she is currently trying to avoid. Perhaps this could be done by going externalist about justification. But on this suggestion, we should forget about reasons altogether and look directly to the causal history of the belief. Whether this is a viable option would depend on how sold the disjunctivist is on out-and-out externalism and on abandoning reasons. I have a suspicion that not many will be that sold.35 But the point here is that for those who aren’t, the present strategy of avoiding the commitment to protean motivating reasons doesn’t look very promising.

6. Conclusion I have argued here that we should not hold the combination of epistemological and metaphysical disjunctivism. As already mentioned, this result can be taken to have different import for different kinds of disjunctivists. If you think that epistemological and metaphysical disjunctivism are logically independent, but you are nonetheless keen on both, the argument shows that you shouldn’t be. If you think that the two principles are logically interdependent, the argument shows that you shouldn’t be keen on either. Of course, you might want to question whether the combination of epistemological and metaphysical disjunctivism really does commit us to disjunctivism about motivating reasons, or the claim that such a commitment is implausible. You will then not accept my conclusion about double disjunctivism. But you will be consolidating the more modest point I have tried to make here: these debates are interconnected in deeper ways than the current literature allows, and we shouldn’t try to decide on the merits of various kinds of disjunctivism without having in place a more general account of reasons.36

Notes 1. There seems to be no consensus amongst metaphysical disjunctivists about the nature of hallucination and illusion or the relationship between the two

210 Veli Mitova

2. 3.

4.

5.

6. 7. 8.

(see Byrne and Logue 2008: 60–61 for a helpful overview). This makes no trouble for my argument, since I don’t rely on any positive claims about the ontology of either. See, e.g., Byrne and Logue (2008), Haddock and Macpherson (2008), Neta (2008). No disrespect intended to Michael Huemer’s appearances, which can be veridical (e.g., Huemer 2007). The advantage of reserving the label ‘perceptual’ for veridical experiences is that we don’t beg any questions against the metaphysical disjunctivist by assuming that veridical and non-veridical experiences share a common element. Though, unlike McDowell, Pritchard does not endorse metaphysical disjunctivism (2012: 29). But see Cunningham (2016), for an argument that any Pritchard-style access-internalist version of epistemological disjunctivism is, in fact, committed to metaphysical disjunctivism. If this is right, then the present argument shows that Pritchard is as vulnerable to the argument I develop here as anyone who explicitly endorses both kinds of disjunctivism. French argues that Pritchard’s positive thesis in fact underpins the negative: it is because perceptual experiences are factive and reflectively accessible that they provide a different kind of support. While I agree with the factivity point, I think that to combine it with accessibility makes for an uneasy grounding of the negative thesis. See, for instance, Madison (2014) for an argument that the insistence on factivity forces us to deny the very motivation for an access requirement—the New Evil Demon Intuition. For a good survey, see Soteriou (2010). See, e.g., Haddock and Macpherson (2008). In Neta’s own words: the only satisfactory explanation of the fact that veridical perception puts us in a better epistemic position than hallucination does is an explanation according to which veridical perception provides one with different reflectively accessible reasons for belief from those provided one by hallucination. And the only explanation that involves that claim, so the argument continues, is an explanation according to which veridical perception is a fundamentally different kind of mental state than hallucination. Since the only satisfactory explanation of the fact that veridical perception puts us in a better epistemic position than hallucination does is an explanation according to which veridical perception is a fundamentally different kind of mental state than hallucination, it follows that the only satisfactory explanation of this fact implies disjunctivism about perception. (2008: 314–315)

9. Despite their widespread popularity, these characterisations of normative and motivating reasons aren’t completely unchallenged (see, e.g., Hieronymi 2005; Broome 1997, respectively). For arguments that such challenges are misguided, see Mitova (2017: section 1.2) 10. Hence, the distinction is sometimes drawn in terms of ‘justifying’ and ‘explanatory’ reasons (Dancy 2000: 20–25; Alvarez 2010: 36). There are other labels, too. Scanlon (1998: 19) dubs motivating reasons ‘operative reasons’, and Davidson calls them ‘the agent’s reasons’ (2001: 3). 11. How is the latter ontologically neutral? If you thought that the ‘what’ is what matters, then you’d think that motivating reasons are facts; if you thought the ‘seeing’ is what matters, then you’d think them psychological states. (As far as I recall, it was Maria Alvarez who first made me see this, in conversation.) You could, of course, think that both matter, and that the

Either Epistemological or Metaphysical Disjunctivism

12. 13.

14.

15. 16.

17. 18. 19.

20. 21.

22.

23.

211

circumstances dictate which we should go for. Then you’d be a disjunctivist about motivating reasons (see section 4). For a good discussion of basing, see Turri (2011). Most epistemologists accept the basing requirement on justification in one form or another (e.g., Goldman 1979; Neta 2002; Turri 2011). And even those who deny it as a requirement on justification accept it as a requirement on knowledge (e.g., Conee and Feldman 2004). I appreciate that formulating things this way gives rise to the so-called Wrong Kind of Reasons Problem (e.g., Hieronymi 2005). The alternative is to say that a normative reason for believing that p is something that favours the truth of p, but that is to take sides in the debate between (broad) evidentialists and pragmatists (e.g., Reisner 2018). I don’t need to do this for present purposes. I do so and offer a tentative solution to the Problem in Mitova (2017). Needless to say, there is a lively debate on this issue. See e.g., Alvarez (2018). Notice that the content of the experiences is propositional. This is how disjunctivists put it (e.g., Pritchard 2012: 21; see also the McDowell quote in subsection 1.1, where he talks of states of affairs). Nothing in my argument turns on this. But see French (2016) for an argument that a non-propositional formulation would overcome certain standard disjunctivist bugbears, such as the Basis Problem. Proponents include Parfit (2001: 18) and Scanlon (1998: ch. 1). For an argument that we should, instead, take oughts to be more basic, see Broome (2013). Some theorists take talk of something’s being a reason and something’s providing a reason to be equivalent (e.g., Parfit 2011: 32). I think that this thought unnecessarily courts confusion. How can the fact that p serve as a reason in the hallucination case when it is not the case that p? Factualists about motivating reasons think that the reason is the fact as the agent takes it to be, where the italicised bit doesn’t amount to positing the psychological state as the reason, but as a mere enabling condition on the reason (e.g., Alvarez 2010; Dancy 2000). Thanks to Casey Doyle for this point. A similar thought-experiment is offered by Johnston (2004) as an argument against metaphysical disjunctivism, a version of the so-called argument from hallucination. (I should mention that I came across it after I’d developed mine. See fn. 27.) This chapter could be seen as a sort of diagnosis of why the argument from hallucination has such force: we simply find it implausible that people’s motivating reasons are shapeshifters. Notice, though, that my argument shows that metaphysical disjunctivism on its own is not the problem, and hence that the argument from hallucination is incomplete without the extra connections I draw here between epistemological and metaphysical disjunctivism, and their relationship to reasons. The double disjunctivist could reply that her epistemological disjunctivism explicitly states that perceptions provide accessible rational support, and hence there is a change in perspective when perceptions and appearances switch. But I think we would be excused if we remained unconvinced. The whole point of talking about my perspective is to isolate the sorts of things that make an epistemic difference from my point of view. This surely means that they make some sort of felt difference to me. But the disjunctivist concedes that perceptions and appearances are phenomenologically indistinguishable. (She just refuses to grant this indistinguishability normative significance.) This is not, of course, universally accepted. For a discussion and arguments that basing is causal, see, e.g., Turri (2011).

212 Veli Mitova 24. Again, this point has its opponents (e.g., Dancy 2000: ch. 8). For my argument that such opposition is implausible, see Mitova (2017: 69–72). 25. I mean this as a perfectly general counterfactual understanding of causation rather than any particular account. For a good discussion of, and some problems for, counterfactual accounts see Broadbent (2007). 26. Thanks to Casey Doyle for making me think about this sort of case in the different context of normative reasons. 27. I develop this argument at greater length in Mitova (2017: 44–45). 28. In case you are seeing red flags with this characterisation of the content of my reason, let me assure you that I am in good company. Here, for instance, is Jennifer Hornsby: ‘Sam is headed for Beech Street because [for the reason that] Ann is there’ (2008: 254, her square brackets). 29. The name is inspired by Jonathan Drake (2018). Other proponents of various versions of the thesis include Gibbons (2010), Littlejohn (2014), Williamson (2017) and Alvarez (2018), to name but a few. 30. A quick response here would be that many theorists use, as a matter of fact, the notions of something’s being a reason and providing a reason interchangeably (e.g., Parfit 2011: 32). But since I think that this is a mistake, I can’t avail myself of this reply. 31. As far as I know, this is Susanna Mantel’s (2014) label. 32. Haddock and Macpherson (2008: 10–11) argue that it isn’t clear whether McDowell thinks that reasons are the experiences or facts about the experiences. But other epistemological disjunctivists are fairly unambiguous. Here, for instance, is Pritchard: when one has paradigmatic perceptual knowledge of a proposition, p, one’s reflectively accessible rational support for believing that p is that one sees that p. (2012: 14, his italics, my underlining) 33. I am personally not a fan of the view (Mitova 2017: 161–170), but the important thing here is that the disjunctivist needs something like it. 34. I leave to one side my general scepticism (Mitova 2017, inspired by Dancy 2000: 125), that such facts are ever good reasons and hence are a good candidate for motivating reasons. 35. I base the suspicion on McDowell’s (1998a: 374) and Pritchard’s (2012: 13) insistence that the rational support perceptions provide is reflectively accessible to the subject. 36. Many thanks for helpful comments to Casey Doyle and my audience at the 2017 Annual Conference of the Philosophical Society of Southern Africa.

References Alvarez, Maria, 2010, Kinds of Reasons: An Essay in The Philosophy of Action. Oxford: Oxford University Press. ———, 2018, False beliefs and the reasons we don’t have, in Veli Mitova, ed. The Factive Turn in Epistemology. Cambridge: Cambridge University Press, pp. 161–176. Broadbent, Alex, 2007, Reversing the counterfactual analysis of causation, International Journal of Philosophical Studies 15(2): 169–189. Broome, John, 1997, Reasons and motivation, Aristotelian Society (supp. vol.) 77: 131–146. ———, 2013, Rationality Through Reasoning. Oxford: Wiley-Blackwell.

Either Epistemological or Metaphysical Disjunctivism

213

Byrne, Alex and Logue, Heather, 2008, Either/or, in A. Haddock and F. Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 57–94. Conee, Earl and Feldman, Richard, 2004, Evidentialism. Oxford: Oxford University Press. Cunningham, J.J., 2016, Reflective epistemological disjunctivism, Episteme 13(1): 111–132. Dancy, Jonathan, 2000, Practical Reality. Oxford: Oxford University Press. ———, 2008, On how to act—disjunctively, in A. Haddock and F. Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 262–279. Davidson, Donald, 2001, Essays on Actions and Events (2nd ed.). Oxford: Clarendon. Drake, Jonathan, 2018, Motivating reason to slow the factive turn in epistemology, in Veli Mitova, ed. The Factive Turn in Epistemology. Cambridge: Cambridge University Press, pp. 177–192. French, Craig, 2016, The formulation of epistemological disjunctivism, Philosophy and Phenomenological Research 92(1): 86–104. Gibbons, John, 2010, Things that make things reasonable, Philosophy and Phenomenological Research 81(2): 335–361. Goldman, Alvin, 1979, What is justified belief? in G. Pappas, ed. Justification and Knowledge. Dordrecht: Reidel. Haddock, Adrian and Macpherson, Fiona, eds., 2008, Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press. Hieronymi, Pamela, 2005, The wrong kind of reason, Journal of Philosophy 102(9): 437–457. Hornsby, Jennifer, 2008, A disjunctive conception of acting for reasons, in A. Haddock and F. Macpherson, eds., Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 244–261. Huemer, Michael, 2007, Compassionate phenomenal conservatism, Philosophy and Phenomenological Research 74(1): 30–55. Johnston, Mark, 2004, That obscure object of hallucination, Philosophical Studies 120(1): 113–183. Littlejohn, Clayton, 2014, The unity of reason, In C. Littlejohn and J. Turri, eds. Epistemic Norms: New Essays on Action, Belief and Assertion. Oxford: Oxford University Press, pp. 135–154. Madison, B.J., 2014, Epistemological disjunctivism and the new evil demon, Acta Analytica 29: 61–70. Mantel, Susanne, 2014, No reason for identity: On the relation between motivating and normative reasons, Philosophical Explorations 17(1): 49–63. McDowell, John, 1998a, Criteria, defeasibility, and knowledge, in his Meaning, Knowledge, and Reality. Cambridge, MA: Harvard University Press, pp. 369–394. ———, 1998b, Singular thought and the extent of inner space, in his Meaning, Knowledge, and Reality. Cambridge, MA: Harvard University Press, pp. 228–259. ———, 2008, The disjunctive conception of experience as material for a transcendental argument, in A. Haddock and F. Macpherson, eds., Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 376–389.

214 Veli Mitova Mitova, Veli, 2017, Believable Evidence. Cambridge: Cambridge University Press. ———, 2018, Introduction: The factive turn, in Veli Mitova, ed. The Factive Turn in Epistemology. Cambridge: Cambridge University Press, pp. 1–12. Neta, Ram, 2002, S knows that P, Nous 36(4): 663–681. ———, 2008, In defence of disjunctivism, in A. Haddock and F. Macpherson, eds., Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 311–329. Parfit, Derek, 2001, Rationality and reasons, in E. Dan et al., eds. Exploring Practical Philosophy. Burlington, VT: Ashgate, pp. 17–39. ———, 2011, On What Matters. New York: Oxford University Press. Pritchard, Duncan, 2012, Epistemological Disjunctivism. Oxford: Oxford University Press. Reisner, Andrew, 2018, Pragmatic reasons for belief, in D. Star, ed. The Oxford Handbook of Reasons and Normativity. Oxford: Oxford University Press, pp. 705–728. Scanlon, Timothy, 1998, What We Owe to Each Other. Cambridge, MA: Harvard University Press. Soteriou, Mathieu, 2010, The disjunctive theory of perception, in Edward N. Zalta, ed. Stanford Encyclopedia of Philosophy (Winter 2010). http://plato. stanford.edu/archives/win2010/entries/perception-disjunctive/. Turri, Jonathan, 2011, Believing for a reason, Erkenntnis 74: 383–397. Wallace, R. Jay, 2003, Explanation, deliberation, and reasons, Philosophy and Phenomenological Research 67(2): 429–435. Williamson, Timothy, 2000, Knowledge and Its Limits. Oxford: Oxford University Press. ———, 2017, Acting on knowledge, in A.J. Carter, E.C. Gordon and B. Jarvis, eds. Knowledge-First. Oxford: Oxford University Press, pp. 163–181. Wright, Stephen, 2013, Review: Duncan Pritchard: Epistemological disjunctivism, Dialectica 67(2): 252–257.

11 Neither/Nor Clayton Littlejohn

1. Introduction On the formulation discussed here, epistemological disjunctivism is the view that in paradigmatic cases of perceptual knowledge, a thinker’s perceptual beliefs constitute knowledge when they are based on reasons that provide them with factive support (i.e., the complete description of the thinker’s reason for believing, say, that it is Agnes curled up on the sofa entails that Agnes is curled up on the sofa).1 A thinker is in a position to know that p perceptually if the thinker sees that p. It is the seeing that p that constitutes the thinker’s reason for believing p and provides the requisite support for that belief. This perceptual relation between a thinker and a fact guarantees that the thinker is in a position to know things about things in her surroundings. Without this kind of support, perceptual knowledge isn’t possible. Some philosophers accept epistemological disjunctivism because they think that alternative accounts of the rational basis of perceptual belief lead to scepticism. Some accept metaphysical disjunctivism because they think that this epistemological proposal requires it.2 The guiding idea seems to be that on the traditional view of experience (i.e., a view on which the experience we have in the good case is of the same fundamental kind as the experience we have in the bad), it wouldn’t be possible for experience to provide a different basis for perceptual belief in the good case and the bad. However, the epistemological disjunctivist proposes that in paradigmatic cases of perceptual knowledge, experience provides us with reasons that provide factive support for some perceptual beliefs about the external world. We cannot have such support for our beliefs in the bad case if, say, our perceptual beliefs are mistaken. Thus, there is a prima facie plausible line of argument from epistemological to metaphysical disjunctivism.3 As someone who has, if anything, too much sympathy for naïve realism, I don’t have any problem with the idea that the experience we have in the good case might provide a kind of contact that some indistinguishable experiences could not hope to provide. As someone who has

216

Clayton Littlejohn

just the right amount of sympathy for the idea that our evidence just is our knowledge (‘E=K’ hereafter), I have no problem with the idea that the evidence we have in good cases is different from and better than the evidence we have in the correlative bad cases. I don’t object to epistemological disjunctivism on the grounds that it clashes with the idea that our reasons for believing what we do couldn’t be better than the reasons we would have if we were BIVs, say.4 My main concern with the disjunctivist account of perceptual knowledge under consideration is that it tries to account for positive epistemic standing in terms of the support provided by the thinker’s reasons. Our non-inferential beliefs do not attain positive standing because of the support provided by our reasons for holding these beliefs because, I have argued, there is nothing that is our reason for forming a non-inferential belief.5 Things go off the rails when we start with this mistaken claim about the rational role of the thinker’s reasons and then use this mistaken claim about reasons to motivate or justify claims about the nature of perceptual experience.6 If you think that these non-inferential beliefs are justified because of the support some thinker’s reasons for forming these beliefs provides, you have to adopt an account of experience on which experience is sufficiently belief-like to make it possible for the non-inferential beliefs to be based on reasons in much the same way our inferential beliefs are. Bad epistemology paves the way for mistakes in the philosophy of mind. Because I accept E=K and think that a thinker’s knowledge just is her evidence, I don’t think that cases of non-inferential knowledge are helpfully modelled on cases where a thinker has some evidence, sees what it supports, and judges that something is so in such a way that the facts that constitute the evidence constitute the thinker’s reasons for judging what she does. Because a thinker’s evidence just is her knowledge, we shouldn’t describe processes by which non-inferential knowledge is acquired as processes that take pieces of evidence as input and give us non-inferential knowledge as an output.7 After I explain the key features of the disjunctivist approach to perceptual knowledge under discussion, I shall argue for two points. The first is that the disjunctivist cannot give us a suitable account of the reasons that support our perceptual belief. If the disjunctivist were right that such reasons were needed for knowledge or for justification, this would be an unfortunate result. Luckily, the claim that we need these reasons for justification and knowledge turns out to be unmotivated. Once you recognise that a belief’s justification turns on whether it constitutes knowledge, you should see that a belief’s justification doesn’t turn on whether the thinker’s reasons provide sufficient support for that belief.

2. Two Ways to Stand in the Space of Reasons According to McDowell (1998, 2002), knowledge is a standing in the space of reasons, in two senses:

Neither/Nor

217

1. For McDowell, knowledge is a standing in the space of reasons in the sense that a thinker who believes what she does not know lacks standing. If she believed only what she knew to be true, she would have standing, but not if she doesn’t: Normativity of Knowledge: If S believes p without knowing p, she should not believe p.8 The situation of the thinker who believes what she doesn’t know isn’t all that different from the situation of someone who is standing on property that they mistakenly believe is theirs. They try to exclude someone from it who they don’t realise has the authority to be there. It might seem that they have standing to exclude (at least, to people who don’t know where the property lines should be drawn), but they don’t. However good their reasons might be for believing they have this right, they lack the right to exclude. However reasonable they might generally be, there is a decisive reason for us to not assist them in trying to exclude others from this property. However reasonable our thinker might be, if she believes what she doesn’t know, there is a decisive reason for our thinker not to use their belief in theoretical reasoning as a tool for excluding or ruling out possibilities. Just as the right to property had better come with ancillary rights (e.g., the right to exclude, to use, to sell, etc.), the right to believe had better come with ancillary rights (e.g., the right to use that belief in deliberation). Once we know that the belief cannot be properly used to rule out possibilities, we know that the apparent right to believe is only that, no more. 2. For McDowell, knowledge is a standing in the space of reasons in the sense that a thinker’s response (in this case, belief) attains the standing it does because it is backed by the thinker’s reasons: Knowledge-Reasons: If S knows that p, S’s being in a position to know supervenes upon the reasons she possesses, and her belief is properly based on sufficient reasons. One of the questions that I want to try to answer here concerns the relations between these two claims. Let’s suppose for the time being that knowledge is a standing in the space of reasons in the first sense. Should we say that knowledge is a standing in the space of reasons in this second sense? The first question is a question about motivation. Why should we accept that knowledge is a standing in the space of reasons in the second sense? Without the Knowledge-Reasons thesis, it isn’t at all clear why we should accept the version of epistemological disjunctivism under consideration. If we don’t need support from the thinker’s reasons for perceptual knowledge, we surely don’t need reasons that provide entailing support for our perceptual beliefs to be knowledge. And if we don’t

218

Clayton Littlejohn

need these entailing reasons, we lose the epistemological motivation for accepting McDowell’s metaphysical disjunctivism. The second question is a question about the tenability of the proposal. Is there any reason to think that there are reasons that can do what McDowell wants them to do? Because he thinks knowledge is a standing in the space of reasons in two senses, he needs the thinker’s reasons to do two things. The thinker’s reasons have to ensure that the thinker’s response is appropriate or justified. In so doing, they have to ensure that the thinker is in a position to know. I don’t think that the proposed account is tenable or well motivated. I’ll explain why later in this chapter. Before I do that, I’ll need to explain why McDowell thinks that epistemological considerations support his metaphysical disjunctivist view.

3. The Epistemological Argument for Metaphysical Disjunctivism Our epistemological disjunctivist thinks that in paradigmatic cases of perceptual knowledge, the thinker’s reason for believing p is that she sees that p. Because this reason is the thinker’s reason for believing p, the thinker’s reason provides factive support for her belief. Because it provides such support, the thinker’s reason is supposed to ensure that the thinker is in a position to know p. For this proposal to work, McDowell thinks that we need to embrace a relational and representational view of experience, one that’s incompatible with the traditional conception of experience that characterises experience as the factor that’s common to perception and indistinguishable hallucination. Otherwise, it is hard to see how perceptual experience could provide us with reasons that provide the required rational support. In combination, McDowell’s two commitments concerning knowledge imply that a thinker doesn’t have the right to believe what she doesn’t know and could only have this right if her beliefs were based on reasons the possession of which entailed that the thinker was in a position to know (and not merely in a position to justifiably believe where justification is understood as compatible with error). Most epistemologists would dismiss this kind of infallibilism out of hand. One reason most epistemologists reject this is that they see in McDowell’s infallibilism an impossibly easy route to scepticism: The strongest view one could take regarding the truth connection is that taken by Descartes. The Cartesian view is that justification logically entails truth. To put it schematically: It is a conceptual truth that, if conditions C justify belief B for subject S, then C logically entails that B is true. . . . The legacy of the Cartesian view is scepticism. Descartes demonstrated this in the first meditation that no such

Neither/Nor

219

connection is forthcoming. . . . Given any plausible specification of C for any S, it will always be logically consistent to suppose that not B. That is what the evil demon argument shows. Where, e.g., C comprises facts about sensory data, and where B is a belief about the truth of some empirical proposition, it is always logically possible that the evil demon has arranged for C to obtain where B is false. (Cohen 1984: 281) McDowell knows why this view is so gripping. People in the grips of it think that our best reasons leave it open whether our beliefs about the external world are true. Recall Conee’s argument for this kind of fallibilism: Suppose you have the belief that someone is speaking. You infer this from your justified belief that Mr. Jones is speaking. Thus, your external world belief that someone is speaking is a belief for which you have an entailing justification, your justified belief that Jones is speaking. However, it is quite plausible that your belief that Jones is speaking must itself be justified in order to justify any other belief. . . . When we consider candidate justifications for entailing justifiers like the belief that Jones is speaking, it becomes plain that at some point there is always a proposition that is justified without being entailed by its justification. In the present instance, the non-entailing justifier may well be your justification for the belief that Jones is speaking. This belief may be justified by the experience of its seeming to you that you hear what you seem to recall to be the sound of Jones’ voice. This experience does not necessitate that Jones, or anyone else, is speaking. But it may be all that you have, and all that you need, in favor of the belief that Jones is speaking. Exactly how this justification works is another matter.  .  . . [I]n any plausible view, at some point in the justification of each external world belief that is justified, there is justification without entailment. When this further assumption is added to the assumption that the entailment account is correct, we have a valid argument for the conclusion that no external world belief is well enough justified to be known. . . . The entailment claim is the argument’s least plausible assumption. So, if the skeptical conclusion is to be avoided, then the entailment account of the truth connection is the best candidate for rejection. (Conee 2004: 245)9 Operating on the assumption that normative standing supervenes upon a thinker’s reasons, they conclude that since our best reasons are compatible with the falsity of most of our beliefs, the right to hold these beliefs cannot turn on whether these beliefs are correct. As such, knowledge cannot be a standing in the space of reasons in the first or the second sense.

220

Clayton Littlejohn

McDowell thinks there are two mistakes in this reasoning. The first mistake is that of thinking that these fallible reasons might be adequate for positive standing. If these reasons really did leave it open whether our beliefs about the external world were correct, they couldn’t give us the right to believe what we do. (He rejects the hybrid view of knowledge that these philosophers assume is correct.) The second mistake is that of thinking that a suitable reconstruction of our epistemic situation and a correct description of our reasons for believing what we do could be things that ‘leave it open’ whether, say, there really is gin in this glass and smoke in the air. This leaves out facts like the fact that we see that there’s smoke in the air or that there is gin in the glass. They’re replaced by something else (e.g., experiences, sense data, facts about the inner, false propositions, etc.). This brings us to the second sense in which knowledge is for McDowell a standing in the space of reasons. A belief’s being knowledge is determined entirely by the thinker’s reasons. If a thinker’s reasons for believing p are sufficient in the normative sense (i.e., they make it the case that the thinker justifiably believes what she does), they must put the thinker in a position to know that p is true. Knowledge stands on supporting reasons that a thinker couldn’t have had unless they, too, were in a position to know. This seems to follow from something that McDowell and his critics mentioned earlier agree on, which is that normative statuses are determined by the reasons that the thinker has in her possession. They disagree about what a belief has to be like to attain this status (i.e., McDowell thinks that the belief has to constitute knowledge, and his critics cash out propriety in different terms), but they seem to agree that whatever status is, it is something a belief has by virtue of the reasons that a thinker has in her possession. Critics of this second idea are targets for McDowell’s criticism because they think of knowledge as a kind of hybrid. On this hybrid view of knowledge, having sufficient or adequate reason and having normative standing is one thing. Having an accurate belief and meeting whatever further conditions must be met for a belief to be knowledge is something further, something that a thinker might lack even if her reasons are perfectly adequate. McDowell rejects this hybrid view. He thinks that it suffers from a kind of incoherence: In the hybrid conception [of knowledge that allows that two subjects might both believe for adequate reasons but differ in what they know], a satisfactory standing in the space of reasons is only part of what knowledge is; truth is an extra requirement. So two subjects can be alike in respect of their satisfactoriness of their standing in the space of reasons, because only in her case is what she takes to be actually so. But if its being so is external to her operations in the space of reasons, how can it not be outside the reach of her rational

Neither/Nor

221

powers? And if it is outside the reach of her rational powers, how can its being so be the crucial element in the intelligible conception of her knowing that it is so? (1998: 403) If knowledge is then taken to be an epistemic standing, we fail to do justice to an internalist insight: [O]ne’s epistemic standing . . . cannot intelligibly be constituted, even in part, by matters blankly external to how it is with one subjectively. For how could such matters be other than beyond one’s ken? And how could matters beyond one’s ken many any difference to one’s epistemic standing? (1998: 390) If we turn the rhetorical questions into premises, we get our argument for the disjunctive conception of experience or appearance: The Epistemological Argument for Metaphysical Disjunctivism P1. If a complete description of your mental profile is neutral on whether p (where ‘p’ throughout will be a promising candidate for being an object of perceptual knowledge), the fact that p obtains or doesn’t obtain cannot be part of your perspective. P2. If the fact that p obtains or doesn’t obtain cannot be part of your perspective, the fact that p obtains or doesn’t obtain cannot be among the reasons that contribute to your epistemic standing. C1. So, if a complete description of your mental states and events is neutral on whether p, the fact that p obtaining or failing to cannot be among the reasons that make a difference to your epistemic standing with respect to p. P3. Knowledge is itself an epistemic standing. C2. So, if a complete description of your mental states and events is neutral on whether p, the fact that p obtaining or failing to cannot make a difference to what you know about p. P4. According to the traditional conception of experience, a complete description of your mental states and events is neutral on whether p. C3. So, if the traditional view of experience were correct, it implies that the fact that p obtaining or failing to cannot make a difference to what you know about p. P5. If the fact that p obtaining or failing to cannot make a difference to what you know about p, you cannot know whether p. C4. So, if the traditional view of experience were correct, you cannot know whether p.10

222

Clayton Littlejohn

Someone who accepts the hybrid conception of knowledge will want to decouple the facts that can contribute to the subject’s epistemic standing by being reasons from the things that determine whether a subject’s beliefs constitute knowledge, but they can do this only by denying (P3). What would be wrong with denying (P3)? McDowell offers us these observations: What I claim yields no satisfactory conception of knowledge is the thought that truth—worthiness of endorsement, which is expressed by the attributor’s undertaking the relevant commitment herself—is needed as an extra condition, over and above whatever entitlement can be attributed to the candidate knower. This thought reflects the idea that entitlement incompatible with falsehood in what one is entitled to cannot be had. Thinking on these lines, one will suppose that if one does no more than attribute whatever entitlement one can, one leaves it open that what the putative knower putatively knows is not even true. In that case, she certainly does not know it. Given this interiorizing of entitlement, the fact that the putative knower’s commitment is to something that is in fact true . . . looks accidental in relation to the putative knower’s entitlement. (2002: 102–103) We start from the idea that one essential difference between mere true belief and knowledge is that the latter cannot be a relation between a thinker and a fact where it is at all accidental that the thinker is right in taking the fact to be a fact (Unger 1968). Suppose that the difference between being in the good case and the bad didn’t tell us what a subject’s reasons could be for believing p. If so, given the thinker’s best reasons for believing p, there is some element of luck and some degree of accidentality that distinguishes the subject in the good case from the bad. But this makes it hard to see what’s good about the good case. To be in the good case, the case of knowledge, we’re supposed to stand in some relation to a fact that shows that it is not at all an accident that we got things right. As we’ve just described the difference between the good case and the bad, it is to some extent lucky or accidental that the thinker who was moved by such and such set of reasons ended up with a true belief as opposed to a false one. So perhaps we have no right to describe the good case as good. It’s good to the extent that a thinker has a true belief, but perhaps it isn’t good in the intended sense. We cannot coherently describe it as a case of knowledge if good fortune is what brings the thinker and the fact together. It’s reasonably clear why McDowell thinks that the hybrid view is problematic. On the hybrid view, a belief needs the support of reasons to be knowledge, but the support it provides leaves it open whether the target belief is knowledge. It seems incoherent, then, to say that positive standing is both ensured by the reasons and requires that the target belief constitutes knowledge.

Neither/Nor

223

While most who defend the hybrid view of knowledge would respond by rejecting the idea that knowledge is a standing in the space of reasons in both of the senses discussed earlier, I would like to consider a slightly different response. I reject the hybrid view of knowledge because, like McDowell, I don’t think that a belief’s normative standing and its status as knowledge are separable. I disagree with McDowell because I deny that knowledge is a standing in the space of reasons in McDowell’s second sense. I reject the idea that the thinker’s reasons determine whether a thinker’s belief attains positive normative standing and that the reasons determine whether a thinker’s belief is knowledge. So, while I agree with McDowell that a belief attains positive normative standing iff it is known, neither positive normative standing nor knowledge is determined entirely by the thinker’s reasons. To see that there is room here for rejecting both the hybrid view and McDowell’s alternative, consider a view that combines the idea that knowledge is normative for belief with Williamson’s (2000) E=K. On this view, a belief’s positive normative standing turns on whether it is knowledge. In every case where a thinker knows, the thinker will have evidence or reasons that aren’t available to thinkers who do not know, however similar their situation and perspective might be. Where these views differ is that the knowledge-first view I’m envisaging says that non-inferential beliefs constitute knowledge without the support of independently possessed reasons. McDowell needs to give us some further argument to rule this option out. If this option is left standing, it doesn’t seem we need epistemological disjunctivism to avoid external world scepticism. And if we don’t need epistemological disjunctivism to avoid scepticism, we don’t need metaphysical disjunctivism to explain how we can come to have the awesome reasons that the epistemological disjunctivist claims provides us with our perceptual knowledge. As I see it, there are two ways that McDowell might try to rule out this knowledge-first alternative. He might appeal to some conceptual truth about knowledge or some conceptual truth about positive normative standing. If he were to go the first route, he might try to show that nothing could be a state of knowledge unless it was adequately supported by independently possessed reasons. If he were to go the second route, he might try to show that nothing could be a rationally evaluable response that attains positive standing unless the thinker’s reasons guarantee that it can attain that standing. We might see some signs of the first rationale in McDowell’s critique of the hybrid view. There is supposed to be some instability in that view because it seems to suggest that the conditions that help to turn a belief into knowledge might be conditions that aren’t ‘internal’ to the subject’s perspective on the world. Why is this problematic? It appears that it’s problematic because this suggests that there is some kind of good luck or fortune that helps to turn our beliefs into knowledge, but it’s hard to see how that status as a knower is one that we can come to have through

224

Clayton Littlejohn

good fortune. That seems right, but notice that this rationale works only if the thinker’s reasons play an essential role in the anti-luck condition on knowledge. I don’t see why they should. If someone has a safe basis for a belief, a complete description of that basis might not entail that the target proposition is true, but it seems strange to me to say that the thinker was lucky to get things right when basing her belief on a safe basis.11 Rather than explore this route further in this chapter, I will just focus on the idea that rationally evaluable responses can only attain positive normative status if supported by the thinker’s reasons. This assumption about normative status and the thinker’s reasons gives us the ancillary argument we need to rule out the knowledge-first alternative that I mentioned earlier: 1. Belief is a rationally evaluable response. 2. A rationally evaluable response can attain positive normative standing iff the thinker’s reasons must provide sufficient support (i.e., support that entails that the response can meet the relevant norms). 3. Belief can attain positive normative standing iff it constitutes knowledge. C. So, belief can attain positive normative status iff the thinker’s reasons guarantee that the thinker is in a position to know the target proposition. The crucial premise in this argument is (2). With it in place, it’s easy to see why someone would think that the knowledge-first position I sketched as an alternative to McDowell’s is mistaken. Without it, it’s hard to see how the considerations that McDowell appeals to could force us to abandon this view. In the next section, I shall argue that perceptual experience will not provide us with the reasons that the epistemological disjunctivist claims we need to have perceptual knowledge. In the section that follows, I shall argue against the crucial premise of this argument. It cannot be true in general that positive status requires support from the thinker’s reasons. Thus, there are no sceptical costs to be paid by denying epistemological disjunctivism.

4. The Visual Basis of Knowledge According to the epistemological disjunctivist, the perceptual beliefs that the thinker forms in good cases are supported by reasons. Consider three candidate facts that could be the thinker’s reason for believing p: (a) That it appears that p. (b) That p. (c) That I see that p.

Neither/Nor

225

It’s clear that (a) wouldn’t serve McDowell’s purposes. It’s too weak. It also seems to me to be a strange thing to propose if you think that perception is a source of reasons in the relevant sense. It seems that if all we get from perception when we see Agnes curled up on the couch is something like (a) is something that introspection provides us about perception. Perception itself seems to give us nothing further. McDowell rejects the idea that (b) could be the basis because he thinks that the transition from p to p couldn’t capture the thinker’s take on why they were convinced that p. This means that we’re left with (c), something that entails (b) and thus seems like a good candidate for playing the rational role that McDowell requires. The proposal differs from the knowledge-first approach sketched earlier because it identifies a reason that could be the thinker’s reason for forming her non-inferential perceptual beliefs. The fact that the thinker sees that p is an input to a process that takes non-knowledge in and spits perceptual knowledge out. (If E=K is correct, there can be no such process that takes reasons or evidence as input.) The possession of this reason and the ability to base beliefs on it cannot require that the thinker knows or believes (c), so the possession of (c) has to be understood in terms of the operation of perceptual capacities. Such capacities have to put us in a position to ϕ for the reason that we see that p without believing p or that we see that p. This acquisition of a reason and being able to treat it as a reason has to be understood in terms of visual relations between the thinker and the things this thinker perceives. To introduce a bit of terminology, visualism is the view that a thinker seeing that p is a visual affair, a matter of a thinker standing in some perceptual relation to the fact that p.12 Visualism isn’t my preferred view. I do not think that we stand in perceptual relations to facts. Even if we did, I do not think that there is any interesting epistemic status that supervenes upon the perceptual relations that we do stand in. I think that part of visualism’s spurious appeal is that the kind of report I used in stating the view is mistakenly assumed to report some visual relation that holds between a perceiver and that which the perceiver perceives. I think this is a mistake about language that helps to explain a mistake about perceptual consciousness. In keeping with epistemicism, I think that reports of the form ‘so and so sees that p’ report a thinker’s knowledge, not perceptual relations that hold between a thinker and the things in her surroundings.13 If epistemicism is correct, it seems that the epistemological disjunctivist hasn’t given us any account of how we acquire perceptual knowledge. At the very least, it hasn’t identified the reason that could serve as an input into a process that takes non-knowledge in and spits non-inferential knowledge out. And since I’m operating on the assumption that this is what epistemological disjunctivism purports to do, I think that the truth of epistemicism shows that epistemological disjunctivism on its current formulation fails. The view would need to be reformulated, and I doubt that any such reformulation will provide the epistemological disjunctivist with what they

226

Clayton Littlejohn

need because, as we’ll see, we have non-linguistic reasons to think that no visual relation between an individual and the things she sees could provide the thinker with reasons that would ensure that this individual was in a position to have a justified belief or knowledge. It’s all to the good that (c) entails (b). This means that (c) provides the factive support the epistemological disjunctivist thinks we need to know (b). One worry about visualism, though, is that (c) doesn’t seem to entail this: (a*) It visually appears that p. Consider, for example, ‘I see that the implications of consequentialism are at odds with many of our intuitive moral judgements’. It surely doesn’t follow from this that it visually appears that our intuitive judgements clash with this moral theory. Since (c) doesn’t seem to entail (a), I think we have some good reason to think that (c) isn’t really about perceptual relations between a thinker and the things that the thinker perceives. It certainly doesn’t identify the light in which the thinker comes to hold the perceptual beliefs that she does. If it did, it would contain the kind of information that (a*) does, information about how things perceptually appear and the things that would be made perceptually manifest when the thinker sees how things are.14 If you think that all uses of ‘sees that’ are univocal, this might seem troubling. The visualists might say that there are purely perceptual uses of ‘sees that’, uses that differ from those that don’t sustain the entailment from (c) to (a*). I still worry that the account will struggle to explain the fact that  the alleged reports of perceptual relations aren’t extensional. Many uncontroversial perceptual reports are extensional. If I see Agnes and Agnes was the one who mixed the drinks, it just follows that I see the one who mixed the drinks even if I don’t think any drinks have been mixed. Epistemicism does a nice job handling linguistic data that alternative accounts struggle with, such as explaining why the ‘sees that’ reports are not extensional. Suppose Agnes sees Agatha and Agnes can see that this is Agatha. Suppose Agatha is the next person to order a martini. Agnes sees the next person to order a martini but it doesn’t follow that she can see that this is the next person to order a martini. Similarly, if Agatha’s drink is a martini and both Agnes and Jack see it, it’s possible that Agnes sees that it’s a martini even if Jack cannot because, say, he is a lapdog, a child, or just doesn’t spend much time in bars. What we fill in for p in a report of ‘S sees that p’ depends, in part, upon the conceptual resources and abilities of the perceiver in a way that certain forms of awareness might not (e.g., the kind of simple seeing that’s common to Agnes’s and Jack’s seeing Agatha’s drink). If visualism is going to account for the fact that these reports are not extensional, they might try to offer us an account of ‘sees that’ that can account for this, one that appeals to the exercise of concepts:

Neither/Nor

227

Visualism: When used perceptually, the report ‘S sees that a is F’ picks out a visual relation between S and the fact that a is F, one that ensures that the thinker is in a position to know that p perceptually. It does this, in part, by exercising the conceptual capacities necessary for representing something as being F.15 What should we make of this account? There is some scepticism about the idea that facts are the right kinds of things to be the objects of perceptual awareness. Since it seems that visualism requires that facts be the objects of visual awareness, this seems like a serious problem for visualism and epistemological disjunctivism. Perhaps this scepticism is misplaced, though. Fish seems to think so: the basic constituents of presentational character are not objects and properties per se. Instead, . . . [we should say that] to see a property is to see it as inhering in some object or other. When we combine this with the thought that we never simply see an object but always see it by seeing one or more of the properties it possesses, it suggests a general theory of what the fundamental constituents of presentational character are. Take a given tract of the environment that I visually perceive. This tract of the environment contains an array of objects and surfaces that possess a vast assortment of properties. Given that, when we see an object in our environment, we don’t ‘just’ see that object, when we see an object what we see must be, at a minimum, that object’s bearing a property. Likewise, given that we always see a property as inhering in an object, when we see a property what we see must be, at a minimum, an object’s bearing that property. Taken together, these considerations suggest that the basic units that feature in presentational character are not properties and objects simpliciter, but rather object—property couples. (2009: 52)16 If we think of facts as abstract, the scepticism about the very idea of perception relating us to facts seems well placed. If, however, we think of facts as Fish’s facts, perhaps the scepticism is unwarranted. Let’s suppose for the time being that facts are Fish’s facts and consider whether perceptual relations to them could play the role that epistemological disjunctivism requires. Could we stand in purely perceptual relations to the fact that, say, this is a tomato or that this tomato is red and thereby be in a position to know such things? Visualism faces a serious objection, one that shows the advantages of epistemicism and, in turn, a knowledge-first approach to perceptual knowledge that tells us in advance that we shouldn’t hunt for a reason that provides us with the basis for our perceptual knowledge. The objection concerns the relationship between seeing that p and being in a position to know that p on the assumption that the former is truly a visual relation between a thinker and some Fish fact. Suppose you see a tomato under ideal viewing

228

Clayton Littlejohn

conditions. You’re in a museum. It’s on a plinth. The light is directed on it just like so. There’s an X for you to stand on and you’re standing on it with your head directed in the right way. There’s a sign that has ‘Tomato’ written on it in a sans serif font attached to the plinth. The tomato rests on a velvet pillow. Fill in the details so that it’s a clear case of simple seeing (i.e., seeing a tomato), epistemic seeing (i.e., seeing that it is a tomato), and knowing (i.e., knowing that the thing on the plinth is a tomato). According to visualism, its being a case of epistemic seeing is partially down to the fact that the understanding is operative and the tomato is represented as being one. The understanding is operative in the experience itself, not just in the judgement. It is also a matter of simple seeing, seeing a tomato. Had the understanding not been operative, the case would have been a case of simple seeing without epistemic seeing, but that’s not what’s happened. Because the tomato seen was represented as a tomato, it seems like as good a case of seeing that something is a tomato as anyone could hope for. And, as luck would have it, it’s a case of being in a position to know. Imagine a series of cases now where the viewer stands slightly farther away from the tomato so that at the last case in the series, a viewer sees a small speck from far off. And let’s suppose that each case has this much in common. The viewer sees a tomato, and the understanding that is operative in our first case is still operative in these latter cases in the same way so that the viewer believes on the basis of her experience that the thing she sees is a tomato and her experience represents the thing as a tomato. Just as we sometimes judge that something is an F from a great distance (perhaps too great a distance to judge reliably), there is no reason to think that the understanding that’s operative in experience (assuming with McDowell, that it is) wouldn’t sometimes represent the thing seen as being a tomato. Now, we have to ask two questions about this case of seeing a tomato at a distance. If the thinker believes that the tomato is a tomato, is this correct belief knowledge? If the thinker’s experience represents it as being a tomato, is this a case of epistemic seeing? When we view things from too far off, the judgement that the thing on the plinth is a tomato might be correct, but its correctness should seem too fortuitous for the thinker’s perceptual belief to constitute knowledge. The interesting question is whether this is a case of epistemic seeing. Here is how I see the situation. If we imagine our series of cases as described earlier, the structure of such cases will look like this: Great Less Less Less Less Less Less Less On the X Distance Simple Seeing Yes Epistemic No Seeing Position to No Know

Yes No

Yes No

Yes No

Yes No

Yes Yes

Yes Yes

Yes Yes

Yes Yes

No

No

No

No

Yes

Yes

Yes

Yes

Neither/Nor

229

On this way of describing things, epistemic seeing correlates with being in a position to know, but not with simple seeing and the representation of the perceived object as a tomato. If we describe things this way, it seems that epistemicism explains why we shouldn’t describe the cases in which a thinker is standing far away as a case of simple seeing where the perceiver cannot see that the thing is a tomato. They cannot see that it is a tomato because they are not in a position to know. (They are not in a position to know, perhaps, because the basis for the correct judgement isn’t reliable or isn’t safe.) Now compare this visualism. Visualism says that this kind of explanation makes no sense. The things that explain why you’re not in a position to know do not explain why you don’t stand in some visual relation to the things perceived. It would seem that this case gives us a straightforward counterexample to visualism. It would seem that if epistemic seeing were a purely perceptual affair, simple seeing and epistemic seeing should correlate in just the way that they appear not to. Someone could, of course, contest this and say that simple and epistemic seeing do correlate: Great Less Less Less Less Distance Simple Yes Seeing Epistemic Yes Seeing Position to No Know

Less Less

Less

On the X

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

No

No

No

No

Yes

Yes

Yes

Yes

If epistemic seeing, though, is understood as simple seeing in which the concept correctly categorises the object perceived, we vindicate the idea that epistemic seeing is indeed a perceptual affair at the cost of maintaining the connection between epistemic seeing and being in a position to know. On this approach (which I think is counterintuitive), visualism is spared from the objection, but it turns out that seeing that p doesn’t guarantee that the thinker is in a position to know p. In turn, this means that the epistemological disjunctivist’s proposal fails. Even if a thinker’s reason for ϕ-ing is that they see that p, it doesn’t follow that they’re in a position to know p. This case generates a dilemma for the target view. On the one hand, if they were to acknowledge that one of our nine cases were a case of epistemic seeing without knowledge, they would have to acknowledge that it doesn’t follow from the fact that some thinker sees that the thing is a tomato that the thinker is in a position to know that it is one. On the other hand, if they were to deny that this is a case of epistemic seeing, they would need a credible explanation as to why this isn’t a case of

230

Clayton Littlejohn

epistemic seeing. I doubt that the visualists could offer a credible explanation as to why this would be. Why wouldn’t a case of simple seeing in which a tomato is correctly classified as such be a case of epistemic seeing? It is surely a matter of being visually aware of one of Fish’s facts if, say, the nearest case of simple and epistemic seeing is.17 It’s obvious why the first horn is a horn. Our epistemological disjunctivist needs to identify some reason that could ensure that the thinker is in a position to know, and (c) was the only possible candidate for that. As for the second horn, I see no good way for the epistemological disjunctivist to tackle this given a commitment to visualism. They cannot deny that the understanding that is operative in the good case is operative in the bad one. The same conceptual capacities are deployed. It’s hard to see how there could be a difference in the visual relations because we’ve chosen our cases by screening out any case that isn’t a case of simple seeing. If they’re going to try to address the issue, they’ll need to show that any case where the thinker isn’t in a position to know that the thing is a tomato isn’t a case of seeing that it is a tomato without appeal to epistemicism. I see no good way for them to do that. The things that would that explain why we cannot know would seem to be modal considerations that explain why it wouldn’t be appropriate to deploy a concept (e.g., it could too easily be misapplied; it would be applied even if the thing were not a tomato). It doesn’t seem that such points could figure in an explanation of what we see or what visual relations obtain between a viewer and things in her environment.

5. The Thinker’s Reasons It’s hard to resist the pull towards visualism and the idea that we need such a view to explain how perceptual knowledge and justification is possible. It seems that such a view is needed because it seems, to some, that it’s impossible for any belief, inferential or non-inferential, to attain positive standing without being supported by reasons. It isn’t always clear why people believe that all belief exhibits this kind of dependence upon the support from the thinker’s reasons, but it is clear that many people do believe this. If they think that this reflects some important insight into the nature of normativity because, say, they believe that reasons always determine normative standing, the explanation as to why belief needs the support of reasons to attain positive standing might appeal to this more general thesis: Necessitarianism: The normative status of a thinker’s rationally evaluable responses supervenes upon the thinker’s reasons. I’d like to shift our focus slightly in this section to discuss necessitarianism because I think that it’s a commitment to necessitarianism that convinces people that the idea that knowledge is a standing in the space

Neither/Nor

231

of reasons in the first sense (i.e., that a belief is justified iff it is knowledge) turns on whether it can be a standing in the space of reasons in the second sense (i.e., that a thinker’s reasons wholly determine whether a thinker’s belief constitutes knowledge). This necessitarianism seems to play a crucial role in the epistemological argument for metaphysical disjunctivism because it would explain why we should think that the possibility of perceptual knowledge requires that perception could provide us with these reasons that guarantee that any thinker with just these reasons would be in a position to know what we know in the good case.18 Necessitarianism is mistaken. The case of non-inferential knowledge is a case in which a thinker believes something without that belief being based on one of the thinker’s reasons. Since the case of perceptual belief is a contested case, I’ll focus on other cases. Let’s take a step back to try to understand the motivation behind necessitarianism. Necessitarianism is appealing to people who think that normative reasons have to play a kind of guiding role: A familiar and intuitive thought is that normative reasons must be able to guide us. That is what reasons seem to be for. Considerations that cannot guide cannot do what reasons are supposed to do. To put the same point differently, it is the job of a reason to recommend that a person perform a certain act or hold some attitude. If it is to do that job, the relevant person must be able to heed and respond to its recommendation. (Way and Whiting 2016: 214) The appeal of necessitarianism is completely mysterious if we reject this idea. If normative reasons don’t have to be suited to play a guiding role, it’s not at all clear why we would think that there’s any interesting relationship between the reasons that the thinker could be moved by and the normative status of the thinker’s rationally evaluable responses. On the other hand, if you think that normative reasons have to be able to play a guiding role, necessitarianism seems an attractive view.19 Because of this connection between necessitarianism and the idea that normative reasons should guide us, let’s examine this thesis more closely: Guidance: A set of normative reasons can require an agent to ϕ only if the agent is capable of being guided by these reasons in ϕ-ing in the circumstances under which they are required (i.e., they can respond correctly on the basis of the right reasons in a way that shows that they are attuned to their demands).20 In the case of perceptual belief, you might think that a thinker in the good case who sees that, say, the bird on the branch is a parrot either ought

232

Clayton Littlejohn

to believe that it is a parrot or, at the very least, ought not believe that it is a penguin. If these requirements require normative reasons, Guidance tells us that these reasons have to be reasons that the thinker is appropriately sensitive to so that they can be moved by them (and not merely in accordance with them) in such a way that the thinker’s rationally evaluable response is correct or appropriate. This couldn’t happen unless the normative reasons that generated the requirement were accessible to the thinker in such a way that those reasons could have been her reasons, so it seems that Guidance supports the idea that the reasons that matter are the thinker’s reasons understood as the reasons that could be the thinker’s reasons for responding as she does. If necessitarianism is mistaken because there can be cases where pairs of thinkers with the very same reasons are required to different things, the reasons that generate some of these requirements wouldn’t satisfy Guidance. If, however, the reasons that require the agent to ϕ satisfy Guidance, it would seem that the connection between these reasons and the justification or propriety of the relevant response couldn’t be contingent. Thus, it would seem that arguments for Guidance should support necessitarianism and objections to Guidance should undermine our confidence in necessitarianism. One concern with Guidance is that it seems to conflict with intuitive verdicts about cases. To see this, we’ll shift our focus away from cases of perceptual belief to a case of inferential belief. Agnes has a very reliable breathalyser, one that she knows never fails to detect drunkenness when a driver is drunk. In only 5% of the cases in which a driver is sober will her breathalyser falsely say that the sober driver is drunk. She also is aware that 1 driver in 1000 will be drunk. Agnes administers the test and it indicates that the driver just stopped at the checkpoint is drunk. She’s now very confident that the driver is very drunk. She has a credence of .95. That is where Agnes is, and you know how she got there. You know what she knows and how she has responded. Is this how she should have responded? Should she have a different credence? It would seem that Agnes’s options would include this credence along with the mathematically kosher credence (roughly .02) and scores more that would not impress your statistics professor. Here is my concern. Suppose that Agnes ought to have some credence or other. The ones that are not mathematically kosher (i.e., her actual credence and those that are not roughly .02) would not seem to be correct responses to her situation. The available reasons do not support this credence. If all the options but the mathematically kosher credence are incorrect responses to the available reasons, a ‘No Dilemma’ constraint tells us that she ought to have the mathematically correct credence for the simple reason that we have eliminated all the other alternatives to this one on the grounds that they are not correct responses to the available reasons. So,

Neither/Nor

233

we have an argument that this response is mandatory. This response cannot be mandatory, however, because this response is not one that meets an attunement condition (i.e., the subject doesn’t meet the conditions necessary for being guided appropriately by the reasons in the way spelled out in Guidance). Agnes cannot respond in the mathematically kosher way while, say, manifesting her competence if that is understood as her know-how to respond correctly to the reasons. If this final response is not required, some other response must be permitted. But those seem to be ruled out because they fail to meet the condition that says that it’s proper to ϕ only if the available reasons support that. There might be an alternative to Agnes’s unwarrantedly high credence where Agnes suspends, withdraws, or just does not take account of the evidence presented to her. If this, however, is the only permitted option left on the table, it is mandatory. The problem is now that this withdrawing response is also one that would fail to meet the attunement condition. So, according to Guidance, it could not be mandatory. So it would seem that there is no feasible response that involves responding correctly to the normative reasons while basing that response on reasons that provide sufficient support for that response. If, as it seems, Guidance clashes with the intuitive judgement that Agnes shouldn’t be confident and should suspend judgement, Guidance clashes with intuition and doesn’t provide support for necessitarianism. There is a kind of structural defect in views that incorporate the attunement conditions. The introduction of such a condition threatens to screen out the requirements generated by the reasons that screen out the incorrect responses but must leave us at least one permitted response if reasons don’t generate dilemmas. What larger lessons can be drawn from this? We should be sceptical of appeals to intuitions about guidance, for a start. Recall the quotation from the opening. The suggestion was that the normative reasons should satisfy certain conditions to be good normative reasons: they have to be such that they can guide creatures like us. Getting the ‘can’ right here is tricky, but on a very demanding reading of ‘can’, this is a deeply disturbing picture. Shouldn’t the idea be that we should be such that we are guided to do, believe, and feel what the reasons require from us? If we’re fully aware of the reasons but moved in the wrong way, isn’t the fault in us? As it happens, I think that it’s important for the epistemological disjunctivist to reject Guidance. Think about the bad case. According to the epistemological disjunctivist, the thinker’s lacking sufficient reason to believe that things are as they appear means that they cannot justifiably believe p when it appears that p. If they shouldn’t believe this, there is a decisive reason for them not to. Can it guide them? It couldn’t guide them if they don’t know that such a reason applies to them. We have to reject negative introspection and acknowledge that their failure to be in a position to know is something that they’re not in a position to know in

234

Clayton Littlejohn

the bad case. If the reason for them not to believe is decisive in this case (much in the way that the reason not to draw the inference is decisive in the case presented earlier), the winning reason isn’t one that can guide the thinker. Again, we have to ask why we should think that the propriety of a response depends upon whether the thinker bases that response on the sufficient reason there is to respond this way. If the epistemological disjunctivist builds Guidance into their framework, they have to abandon this idea that the thinker in the bad case shouldn’t believe because they have to abandon the idea that the thinker has a decisive reason not to believe what she does not know.

6. Why Is the Good Case Better? Let’s return to this question about the relationship between Normativity of Knowledge and Knowledge-Reasons. There are at least three ways in which things are better in the good case: 1. The good case is better than the bad because the thinker’s reasons for believing p are better in the good than in the bad. 2. The good case is better in the sense that the belief that p is justified. 3. The good case is better in the sense that the thinker knows p. I agree with McDowell that the hybrid view of knowledge is mistaken and that it’s a mistake to say, as most epistemologists do, that (2) is false on the grounds that the reasons we have in the bad case are good enough to show that the beliefs formed in that case are justified. Like McDowell, I think that (2) and (3) are correct. Our disagreement concerns (1). As I would explain (2), (2) is true because (3) is.21 If people insist on using the language of reasons, I would say this: because ‘ought’ implies ‘reason’ and (2) implies that the thinker should not believe p. This reason is decisive. This just follows from (3) and the idea that knowledge is the norm of belief, the idea that knowledge is a standing in the space of reasons in the first sense. Since I reject that all such standings are determined by the thinker’s reasons, I see no need to appeal to (1) to explain (2) or (3). The explanation of (2) goes no farther than an appeal to (3). In stating his epistemological case for metaphysical disjunctivism, McDowell claims that the difference between the good case and the bad cannot be external to the subject’s perspective on things. In a way, the view offered respects that. The thinker who knows is aware of something that the thinker in the bad case is not—to know p is to be aware of the fact that p, and without such knowledge we don’t have such awareness. To rule out this shallow knowledge-first explanation of (2), McDowell would need to show that (2) requires (1). I don’t think that it’s a

Neither/Nor

235

general truth that the propriety of a response requires access to the normative reasons that determine the status of our potential responses. The case of the fallacious inference seems to suggest that a thinker ought to suspend judgement when she’s out of her depth even when she’s in no position to appreciate that this is so. We find something similar happens in the bad case where a thinker hallucinates. The thinker has, according to McDowell, a decisive reason not to believe p. The thinker has no access to the reasons in light of which this is so. So even McDowell has to agree, given his commitment to (2) and (3), that some rationally evaluable responses can be proper or required even when the thinker cannot base her response on the reasons that determine the response’s normative status. That is to say, even McDowell has to deny Guidance. Without some reason to think that belief is different from suspension of judgement in that proper belief always requires being based on reasons that provide sufficient support, there is little reason to think that our commitment to the Normativity of Knowledge should incline us towards Knowledge-Reasons. Since I lack the imagination to invent further arguments for the mistaken view that all justified beliefs are justified because of the support provided by the thinker’s reasons, I should stop here. Neither epistemological disjunctivism nor a suitably matched metaphysical disjunctivism is necessary for accounting for the justification of perceptual belief. Hence the title.22

Notes 1. See McDowell (1998), Pritchard (2012). McDowell’s views have evolved, and the criticisms of epistemological disjunctivism discussed here do not apply straightforwardly to his new views. See McDowell (2009) for details. It is not clear whether his new views concerning the objects of visual awareness sit well with the epistemological assumptions operative in the epistemological arguments for disjunctivism discussed in this chapter. My target is not McDowell, but an idea about the rational role of visual contact with things in our surroundings that is inspired by McDowell (1998). 2. McDowell (1998) thinks that we need both forms of disjunctivism to avoid scepticism, but Pritchard (2012) is agnostic about metaphysical disjunctivism. For a helpful discussion, see Byrne and Logue (2008). 3. For further discussion of the connection, see Schmidt (2018). 4. Although Cohen (1984), Conee (2004), and Conee and Feldman (2008) all find this objectionable. 5. See Littlejohn (2017) and McGinn (2012) for arguments that our perceptual beliefs are not based on the thinker’s reasons. 6. Some philosophers like to say that experience is a source of evidence or a source of reasons. There is a kernel of truth in this idea. Experience is a source of knowledge, so it must be a source of evidence. I don’t think we should also say that experience is a source of evidence in the sense that it provides an evidential basis for knowledge. The process by which we acquire knowledge is the process by which we acquire evidence. These are not two processes, but one process under two descriptions.

236

Clayton Littlejohn

7. Although the idea that cases of perceptual knowledge and/or justified perceptual belief are cases in which there is something that is the thinker’s reason for believing what she does is popular, it is hard to find arguments for this claim. For accounts of perceptual knowledge and/or justified belief where reasons and/or evidence play a central role, see McDowell (1998), Siegel (2010), Ginsborg (2011), Pritchard (2012), Logue (2014), Schellenberg (2011, 2018), Schroeder (forthcoming). For criticism of this idea, see Brewer (2011), McGinn (2012), Travis (2013), French (2016), Littlejohn (2017). 8. In addition to McDowell, see Sutton (2007), Littlejohn (2013), Williamson (forthcoming). 9. Infallibilism is often taken to be a sceptical view, but see Dutant (2016), McDowell (1998), Williamson (2000) for helpful perspective on the (alleged) sceptical consequences of infallibilism. 10. This reconstruction draws on van Cleve’s (2004) exegetical work. 11. For defence of a safety condition as an anti-luck condition, see Luper (1984), Sosa (1999), Williamson (2000). 12. Prominent visualists include McDowell (1998), McGinn (1999), Ginsborg (2011), Turri (2010), and Pritchard (2012). The visualists that I have in mind think the fact that p is seen when we see that p. For a helpful discussion of what it might mean to be visually aware of facts, see Longworth (2018). For an interesting alternative proposal about reasons and the objects of visual awareness, see Logue (2014), Cunningham (2018). 13. Williamson (2000), French (2012). One source of confusion about visualism and epistemicism is that some authors seem to conflate the question, ‘Does perceiving require knowing?’ with a linguistic question about reports or attributions that use ‘sees’ and related verbs. Epistemicism does not tell us that propositional knowledge is necessary for perception. It tells us only that ‘S sees that p’ functions to attribute propositional knowledge. Since this is a claim about language, arguments for visualism and/or against epistemicism that appeal to psychological claims (e.g., perceiving does not require belief or knowledge) miss the point. 14. For further arguments that these reports of epistemic seeing do not report perceptual relations between an individual and the facts, see Travis (2013). 15. I suspect that visualism is largely responsible for the (mistaken) belief that fake barn cases show that it’s possible to have reasons or evidence that aren’t things that we know. For the use of such examples to argue that the thinker’s reasons for ϕ-ing needn’t be things the agent knows see, for example, Hughes (2014), Locke (2015). For responses to the use of such cases to criticise knowledge accounts of the possession of reasons, see Carter (2013), Littlejohn (2014). 16. The objection to a visualist view that takes perceptual contact with Fish’s facts to be sufficient for being in a position to know should apply, inter alia, to views that take the relata of perceptual consciousness to include objects and properties. Even if perceptual consciousness relates us to objects and properties and has some kind of intuitional content because of the exercise of conceptual capacities, the case described in this section suggests that there can be pairs of cases with the same object, same properties, and same intuitional content where the perceptions differ with respect to whether they put the thinker in a position to know that the object in question has the relevant properties. 17. In this argument, the operative idea of being in a position to know is (roughly) the idea that you would know if you were to believe given the way of coming to believe available to you. Someone might try to blunt the force of the objection by adopting a slightly different conception of being in a position to know,

Neither/Nor

237

one that says that a thinker is in a position to know by basing her belief on a reason R when the thinker would come to know by basing her belief on R provided that there are no defeaters. If, as someone might naturally think, there are defeaters here that defeat knowledge if you believe that there is a tomato on the plinth, the visualist can preserve the idea that there is a perceptual relation that’s sufficient for being in a position to know. I don’t think that this proposal will suit the needs of someone like McDowell because it introduces something that is necessary for knowledge that is independent from the reason that you’d be aware of perceptually. This would, in turn, result in a hybrid picture of knowledge according to which the difference between knowing and failing to know would be ‘extrinsic’ to perceptual consciousness (i.e., the presence/absence of the knowledge defeater would not supervene upon the awareness of the reason R). But McDowell rejects this hybrid picture, and the rejection of the hybrid picture is the essential idea in his argument for metaphysical disjunctivism. If we’re allowed to adopt a hybrid view of knowledge where a belief’s status as knowledge doesn’t supervene upon facts about the thinker’s total reasons and the reasons on which her belief is based, we lose all motivation for thinking that the relevant reason has to be something the awareness of which is sufficient for all the things packed into my conception of being in a position to know. Thanks to Joe Milburn for raising the issue. I think it’s an important one, and one that I had overlooked previously. 18. Necessitarianism supports the idea that a perceptual belief is justified only if it is based on a sufficient reason. Suppose Agnes forms the perceptual belief that this is a tomato on the basis of her experience. Suppose she comes to know that this is a tomato. If Agnes’s belief was justified without being supported by reasons to believe this proposition, there wouldn’t be a difference in the support she had for believing this or for believing that it is not a tomato. To account for the fact that the cases where Agnes has justification to believe it is a tomato are never cases where she has justification to believe that it is not, the necessitarian wants to say that her reasons support one without supporting the other. It doesn’t seem they can account for this without insisting that the justification to believe one thing rather than the other requires the possession of reasons that provide sufficient support for one and not the other. 19. Joe Milburn wanted to know what connection, if any, there was between necessitarianism and Guidance. Might the necessitarian try to motivate their view without any appeal to Guidance? I think so. As Milburn reminded me, they might appeal to some sort of reasons-first view and/or the arguments for such a view. On this view, normative reasons wholly determine the normative status of any rationally evaluable response. I have no objections to the reasons-first idea, per se, but we should note that the reasons-first view might not be well suited for the epistemological disjunctivist. Many reasons-first philosophers (e.g., Kiesewetter 2017; Lord 2018) think that the reasons that determine what we ought to believe, feel, and do are always accessible to the agent. (Part of their reason for thinking this is that they accept Guidance, but they could reject Guidance and accept the accessibility constraint.) But, as I’ve argued here, the epistemological disjunctivist has to reject this idea since they have to acknowledge that there are decisive reasons for us not to believe that are not accessible to the agent. If reasons-first philosophers were to relax the access requirement on normative reasons, their views might seem to be compatible with epistemological disjunctivism; but then it’s not clear why in this framework we would need to have a rational basis for perceptual beliefs constituted by reasons of the kind that McDowell and Pritchard think are necessary for perceptual knowledge.

238

Clayton Littlejohn

20. Kiesewetter (2017) and Lord (2018) appeal to the idea that normative reasons are supposed to be able to guide us to objections to the idea that rationality consists in responding correctly to reasons. One of the aims of Littlejohn (2012) was to argue that there are no norms that could provide us with such reasons since, in principle, there is no interesting application condition that we could not in principle be misled about. If all potential application conditions are things that we can rationally but mistakenly believe to obtain (or rationally but mistakenly believe not to obtain), it would seem that Guidance would have to be mistaken. For similar sceptical arguments against Guidance, see Srinivasan (2015). 21. In Littlejohn (2012), I also argued that something close to (1) explains why the good case is better. There I argued that the good case is better because the bad case is worse. It is worse because that is a case where there is a decisive reason not to believe, one that is provided by the fact that the thinker violates a norm. (There is no such reason not to believe in the good case because there is no norm violated in that case.) But the difference in reasons not to believe differs from the difference in the quality of the reasons to believe. I think the right explanation will accentuate the negative, not the positive, and focus on the presence/absence of reasons not to believe rather than the presence/ absence of reasons to believe. This is not an explanation that is available to the disjunctivist. 22. I would like to thank Casey Doyle and Joe Milburn for helpful feedback on a previous draft. I would also like to thank Maria Alvarez, Bill Brewer, Adam Carter, Christina Dietz, Julien Dutant, Craig French, John Hawthorne, Nick Hughes, Dustin Locke, Guy Longworth, Errol Lord, Susanne Mantel, David Papineau, Susanna Schellenberg, Eva Schmidt, Matthew Soteriou, and Charles Travis for discussing these issues with me.

References Brewer, Bill, 2011, Perception and Its Objects. Oxford: Oxford University Press. Byrne, Alex and Logue, Heather, 2008, Either/or, in Adrian Haddock and Fiona Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press. Carter, Adam, 2013, A problem for Pritchard’s anti-luck virtue epistemology, Erkenntnis 78: 253–275. Cohen, Stewart, 1984, Justification and truth, Philosophical Studies 46: 279–295. Conee, Earl, 2004, The truth connection, in E. Conee and R. Feldman, Evidentialism, Oxford University Press, Oxford, pp.242-254. ———, 2008, Evidence, in Quentin Smith, ed. Epistemology: New Essays. Oxford: Oxford University Press. ——— and Feldman, Richard, 2004, Evidentialism. Oxford: Oxford University Press. Cunningham, J.J., 2018, Are perceptual reasons the objects of perception? in Johan Gersel, Raysums Jensen, Morton Thaning and Søren Overgaard, eds. In the Light of Experience. Oxford: Oxford University Press. Dutant, Julien, 2016, How to be an infallibilist, Philosophical Issues 26: 148–171. Fish, William, 2009, Perception, Hallucination, and Illusion. Oxford: Oxford University Press. French, Craig, 2012, Does propositional seeing entail propositional knowledge? Theoria 78: 115–127.

Neither/Nor

239

———, 2016, The formulation of epistemological disjunctivism, Philosophy and Phenomenological Research 92: 86–104. Ginsborg, Hannah, 2011, Perception, generality, and reasons, in Andrew Reisner and Asbjørn Steglich-Petersen, eds. Reasons for Belief. Cambridge: Cambridge University Press. Hughes, Nick, 2014, Is knowledge the ability to ϕ for the reason that p? Episteme 11: 457–462. Kiesewetter, Benjamin, 2017, The Normativity of Rationality. Oxford: Oxford University Press. Littlejohn, Clayton, 2012, Justification and the Truth-Connection. Cambridge: Cambridge University Press. ———, 2013, The Russellian retreat, Proceedings of the Aristotelian Society 113: 293–320. ———, 2014, Fake barns and false dilemmas, Episteme 11: 369–389. ———, 2017, How and why knowledge is first, in J. Adam Carter, Emma Gordon and Benjamin Jarvis, eds. Knowledge First. Oxford: Oxford University Press. Locke, Don, 2015, Knowledge, explanation, and motivating reasons, American Philosophical Quarterly 52: 215–232. Logue, Heather, 2014, Experiential content and naive realism: A reconciliation, in Berit Brogaard, ed. Does Perception Have Content? Oxford: Oxford University Press. Longworth, Guy, 2018, Surveying the facts, in John Collins and Tamara Dobler, eds. The Philosophy of Charles Travis: Language, Thought, and Perception. Oxford: Oxford University Press. Lord, Errol, 2018, The Importance of Being Rational. Oxford: Oxford University Press. Luper-Foy, Steven, 1984, The epistemic predicament: Knowledge, Nozickian tracking, and skepticism, Australasian Journal of Philosophy 62: 26–48. McDowell, John, 1998, Meaning, Knowledge, and Reality. Cambridge, MA: Harvard University Press. ———, 2002, Knowledge and the internal revisited, Philosophy and Phenomenological Research 64: 97–105. ———, 2009, Having the World in View: Essays on Kant, Hegel, and Sellars. Cambridge, MA: Harvard University Press. McGinn, Colin, 1999, The Concept of knowledge, in his Knowledge and Reality: Selected Essays. Oxford: Oxford University Press. McGinn, Marie, 2012, Non-inferential knowledge, Proceedings of the Aristotelian Society 112: 1–28. Pritchard, Duncan, 2012, Epistemological Disjunctivism. Oxford: Oxford University Press. Schellenberg, Susanna, 2011, Perceptual content defended, Nous 45: 714–750. ———, 2018, The Unity of Perception: Content, Consciousness, Evidence. Oxford: Oxford University Press. Schmidt, Eva, 2018, A dilemma for epistemological disjunctivism, in Robert French and John Smythies, eds. Direct Versus Indirect Realism: A Neurophilosophical Debate on Consciousness. London: Elsevier, pp. 141–162. Schroeder, Mark, 2015, Knowledge is belief for sufficient (objective and subjective) reason, in John Hawthorne and Tamar Gendler, eds. Oxford Studies in Epistemology (vol. 5). Oxford: Oxford University Press, pp. 226–252.

240

Clayton Littlejohn

Siegel, Susanna, 2010, The Contents of Visual Experience. Oxford: Oxford University Press. Sosa, Ernest, 1999, How must knowledge be modally related to what is known? Philosophical Topics 26: 373–384. Srinivasan, Amia, 2015, Normativity without Cartesian privilege, Philosophical Issues 25: 273–299. Sutton, John, 2007, Without Justification. Cambridge, MA: MIT Press. Travis, Charles, 2013, Perception: Essays After Frege. Oxford: Oxford University Press. Turri, John, 2010, Does perceiving entail knowing? Theoria 76: 197–206. Unger, Peter, 1968, An analysis of factual knowledge, Journal of Philosophy 65(6): 157–170. Van Cleve, James, 2004, Externalism and disjunctivism, in R. Schantz, ed. The Externalist Challenge. Berlin: De Gruyter, pp. 481–495. Way, Jonathan and Whiting, Daniel, 2016, Reasons and guidance (or, surprise parties and ice cream), Analytic Philosophy 57: 214–235. Williamson, Timothy, 2000, Knowledge and Its Limits. Oxford: Oxford University Press. ———, forthcoming, Justifications, excuses, and sceptical scenarios, in Fabian Dorsch and Julien Dutant, eds. The New Evil Demon. Oxford: Oxford University Press.

12 Disjunctivism and Credence Ram Neta

1. Specifying the Disjunctivist Thesis of Interest Here The term “disjunctivism” has been used to denote a variety of positions in epistemology and the metaphysics of perception, mind, and action. In this chapter, I will use the term to denote the epistemological view that John McDowell and Duncan Pritchard have recently defended: a view according to which an agent’s successful perception of a mindindependent object can provide her with a kind of epistemic ground for belief concerning that object that she cannot possess unless the belief is true. For future reference, I will give this view a label and a canonical formulation (I use the word “basis” in the canonical formulation, but I take it to be equivalent to “ground”): BETTER BASIS: Successful perception of a mind-independent object can provide a better basis for belief about that object than can an indistinguishable illusion or hallucination. Although disjunctivists like McDowell and Pritchard defend BETTER BASIS, so too do self-professed non-disjunctivists like Timothy Williamson and Susanna Schellenberg. I will therefore refer to proponents of this thesis not as “disjunctivists”, but rather as proponents of BETTER BASIS. We could say, following Pritchard, that the epistemic ground in question is “factive” (2013), so long as this is not taken to imply that the epistemic ground has the metaphysical nature of a fact: describing the epistemic ground in question as “factive” is meant only to signal that a necessary condition of any agent’s having the epistemic ground is that the belief for which it serves as a ground is true. Thus, “factive” denotes a relation between a particular epistemic ground, on the one hand, and a belief for which it serves as a ground, on the other: it is a relation that obtains only if it is impossible for any agent to possess that epistemic ground unless that belief is true. If we individuate epistemic grounds in such a way that the very same epistemic ground can serve as grounds for

242

Ram Neta

many different beliefs, then that ground may be factive with respect to some of those beliefs and not factive with respect to others. As long as we keep this in mind, we can express the disjunctivist thesis of interest to us by saying: successful perception of a mind-independent object can at least sometimes provide the perceiver with factive grounds for belief concerning that object. In contrast to Pritchard, McDowell expresses his version of BETTER BASIS by speaking not of factive grounds but rather of indefeasible grounds.1 But what does “indefeasible” amount to here? McDowell cannot mean that a belief based on such grounds cannot eventually come to be defeated by the acquisition of additional evidence—any belief could be so defeated if the acquisition of that additional evidence necessitates a change in the grounds on which the belief is based (and some of the things that McDowell says grant as much).2 So McDowell must instead mean that a belief based on such grounds cannot come to be defeated by the acquisition of additional evidence so long as the belief remains based on those same grounds. How stringent a condition this is depends upon how finely we individuate epistemic grounds. Consider, for instance, Timothy Williamson’s example of successful perception that supplies defeasible grounds for belief: I see one red and one black ball put into an otherwise empty bag [call this “e”]. . . . Now suppose that on the first ten thousand draws a red ball is drawn each time, a contingency which my evidence does not rule out in advance, since its evidential probability is non-zero. But when I have seen it happen, I will rationally come to doubt e; I will falsely suspect that the ball only looked black by a trick of the light. (Williamson 2000: 219) Williamson takes this example to show that the epistemic grounds supplied by successful perception are always defeasible: the case of seeing one red and one black ball being put into the bag is taken to be representative of cases in which we gain knowledge by means of successful perception. But McDowell could claim that Williamson’s case can be spelled out in either of two ways. Either (a) the perceiver continues to see the red and black ball in the empty bag (this would be the case if the bag were transparent), or (b) the perceiver does not continue to see the red and black ball in the empty bag, but only, at most, remembers having seen them. In case (a), the perceiver could not rationally come to conclude that there is no red ball in the bag: instead, she would wonder why the draws from the bag constantly reach for the black ball instead of the red ball. In case (b), the perceiver would rationally come to conclude that there is no red ball in the bag; but that is because her grounds for believing that there is such a ball in the bag have changed—she originally believed it because she saw the ball in the bag, and now she believes it only because

Disjunctivism and Credence

243

she recalls having seen it. The defeasibility of the second ground does not imply the defeasibility of the first. On neither interpretation does Williamson’s example challenge McDowell’s indefeasibility thesis, but these interpretations are exhaustive. BETTER BASIS is a thesis with which Williamson himself can agree. For Williamson, our evidence includes only those facts that we know, and successful perception is a way of knowing facts about the mindindependent objects that we perceive. If an agent’s evidence includes all those facts that can serve as her epistemic grounds for belief, then Williamson’s view is not simply consistent with BETTER BASIS, but entails it. Of course, Williamson would not call himself a disjunctivist, but his reasons for refusing the label concern his objection to explaining nonfactive states like belief or seeming in terms of a disjunction of a factive state and something else. Such explanatory attempts are completely independent of BETTER BASIS, which says nothing about how non-factive states are to be explained, either metaphysically or conceptually. BETTER BASIS is also accepted by Susanna Schellenberg. Schellenberg describes a pair of examples that make for the clearest statement of BETTER BASIS, though she speaks not of bases for belief, but rather of evidence (which she takes to constitute a basis for belief): “Percy, the perceiver, accurately perceives a white cup on a desk. Hallie, the hallucinatory, suffers a subjectively indistinguishable hallucination as of a white cup on a desk: that is, it seems to her that there is a white cup where in fact there is none” (Schellenberg 2013: 699).3 If we stipulate that Percy occupies a paradigmatically good case of perception (nothing unusual either about the environment or Percy’s state of mind), then we can state BETTER BASIS as follows: Percy’s grounds for believing at least some propositions about the white cup include some grounds that Hallie cannot have, even though Hallie’s state is subjectively indistinguishable from Percy’s. What does it mean to say that Hallie’s state is “subjectively indistinguishable” from Percy’s? I take it that it is to say at least this much: introspection of her perceptual experience cannot reveal to Hallie that she is not in Percy’s situation. This does not imply that there cannot be some other way for Hallie to discover that she is not in Percy’s situation. Nor does it imply that introspection of her perceptual experience cannot reveal to Percy that she is not in Hallie’s situation. In short, we are under no obligation to interpret claims of “subjective indistinguishability” symmetrically: if Hallie’s experience is subjectively indistinguishable from Percy’s, this does not imply that Percy’s experience is subjectively indistinguishable from Hallie’s.4 This point about asymmetry allows us to distinguish two categories of philosopher who accept BETTER BASIS. Recall the thesis: Percy’s grounds for believing at least some propositions about the white cup include some grounds that Hallie cannot have, even though Hallie’s state is subjectively indistinguishable from Percy’s. Schellenberg thinks of this as an “externalist”

244

Ram Neta

thesis about evidence, but McDowell and Pritchard both deny that they are externalists. They both take it that, even though Hallie lacks any kind of privileged access to the differences between her perceptual state and Percy’s, Percy enjoys a kind of privileged access to the difference between her perceptual state and Hallie’s. Both McDowell and Pritchard speak of the access in question as “reflective”, but they do so only to mark its similarity to the epistemic access that we enjoy to our various practical and theoretical commitments as well as to our various reasons for those commitments. Both McDowell and Pritchard insist that Percy does enjoy such access to the fact that she perceives (and does not merely hallucinate) a white cup. In this respect, McDowell and Pritchard disagree with Schellenberg and Williamson. We can thus distinguish the latter as “externalist” proponents of BETTER BASIS while distinguishing the former as non-externalist proponents of BETTER BASIS. (I don’t call them “internalist” since both of them think that the terminology of “internalism” and “externalism” is itself misleading.) This distinction will become relevant in what follows, since the objection that I consider in the next section may seem to pose a much more serious problem for non-externalist proponents of BETTER BASIS than for its externalist proponents.

2. An Objection to Better Basis I just saw the headline on Bloomberg.com; it says that the S&P 500 fell six points today. I therefore believe that the S&P 500 fell six points today. My grounds for believing this are that it’s so reported on Bloomberg.com, and while that site does sometimes make mistakes in its reporting, news about S&P movement is not the kind of news about which it would normally make a mistake without almost instantly correcting it. So I have a belief, and I have strong grounds for that belief. If I gain further grounds for the same belief—if, for instance, I see the same news reported on other trustworthy news outlets—then I should become more confident that the S&P 500 fell six points today. If I lose grounds for the same belief—if, for instance, I learn that Bloomberg has recently made some mistakes in its reporting of S&P movement—then I should become less confident that the S&P 500 fell six points today. In short, one normal effect of gaining or losing grounds for some belief is that my confidence in the truth of the belief increases or decreases. As I have argued elsewhere (Neta 2008a), this relation between evidence and rational confidence is not merely contingent but necessary, and indeed, constitutive of what it is for an agent to possess some evidence. But I won’t assume this latter point in what follows. I will assume only that, whether or not it is constitutive or of an agent’s possessing some evidence, there is a correlation between evidence and rational confidence. In the preceding paragraph, I spoke of becoming more or less confident of the truth of a belief. I assume that believing a proposition can

Disjunctivism and Credence

245

persist through such changes in one’s level of confidence: an agent can believe a proposition for years, while being more confident of it on some days than on others. How is belief related to confidence?5 For an agent A to believe some proposition p requires A to be committed to the truth of p. It requires more than this, since an agent may be committed to propositions that she nonetheless fails to believe. But whatever else belief involves, it involves at least such a commitment to the truth of what one believes. Indeed, it’s only because belief involves such commitments that we can rightly criticize someone’s beliefs by pointing out what propositions she is committed to by virtue of holding those beliefs. In contrast, for an agent A to have some degree of confidence n:(1 − n) in a proposition p requires A to be committed to the fairness of a bet on whether p at odds of n:(1 − n). Again, having a particular degree of confidence in p may require more than this, but it must require at least this much, or else we could not criticize someone’s degrees of confidence by pointing out those degrees of confidence commit her to the fairness of a Dutch Book. To believe a proposition is to be committed to its truth. To have grounds for believing a proposition is to have reason to commit oneself to its truth. To be confident to degree n:(1 − n) in a proposition is to be committed to the fairness of a bet on the truth of that proposition at odds of n:(1 − n). To have grounds for being confident to a certain degree in a proposition is to have reason to commit oneself to the fairness of a bet on the truth of that proposition at corresponding odds. Of course, if you are committed to the truth of a proposition, then you are also committed to the fairness of betting on the truth of that proposition at long odds. But this is just to say that high confidence is necessary for belief, even though—as lottery cases show—high confidence is not sufficient for belief. Thus, having grounds for belief is sufficient for having grounds for high confidence, though having grounds for high confidence is not sufficient for having grounds for belief. These points about the relations between belief and confidence—and between grounds for belief and grounds for confidence—will be important in what follows. We can use the relation between an agent’s possession of evidence on the one hand, and her grounds for belief on the other, to develop a test for whether or not a particular agent’s grounds for a belief have increased or decreased: if it is rationally incumbent on the agent to become more confident in the belief, this indicates that she has gained grounds for that belief, or else lost some grounds she may have had for not believing; whereas if it is rationally incumbent on the agent to become less confident in the belief, this indicates that she has either lost grounds for that belief, or gained some grounds for not believing. But suppose it is not rationally incumbent on the agent to change her level of confidence in either direction: can we draw any conclusions concerning changes in the agent’s grounds from such a supposition? No: if it is not rationally incumbent on the agent to change her level of confidence in some hypothesis, that could

246

Ram Neta

be because there was no change in the agent’s grounds, or it could be for other reasons. For instance, suppose that you review 1,000 Bloomberg stories and discover five errors among them. Given only this information, how confident should you now be that the next Bloomberg story will contain an error? Clearly, you should be .005 confident of this. But now suppose that you review 100,000 more Bloomberg stories and discover 500 errors among them. Now how confident should you be that the next Bloomberg story will contain an error? Once again, you should be .005 confident of this. So it is rationally incumbent on you to keep your confidence at the same level as before, but clearly your grounds for that degree of confidence are much greater in the second case than in the first. From the fact that your rational degree of confidence does not change, nothing can be immediately inferred concerning any change in your grounds for the credal state: your grounds could have increased greatly, as in the example just given, or they could have remained completely unchanged. The test just offered uses facts about which credal changes are rationally incumbent on an agent to determine how the agent’s grounds have changed. But notice that my example of a case in which no credal change is rationally incumbent on an agent, despite a big change in the agent’s grounds, the grounds at issue were grounds for the agent to hold a particular credal state—viz., to be .005 confident that there is an error in the next Bloomberg story. Are there examples of the same phenomenon—examples in which no credal change is rationally incumbent on an agent, despite a big change in her grounds—where the grounds at issue are grounds for a belief, and not just for a credal state? It’s not obvious that there are, since the example just offered works because the grounds it involves are statistical, and it’s not obvious that we can have statistical grounds for a belief, as opposed to a credence.6 Typically at least, if we get additional grounds for a belief, then it is rationally incumbent upon us to become more confident in that belief; and if we lose grounds for a belief, then it is rationally incumbent upon us to become less confident in that belief. But now consider the situations of Percy and Hallie, and notice that it’s possible for a subject to transition from one of their situations to the other. A subject who’s enjoying paradigmatically successful perceptual access to a mind-independent object could eventually come to be hallucinating, and vice-versa. Of course, it’s possible that such a transition would have to be gradual in order for the subject on either side of the transition to count as well-situated: I leave it open that complete perceptual success requires not only the proper workings of one’s perceptual competence, but also the exercise of that competence in a situation in which one has not just recently been hallucinating, or in which one is not soon to begin hallucinating. Perhaps knowing about mind-independent objects by looking at them requires not only that one’s spatial surroundings be free of facades, but also that one’s temporal surroundings be free of facades as well.7

Disjunctivism and Credence

247

So suppose you are a subject who is transitioning from Percy’s situation to Hallie’s, or vice-versa. When you are in Percy’s situation, you will—at least according to McDowell and Pritchard—know that you are in Percy’s situation. But when you are in Hallie’s situation, you will not know that you are in Hallie’s situation—you will instead believe falsely that you are in Percy’s situation. As you transition from Hallie’s situation to Percy’s, or vice-versa, how should you adjust your confidence that you are in Percy’s situation? The answer to this question seems obvious: since Hallie’s situation is (by hypothesis) subjectively indistinguishable from Percy’s, unless you have some so-far-unmentioned way of knowing that you are making the transition from one state to the other, you should not alter your confidence that you are in Percy’s situation at all. It is not rationally incumbent on you to alter your confidence that you are in Percy’s situation—indeed, it seems rationally incumbent on you to not alter your confidence that you are in Percy’s situation, since the transition itself is, by hypothesis, subjectively undetectable to you. So far as you can tell as you undergo this transition, your situation is not changing in any way that is relevant to the issue of whether you are in Percy’s situation or in Hallie’s situation. If you know (and so believe) that you are in Percy’s situation when you are, then it seems rational for you to continue to hold this same belief, and to hold it with the same degree of confidence, even after you’ve transitioned into Hallie’s situation. Or, if you believe that you are in Percy’s situation when you are not, then it seems rational for you to continue to hold this same belief, and hold it with the same degree of confidence, even after you’ve transitioned into Percy’s situation. But if the transition from Hallie’s situation to Percy’s involves gaining additional grounds for the belief that you are in Percy’s situation, or if the transition from Percy’s situation to Hallie’s involves losing some grounds for the belief that you are in Percy’s situation, then how can it be rationally incumbent on you to not alter your degree of confidence that you are in Percy’s situation throughout the transition? If transitioning from Percy’s situation to Hallie’s involves losing grounds for believing that you are in Percy’s situation, then why should you not lower your confidence that you’re in Percy’s situation throughout this transition? If transitioning from Hallie’s situation to Percy’s involves gaining grounds for believing that you are in Percy’s situation, then why should you not raise your confidence that you’re in Percy’s situation throughout this transition? The proponent of BETTER BASIS must answer these questions if her view is to be credible. But how can she answer them? Another way to raise this challenge for the defender of BETTER BASIS is by means of the following argument: BETTER BASIS: Successful perception of a mind-independent object can provide a better basis for belief about that object than can an indistinguishable illusion or hallucination.

248

Ram Neta BETTER BASIS STRONGER EVIDENCE: For any two subject timeslices, S1 and S2, if S1 has a better basis for p than S2, then, all else equal, S1 has stronger evidence for p than S2. STRONGER EVIDENCE HIGHER RATIONAL CONFIDENCE: For any two subject time-slices, S1 and S2, if S1 has stronger evidence for p than S2, then S1’s rational level of confidence in p is higher than S2’s rational level of confidence in p. BETTER BASIS HIGHER RATIONAL CONFIDENCE: For any two subject time-slices, S1 and S2, if S1 successfully perceives a mindindependent object and S2 has an indistinguishable hallucination of that object, then, all else equal, S1’s rational level of confidence in the existence of that object should be higher than S2’s rational level of confidence in the existence of that object.

The conclusion of this argument appears to be unacceptable, but the argument is valid. If BETTER BASIS is true, then one of the other two premises must be false: but which one? Neither McDowell nor Pritchard address these questions about rational confidence, nor do they say anything about rational credence that indicates how they would address this question. But Schellenberg (2016) does address a related question (raised by Neta 2016) about rational credence, and what she says in addressing that question indicates how she might address the questions just raised about the agent who transitions from Hallie’s situation to Percy’s, or vice-versa. Specifically, it indicates how she would try to defend the conclusion of the argument. Here is Schellenberg: Neta asks how evidence bears on the rationality of mental states other than beliefs, such as credences. He imagines a case in which one is first perceiving a white cup (cup1) and then starts [also] hallucinating a cup. Neta asks “what implications does this new hallucination have for the rationality of Percy’s states of comparative confidence, or for her rationality of her degrees of confidence?” In response, I would say that the subject in this case has phenomenal and factive evidence for the presence of cup 1 and phenomenal evidence for the cup he is hallucinating. In contrast to Neta, I see no reason for thinking that his evidence for the presence of cup1 changes after he has started to hallucinate an additional cup—while still perceiving cup1. After all, he does not know he is hallucinating an additional cup. So he has no reason to doubt his over all epistemic standing. Moreover, there are good reasons to treat separately the fact that he is veridically perceiving cup1 while hallucinating an additional cup. We do not have to treat these two aspects of his current mental state as interfering with one another.

Disjunctivism and Credence

249

I would say that his rational confidence in cup1 being present is 1 (before and after he starts hallucinating the second cup), but that his rational confidence in an additional cup being present is lower. After all, he is hallucinating rather than perceiving that additional cup. Parallel to what I said about evidence above, I see no reason to think that hallucinating an additional cup should lower his rational confidence in the presence of cup1. This approach goes hand in hand with arguing that the rationality of his degree of confidence will change as his environment changes. In that sense, I am following the standard externalist approach about rationality in holding that the amount of rational confidence one has can change due to external factors. (Schellenberg 2016: 945) For Schellenberg, then, if an agent enjoys fully successful perception of one cup while simultaneously hallucinating a second cup, then, even if she has no way of telling which apparent cup is perceived and which is hallucinated, it is nonetheless rationally incumbent on her to be more confident of the presence of the perceived cup than about the presence of the hallucinated cup. This suggests that Schellenberg would adopt an equally externalist stance with regards to the perceiver who’s transitioning from Percy’s situation to Hallie’s, or vice-versa: it is rational for such a perceiver to be more confident in the presence of a perceived cup when she’s in Percy’s situation than when she’s in Hallie’s, even though her transition from one situation to the other is not noticeable in any way to her. This verdict seems obviously wrong: An agent who maintains her level of confidence throughout the transition described is not thereby rationally criticizable in any way. Indeed, an agent who raises or lowers her level of confidence throughout the transition described, despite being unaware of any reason whatsoever for doing so, is thereby rationally criticizable for doing so. But perhaps I’ve misunderstood Schellenberg’s position: perhaps her position concerns not how it’s rational for an agent to change her credences in certain conditions, but rather what synchronic credal states are themselves rational under what conditions. And perhaps, just as it follows from probabilistic views of credence that an assignment of maximum probability to all logical truths is rational under all evidential conditions, it also follows from disjunctivist views of evidence that an assignment of maximum probability to all truths in our evidence set is also rational under all evidential conditions. On this view, independently of whether the agent is rational to raise or lower her credence during the transition from Percy’s situation to Hallie’s (or vice-versa), the credal state that is rational while she’s in Percy’s situation differs from the credal state that is rational while she’s in Hallie’s situation: if the act of transitioning from

250

Ram Neta

one to the other is itself not rational, that means only that an agent who finds herself in such a situation is not in a position to comply with all the demands of rationality—but all probabilists must issue verdicts of this sort about some cases (e.g., those who become more or less confident about some proposition being a logical truth). If this is how to understand Schellenberg’s view, then it makes no implausible predictions about how an agent should adjust her credences, because it makes no predictions whatsoever about how an agent should adjust her credences. But now let’s consider whether this view about the rationality of synchronic credal states is correct. The case against the view is simple and obvious: maximal confidence concerning the presence of a white cup does not appear to be more rational for an agent in Percy’s situation than for an agent in Hallie’s situation—rather, it appears to be no more rational. Of course, this simple, obvious argument against the view is defeasible: perhaps the case in favor of the view can explain away the aforementioned appearance. But then, what is the case in favor of the view? McDowell, Pritchard, and Schellenberg don’t address this question at all, or even say anything to suggest how they might answer it. Does Williamson say anything to suggest an answer? It may at first appear that he does. Sometimes, he treats evidential probabilities as measures of “how well off a proposition is on the evidence” (Williamson 2009: 333). So construing evidential probabilities, Williamson’s claim would then be that propositions concerning the presence of a white cup are better off on Percy’s evidence than on Hallie’s evidence. Better off how? To judge from remarks that Williamson makes elsewhere,8 it seems that evidential probability measures how strongly evidentially supported a proposition is for an agent at a time. Whatever we think of evidential probability as a measure of evidential support, this has no obvious implications for the degree of confidence that it is rational for an agent to have in a particular proposition. In fact, Williamson explicitly allows that a proposition may be maximally evidentially supported, even if an agent should not have a maximal degree of confidence in it.9 This suggests what is anyway clear from much of Williamson’s writings on this issue, which is that the degree to which a proposition is evidentially supported for an agent at a time, and the degree of confidence that it is rational for the agent to invest in that proposition at that time, needn’t in general be proportional—at least rationality does not (on Williamson’s view) require any such proportionality. So Williamson’s discussion of evidential support doesn’t tell us anything directly about the degrees of confidence that it is rational for our transitioning subject to invest in propositions about the white cup—either before or after the transition. Williamson says nothing to provide needed support for the view (concerning differences in rational synchronic credence) that I considered earlier on Schellenberg’s behalf. Williamson does, however, indicate that there is some relation between the degree of evidential support that a proposition has, on the one hand,

Disjunctivism and Credence

251

and what he calls the rational degree of “outright belief”, on the other. To believe a proposition is to occupy a particular mental state—but just as states in general can be occupied more or less fully (the car may be almost red, somewhat red, very red, etc.), so too can belief-states be occupied more or less fully. But the degree to which one believes a proposition is not the same as the degree of confidence that one invests in the proposition. As Williamson points out, one may invest confidence of ½ that a fair coin will land heads on the next toss—but to invest such confidence is not to half-believe that it will land heads. On the contrary, it is to not believe at all that it will land heads. Perhaps Williamson would apply this notion of rational degree of outright belief to explain what is rationally incumbent on our transitioning subject: when she’s in Percy’s situation, it is rationally incumbent on her fully to believe propositions about the white cup; whereas once she’s transitioned to Hallie’s situation, it is rationally incumbent on her to believe these same propositions less than fully. This could be understood either diachronically—as claiming that it is rationally incumbent on her to lower her degree of outright belief—or synchronically—as claiming nothing about what process it is rationally incumbent on her to undergo, but rather as claiming something about a difference in the degree of outright belief that it is rationally incumbent on her to have at each moment. But once again, this verdict is not credible. If the subject’s transition is not detectable to her in any way at all, and if it is rationally incumbent on her to fully believe some propositions about the white cup when she’s in Percy’s situation, it must also be rationally incumbent on her to fully believe those same propositions when in Hallie’s situation. Or, at the very least, it is not rationally incumbent on her to believe those same propositions to a lower degree when in Hallie’s situation. Williamson might wish to resist this judgment about our transitioning subject on the grounds that it is a manifestation of the internalist prejudice that completely undetectable differences cannot make a difference to what is rationally incumbent on a subject. This is a prejudice that Williamson takes himself to have called into question by means of his famous argument to the effect that no non-trivial mental state is “luminous” to its subject; i.e., no non-trivial mental state is such that, whenever you’re in it, you’re in a position to know that you’re in it. But even if we grant the soundness of Williamson’s anti-luminosity argument, the conclusion of that argument is consistent with the claim that completely undetectable differences cannot make a difference to what is rationally incumbent on a subject. For the claim (ANTI-LUMINOSITY) No non-trivial mental state is such that, whenever you’re in it, you’re in a position to know that you’re in it is consistent with the conjunction of

252

Ram Neta (KNOWABILITY) Some non-trivial mental states are such that, on some of the occasions when you’re in them, you’re in a position to know that you’re in them

and (ACCESSIBILIST MENTALISM) Mental states fix what is rationally incumbent on a subject—but they do so on only those occasions in which the subject is in a position to know that she’s in them. Williamson’s anti-luminosity argument does nothing to tell against the common judgment that completely undetectable differences cannot make a difference to what is rationally incumbent on a subject—for that judgment may manifest a commitment to the conjunction of KNOWABILITY and ACCESSIBILIST MENTALISM, the conjunction of which is fully consistent with ANTI-LUMINOSITY. I conclude that there is no case to be made against this common judgment. And the case in favor of it is simple and obvious. To review: BETTER BASIS says that an agent’s successful perception of a mind-independent object can provide her with a kind of epistemic ground for belief concerning that object that she cannot possess unless the belief is true. Whatever epistemic grounds Hallie has for believing propositions about a white cup before her, Percy has all those same grounds, and more. But this raises the question: why isn’t it then rationally incumbent on a subject who’s transitioning from Percy’s situation to Hallie’s to have a lower degree of confidence in the truth of those propositions at the end of the transition than at the beginning? We’ve examined various ways of undermining the presupposition of this last question, and found them unsuccessful. I conclude that the question is a good one, and the proponent of BETTER BASIS needs an answer. The proponent of BETTER BASIS had better be prepared to explain either why, contrary to appearances, BETTER BASIS STRONGER EVIDENCE is false, or STRONGER EVIDENCE HIGHER RATIONAL CONFIDENCE is false, or BETTER BASIS HIGHER RATIONAL CONFIDENCE is true. In the next section, I offer the proponent of BETTER BASIS an answer to these questions. It’s an answer that embraces the argument from BETTER BASIS to BETTER BASIS HIGHER RATIONAL CONFIDENCE, but avoids implausible commitments concerning the transitioning subject, by insisting that, in the envisaged cases of transitioning subjects, the “all else equal” clause of BETTER BASIS HIGHER RATIONAL CONFIDENCE is not satisfied.

3. Percy, Hallie, and Rational Degrees of Confidence Suppose that the white cup in front of Percy is my favorite white cup, but Percy doesn’t know this and doesn’t even know me. When Percy sees

Disjunctivism and Credence

253

the white cup, she is seeing my favorite white cup, but the fact that it is my favorite white cup is not a fact disclosed to her by her perceptual state. In fact, her perceptual state itself may provide her with no evidence whatsoever that she sees my favorite cup: if she has no idea who I am, and no evidence of my own existence, then she would have no evidence of anything that would very obviously entail my existence, including the fact that the white cup in front of her is my favorite white cup. In general, even though successful perception involves a relation between a perceiver and something perceived, not every true description of that relation is a description that the perception justifies the perceiver in accepting as true. What the perception justifies the perceiver in accepting as true depends not only on the obtaining of the perceptual relation, but also on the particular way in which that relation is presented to the perceiver—a way that may be available to some perceivers but not others in virtue of their background knowledge, or their conceptual repertoire, etc. Suppose that Percy looks at the same white cup for a long time. She starts looking at it before I am born, and she continues looking at it steadily for a decade, during which period I am born, I am shown the same white cup, and that white cup eventually becomes my favorite white cup. Percy has undergone a transition: she began the decade looking at a white cup that was not my favorite white cup, and she ended the decade looking at a white cup that was my favorite white cup. It’s numerically the same white cup, of course, but its properties are different. This difference may be undetectable to Percy, and if it is, then it is not rationally incumbent on Percy to change any of her opinions about the white cup that she’s looking at, nor need there be any change in the opinions about the white cup that it is rationally incumbent on Percy to have at the beginning of the decade and at the end of it. (It may, of course, be rationally incumbent on her to change her opinions about all sorts of other things, e.g., how bored she is of still looking at the same white cup.) Now suppose that, instead of the white cup changing its relational properties during Percy’s decade-long view of it, something different happens: halfway through the decade, one white cup is very quickly replaced by a numerically distinct but visually indistinguishable white cup, and the replacement occurs in just a few nanoseconds—too quickly for Percy to see it happen. Percy spends the first half of the decade seeing cup 1, and spends the last half of the decade seeing cup 2. This is a difference in the object to which Percy is perceptually related, but once again, this difference makes no difference to what opinions it is rationally incumbent on Percy to have at the beginning and at the end of the decade. Given the stipulations of the case, it is rationally incumbent on Percy to think that she’s been seeing the same cup throughout the decade (unless she has background evidence to the effect that such indiscernible switches happen around here). But what if Percy begins the decade by seeing what she knows to be her favorite white cup. Since she knows that it is her favorite white cup

254

Ram Neta

that she’s seeing, it is rationally incumbent on her to be very confident that she is seeing her favorite white cup. But halfway through the decade, her favorite white cup is indiscernibly switched with a visually indistinguishable white cup. By the reasoning in the preceding paragraph, it will continue to be rationally incumbent on Percy to be very confident that she is seeing her favorite white cup—even though it is now false that she is seeing her favorite white cup. In other words, the reasoning that we’ve just endorsed implies that it can be rationally incumbent on a perceiver to invest a high degree of confidence in falsehoods about what she’s perceiving. Notice, though, that this does not imply that it can be rationally incumbent on a perceiver to believe falsehoods about what she’s perceiving: recall the distinction drawn earlier between belief and high confidence. What is rationally incumbent on Percy to believe may change across visually indiscernible changes in the identity of the cup that she sees, even if nothing about her rational degrees of confidence changes. Rational degrees of confidence are determined by evidential support, but rational belief may be determined by other factors in addition. In each of the transitions described here, what changes across the transition is the identity or the properties of the cup that Percy sees. But the same verdicts should hold for other indiscernible changes, including a change from Percy’s situation to Hallie’s: in such a case, the transitioning subject’s rational degrees of confidence at the beginning of the transition should be just as they should be at the end. But, since rational confidence is determined by evidential support, how does this fit with the implication of BETTER BASIS that the subject has better evidence at one end of the transition than at the other? To see how this can work, consider the following two situations: You see a seemingly fair coin flipped 10 times, and it comes up heads five times and tails five times. The track record confirms that the coin is fair, and so you should have confidence .5 that the coin will come up heads on the next toss. You see a seemingly fair coin flipped 10,000 times, and it comes up heads 5,000 times and tails 5,000 times. The track record confirms that the coin is fair, and so you should have confidence .5 that the coin will come up heads on the next toss. In both of these two situations, you should have confidence .5 that the coin will come up heads on the next toss, but this confidence is based on stronger evidence in the second situation than in the first. Additional evidence concerning some hypothesis need not always alter the degree of confidence that the evidence-possessing agent should invest in the hypothesis. But the contrast between Percy’s perceptual evidence and Hallie’s is not a contrast in quantity, but in kind: Percy’s evidence concerning the

Disjunctivism and Credence

255

white cup is of a different kind than Hallie’s, and not simply more of the same. Percy’s evidence set includes, or entails, the presence of the white cup, whereas Hallie’s doesn’t. So how can Percy’s rational degree of confidence concerning the white cup be the same as Hallie’s? We’ve assumed here that an agent’s rational degree of confidence in a hypothesis is fixed by the support that that hypothesis receives from the agent’s evidence. We’ve also granted that Percy has evidence that Hallie lacks, and that Percy’s evidence entails or includes the presence of the white cup, whereas Hallie’s doesn’t. How can both of these claims be true? Here’s how: the degree to which an agent’s evidence supports a proposition is a product not just of what evidence the agent has, and the normative relation between that evidence and the hypothesis in question. It is also a product of how confident the agent should be of that evidence itself. We can illustrate this point by specifying further features of Percy’s case and Hallie’s case. Suppose that Hallie hallucinates a white cup, and that, given her total evidence, it is rational for her to be 95% confident that she successfully perceives a white cup, and 99% confident that it appears to her as if she perceives a white cup. Now, suppose that Percy successfully perceives a white cup, and that her evidence therefore includes the fact that she perceives a white cup. Does it follow that she must be more than 95% confident that she successfully perceives a white cup? That follows only if we must be rationally certain of each element of our evidence. But why should anyone assume that? Much of what we ordinarily take to be our evidence consists of facts concerning which we have slightly less than maximal confidence: why should we assume that appearances are here misleading, and that it is rationally incumbent on us to be maximally certain of everything in our evidence? Of course, it may be rationally incumbent on us to be highly confident of the conjunction of our evidence—but satisfying that requirement (if it is, as I believe, a requirement) does not require that we have maximal confidence of each element of our evidence: we could have less than maximal confidence in each element, so long as those elements are probabilistically interrelated in such a way that my rational confidence in their conjunction is not much lower than the lowest level of rational confidence in any individual element. Percy’s evidence set can include facts about the white cup missing from Hallie’s evidence set, but this does not entail that Percy should be more confident of propositions about the white cup than Hallie is: Percy’s rational degree of confidence is determined by evidential support, but evidential support is determined not just by what it is in Percy’s evidence set, but also by how confident it is rational for Percy to be concerning those contents of her evidence set. The proponent of BETTER BASIS can give a satisfactory answer to the present question concerning rational confidence so long as she grants

256

Ram Neta

that it is not rationally incumbent upon an agent to be maximally confident of every proposition in her evidence set. But granting this will raise a worry:10 if we grant that Percy’s rational degrees of confidence need not differ from Hallie’s, then why so much as bother distinguishing a subject’s evidence, on the one hand, from any other factors that determine that subject’s rational degrees of confidence, on the other? If a diminution of evidence can be traded off against an increase in the subject’s rational confidence in the evidence she maintains, or if an augmentation of evidence can be traded off against a decrease in the subject’s rational confidence in that evidence, then why bother trying to separate out these two components that determine a subject’s rational confidence? And if we have no reason to separate out these two components, then why engage in the kind of theorizing about evidence that involves claims like BETTER BASIS in the first place? Although I think BETTER BASIS has no immediate implications concerning the degree of confidence that it is rationally incumbent on us to have in various hypotheses, determining such degrees of rational confidence is not the only work that having evidence is supposed to do. What evidence you have has implications not only for your degrees of rational confidence, but also, as I’ve argued elsewhere, for what it is rational for you to believe, and what it is possible for you to know.11 But why do belief and knowledge matter? Why should epistemologists talk about anything other than rational degrees of confidence, and the evidence that fixes those? These are questions for another paper—or rather, several other papers. In brief, my answer to them is similar to, but not quite the same as, Williamson’s. As I have argued elsewhere, the evidence that fixes an agent’s rational distribution of confidence across hypotheses consists of all and only those facts that the agent is in a position to know non-inferentially (2018). If we are interested in rational degrees of confidence, then we must be interested in evidential support, and so in what evidence an agent possesses, and so in what it is that puts us in a position to gain noninferential knowledge. Even if the proponent and the opponent of BETTER BASIS can make all the same predictions concerning our rational degrees of confidence, they are committed to offering differing explanations of these predictions. The reason for accepting BETTER BASIS is not that it makes more accurate predictions than other views, but rather that it can do a better job of explaining those same predictions.1213

Notes 1. See McDowell (2011: esp. sec. 8). 2. McDowell (2011: sec. 11) discusses a case of an experimental subject who possesses a fully normal capacity to know the colors of things by looking at them, but is given misleading information that impedes her successful exercise of this capacity: once she possesses the misleading information, she can continue to see things by looking at them, but she can no longer come to know their colors by seeing them.

Disjunctivism and Credence

257

3. See Schellenberg (2016) for more discussion of the difference between Percy and Hallie. 4. See Neta (2008b). 5. The answer I provide in the remainder of this paragraph is an abbreviated form of the answer given in Neta (2015). 6. Nelkin (2000), Weiner (2005), and Smith (2010) all argue that statistical evidence in favor of p cannot serve as adequate grounds for believing that p. 7. See Neta (2008b) for an elaboration of this thought in response to the criticism of disjunctivism in Johnston (2004). 8. E.g., Williamson (2013: 92): “In a sceptical scenario, beliefs only partially supported by one’s evidence appear to be fully supported by it”. 9. Williamson (2009: 336): “it is not reasonable to have the greatest possible degree of confidence in a complex unproved logical truth on present evidence. We should rightly regard someone who bet everything he held most dear for trivial gain on such a logical truth as far more deserving of criticism than someone who refused the bet”. 10. In conversation, it’s raised a technical worry in addition to the one discussed in the text, since Jeffrey Conditionalization is the standard model of credal updating with uncertain evidence. What about the worry raised by Weisberg (2009) that Jeffrey Conditionalization cannot accommodate both holism and commutativity? This worry is based on a pair of confusions. First, as Weisberg notes, this same worry applies to regular conditionalization as well, so letting evidence be less than certain has nothing to do with generating the worry. Second, the way to accommodate both commutativity and holism is to follow Field’s strategy of mapping experiences onto evidence, but letting the right mapping depend on the whole temporal sequence—past, present, and future—of the subject’s experience. Understood thus, conditionalization is not a usable “update rule”. But it was never meant to be: it’s a wide-scope diachronic constraint on credences, not a narrow-scope constraint on activity, let alone a guide to intellectual conduct. 11. See Neta (2008a) on the implications of evidence for what it is rational for you to believe. See Neta (forthcoming) on the implications for what it is possible for you to know. 12. Neta (2011) argues on these grounds for an internalist version of BETTER BASIS. 13. I am grateful to Susanna Schellenberg, Casey Doyle, and Joe Milburn for comments on an earlier version of this chapter.

References Johnston, Mark, 2004, That obscure object of hallucination, Philosophical Studies 120: 113–183. McDowell, John, 2011, Perception as a Capacity for Knowledge. Milwaukee, WI: Marquette University Press. Nelkin, Dana, 2000, The lottery paradox, knowledge, and rationality, The Philosophical Review 109: 373–409. Neta, Ram, 2008a, What evidence do you have? British Journal for the Philosophy of Science 59: 89–119. ———, 2008b. In defense of disjunctivism, in Fiona MacPherson and Adrian Haddock, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press. ———, 2011, A refutation of Cartesian fallibilism, Nous 45: 658–695. ———, 2015, Coherence and deontology, Philosophical Perspectives 29: 284–304.

258

Ram Neta

———, 2016, Perceptual evidence and the capacity view, Philosophical Studies 173: 907–914. ———, 2018, Your evidence is the set of facts that are manifest to you, in Veli Mitova, ed. The Factive Turn in Epistemology. Cambridge: Cambridge University Press. Pritchard, Duncan, 2013, Epistemological Disjunctivism. Oxford: Oxford University Press. Schellenberg, Susanna, 2013, Experience and evidence, Mind 122: 699–747. ———, 2016, Phenomenal evidence and factive evidence defended: A reply to McGrath, Pautz, and Neta, Philosophical Studies 173: 875–896. Smith, Martin, 2010, What else justification could be, Nous 44: 10–31. Weiner, Matthew, 2005, Must we know what we say? The Philosophical Review 114: 227–251. Weisberg, Jonathan, 2009, Commutativity or holism? A dilemma for conditionalizers, British Journal for the Philosophy of Science 60: 793–812. Williamson, Timothy, 2000, Knowledge and Its Limits. Oxford: Oxford University Press. ———, 2009, Williamson on Knowledge. Eds. Patrick Greenough and Duncan Pritchard. Oxford: Oxford University Press. ———, 2013, Response to Cohen, Comesana, Goodman, Nagel, and Weatherson on gettier cases in epistemic logic, Inquiry 56: 77–96.

13 Disjunctivism, Skepticism, and the First Person Adrian Haddock*

1. Here is John McDowell. One can sometimes tell what someone else feels or thinks by seeing and hearing what he says and does. It is very common for philosophers to interpret this idea so that “what he says and does” is taken to allude to a basis for knowledge of what the person feels or thinks. The thought of a basis here has two elements. The first is that the basis is something knowable in its own right . . . knowledge of the basis could have this status—be knowledge—independently of the status of what it is a basis for. The second is that judgments about what the person feels or thinks emerge as knowledgeable in favourable cases because of an inferential relation in which they stand to the basis. The notion of a criterion, as used by Wittgenstein in connection with this sort of knowledge, is often interpreted on these lines. (McDowell 1988: 455)1 McDowell recommends replacing this interpretation of the notion of a criterion with the following. Knowledge that one is confronted by a criterion for the person’s feeling as one judges him to . . . can be an exercise of the very same (of course fallible) capacity that we speak of when we say that one can tell what someone feels from what he says and does: not an epistemically independent capacity whose deliverances ground the epistemic status of judgments as to how people feel. (Ibid.: 456) “Knowledge that one is confronted by a criterion for [something]” is self-knowledge. That is reflected in how the embedded pronoun (“one”) is functioning in this formulation. Although it is sometimes possible to replace “one” with “someone” without affecting the sense, that is obviously not how it is here; if McDowell had written “Knowledge that

260 Adrian Haddock someone is confronted by a criterion for [something]”, he would have made it look as if the knowledge concerns someone who might not herself have any knowledge of being confronted by a criterion, because he would have made it look as if the knowledge concerns someone who, for all that the knowledge itself determines, might be other than the knowing subject. McDowell is speaking of knowledge that the knowing subject could express by saying “I am confronted by a criterion for something”. And the criterion might be specified, in the case at issue, by speaking generically of “what [the other] says and does”: that “can be [the subject’s] criterion for what [the other] feels, in the quite non-technical sense of ‘way of telling’—without any implication that ‘what [the other] says and does’ stands in for something more specific” (Ibid.: 455–456). But the crucial thought is that the subject’s self-knowledge of being confronted by this criterion—knowledge that she could express by saying “I am confronted by the criterion”, or “I see what he says and does”—is an exercise of the very same fallible capacity whose exercise constitutes knowledge of what the other feels. And this sameness is reflected in the fact that the criterion, embraced in the self-knowledge constituted by the exercise of this capacity, can be further specified so as to reveal it to not fall short of what is known: by saying “I see him expressing what he feels in what he says and does”, for example. 2. This thought is at the heart of McDowell’s “disjunctivism”. And although McDowell initially presented his disjunctivism as a contribution to the epistemology of “other minds”, it is clear from subsequent writings that it is intended to constitute a response to Cartesian skepticism, and as such is intended to contribute to epistemology more generally.2 In general, McDowell’s disjunctivism is the idea that the capacity whose exercise constitutes the self-knowledge that is necessary knowledge is not epistemically independent of the capacity whose exercise constitutes the knowledge for which it is necessary, but the very same capacity. The aim of this chapter is to clarify what this idea comes to, by getting clear about the response to Cartesian skepticism that it affords. 3. This will require saying something about Cartesian skepticism. Although it is possible to gripped by Cartesian skepticism without having read a word of philosophy, it does not emerge in a vacuum; it emerges, and seems inevitable, only in the context of a certain intellectual project, which Descartes founded—the project of “modern” epistemology. To be gripped by skepticism, of this form, is to be in the grip of this project. But a central thought of this chapter is that this project is responsive to a more general epistemological ambition. Understanding the project in this way makes it possible, not merely to isolate what it is about the project that generates the skeptical consequence, but to point to the possibility of a position that dismantles the project, and in dismantling it avoids the consequence, but without therein abandoning the general ambition to

Disjunctivism, Skepticism, and the First Person

261

which the project is responsive. That enables a path to be trod between the Scylla of the modern project, and its attendant skepticism, and the Charybdis of contemporary epistemology, which also rejects the project, but in rejecting it equally abandons the general ambition. Insofar as this general ambition is not kept in view, it will be tempting either to understand McDowell’s disjunctivism as a contribution to the modern project, or to try to understand it as a contribution to contemporary epistemology. Both will drain disjunctivism of its anti-skeptical import: the first, because insofar as it is understood as a contribution to this project, it is powerless to avoid Cartesian skepticism; and the second, because as contemporary epistemology has abandoned the general ambition, when any contemporary position speaks of “knowledge” it changes the subject, and as such is incapable of engaging with Cartesian skepticism. 4. The plan is as follows. The essay will begin by presenting the project of epistemology at a very general level (sections 5–14). That will allow the project of modern epistemology to be seen as a specific version of this general project—one that inevitably leads to skepticism (sections 15–20). And it will show that, insofar as disjunctivism is understood as a contribution to this more specific project, it is powerless to resist the skeptical conclusion (sections 21–24). But the project of modern epistemology is incoherent. And the essay will bring this out, by pointing up an assumption that holds the project in place, but which the project cannot sustain (sections 25–34). This assumption is a certain way of understanding the first person: the project both must understand the first person in this way; and yet by its own lights it cannot understand the first person in this way. Without this way of understanding the first person, the project of modern epistemology falls to the ground. But the project of epistemology remains, insofar as it is understood in the very general terms with which this essay began. And insofar as disjunctivism is understood as a contribution to this purified project, it yields a satisfactory response to skepticism (sections 35–40). If the modern project is taken for granted, then the anti-skeptical import of disjunctivism will remain obscure. Only by dismantling this project, in the way explained here, can the fundamental insight in disjunctivism emerge. * 5. The founding text of modern epistemology is Descartes’s Meditations on First Philosophy. This begins by describing something that happened to its author at some point in his life: “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” (1996: 12). This sentence uses the first person. That is no accident, because the experience that it reports—“the experience of falsehood”—is at once part of its author’s autobiography,

262 Adrian Haddock and something that anyone is in a position to report by using “I”. The sentence exhibits the unity of particularity and universality that characterizes the first person. This unity is fundamental to the Meditations as a whole. Although it is possible to read the Meditations as autobiography, its author intends it to be read in a different way—not by disregarding its use of the first person, but by exploiting the universality that “I” contains. This is something that G.E.M. Anscombe notes, in her discussion of one of its central arguments: “The first-person character of Descartes’s argument means that each person must administer it to himself in the first person” (1975: 21). And this character is central to the project of modern epistemology inaugurated by Descartes. 6. To bring this out, imagine three books, each of which is primarily distinguished by how its author intends it to be read. The author might express his intention, in each case, as follows: (1) “The reader is to come to understand that I am like this”; (2) “The reader is to come to understand that everyone is like this”; (3) “The reader understands (as she would put it) ‘I am like this’”. Book (3) is like book (1) but unlike book (2) in that it is written in the first person rather than the third—but whereas book (1) seeks to communicate an understanding that is to be received in the third person, by the reader coming to understand (as she would put it) “The author is like this”, book (3) invites the reader to articulate in the first person the very self-understanding that it seeks to articulate. By contrast, book (3) is like book (2), but unlike book (1), in that the understanding that it seeks to articulate is universal rather than particular. But whereas in the case of book (2), this is because it propounds a generalization by using “everyone”, in the case of book (3), it is because it invites every one of its readers to articulate this understanding by using “I”. A text in epistemology in the tradition inaugurated by Descartes is in these respects akin to book (3), rather than to books (1) or (2). 7. That brings out the central way in which this tradition differs from contemporary epistemology. A typical text of the latter is akin to book (2), in that it seeks to communicate an understanding that says that anything of a certain kind—a “subject”, or a “self”—that satisfies a certain condition—which it goes on to specify—exemplifies a certain property— which it signifies by the word “knowledge”. And it is quite possible that this understanding concerns a “subject”, or a “self” who does not herself have this understanding: one of the doctrines of this kind of epistemology is that those it concerns (as it likes to say) “need not even have the concept of knowledge”. By contrast, any instance of the self-understanding that a text in modern epistemology seeks to articulate will be had by the very one that it concerns: it is not possible that an instance of this understanding concerns someone who does not herself have this understanding. It is interesting that much contemporary epistemology is written either in the first person, or by using a pronoun that is naturally taken to issue the same invitation—such as “one”, or even “you”. In a text of

Disjunctivism, Skepticism, and the First Person

263

contemporary epistemology, these devices cannot be read as issuing this invitation; they function at best as a plangent reminder of the kind of understanding that it is not seeking to articulate. * 8. I have just spoken of articulating an instance of self-understanding; and I might equally have spoken of self-knowledge, or self-consciousness: at the present level of generality, these come to the same. To understand the project of epistemology, it is first necessary to understand what the idea of articulating an instance of self-understanding comes to. 9. At the most general level, to articulate an instance of self-understanding just is to express a thought. This could be any thought: it does not matter what it is about. 10. That might seem wrong. It is tempting to think that to articulate an instance of self-understanding is to articulate a thought that is about something of a certain kind, a “self”, or a “subject”. As it is natural to say: “self-knowledge is knowledge of oneself”. And it is tempting to think that this explains self-knowledge as knowledge of a certain kind of thing—signified here by the reflexive pronoun. But at the present level of generality, nothing like that is built into the idea of self-knowledge. It is tempting to think that it is not ruled out, either—that, when the descent is made to a more specific level, the articulation of an instance of self-understanding will at least include the expression of a thought about a self. This is the position of modern epistemology. And it is the source of its difficulties—as will be seen. But to bring these difficulties into focus, and to bring out the contrast between modern epistemology, and the project of contemporary epistemology, it will help to remain at the present level of generality, at least at the outset. 11. The project of articulating an instance of self-understanding could be described as that of articulating an understanding of a thought “from within”. To articulate such an understanding just is to express the thought that the understanding is an understanding of. This contrasts with the project of articulating an understanding of a thought “from outside”. To articulate such an understanding is to express not the thought that the understanding is an understanding of, but a distinct thought that is directed towards this thought. There will later be cause to return to the putative project of articulating an understanding of a thought “from outside” (see sections 30–31). But the very idea that articulating an instance of self-understanding is articulating an understanding “of” a thought is liable to mislead here, because it is liable to suggest that the understanding itself consists in a thought that is directed towards a further thought: the first thought is the understanding, and the second thought is the thought that the understanding is “of”. But the idea of the thought that the understanding is “of” is merely the idea of the thought that the articulation of the understanding expresses. We might say that the thought,

264 Adrian Haddock and the understanding “of” the thought, are the same. But it is better to avoid the misleading word “of” altogether, by calling the understanding “self-understanding”. 12. At the present level of generality, the project of epistemology is a species of the project of articulating an instance of self-understanding, in that it seeks to express a thought. But as it is a species of this project, this is not all it seeks to do. It seeks not merely to express a thought, but to express a thought in a manner that reveals it to be true. That is what makes it distinctively epistemological. Thoughts are expressed in language, specifically in sentences. And to use a sentence is to grasp the thought that the sentence expresses. The idea of expressing a thought that reveals it to be true is to be understood in this light. To express a thought in a manner that reveals it to be true is to use a sentence of which the following holds: to use this sentence is to grasp not merely the thought that it expresses, but the truth of this thought. And insofar as the truth of this thought is revealed, so is that of every thought whose truth is understood to follow from the truth of this thought. At the present level of generality, the idea of knowledge just is the idea of the capacity to express a thought in a manner that reveals it to be true. The idea of skepticism about knowledge in general just is the idea that there is no such thing as this capacity. And the idea of skepticism about knowledge of the truth of thoughts of a specific sort is the idea that there is no such thing as the capacity to express thoughts of this sort in this manner. 13. It is worth bringing out how the idea of self-understanding that is internal to this project differs from the idea of self-understanding presupposed by contemporary epistemology. It belongs to this branch of epistemology to reject the KK thesis: that if someone knows something, then she knows that she* knows it—where the function of the star is to indicate that the knowledge reported by the consequent is self-knowledge.3 The standard objections to this thesis all assume that the thesis is an instance of the (absurd) idea that, if someone knows something, then she knows something else: the knowledge reported by the consequent is knowledge of the truth of a further thought, distinct from the thought that constitutes the knowledge reported by the antecedent. This assumption is (to my knowledge) never made explicit, let alone argued for; but because the knowledge reported by the consequent is self-knowledge, the effect of the assumption is that self-knowledge is itself pictured as knowledge of the truth of a further thought. This explains why contemporary epistemology shows no interest in the possibility of self-knowledge: what reason could there be to think that the possibility of this further knowledge has any bearing on the possibility of the knowledge that the antecedent reports—short of the truth of the (absurd) KK thesis? But selfknowledge does not consist in a further thought. For someone to have the self-knowledge with which the project of epistemology is concerned is, simply, for her to have the capacity to express a thought in a manner that reveals it to be true. Her having this capacity is not a condition over and

Disjunctivism, Skepticism, and the First Person

265

above her having the knowledge that the antecedent of the thesis reports: it is nothing other than her having this knowledge.4 And for this reason, it is possible to cast the idea of the kind of self-understanding with which the project is concerned in terms of the KK thesis properly understood: if someone knows something, then she knows that she* knows it—simply in that she has the capacity to expresses it in a manner that reveals it to be true. Contemporary epistemology thinks that the way to reject skepticism is to reject KK. But to reject KK properly understood is to reject the very idea of knowledge, whereas to reject—or endorse—what contemporary epistemology understands by KK has no bearing on skepticism at all. 14. Contemporary epistemology misunderstands the KK thesis because it takes for granted that the self-knowledge reported by the consequent is knowledge to the effect that a certain property is exemplified by a “self”, and as such is knowledge of something other than what is known in the knowledge reported by the antecedent—which may not concern a “self” at all (and, even if it does, is not to the effect that this “self” exemplifies the property that figures in the content of the knowledge reported by the consequent). But at the present level of generality, the idea of selfknowledge internal to the KK thesis is not the idea of knowledge of a “self”. The office of the formulation that constitutes the consequent of the thesis—and, specifically, the office of the star on the pronoun—is to show that the knowledge that the consequent reports is self-knowledge. At the present level of generality, this is merely to show that the knowledge reported by the consequent consists in the capacity to express the thought that figures as the content of the knowledge reported by the antecedent in a manner that reveals this thought to be true. And this capacity just is the knowledge reported by the antecedent. The idea of knowledge is the idea of self-knowledge, in that it is the idea of a capacity to express a thought in this manner. At the present level of generality, this is all that the idea of self-knowledge internal to the KK thesis comes to. 15. Within modern epistemology, the case for skepticism comes into focus once a descent is made, from the present level of generality to a level at which self-understanding does concern a “self”, or a “subject”. This level was operative earlier in this essay, when an instance of selfunderstanding was said to concern someone—the very one who has the understanding. At this level, the articulation of an instance of selfunderstanding is not merely a matter of expressing a thought; it is also a matter of identifying the one who thinks the thought as its thinker. Suppose that the thought is expressed by the sentence “p”. Although one way to articulate an instance of self-understanding is to use this sentence, its use does not fully articulate the instance of self-understanding, because the thought that its use expresses does not itself identify the one who thinks the thought as its thinker. Fully articulating the instance of self-understanding requires using the distinct sentence “I think p”: the thought expressed by this use contains the thought expressed by “p”, and as such equally expresses this thought; but it also goes beyond this

266 Adrian Haddock thought, because it also identifies the one who thinks it as its thinker. And it is in this light that modern epistemology is to be understood. 16. The experience of falsehood with which Descartes begins the Meditations specifies the kind of thoughts with which modern epistemology is primarily concerned: thoughts that are capable of being true, or false. And “p” expresses such a thought. But a use of “I think p” does not. Insofar as someone uses this sentence, its use expresses the thought expressed by “p”. But her use of this sentence does not merely express this thought. It equally expresses a further thought that contains the thought expressed by “p”. So, it expresses two thoughts: the thought expressed by “p”, and a further thought that contains this thought. In what follows, I will call the second of these thoughts “the thought that a use of ‘I think p’ expresses”, and the first “the thought that a use of ‘I think p’ contains”. The crucial point, for present purposes, is that the thought that a use of “I think p” expresses cannot be false. It is selfverifying, for the following reason. In using “I think p”, the user thinks the thought that her use of this expresses. And as this thought contains the thought expressed by “p”, in thinking this thought, she equally thinks the thought expressed by “p”. The thought that her use of “I think p” expresses is true just in case she thinks the thought expressed by “p”. So, the thought that her use of “I think p” expresses is true. Insofar as “I think p” is used at all, the thought that its use expresses is true. It is incapable of being false. And in using “I think p”, its user understands this: she grasps the thought that her use of this sentence expresses, and in grasping it, she grasps that it is incapable of being false. In using “I think p”, therefore, she grasps the truth of the thought that her use of this sentence expresses. And as such, her use of this sentence may be said to reveal the truth of this thought. 17. This sets the agenda for modern epistemology. A use of “I think p” reveals the truth of the thought that it expresses. As such, the thought that it expresses is something that its user knows. But the same is not true of “p”: if the thought that this sentence expresses is to count as something that its user knows, then there must be a way of expressing it that equally reveals it to be true. A use of “p” does not do this. Neither does a use of “I think p”: it reveals the truth of the thought that it expresses, but it does not reveal the truth of the thought that it contains. The search is on, then, for a way of expressing the thought expressed by “p” that does reveal its truth. 18. At this point a distinctively Cartesian assumption comes onto the scene: the sought-for manner of expression must be the same as a use of “I think p”, not only in that it must contain the thought expressed by “p”, but in that the thought that it expresses must equally be selfverifying. It is a question, however, how it can be. To reveal the thought that “p” expresses to be true, it is not enough simply to express it; it seems that its expression must somehow reveal it to be in a nexus with something else, and, through revealing it to be in this nexus, reveal it to

Disjunctivism, Skepticism, and the First Person

267

be true. And it seems that this “something else” cannot simply be another thought. On the contrary: it seems that it must be something given to the senses: only through being revealed to be related to this can the truth of the thought be revealed. It is tempting to think that a sentence beginning with “I perceive .  .  .” is a form for expressing a thought that bears on what is given to the senses. As such, it is tempting to think that “I think p because I perceive . . .” is a form for revealing the thought expressed by “p” to be in a nexus with what is given to the senses, and as such for revealing it to be true. But given the Cartesian assumption, it can do this only insofar as the thought expressed by a use of a sentence beginning with “I perceive . . .” is self-verifying: only then can the thought expressed by “I think p because I perceive . . .” be self-verifying, and only insofar as this thought is self-verifying can the Cartesian assumption be satisfied. So, the question arises as to how the thought expressed by a use of a sentence beginning with “I perceive . . .” can be self-verifying. 19. At this point it is tempting to fall into the Myth of the Given. It might seem that a sentence beginning with “I perceive . . .” can express a thought that is self-verifying, insofar as the meaning of whatever takes the place of the ellipsis is determined simply by what is given to the senses at the time of its use. To work with an easy example: suppose that the word “red” takes its place. And suppose that it is possible to explain what is meant by this word at the time of its use simply by pointing to what is given to the senses at the time. As it is tempting to say: “By ‘red’ I mean this”—where the bare demonstrative manifests the act of pointing to what is given that purportedly serves to explain the meaning that the word “red” bears at the time, and as such expresses how what is given is given to the senses at the time. It seems that there is no possibility of what is expressed by a use of “I perceive red” being false, if the meaning that “red” bears at the time of its use is explicable in this manner. 20. The conception of meaning that this presupposes is notoriously problematic—it is Wittgenstein’s target in his remarks on private ostensive definition, and Wilfrid Sellars famously awarded it the title of a myth. But, for now, there is no need to bring out these difficulties.5 It is enough to note that no attempt at revealing the truth of what is expressed by “p” which conforms to the Cartesian assumption can succeed. No expression of a thought that is incapable of being false can reveal the truth of a thought that is capable of being false. This is because the expression of one thought can reveal the truth of another only if the truth of the first entails the truth of the second. But if the truth of the first does entail that of the second then, insofar as the second is capable of being false, so is the first. Given the Cartesian assumption, it is impossible to know the truth of thoughts that are capable of being false. 21. At this point McDowell’s disjunctivism enters the fray. It rejects the Cartesian assumption that what is expressed by the sought-for manner of expression must be self-verifying. Consider a use of “I perceive p”. A use of this sentence expresses a thought that contains the thought expressed

268 Adrian Haddock by “p”. And the truth of the thought that it expresses consists in the successful exercise of a fallible capacity that is possessed by the subject that the thought serves to identify. This capacity is fallible because it admits not only of successful but also of unsuccessful exercises: when it is successfully exercised, the thought expressed by a use of “I perceive p” is true; when it is not, this thought is false. The hope is that, insofar as this thought is true, expressing it will not merely express the thought that it contains—the thought expressed by “p”—but in so doing will reveal this thought to be true. 22. But it is hard to see how a use of “I perceive p” can reveal the truth of the thought expressed by “p”. Let it be granted that a use of “I perceive p” expresses a thought that contains the thought expressed by “p”, and that if the thought that this use expresses is true, then so is the thought that this thought contains. It remains the case that no use of “I perceive p” can reveal the truth of the thought that it contains. That would require it to reveal the truth of the thought that it expresses. And it can no more do this than a use of “p” can reveal the truth of the thought that it expresses. Recognition of this point underwrites the Cartesian assumption. Disjunctivism proceeds as if the Cartesian assumption is some sort of disposable accretion. But it seems to be compulsory, if there is to be a capacity to express thoughts in a manner that reveals them to be true at all. The problem is that this assumption makes it impossible to know the truth of thoughts that are capable of falsehood. 23. It might seem that disjunctivism has a response to this difficulty. So far, disjunctivism has been presented as holding that the truth of the thought expressed by a use of “I perceive p” consists in the successful exercise of a fallible capacity. But this is only part of the story. More fully, it holds that the successful exercise of this capacity constitutes this thought, not merely as true, but as a thought whose truth the subject who successfully exercises the capacity knows. The subject knows not merely the truth of the thought expressed by “p”, but the truth of the thought expressed by her use of “I perceive p”. This is what it means to say that the capacity is, not merely fallible, but self-conscious. But how is this supposed to help? Insofar as she knows the truth of this last thought, she must be capable of expressing it in a manner that reveals it to be true. And her use of “I perceive p” does not do this. So, she needs to use a further sentence that does. “I know that I perceive p” might seem to be the obvious candidate. But the same difficulty will arise again at this level. And insofar as disjunctivism gives the same response to deal with this difficulty, and any subsequent difficulty of the same shape, it will endorse the misbegotten version of the KK thesis that was considered earlier. It will endorse the thesis that, if someone knows something, then she knows something else—in the shape of the truth of the further thought that arises at each subsequent level. And that is not merely absurd; it is hopeless to deal with the present difficulty. If a use of a

Disjunctivism, Skepticism, and the First Person

269

sentence cannot reveal the truth of a thought contained in the thought that it expresses, on account of its not revealing the truth of the thought that it expresses, then it does not matter how many further thoughts are added—so long as what is added at each subsequent level is equally incapable of being expressed in a manner that reveals it to be true, the revelation of the truth of the first thought will not arrive, and this thought will not count as knowledge. 24. The project of modern epistemology leads inexorably to skepticism. And disjunctivism is powerless to stop it. It might seem that the only way to avoid skepticism is to give up the idea that knowledge is the capacity to express a thought in a manner that reveals it to be true. But that is not a way to avoid skepticism at all. It changes the subject and, as such, leaves whatever case there might be for skepticism in place: it might conceive of itself as “anti-skeptical”, but its effect will be “precisely what the skeptic most desires”.6 The only possible positions seem to be skepticism, and a position that understands (what it calls) “knowledge”, not as the capacity to express a thought in a manner that reveals it to be true, but as a property that “subjects” or “selves” exemplify just in case they satisfy certain conditions, none of which amounts to the possession of this capacity. This is the position of contemporary epistemology. And it is not opposed to skepticism: because it understands (what it calls) “knowledge” in this way, it does not merely reject the KK thesis on the misunderstanding of what this comes to considered earlier; it rejects KK even when properly understood, because it rejects the idea that knowledge is the capacity to express a thought in a manner that reveals it to be true. The reason why contemporary epistemology shows so little interest in skepticism is because it has abandoned the very idea of knowledge internal to the epistemological project, and in so doing has abandoned not merely the project of modern epistemology, but the project of epistemology tout court: the similarity between “contemporary epistemology” and the epistemological project is merely verbal. (It is perhaps worth noting that there are those who have tried to understand McDowell’s disjunctivism within the contemporary framework:7 they present themselves as doing justice to (what they take to be) the central insight in disjunctivism, and as such, are happy to employ the language of fallible capacities; but the position they defend is merely a version of the contemporary position, which is distinguished from other versions only by its refusal to give a reductive specification of the conditions under which the property they call “knowledge” is exemplified. It equally rejects KK even when properly understood, and so equally changes the subject, and so is equally devoid of anti-skeptical import.)8 * 25. The upshot is that skepticism seems to be inevitable. Contemporary epistemology passes it by. And so long as the project of modern

270 Adrian Haddock epistemology remains in place, it is inevitable. But the modern project is incoherent: exposing its incoherence will enable the insight in disjunctivism to emerge; and as it does, skepticism will fall to the ground. 26. The form for the articulation of an instance of self-understanding was given as the sentence “I think p”. A use of this sentence was said to express not just the thought expressed by “p”, but a further thought that contains this thought. And this further thought was said to be selfverifying. That a use of “I think p” expresses a further thought is something that modern epistemology takes to be obviously true. But it is not obviously true. It is hard to see how a use of “I think p” can express any thought over and above the thought expressed by “p”. That is a central moral of G.E.M. Anscombe’s great essay “The First Person”. Extracting this moral from Anscombe’s essay, and seeing how it bears on the project of modern epistemology, will expose the incoherence of the project. 27. It might seem obvious that a use of “I think p” expresses a thought over and above the thought expressed by “p”. Suppose that “p” is “snow is white”. That sentence refers to something in a certain way (through the word “snow”), and says something of this thing (through the words “is white”). In the same way, a use of “I think p” refers to someone in a certain way (through the word “I”), and says something of her (through the words “think p”). Or so it might be natural to think. But this assumes that it is possible to explain “I” as a referring term. And this is not possible— as Anscombe’s essay shows. 28. Anscombe begins her essay by pointing up two requirements on explaining “I” as a referring term that cannot be satisfied together. The first requirement is this. To explain any word as such a term, it is not enough merely to identify, first, a word, and, second, a thing, and to say of the word that it refers to the thing. Any referring term refers to its referent in a certain way: this way is the sense of the word, and any explanation of a word as a referring term must express its sense—as it might be put: it must show its sense—through referring to what it specifies as the referent of the word in the very way that the word refers to it, paradigmatically, through using this very word. Snow-bound trivialities supply an example: (1) “Snow” refers to snow. This identifies the referent of a word through using this very word, and in so doing expresses, or shows, the word’s sense. The second requirement, by contrast, is taken to be specific to certain words, paradigmatically to “I”, “now”, and “here”. It says that any explanation of such a word must account for its indexical character—in the case of “I”, for example, it must account for the fact that its referent always varies with its user.

Disjunctivism, Skepticism, and the First Person

271

29. It is easy to see that these requirements cannot be satisfied together. Consider an explanation of the following familiar sort: (2) On everyone’s lips, “I” refers to herself. This satisfies the second requirement. But as it generalizes over everyone, it does not refer to anyone. And as it does not refer to anyone, it cannot express the sense that “I” bears on anyone’s lips. It might seem that instances of (2) will avoid this difficulty, as they will refer to someone. But suppose that (2) is instantiated for someone called “NN”: (3) On NN’s lips, “I” refers to herself. From (3) it is possible to infer: (4) On NN’s lips, “I” refers to NN. And it might seem that (4) does specify the referent of “I” on NN’s lips, through using the name “NN”. But this name cannot express the sense that “I” bears on NN’s lips. This is because no name conforms to the present explanation of “I”; insofar as “I” does conform to this explanation, no name can express the sense that “I” bears on anyone’s lips. It is possible to bring into this explanation the idea that the sense of “I” on NN’s lips is distinct from the sense of any word that does not have the indexical character determined by (2), by marking the reflexive with a star to indicate its specialness, and explaining this special reflexive as follows: for a word on anyone’s lips to refer to herself* is for it to refer to herself (ordinary reflexive) in the way that “I” does on her lips. That would yield, in place of (2): (5) On everyone’s lips, “I” refers to herself*. Which in turn would yield, in place of (4): (6) On NN’s lips, “I” refers to NN in the way that “I” does on NN’s lips. But this can no more express the sense of “I”, on NN’s lips, than (4) can. As it might be put, it contains one set of quotation marks too many:9 whereas the problem with (4) is that it uses the wrong word, the problem with (6) is that it does not use the right one. 30. In this light, Anscombe rejects the second requirement. But she does not reject the first. This contrasts with contemporary treatments of the first person, which take the second requirement for granted, and in so doing suppress the first. Its suppression is reflected in the widespread

272 Adrian Haddock tendency to think that the argument of Anscombe’s that I have just rehearsed rejects the explanation of “I” as an indexical on the grounds that this explanation is “circular”. Gareth Evans is representative here. He glosses what he takes to be her argument as follows: [Anscombe] argues that either “John refers to himself” is simply equivalent to “John refers to John” [cf. (4)], in which case the observation is wrong; or else it means something which can be elucidated only by reference to the first person pronoun [cf. (6)] in which case the observation is useless. Evans (1982: 258) Although Evans says that the problem with the second of these accounts of “John refers to himself” is that it involves reference to “I”, his way of dealing with this problem shows that he thinks that the problem is not that “I” is mentioned rather than used, but that “I” shows up in the account at all, for he goes on to give a different account of the special reflexive, which neither mentions nor uses the first person. This account claims that the truth of (5) consists in the fact that “I” is associated with a reflexive way of referring, which fixes it that the one on whose lips it refers is the one it refers to. That will satisfy the second requirement. And as such, it cannot satisfy the first. Consider what it would yield in place of (6): (7) On NN’s lips, “I” refers to NN in a reflexive way that fixes it that the one on whose lips it refers is the one it refers to. This no more shows the sense of “I”, on NN’s lips, than (4) or (6) does. Articles and even books have been devoted to defending, or trying to criticize, this explanation.10 But they all miss the point, for they all suppress the first requirement. And in suppressing it, they reveal something about themselves—that they have abandoned the project of articulating an understanding of thought and language from within: of expressing the thought, or using the element of language, which the understanding is an understanding “of”. Any explanation that treats “I” as an indexical abandons this project, for what is expressed by such an explanation instantiated for NN—such as (4), (6), or (7)—is not what NN expresses in using “I”. These instances of the explanation are all given, not from within what NN expresses in using the first person, but from outside. Anscombe takes it for granted that a philosophical understanding of thought and language is given from within. And this is why it does not occur to her to abandon the first requirement. (In this light, it is ironic that she is said to reject the explanation of “I” as an indexical because it is “circular”. The truth is that she rejects it because it is

Disjunctivism, Skepticism, and the First Person

273

not circular, as any explanation of an element of language that is given from within, and as such shows the element’s sense, must be; (1) is an example.)11 31. These contemporary treatments of the first person proceed, not from within what is expressed in using a sentence with “I” as subject, but from outside. By contrast, modern epistemology is engaged in the general project of articulating self-understanding: it seeks to proceed, not from outside what is expressed in the uses of language that express the thoughts with which it is concerned, but from within. But insofar as modern epistemology appeals to the idea of “I” as an indexical to make out its assumption that “I” is a referring term, it abandons this project at a crucial point. It purports to be articulating self-understanding; but in taking it for granted that a use of “I think p” expresses a thought over and above the thought expressed by “p”, it proceeds from outside whatever is expressed in using “I think p”. The idea of “I” as an indexical can be taken for granted only by exiting the project of articulating selfunderstanding: from within this project, this idea can have no foothold. 32. The project of modern epistemology rests on the assumption that “I” is a referring term: only given this assumption can it make good on its fundamental idea that a use of “I think p” expresses a thought over and above the thought expressed by “p”. But it seems that this idea cannot be sustained from within this project, because it is a project of articulating self-understanding. And then the project is incoherent: it takes for granted that “I” is a referring term—and yet it is unable to make sense of this assumption, by its own lights. 33. If the project is to be saved from incoherence, then it needs to explain “I” as a referring term from within what is expressed in using “I”: it needs to give an explanation of “I” that not merely assigns a referent to the word but, in assigning its referent, expresses its sense. And it is far from clear that it can do this. To bring this out, suppose that the attempt is made to do it. It is natural to think that there are two basic ways for words to refer: through being attached to something that is given to the senses in a certain way; or through concepts alone, as set against what is given. Frege thought that the second of these ways could work for numerals—but even he thought that this was an exceptional case. So, that leaves the first way. Insofar as a word refers in this way, its explanation must proceed through the use of a demonstrative—specifically, through something with the significance of: “W refers to this G”, where “W” signifies the word, and “G” signifies the kind to which the word’s putative referent is taken to belong. This kind of explanation matches an explanation of the “‘Snow’ refers to snow” variety, in that it equally seeks to show the sense of the word; but it operates at a more fundamental level, in that it seeks to show the sense that the word bears in its most basic attachment to reality, in which its sense matches that of a demonstrative.

274 Adrian Haddock As the use of a word develops, so does its sense, as the word frees itself from its basic attachment to the real, by being used with the same referent outside of situations in which its referent is available to be demonstrated. But the potentiality of explaining a word as a referring term through using a demonstrative belongs to any word whose reference is secured directly through being attached to something that is given to the senses in a certain way. For this reason, any explanation of the present form will show at least an aspect of the sense of the word that it explains. But there is no prospect of bringing “I” within an explanation of this form. This is because any explanation of this form presupposes that something is given to the senses in a certain way. And any way in which an object is given to the senses has a first-person character, in that it is expressible by saying something with “I” in subject-position, paradigmatically something beginning “I perceive . . .”—such as “I perceive this G”. But then there is a use of “I” that any explanation of this form presupposes and, as such, cannot explain. The explanation’s inability to explain this use comes out in the possibility of asking, with respect to something given: “Am I this thing”? Or even, if the given object is called “I”: “Am I I?” In the first of these questions, and in its first appearance in the second question, the word “I” is figuring in the use that the explanation presupposes. And the sheer possibility of asking these questions shows that no explanation of this form is capable of explaining this use: insofar as the questions can be asked, what “I” expresses in this use cannot be the sense of any expression whose reference is explicable through the use of a demonstrative. Insofar, then, as “I” is not equivocal, it is not possible to explain “I” as an expression that refers through being attached to something that is given to the senses in a certain way. 34. Modern epistemology cannot make good on its assumption that “I” is a referring term. And as it cannot, it cannot descend from the general level at which the project of epistemology was first introduced. But the very idea of the project of modern epistemology is the idea of an epistemological project that makes this descent. So, the project of modern epistemology lapses, and melts away. But the project of epistemology remains—at the most general level. And at this level, the project says that it is possible to know the truth of a thought that is capable of falsehood only insofar as it is possible to express the thought in a manner that reveals it to be true. That might still look as hopeless as ever, insofar as the thought is capable of truth and falsity. So, it might not be clear what consequence the demise of the modern project has for skepticism. 35. But as the assumption that “I” is a referring term fades away, a hitherto suppressed insight in disjunctivism comes to light. When that assumption was in place, it seemed clear that a use of “I perceive p” expresses a thought that goes beyond but contains the thought expressed by “p”. Short of this assumption, however, the only thought that is

Disjunctivism, Skepticism, and the First Person

275

expressed by a use of “I perceive p” is the thought expressed by “p”: a use of “I perceive p” does not express a further thought that must itself be revealed to be true, in order to reveal the truth of the thought expressed by “p”. And that removes an obstacle to understanding “I perceive p’” as fit, not merely to express the thought expressed by “p”, but to reveal its truth. 36. A popular word in contemporary epistemology is “factivity”. And it is said that the locution “I perceive p” is “factive”. But there are two things that it could mean to say this locution is “factive”. It could mean that a use of “I perceive p” expresses a thought, over and above the thought expressed by “p”, whose truth entails the truth of the thought expressed by “p”. And a lesson of skepticism is that “factivity”, in this sense, is of no significance for the project of epistemology. But that is not the only thing it could mean to say this. It could mean that a use of “I perceive p” expresses nothing over and above the thought expressed by “p”, but in a manner that reveals this thought to be true. A use of “I perceive p” could not perform this revelatory function if it expressed a further thought over and above the thought expressed by “p”, because then it would be necessary to use a further sentence to reveal the truth of this further thought in turn. But as it does not express any further thought, there is no obstacle to its performing this function. And this is of real epistemological significance. 37. It is crucial that the revelation of the truth of the thought expressed by “p” is had in using “I perceive p”. It is tempting to think that it is possible to think of someone who is using this last sentence, but without using the sentence oneself. And now it might seem obvious that her assertion of the sentence might be false, because the thought expressed by “p” might be false. But that is to proceed from outside what she expresses in using the sentence, because it is not itself to use the sentence; and that is not the way of proceeding that characterizes the project of epistemology. Although modern epistemology is wrong to assume that “I” is a referring expression, it is right to assume that epistemological reflection has a first-person character. To engage in this reflection is not to think of someone who is using a sentence beginning with “I”; it is to use that sentence oneself. And to use the sentence “I perceive p” is not to think of someone in some way (through the word “I”), and to say something of her (through the words “perceive p”). It is not to think anything of anyone—or at least, of anyone other than the one signified in “p” (if anyone is). To use “I perceive p” is to reveal the truth of the thought it expresses, which is nothing other than the thought expressed by “p”. 38. This then is the insight in disjunctivism: there is a sentence the use of which does reveal the truth of the thought it expresses, even though this thought is capable of being false. And “I perceive p”—more

276 Adrian Haddock generally: “I know p”—is an example of such a sentence. That insight, however, cannot come into focus when the assumption that “I” is a referring expression is in place. Modern epistemology takes this assumption for granted—but it cannot make good on it from within its project. Once this is seen, and the assumption abandoned, the insight in disjunctivism will emerge. And once it has emerged, skepticism will fall away. It is possible to express a thought that is capable of being false, but in a manner that reveals its truth. 39. As skepticism dissolves, it is possible to understand more fully what the KK thesis comes to. The thesis says that if someone knows something, then she knows that she* knows it. And now that the project of modern epistemology has lapsed, it is clear that the self-knowledge reported by the consequent does not ascribe a property to a “self”, but consists merely in the capacity to express the thought that constitutes the content of the knowledge reported by the antecedent in a manner that reveals it to be true. What has emerged from the present reflection is that, insofar as the only thought expressed by a factive locution such as “I perceive p”—or more generally, “I know p”—is the thought expressed by “p”, this capacity consists in the capacity to use such a locution. At the most general level, knowledge is the capacity to express the thought expressed by “p” in a manner that reveals it to be true; and this capacity just is the capacity to say: “I know p”. As skepticism falls away, it becomes clear that, at the most general level, the KK thesis comes to the following: if someone knows the truth of a thought expressed by “p”, then she can express this thought by saying “I know p”. And in this light, it becomes equally clear that the idea of knowledge is the idea of self-knowledge, not merely because knowledge consists in the capacity to express a thought in a manner that reveals it to be true, but because this is the capacity to use a factive locution that contains the first person. Knowledge is a selfconscious capacity, because it is the capacity to express a thought in this manner, through using such a locution. And it is fallible, insofar as this thought is capable of being false.12 * 40. The result of the present reflection is that it is possible to know the truth of thoughts that are capable of being false. And that is enough to refute Cartesian skepticism. It is another matter whether a specific thought that is capable of being false is—in fact—true. But knowledge of the truth of such thoughts is not something that philosophy can yield. Knowledge of the truth of such thoughts can be acquired only through  knowledge of the truth of such thoughts: it cannot be acquired through knowledge of the truth of thoughts of some other kind (if such there be). And philosophy does not itself consist in knowledge of the truth of thoughts that are capable of being false. But it can remove

Disjunctivism, Skepticism, and the First Person

277

obstacles that stand in the way of seeing how it is possible to know the truth of such thoughts. That is what this essay has sought to do.

Notes * I am deeply grateful to Andrea Kern for many illuminating conversations. I would also like to thank Jennifer Hornsby for providing me with very helpful comments, and the Alexander von Humboldt Foundation for awarding me a Humboldt Research Fellowship for Experienced Researchers, during which this essay was completed. 1. John McDowell, “Criteria, Defeasibility, and Knowledge” (1982), Proceedings of the British Academy, vol. 68 (1982): 455–479, reprinted with revisions, additions, and subtractions, in Jonathan Dancy, ed., Perceptual Knowledge (Oxford: Oxford University Press, 1988), p. 455. (To my knowledge, the passages cited here appear only in this modified reprint of McDowell’s essay.) 2. See, in particular, McDowell (2005, this volume). For a development of disjunctivism as a response, not merely to Cartesian skepticism, but equally to ancient, and to Kantian forms of skepticism, see Kern (2017). 3. I owe the device of the star to Castañeda (1966). 4. So understood, KK is not a target of the “anti-luminosity” argument. This argument rests on the principle that if at one time someone knows X, then at a very slightly later time X equally obtains. Whatever plausibility this principle has rests on the idea that someone’s knowing X, and X, are distinct conditions: it would be absurd to claim that if at one time X obtains then at a very slightly later time X equally obtains. And it is an assumption of the argument that someone’s knowing that she* knows something, and her knowing it, are examples of someone’s knowing X, and X respectively, and as such are themselves distinct conditions. But properly understood, KK says that someone’s knowing that she* knows something, and her knowing it, are not distinct conditions at all. Her knowing that she* knows something, and her knowing it, are each simply a matter of her having the capacity to express a thought in a manner that reveals it to be true. (For the “anti-luminosity” argument, see Williamson 2000). 5. For a splendid discussion, see Anscombe (2015). 6. Kant (2001: B168). 7. These include Greco (2004) and Millar (2008). 8. The position they defend is merely a verbal variant on the position of Williamson, op. cit. (to which the same judgment applies). 9. I am grateful to Christian Kietzmann for this way of putting things. 10. The most significant examples are Rumfitt (1994) and Doyle (2018). Rumfitt defends the explanation, whereas Doyle tries to criticize it by defending what he calls Anscombe’s “circularity argument”. It is no surprise that Doyle cannot sustain this criticism and is led to conclude that there “does not seem to be any important consideration yet adduced [by Anscombe or Rumfitt] that would decide the issue [between them] either way” (130). There could not be any such consideration, because Anscombe does not advance a “circularity argument”, and Rumfitt does not engage with her actual argument, because he defends the explanation. 11. I am grateful to Sebastian Rödl for this way of putting things. 12. This is the core of the idea of knowledge as a fallible, self-conscious capacity, and it is sufficient for the purpose of bringing out how disjunctivism bears on Cartesian skepticism. There might be other purposes for which the idea needs further development; and for a further development, see Kern (2017).

278 Adrian Haddock

References Anscombe, Gertrude Elizabeth Margaret, 1975 [1981], The first person, reprinted in her Metaphysics and the Philosophy of Mind: Collected Papers (vol. II). Oxford: Basil Blackwell. ———, 2015, Private ostensive definition, in Mary Geach and Luke Gormally, eds. Logic, Truth, and Meaning: Writings by G.E.M. Anscombe. Exeter: Imprint Academic. Castañeda, Hector-Neri, 1966 [1981], He: A study in the logic of self-consciousness, in his The Phenomeno-Logic of the I: Essays on Self-Consciousness. Indianapolis: Indiana University Press. Descartes, René, 1996, Meditations on First Philosophy. Ed. and trans. John Cottingham. Cambridge: Cambridge University Press. Doyle, James, 2018, No Morality, No Self: Anscombe’s Radical Skepticism. Cambridge, MA: Harvard University Press. Evans, Gareth, 1982, The Varieties of Reference. Ed. John McDowell. Oxford: Clarendon Press. Greco, John, 2004, Externalism and skepticism, in Richard Schanz, ed. The Externalist Challenge. Berlin: De Gruyter. Kant, Immanuel, 2001, Kritik der reinen Vernunft. Trans. Paul Guyer and Allen Wood. Cambridge: Cambridge University Press. Kern, Andrea, 2006 [2017], Sources of Knowledge: On the Concept of a Rational Capacity for Knowledge. Trans. Daniel Smyth. Cambridge, MA: Harvard University Press. McDowell, J., 1982 [1988], Criteria, defeasibility, and knowledge, Proceedings of the British Academy 68: 455–479, reprinted in Perceptual Knowledge. Ed. Jonathan Dancy. Oxford: Oxford University Press. ———, 2005 [2009], The disjunctive conception of experience as material for a transcendental argument, in his The Engaged Intellect. Cambridge, MA: Harvard University Press. ———, this volume, Perceptual experience and empirical rationality, Millar, Alan, 2008, Disjunctivism and skepticism, in John Greco, ed. The Oxford Handbook of Skepticism. Oxford: Oxford University Press. Rumfitt, Ian, 1994, Frege’s theory of predication: An elaboration and defense, with some new applications, The Philosophical Review 103(4): 599–637. Williamson, Timothy, 2000, Knowledge and Its Limits. Oxford: Oxford University Press.

Part IV

Disjunctivism in Other Domains

14 Two Forms of Memory Knowledge and Epistemological Disjunctivism Joe Milburn and Andrew Moon

Recent discussion of epistemological disjunctivism has centered on perceptual knowledge. For instance, Duncan Pritchard describes the core claim of epistemological disjunctivism as follows: Core Thesis In paradigmatic cases of perceptual knowledge an agent, S, has perceptual knowledge that Φ in virtue of being in possession of rational support, R, for his belief that Φ which is both factive (i.e., R’s obtaining entails Φ) and reflectively accessible to S. (Pritchard 2012: 13)1 Earlier presentations of epistemological disjunctivism, however, concerned not only perceptual knowledge, but knowledge of other minds and testimonial knowledge (cf. McDowell 1998a, 1998b). In fact, John McDowell’s arguments for epistemological disjunctivism involve considerations about the nature of knowledge as such and do not depend upon anything particular to perceptual knowledge (Cf. McDowell 1998a, 1998b, 1998c). If his arguments are correct, then we should accept disjunctivist accounts of any kind of knowledge in which it makes sense to speak of “appearances” deceiving us through no fault of our own. Of course, many philosophers have not been moved by McDowell’s arguments. In this way, Pritchard’s treatment of epistemological disjunctivism offers a new approach. By making epistemological disjunctivism a claim exclusively about perceptual knowledge, Pritchard is able to argue for it by appealing to special features of perceptual knowledge. Given plausible assumptions about the nature of perceptual knowledge, Pritchard’s Core Thesis becomes quite attractive. In this chapter, we will push the discussion forward by evaluating the possibility of establishing an epistemological disjunctivist account of memory knowledge along Pritchardean lines. We argue that, mutatis mutandis, the case that Pritchard makes for epistemological disjunctivism regarding perceptual knowledge can also be made for a certain type of memory knowledge. We also argue that the prospects for a disjunctivist

282

Joe Milburn and Andrew Moon

account of another form of memory knowledge are bad. The final result is that Pritchardean disjunctivism is a live option for one kind of memory knowledge but not another kind.2 The chapter proceeds as follows. In section 1, we distinguish between two different types of memory knowledge: experiential memory knowledge and stored memory knowledge. In section 2, we defend a parity thesis. According to it, arguments analogous to Pritchard’s for epistemological disjunctivism regarding perceptual knowledge serve equally well for epistemological disjunctivism regarding experiential memory knowledge. In section 3, we argue that such a parity thesis fails for a disjunctivist account of stored memory knowledge.

1. Memory Knowledge: Experiential and Stored In this section, we’ll clarify our terms. Very broadly, our topic is propositional knowledge, that is, knowledge that p. (From here on, any use of “knowledge” will refer to propositional knowledge unless otherwise stated.) More specifically, our topic is memory knowledge. We’ll discuss two types.3 Suppose that John is trying to remember where he met Lorraine. At first, he cannot remember where they met. He then has a clear memory of meeting her at the Hesburgh Library. He then comes to believe that he met Lorraine at the Hesburgh Library on the basis of this memory of meeting her there. The memory, which involves an experience or image, provides the grounds for John’s knowledge. We will call this experiential memory knowledge. Contrast this with stored memory knowledge. Even while John is focused on his breakfast, he continues to know that his name is John. This knowledge is stored in his memory. Note that this knowledge does not depend on any kind of memory image or experience; he isn’t having any memory experiences as he devours his waffles. This example distinguishes stored memory knowledge from experiential memory knowledge. Before moving on, we will distinguish between three forms of memory, which can be picked out linguistically. Consider: 1. John remembers the 13th floor of the Hesburgh Library. 2. John remembers reading philosophy books on the 13th floor of the Hesburgh Library. 3. John remembers that the philosophy books were held on the 13th floor of the Hesburgh Library. In sentences 1 and 2, the verb “remember” takes as its complement a direct object. In sentence 1, the direct object is a kind of physical object, the thirteenth floor of the Hesburgh Library. In sentence 2, the direct object is an event: reading philosophy books on the 13th floor of the

Two Forms of Memory Knowledge

283

Hesburgh Library. We can call these cases of remembering a case of object memory and a case of event memory respectively. In sentence 3, on the other hand, the verb “remember” takes a that-clause for its complement. Here John is described not as having object or event memory but as having propositional memory. These three types of memory will help us evaluate the prospects of a disjunctivist account of memory knowledge.

2. Experiential Memory Knowledge and the Parity Thesis 2.1 Pritchard on the Core Thesis In this section, we consider the prospects for establishing a disjunctivist account of experiential memory knowledge along Pritchardean lines. Recall his Core Thesis: In paradigmatic cases of perceptual knowledge an agent, S, has perceptual knowledge that Φ in virtue of being in possession of rational support, R, for his belief that Φ which is both factive (i.e., R’s obtaining entails Φ) and reflectively accessible to S. Pritchard (2012: 13) Now consider: E-Memory Core Thesis In paradigmatic cases of experiential memory knowledge, S has experiential memory knowledge that p in virtue of being in possession of rational support, R, for his belief that p, which is both factive (i.e., R’s obtaining entails p) and reflectively accessible to S. This claim is the equivalent of Pritchard’s Core Thesis, adapted to the case of experiential memory knowledge. Are there good grounds for accepting this claim? We think there are. Or rather, we think that arguments analogous to Pritchard’s for the Core Thesis can be given with equal effect for the E-Memory Core Thesis. Thus, if Pritchard’s arguments are sufficient to establish epistemological disjunctivism regarding perceptual knowledge, these analogous arguments should be sufficient for establishing epistemological disjunctivism regarding experiential memory knowledge. We will call this claim the parity thesis. To prove this thesis, we must first consider Pritchard’s case for the Core Thesis. That is the topic of this subsection. As Pritchard puts it, he aims to [motivate] epistemological disjunctivism [i.e., the Core Thesis] by showing that this is an attractive position which we should want to hold if it were theoretically available, and furthermore showing

284

Joe Milburn and Andrew Moon that it is theoretically available, contrary to prevailing conventional wisdom in epistemology. (2012: 18; emphasis in original)

Why should we want to hold epistemological disjunctivism if it is theoretically available? Pritchard provides three reasons. First, he notes that if the Core Thesis were true, it would allow us to understand paradigmatic cases of perceptual knowledge in a way that respected both internalist and externalist intuitions. On the one hand, paradigmatic cases of perceptual knowledge would be grounded on reflectively accessible rational support. Whatever grounds we have for our belief would not be beyond our ken. This would respect internalist intuitions. On the other hand, paradigmatic cases of perceptual knowledge would be grounded on factive, rational support. And this would respect externalist intuitions that stress the necessary connection between truth conducive belief formation and knowledge (cf. Ibid.: 3–5). Second, Pritchard notes that it is common practice to justify one’s perceptual belief that p (or one’s claims to perceptual knowledge that p) by appealing to the fact that one sees that p. Given this, the default should be to think of our perceptual knowledge that p as being rationally supported by our seeing that p. If this is correct, then we have strong prima facie grounds for accepting the Core Thesis, for we cannot see that p unless p is true; furthermore, it seems that in paradigmatic cases of perceptual knowledge that p, we can know on the basis of reflection alone that we see that p (Ibid.: 17). Third, Pritchard holds that epistemological disjunctivism allows us to deal with radical skepticism in a satisfying way.4 In particular, epistemological disjunctivism provides an undercutting response to the underdeterminationbased skeptical paradox. This paradox arises when we accept the following three claims. U1. One cannot have rational support that favors one’s belief in an everyday proposition over an incompatible radical skeptical hypothesis. U2. If S knows that p and q describe incompatible scenarios, and yet S lacks a rational basis that favors p over q, then S lacks rationally grounded knowledge that p. U3. One has widespread rationally grounded knowledge (cf. Pritchard 2016: 133–134). Accepting epistemological disjunctivism gives us grounds for rejecting U1. Suppose that I have perceptual knowledge that I have hands. Given the Core Thesis, I have this knowledge in virtue of having factive, reflectively accessible rational support for my belief that I have hands. But given that this rational support is factive, it obviously favors believing that I have hands over the skeptical hypothesis that I am a handless brain in a vat (cf. Ibid.: 133–134).

Two Forms of Memory Knowledge

285

Pritchard’s view is that these considerations provide prima facie reasons for accepting epistemological disjunctivism. So, unless we have good grounds for rejecting epistemological disjunctivism, we should accept it as true. Of course, many epistemologists think that we have good grounds for rejecting epistemological disjunctivism. Pritchard’s strategy is to survey the most pressing problems for epistemological disjunctivism and show that, after all, these are not sufficient grounds for rejecting the position. The problems that Pritchard identifies are the access problem, the distinguishability problem, and the basis problem (cf. Pritchard 2012: 19–22, 2016: 127–132). In brief, the access problem arises from the claim that we have perceptual knowledge by virtue of having rational support that is reflectively accessible. But if we can know by reflection alone that we see that p, and by reflection alone that if we see that p then p, then it seems that we could know by reflection alone that p, where p is some empirical fact. Thus, the Core Thesis implies that we can by reflection alone have empirical knowledge, which is incorrect. The distinguishability problem starts from the idea that we are not able to distinguish illusory or hallucinatory states from states of seeing that p by reflection alone. But if the Core Thesis is true, then this shouldn’t be the case; we are able to know by reflection alone that we see that p. And if this is true, then we should be able to deduce that we are not in an illusory or hallucinatory state. Thus, contrary to the initial hypothesis, we can distinguish by reflection alone illusory or hallucinatory states from states of seeing that p. Finally, the basis problem arises because it seems that to see that p is not only factive, but epistemic. That is, it seems that seeing that p is merely a way of knowing that p. But if seeing that p is a way of knowing that p, then it is hard to see how one could know that p in virtue of the rational support one has from seeing that p. In this case, one would know that p in virtue of the rational support that one has from knowing that p, which seems absurd. In the following two subsections, we will consider Pritchard’s responses to these problems. 2.2 Parity Between the Core Thesis and the E-Memory Core Thesis Our claim is that arguments analogous to Pritchard’s for the Core Thesis can be given with equal effect for the E-Memory Core Thesis. To see this, start by considering the grounds we possess for finding the E-Memory Core Thesis theoretically attractive. The E-Memory Core Thesis respects internalist and externalist intuitions to the same extent that the Core Thesis does. It claims that just as we have factive, reflectively accessible rational support for our beliefs in paradigmatic cases of perceptual knowledge, we also have it in paradigmatic cases of experiential memory knowledge. If the E-Memory Core Thesis is true, then

286

Joe Milburn and Andrew Moon

in paradigmatic cases of experiential memory knowledge our rational support for our beliefs is not beyond our ken, respecting internalist intuitions. At the same time, our rational support is factive, respecting externalist intuitions that knowledge must be tied to truth conducive belief formation. Likewise, common practice supports the E-Memory Core Thesis in the same way that common practice supports the Core Thesis. Suppose that you have left your office for lunch, and the question arises as to whether you have locked your office door. (One of your colleagues wants to go in and borrow a book from your desk, but you won’t be returning to your office for a couple of hours.) You have purposely left your door unlocked so that your colleague can go by and collect the book, but your colleague has some doubts. (She believes that you are a creature of habit and always lock your door.) In response to your colleague, it would be perfectly natural to say something like the following: “I distinctly remember that I left the door unlocked because I wanted you to be able to get the book.”5 Part of our common practice in justifying our beliefs about our past actions and experiences is to appeal to the fact that we remember that we did such and such. Furthermore, “remember that”, just like “see that”, is factive;6 i.e., you cannot remember that p, unless p. You cannot remember that you left the door unlocked, unless as a matter of fact you left the door unlocked. On top of this, in paradigmatic cases of experiential memory knowledge, the fact that one remembers that p is reflectively accessible. Take for example one’s experiential memory knowledge that one did not lock one’s office door. Presumably, in such a case, one can know that one remembers this without obtaining further information (e.g. double checking that the door is in fact unlocked). Rather, one’s knowledge that one remembers that one left the door unlocked is the result of reflection. If this is correct, then we have strong prima facie grounds for accepting the E-Memory Core Thesis. Finally, the E-Memory Core Thesis gives us grounds for rejecting certain forms of radical skepticism about the past. Consider the following Underdetermination-based skeptical paradox regarding our past. U1. One cannot have rational support that favors one’s belief in an everyday proposition about their personal history over an incompatible radical skeptical hypothesis (e.g. that the world is five minutes old, and all of one’s memories of events prior to this time are illusory). U2. If S knows that p and q describe incompatible scenarios, and yet S lacks a rational basis that favors p over q, then S lacks rationally grounded knowledge that p. U3. One has widespread rationally grounded knowledge about their personal history.

Two Forms of Memory Knowledge

287

Given the E-Memory Core Thesis, if I have paradigmatic experiential memory knowledge that I bicycled to work this morning, then I know this in virtue of having factive, reflectively accessible rational support for my belief. And since this rational support is factive, it obviously favors my believing that I bicycled to work this morning over the skeptical hypothesis that the world is five minutes old.7 In these ways, then, we can see that there is a parity between Pritchard’s Core Thesis and the E-Memory Core Thesis. The same kinds of considerations that provide prima facie reason for accepting the Core Thesis also provide prima facie reason for accepting the E-Memory Core Thesis. This brings us to the question of whether the grounds for rejecting the E-Memory Core Thesis and the Core Thesis are the same. To answer this question, we start by looking at the problems facing the Core Thesis: the access problem, the distinguishability problem, and the basis problem. Plausibly, all three of these problems for the Core Thesis have analogues for the E-Memory Core Thesis. So, for example, we can formulate a version of the access problem as follows: if the E-Memory Core Thesis is correct, then it seems that one could, on the basis of reflection alone, know that some contingent event took place in the past; but this is false. Therefore, the E-Memory Core Thesis is false. Likewise, we can formulate a version of the distinguishability problem as follows: prima facie, we cannot distinguish between states of remembering that something is the case (veridical memory experiences), from illusory or hallucinatory memory experiences. But if the E-Memory Core Thesis is true, then we should be able to do so. Therefore, the E-Memory Core Thesis is false. Finally, we can formulate a version of the basis problem in the following terms. Remembering that p is merely a way of knowing that p (e.g. remembering that one bicycled to work this morning is merely a way of knowing that one bicycled to work this morning). But if remembering that p is a way of knowing that p, then it is hard to see how one could know that p in virtue of the rational support one has from remembering that p. In this case, one would know that p in virtue of the rational support that one has from knowing that p, which seems absurd. Our claim is that while there are analogues to the access, distinguishability, and basis problems for the E-Memory Core Thesis, these problems are no more of a barrier for accepting the E-Memory Core Thesis than the original problems are for accepting the Core Thesis. Consider first the access problem and the distinguishability problem. They arise for the Core Thesis, not because it is a thesis about perceptual knowledge in particular, but because it claims that we can have factive and reflectively accessible reasons for believing contingent truths. The E-Memory Core Thesis makes a similar claim, which is why versions of the access problem and the distinguishability problem arise for the E-Memory Core Thesis. But just as these problems are not particular problems for perceptual

288

Joe Milburn and Andrew Moon

knowledge, neither are Pritchard’s solutions to these problems particular to perceptual knowledge. For example, consider Pritchard’s reply to the access problem. He argues that all that the Core Thesis implies is that one can have knowledge that one sees that p by reflection alone in just those cases in which one already has paradigmatic perceptual knowledge that p. But this claim is not so absurd since it does not allow that one could come to know that p on the basis of reflection alone (cf. Pritchard 2012: 50–52, 2016: 130). Mutatis mutandis, this same solution could be given to the problem as it arises for experiential memory knowledge. The E-Memory Core Thesis only implies that one can have reflective knowledge that e.g. one bicycled to work today, when one already has experiential memory knowledge that one bicycled to work today. It does not allow that one could come to know that one bicycled to work today on the basis of reflection alone. Likewise, Pritchard’s way of dealing with the distinguishability problem relies on nothing that is particular to perceptual knowledge. Simplifying things, we can understand Pritchard’s response to the distinguishability problem as follows. He grants that we cannot discriminate between cases of seeing that p from cases of illusion and hallucination on the grounds given to us by perception. We are, after all, liable to fall into the grips of an illusion. But this does not imply that in good cases we cannot know that we see that p instead of merely seeming to see that p. While it is true that we cannot discriminate between cases of seeing that p and cases of illusion and hallucination just on perceptual grounds, we can in some sense distinguish between these states (cf. Pritchard 2012: 96; Pritchard 2016: 131–132). Mutatis mutandis, the same solution can be given to the distinguishability problem for experiential memory knowledge. It can be admitted that we cannot discriminate between veridical memory experiences from non-veridical experiences on the mere basis of our memory experiences, but we can, in the good case, know that we are in the good case and not in the bad one. 2.3 A Potential Disparity: The Basis Problem Things are slightly different with the basis problem. The original basis problem for the Core Thesis arises because the following seems to be true. Visual Entailment Thesis If S sees that p, then S knows that p. It is possible that the basis problem arises for the Core Thesis because of particularities regarding the notion of “seeing that p”. The basis problem for the E-Memory Core Thesis arises because the following seems true.

Two Forms of Memory Knowledge

289

Memory Entailment Thesis If S remembers that p, then S knows that p. As a result, it is possible that the basis problem for the E-Memory Core Thesis arises from particularities regarding the notion of “remembering that” that are not present with “seeing that”. If this is true, then solutions to the basis problem for the Core Thesis might not work for the E-Memory Core Thesis. Despite this abstract possibility, it seems that responses to the basis problem for the E-Memory Core Thesis, in fact, can run parallel to responses to the basis problem for Pritchard’s Core Thesis. So, for example, Pritchard tries to solve the basis problem by attacking the visual entailment thesis. He thinks there are cases in which one sees that p without knowing that p. His cases involve a person whose perceptual faculties are working properly in the appropriate sort of environment (and so one sees that p), but the person has either the false belief that their faculties are not working properly or the false belief that they are not in the appropriate environment. The person would thereby fail to know that p (see Pritchard 2012: 26–28, 2016: 127–128). A similar strategy could be used to solve the basis problem for the E-Memory Core Thesis. Bernecker (2010: 74–88) thinks there are cases in which one remembers that p without knowing that p. His cases involve a person whose memory faculties produce a thought that p (and so one remembers that p) but the person fails either to believe p, to have a justified belief that p, or to not be in a Gettier situation with respect to p. He would thereby fail to know that p. Now, despite the broad similarities in Pritchard’s and Bernecker’s arguments against the respective entailment theses, we admit that differences in the details allow for potential disparities.8 Does this undermine our parity thesis? No, because we will show in the following paragraphs that a solution to the two basis problems does not depend on the falsity of the entailment theses. The basis problem for the Core Thesis assumes that paradigmatic rational support for visual knowledge that p is seeing that p. However, as Craig French (2016) insightfully notes, Pritchard’s Core Thesis is silent about what the rational support is. While Pritchard himself holds that seeing that x is F provides the rational support for believing that x is F, this is not the only plausible candidate. Perhaps, as French suggests, it is not seeing that x is F, but seeing an F-object. Suppose, using French’s example, you have paradigmatic perceptual knowledge that a particular lemon is yellow. On this alternative view, you have this knowledge in virtue of the factive, reflectively accessible rational support you receive from seeing a yellow lemon. Likewise, in the case of perceptual knowledge of an event occurring, perhaps it is not seeing that x is A-ing, but seeing x A-ing that provides one’s rational support. Suppose you have

290

Joe Milburn and Andrew Moon

paradigmatic perceptual knowledge that someone is running down your hallway. On the view under examination, you have factive, reflectively accessible rational support from seeing someone running down your hallway. Seeing an F-object and seeing an x A-ing do not entail knowing, respectively, that x is F and that x is A-ing, and so the basis problem is avoided. One can see a yellow lemon without knowing that the lemon is yellow; likewise, one can see someone running down their hallway without knowing that someone is running down their hallway. This view is in line with the common practice of justifying one’s beliefs and claims to knowledge by appealing to the fact that one has seen the relevant kind of object or event. How do I know that a particular lemon is yellow? A natural response is to say that I see it (the yellow lemon). How do I know that someone is running down my hallway? A natural response is to say that I see someone running down my hallway. The view fits with common linguistic practice. This view also supports internalist intuitions. It is plausible that the rational support for our paradigmatic cases of perceptual knowledge— e.g. that we see a yellow lemon—can be known by reflection alone. The view also supports externalist intuitions. On this view, it is plausible that seeing an F-object entails that there is an F-object, and seeing an x A-ing entails that that there is an x A-ing. But if this is the case, then we can understand seeing an F-object as being factive, in the sense that if seeing an F-object is one’s rational support for believing that there is an F-object, then one’s rational support guarantees the truth of one’s belief.9,10 Now, the E-Memory Core Thesis also fails to specify exactly what provides factive, reflectively accessible rational support in paradigmatic cases of experiential memory knowledge. Our claim is that a view analogous to French’s is available for a disjunctivist account of experiential memory knowledge. Just as it is possible to see a yellow lemon or to see someone running down one’s hall, it is also possible to remember a yellow lemon (object memory) or to remember someone running down one’s hall (event memory). And just as it is common practice to appeal to seeing F-objects in order to justify the relevant belief, so too it is common practice to appeal to remembering F-objects in order to justify the relevant belief. So, I might justify the claim that a particular lemon was yellow by saying that I remember a yellow lemon; likewise, I might justify the claim that someone was running down my hallway by saying that I remember someone running down my hallway. So, instead of appealing to propositional memory as what provides rational support for our experiential memory knowledge, perhaps we should instead appeal to object memory and event memory. Plausibly, in paradigmatic cases of experiential memory knowledge, one can know by reflection alone that one remembers the relevant object or event. That is, in paradigmatic cases of experiential memory knowledge, one does not need more information than what one already possesses to know that one remembers the object or event. So, for example, in the

Two Forms of Memory Knowledge

291

case in which you have experiential memory knowledge that you left your door unlocked for your colleague, you don’t need to receive further information in order to know that you remember leaving your door unlocked. So this view allows that the rational support we have in paradigmatic cases of experiential memory knowledge is reflectively accessible. This view affirms the factivity component of the E-Memory Core Thesis. Plausibly, one cannot remember a yellow lemon unless the lemon was yellow; one cannot remember someone running down their hallway, unless someone was running down their hallway. So object and event memory are relational. In this way, they can also serve as factive rational support for the relevant belief.11 2.4 The Parity Thesis Vindicated We have strong grounds for accepting the parity thesis. First, every kind of consideration that gives us prima facie reason for accepting the Core Thesis also gives us prima facie reason for accepting the E-Memory Core Thesis. Second, the analogous problems for the Core Thesis and the E-Memory Core Thesis can be solved in analogous ways. Of course, this does not show that there are no additional problems that arise only for the E-Memory Core Thesis. To see this, consider that Pritchard hasn’t proven that the only possible problems for the Core Thesis are the access, distinguishability, and basis problems. It is just that there are no further obvious problems for the view. As a result, we take it that the parity thesis stands. To sum up, the analogies between seeing and remembering are such that we should expect the Pritchardean case for a disjunctivist account of experiential memory knowledge to be as plausible as the Pritchardean case for a disjunctivist account of perceptual knowledge.

3. Stored Memory Knowledge and the Parity Thesis 3.1 Stored Memory Knowledge Having discussed the prospects of establishing a disjunctive account of experiential memory knowledge on Pritchardean lines, we now consider the prospects for a disjunctivist account of stored memory knowledge. Consider the following claim. S-Memory Core Thesis In paradigmatic cases of stored memory knowledge an agent, S, has stored memory knowledge that p in virtue of being in possession of rational support, R, for his belief that p which is both factive (i.e., R’s obtaining entails p) and reflectively accessible to S.

292

Joe Milburn and Andrew Moon

We think that there are good grounds for rejecting any parity thesis between the S-Memory Core Thesis and the Core Thesis. Whereas there are no obvious undealt-with objections to either the Core Thesis or the E-Memory Thesis, there are serious undealt-with objections to the S-Memory Thesis. In fact, we think these objections show the S-Memory Thesis to be false. Stored memory knowledge is typically derived from another source, such as perception, testimony, deductive inference, experiential memory, etc. By perception, Michael comes to know that snow is white. Later on, even when his mind is focused on other things, he continues to have stored memory knowledge that snow is white. By testimony, Sam comes to know that everyone is meeting to read a paper next Friday. Later in the day, even when his mind is focused on other things, he has stored memory knowledge that everyone is meeting to read a paper next Friday. A person’s stored memory knowledge will either be noninferential or inferential. If it is inferential, it will be either deductive inference or nondeductive inference. We will consider whether, in each of these three options—noninferential, deductive inference, and nondeductive inference—a person’s stored memory knowledge has factive, reflectively accessible support. The result will be that only in one of the options— deductively inferred, stored memory knowledge—can there be factive, reflectively accessible support. 3.2 Noninferential Stored Memory Knowledge Let us begin by considering noninferential, stored memory knowledge. Such knowledge often has its origin in a non-memorial source. For example, if pressed as to why he believes that snow is white, Michael is likely to appeal to his perception; he has seen that snow is white, or he has seen white snow. He might also appeal to his memory; he remembers that snow is white, or he remembers white snow. So, he will tell his interlocutor that he believes that snow is white either because he has seen that snow is white (or has seen white snow) or because he remembers that snow is white (or snow being white). Here is the problem.12 It seems that there is no good candidate for what will be the reflectively accessible, factive rational support for this stored memory knowledge. Since Michael does not currently see the white snow, he cannot appeal to his seeing the white snow or his seeing that snow is white; neither seeing can be his reflectively accessible, factive rational support. They are factive, but they are not reflectively accessible. The rational support must therefore be either his remembering snow being white or his remembering that snow is white. But these are also not good candidates. Note that much of our stored knowledge is unconscious; Michael might even be taking a nap! During his dreamless nap, Michael is not remembering snow being white. This is because Michael is not having any experiences as he dreamlessly sleeps; he isn’t

Two Forms of Memory Knowledge

293

remembering anything at the moment. Neither can Michael’s rational support be his remembering that snow is white. Remembering that p entails belief that p; belief is a component of the remembering.13 Hence, given the plausible thesis that rational support is irreflexive, remembering that snow is white cannot be the rational support for believing that snow is white.14 Notice that what we have said about stored memory knowledge with an original source in perception will apply to other sources of basic belief: introspection, a priori intuition, moral intuition, etc. I may first come to believe I feel pain, or 1 + 1 = 2, or gratuitous suffering is bad by way of “seeing” these facts. Later, I continue to know these facts when I dreamlessly sleep. But there seems to be no reflectively accessible, factive rational support for these beliefs at these later times. Given this, the S-Memory Core Thesis fails. We have paradigmatic instances of stored memory knowledge that we do not have in virtue of having factive, reflectively accessible rational support. 3.3 Inferential Stored Memory Knowledge Let us consider a case of deductive inference. Suppose Chris knows via testimony that Joyce is either at the party or at home; he then learns via perception that Joyce is not at home; he deductively infers that Joyce is at the party. Chris then falls into dreamless sleep. At this later time, Chris does have reflectively accessible rational support for his belief that Joyce is at the party: his knowledge that Joyce is either at the party or at home, and his knowledge that Joyce is not at home. Furthermore, both instances of knowledge entail that Joyce is at the party, so the rational support is factive.15 So, we think that cases of stored memory knowledge that has its origin in deductive inference is not a problem for the S-Memory Core Thesis.16 Things are different with nondeductive inference. Plausibly, inductive knowledge can also be stored knowledge. Consider inductive knowledge that is based on an ampliative inference, which involves premises that do not entail their conclusion. Presumably, much of our everyday and scientific knowledge rests on ampliative inference. Suppose Josh knows that he has mice in his kitchen on the basis of seeing the usual symptoms of a mouse infestation: droppings, tracks, and signs of gnawing. It is natural to think of Josh as inferring, from the fact that he sees the various symptoms, that there are mice in his kitchen. Of course, Josh’s seeing droppings, tracks, and signs of gnawing does not entail that there are mice in his kitchen. So Josh’s knowledge that he has mice in his kitchen rests on an ampliative inference. If pressed as to why he believes that he has mice in his kitchen, Josh will appeal to his seeing the symptoms of a mouse infestation. This inductive knowledge can also become stored memory knowledge.

294

Joe Milburn and Andrew Moon

This is a problem for the S-Memory Thesis. Josh does not possess his initial inductive knowledge in virtue of possessing factive rational support, since his support does not entail the truth of his belief. But then, when this knowledge becomes stored, his rational support will still only make his belief probable and not entail the truth of his belief. So, this standard case of stored memory knowledge, originally arrived at by induction, counts as a counterexample to S-Memory Thesis. One might object by claiming that before Josh gained his inductive knowledge, he lacked factive rational support for his belief; but after he gained this knowledge, he acquired factive rational support for it.17 In this case, while Josh does not have inductive knowledge in virtue of having factive rational support, Josh could have stored knowledge in virtue of possessing factive rational support. However, on the plausible assumption that rational support is irreflexive, there simply is no good candidate for what that rational support might be. Note that, when pressed as to why he believes that he has mice in his kitchen, Josh is likely to appeal to the various symptoms of mice infestation that he has seen. But this is to appeal to the same non-factive rational support he had when he first formed his belief. So our common linguistic practice does not provide a clue as to where the new, factive rational support is supposed to come from. Furthermore, the objection implies that we gain rational support for our inductive knowledge once it becomes stored memory knowledge. This is odd. Suppose at 9:00 a.m., Josh sees the symptoms of mice infestation, and at 9:01 he makes the relevant inference, and so comes to know that he has mice in his kitchen. Suppose that Josh receives no more information bearing on the question of whether he has mice in his kitchen throughout the day, but his knowledge that he has mice in his kitchen becomes part of his stored memory knowledge. To claim that Josh now has more rational support for his belief that there are mice in his kitchen than he did at 9:00 a.m. is absurd. If there is any change in the rational support for Josh’s belief, it seems that, given the fallibility of memory, Josh would lose rational support for his belief, not gain it. Given that we can have paradigmatic instances of stored memory knowledge that have their source in nondeductive inference, the S-Memory Thesis fails.

4. Conclusion In this chapter, we distinguished between two sorts of propositional memory knowledge: experiential memory knowledge and stored memory knowledge. In the case of experiential memory knowledge, we argued for the parity thesis: the prospects for a disjunctivist account of experiential memory knowledge are as good (or roughly as good) as the prospects for a disjunctivist account of perceptual knowledge. In the case of stored memory knowledge, we denied this. There are good reasons for believing that a disjunctivist account of stored memory knowledge fails. On

Two Forms of Memory Knowledge

295

the one hand, there are paradigmatic cases of stored memory knowledge in which one lacks rational support for the relevant beliefs (not to mention factive, reflectively accessible rational support!) On the other hand, even if one were always to possess stored memory knowledge in virtue of having rational support, there are paradigmatic cases of stored memory knowledge that would lack factive rational support (those instances of memory knowledge that have their source in inductive inference). Either way, we should reject the parity thesis for stored memory knowledge. Two results of this chapter are worth stressing. The first is that Pritchard’s style of disjunctivism is applicable beyond the cases of perceptual knowledge. The second is that there are kinds of knowledge for which it is not applicable. These results call for a potentially fruitful question for the disjunctivist. Is it a merely contingent fact that we possess certain kinds of knowledge in virtue of possessing factive, reflectively accessible rational support and certain other kinds of knowledge without it? Could it have been the case that we possess e.g. paradigmatic perceptual knowledge without having factive, reflectively accessible rational support? Or is there something special about the role of perceptual knowledge or experiential memory knowledge, as opposed to stored memory knowledge, that explains why this is the case? Providing an answer to these questions will help us to better understand epistemological disjunctivism.18

Notes 1. One might wonder how this claim is at the core of epistemological disjunctivism. It follows from Pritchard’s core thesis that the reflectively accessible rational support we possess for our perceptual beliefs is of two kinds. Either it is factive rational support, as in the good case, or it is non-factive rational support, as in the bad case. 2. As should be clear from the following discussion, nothing we say hangs upon the nature of veridical and non-veridical memory experiences. Following Pritchard (2012: 24), we assume throughout this chapter that epistemological disjunctivism is logically distinct from metaphysical disjunctivism. For a discussion of memory and metaphysical disjunctivism see Schwarz (2018). 3. The taxonomy for memory knowledge in this section roughly follows Moon (2012: 310–323). See also Bernecker (2010: ch. 1) and Moon (2013: 2718–2719). 4. Pritchard’s views as to how epistemological disjunctivism provides a satisfying response to radical skepticism differ through time. In his (2012), Pritchard takes epistemological disjunctivism to provide the grounds for a satisfying Neo-Moorean response to skepticism. In his (2016), Pritchard rejects NeoMoorean responses to skepticism but holds that epistemological disjunctivism provides the resources for undermining the underdetermination-based skeptical paradox. 5. Or, alternatively, one could naturally say, “I distinctly remember being about to lock the door, but not doing it, because I wanted you to be able to get the book”. This is important for dealing with the basis problem. 6. For more defense of the factivity of propositional memory, see Bernecker (2008: ch. 8).

296

Joe Milburn and Andrew Moon

7. This response to skepticism was not considered for how to close the appearancereality gap in section 4.1 of Moon (2017). 8. At least one of the authors of this chapter—you can guess which one!— thinks that Bernecker fails to show that the memory entailment thesis is false (see Moon 2013 for a reply to Bernecker’s cases and a general defense of the memory entailment thesis). Moon’s strategy might apply to Pritchard’s arguments against the visual entailment thesis, although that exploration must await another paper. 9. One might object that object and event seeing are not factive in the relevant sense. Suppose Michelle and Hugh know that the YL-pill causes hallucinations. Hugh takes the YL-pill, and they wait; he hallucinates a yellow lemon. If Michelle asks Hugh, “What do you see?” and Hugh responds, “I see a yellow lemon!” This is natural for him to say. One might conclude that one can see an F-object without there actually being an F-object. In reply, retraction data indicates that Hugh’s statement is not meant literally. Were Michelle to say, “But literally, you don’t see a yellow lemon”, Hugh would naturally reply, “You’re right, Michelle. Literally, I don’t see a yellow lemon.” This is evidence that the original statement shouldn’t be taken literally. 10. We have presented only a partial defense of the sort of view French defends. For a more complete defense, see French (2016: 95–102 this volume, Chapter 9). 11. Moon (2017: 346) has denied the relational nature of event memory by pointing out that the following sentence, (1) “Old man Nelson remembers the fish being this big, but he’s certainly wrong,”

12.

13. 14.

15.

does not seem inconsistent. (We can imagine that the fish gets bigger every time he tells the story.) If (1) is consistent, then one can remember the event of an x being F even if x was not, in fact, F. We, including Moon, are inclined to think this is a mistake. We think that expressions of “S remembers x A-ing” are ambiguous between two readings. On one reading, S actually remembers the event of x A-ing; on this reading, S remembers x A-ing only if an event of x A-ing actually occurred. On another reading, S seems to remember the event of x A-ing; on this reading, S can seem to remember x A-ing even if there was no event of x A-ing. Seeming to remember should be understood in terms of actual remembering; one seems to remember x A-ing if and only if one seems to be in a state in which one actually remembers x A-ing. Our expressions of “S remembers x A-ing” are ambiguous between seeming to remember and actual remembering. And it’s actually remembering x A-ing that is the factive, reflectively accessible support for the proposition that x was A. Here, we are employing Moon’s (2012) argument that one can have knowledge without evidence. His example is an instance of stored memory knowledge that had its origin in a priori intuition. See that reference for further exploration and defense of this type of argument. See Moon (2013) for defense of the claim that remembering entails believing. One might argue that rational support is not irreflexive: one’s belief that p can be rationally supported by one’s remembering that p (which includes one’s believing that p). However, this comes at high theoretical costs. First, it is counterintuitive. Second, it would be impossible for one to believe on the basis of one’s rational support. This follows from the plausible claims that basing requires causation and that self-causation is impossible. Although this entailment might not amount to factivity, in its ordinary sense, the rational support here entails the truth of the relevant p, which is what is required by the S-Memory Core Thesis.

Two Forms of Memory Knowledge

297

16. However, despite Fred’s having rational support for his belief, there are good reasons to doubt that Fred’s belief can be based on that support. See Moon (forthcoming) for a defense of this claim. 17. Consider for instance, views that take all knowledge that p to be evidence that p (e.g. Williamson 2000). If we take one’s evidence for p to provide one with rational support for believing that p, then one will have factive rational support for everything one knows. For criticisms of the view that all knowledge that p is evidence that p see Jessica Brown (2018). 18. Thanks to Casey Doyle and Clayton Littlejohn for providing comments on an earlier draft of this chapter.

References Bernecker, Sven, 2008, The Metaphysics of Memory. Dordrecht: Springer. ———, 2010, Memory: A Philosophical Study. New York: Oxford University Press. Brown, Jessica, 2018, Fallibilism: Evidence and Knowledge. Oxford: Oxford University Press. French, Craig, 2016, The formulation of epistemological disjunctivism, Philosophy and Phenomenological Research 92: 86–104. McDowell, John, 1998a, Criteria, defeasibility, and knowledge, in his Meaning Knowledge and Reality. Cambridge, MA: Harvard University Press, pp. 369–394. ———, 1998b, Knowledge and the internal, in his Meaning Knowledge and Reality. Cambridge, MA: Harvard University Press, pp. 395–413. ———, 1998c, Knowledge by hearsay, in his Meaning Knowledge and Reality. Cambridge, MA: Harvard University Press, pp. 414–443. Moon, Andrew, 2012, Knowing without evidence, Mind 121: 309–331. ———, 2013, Remembering entails knowing, Synthese 190: 2717–2729. ———, 2017, Skepticism and memory, in Sven Bernecker and Kourken Michaelian, eds. The Routledge Handbook of Philosophy of Memory. New York: Routledge, pp. 335–347. ———, forthcoming, All evidential basing is phenomenal basing, in P. Bondy and J.A. Carter, eds. The Epistemic Basing Relation. New York: Routledge. Pritchard, Duncan, 2012, Epistemological Disjunctivism. Oxford: Oxford University Press. ———, 2016, Epistemic Angst: Radical Skepticism and the Groundlessness of Our Believing. Princeton: Princeton University Press. Schwarz, Arieh, 2018, Memory and disjunctivism, Essay in Philosophy 19: 2 (Essay 4). Williamson, Timothy, 2000, Knowledge and Its Limits. Oxford: Oxford University Press.

15 Testimonial Disjunctivism Stephen Wright

1. The central thesis of epistemological disjunctivism is that it is possible for a subject’s knowledge to be grounded in a reason that is both (i) factive and (ii) reflectively accessible to the subject. Traditionally, discussions of epistemological disjunctivism have focused on the question of perceptual knowledge. In discussing epistemological disjunctivism, Duncan Pritchard makes the following suggestive comment: It may well be possible to offer a variant of epistemological disjunctivism which is applicable to knowledge in general. (Pritchard 2012: 13) A notable exception to the rule that discussions of epistemological disjunctivism focus on knowledge from perception is John McDowell’s (1998b) discussion of the epistemology of testimony. McDowell offers a disjunctivist view of knowledge from testimony that is based on a disjunctivist view of perceptual knowledge. This chapter examines the plausibility of McDowell’s disjunctivist epistemology of testimony. It begins with an identification of the key elements of McDowell’s view before moving to the question of where it fits into the landscape of traditional reductionist and anti-reductionist theories. Contrary to the prevailing view, it is argued that McDowell’s testimonial disjunctivism has just as much in common with traditional reductionist views in the epistemology of testimony as it does with traditional antireductionist views. This point becomes particularly significant when it transpires that the testimonial disjunctivist’s distinctive commitments give her a line of response to arguments prominently developed against anti-reductionism that is not available to traditional anti-reductionists.

2. According to epistemological disjunctivism, a subject’s knowledge can be grounded in a reason that is both factive and reflectively accessible to the

Testimonial Disjunctivism

299

subject.1 Applied to the epistemology of testimony, this amounts to the view that a listener can come to know the truth of what a speaker says on the basis of some reason that is incompatible with the falsity of her belief and that the listener is aware of. The first part of McDowell’s testimonial disjunctivism involves the identification of a factive reason that can be made available by a speaker’s testimony: The epistemic standing one can acquire in conversation is that of having heard from one’s interlocutor that things are thus and so. One cannot count as having heard from someone that things are thus and so unless, by virtue of understanding what the person says, one is in a position to know that things are that way. If it turns out that things are not that way, or that although they are, the person from whom one took oneself to have heard it did not know it, one cannot persist in the claim that one heard from him that things are that way, but must retreat to the claim that one heard him say that they are. (McDowell 1998b: 433–434) A listener’s knowledge of the truth of what a speaker says can thus be grounded in the fact that she heard it from the speaker that things are thus and so. Having heard from someone that things are thus and so is factive and thus grounds knowledge. By contrast, having heard someone say that things are thus and so is not and thus does not ground knowledge. Testimony thus sometimes makes available a factive reason that grounds knowledge and sometimes makes available a non-factive reason that does not ground knowledge. The obvious next question is when a listener has a factive reason available to her and when she does not. In McDowell’s terms, the question is when a listener is in a position of having heard from someone that things are thus and so and when she is merely in a position of hearing from a speaker that things are thus and so and when she is merely in a position of having heard a speaker say that things are thus and so. McDowell’s account of this is given in the following: The idea of knowledge by testimony is that if a knower gives intelligible expression to his knowledge, he puts it into the public domain, where it can be picked up by those who can understand the expression, as long as the opportunity is not closed to them because it would be doxastically irresponsible to believe the speaker. (McDowell 1998b: 438) There are two central components to this. The first is that knowledge from testimony involves the speaker’s testimony facilitating the transfer of her knowledge to a listener through the expression of her knowledge. A speaker, according to McDowell, gives expression to her knowledge,

300

Stephen Wright

and it is this same knowledge that is picked up by the listener. Giving expression to one’s knowledge is plausibly more than simply saying what one knows. Rather, it involves saying what one knows because one knows it.2 The second is that the listener comes to acquire the speaker’s knowledge only when believing the speaker’s testimony is not doxastically irresponsible. If a listener is aware of reasons for thinking that the speaker is untrustworthy, these prevent the listener from acquiring the speaker’s knowledge, even if it is made available by the speaker’s testimony.3 It is important that McDowell’s disjunctivism does not maintain that factive reasons are always made available to a listener in a case of testimony. Rather, they are made available only in a limited range of cases. In a similar spirit, Pritchard’s discussion of epistemological disjunctivism claims that a perceiving subject’s knowledge is grounded in a factive reason that is reflectively accessible to her only in cases where the situation is objectively good in that the subject’s perceptual processes are reliable and situated in an environment conducive to their reliable functioning and subjectively good in that the subject is not aware of any reasons for doubting the belief in question or the reliability of her perceptual processes (Pritchard 2012: 29–30). With the nature of the factive reason that McDowell believes can ground knowledge from testimony and an account of the conditions under which it can do so in view, the remaining question is why this reason should be thought to be reflectively accessible to the listener. McDowell’s reason for thinking that this is the case is broadly transcendental. If factive reasons cannot be reflectively accessible to a listener, then the consequence is not merely that knowledge from testimony is impossible. Rather, the consequence is that our thoughts fail to be about the world at all. McDowell makes this point in the discussion of epistemological disjunctivism about perceptual knowledge: If we conceive what we want to think of as the space of concepts, the realm of thought, in a way that alienates it so radically from the merely material that we seem to be faced with those familiar modern problems of reconciling the subjective with the objective, we undermine our right to think of it as the realm of thought at all. When we set it off so radically from the objective world, we lose our right to think of moves within the space we are picturing as contentinvolving. So we stop being able to picture it as the space of concepts. Everything goes dark in the interior as we picture it. (McDowell 1998b: 409) Factive reasons being reflectively accessible to a subject is thus a corollary of our thoughts being about the world at all. One might think that, even if this is plausible with respect to perceptual knowledge, it remains an open question whether or not this is plausible with respect to knowledge

Testimonial Disjunctivism

301

from testimony. Perceptual knowledge, one might think, is fundamental in a way that knowledge from testimony is not. Hence it does not follow from the fact that our thoughts being about the world at all depend on perceptual knowledge being grounded in factive reasons that are reflectively accessible that something similar is true of testimony. On closer inspection, however, this becomes difficult to maintain. As Strawson points out, our comprehension of what we perceive depends on concepts acquired through testimony: In any community of language-users, perception, memory, and testimony are not only equally essential to the construction of the beliefor-knowledge-systems of its members. It is also true that all three are on an equal footing in that there is no possibility of a general reductive analysis of any one of the three in terms of the others, supplemented by inference. The interdependence of all does not entail the reducibility of any. If we (often) know, directly and immediately, what our eyes tell us, then we (often) know, no less directly and immediately, what other people tell us. (Strawson 2011: 190) Insofar as testimony is no less fundamental than perception, the transcendental argument developed with respect to perceptual knowledge applies equally to knowledge from testimony. We thus have the contours of McDowell’s testimonial disjunctivism in view. According to McDowell’s disjunctivist epistemology of testimony, knowledge from testimony can be grounded in a factive reason. The factive reason is the fact that the listener heard from the subject that things are thus and so. Moreover, this reason should be thought to be reflectively accessible to the listener because this is required not merely for the possibility of knowledge, but for the possibility of our thoughts being about the world at all. The range of cases in which a listener’s knowledge is grounded in such a reason is limited, however, to cases in which a speaker’s testimony is an expression of her knowledge and the listener is not irresponsible in believing the speaker’s testimony.

3. So much for the outline of McDowell’s view. The next order of business is to establish the place that McDowell’s testimonial disjunctivism occupies in the landscape of theories in the epistemology of testimony. Existing discussions from Jennifer Lackey (2008) and Paul Faulkner (2011) situate McDowell’s view squarely in the anti-reductionist tradition. The case for doing so rests on two observations. The first is that, like traditional anti-reductionism, McDowell’s testimonial disjunctivism maintains that knowledge from testimony involves the transmission of knowledge from

302

Stephen Wright

speaker to listener. The second is that both traditional anti-reductionism and testimonial disjunctivism reject the claim that knowledge from testimony depends on a listener being aware of the kind of reasons that traditional reductionist views emphasise the importance of. The assimilation of McDowell’s view and anti-reductionism on these grounds, however, is premature. On closer inspection, the apparent similarities between testimonial disjunctivism and anti-reductionism turn out to be superficial—a veil for disagreement between the two views. Moreover, for all the similarities between McDowell’s view and traditional anti-reductionism, there are important points of agreement between McDowell’s view and traditional reductionism. The result is that, in much the same way that disjunctivism about knowledge in general occupies a space between traditional internalist and externalist theories, McDowell’s testimonial disjunctivism occupies a space between traditional reductionism and anti-reductionism. In the epistemology of testimony, the terms ‘reductionism’ and ‘antireductionism’ are terms of art that resist precise definition. They were introduced in the 1990s—around the time of the first publication of McDowell’s presentation of testimonial disjunctivism—to distinguish theories that claim that knowledge from testimony depends on a listener being aware of reasons for believing what a speaker says from those that do not make this claim. Since then, however, the terms have been used in various different ways by various different authors to make various different points about knowledge from testimony. Most recently, the epistemology of testimony has sought to explore the territory between and beyond the traditional positions.4 Nonetheless, an idea of what is at stake between traditional reductionist and anti-reductionist views of testimony can be gained by considering two views of testimony from the time of the initial publication of McDowell’s disjunctivist approach. The first is the reductionist view developed by Elizabeth Fricker (1994, 1995). The second is the antireductionist view developed by Tyler Burge (1993). Both views have been developed since their initial formulations.5 And other theories along similar lines have been developed.6 The views developed by Fricker and Burge respectively, however, constitute paradigmatic versions of reductionism and anti-reductionism. The disagreement over exactly what the difference between reductionism and anti-reductionism notwithstanding, it is generally agreed that the early statement of Fricker’s view constitutes a reductionist theory and the early statement of Burge’s view constitutes an anti-reductionist theory. These can thus be used as landmarks to situate McDowell’s testimonial disjunctivism. According to Fricker’s view, knowledge from testimony is the product of a listener formulating a theory of why a speaker said what she did on an occasion in question. As Fricker puts it:

Testimonial Disjunctivism

303

[T]he primary task for the hearer is to construct enough of a theory of the speaker, and relevant portions of her past and present environment, to explain her utterance: to render it comprehensible why she made that assertion, on that occasion. (Fricker 1994: 149) When the listener’s theory of why the speaker said what she did on the occasion in question features the truth of what the speaker says, the listener’s theory provides her with a reason for believing what the speaker says. It is this reason that grounds knowledge from testimony. But the listener’s theory cannot be constructed in a vacuum. Rather, it is to be constructed from the listener’s monitoring the speaker’s testimony for signs of truth or falsity. A speaker’s testimony might be false either because the speaker is insincere or because the speaker is incompetent. According to Fricker, a listener is typically able to discern signs of insincerity and incompetence when she is told something by a speaker. As a result, a listener monitoring a speaker for signs of insincerity and incompetence and finding none provides her with a reason for thinking that the speaker’s testimony is true. The reliability of the monitoring process grounds the listener’s theory of why the speaker said what she did, and the listener’s theory of why the speaker said what she did grounds the listener’s knowledge. Fricker’s view maintains that knowledge from testimony is grounded in reasons for believing the speaker’s testimony that a listener is aware of. A consequence of this is that if the listener is not aware of reasons for believing the speaker’s testimony, then she cannot come to know the truth of what the speaker says by believing her testimony. Hence, Fricker maintains that believing what a speaker says without being aware of reasons for doing so exhibits gullibility, which is an epistemic vice and does not result in knowledge. By contrast, Burge’s anti-reductionism maintains that a listener coming to know the truth of what a speaker says does not depend on her being aware of reasons for believing the speaker’s testimony. Burge encapsulates this in the Acceptance Principle, which states that: A person is apriori entitled to accept a proposition that is presented as true and that is intelligible to him, unless there are stronger reasons not to do so, because it is prima facie preserved (received) from a rational source, or resource for reason; reliance on rational sources—or resources for reason—is, other things equal, necessary to the function of reason. (Burge 1993: 469) Importantly, a listener being entitled to believe what a speaker says does not depend on the listener being aware of why she is so entitled, or even

304

Stephen Wright

that she is so entitled. Contrary to Fricker’s reductionism, Burge’s antireductionism maintains that a listener can come to know the truth of what a speaker says even if she is not aware of reasons for believing the speaker’s testimony. A consequence of this is that Burge’s view cannot claim that knowledge from testimony is grounded in reasons for believing a speaker’s testimony that a listener is aware of. Where Fricker’s reductionism takes it that knowledge from testimony is grounded in reasons that a listener is aware of, Burge’s anti-reductionism takes it that knowledge from testimony can be a matter of transmission. The idea is that a listener can acquire a speaker’s knowledge in a case where her belief is not the product of her reasoning about the relationship between the speaker’s testimony and the truth of what she says. Burge does not maintain that this is what usually happens in cases of testimony, but insists that this is at least possible. The central difference between Fricker’s reductionism and Burge’s anti-reductionism thus concerns cases in which a listener is not aware of reasons for believing a speaker’s testimony. According to Fricker’s reductionism, a listener believing what a speaker says in such a case amounts to gullibility and is not a way of coming to know the truth of what the speaker says. Knowledge from testimony is grounded in reasons that a listener is aware of, so the absence of such reasons means an absence of knowledge. According to Burge’s anti-reductionism, a listener can come to know the truth of what a speaker says in such a case. The speaker’s knowledge can be transmitted to the listener, the result being that the listener’s knowledge is grounded in reasons beyond those that she is aware of. It is easy enough to see why McDowell’s testimonial disjunctivism is often placed in the anti-reductionist tradition. Like Burge’s anti-reductionism, McDowell’s testimonial disjunctivism maintains that knowledge is transmitted from speaker to listener. Moreover, like Burge’s anti-reductionism, McDowell’s testimonial disjunctivism maintains that a listener can come to know the truth of what a speaker says without being aware of the kind of reasons for believing the speaker’s testimony that Fricker’s reductionist view identifies. These are the two main reasons for situating McDowell’s testimonial disjunctivism in the anti-reductionist tradition. But to do so on these grounds is premature. Lackey notes that a consequence of transmission theories, such as the one that Burge develops, is that a listener’s knowledge can be explained in terms of transmission only if the speaker is in a position to know the truth of what she says. But this is equally true of McDowell’s disjunctivist theory. A listener’s knowledge being grounded in the distinctive factive reason that McDowell’s testimonial disjunctivism identifies depends on the speaker’s testimony being an expression of her knowledge. This is possible only if the speaker knows the truth of what she says. Hence, Lackey observes that there is an important similarity between McDowell’s testimonial disjunctivism and the anti-reductionism developed by Burge (Lackey 2008: 39).

Testimonial Disjunctivism

305

Faulkner makes a similar point in identifying McDowell’s testimonial disjunctivism as an anti-reductionist theory. According to McDowell’s view, a speaker’s testimony puts a listener in the same state of informedness as the speaker with respect to what the speaker says. Hence, the speaker’s testimony making a factive reason available to a listener depends on the speaker herself being in possession of a factive reason for believing what she says. In much the same way, the transmission of knowledge depends on the speaker being in a position to know the truth of what she says. Faulkner thus identifies McDowell’s view as similar to Burge’s anti-reductionism (Faulkner 2011: 106). Viewed in a certain light, there is clearly an important point of agreement between Burge’s anti-reductionism and McDowell’s testimonial disjunctivism here. Whilst it is true that both theories maintain that testimony can transmit knowledge from speaker to listener, their reasons for thinking this are quite different. And this difference amounts to an equally important point of disagreement between the anti-reductionist and the testimonial disjunctivist. For the anti-reductionist, the reason for endorsing transmission comes from the idea that knowledge can be grounded in something that is not reflectively accessible to a subject. The anti-reductionist accepts that a listener can come to know the truth of what a speaker says by believing her testimony without being aware of reasons for doing so. Hence, the anti-reductionist accepts transmission on the grounds that the epistemic grounds that underpin knowledge from testimony can extend beyond that which the listener is reflectively accessible to the listener. The case with the disjunctivist, however, is different. The disjunctivist endorsement of transmission is the product of thinking that a speaker’s testimony can make a factive reason reflectively accessible to a listener. Unlike the anti-reductionist, the disjunctivist’s endorsement of transmission is not the product of thinking that the epistemic grounds that underpin knowledge from testimony might extend beyond that which is reflectively accessible to the listener. The testimonial disjunctivist maintains that only that which is reflectively accessible to the listener can ground knowledge. Not only does this point mark an important difference between the testimonial disjunctivist and the anti-reductionist, it marks an important similarity between the testimonial disjunctivist and the reductionist. Both the testimonial disjunctivist and the reductionist maintain that knowledge from testimony is grounded only in reasons that are reflectively accessible to the listener. The difference between the testimonial disjunctivist and the reductionist is a disagreement over whether or not factive reasons are reflectively accessible to a listener in a case of knowledge from testimony. Where the reductionist thinks that they are not, the disjunctivist thinks that they are. This is connected to the second point of similarity between the antireductionist and the testimonial disjunctivist. Both the anti-reductionist

306

Stephen Wright

and the testimonial disjunctivist maintain that a listener can come to know the truth of what a knowledgeable speaker says on the basis of an entitlement to believe the speaker’s testimony. Equally, both the antireductionist and the testimonial disjunctivist maintain that the epistemology of testimony is not to be explained in terms of the kind of reasons for believing the speaker’s testimony that Fricker’s reductionism identifies. Both Lackey and Faulkner identify this as a point of commonality between Burge’s anti-reductionism and McDowell’s testimonial disjunctivism. Lackey points out McDowell’s claim that a listener can pick up knowledge that is made available by a speaker’s testimony as long as she is not doxastically irresponsible in believing what the speaker says (Lackey 2008: 40). This observation is given additional force by the observation that McDowell’s vocabulary does make reference to entitlements. According to McDowell: [s]omeone who can truly make a claim of that [‘I see that . . .’] form has an entitlement, incompatible with any possibility of falsehood, to a claim whose content is given by the embedded proposition. The entitlement consists in the visual availability to her of the fact she would affirm in making that claim. (McDowell 2002: 98) Individuals thus enjoy an entitlement to appeal to factive premises, on McDowell’s view. Faulkner also notes that this aligns McDowell’s view with the anti-reductionist tradition (Faulkner 2011: 101–102). As with the previous point of agreement, however, the similarity masks a deeper point of disagreement. Whilst it is true that both the anti-reductionist and the testimonial disjunctivist accept that a listener can be entitled to believe the truth of what a speaker says without being aware of the kind of reasons that Fricker’s reductionism identifies, their entitlements differ in important ways. Most importantly, Burge’s Acceptance Principle maintains that a listener can be entitled to believe what a speaker tells her even if the speaker’s testimony is false. The listener’s entitlement does not derive from particular facts about the speaker’s testimony and thus constitutes a general entitlement to believe what speakers say unless one is aware of reasons against doing so. McDowell’s entitlement is different. Note that, for McDowell, someone has an entitlement that consists in the visual availability to her of the fact that she would affirm in making the claim. In other words, it is exactly because the case is one that makes a factive reason reflectively accessible to a listener that she has the entitlement McDowell mentions. In discussing a case where a tourist is told the whereabouts of the cathedral by a local, McDowell states that ‘the tourist is entitled to his belief about where the cathedral is . . . but I do not think that is because he is exercising a general presumption of sincerity and competence’ (McDowell

Testimonial Disjunctivism

307

1998b: 218). Where the anti-reductionist accepts that a listener might have an entitlement to believe what a speaker says when her testimony is not true, the testimonial disjunctivist does not. Again, this is indicative of a deeper point of similarity between the disjunctivist and the reductionist. Neither the reductionist nor the disjunctivist thinks that a listener is entitled to believe what a speaker says if she is not aware of reasons for doing so. Whilst the disjunctivist accepts that a listener might come to know the truth of what a speaker tells her on the basis of an entitlement, the cases in which the disjunctivist claims that the listener has such an entitlement are all and only those cases in which the disjunctivist claims that the listener has a factive reason reflectively accessible to her. Hence, the reductionist and the anti-reductionist agree that a listener is not entitled to believe a speaker’s testimony in a case where she is not aware of reasons for doing so. And this is an important point of agreement. Lastly, whilst it is true that both the anti-reductionist and the testimonial disjunctivist maintain that the epistemology of testimony is not to be explained in terms of the reasons for believing a speaker’s testimony that Fricker’s reductionism identifies, their reasons for doing so are— again—quite different. Where Burge’s anti-reductionism maintains that a listener can come to know the truth of what a speaker says on the basis of the kind of reasons that Fricker’s reductionism identifies, McDowell’s testimonial disjunctivism does not. According to the anti-reductionist, the kind of reasons Fricker’s reductionism identifies can ground knowledge, but not all knowledge from testimony is grounded in this way. Reductionist reasons are, for the anti-reductionist, sufficient but not necessary for knowledge from testimony. For the testimonial disjunctivist, they are neither necessary nor sufficient. According to McDowell, the kind of reasons that Fricker’s reductionism identifies are not sufficient for knowledge. In the case of the tourist coming to know the whereabouts of the cathedral, McDowell states that the kind of reasons that Fricker identifies cannot ground knowledge: If the tourist’s title to know depends on the best argument he can muster for the proposition he believes, and the premises of the argument leave it open that his supposed informant is not giving expression to knowledge, then surely the verdict ought to be that for all the tourist knows the cathedral is somewhere else. (McDowell 1998b: 419) Whilst there is agreement between the anti-reductionist and the testimonial disjunctivist, this agreement masks substantial disagreement. The testimonial disjunctivist’s reason for thinking that the epistemology of testimony should not be understood in terms of the reasons that Fricker’s reductionism identifies is that to think of the epistemology of testimony

308

Stephen Wright

in these terms is to settle for non-factive reasons when factive reasons are available. McDowell explicitly states agreement with Fricker’s view on the subject of knowledge from testimony being grounded only in reasons that are reflectively accessible to the listener. The difference between McDowell’s disjunctivism and Fricker’s reductionism is a disagreement over the scope of what reasons are reflectively accessible to the listener. But on this point, Fricker’s reductionism agrees with Burge’s anti-reductionism. Ultimately, whilst it is true that there are points of agreement between McDowell’s testimonial disjunctivism and the traditional anti-reductionism developed by Burge, these do not amount to a decisive case for thinking of testimonial disjunctivism as a version of anti-reductionism. This is because there are also important points of similarity between the testimonial disjunctivist and the traditional reductionist view developed by Fricker. In the same way that disjunctivism about knowledge in general aligns precisely with neither traditional internalist nor traditional externalist views of knowledge, testimonial disjunctivism occupies a space in between traditional reductionist and anti-reductionist views of testimony.

4. The purpose of establishing the place of testimonial disjunctivism in the landscape of traditional theories goes beyond exegesis. Understanding where testimonial disjunctivism fits into the broader landscape of reductionist and anti-reductionist views is crucial to understanding how testimonial disjunctivism withstands the arguments that have been developed against traditional anti-reductionism. The differences between testimonial disjunctivism and anti-reductionism illustrate why the disjunctivist position is not so easily dismissed as its critics maintain. Faulkner argues against testimonial disjunctivism on the grounds that the kind of factive reasons that the disjunctivist describes are straightforwardly not available to a listener in a case of testimony. Faulkner’s objection is based on the following statement: Let me—the author, Paul Faulkner—tell you something about myself. When I was schoolboy I was a member of Phoenix Athletic Club. And I came third in the school senior cross-country race. (Faulkner 2011: 132) One of the statements is true and an expression of knowledge. The other is a lie. But, as Faulkner points out, it is not obvious which is which. Each statement seems as plausible as the other. Contra testimonial disjunctivism, then, it seems that there is no reason available to the listener that is both reflectively accessible and distinctive to the case that is an expression of knowledge.

Testimonial Disjunctivism

309

This seems to impugn the testimonial disjunctivist claim that a listener’s knowledge can be grounded in reflectively accessible factive reasons. Since one of the statements is true and an expression of knowledge, it seems to make a factive reason available to a listener. But no such reason seems to be available. Faulkner thus concludes that, even if epistemological disjunctivism is plausible with respect to knowledge from perception, it is not plausible with respect to knowledge from testimony. In Faulkner’s words: The testimonially presented fact just does not ‘impress itself’ in the way that perceptually presented facts do. As such, it is just not plausible to suppose the knowledge of my high school years that the former testimony makes available to you makes it reasonable for you to believe that I am telling the truth with this statement. (Faulkner 2011: 132) Faulkner’s description of the phenomenology is surely accurate. But there is reason to think that the case does not undermine McDowell’s testimonial disjunctivism. The central point of McDowell’s observations in motivating testimonial disjunctivism is to show that it is possible for a listener’s knowledge to be grounded in a factive reason that is reflectively accessible to her. The fact that no such reason is available in Faulkner’s case undermines McDowell’s testimonial disjunctivism only if the case is one in which McDowell’s disjunctivist view maintains that there is a factive reason made reflectively accessible reason to the listener. But there is reason to think that it is not. Note that, in the case Faulkner describes, the true statement is presented with an accompanying false statement that is indistinguishable from it. One might think that this fact means that the case is not one in which a factive reason is made reflectively accessible to a listener. Compare the case that Faulkner presents with the case presented by Alvin Goldman (1976) in which someone sees the only barn in an environment containing various convincing facsimiles. In Goldman’s case, it is intuitive that the subject does not know that there is a barn in front of her because the environment is such that she might easily have formed a false belief in this way. This is so because of the presence of the facsimiles. When Faulkner presents us with two statements, one of which is an expression of knowledge and the other a lie, the case is similar. There is a true statement and there is a lie disguised as a truth. In such a situation, one might think that it is doubtful that a listener can come to know the fact reported in the true statement in much the same way that it is doubtful that the subject can come to know that there is barn before her in the case Goldman describes. As such, the case does not undermine McDowell’s testimonial disjunctivism. Even if it is true that no factive reason is made reflectively accessible in the case Faulkner presents, it does not

310

Stephen Wright

follow from this that a factive reason is never made reflectively accessible to the listener. This point becomes vivid in Pritchard’s discussion of epistemological disjunctivism. According to Pritchard, the cases in which a subject’s knowledge is grounded in a reason that is both factive and reflectively accessible to her are those cases that are paradigm cases of perceptual knowledge. A case is a paradigm case, according to Pritchard, just in case the epistemic situation is both objectively and subjectively good. In the context of perception, a case is objectively good just in case the cognitive processes in the subject are ones that reliably give rise to true beliefs and are situated in an environment that is conducive to their doing so and is subjectively good just in case the subject is not in possession of reasons that entail that she ought to doubt the belief in question (Pritchard 2012: 29–30). The idea is that this type of case is, by anyone’s standards, a central case of perceptual knowledge. Applied to the case of the barn facsimiles, the idea is that the subject does not come to know on the basis of a factive reason that is reflectively accessible to her because the epistemic situation is not good objectively. The subject is not in an environment that is conducive to her cognitive processes functioning reliably. Something similar is plausibly true of the case Faulkner presents. The epistemic situation is not good objectively, since the listener is faced by a true statement and a false statement and is unable to distinguish between them. As such, the disjunctivist can reject the claim that the speaker’s testimony makes available to the listener a factive reason that is reflectively accessible to her. Something similar is true of the objection to traditional anti-reductionist views presented by Lackey. The argument Lackey develops involves a case of someone finding what appears to her to be a diary dropped by an alien. Inside, it appears to be written that, whilst exploring Earth, some of the inhabitants of the alien’s home planet have been eaten by tigers. The subject treats the apparent diary entry as an instance of testimony and takes it at face value, thereby coming to believe that tigers have eaten some of the inhabitants of the alien’s planet. Against anti-reductionism, Lackey notes that the view that the subject can come to know the truth of what she reads in this way is highly unintuitive. Whilst the case against disjunctivism specifically is implicit, one might note that the fact that the subject cannot come to know entails that the subject cannot come to know on the basis of a factive reason that is reflectively accessible to her. Intuitively, the subject cannot come to know the truth of what she reads because there are various problematic possibilities that she is unable to rule out. The possibility of standard alien practice being to produce works that would appear to be diaries but are actually works of science fiction is one. The possibility of aliens using a language superficially similar to English but with different semantic content is another. Whilst the writing is in fact a statement and is in fact an

Testimonial Disjunctivism

311

expression of the alien’s knowledge, the listener has no reason for ruling out these various possibilities. Hence, Lackey concludes that ‘it seems plainly irrational epistemically for him to form the belief in question on the basis of the alien’s testimony’ (Lackey 2008: 169). In responding to this objection, the disjunctivist has resources available that her anti-reductionist counterpart does not. The anti-reductionist must maintain that the claim that the subject cannot come to know in this case by reading the diary is mistaken. By contrast, the disjunctivist can deny that the case in question is one in which the listener has an entitlement to believe what she reads. Since McDowell’s disjunctivism maintains that someone has an entitlement to her belief only in those cases in which a factive reason is reflectively accessible to her, the disjunctivist can maintain that the case in question is not one in which the subject is entitled to her belief. Various features of the case show that it is not the kind of paradigm case in which the disjunctivist is committed to thinking that the listener is entitled to her belief. These are just the same features that Lackey draws attention to in motivating the intuition that the subject does not come to know based on reading the diary. The subject’s unfamiliarity with the type of testimony produced gives us reason for thinking that this is not the kind of case in which the disjunctivist thinks that a listener’s knowledge is grounded in a reflectively accessible factive reason. It seems as though McDowell’s central thesis—that there are cases in which knowledge from testimony can be grounded in a factive reason that is reflectively accessible to the listener—survives the observation that the case that Lackey describes is not such a case. As such, it might seem as though the objection from Lackey misses its intended target. One might point out that the case that Lackey describes meets the conditions that McDowell sets out. In writing the diary entry, the speaker does give expression to her knowledge and, in doing so, puts it into the public domain. And the listener is not aware of any consideration against believing that some of the inhabitants of the alien’s planet have been eaten by tigers, the opportunity to acquire the speaker’s knowledge is not closed to the listener because believing what is written in the diary is doxastically irresponsible. In light of this, one might doubt whether McDowell’s disjunctivism can resist the objection by maintaining that the case that Lackey describes is outside the scope of cases in which a listener’s knowledge is grounded in a factive reason that is reflectively accessible to her. Even if this counter were successful, it would only establish that McDowell misidentifies the conditions under which a listener’s knowledge is grounded in a factive reason that is reflectively accessible to her. This, however, is a relatively minor objection. The central thesis of testimonial disjunctivism is that a listener’s knowledge can be grounded in a reason that is both factive and reflectively accessible to her. And it would

312

Stephen Wright

seem that this can be maintained even if McDowell’s account of the conditions under which this is the case is mistaken. It therefore seems as though this strategy for undermining testimonial disjunctivism is unlikely to be successful. Testimonial disjunctivism is not to be undermined on the grounds of a straightforward counterexample involving a case in which a listener is confronted by a true statement and an indistinguishable false statement. Nor is it to be undermined on the basis of a case in which a listener intuitively does not come to know what the speaker says but, according to McDowell’s testimonial disjunctivism, is in possession of a reflectively accessible reason for her belief. Whilst these cases might show that the details of McDowell’s view are problematic, they do not show that the general idea behind testimonial disjunctivism is false.

5. The foregoing discussion indicates that there might be grounds for optimism concerning the viability of testimonial disjunctivism. There are important differences between testimonial disjunctivism and anti-reductionism, and these differences mean that the former is equipped with resources for responding to objections that the latter lacks. Nonetheless, I am unable to give epistemological disjunctivism a wholehearted endorsement. Like other epistemological disjunctivist views of knowledge from other sources, the epistemological disjunctivist view of testimony maintains that it is possible for a subject’s knowledge to be grounded in a factive reason that is reflectively accessible to her. That the disjunctivist maintains that this is only so in a limited range of cases allows her to respond to counterexamples by maintaining that they are not the kind of case in which a subject’s knowledge is grounded in a factive reason that is reflectively accessible to her. A problem for testimonial disjunctivism, however, emerges with the difficulty with giving a positive account of the type of case in which a listener’s knowledge is grounded in a factive reason that is reflectively accessible to her. As we have seen, McDowell describes the kind of case in question as one in which a speaker’s testimony is an expression of her knowledge, and the listener is not doxastically irresponsible in believing the speaker’s testimony. This is the application to the epistemology of testimony of a more general idea that Pritchard expresses as the view that the case must be good both subjectively and objectively. The idea is that, given that the case is good both subjectively and objectively, the fact that the subject is in a position to know will be accepted by any nonsceptical epistemology. But there are problems with specifying what a case being good both subjectively and objectively amounts to in the case of the epistemology of testimony. More specifically, there are problems with specifying what it means to say that a case is good subjectively. McDowell’s idea is that a case

Testimonial Disjunctivism

313

being subjectively good is a matter of the listener believing the speaker’s testimony without thereby being doxastically irresponsible. And this is naturally taken to involve the listener not being aware of countervailing considerations. Understood in this way, however, it is highly controversial that avoiding doxastic irresponsibility is sufficient to make a case subjectively good. Those who reject traditional anti-reductionism deny that a listener being unaware of reasons against believing what the speaker says is sufficient for the avoidance of doxastic irresponsibility. Faulkner, for example, argues that a listener cannot come to know the truth of what a speaker says so unless she is in possession of the kind of reasons for believing the speaker’s testimony that the reductionist describes. According to Faulkner, speakers and listeners enter into conversations with practical interests—speakers want to be believed, and listeners want to find out the truth. More generally, however, the interests of speakers and listeners conflict—speakers want to be believed whether or not they are telling the truth, and listeners want to believe speakers only when they tell the truth. This conflict of interests, Faulkner argues, means that a listener believing what a speaker says without being aware of reasons for doing so is irrational (Faulkner 2011: 6). In light of this, it is tempting to think McDowell’s notion of doxastic irresponsibility might be understood in the context of testimony as a listener being aware of reasons for believing what the speaker says. Understood in this way, however, a different problem arises. In a situation where the listener comes to know the truth of what a speaker says on the basis of reasons for believing the speaker that she is aware of— the kind of reasons that Faulkner’s argument concerning practical interests motivates—it seems that the listener’s knowledge is not grounded in the speaker’s testimony, but in the reasons that she uses in forming her belief. Lackey describes a case in which someone forms a belief in what a speaker says on the basis of reasons for thinking that she (the listener) is capable of distinguishing between true and false statements. But in such a situation, Lackey notes, the speaker’s testimony ‘simply drops out of the epistemic picture’ (Lackey 2008: 92). The result is that there is a dilemma for the testimonial disjunctivist. Understanding testimonial disjunctivism as an epistemological theory of testimony at all depends on the testimonial disjunctivist being able to say something about the kind of case in which a listener’s knowledge is grounded in a reason that is both factive and reflectively accessible to her. The idea that the kind of case might be one in which the epistemic situation is both subjectively good and objectively good is an attractive starting point. But disagreement over what makes a testimonial situation subjectively good prevents us from seeing with sufficient clarity the kind of case that the disjunctivist has in mind. This concern means that, whilst there is reason for optimism concerning the application of epistemological disjunctivism to the domain of

314

Stephen Wright

testimony, this optimism is to be tempered with a degree of caution. Since the considerations invoked by Faulkner in support of the claim that a listener coming to know the truth of what a speaker tells her depends on the listener being aware of reasons for believing the speaker’s testimony are distinctive to the epistemology of testimony or, at any rate, do not apply in the case of perception, the problem for testimonial disjunctivism has no analogue in the case of perception. Thus, whilst there are good reasons for thinking that an epistemological disjunctivist theory of knowledge from testimony is worthy of serious consideration, there are also reasons for thinking that such a view faces problems beyond those that its perceptual analogue faces.

Notes 1. Epistemological disjunctivism is to be distinguished from metaphysical disjunctivism, according to which a perceptual experience is to be understood fundamentally as either an instance of seeing or an instance of hallucination. Classic statements of metaphysical disjunctivism are given by Hinton (1973), McDowell (1998a, 1998c), Martin (2002, 2006), and Snowdon (1980–1981, 1990). My own view is that, in metaphysical terms, disjunctivism about testimony is substantially more plausible than disjunctivism about perception. I take this to be exemplified in the work of Ross (1986), Hinchman (2005), and Moran (2005), who argue for important differences between testimony and other paradigmatic types of evidence, though I shall not argue for this here. For more on this view, see Wright (2018). 2. See Owens (2006), Wright (2018). 3. See Ginsborg (2006). 4. For varying definitions of reductionism and anti-reductionism, compare Fricker (1994), Lackey (2008), and Faulkner (2011). Where Faulkner and Lackey seek to explore the ground between the traditional positions, Hinchman (2005), Moran (2005), and McMyler (2011) seek to explore the territory beyond reductionism and anti-reductionism. 5. See Fricker (2006a, 2006b, 2015a, 2015b, 2017), Burge (1997, 2013). 6. Other reductionist views are given by Adler (1994), Barnett (2015), Fumerton (2006), and Malmgren (2006). Other anti-reductionist views are given by Dummett (1996), Owens (2000), and Wright (2018).

References Adler, Jonathan, 1994, Testimony, trust, knowing, Journal of Philosophy 91(5): 264–275. Barnett, David James, 2015, Is memory merely testimony from one’s former self?, Philosophical Review 124(3): 353–392. Burge, Tyler, 1993, Content preservation, Philosophical Review 102(4): 457–488. ———, 1997, Interlocution, perception, and memory, Philosophical Studies 86(1): 21–47. ———, 2013, Postscript: ‘Content preservation’, in his Cognition Through Understanding: Self-Knowledge, Interlocution, Reasoning, Reflection: Philosophical Essays (vol. 3). Oxford: Oxford University Press, pp. 254–284. Dummett, Michael, 1996, The Seas of Language. Oxford: Oxford University Press.

Testimonial Disjunctivism

315

Faulkner, Paul, 2011, Knowledge on Trust. Oxford: Oxford University Press. Fricker, Elizabeth, 1994, Against gullibility, in Bimal Matilal and Arindam Chakrabarti, eds. Knowing From Words. Dordrecht: Kluwer Academic Publishers, pp. 125–161. ———, 1995, Critical notice: Telling and trusting: Reductionism and antireductionism in the epistemology of testimony, Mind 104(414): 393–411. ———, 2006a, Second-hand knowledge, Philosophy and Phenomenological Research 73(3): 592–618. ———, 2006b, Testimony and epistemic autonomy, in Jennifer Lackey and Ernest Sosa, eds. The Epistemology of Testimony. Oxford: Oxford University Press, pp. 225–250. ———, 2015a, How to make invidious distinctions amongst reliable testifiers, Episteme 12(2): 173–202. ———, 2015b, Know first, tell later: The truth about Craig on knowledge, in John Greco and David Henderson, eds. Epistemic Evaluations. Oxford: Oxford University Press, pp. 46–86. ———, 2017, Inference to the best explanation and the receipt of testimony: Testimonial reductionism vindicated, in Kevin McCain and Ted Poston, eds. Best Explanations: New Essays on Inference to the Best Explanation. Oxford: Oxford University Press, pp. 262–294. Fumerton, Richard, 2006, The epistemic role of testimony: Internalist and externalist Perspectives, in Jennifer Lackey and Ernest Sosa, eds. The Epistemology of Testimony. Oxford: Oxford University Press, pp. 77–92. Ginsborg, Hannah, 2006, Reasons for belief, Philosophy and Phenomenological Research 72(2): 286–318. Goldman, Alvin, 1976, Discrimination and perceptual knowledge, Journal of Philosophy 73(11): 771–791. Hinchman, Edward, 2005, Telling as inviting to trust, Philosophy and Phenomenological Research 70(3): 562–587. Hinton, John Michael, 1973, Experiences. Oxford: Oxford University Press. Lackey, Jennifer, 2008, Learning From Words: Testimony as a Source of Knowledge. Oxford: Oxford University Press. Malmgren, Anna-Sara, 2006, Is there a priori knowledge by testimony? Philosophical Review 115(2): 199–241. Martin, Michael, 2002, The transparency of experience, Mind and Language 17(4): 376–425. ———, 2006, On being alienated, in Tamar Gendler and John Hawthorne, eds. Perceptual Experience. Oxford: Oxford University Press, pp. 354–410. McDowell, John, 1998a, Criteria, defeasibility, and knowledge, in his Meaning, Knowledge, and Reality. Cambridge, MA: Harvard University Press, pp. 369–394. ———, 1998b, Singular thought and the extent of inner space, in his Meaning, Knowledge, and Reality. Cambridge, MA: Harvard University Press, pp. 228–259. ———, 1998c, Knowledge by hearsay, in his Meaning, Knowledge, and Reality. Cambridge, MA: Harvard University Press, pp. 414–443. ———, 2002, Knowledge and the internal revisited, Philosophy and Phenomenological Research 64(1): 97–105. McMyler, Benjamin, 2011, Testimony, Trust, and Authority. Oxford: Oxford University Press.

316

Stephen Wright

Moran, Richard, 2005, Getting told and being believed, Philosophers’ Imprint 5: 1–29. Owens, David, 2000, Reason Without Freedom: The Problem of Epistemic Normativity. London: Routledge. ———, 2006, Testimony and assertion, Philosophical Studies 130(1): 105–129. Pritchard, Duncan, 2012, Epistemological Disjunctivism. Oxford: Oxford University Press. Ross, Angus, 1986, Why do we believe what we are told?, Ratio 28(1): 69–88. Snowdon, Paul, 1980–1981, Perception, vision and causation, Proceedings of the Aristotelian Society 81: 175–192. ———, 1990, The objects of perceptual experience, Proceedings of the Aristotelian Society (supp. vol.) 64: 121–150. Strawson, Peter, 2011, Philosophical Writings. Oxford: Oxford University Press. Wright, Stephen, 2018, Knowledge Transmission. London: Routledge.

16 Epistemological Disjunctivism Perception, Expression, and Self-Knowledge Dorit Bar-On and Drew Johnson

1. Introduction We seem to have privileged knowledge of our own present states of mind. As I say (out loud or to myself): “What a terrible headache I have!”, or “I’d love a cup of tea!”, I know as no one else does that I have a headache or would like a cup of tea. As a subject of mental states, I seem to be in a better position than anyone else to know how I feel or what I want or what’s on my mind at a given moment. And I seem to enjoy this privileged epistemic position despite not having any justification (arising from, e.g. inference, evidence, or observation) for the relevant self-beliefs. The relevant self-beliefs can count basic.1 We are here interested in views of so-called basic self-knowledge according to which, although basic selfbeliefs are indeed not arrived at on any particular epistemic basis, they are nevertheless—in a sense that will be made clear—epistemically warranted; what provides the warrant for an individual’s basic mental selfbelief is the very mental state the belief is about.2 (So, for example, my self-belief that I’m feeling anxious right now is not only made true by my presently feeling anxious; my belief is also epistemically warranted by my being in that state.) Indeed, it is this that enables my self-belief to be an instance of knowledge.3 On this view of self-knowledge, false self-beliefs are not only false; they are also epistemically less warranted than the corresponding true selfbeliefs, because there are no first-order mental states either to make them true or to supply them with warrant. Views that follow this approach are committed to a form of epistemological disjunctivism about self-knowledge, because they entail that (internalist) epistemic warrant can vary between a true self-belief and its corresponding false self-belief. Epistemological disjunctivism is usually discussed in the context of perceptual knowledge. According to one such account (Pritchard 2012), in paradigmatic cases of perceptual knowledge, S’s perceptual belief that p is warranted by S’s seeing that p (in the case of visual perception), where ‘seeing that’ is factive (that is, ‘seeing that p’ entails p), and where S has reflective access to this warrant for her belief. When S’s perceptual belief

318

Dorit Bar-On and Drew Johnson

is false, S does not see that p (because that would entail p), but only seems to see that p. Thus, epistemological disjunctivism about perceptual knowledge, like the approach to self-knowledge mentioned earlier, hold that S’s veridical belief that p has a different epistemic warrant from the warrant it would have had if p had been false. In section 2, we briefly contrast metaphysical disjunctivism about perceptual states (MDP) with epistemological disjunctivism about perceptual knowledge (EDP). We go on to present what we take to be a serious challenge to metaphysical disjunctivism about perception that also poses a prima facie challenge to epistemological disjunctivism—a challenge due to Tyler Burge (2005, 2011). In section 3, we offer some general comparisons between basic self-knowledge and perceptual knowledge, and we discuss the sort of warrant appropriate for basic mental self-belief. In section 4, we consider a version of epistemological disjunctivism about basic self-knowledge (EDSK) that is associated with constitutivist views of self-knowledge and argue that it is not subject to the same challenge that faces EDP. However, there are difficulties with constitutivism, which speak in favor of considering an alternative. In section 5, we consider an alternative, neo-expressivist account that—we argue—has several advantages over constitutivism when it comes to articulating an acceptable version of EDSK. We conclude (in section 6) by briefly considering whether an acceptable disjunctivism about mental self-beliefs can point the way toward a sensible disjunctivism about perceptual beliefs.4

2. Metaphysical and Epistemological Disjunctivism about Perception It is standard, in discussions of disjunctivist accounts of perception and perceptual knowledge, to distinguish the following two cases. (For simplicity, we will focus primarily on visual perception.) The Good Case: S seems to see an orange before her. S’s sense organs and perceptual system are functioning properly, and the lighting and other environmental conditions are conducive to accurate vision. There is an orange before S, and S does in fact see the orange. And so, if S has a perceptual belief that there’s an orange in front of her, that belief will be true. The Bad Case: S seems to see an orange before her. S’s sense organs and perceptual system are functioning properly. However, S is undergoing a perceptual referential illusion where there is no orange (suppose there is a hologram of an orange instead, for instance), or is perceiving some other object that S cannot perceptually distinguish from an orange (e.g. a red grapefruit from a distance). There is no orange before S; S does not in fact see an orange.5 And so, if S has a

Epistemological Disjunctivism

319

perceptual belief that there’s an orange in front of her, that belief will be false. According to metaphysical disjunctivism about perception (MDP), as between a veridical perception of an orange and a perceptual experience as of an orange caused by, say, a hologram, there is no metaphysically relevant kind of perceptual experience in common. Whatever commonalities there are between the experiences, they are not to be taken as underwritten by a single type of perceptual experience that the subject undergoes.6 2.1 Burge’s Objection to Disjunctivism Tyler Burge (2005, 2011) has argued against disjunctivism about perception, which he takes to be the following view: [T]here is never an explanatorily relevant mental state type in common between (and specific to) a veridical perception and a referential perceptual illusion . . . [a]nd . . . there is never a mental state type in common between (and specific to) perception of an object and perception of a would-be duplicate substitute for the object that would be, in the context, perceptually indiscernible to the perceiver. (Burge 2005: 25)7 Burge’s target seems to be metaphysical disjunctivism about perception (MDP). He goes on to argue that MDP is incompatible with what the science of perception tells us about perceptual experience. The relevant point that Burge takes the science to support is what he calls the Proximality Principle: The Proximality Principle Holding constant the antecedent psychological set of the perceiver, a given type of proximal stimulation (over the whole body), together with associated internal afferent and efferent input into the perceptual system, will produce a given type of perceptual state, assuming that there is no malfunctioning in the system and no interference with the system. On any given occasion, given the total antecedent psychological state of the individual and system, the total proximal input together with internal input into the system suffices to produce a given type of perceptual state, assuming no malfunctioning or interference. (Burge 2005: 22, emphasis in original) Consider again the good and the bad case. Assuming the conditions of the Proximality Principle are met, the Principle tells us that the perceptual state kind in the good case is the same as in the bad case. So

320

Dorit Bar-On and Drew Johnson

disjunctivism is committed to accepting, while perceptual psychology denies, “that types of perceptual state can vary even when there is no change in proximal stimulation, internal input, and antecedent psychological states remain the same” (Burge 2005: 22).8 If, as Burge claims, MDP is indeed incompatible with well-established empirical research in perceptual psychology, then MDP should be rejected.9 Note, however, that even if one acknowledges a metaphysical commonality in perceptual experience between the good and the bad case, it remains to be seen what—if any—epistemological significance should be assigned to this commonality. Indeed, some self-proclaimed disjunctivists set aside metaphysical disjunctivism and instead focus directly on epistemological disjunctivism, which is specifically concerned with how the epistemic status of perceptual belief might vary between good and bad cases. For example, in articulating his preferred (McDowell-inspired) form of disjunctivism, Duncan Pritchard makes clear from the outset that his is a form of EDP, which he takes to be independent of MDP. The core thesis he puts forward is this: EDP: “In paradigmatic cases of perceptual knowledge an agent, S, has perceptual knowledge that ϕ in virtue of being in possession of rational support, R, for her belief that ϕ which is both factive (i.e., R’s obtaining entails ϕ) and reflectively accessible to S”. (Pritchard 2012: 13) In the good case, according to (this version of) EDP, S can see that p. ‘Seeing that’ is factive, so S’s seeing that p entails p. Moreover, disjunctivism adds, in the good case, S has reflective access to this factive reason, and it is in virtue of this reflective access that S’s belief is justified in a way conducive to knowledge. When I believe there is an orange before me in the good case, I can know by reflection alone that what justifies my belief is that I see that there is an orange before me. Pritchard maintains that EDP represents the commonsense view of perceptual knowledge, remarking that when one reports a perceptual belief, and someone challenges this belief (“How do you know there is an orange before you?”), a natural response to this challenge is to cite one’s factive reason (“because I can see that there is an orange before me”) (Pritchard 2012: 17–18). Now, it’s true that one would attempt to offer the same reason in the bad case as well, though in that case one’s reason won’t be factive (since there’s no orange to be seen). But the point is that, even in bad cases, we advert to what would be factive reasons in support of our knowledge claim. We do not offer as reasons what might in fact be common to both the good and the bad cases, namely, its seeming to us that we see that there is an orange. 2.2 EDP’s Relation to MDP If Burge is right that MDP is incompatible with empirical research in perceptual psychology, then, if it turned out that EDP entails MDP, EDP

Epistemological Disjunctivism

321

would have to be rejected, too. Yet it seems that EDP is at least logically compatible with the denial of MDP.10 At the same time, as Pritchard himself observes, it is difficult to see how EDP could be plausible without MDP (2012: 24). The difficulty in accepting EDP while denying MDP stems from the following lingering intuition. Suppose we hold fixed S’s psychological states, the proximal stimulations of S’s sense organs, etc., and only vary the distal causes of the proximal stimulations (thereby affecting whether S’s resulting perceptual state is veridical or not). Then there seems to be no principled way for the epistemological disjunctivist to explain in virtue of what the warrant for the perceptual belief could vary between the good and the bad case. Accepting MDP makes EDP much more plausible, because then one can say, quite naturally, that the difference in the warrant for belief in the two sorts of cases supervenes on differences in the nature of the perceptual state involved.11 In the case of perceptual knowledge, it seems clear that one’s perceptual experience has an important role to play in supplying the warrant for perceptual belief. As Pritchard recognizes, it is natural to offer the reason that one sees that p in support of one’s belief that p, as opposed to simply offering p itself as the reason. Indeed, it seems that the involvement of one’s perceptual experience is crucial for the relevant belief being perceptual in the first place. This suggests to us the following principle: The Perceptual Warrant Principle (PWP) For a perceptual belief B formed on the basis of perceptual experience e, the epistemic warrant for B is wholly constituted by e. PWP seems to us like an intuitively plausible principle, not to be rejected without some argument. Now, one might suggest, as an alternative, that the warrant for a perceptual belief B is secured through some inference of B from e. That is, one’s perceptual beliefs, it might be thought, are based on one’s perceptual experiences in the sense of being inferred from them. But this over-intellectualizes perceptual belief. It seems reasonable to credit non-human animals and young children with warranted perceptual beliefs, yet it may well not be plausible to credit them with the conceptual and rational capacities necessary for making the relevant inferences. Instead, it is more plausible to think (along the lines of PWP) of perceptual experience itself as providing the warrant for perceptual belief. This is something that epistemological disjunctivism can (and we think should) accept.12 However, if EDP accepts this principle, it must add that perceptual experience in good cases is factive and reflectively accessible, while perceptual experience in bad cases is not factive. And this will likely involve commitment to MDP.13 The anti-disjunctivist point here is that it is mysterious how the contribution one’s perceptual experience makes to one’s warrant for belief could vary when the type of perceptual experience one is in is held fixed.

322

Dorit Bar-On and Drew Johnson

If one accepts the Proximality Principle (see subsection 2.1), and thus rejects MDP, then, if one also accepts PWP (i.e., that the perceptual experience constitutes the epistemic warrant for the perceptual belief), there is simply no room for pulling the good and the bad cases apart in terms of epistemic warrant. Our tentative conclusion is that EDP is a problematic view.14 In what follows, we aim to articulate a plausible version of epistemological disjunctivism about basic self-knowledge (EDSK). If we are to be successful, we shall have to show that EDSK at least is not susceptible to the challenge we presented for EDP.

3. Epistemological Disjunctivism and Basic Self-Knowledge: Some Preliminaries Epistemological disjunctivism about basic self-knowledge (EDSK) is the view that the warrant one has for a basic self-belief about a (current) state of mind—e.g. that one has a headache, or would like a cup of tea, or is thinking about the third premise of an argument—varies between veridical and non-veridical cases. In a veridical case, when one’s self-belief that one is in M is true, because one is in M, one has knowledge, since the belief is not only true but is also warranted by the state M itself. Being in M in some way provides one with the appropriate epistemic warrant for one’s self-belief. In a non-veridical case, one has no knowledge, not only because one’s self-belief is not true, but also because it is not warranted (or, at least, it lacks the kind of warrant had in the veridical case). In sections 4 and 5, we consider two specific versions of EDSK: one constitutivist, the other neo-expressivist. We will be arguing that: 1. These versions of EDSK are not susceptible to the challenge earlier presented for EDP (subsection 2.1), so one cannot object to EDSK based on this challenge; and 2. As between the two versions of EDSK—constitutivism and neoexpressivism—neo-expressivism is the more plausible view, insofar as it avoids other problems facing constitutivism. However, before turning to these versions, some clarifications are in order. 3.1 Perceptual Beliefs vs. Basic Mental Self-Beliefs We begin by offering some general comparisons between perceptual beliefs and basic mental self-beliefs. •

First, if basic self-knowledge were thought to be simply a species of perceptual knowledge—as per certain views of introspection15—then EDSK would be just a special case of EDP.

Epistemological Disjunctivism •

•

•

323

However, second, the literature is rife with arguments against the applicability of the perceptual model to basic self-knowledge.16 In any event, there is a clear disanalogy between the perceptual case and the case of basic self-knowledge. It is intuitively unobvioius what environmental conditions could be relevant to one’s possession of mental self-beliefs in the first place; whereas, in the perceptual case, certain specific environmental conditions must be met in order for one to have the relevant belief. Moreover, it is unobvioius what would be the analog in the self-knowledge case of the condition that “S’s sense organs and perceptual system are functioning properly”. Since the two versions of EDSK we go on to consider both reject the perceptual model, we will simply set it aside here. (For the same reason, we will further assume that inferential models of basic self-knowledge should also be rejected.)17 Third, views that reject the perceptual model of basic self-knowledge must deny that there is a perceptual experience in common between the good and the bad case of basic self-knowledge, because they hold that there is no perceptual experience involved in this type of knowledge at all. The views we go on to consider deny, in addition, that basic mental self-beliefs have any other source of epistemic justification (such as inference). A natural remaining candidate for warrant for basic self-beliefs on these views would seem to be the very mental states the beliefs concern—which is obviously not something in common between veridical and non-veridical belief. And this would commit them to EDSK.18 Yet, fourth, if the perceptual model is rejected, at least one commonsense motivation for EDP will not carry over to the case of selfknowledge. Recall Pritchard’s observation that the reason we would normally offer for a perceptual belief is a factive one: “How do you know that p?”—“I see that p”. Pritchard takes this to provide some positive support EDP. But if I were to ask you “How do you know you’d like a cup of tea?”, the natural answer would not be “I see/feel that I’d like a cup of tea . . .”. To the extent that the question merits an answer, the sensible answer would be: “Well, because I would like a cup of tea”.19 Still, this last answer itself can point toward a (related) commonsense support for EDSK. For it suggests that the reasons we would ordinarily cite for basic mental self-beliefs are the very facts that would make our self-beliefs true; we do not invoke some selfexperience (e.g. “I seem to be/feel as though I am in pain”), in contrast with the perceptual case, where we do invoke our perceptual experiences. And this, if taken at face value, suggests that one’s belief that one is in M is warranted simply in virtue of one’s being in M. But, as this is a factive reason—one cannot have that same reason if one is not in M—EDSK follows.

324 •

Dorit Bar-On and Drew Johnson Finally, fifth, although we are concerned to explain the special epistemic security of basic self-knowledge, we by no means take mental self-beliefs to be infallible. Indeed, we take it that, at a minimum, a disjunctivist view of basic self-knowledge, just like a disjunctivist view of perception, would need to make room for the possibility of veridical and non-veridical cases of basic self-beliefs that are indistinguishable from the self-believer’s perspective.

3.2 Varieties of Epistemic Warrant Given the focus of epistemological disjunctivism on the question of warrant for belief, it will be helpful to mark some distinctions among species of epistemic warrant. These distinctions will allow us to clarify how the views we consider in the next two sections explain the distinctive nature of self-knowledge. First, it should be noted that EDP (at least on Pritchard’s and McDowell’s versions) adopts an internalist notion of warrant for veridical perceptual beliefs. A central feature of Pritchard’s EDP is the claim that in paradigm cases of perceptual knowledge, one has reflective access to the rational support for one’s perceptual belief.20 This commitment to reflective access is what qualifies EDP as an internalist view.21 The internalist component of EDP is part of what makes it an interesting view; what is interesting and distinctive about EDP is the combination of these ideas: (i) reasons for perceptual beliefs are reflectively accessible, and (ii) the reasons for veridical perceptual beliefs are factive. This combination would be denied by traditional internalism, since that view would contend that reflectively accessible reasons cannot be factive in the case of perception, even if the reasons are sufficient for knowledge. (Pritchard 2012: 38).22 The relevant point here is that if the views about basic self-knowledge that we go on to consider are to count as forms of epistemological disjunctivism (on the model of Pritchard’s EDP), we should expect them to invoke an internalist conception of warrant as well. So, in articulating the view of basic self-knowledge we prefer—the neo-expressivist view—we will assume a form of internalism that (we hope) should be acceptable to internalist epistemological disjunctivists. According to this form of internalism, to anticipate, for S to be internalistically warranted in believing p, S must be capable of citing, upon reflection, an essential reason R for her belief that p. (The notion of an essential reason here—adapting Pappas’ notion of ‘essential justifier’ (2014)—is that of a reason without which there would be insufficient warrant for belief.) Second, we would like to reject the epistemological assumption that when a belief is warranted, its warrant must consist in some particular epistemic basis, or method, that an agent employs in order to arrive at the

Epistemological Disjunctivism

325

belief. This assumption is what Bar-On calls the Distinct Epistemic Basis requirement, which (as applied to self-knowledge) states: If I can be said to have privileged knowledge that I am in a certain state of mind, then this knowledge must have some distinct epistemic basis; there must be a special epistemic method or route (a special ‘way of knowing’) that I use to obtain this knowledge. (2004b: 344; emphasis in original) Applied to knowledge generally, the Distinct Epistemic Basis requirement implies that it would be incorrect to say that one knows something unless there is a way that one knows—some distinct method one employs, or some ‘evidential work’ one engages in, in arriving at the relevant belief.23 Distinct epistemic bases for belief include: inference, testimony, observation, memory, etc.—in short, methods of forming a belief that confer justification on that belief. In rejecting the Distinct Epistemic Basis requirement, we open the possibility that some beliefs may count as warranted, even if the believer has not employed any distinct method or done any evidential work. One can be warranted in holding beliefs that one just finds oneself with, without having engaged (now or at any point in the past) in some justifying belief-forming method. So, although employing a distinct epistemic method in arriving at a belief is sufficient for warranting the belief to some extent, it is not necessary for warrant. We thus follow several authors in distinguishing two species of epistemic warrant: justification and entitlement.24 We understand justification internalistically, as requiring (i) that one employ a distinct epistemic basis in forming a belief, and (ii) that the believer must have reflective access to the basis for her belief.25 We understand entitlement to represent a way for belief to be warranted in the absence of (i): one can be entitled even if one has done no ‘evidential work’ (or has employed no special method) to earn that entitlement (Wright 2004: 174). This way of thinking about entitlement captures the rejection of the Distinct Epistemic Basis requirement articulated in the previous paragraph. On Dretske’s understanding of entitlement, we are entitled to beliefs that we cannot avoid having (Dretske 2000). Wright (2004) holds that although one may accept the propositions to which one is entitled, entitlement does not provide any justificatory evidence for the truth of those propositions. A worry for these views of entitlement is that they do not explain what makes entitlement an epistemically good-making feature— they do not assign a clear positive epistemic status to entitlements—one that, for example, systematically connects the entitlement with truth. However, we think there is a category of entitled belief that is appropriately connected with truth: we call this category grounded belief. Here ‘being grounded’ is to be understood in the sense of being anchored,

326

Dorit Bar-On and Drew Johnson

rather than in the sense of being made on some specific grounds.26 The grounding for one’s belief in this sense can provide reason for belief, albeit not the sort of reason envisaged by the Distinct Epistemic Basis requirement. The views of self-knowledge we consider (constitutivism and neoexpressivism) will take basic self-beliefs to provide paradigm examples of grounded (in the present sense) yet base-less belief. One might worry that views that take basic self-beliefs to be beliefs we are entitled to (in virtue of their being grounded) but for which we have no justification (because they do not satisfy the Distinct Epistemic Basis requirement) cannot count as forms of epistemological disjunctivism. As discussed earlier, EDP employs an internalist notion of epistemic warrant, but on some views of entitlement (e.g. Burge 1993, 1996, 2003, 2013), entitlement is an externalist type of warrant. However, the notion of entitlement we employ in this chapter is neutral with respect to the (epistemological) internalism/externalism debate. It is at least a theoretical possibility for grounded beliefs to be internalistically warranted. While (in our terminology) such beliefs would not count as epistemically justified, they would still be internalistically warranted. In the case of basic self-knowledge, all that would be required for self-belief to be internalistically warranted is for one to have reflective access to that which provides the grounding for one’s self-belief, and all that requires is that one be able to provide her reason for the belief in terms of that grounding. Granting that one’s self-beliefs are grounded in the mental states that those self-beliefs concern, and granting (as is plausible) that one can typically cite her mental states when providing reasons for her mental self-beliefs, it will turn out that we have reflective access to the grounding for our basic self-beliefs. We are now in a position to evaluate two different views that embrace disjunctivism about basic self-knowledge.

4. Constitutivism About Basic Self-Knowledge 4.1 Constitutivism Proponents of constitutivism maintain that spontaneous mental selfbeliefs count as knowledge courtesy of constitutive connections they bear to the states of mind they are about.27 Believing and other states of mind—both ‘evaluative’ attitudes, such as intending, desiring, hoping, and ‘receptive’ experiential states—are ‘intrinsically known’. We have privileged knowledge of such states “not because we are in a specially good position to form second-order beliefs about them that reliably track their existence, but because their existence is normally constituted by our knowing assessment” of them (Boyle 2011: 237). Thus, “in the normal and basic case, [e.g.] believing P and knowing oneself to believe P are not two cognitive states; they are two aspects of one cognitive state” (2011:

Epistemological Disjunctivism

327

228).28 This means that the constitutivist is committed to a strong thesis of self-intimation: it is metaphysically impossible for you to be in a state of mind M and not know it, since your M, and your self-belief that you are in M, are in effect not ‘distinct existences’. (For this reason, it is a mistake to think of basic self-knowledge, as do introspectionist views, in terms of a kind of perception-like tracking of “some independent condition” (2011: 239)). Constitutivism, then, takes the epistemic status of (true) self-beliefs as knowledge to be underwritten by the metaphysics of mental states: these states have beliefs ‘built into’ them. Possessing a basic self-belief that you are in a mental state is simply guaranteed by your being in that state. In particular, mental self-beliefs are not understood as ones you arrive at using observation, inference, or any other epistemic method; for having these (second-order) beliefs is a constituent or aspect of being in the relevant states. But, despite having no epistemic basis, these beliefs—according to the version of constitutivism of interest to us here—are still epistemically warranted. For they can be said to be epistemically grounded (in our sense—see subsection 3.2) in the first-order mental states they are about. Thus, on the constitutivist view we consider here, the first-order states make the second-order beliefs both true and epistemically warranted.29 This form of constitutivism is committed to a form of epistemological disjunctivism: when one has a second-order belief, either that belief is epistemically grounded in the first-order mental state that makes it true, or there is no first-order mental state that makes the belief true, and so no first-order mental state to provide epistemic grounding for the belief. So a true and a false self-belief that one is in M have different epistemic grounding; the true belief will be grounded in the state it is about, while the false belief cannot be. If a self-belief’s epistemic warrant were to depends only on its epistemic grounding, a true and a false self-belief could not share their epistemic warrant; hence epistemological disjunctivism. As noted earlier, constitutivists reject perceptual (as well as inferential) models of basic self-knowledge. There is, on this view, a contrast between basic self-knowledge and perceptual knowledge, in that mental self-beliefs are not metaphysically independent of the states they are about the way perceptual beliefs are independent of the observable facts they are about.30 Whereas in the perceptual case, given the metaphysical independence of the relevant belief and fact, there is an ‘epistemic distance’ one must traverse to attain knowledge (one’s perceptual beliefs must hook up to the worldly facts in the right way), there is no such distance to traverse in the case of mental self-beliefs. 4.2 Constitutivism and EDSK We are now in a position to see how the constitutivist avoids the challenge presented earlier for EDP. On our diagnosis, the challenge arises because,

328

Dorit Bar-On and Drew Johnson

if we follow Burge in insisting that we must acknowledge the presence of an explanatorily relevant perceptual state in common between the good and the bad cases (to avoid contradicting empirical psychology), it becomes very difficult to deny the epistemic relevance of such a state. Even if perceptual beliefs are not arrived at via inferences from perceptual experiences, it seems difficult not to regard the perceptual beliefs as epistemically warranted by these experiences. The difficulty for EDP is then to explain how the good case and the bad case can differ in terms of their epistemic warrant, given that they are based on perceptual experiences of the same type. But, according to the constitutivist, mental selfbeliefs are not epistemically based on anything; in particular, they are not based on any (possibly shared type of) experience concerning how one’s mental state appears to one. So there is no pressure to acknowledge an epistemically relevant psychological commonality between the true and false self-believer in terms of some perceptual self-experience. But then the road is clear for the constitutivist to propose that the good and bad cases differ in terms of their epistemic warrant. In the good case, S’s basic self-belief is epistemically warranted, because grounded in the relevant first-order mental state (which is also what makes the self-belief true). Not so in the bad case. Whereas EDP faces an apparently insurmountable challenge, the constitutivist’s EDSK does not. However, the constitutivist idea that being in a mental state is metaphysically inseparable from believing that one is in it is problematic, and there are several extant objections to constitutivism both as a view about the metaphysics of mental states and as a view of basic self-knowledge. For reasons of space, we mention only one problem that is the most relevant to our concerns here.31 Consider a false belief that one is in M. A false mental self-belief cannot be properly ‘built into’—or be inseparable from—the state of mind it is about. Being false, there is no state of mind for it to be ‘built into’! So the constitutivist needs to tell us what to make of the straightforward possibility of false basic mental self-beliefs. However the constitutivist accommodates such self-beliefs, we think she faces the following prima facie difficulty. (One main advantage we see for the alternative account we consider in the next section is its potential for avoiding this difficulty.) A key component in the commonsense explanandum of basic self-knowledge is the privileged epistemic status of basic mental self-beliefs—the fact that, unlike other beliefs, they are strongly presumed to be true and to enjoy a special entitlement (even though mental self-believers are not expected to possess, or required to offer, any evidence for the beliefs).32 The constitutivist purports to explain this special epistemic status by appealing to the metaphysics of mental states: the fact that mental states are partially constituted by self-beliefs. But that explanation can now be seen as insufficient, in light of the possibility of false self-beliefs. A false basic mental self-belief, on the constitutivist account, is not only false; it also lacks

Epistemological Disjunctivism

329

all warrant, since the only source of epistemic warrant for a basic belief that one is in M (given EDSK) is the state that grounds it (M). But this means that false and true mental self-beliefs cannot be said to enjoy the same secure epistemic status—the status we ordinarily accord them. That status, it seems, is indifferent to whether the belief is true or false. A false (basic) mental self-belief would still seem to enjoy the same presumptive privileged status that would be accorded to a corresponding true belief— simply in virtue of being a (basic) mental self-belief. (This is a point of disanalogy between ordinary mental self-beliefs, on the one hand, and perceptual as well as third-person mental beliefs, and bodily self-beliefs, on the other. See section 5 of this chapter.) This means that an acceptable version of EDSK has to make room for some epistemic commonality between true and false mental self-beliefs. Thus, whereas EDP (as we saw earlier) struggles to explain in virtue of what perceptual beliefs in the good and the bad cases vary in terms of epistemic warrant, constitutivism faces a difficulty in explaining what mental self-beliefs in good and bad cases have in common in terms of their epistemic warrant. (As we shall argue in the next section, neoexpressivism is—by contrast—well placed to capture both the epistemic similarities and the epistemic differences between true and false mental self-beliefs.)

5. Neo-Expressivism and Basic Self-Knowledge We now turn to an alternative view of basic self-knowledge— neoexpressivism—which, like constitutivism, endorses a form of EDSK, but which avoids the problem facing constitutivism mentioned in the previous section. Like constitutivism, neo-expressivism rejects the perceptual model of basic self-knowledge and denies that basic self-beliefs must be arrived at via some distinct epistemic route or formed on some distinct epistemic basis in order to qualify as instances of genuine (and privileged) knowledge. This allows neo-expressivism also to avoid the main challenge facing EDP. However, unlike constitutivism, neo-expressivism is not committed to the idea that there is a metaphysical dependence between mental self-beliefs and the states they are about, and so it can straightforwardly acknowledge the possibility of false basic self-beliefs and can account for the epistemic commonality between true and false mental self-beliefs. 5.1 Neo-Expressivism Bar-On (2004b, 2011, 2009, 2012, and elsewhere) develops an account of the distinctive epistemic security of avowals (where avowals are understood to be spontaneous, non-reflective, or ‘unstudied’, self-attributions of occurrent mental states that are made on no distinct epistemic basis).

330

Dorit Bar-On and Drew Johnson

On Bar-On’s view, avowals are distinctive acts that serve to express in the act sense (‘a-express’) the very mental state that the avower self-attributes, in addition to expressing the (higher-order) belief that she is in the state. So, for example, in avowing (out loud or silently) “I am so glad to see you” you a-express your delight at seeing your addressee, and not just your belief that you are happy to see her. This renders the avowal—as an act—similar to expressive acts such as saying (or thinking) “It’s so good to see you”, or just giving a hug, though understood as a product— or in terms of the expressive vehicles used—it is different semantically from these other expressions. (“I’m so glad to see you” expresses in the semantic sense—‘s-expresses’—a self-ascription, true iff you are glad to see your addressee, whereas “It’s so good to see you” s-expresses a nonself-ascriptive proposition, and the hug does not s-express any proposition.) Bar-On argues that her account enables us to capture the intuitive epistemic contrasts between avowals and all other attributions that merely report contingent matters (including third-person mental attributions, perceptual judgments, and all bodily self-attributions as well as all evidential mental self-attributions). Notably, the account helps explain why avowals are not open to ordinary doubts, epistemic challenges, or requests for reasons; they are not subject to simple correction or defeat. (This is a point the constitutivist apparently struggles to explain in the case of false avowals). At the same time, it also accommodates the semantic continuities between avowals and truth-evaluable pronouncements, as well as semantic differences between avowals and other expressions. Now, the question often at the center of discussions of basic selfknowledge is how our mental self-beliefs can be taken to manifest things we know about ourselves, and know in a unique and privileged way. This question is especially pressing for any view that—like neo-expressivism— rejects the Distinct Epistemic Basis requirement and maintains that the special epistemic status of our avowals is not due to the fact that the selfbeliefs they express are arrived at through the use of a special epistemic method (such as perception, evidential inference, self-interpretation, etc.). 33 As regards accounts of basic self-knowledge that, specifically, explain the special status by appeal to the reliability of an internal perception-like mechanism, Bar-On argues that the best such accounts can do is treat the contrasts between avowals and other ascriptions as being on a par with the epistemic contrasts between first-person and third-person attributions of certain bodily states. However, assimilating the epistemic status of avowals to that of bodily self-reports fails to do justice to the distinctive security of basic mental self-beliefs as ordinarily understood.34 5.2 Immunity to Error35 Consider self-attributions issued through proprioception or kinesthesis. Such self-attributions share a certain epistemic feature with avowals: they

Epistemological Disjunctivism

331

are ‘identification-free’. In normal circumstances, if I say or think: “My legs are crossed” or “I’m spinning around”, then, epistemically speaking, my self-attribution does not rely on my recognizing some individual as myself and taking that person’s legs to be crossed, etc. (Indeed, I have no more reason for thinking that someone’s legs are crossed than whatever reason I have for thinking that my legs are crossed.) To use a phrase due to Gareth Evans and Sydney Shoemaker, bodily self-attributions of this kind are “immune to error through misidentification” (IEM, for short).36,37 When a self-attribution of the form “I am F” is IEM, then, although I may fail to be F—so my self-attribution may be false—there is no room for me to think: Someone is F, but is it me? I cannot sensibly doubt that it is me who has the relevant property F without doubting that someone has it. (Contrast this with a case in which I e.g. tell that I have $500 in my bank account by consulting the bank teller’s screen. Here, I can wonder whether someone—though not me—has $500 in her account.) Importantly, immunity to error through misidentification does not reflect a special recognitional success, that is, success in identifying— by recognizing—the “right” individual of whom to predicate F. Quite the opposite: if anything, this kind of immunity reflects the absence of recognitional identification.38 Still, on Evans’s analysis, self-attributions that are IEM can represent knowledge that we gain about ourselves in a distinctive way. Evans notes that we possess two general capacities for gaining information about some of our own states and properties—the capacity to “perceive our own bodies” (which includes “our proprioceptive sense, our sense of balance, of heat and cold, and pressure”), and the capacity to determine our own “position, orientation, and relation to other objects in the world . . . upon the basis of our perceptions of the world”.39 When a subject gains information of the form “I am F” (for the relevant range of F’s) using one of these capacities, Evans remarks, [t]here just does not appear to be a gap between the subject's having information (or appearing to have information), in the appropriate way, that the property of being F is instantiated, and his having information (or appearing to have information) that he is F; for him to have, or to appear to have, the information that the property is instantiated just is for it to appear to him that he is F. (1982: 221) So self-attributions of the form “I am F” that are IEM can represent secure knowledge I have that I myself am F, even though they do not rest on recognitional identification of the one who is F. Now, Bar-On’s neo-expressivist account of basic self-knowledge begins with the suggestion that paradigmatic, present-tense, non-reflective mental self-attributions of the form “I am in M” or “I’m M(ing) (that) c” (where M is a mental state and c is its putative intentional content)40 are

332

Dorit Bar-On and Drew Johnson

issued ‘in the avowing mode’: they are not made on the basis of (inner) observation, tacit deliberation, evidence, or interpretation. And they are immune to error through misascription—IEM2—in addition to being immune to error through misidentification—(which we’ll now label) IEM1.41 Suppose I produce a self-attribution such as “I am nervous about this dog”, or “I am hoping that you’ll join us tonight”, or “I am feeling tired”. Epistemically speaking, my self-attribution does not rely on my recognition of a state in me as having a certain character and content, based on how it appears to me (say, as a result of introspection). I do not recognize a state of mine as being M (as opposed to M’) and having content c (rather than c’). (Indeed, I have no reason—specific to the occasion on which I issue the avowal—for thinking that I am in some state or other, and that it has some content, other than whatever reason I have for just thinking that I am in mental state M with content c.)42 Consequently, in making the self-attribution, I am not subject to the possibility of a brute recognitional error—where a brute error is one that is simply due to the world failing to cooperate, rather than being due to some kind of failure of the subject’s faculties.43 This is in good part what makes the perceptual model inappropriate for mental self-beliefs. Note, however, that to say that (basic) mental self-attributions are not open to brute recognitional errors of identification or ascription is not to say that they are absolutely infallible or incorrigible. It is just to say that they are protected from a much wider array of epistemic errors, doubts, and corrections than other attributions, including, specifically, proprioceptive and kinesthetic self-reports.44 Interestingly, perceptual beliefs are open to brute error; a perceptual belief can be false even when the believer’s perceptual faculties are working perfectly. Accordingly, perceptual beliefs are not immune to brute recognitional error in the way that basic mental self-beliefs are. An adequate disjunctivist account of basic self-knowledge should explain the immunities to error of basic mental self-beliefs, but there is no pressure for EDP to provide a parallel explanation for perceptual beliefs. The characterization of avowals’ distinctive security as a matter of their being IEM2 (in addition to IEM1) provides a suitably tempered interpretation of the familiar claim that our own basic mental self-beliefs are absolutely protected from epistemic challenge (doubt, correction, falsification). The dual immunity to error can in part explain why, if we consider your present thought that you are, say, feeling disappointed by your friend, the self-belief it manifests appears at once to lack any distinct epistemic basis and yet to be especially likely to be something you know (as no one else does). And it explains why basic mental self-attributions are not subject to perception-like brute recognitional errors.45 The above characterization does not commit one either to Cartesian privileged access, or, indeed, to any distinctively secure epistemic basis on which avowals supposedly are made. Moreover, it does not require supposing that the secure epistemic status of these self-attributions is

Epistemological Disjunctivism

333

vouchsafed via some conceptual or metaphysical guarantee, as per constitutivism. However, by itself, the characterization does not yet give us a full account of avowals’ epistemic status. For one thing, we need to understand the source of the additional immunity to error enjoyed by basic mental self-attributions.46 Why is it that they are not only IEM1 but also IEM2? In addition, we have seen that immunity to error, in general, is no guarantee of truth. So we need to understand why basic mental self-attributions contrast with all other ascriptions in being so strongly presumed to be true; and, given that they are made on no epistemic basis, we need to know what provides positive epistemic warrant for them. 5.3 Immunity to Error, Expression, and Knowledge47 By way of appreciating the expressive character of spontaneously produced mental self-attributions, consider a case in which you falsely (though sincerely) avow “My tooth hurts!” (say, at the dentist’s chair, as the drill approaches your mouth). Under the circumstances, you might have equally said: “Ow!”, or emitted a yelp, or winced. Given that, it does not seem plausible to regard the avowal—but not the “Ow!” or the wince—as an upshot of mistaking (say) fear of the approaching drill for pain, resulting in a brute error about your state. You falsely think or judge that your tooth hurts, but you didn’t come by this judgment in consequence of being fooled by the appearance of your internal state. It seems much more plausible to regard the avowal as, just like the yelp, something forced out of you, though in this unusual case not by an actual toothache but rather by the priming effect of fear (fueled, perhaps, by painful dental history). The avowal, as an act, is on a par with the “Ow!” or the yelp/ wince, in being an expressive act. It is no more plausible to regard the avowal as manifesting an evidence- or observation-based self-belief than it is to regard the spontaneous yelp/wince as so based. Although we say that a true basic self-belief is grounded in the mental state the belief concerns, this is not the same as saying that the self-belief has the mental state as a distinct epistemic basis. As we have understood the notion, having a distinct epistemic basis involves employing an epistemic method in forming the relevant belief that one can cite in support of that belief. This is absent in the case of basic self-belief. The self-belief one has when avowing her mental state is a belief that she, in a sense, simply finds herself with. On the neo-expressivist account, one who avows her state of mind is speaking from that state (using a self-ascriptive vehicle), rather than merely reporting (or otherwise informing of) a self-belief she has arrived at on the basis of the way her state appears to her. Insofar as the selfattribution made when avowing is not made on any recognitional basis, it is not subject to brute errors of misidentification or misascription. This is also why self-attributions produced in the avowing mode are so strongly presumed to be true, since to take someone to be avowing M is just to

334

Dorit Bar-On and Drew Johnson

take it that she has given voice to her M—which means taking the selfattribution to be true. The relevant presumption of truth does not amount to a conceptual guarantee of truth; it is defeasible. We can make sense of a spontaneous self-attribution of an occurrent state of mind being false, as in the dentist case, or certain cases of self-deception, implicit bias, and so on. Still, as noted earlier, in such cases we do not suppose the avower to be subject to a purely brute ‘recognitional’ error, of the sort that afflicts one who is subject to a perceptual illusion. Falsity in these cases is not plausibly taken to be due to a subject’s simply mistaking of one of her mental states for another; we expect it to have its source in some psychological irregularity, impairment, failure, or interference. Note that on the present account, there is no special difficulty in accommodating the straightforward possibility of false basic self-beliefs and the doxastic commonality between the good and bad cases of self-knowledge. The person who is subject to self-deception, or to some priming effect, would be just as disposed to (sincerely) avow: “I am in M” as the person who truly believes she is in M. By ordinary standards, that person would be straightforwardly credited with having a self-belief as naturally as the person who has a true self-belief. The neo-expressivist, in contrast with the constitutivist, can make room for this, since she allows what the constitutivist denies: that there is (in general) an ontological independence between mental self-beliefs and the mental states that would make such beliefs true or false. At the same time, like the constitutivist, the neoexpressivist denies that (basic) mental self-beliefs must be epistemically based on internal experiences of appearances, which experiences, moreover, explain the beliefs’ positive epistemic status. 5.4 Neo-Expressivism and EDSK If basic mental self-beliefs indeed have no epistemic basis, must we suppose—as do constitutivists—that such beliefs can only be epistemically warranted insofar as they are epistemically grounded, and that they can only be epistemically grounded if they are true—that is, if the believer is in the state that her belief is about? If so, then epistemic disjunctivism about basic self-knowledge follows. However, our discussion of neo-expressivism so far opens up another possibility.48 Insofar as basic mental self-beliefs are not based on any inference, evidence, or self-experience, it doesn’t seem that one could have justification for the relevant self-attributions (see the earlier discussion of justification in section 3.2). However, this does not mean that one cannot have epistemic warrant for the self-belief, in the sense of being entitled to it. As a subject of bodily states with normal physiological capacities, if I think, in the ordinary way, that my legs are crossed, I can be said to be entitled to believe that it is my legs that are crossed, even if I have relied on no evidence and have done nothing to ascertain the identity

Epistemological Disjunctivism

335

of the person whose legs are crossed, say, by ruling out imposters. I can perhaps be said to be entitled by default to this belief.49 Similarly, given the neo-expressivist analysis of avowals’ security (in terms of a distinctive immunity to error), perhaps, as a subject of mental states with normal psychological capacities, I am in a unique position to give voice to a state of mind by self-attributing it, and this renders me entitled by default to the relevant self-belief, even if I have not done any evidential work or have not ruled out alternatives to the attribution I make.50 If it is correct to think that basic mental self-beliefs enjoy entitlement by default, however, one might wonder whether neo-expressivism represents a form of disjunctivism at all. After all, both true and false selfbeliefs enjoy this sort of entitlement, and so in this respect are alike in epistemic warrant.51 Yet the defining feature of epistemological disjunctivism is the commitment to there being a difference in epistemic warrant between belief in the good vs. the bad case. We think that despite positing an epistemic commonality between true and false basic self-belief in terms of entitlement, neo-expressivism can still count as a form of epistemological disjunctivism. This is simply because the epistemic warrant for a true self-belief is not exhausted by entitlement by default; a true self-belief enjoys another sort of epistemic warrant that is not had by the relevant false self-beliefs. Like the constitutivist, the neo-expressivist can insist that a true basic mental self-belief will have epistemic grounding—in the very state the belief is about (in addition to the believer being entitled to it by default). Not so when it comes to the corresponding false self-belief. When one is not in M, one can still believe that one is in M, and even be entitled by default to that belief; but one is not (fully) warranted, since one’s belief fails to be epistemically grounded.52 Note that this resulting view constitutes what one might call a weak epistemological disjunctivism. A strong epistemological disjunctivism would deny that there is any epistemic warrant shared by beliefs in good and bad cases. A weak epistemological disjunctivism would allow that belief in good and bad cases can share some sources of warrant, yet still have different epistemological statuses: beliefs in bad cases will not be fully warranted (in a way that allows for knowledge), while beliefs in good cases are fully warranted (in a way that renders them instances of knowledge). One advantage of the weaker version is that it straightforwardly accommodates the possibility of false basic mental self-beliefs, and allows that such self-beliefs can be as epistemically ‘innocent’—and even enjoy some of the same entitlement—as their true counterparts, without invoking any epistemically relevant self-experiences. At the same time, it acknowledges that being in the relevant mental state has bearing on the epistemic support one has for one’s self-belief (without bringing the self-belief in its train). This can be seen as an advantage, inasmuch as it captures something about subjects’ epistemic position in

336

Dorit Bar-On and Drew Johnson

ordinary circumstances: both the true and the false believer, if pressed, would offer M itself as the reason for their belief that they are in M (as well as for avowing M).

6. Some Concluding Remarks One of our central aims in this chapter has been to show that EDSK is better placed to avoid a serious challenge that faces EDP (presented in subsection 2.1). But one might now wonder whether our discussion of EDSK points the way to a version of EDP that can also avoid the challenge. Unlike in the case of basic self-knowledge, in the case of perceptual knowledge, adopting a non-perceptual model is obviously not an option. Still, the proponent of EDP may not be left without response. By way of conclusion, we offer some reflections on behalf of this proponent, leaving further discussion for future work. First, when posing the problem for EDP (in subsection 2.1), we noted that for EDP to accept the Perceptual Warrant Principle (PWP) (according to which the warrant for perceptual belief is constituted by perceptual experience) likely involves commitment to MDP. We say that it likely involves MDP, rather than requiring such commitment, because there is another theoretical option open to epistemological disjunctivism. One can accept both EDP and PWP without accepting that perceptual experience in the good case is radically different from perceptual experience in the bad case. One need only accept, instead, that while perceptual experience is of the same fundamental metaphysical type in the good and the bad case, there are nevertheless significant differences between the experiences (this is what Byrne and Logue call the Moderate View about perceptual experience—note that this view is not itself a version of MDP, but is instead a competitor; 2009: x–xi). These differences, one might argue, could then underwrite the difference in warrant between belief in good and bad cases that epistemological disjunctivism claims. We are not familiar with anyone who explicitly endorses the Moderate View of perceptual experience as a supplement to epistemological disjunctivism, but this combination of views may be an interesting way forward for epistemological disjunctivists concerned to avoid the criticism presented in section 2.1.53 Second, recall from our discussion of neo-expressivism, that the dual immunity to error in the case of avowals provides a default entitlement for basic mental self-beliefs, where this default entitlement is shared between true and false basic self-beliefs. A true basic self-belief gains additional epistemic warrant because it is epistemically grounded in the mental state it is about. This feature of the view qualifies it as a form of epistemological disjunctivism, since, given this feature, a true self-belief would still have a different warrant from a corresponding false one.

Epistemological Disjunctivism

337

Perhaps a similar move is available to EDP. It may be that one is epistemically entitled to one’s perceptual beliefs because of the nature of one’s perceptual experience. Granting that one’s perceptual beliefs in the good and the bad case are epistemically based on the same type of perceptual experience, one is equally entitled to one’s belief in the good case as one is in the bad case. Nevertheless, the epistemological disjunctivist might add that a further source of warrant is available in the good case that is not available in the bad case; namely, the factive reason that one sees that such-and-such. However, while in the case of self-belief it is clear that the epistemic grounding for one’s believing comes from the very mental state that the belief concerns, the point that is at issue when it comes to assessing EDP is whether or not perceptual beliefs enjoy any kind of epistemic warrant other than what is provided by the nature of one’s perceptual experience—this cannot simply be assumed. But our observations at least point the way toward a weaker epistemological disjunctivism about perception, which—like the weaker disjunctivism about basic selfknowledge—accepts that beliefs in good and bad cases may share some kind of epistemic warrant.54

Notes 1. By ‘basic’ self-belief, we mean the ordinary, non-reflective beliefs one has about one’s own present mental states—paradigmatically, these are spontaneous beliefs one has when avowing (out loud or to oneself) e.g. “I’m so tired”, “I’m fed up with this”, “I’d like to leave now”, etc. Excluded are, for example, self-beliefs formed as a result of deliberate investigation or testimony, as when one forms a self-belief on the basis of discussion with a therapist or a friend’s input. 2. For simplicity, we will here be assuming a standard analysis of knowledge, according to which knowledge requires true belief that is warranted. For reasons abundantly discussed in the literature, knowledge may require something more. This should not matter for our concerns here. We set aside ‘knowledge first’ views such Williamson (2000). 3. In section 3.2 we articulate more thoroughly in what sense self-beliefs can count as warranted on these accounts, by elucidating a notion of epistemic grounding, distinct from epistemic justification. 4. Throughout, unless otherwise noted, when we refer to self-knowledge/beliefs we are concerned with basic knowledge/beliefs concerning one’s present states of mind. 5. Typical formulations of the bad case appeal to hallucinations or illusions, without distinguishing the cause of illusion, sometimes alluding to evil demons, or brain-in-a-vat scenarios, or else not describing in much detail what sort of hallucination or illusion is under consideration. Our construal of the bad case is designed to accommodate Burge’s reservation about not counting certain kinds of hallucination (such as hallucinations caused by drugs or brainticklings) as cases of perception (see Burge 2005: 42). 6. See e.g. Martin 2009, Snowdon 2009. Here is a representative statement of MDP: [T]he basic claim of disjunctivism can be put as follows: the experiences in the good case and the hallucinatory bad cases share no mental

338

Dorit Bar-On and Drew Johnson core, that is, there is no (experiential) mental kind that characterizes both cases. (Byrne and Logue 2009: ix)

7. As Burge (2005: 42) notes, it may be mistaken for the metaphysical disjunctivist to regard all cases of illusion and hallucination as genuine perception. Their focus should be on cases that are genuinely perceptual (see note 5). But when it comes to epistemological disjunctivism, what matters is that experiences in good and bad cases are subjectively indistinguishable, whether or not those experiences are genuine perceptions. 8. Ram Neta has objected that, whether the Proximality Principle makes trouble for MDP depends on how the notion of proximality is understood. One would need to give sense to “proximal” that renders the Principle both plausible and inconsistent with MDP—something Burge does not seem to provide, and which we are not able to provide here, either. 9. Thus, proponents of the view would have to show that perceptual psychology is somehow not relevant to the evaluation of MDP. Perhaps they could argue, for instance, that perceptual psychology concerns states attributable only to perceptual subsystems, whereas MDP concerns perceptual states attributable to (whole) perceivers (see e.g. McDowell 2010, 2013). Burge disagrees. The theory of vision he describes “attributes states that are recognizably perceptual and recognizably states of individuals, not merely of subsystems” (2005: 22). 10. Pritchard concurs: “It is reasonably clear that epistemological disjunctivism does not in itself entail metaphysical disjunctivism. For that the rational standing available to the agent in the normal veridical perceptual experiences and corresponding (introspectively indistinguishable) cases of illusion and hallucination are radically different does not in itself entail that there is no common metaphysical essences to the experiences of the agent in both cases” (2012: 24). 11. One might think that the difference in the warrant S has for her perceptual belief in the good and bad cases is explained in terms of the reasons available to S in those cases, where what reason S has in turn depends on the relation between S’s mental states and the world. So, in the good case, S’s reason is a good one because S’s mental states (including her perceptual experience) are related to the world in the right way, whereas in the bad case, S’s mental states (which are of the same type as in the good case) are not appropriately related to the world. (Thanks to Matthew McGrath for suggesting this possible response.) A worry for this response is that it seems to present an externalist picture of warrant (as the nature of the relation of one’s experience to the world is apparently not available for reflective access), whereas EDP (at least of Pritchard’s variety) is an internalist view. How could the internalistically available reasons one has for belief vary between good and bad cases? 12. One might reject the idea that our perceptual experiences exhaust the internalist warrant for perceptual belief; perhaps there is something else, in addition to one’s perceptual experience, that can play a warranting role for perceptual belief—for instance, the relation between that experience and states of affairs in the world. But (as mentioned in note 9) it is unclear to us how such an additional source of warrant could be something both reflectively accessible to an agent and something that can vary between good and bad cases. 13. For an alternative argument that Pritchard’s EDP is committed to MDP, see Cunningham (2016). 14. But see section 6, where we revisit this issue. 15. See Shoemaker (1986), Fricker (1998), Macdonald (2007).

Epistemological Disjunctivism

339

16. In a series of influential articles, Shoemaker (1994) offered a detailed critique of various versions of the perceptual model of basic self-knowledge, which we will not rehearse here. 17. But see Byrne (2005, 2018), Cassam (2014), Carruthers (2011) for recent defenses. 18. A non-disjunctivist view that would be fashioned after Burge’s view of perceptual entitlement (see note 22) would not be available to them. 19. The analog of this in the perceptual case—viz. “well, p!”—would seem tantamount to saying “I just know!”. 20. Pritchard considers two ways of filling out the reflective access condition (2012: 36). First, there is accessibilism: “S’s internalist epistemic support for believing that ɸ is constituted solely by facts S can know by reflection alone.” The second version is mentalism: “S’s internalist epistemic support for believing that ɸ is constituted solely by S’s mental states.” We shall restrict our understanding of reflective access to accessibilism, though it is worth considering what a disjunctivist account of self-knowledge might look like with a mentalist understanding of the internalist reflective access condition. If, as we go on to consider, self-beliefs about one’s mental states are epistemically warranted by those mental states themselves, then a mentalist understanding of reflective access is clearly, and perhaps trivially, satisfied. Pritchard also notes, interestingly, that if EDP accepts mentalism, then it may also be committed to MDP (2012: 36–37). Given our earlier argument, this is reason then for EDP not to accept the mentalist version of the reflective access condition. 21. As a starting point, we can consider a weak form of epistemic internalism, which Pappas (2014) calls ‘Weak AJI’: “One has a justified belief that p only if one can become aware by reflection of some essential justifier one then has for p.” An essential justifier, here, is a justification for one’s belief without which the belief would fail to be justified. Note, moreover, that this form of internalism only requires that one become aware of an essential justifier; one need not recognize this justifier as an essential justifier. 22. Pritchard (2012), advertises his disjunctivism as combining certain features of traditional internalism and traditional externalism. Neta and Pritchard (2007) also argue that McDowell’s position combines internalist and externalist elements. The internalist conception of warrant that EDP employs is also recognized in Madison’s (2010) overview of epistemic internalism. 23. Zimmerman also identifies and rejects a similar assumption in addressing a purported puzzle about self-knowledge: “Our first-order beliefs themselves provide grounds for our second-order introspective beliefs. One’s reason for thinking that one believes that p is the very fact that one believes that p. Why haven’t philosophers embraced this simple answer? Many epistemologists have assumed that we can justify our beliefs only through observation or inference” (2006: 338, emphasis added). 24. See Dretske (2000), Burge (2003), Wright (2004), Altschul (2011). 25. This way of understanding justification is meant to align with a Cartesian, rationalist view of justification. 26. There is a similar use of the notion of ‘grounding’ in contemporary discussions of theories of truth, as well as metaphysics. 27. See Peacocke (1996), Zimmerman (2006), Coliva (2012). For critical discussion, see Bar-On (2004b: 388ff., 2009, 2018). 28. For a more functionalist version, see Zimmerman (2006: 343f), which expounds Shoemaker’s view. 29. See Zimmerman (2006: 343ff). Zimmerman frames his discussion primarily in terms of second-order beliefs about first-order beliefs. But constitutivists

340

30. 31. 32.

33. 34. 35. 36. 37.

38.

39. 40. 41.

42.

43.

Dorit Bar-On and Drew Johnson (including Zimmerman and Boyle) often extend the constitutivist claim to other first-order mental states. See e.g. Zimmerman (2006: 342), Boyle (2010). For a critical discussion of constitutivism, see Bar-On (2004b, 2009, 2018). (See also Bar-On, “No ‘How’ Privileged Self-Knowledge”, in progress.) Bar-On (2004b: esp. chs. 1 and 9) argues that this special epistemic status (aka ‘first-person authority’) constitutes a more salient and stable component of our commonsense view than the self-intimation that constitutivism tries to capture and that it provides a more feasible explanandum for a philosophical theory of basic self-knowledge. For Bar-On’s rejection of the perceptual view, and of the Epistemic Approach in general, see her (2004b: chs. 3 and 4, 2009), Bar-On and Long (2001). See Bar-On (2004b), esp. ch. 4 and passim. For fuller discussion, see Bar-On (2004b, 2009, 2012). See later in this subsection, where we distinguish IEM1 from IEM2. For discussion of the phenomenon of immunity to error through misidentification, see Wittgenstein (1958: 66–67), Shoemaker (1968), Evans (1982: esp. ch. 7, sec. 2), Wright (1998: 18–20). For a very useful analysis, see Pryor (1999). There is a “thin” sense in which I do identify myself as the subject of the ascription, for I do manage to refer to myself, and referring is picking out, and picking out is a form of identifying. It may be useful to distinguish between the referential notion of identifying (a semantic notion) and the recognitional notion of identifying (which is an epistemic notion). Compare Evans (1982: 218). For a related discussion of the differences between semantic and epistemic aspects of perception, see Burge (2005: 6–9). Evans (1982: 220, 222). The phrase “intentional content” is here used to cover both intentional object (e.g. “I’m afraid of the dog”) and propositional content (e.g. “I’m hoping that it won’t rain today”). Bar-On first introduces the notion of ascriptive immunity to error to characterize avowals’ distinctive security in 2000 and develops it in 2004b, 2004a, 2009, 2012. In his 2012, Wright offers a discussion of immunity to error that has several notable points of contact with Bar-On’s, though he does not make the connection—crucial to Bar-On’s neo-expressivism—between avowals’ distinctive immunity to error and their expressive character. Assuming that avowals are taken to be made on no epistemic basis, Bar-On takes it to be necessary to supplement the negative characterization in terms of immunity to error with a positive epistemological account of what qualifies them as knowledge. In the case of mental attributions to others, I do typically have such independent reasons. And mental self-attributions can also take a more evidential character—when one needs to rely on e.g. testimony, observation, or interpretation in order to determine, e.g., whether one is scared of something or what one is scared of. But this is not the case when one is simply avowing being in a mental state. For discussion of the notion of brute error and its connection to immunity to error, see Bar-On (2004b: 9f., 183, 200f., 332f). Bar-On’s neo-expressivist view (unlike standard expressivist views) makes room for the possibility of false avowals. However, on her view, false avowals constitute species of expressive failures; they are not examples of brute errors. Expressive failures, Bar-On argues, are to be understood in psychological, rather than epistemological terms. A false avowal, thus understood, involves a self-attribution that is wrong (because false) but not because the avower has gone wrong, in

Epistemological Disjunctivism

44. 45.

46.

47. 48. 49. 50.

51. 52. 53.

341

moving from an internal ‘judgment of appearance’ to the relevant self-belief (See Bar-On 2004b: ch. 8, where an explanation of e.g. self-deception and wishful thinking is provided along these lines). Doyle (this volume) argues that accommodating the possibility of false basic self-beliefs (at least ones that concern one’s first-order beliefs) requires allowing that there are ‘ringers’ for the states the beliefs are about—viz., states of the individual that are “subjectively indistinguishable between good and bad cases” and that “mislead us when we err about our own minds” (p. 362). However, this ignores the earlier-mentioned option of explaining false self-beliefs on the model of expressive failures. That model rejects the idea that when a subject errs about her own mind, this must be because she has been misled, or fooled by a ‘ringer’ (for a directly relevant discussion of the status of internal ‘impostors’, see Bar-On 2004a). Doyle argues that appeal to an epistemically relevant element that is subjectively indistinguishable between the good and the bad case (what he refers to as a ‘conscious judgment’, an ‘inner assertion’) is necessary in order to explain “how the possession of self-knowledge is intelligible from the subject’s point of view” (p. 348). We think that the dual immunity to error discussed here should help in providing such an explanation, insofar as this immunity is something to which self-believers have reflective access. A fuller explanation will have to await another occasion. For discussion of the possibility of false avowals, see Bar-On (2004b: esp. 320–335 and 394ff). On occasion, one might ‘second-guess’ an earlier avowal—“I thought I was annoyed at you, but really I was frustrated by this puzzle.” Note, however, that when this occurs, one is (on the occasion of second-guessing) not expressing frustration in the avowing mode, but offering a self-report from a third-personal stance. The later self-attribution does not represent a basic self-belief; and its occurrence does not reveal the earlier avowal to involve a brute error. In the case of proprioceptive reports, the source of their immunity to error through misidentification has to do with our possessing special mechanisms for obtaining information concerning our own bodies. See Evans (1982: ch. 7). For a more extended presentation of the ideas summarized here, see Bar-On (2011) (see also 2009, 2012). Full discussion appears in Bar-On (2004b: Chs. 6–8). For discussion, see Bar-On (2004b: 381ff). See Bar-On (2004b: 374ff). for discussion of entitlement by default. To say that one is entitled by default is, of course, not to say that one’s selfbelief is true by default. Truth, on the present proposal, is simply a matter of whether one is in M or not, which may in general be independent of whether one thinks (or judges, or believes) that one is in M. One may also worry, in addition, that entitlement by default is not sufficient to constitute a positive epistemic good (see subsection 3.2). Bar-On (2004b: 388ff). spells out a view along these lines and offers some neo-expressivist motivations for it. It should be noted that, on McDowell’s brand of disjunctivism, it seems that in the case of veridical perceptual belief, the perceptual state itself contains the fact that the belief concerns (see e.g. 2009). On this view, there is no ‘epistemic distance’ one needs to traverse for one’s (true) perceptual belief to amount to knowledge. And this parallels the neo-expressivist view proposed here. (Thanks to Ram Neta for pointing out this parallel.) However, it is precisely this feature of the McDowellian view that makes it controversial as an account of perceptual knowledge.

342

Dorit Bar-On and Drew Johnson

54. We thank Casey Doyle, Joe Milburn, Matt McGrath, and Ram Neta for helpful comments on earlier drafts of this chapter. We thank the Language and Mind group at Arché for questions and discussion on a presentation of this chapter.

References Altschul, Jon, 2011, Epistemic entitlement, The Internet Encyclopedia of Philosophy. www.iep.utm.edu/ep-en/ Bar-On, Dorit, 2000, Speaking my mind, Philosophical Topics 28(2): 1–34. doi:10. 5840/philtopics200028216 ———, 2004a, Externalism and self-knowledge: Content, use, and expression, Nous 38(3): 430–455. doi:10.1111/j.0029-4624.2004.00477.x ———, 2004b, Speaking My Mind: Expression and Self-Knowledge. Oxford: Oxford University Press. ———, 2009, First-person authority: Dualism, constitutivism, and neo-expressivism, Erkenntnis 71(1): 53–71. doi:10.1007/s10670-009-9173-y ———, 2011, Externalism and skepticism: Recognition, expression, and selfknowledge, in Annalisa Coliva, ed. Self-Knowledge and the Self. Oxford: Oxford University Press. ———, 2012, Expression, truth, and reality: Some variations on themes from Wright, in Annalisa Coliva, ed. Mind, Meaning, and Knowledge: Themes From the Philosophy of Crispin Wright. Oxford: Oxford University Press, pp. 162–192. ———, 2018, Minding the gap: in defense of Mind-mind continuity, in Kevin M. Cahill and Thomas Raleigh, eds. Wittgenstein and Naturalism. New York: Taylor and Frances, pp. 177–203. Bar-On, Dorit and Long, Douglas, 2001, Avowals and first-person privilege, Philosophy and Phenomenological Research 52(2): 311–335. doi:10.1111/j.19331592.2001.tb00058.x Boyle, Matthew, 2010, Bar-On on self-knowledge and expression, Acta Analytica 25(1): 9–20. doi:10.1007/s12136-009-0075-z ———, 2011, Tranparent self-knowledge, Proceedings of the Aristotelian Society (suppl. vol.) lxxxv: 223–241. doi:10.1111/j.1467-8349.2011.00204.x Burge, Tyler, 1993, Content preservation, The Philosophical Review 102(4): 457– 488. doi:10.2307/1523046 ———, 1996, Our entitlement to self-knowledge, Proceedings of the Aristotelian Society 96(1): 91–116. doi:10.1093/aristotelian/96.1.117 ———, 2003, Perceptual entitlement, Philosophy and Phenomenological Research 67(3): 503–548. doi:10.1111/j.1933-1592.2003.tb00307.x ———, 2005, Disjunctivism and perceptual psychology, Philosophical Topics 33(1): 1–78. doi:10.5840/philtopics20053311 ———, 2011, Disjunctivism again, Philosophical Explorations 14(1): 43–80. doi: 10.1080/13869795.2011.544400 ———, 2013, Epistemic warrant, in his Cognition Through Understanding: Philosophical Essays (vol. 3). Oxford: Oxford University Press, pp. 521–533. Byrne, Alex, 2005, Introspection, Philosophical Topics 33(1): 79–104. ———, 2018, Transparency and Self-Knowledge. Oxford: Oxford University Press.

Epistemological Disjunctivism

343

Byrne, Alex and Logue, Heather, 2009, Either/or, in Adrian Haddock and Fiona Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 57–94. ———, 2009, Introduction, in Alex Byrne and Heather Logue, eds. Disjunctivism: Contemporary Readings. Cambridge, MA: MIT Press, pp. vii–xxix. Carruthers, Peter, 2011, The Opacity of Mind: An Integrative Theory of SelfKnowledge. Oxford: Oxford University Press. Cassam, Quassim, 2014, Self-Knowledge for Humans. Oxford: Oxford University Press. Coliva, Annalisa, 2012, One variety of self-knowledge: Constitutivism as constructivism, in Annalisa Coliva, ed. The Self and Self-Knowledge. Oxford: Oxford University Press, pp. 212–242. Cunningham, Joseph, 2016, Reflective epistemological disjunctivism, Episteme 13(1): 111–132. doi:10.1017/epi.2015.62 Dretske, Fred, 2000, Entitlement: Epistemic rights without epistemic duties? Philosophy and Phenomenological Research 60(3): 591–606. doi:10.2307/2653817 Evans, Gareth, 1982, The Varieties of Reference. Ed. John McDowell. Oxford: Oxford University Press. Fricker, Elizabeth, 1998, Self-knowledge: Special access versus artefact of grammar— a dichotomy rejected, in Crispin Wright, Barry Smith and Cynthia Macdonald, eds. Knowing Our Own Minds. Oxford: Oxford University Press. Macdonald, Cynthia, 2007, Introspection and authoritative self-knowledge, Erkenntnis 67(2): 355–372. doi:10.1007/s10670-007-9072-z Madison, Brent., 2010, Epistemic internalism, Philosophy Compass 5(10): 840– 853. doi:10.1111/j.1747-9991.2010.00333.x Martin, Michael., 2009, The reality of appearances, in Alex Byrne and Heather Logue, eds. Disjunctivism: Contemporary Readings. Cambridge, MA: MIT Press, pp. 91–116. McDowell, John, 2009, Selections from Criteria, defeasibility, and knowledge, in Alex Byrne and Heather Logue, eds. Disjunctivism: Contemporary Readings. Cambridge, MA: MIT Press, pp. 75–90. ———, 2010, Tyler Burge on disjunctivism, Philosophical Explorations 13(3): 243–255. doi:10.1080/13869795.2010.501905 ———, 2013, Tyler Burge on disjunctivism (II), Philosophical Explorations 16(3): 259–279. doi:10.1080/13869795.2013.808693 Neta, Ram and Pritchard, Duncan, 2007, McDowell and the new evil genius, Philosophy and Phenomenological Research 74(2): 381–396. doi:10.1111/j.19331592.2007.00023.x Pappas, George, 2014, Internalist vs. externalist conceptions of epistemic justification, in Edward N. Zalta, ed. The Stanford Encyclopedia of Philosophy (Fall 2017 ed.). https://plato.stanford.edu/archives/fall2017/entries/justep-intext/ Peacocke, Christopher, 1996, Our entitlement to self-knowledge: Entitlement, self-knowledge, and conceptual redeployment, Proceedings of the Aristotelian Society 96(1): 117–158. doi:10.1093/aristotelian/96.1.117 ———, 1999, Being Known. Oxford: Oxford University Press. Pritchard, Duncan, 2008, McDowellian neo-Mooreianism, in Adrian Haddock and Fiona Macpherson, eds. Disjunctivism: Perception, Action, and Knowledge. Oxford: Oxford University Press, pp. 283–310. ———, 2012, Epistemological Disjunctivism. Oxford: Oxford University Press.

344

Dorit Bar-On and Drew Johnson

Pryor, James, 1999, Immunity to error through misidentification, Philosophical Topics 26(1–2): 271–304. Shoemaker, Sydney, 1968, Self-reference and self-awareness, Journal of Philosophy 65(October): 555–567. doi:10.2307/2024121 ———, 1986, Introspection and the self, Midwest Studies in Philosophy 10(1): 101–120. doi:10.1111/j.1475-4975.1987.tb00536.x ———, 1994, Self-knowledge and ‘inner sense’, Philosophy and Phenomenological Research 54: 249–314. Snowdon, Paul, 2009, The objects of perceptual experience, in Alex Byrne and Heather Logue, eds. Disjunctivism: Contemporary Readings. Cambridge, MA: MIT Press, pp. 49–74. Williamson, Timothy, 2000, Knowledge and Its Limits. Oxford: Oxford University Press. Wittgenstein, Ludwig, 1958, The Blue and Brown Books. New York: Harper & Row. Wright, Crispin, 1998, Self-knowledge: The Wittgensteinian legacy, Royal Institute of Philosophy Supplement 43: 101–122. doi:10.1017/S135824610000432X ———, 2004, Warrant for nothing (and foundations for free)? Aristotelian Society Supplementary Volume 78(1): 167–212. doi:10.1111/j.0309-7013.2004. 00121.x ———, 2012, Reflections on Francois Recanati’s ‘immunity to error through misidentification: What it is and where is comes from’, in Simon Prosser and Francois Recanati, eds. Immunity to Error Through Misidentification: New Essays. Cambridge: Cambridge University Press, pp. 247–280. Zimmerman, Aaron, 2006, Basic self-knowledge: Answering Peacocke’s criticisms of constitutivism, Philosophical Studies 128: 337–379. doi:10.1007/ s11098-004-7797-y

17 Ringers for Belief Casey Doyle1

1. Introduction Epistemological Disjunctivism in the epistemology of perception holds that a subject can know that p on the basis of a reason, seeing that p, which is both reflectively accessible and factive, indefeasible, or truthguaranteeing, as it is variously put. That way of stating the view leaves out the disjunction, though. It is between good cases, cases of seeing that p, and bad cases, such as hallucination, where it merely seems to one that one sees that p.2 Against those who accept a “highest common factor” conception of justification for perceptual beliefs, the Disjunctivist insists that the good and bad cases differ in the sort of justification they provide.3 The good case provides better justification than the bad case, specifically truth-guaranteeing justification.4 In John McDowell’s hands, Disjunctivism is not restricted to perceptual knowledge; he has defended versions of Disjunctivism for knowledge of other minds (1998a), testimony (1998b), and knowledge generally (1998c). This paper considers a Disjunctivist treatment of self-knowledge of belief, which I will call Disjunctivism about Doxastic Self-Knowledge (DDSK). Specifically, the chapter examines whether such a Disjunctivism faces a well-known objection: that one cannot possess a factive, reflectively accessible reason since the subject cannot distinguish, by reflection alone, whether she is in a good or bad case. Some recent defenders of DDSK have insisted that it doesn’t face this worry. But I don’t things are so straightforward. Like other contributions to this volume, this chapter aims to consider a familiar Disjunctivist view in a domain other than perceptual knowledge. And it aims to determine whether the considerations that have been brought to bear against Disjunctivism apply in this other domain. But its aims are also broader than this. I will argue that whether a familiar objection to Epistemological Disjunctivism applies to a Disjunctivist treatment of doxastic self-knowledge depends on how we conceive of conscious judgment. My suggestion at the end will be that the objection can be avoided if we opt for the conception of judgment that has been adopted by some Disjunctivists about perceptual knowledge, but that this

346

Casey Doyle

conception faces serious difficulties. This is to say that an additional goal of the chapter is to show how reflection on self-knowledge can bring out something about at least some versions of Epistemological Disjunctivism about perception.5 The next section sets out the objection and Disjunctivism about doxastic self-knowledge. Section 3 looks at arguments purporting to show that this Disjunctivism doesn’t face the objection. Section 4 articulates a constraint on an adequate theory of self-knowledge: it must explain how the possession of self-knowledge is intelligible from the subject’s point of view. Section 5 proposes that self-knowledge is intelligible in light of the experience of conscious judgment. It also argues that, on perhaps the standard conception of conscious judgment, what I call the “inner assertion model”, this gives rise to a version of the objection. I conclude by considering an alternative conception of judgment defended by some Epistemological Disjunctivists and suggest that it might fare better. That conception of judgment, and the package of views about self-consciousness and knowledge of which it is a part, are quite radical. The tentative suggestion at the end, then, is that, if our account of selfknowledge is to avoid the familiar objection, we might need to accept the more radical view.

2. Disjunctivism, Indistinguishability, Self-Knowledge The Disjunctivist position about perceptual knowledge can seem puzzling. Suppose that what is reflectively accessible to one at a time is what one is in a position to know at that time (Gibbons 2006). The Disjunctivist holds that, when you see that p, when all goes well, you are in a position to know both that p and that you see that p. One worry is that, given that good and bad cases are subjectively indistinguishable, it is hard to see how one is in a position to have that self-knowledge when one is in the good case. If the appearances cannot rule out that one is in a bad case, then being in the good case is not sufficient to put one in a position to know that one is. Now John McDowell has long insisted that a false conception of self-consciousness is at the root of this familiar source of resistance to Disjunctivism (1998a, 1998d, 2013, this volume). And, of course, others have offered responses.6 For now, it will suffice to note that the “objection from subjective indistinguishability”, as we can call it, arises because the knowledge in question, perceptual knowledge, is based on an appearance. Therefore, we should expect the objection to arise in other domains where knowledge is based on an appearance, such as memory, another’s say-so, or their behaviour. It is for this reason that DDSK is thought to be better positioned than its cousins in other areas. Here is a statement of DDSK. DDSK: One can know that one believes that p because one believes that p itself. Second-order beliefs are based directly on their truthmakers.

Ringers for Belief

347

DDSK is a thesis about doxastic self-knowledge, but not yet a fullfledged theory since it doesn’t articulate the precise way in which secondorder beliefs are based on first-order ones. There are different options here. According to Constitutivists, part of what it is to believe that p, in a non-trivial range of cases, is to believe or know that you believe that p (Shoemaker 1996, 2012; Zimmerman 2006; Rödl 2007; Boyle 2011; Travis 2012; Marcus 2016). On this view it is wrong to suppose that one makes a transition in thought between the first and second order, or that there is a causal relation between them. Belief is itself a selfconscious condition. Others accept that first and second-order beliefs are, in Hume’s terms “distinct existences”. Why is the transition from one to the other licit? There are options here, as well. Some philosophers appeal to doxastic agency (Moran 2001), others to the requirements of rationality (Roessler 2013b), and Dorit Bar-On suggests that we are “entitled by default” to self-ascriptions given that they are grounded in their truthmakers, when all goes well (Bar-On 2004: 384).7 The Belief Account is a thesis accepted by a family of theories of self-knowledge. Our focus here will be on the thesis. The thesis should be contrasted with views on which self-knowledge is arrived at by inference from a premise about the world (Byrne 2005), by inference from a premise about one’s behaviour about one’s behaviour or sensory evidence (Cassam 2014), by inner sense (Armstrong 1968 [1993]), or by introspecting one’s conscious experiences (Pitt 2004). On DDSK, self-knowledge of belief is based on belief itself, conceived of as a nonconscious, standing attitude. As I will understand it, DDSK denies that self-knowledge is based on evidence or reasons at all. As we’ll see, DDSK holds that self-knowledge is baseless. As others have noted (Bar-On 2004; Zimmerman 2006), DDSK is Disjunctivist because, on it, one can know that one believes that p in virtue of possessing a ground that is both factive and reflectively accessible. Indeed, both conditions are trivially satisfied. According to DDSK, one can form a self-ascription noninferentially on the basis of the selfascription’s truthmaker. Since p entails p, one’s ground is factive. And since the method outlined by DDSK just is a method of reflectively accessing one’s beliefs, the ground is reflectively accessible. Furthermore, there is an asymmetry in the support one receives between good and bad cases. If you believe that you believe that p because you believe that p, then you are in possession of a truth-guaranteeing ground. But if you spontaneously but falsely believe that you believe that p, then, obviously, you are not (Bar-On and Johnson this volume). Now it is not entirely clear how to make sense of a bad case on this view, given that the first-person method is simply forming a self-ascription on its truthmaker. What could be an instance of that process, or an exercise of that capacity, gone wrong? Yet we are surely fallible about our beliefs. And we can get things wrong in circumstances where we take ourselves to have

348

Casey Doyle

first-person knowledge, that is, where we take ourselves to know what we believe in the way set out by a theory of self-knowledge like DDSK.8 We can be wrong about what we believe when we occupy the first-person perspective on our mental states. I will return to this issue later. Despite being Disjunctivist in the sense outlined, DDSK differs from other versions of Epistemological Disjunctivism because the form of knowledge it specifies isn’t possessed on the basis of an appearance. As Bar-On puts it: When I avow being in M, my epistemic position is not one of somehow moving from a judgment that someone, who appears to me to be me, is in some mental state, which appears to me to be M, to the self-ascription that I am in M. I have no reason for thinking that someone is in M, or that I am in some mental state—other than whatever reason I have for thinking, simply, that I am in M. (2004: 393) The objection we are considering alleges that when one’s ground for believing that p is an appearance that p, then one cannot rule out, by reflection alone, that one is not fooled by the appearance. But on DDSK, one’s ground is not an appearance: it is simply the fact that one believes that p, the self-ascription’s truthmaker. So there is no question of being fooled by the ground one possesses. That it simply falls out of DDSK that it doesn’t face this objection doesn’t clearly speak in its favor, since it might be thought that what looks like a benefit is actually a cost. That’s because it might be thought that the view is psychologically unrealistic in failing to make sense of our fallibility in knowing our beliefs. You might think that self-knowledge is like other forms of knowledge in that one’s subjective condition could lead one into error or leave one in doubt about whether one is really in a position to know what one believes. However, defenders of DDSK have insisted that this is not so. Here is how Charles Travis puts it: [H]allucinating the wind blowing your hair is, while improbable, an unproblematic idea as such. Whereas it is difficult to make sense of the idea of hallucinating believing, for example, that Chez Fred has changed its menu. Someone might be presented, in a way he is presented to himself alone, with a ringer for the wind blowing his hair. It is at best difficult to imagine one being presented with a ringer for his believing that Chez Fred has changed the menu—especially in some way he is presented to himself alone. (Travis 2012: 393) Now Travis has principled reasons for thinking there isn’t a state subjectively indistinguishable from believing that p. I will return to these

Ringers for Belief

349

reasons. In this passage, though, the thought goes that we can’t imagine a case where there is a “ringer” for believing that p, something about one’s condition in virtue of which it would seem to you, in the first-personal way, that you believe that p, though you don’t. Aaron Zimmerman makes a similar point and explains why we shouldn’t expect there to be ringers for belief. His argument goes like this. In the good case one believes that p, in the bad case one doesn’t. There can’t be a ringer for the good case because believing that p and not believing that p are subjectively distinguishable: when you don’t believe that p you can tell that you don’t. That’s because believing that p constitutively involves a disposition to consciously judge that p, whereas not believing that p doesn’t. [A] subject with [a conscious] belief must be so disposed that were she to consider the matter she would occurrently and consciously judge that p, then any alteration in a subject’s set of [conscious] beliefs will bring with it changes in the kinds of experience to which that subject is prone. There is therefore no reason why a constitutivist should allow that the absence of [a conscious] belief might be subjectively indistinguishable from its presence. (Ibid.)9 Believing that p and not believing that p seem different from the subject’s point of view, because in one case, but not the other, the subject is disposed to consciously judge that p.10

3. Evaluating the Arguments We have seen two reasons for thinking that there aren’t ringers for belief. The first is that there can only be ringers when one’s belief is based on an appearance, which it isn’t in self-knowledge. The second is that, from the subject’s point of view, cases of believing that p and not believing that p are subjectively distinguishable. They are distinguishable on the grounds that one involves a disposition to judge and the other does not. You can tell whether you believe that p or not based on the experience of conscious judgment. Notice that to say this much is not to say that one possesses self-knowledge on the basis of introspecting the experience of conscious judgment. All versions of DDSK agree that one knows what one believes simply on the basis of the belief itself. The point is rather that any purported ringer for belief would actually be a case of belief. Is this right? Are there really no ringers for belief? I’m not so sure. There are many attitudes that are not belief, but which involve many of the same dispositions as belief. Imagining is an obvious example. More controversial ones are delusion and self-deception. Tamar Gendler (2007) has argued that in cases of self-deception, a subject engages in imaginative pretense that p, while believing that not-p. Here is a stock example

350

Casey Doyle

(Funkhouser 2005). Carol has an overwhelming amount of evidence that her husband is having an affair with her friend. Other friends report seeing his car at the friend’s house at night. Much of Carol’s behaviour suggests that she is aware of this; for example, at night, when her husband is supposed to be at work, she avoids driving by the friend’s house, even when it requires her to go out of her way. Yet, if asked, Carol assures people that her husband is faithful and she thinks this to herself as well, whenever the issue arises. On Gendler’s proposal Carol is simply pretending that her husband is faithful. Pretense differs from belief in that it isn’t responsive to evidence and doesn’t have the same behavioural manifestations (Velleman 2000). (For example, in high-stakes cases one will drop the pretense but act on one’s belief.) Despite these differences, pretense disposes one to consciously entertain the content one is imagining. Gendler’s view of self-deception is controversial, of course. But the more general point—that in cases of imagining or pretense, a subject is disposed to entertain the content imagined—is highly plausible. It is also plausible in other cases, such as monothematic delusion. Some philosophers think that it is wrong to conceive of states like the Capgras delusion as belief (Currie and Jureidini 2001). But whatever state it is, these subjects are disposed to entertain the content of their state in consciousness and to assert it. This strongly suggests that, contra Zimmerman, there are states other than believing that p that dispose a subject to consciously judge that p.  Delusion, imagination, pretense, and perhaps self-deception are all plausible, if not wholly uncontroversial cases. If you are fooling yourself, as we say, about whether a spouse is faithful or whether you are well prepared for a job interview, then you are disposed to affirm such things in conscious thought, even though you don’t really believe them. But affirming the content in consciousness is subjectively indistinguishable from consciously judging so, when that judgment is the manifestation of one’s belief. Further, it is plausible to suppose that in just these cases one will be led astray about what one believes. Delusional subjects take themselves to believe their delusions. Likewise with cases of self-deception. Carol is likely not only to imagine that her husband is faithful, in spite of the evidence to the contrary, but to self-ascribe the belief that he is.11 It may after all not be too difficult to imagine ringers for belief, contra Travis, precisely because other cognitive attitudes dispose one to make conscious judgments, contra Zimmerman. And of course, this is what we should expect, given that we are in need of some explanation of how we can be led astray about what we believe. One seeming way to resist this conclusion would be by holding that judgment entails belief. If judgment entails belief, then judgments can’t be ringers for belief. But this response won’t do. Either judgment entails belief or it doesn’t. If it doesn’t, then there are cases of conscious judgment without belief, and these cases, whatever underlies them, would,

Ringers for Belief

351

from the subject’s point of view, count as ringers. If judgment does entail belief, then we can rephrase the point as follows: there are episodes in consciousness—telling oneself, fooling oneself, imagining—that are subjectively indistinguishable from judging. These episodes are ringers for belief. In such cases, one’s thoughts are presented to one in a way they are only presented to one, as beliefs, and yet, one does not believe what one takes oneself to believe. It seems, then, that we should reject Zimmerman and Travis’s arguments. It might be objected that whether there are ringers for belief depends on how we conceive of the phenomenology of conscious judgment. Is there a proprietary phenomenology for each attitude type? Is there something it is like to judge that p that is different from what it is like to imagine that p or to simply entertain the proposition that p? Some philosophers think the answer is yes (Pitt 2004). In their view, each attitude type has a distinctive phenomenology. This topic deserves more attention than I can pay it here. The following will have to do. For present purposes I assume a particular picture of conscious judgment, what I will call the “inner assertion” model of judgment (cf. Peacocke 1998; Cassam 2010). On this model a conscious judgment is an event in phenomenal consciousness, the sort of thing that makes a difference to what it is like for the subject. Conscious judgment is an act, manifest in conscious experience, of affirming a proposition. It is “inner assertion” because it is an analogue of the public act of assertion. Judgment is a private act of affirming a proposition for oneself. For our purposes it won’t matter whether judgment, so conceived, is an action or deed of the agent’s (Peacocke 1998). And it won’t matter how precisely we think of the phenomenology of conscious judgment, for example, whether it is manifest in auditory or cognitive phenomenology.12 Now I am not sure how to establish a negative phenomenological claim, and I am not going to run through examples insisting that I don’t find a phenomenology of attitude type when I reflect on them. Instead, here are two principled reasons for doubting that there is such a phenomenology. The first comes from taking the analogy with public assertion seriously. It is plausible to suppose that, qua inner assertion, the phenomenology of judgment is more or less like the phenomenology of outer speech.13 It represents a particular content to the subject in a particular medium, in this case some form of imagery. Asserting that p sounds the same when it is a lie and when it expresses the speaker’s belief. Likewise, one might think that asserting to oneself that p will “sound” the same whether this expresses one’s belief or is a bad case, like delusion or self-deception. Indeed, it will “sound” the same when it is simply a different kind of act, such as imagining an interlocutor. The second consideration is this. I argued that there are cases where a subject entertains a proposition in thought, but where this is the manifestation of pretense, delusion, or selfdeception instead of belief. And in just these cases, the subject is likely

352

Casey Doyle

to be in error about what she believes. This strongly suggests that the conscious thought she entertains is subjectively indistinguishable from a judgment that expresses her belief. If we want a plausible explanation of first-person error, we should grant that there are ringers for belief. But perhaps this is beside the point. Bar-On concedes the possibility we are considering, but argues that it doesn’t give rise to a version of the objection from subjective indistinguishability. As she puts it: “We may remain neutral on the psychological question whether, when I avow, there is some psychological state I am in that does remain invariant regardless of whether I am in the avowed state or not” (Bar-On 2004: 383 fn. 40, emphasis in original). On DDSK one’s ground for self-ascription is one’s belief. The objection from subjective indistinguishability gets going when the ground is subjectively indistinguishable from a corresponding bad case. But the ringers we have considered are not like that. At the level of grounds, then, there is no cause for concern. Indeed, it might be thought that ours is a happy development for the Disjunctivist, since, as we saw, she needs to explain what happens in bad cases. We now have an explanation: in a bad case one is misled by one’s conscious judgment (or an episode in consciousness indistinguishable from judgment) into thinking that one believes that p, when one doesn’t. But in the good case, one’s ground is simply the fact that one holds the belief. There isn’t a ringer for this. This is right, as a claim about the consequences of DDSK, and further, unlike Travis and Zimmerman’s claim, it is backed by something psychologically realistic. Stuart Hampshire observed that when someone makes a claim to self-knowledge, for example, by saying “I believe that p”, it is typically inappropriate to ask her how she knows it (Hampshire 1979). While we ask people to defend and justify their beliefs, we don’t ask for evidence or reasons for a subject’s self-ascriptions. Now Hampshire implausibly concluded from this observation that there isn’t any explanation at all for one’s self-knowledge. We needn’t follow him there (after all, DDSK is such an explanation).14 Still, the most natural conclusion to draw from Hampshire’s observation is that we don’t ask for reasons or evidence because self-ascriptions aren’t based on evidence, at least of the sort that would be apt in conversation. In this spirit, philosophers have claimed that first-person knowledge is baseless, groundless, or silent (McDowell 1998d; Wright 1998; O’Shaughnessy 2000). As Johannes Roessler puts it, “the first-person perspective affords no insight into how we know what we believe” (2013b: 1). The Disjunctivist’s claim that self-ascriptions of belief aren’t based on appearances sits well with our practices of avowing mental states. I have argued that there is something subjectively indistinguishable between good and bad cases, specifically the judgment that p (or an episode in consciousness indistinguishable from judging that p). But whether that matters—that is, whether it gives rise to a version of the objection

Ringers for Belief

353

from subjective indistinguishability—depends on the role of conscious judgment in the possession of self-knowledge. It might be thought that, since conscious experience is not the basis of self-knowledge of belief, and since it is inappropriate even to ask the avowing subject for her reasons, conscious judgment doesn’t play a significant role. But that is too quick. Even if we concede that the objection doesn’t arise at the level of grounds, it might arise elsewhere. In the next section I explain what role conscious judgment plays in the possession of self-knowledge and how this gives rise to a version of the objection.

4. The Intelligibility of Self-Knowledge DDSK denies that, when one knows what one believes in the first-personal way, one’s ground is a conscious judgment. But I think conscious judgment still plays an important role in the possession of self-knowledge, such that the fact that there are ringers for conscious judgment poses a problem for DDSK. That role is this: rendering the possession of selfknowledge intelligible to the subject. Given some of the peculiarities of self-knowledge (it is based on its truthmaker, it is groundless, it is not subject to the “how do you know?” question, etc.) it should not be surprising that we need to invoke less familiar ideas to make sense of it. The basic idea I have in mind is brought up by Johannes Roessler in a defense of Stuart Hampshire’s account of our knowledge of our future intentional actions (2013a). On Hampshire’s view, “your practical reasoning can make it intelligible how you are in a position to know what you will do” (Ibid.: 51). He writes: Even a non-philosopher who declares she will come to the meeting may find her possession of the knowledge expressed in her statement unmysterious in the light of the practical reasoning on which the statement is based. This should not be taken to mean that she has a good answer to the question “How do you know you’ll come to the meeting?” For that question is naturally heard as a question about the means by which she obtained her knowledge; and as indicated, in the normal case she will not have employed any means. Still, it is clear to her, given her awareness of her grounds for the statement, that she did not make her statement irresponsibly. It is not, for example, as if the statement were merely a matter of speculation or wishful thinking. On the one hand, then, she is aware of having good grounds for her statement; on the other hand, clearly her grounds are not adequate evidence for the truth of the statement. (2013a: 52; emphasis in original) I want to bring out four related features of the idea of intelligibility at play here. First, the fact about the subject’s condition in the light of which

354

Casey Doyle

her possession of self-knowledge is intelligible needn’t be evidence, that is, something that is taken to support the truth of what is believed. Your practical reasons don’t support the proposition that you will act in that way. Second, self-knowledge isn’t subject to the “how do you know?” question, where an answer cites a non-circular reason for belief. So the idea of rendering self-knowledge intelligible isn’t the same as justifying it or defending it with reasons. Third, what renders the self-ascription intelligible reveals to the subject that she is responsible in making it, and, fourth, this is manifest in her being able to rule out alternatives to her possessing self-knowledge, such as that she is confused, self-deceived, or engaging in wishful thinking. Here is how this applies to our case. Suppose I take myself to know that I believe that p and I tell you this. My claim to know cannot be challenged on the basis of my reasons, evidence, or sources. Self-ascriptions aren’t subject to ordinary forms of criticism and doubt (Bar-On 2004). But one shouldn’t get carried away with first-person authority. Even if we cannot challenge self-ascriptions in that way—by asking the “how can you tell?” question—there are others. You might challenge me by suggesting that I am insincere, confused, or deluded. You suggest that I am not in a position to know my own mind. And if you can issue me a challenge like that, I can issue it to myself. I can wonder whether I am confused or self-deceived, if I am really in a position to know. The thought, then, is that there must be something about my condition in virtue of which I can rule out these alternatives. We can get a handle on the idea of intelligibility by considering how challenges like this are met in the interpersonal case. Suppose a friend tells you that she believes that climate change is a hoax perpetuated by the Chinese government. Surprised, you ask: “Do you really believe that?” This is a way of challenging the self-ascription: you imply that your friend is insincere or confused. What might she say in response? Hampshire is surely right that she won’t go on to provide reasons that support the claim that she believes that climate change is a hoax. We could say: she won’t justify her self-ascription. But that doesn’t mean that there is nothing for her to say. Instead, what she is likely to do is offer support for the first-order belief itself. She’ll provide reasons for believing that climate change is a hoax, respond to some well-known objections, and maybe defend others who hold it. In doing this, she isn’t providing you with reasons for the proposition that she believes that climate change is a hoax, since the considerations adduced aren’t evidence about what she believes at all. (She doesn’t say: “I have a real weakness for wild conspiracies and right wing propaganda”.) She doesn’t provide evidence that she holds the belief; she manifests the belief or expresses it in her behaviour. What she says doesn’t provide you with reasons for believing that she believes that climate change is a hoax, but the fact that she says these things does. In the interpersonal case, then, a subject can

Ringers for Belief

355

reveal that she is in a position to know, and so rule out that she is selfdeceived, by expressing her belief or providing a rationalizing explanation of it. That’s the interpersonal case. How does it work in the intrapersonal case? How can a subject render her possession of self-knowledge intelligible to herself? We are looking for something about the subject’s condition in virtue of which it makes sense to her that she is in a position to know what she believes. Here is an answer: the fact that she believes. The firstorder belief itself figures in one’s subjective point of view and renders one’s possession of self-knowledge intelligible. The response here involves treating doxastic self-knowledge on a par with knowledge of one’s sensations and phenomenally conscious events more generally. But beliefs are nonconscious states. So I don’t think the quick response will work; we cannot treat doxastic self-knowledge on a par with our knowledge of our conscious experiences (Moran 2001; Boyle 2009). Indeed, this is what we should expect if we take seriously the idea that we don’t know what we believe by any method or conscious procedure. Wittgensteinian quibbles notwithstanding, there doesn’t seem to be anything wrong with responding to challenges to self-ascriptions of sensations by saying: “yes, I am nauseous; I can feel it!”15 In the case of sensations, we can safely assume that the same fact that provides one with a reason for self-ascription—the conscious experience—renders that self-ascription intelligible from your point of view. But it can’t work that way for one’s beliefs since beliefs aren’t episodes in conscious experience.

5. Subjective Indistinguishability Again It should be clear where we are headed. The experience of conscious judgment renders the possession of self-knowledge intelligible from the subject’s point of view. Conscious judgment plays the same role, for the subject, that manifesting one’s belief by justifying does for others. It makes sense to you that you are in a position to know what you believe because, from your point of view, it strikes you that p is the thing to think. You can, as it were, feel yourself affirming it. Notice, again, this has nothing to do with the grounds for one’s self-ascription. One knows what one believes because one believes it. But when a challenge to one’s position to know arises, this can be met by attending to the experience of judgment. This is why it matters that there are ringers for belief. Think back to Carol. Suppose that you confront her with her self-deception. You suggest that she has good reason for avoiding the evidence and trying to escape from seriously entertaining the question whether her spouse is having an affair. You survey the evidence that she herself possesses, and you suggest that, deep down, she really believes that he is having an affair, perhaps even that she knows this. But, as a committed selfdeceiver, Carol resists. She undermines your evidence, perhaps not in an

356

Casey Doyle

entirely convincing way, and she avows that she believes that he is faithful. Afterward, she leaves, thinking to herself that he is, surely faithful. Now, ex hypothesi, Carol does not believe this. She is merely telling it to herself, engaging in some soothing imaginative pretense or fantasy. But, from her point of view, she is (or seems to be) judging that p. She is not in a position to know her own mind, and yet, from her point of view, it seems that she is, it seems that “my husband is faithful” is the thing to think, and that she affirms this. That’s the bad case, but things are the same in the good case. How can we respond to the challenge that we aren’t in a position to know our minds? We know from personal experience and experimental psychology that we are often self-deceived. We are often irrational, confused, and so on. How can we assure ourselves that we are in a position to know? I have claimed that the experience of conscious judgment does this for us. But since there are ringers for belief, it seems that we cannot take ourselves to have hold of a factive reason for belief. We cannot say, “of course I believe that p because I believe that p!” since that is precisely how things would strike us if we were self-deceived or confused. We can grant that the objection from subjective indistinguishability does not arise at the level of grounds or reasons, but it does arise for intelligibility, our ability to rule out that we aren’t in a position to know our minds. This is enough to show that DDSK is faced with an analogous, and serious, objection.16 One might object as follows.17 Suppose we assume a Functionalist conception of belief. We are considering two experiences, sincerely manifesting one’s belief in judging that p and its merely seeming as if one is doing that. These experiences are either associated with the same functional role, or they are not. If they are, then they are the same state, and there is no subjective indistinguishability between distinct states. If they are not, if they are manifestations of different states, then there is bound to be some introspectible difference between them. We can happily grant that any difference in functional role entails a difference in what is introspectibly discernible, but only in the following sense. If two states are functionally different, then there will be some circumstance in which the states are manifest in ways that render them introspectibly distinct. There will be some circumstances in which, by reflection alone, one can tell that one doesn’t believe that p, if one doesn’t. But that is irrelevant to the present point. The objection I am pressing against DDSK doesn’t require that there are two states that are subjectively indistinguishable in every situation; it requires only that for any conscious judgment, there is a state of appearing to judge that is subjectively indistinguishable from that experience. That is, the objection requires only subjective indistinguishability of an experience at a particular time. Still there may be an important difference between perceptual knowledge and self-knowledge. You might think that what’s especially vertigo-inducing

Ringers for Belief

357

about the perceptual case is that one can be fooled by a ringer through no fault of one’s own. That is, no matter how well you attend, reflect, and introspect, you can’t tell that you are in the bad case when you are. By contrast, you might think that self-knowledge is immune from “brute error”: any errors in forming beliefs about your beliefs are attributable to irrationality or some kind of mistake on your part (Burge 1996). Even if there are ringers for belief, we shouldn’t be troubled by them in the way that we are troubled by hallucination. Let’s grant that self-knowledge is immune from brute error and that ringers for belief always involve irrationality, insincerity, or something else going wrong. The worry is that this undermines the claim that good and bad cases are subjectively indistinguishable. That’s because you might suppose that two items are indistinguishable if and only if a subject cannot tell the difference between them no matter how well she tries. That is how it works in the perceptual case. Hallucinations are subjectively indistinguishable from perceptions because, if I am under a hallucination, no amount of focused introspection of my perceptual state can reveal to me that I am not perceiving. But if ringers for belief require a failure on my end, then it would seem to follow that there is something I can do to reveal to myself that I am in the bad case when I am. If ringers for belief require irrationality, then there is something one can do to recognize that one is in the bad case, namely overcome irrationality, see through one’s own self-deception, and so on. But we should reject the proposed interpretation of subjective indistinguishability. Subjective indistinguishability concerns how two conscious experiences seem from the subject’s point of view. Overcoming irrationality is not a matter of attending better to the subjective character of one’s experience. This difference between the cases of self-knowledge and perceptual knowledge doesn’t undermine the parallel point about subjective indistinguishability.18

6. The “Creative Nature” of Judgment Perhaps it shouldn’t come as a surprise that the objection has reemerged at the level of intelligibility. DDSK denies that self-knowledge is based on an appearance: one’s ground is the self-ascription’s truthmaker. But selfknowledge is intelligible in light of the experience of conscious judgment. I have assumed what I called the “inner assertion” model of conscious judgment. And on that model, conscious judgment is a sort of appearance. It is something one can observe or “hear” in one’s head. Since we have reintroduced the idea of an appearance, we shouldn’t be surprised that the objection has returned. That introduction was well-motivated, though, for four reasons. First, something must explain how false selfascriptions are arrived at in the first-personal way. Second, examples like self-deception reveal that there must be something psychological shared between good and bad cases. Third, it is plausible to suppose that

358

Casey Doyle

conscious judgment renders the possession of self-knowledge intelligible to the subject. Finally, fourth, it was appropriate to introduce the particular model of conscious judgment we have looked at, the inner assertion model, because, arguably, it is the dominant one in the literature. I want to end by considering whether we can avoid the objection by adopting a different conception of conscious judgment. We’ll see that perhaps we can, but that there are costs of doing so. Charles Travis thinks there aren’t ringers for belief. Earlier, we considered the idea that this claim is supported by the absence of examples. But he also offers principled reasons. His suggestion is that there cannot be ringers for belief because belief has a “creative nature”. Here is what he means. In Travis’s view, belief involves a feeling of rational compulsion. To believe is to be under the sway, or influence, of something: (one’s encountering of) things being as they are. To be under such influence is to feel it—to feel compelled to think as one thus does. One might think of belief in this respect as Lutheran: one so stands towards things, unable to do other. (2012: 396) Talk of feeling here certainly suggests that this is an event in conscious experience, though, as he emphasizes the feeling is of rational rather than psychological compulsion (Ibid.). At a minimum it seems that what Travis has in mind is a certain event, that of finding oneself under a compulsion. For this reason, it seems plausible to suppose that what he says here applies to an act of judgment as well as the state of belief; indeed, it is easier to understand it in that case, since it is hard to see how the state of believing that p, which one is in while sleeping, would always involve a feeling. On the picture, judgment is self-conscious: to be under a rational compulsion is to feel it, that is, to be aware of oneself as under it. But there is more: the self-conscious element is constitutive of or sufficient for the act of judging. As he puts it: I cannot, while seeing nothing else for me to think, treat the question as open whether so to think. Nor can I treat the question as closed without believing. . . . My so seeing myself is my having nothing else to think. So it just is my believing things to be the way in question. My view of myself thinking as I do is in this way decisive. If I so see myself, the question what to think . . . is settled for me. (Ibid.: 398, emphasis in original) If I am aware of myself as under a compulsion to affirm that p, in the first-personal way, then it just follows that I do believe that p. The creativity lies in this: the self-conscious element—self-awareness of an act of judgment—is sufficient for the state of which it is conscious. There is no

Ringers for Belief

359

gap between having that sort of awareness and believing, so there is no possibility of a ringer. I want to draw attention to the fact that a view very much along these lines has been invoked by some Epistemological Disjunctivists about perceptual knowledge to rebut the subjective indistinguishability worry. Here are some representative passages. According to the internalism I have been elaborating, someone who makes a perceptually knowledgeable judgment knows, at least implicitly, that her ground for her judgment is an experience in which it is manifest to her that things are as she judges them to be. And that knowledge of her ground is an act of the same capacity for knowledge that is in act in the perceptually knowledgeable judgment she makes on that ground, the capacity for perceptual knowledge of how things are in the world. Her knowledge of what grounds her judgment is part of the self-consciousness with which she engages in the act of the capacity for perceptual knowledge that her perceptually knowledgeable judgment is. McDowell (pp. 34–35) Her belief that p on the basis of her perception that p realizes two kinds of knowledge in one and the same act: knowledge about how things are around her as well as self-knowledge about how she knows these things around her. Kern (2017: 122) [I]f my first person thought is true, then I know that it is true from spontaneity. When I know that p, exercising my power of receptive knowledge, my receptively knowing that p and my knowing that I receptively know it are one reality. Rödl (2007: 145) On this view, perceptual judgment is a self-conscious act involving awareness of one’s judgment thanks to awareness of its ground in perceptual experience. One knows that one judges that p thanks to one’s encountering, in perception, things being that way. My claim is not that these writers commit to Travis’s picture of judgment in all its details. But there are strong affinities with Travis’s view here, even if these philosophers don’t make use of the concepts of feeling or rational compulsion.19 If, as Rödl puts it, knowing that p and knowing that I know that p are “one reality”, then it would seem that one couldn’t be confronted in the first-personal way with what seems to be an act of knowing that p, say, thanks to one’s capacity for perceptual knowledge, which wasn’t an instance of at least, believing that p. For, as they are quick to point out, these philosophers do not assume that we cannot be wrong about whether we know that p. In a bad case, when under a hallucination, both one’s perceptual

360

Casey Doyle

experience and one’s self-consciousness is “defective”: one is mistaken both in believing that p and in believing that one knows and sees that p (De Bruijn et al. 2018: 100–101). Still, if this self-consciousness is part of what it is to exercise one’s capacity for perceptual judgment, in good and bad cases, then it is hard to see how one could have the relevant form of self-consciousness without it being self-consciousness of an act of judgment. Like Travis, these Disjunctivists deny that a judgment is a certain way of representing that p, in inner speech rather than outer; instead, it is an implicitly self-referential act of affirming a proposition in awareness of its ground. In judging that p, one represents oneself (implicitly) as under the authority of what one takes to be a reason.20 Hence, it is thought, being in a position to know that p, by seeing that p, is sufficient to put one in a position to know that one sees that p, and so know that one knows that p. It is this picture of judgment and self-consciousness in judgment to which McDowell and others have appealed as part of their response to the objection from subjective indistinguishability for the case of perceptual knowledge (McDowell 1998a, 1998d, 2013, this volume; Rödl 2007; Kern 2017; Haddock this volume). According to them, the objection gets going by assuming that knowing that one sees that p must be achieved by some capacity other than one’s capacity for perceptual knowledge, something like an infallible capacity of inner sense. But on the alternative view, one knows that one sees that p by the very same means as one knows that p: through one’s capacity for perceptual knowledge. Having justification for my belief that p brings with it, for free, justification for believing that I see that p and believe that p. It isn’t my aim here to assess this view on its merits, at least as a response to the objection from subjective indistinguishability for perceptual knowledge. Instead, I want to consider the view of self-consciousness in judgment. There is something to be said for this conception and something against it. I’ll end by mentioning both. We already know what is to be said for it. If there is no gap between awareness of conscious judgment and belief, then, as Travis claims, there are no ringers for belief. And if there are no ringers for belief, then the objection from subjective indistinguishability doesn’t face the Disjunctive conception of doxastic self-knowledge. That this should be important to Disjunctivists about self-knowledge should be clear.21 But avoiding the objection comes at the cost of rejecting our earlier claim: that our capacity for first-person knowledge is fallible and thus that there is something subjectively indistinguishable between good and bad cases, something that misleads us when we err about our own minds. To deny this seems, on the face of it, quite implausible. It seems to require denying the familiar (and less familiar) examples of self-deception, imagination, and delusion. More generally, the worry is that denying that there

Ringers for Belief

361

is a gap between the experience of judgment and the reality of belief requires accepting that we are infallible about our beliefs. Of course, it must be said that defenders of this picture of judgment largely deny that we are infallible.22 But they avoid that commitment by adopting other radical claims. Zimmerman argues that errors in self-ascriptions are always attributable to other forms of evidence (like observation or inference from behaviour) and so do not genuinely arise from the firstperson perspective (2006). Otherwise, one might reject the commitment by rejecting the view that “I” is a referring term (Haddock this volume). Travis holds that self-ascriptions of belief (and other mental states) aren’t really the expressions of knowledge at all. Such claims aren’t liable to error because they are not “environmentally embedded” (Travis 2012: 407–408). If according a status (such as “believing that p”) to an object cannot come apart from the object’s having that status, then no “question of truth turns on claims according the status” (Ibid.).23 That these are highly controversial claims should, I take it, be clear. It appears, then, that we face a dilemma, for the following two claims are inconsistent: FALLIBILITY: Our capacity for first-person knowledge is fallible, in the sense that it is possible to be presented, in the first-personal way, with what seems like a belief, but is not. CREATIVITY:24 Necessarily, if you take yourself to believe that p in the first-personal way, you believe that p. If Fallibility is true, then there are ringers for belief, and the objection from subjective indistinguishability gains purchase. This is true even if one accepts DDSK, since there are ringers for that which we would appeal to in meeting challenges to our self-ascriptions. On the other hand, if Creativity is true, then there aren’t ringers for belief. If you are presented to yourself as believing that p in the first-personal way, by judging that p, then it just follows that you do believe that p. But this seems out of step with our common sense understanding of our own fallibility, and, perhaps worse, it may require that we accept a radical view of self-knowledge as either infallible or as failing to be genuine knowledge altogether.25 I am not sure which horn of this dilemma to grab, but I do think it confronts the Disjunctivist about self-knowledge.

Notes 1. I have benefitted from written comments from Dorit Bar-On, Lucy Campbell, Joe Milburn, Ben Sorgiovanni, and Preston Stovall. Some of the material was presented at the University of Helsinki and the Oxford Mind Work in Progress Seminar. Thanks to the participants and especially to Anil Gomes, Alexander Green, and Nick Shea. Finally, as will be clear, I have been helped a great deal by the other contributions to this volume.

362

Casey Doyle

2. Nothing here hangs on the characterization of the content of perception as propositional, nor on which side of the disjunction we place illusion. 3. This is what French (2016) calls the Disjunctivist’s negative thesis, it is the rejection of the New Evil Genius Hypothesis. 4. This specification of the way in which the good case lends better support is what French calls the Disjunctivist’s “positive thesis”, ibid. 5. I say “some” since, for example, Pritchard (2012) doesn’t develop any particular view about introspection or reflective access to one’s mental states. 6. Neta (2008) argues that subjective indistinguishability is asymmetric: when you are in the bad case you can’t tell that you aren’t in the good case, but when you are in the good case you can tell that you are. Pritchard (2012) responds to the objection by distinguishing between ‘favouring’ and ‘distinguishing’ evidence. One’s total evidence can favor the good case even though one cannot distinguish them. 7. It is worth emphasizing that Bar-On suggests but does not commit to this view. See her (2004: ch. 9), as well as Bar-On and Johnson, this volume. 8. For this reason it does not seem to me plausible that false self-ascriptions can always be chalked up to “third-personal” sources of information about oneself, like observation or inference, as Zimmerman (2006) claims. 9. Zimmerman couches his point in terms of access conscious beliefs, Constitutivism, and a background commitment to Functionalism. This is because he defends Shoemaker’s Functionalist Constitutivist account of self-knowledge. But none of these commitments are necessary for his general point, which I have captured in dispositional terms. 10. For our purposes we can treat withholding belief as a case of not believing. 11. That self-deception involves a failure of self-knowledge is convincingly argued by Holton (2001), Funkhouser (2005). 12. See the essays collected in Bayne and Montague (2011). 13. ‘Phenomenology’ is used here to refer to the distinctive “what it’s likeness” of an experience, as opposed to the first-person investigation of the structure of conscious mental life. 14. See Cassam (2009) for discussion. 15. Falvey (2000) makes a similar point. 16. It is worth emphasizing that this issue arises because belief is a nonconscious standing attitude. It seems to me that it would make sense to respond in this way to challenges to self-ascriptions of sensations like pain. 17. I am grateful to Ben Sorgiovanni for raising this issue. 18. Thanks to Lucy Campbell and Ben Sorgiovanni for raising these worries. 19. The view is anticipated by Cook Wilson: “The consciousness that the knowing process is a knowing process must be contained within the knowing process itself” (1926: 107). 20. On “implicitly”, see Boyle (2011). 21. To be clear: the discussions from which the passages are quoted earlier are not concerned with the question of whether there are ringers for belief. I am not assuming that McDowell, Kern, and Rödl would necessarily agree with Bar-On, Travis, and Zimmerman. I am only drawing attention to the affinity between Travis’s view of judgment and theirs, and so the relevance of the current discussion to the analogous issue in the case of perceptual knowledge. 22. Cook Wilson is an exception. 23. Travis tries to explain away cases of self-deception (see 2012: 401–402). I won’t go into his discussion here. It is sufficient for our purposes that he denies that one can be presented in the first-personal way as if one believed that p when one doesn’t. And that is at odds with our earlier discussion.

Ringers for Belief

363

24. So named for Travis’s way of putting the point. Creativity is not quite the same thing as Constitutivism, a view mentioned earlier. The latter view holds that, necessarily, if you believe that p, you believe that you believe that p (Shoemaker 1996; Boyle 2011). Others hold that, necessarily, if you believe that you believe that p, then you believe that p (Heal 1994; Bilgrami 2006; Coliva 2016). Both are claims about the nature of the state of first- or secondorder belief. Creativity is a thesis about an act of ‘taking’ oneself to believe that p, i.e. being presented with what one takes to be a belief in the firstpersonal way, which I have claimed is realized in an act of self-conscious judgment. One could accept that a self-conscious feeling of rational compulsion is sufficient for belief while denying that first-order belief entails secondorder belief and that second-order belief entails first-order belief. 25. There is another reason for someone of a Disjunctivist temperament to be suspicious of Creativity. Some Disjunctivists diagnose their opponents as assuming what Martin calls a ‘substantive epistemic principle’, namely that there are aspects of the subjective realm about which we are infallible (2004: 50). (See also the Introduction to this volume, sec. 2, and McDowell 2013). The Disjunctivist who denies that there is anything about the subjective about which we are infallible should reject Creativity. Our minds are a part of the environment, and we are liable to error about them.

References Armstrong, David, 1993 [1968], A Materialist Theory of the Mind. London: Routledge. Bar-On, Dorit, 2004, Speaking My Mind. Oxford: Oxford University Press. Bar-On, Dorit and Johnson, Drew, this volume, Epistemological disjunctivism: Perception, expression, and self-knowledge. Bayne, Tim and Montague, Michelle, eds., 2011, Cognitive Phenomenology. Oxford: Oxford University Press. Bilgrami, Akeel, 2006, Self-Knowledge and Resentment. Cambridge, MA: Harvard University Press. Boyle, Matthew, 2009, Two kinds of self-knowledge, Philosophy and Phenomenological Research 78(1): 133–164. ———, 2011, Transparent self-knowledge, Proceedings of the Aristotelian Society (supp. vol.) 85(1): 223–241. Burge, Tyler, 1996, Our entitlement to self-knowledge, Proceedings of the Aristotelian Society 96: 91–116. Byrne, Alex, 2005, Introspection, Philosophical Topics 33(1): 79–104. Cassam, Quassim, 2009, The basis of self-knowledge, Erkenntnis 71(1): 3–18. ———, 2010, Judging, believing and thinking, Philosophical Issues 20: 80–95. ———, 2014, Self-Knowledge for Humans. Oxford: Oxford University Press. Coliva, Annalisa, 2016, The Varieties of Self-Knowledge. London: Palgrave Macmillan. Cook Wilson, John, 1926, Statement and Inference. Ed. A.S.L. Farquharson. Oxford: Clarendon Press. Currie, Gregory and Jureidini, Jon, 2001, Delusions, rationality, empathy, Philosophy, Psychiatry and Psychology 8(2–3): 159–162. De Bruijn, David, Goldhaber, Charles, Kern, Andrea, McDowell, John, Smithies, Declan, Springle, Alison and Yoon, Bosuk, 2018, Discussion of John McDowell’s “Perceptual experience and empirical rationality”, Analytic Philosophy 59(1): 99–111.

364

Casey Doyle

Falvey, Kevin, 2000, The basis of first-person authority, Philosophical Topics 28(2): 69–99. French, Craig, 2016, The formulation of epistemological disjunctivism, Philosophy and Phenomenological Research 92(1): 86–104. Funkhouser, Eric, 2005, Do the self-deceived get what they want? Pacific Philosophical Quarterly 86(3): 295–312. Gendler, Tamar, 2007, Self-deception as pretense, Philosophical Perspectives 21: 231–258. Gibbons, John, 2006, Access externalism, Mind 115(457): 19–39. Haddock, Adrian, Disjunctivism, skepticism, and the first person. 259–279. Hampshire, Stuart, 1979, Some difficulties in knowing, in Ted Honderich and Myles Burnyeat, eds. Philosophy as It Is. Harmondsworth: Penguin. Heal, Jane, 1994, Moore’s paradox: A Wittgensteinian approach. Mind 103(409): 5–24. Holton, Richard, 2001, What is the role of the self in self-deception? Proceedings of the Aristotelian Society 101(1): 53–69. Kern, Andrea, 2017, Sources of Knowledge. Trans. D. Smyth. Cambridge, MA: Harvard University Press. Marcus, Eric, 2016, To believe is to know that you believe, dialectica 70(3): 375–406. Martin, Michael, 2004, The limits of self-awareness, Philosophical Studies 120: 37–89. McDowell, John, 1998a, Criteria, defeasibility and knowledge, in his Meaning, Knowledge, and Reality. Cambridge, MA: Harvard University Press. ———, 1998b. Knowledge and the internal, in his Meaning, Knowledge and Reality. Cambridge, MA: Harvard University Press. ———, 1998c, Knowledge by hearsay, in his Meaning, Knowledge and Reality. Cambridge, MA: Harvard University Press. ———, 1998d, Reply to Wright in Crispin Wright, Barry Smith and Cynthia Macdonald, eds. Knowing Our Own Minds. Oxford: Oxford University Press. ———, 2013, Perceptual experience: Both relational and contentful, European Journal of Philosophy 21(1): 144–157. ———, this volume, Perceptual experience and empirical rationality. Moran, Richard, 2001, Authority and Estrangement. Princeton, NJ: Princeton University Press. Neta, Ram, 2008, In defense of disjunctivism, in Haddock, Adrian and MacPherson, Fiona, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 311–329. O’Shaughnessy, Brian, 2000, Consciousness and the World. Oxford: Oxford University Press. Peacocke, Christopher, 1998, Conscious attitudes, attention and self-knowledge, in C. Wright, B. Smith and C. Macdonald, eds. Knowing Our Own Minds. Oxford: Oxford University Press. Pitt, David, 2004, The phenomenology of cognition, or, what is it like to think that p? Philosophy and Phenomenological Research 69: 1–36. Pritchard, Duncan, 2012, Epistemological Disjunctivism. Oxford: Oxford University Press. Rödl, Sebastian, 2007, Self-Consciousness. Cambridge, MA: Harvard University Press.

Ringers for Belief

365

Roessler, Johannes, 2013a, The epistemic role of intentions, Proceedings of the Aristotelian Society 113: 41–56. ———, 2013b, The silence of self-knowledge, Philosophical Explorations 16(1): 1–17. Shoemaker, Sydney, 1996, The First-Person Perspective and Other Essays. Cambridge: Cambridge University Press. ———, 2012, Self-intimation and second-order belief, in Declan Smithies and Daniel Stoljar, eds. Introspection and Consciousness. Oxford: Oxford University Press. Travis, Charles, 2012, While under the influence, in his Perception. Oxford: Oxford University Press. Velleman, David, 2000, On the aim of belief, in his The Possibility of Practical Reason. Oxford: Oxford University Press. Wright, Crispin, 1998, Self-knowledge: The Wittgensteinian legacy, in Crispin Wright, Barry Smith and Cynthia Macdonald, eds. Knowing Our Own Minds. Oxford: Oxford University Press. Zimmerman, Aaron, 2006, Basic self-knowledge: Answering Peacocke’s criticism of constitutivism, Philosophical Studies 128: 337–379.

18 Disjunctivism and Other Minds Anita Avramides

1. Introduction Towards the later part of the 20th century, a position emerged in philosophy that came to be known as disjunctivism. The position has been advocated by a number of philosophers, all of whom were in some way or another thinking about the philosophy of perception.1 However, while some were primarily interested in the philosophy of perception, there is one philosopher whose work has come to be closely associated with disjunctivism whose interest in perception forms only part of a much wider concern.2 That philosopher is John McDowell, and understanding the wider concern that drives his disjunctivism is one of the main aims of this chapter. But my interest in understanding the wider context out of which McDowellian disjunctivism emerges is not driven by concerns arising in the philosophy of perception. Rather, my main interest is in concerns arising in connection with other minds. Of those whose work is often associated with disjunctivism, only McDowellian disjunctivism is brought to bear (directly) on issues in connection with other minds. What precisely is the connection between disjunctivism and matters having to do with other minds? How is a chapter devoted to these topics different from one devoted to disjunctivism and the world of objects or bodies? I want to suggest that the former is an extension of the latter. While I concede that one needn’t understand things in this way, I would urge that doing so helps one to appreciate the way in which disjunctivism, at least in McDowell’s work, is an idea that can be put to work in dislodging a picture of mind in relation to the world, a picture which also has ramifications for how we think of mind in relation to other minds. In so far as one does understand disjunctivism in connection with other minds as an extension from its application in connection with bodies, it is obvious that we must first have a firm understanding of disjunctivism as it arises in connection with ordinary perception of objects, before we turn to ask how appeal to this view can help us with issues concerning other minds. We also need to appreciate not just that an application of disjunctivism to the world of bodies is insufficient on its own to do any work in connection with other minds, but also why this is so.

Disjunctivism and Other Minds

367

Appreciating why this is so requires that we understand just what work needs doing in connection with other minds. In this chapter I will explain what work McDowell thinks needs doing in this connection, before turning to explain how he believes disjunctivism can help here. As I have introduced the topics that will be of my concern here, it should be apparent to the discerning reader that an application of a position (disjunctivism) connected to considerations in the philosophy of perception to issues connected with other minds will have what to some will seem a curious outcome: it may be thought to bring us to talk of perception in connection with other minds. While it is undeniable that my relation to, and knowledge of, the world of bodies is through perception, it has not been widely thought that my relation to, and knowledge of, the mind of another is through perception. But recently the idea has gained currency, fuelled in part by McDowell’s work.3 It is interesting to ask whether adopting a disjunctive account of perception experience can be restricted to consideration of bodies, rejecting its extension to other minds. If one thinks of disjunctivism quite generally, this may be possible. However, if one restricts consideration only to McDowellian disjunctivism, an extension to other minds, I shall argue, follows naturally.4 One of the things I hope to show in this chapter is that it is dubiously coherent to adopt McDowellian disjunctivism in connection with perception of bodies while refusing its extension to other minds. This follows from an understanding of the larger concerns that drive McDowellian disjunctivism. Furthermore, if I am right about the demands of the McDowellian project here, then that project can also be understood to be committed to a very particular understanding both of our relations to the world and of our relations to the minds of others. This chapter will proceed as follows: in section (i) will introduce McDowellian disjunctivism and explain the wider context out of which it emerges. In section (ii) I introduce other minds and explain why experiential disjunctivism alone has no particular ramifications for issues in connection with them. In section (iii) I explain how the wider context that drives McDowell’s work also informs his interpretation of the notion of criteria, which notion is designed, when coupled with his disjunctive account of experience, to have significant ramifications for issues concerning other minds. I will end by considering the way in which application of the disjunctive move in connection with other minds affords an opportunity for re-assessing our relations to others. As the main trust of this chapter is to explain how experiential disjunctivism can be extended to other minds, I will not be concerned here to raise questions or problems for McDowell’s approach.

2. Experiential Disjunctivism Disjunctivism emerges in several of McDowell’s early papers. Much of the recent discussion of McDowellian disjunctivism tends to concentrate

368 Anita Avramides on the strand that is rooted in McDowell’s preoccupations with the work of Wilfred Sellars, in particular the Sellarsian idea of knowledge (at least in the case of rational animals) being bound up with a certain standing in the space of reasons. But it is worth remembering another strand in the development of McDowell’s disjunctivism, one that may be appreciated by reflection on some ideas in the work of Bertrand Russell. Complex though this work is, it is possible to extract from it a core, which can serve to highlight McDowell’s central concerns when proposing his disjunctivism.5 Russell’s idea is that it is possible to combine genuinely referring expressions (logically proper names) and singular propositions. The characteristic of such a combination is that the proposition expressed by it would not be available to be expressed at all if the objects referred to did not exist. McDowell does two things with Russell’s idea: firstly, he drops the restriction to logically proper names; and secondly, he gives the idea a psychological application. Consider McDowell’s first move first. McDowell explains that Russell’s reasons for restricting singular propositions to combinations restricted to logically proper names can be traced to the fact that Russell believed that genuinely singular propositions can only be entertained where there can be no illusions as to the existence of an object of the appropriate kind. In cases where there can be such illusions, Russell employs the apparatus of the Theory of Descriptions to equip such sentences (utterances) with non-singular propositions that those sentences (utterances) can be understood to express. But McDowell thinks that there is another way around the problem Russell here identifies. Rather than ‘devising a kind of non-singular proposition to be associated with all [sentences/utterances] alike’, McDowell suggests that we say the following: ‘some sentences (or utterances) of a given range express singular propositions, whereas others present the illusory appearance of doing so (to those not in the know)’ (McDowell 1986: 229). We see here the emergence of what might be labelled a disjunctive account of singular propositions. The point of the disjunctive move is to avoid the need to postulate non-singular propositions (courtesy of the Theory of Descriptions) to account for what is expressed by sentences where the suitably related object does not exist.6 Next, McDowell takes the Russellian idea of a singular proposition, suitably adjusted, and gives it a psychological application. He suggests that such an application would allow us to say that ‘which configuration a mind can get itself into is partly determined by which objects exist in the world’ (McDowell 1986: 230). As we have just seen, Russell restricts his view to logically proper names because error cannot occur when these are constituents of singular propositions. McDowell suggests that Russell insists on this restriction because in the case of logically proper names there is (what Russell calls) acquaintance with the object in question. McDowell tells us that Russell takes perception to be the ‘paradigm

Disjunctivism and Other Minds

369

of acquaintance’ and that, in order to ensure the impossibility of error, Russell takes acquaintance to be with sense data.7 But McDowell questions Russell’s restriction here and suggests that we can adapt Russell’s idea to allow that acquaintance be not with sense data but with objects in the world. He then extracts the notion of acquaintance from Russell’s epistemological framework (involving sense data) and applies it to perceptual relations between minds and objects in the world external to them. Immediate acquaintance with objects in the world through/in perceptual experience makes possible singular thoughts targeted on particular objects. McDowell not only questions Russell’s restriction of acquaintance to sense data, but he also questions the motivation for it. As we have seen, that motivation, in the first instance at least, is a commitment to subjective infallibility. For Russell, it would not make sense to allow that a subject could be mistaken about how things are regarding her inner life. It is here that McDowell identifies in Russell’s work ‘a conception of the inner life, and the subject’s knowledge of it, that it seems fair to label “Cartesian”’ (Ibid.) This conception becomes the central target of McDowell’s philosophical preoccupations. McDowell insists that if we reject this conception of the inner life, we will be freed up to embrace, inter alia, the suggested adaptation of Russell’s original idea about singular propositions. At this point I want to leave to one side McDowell’s concern with singular thought, and concentrate on the conception of mind he aims to reject. What McDowell takes to be distinctively Cartesian about the inner life is that it is taken to be an autonomous realm, entirely transparent to its subject.8 One problem with this conception of the inner life is the difficulty of reconciling it with an account of the access such a realm can have to an outer, external, world—a world that includes other minded individuals. Descartes’s conception of mind is plagued by the possibility of an evil demon who threatens to take away the world and all others like myself; in the case of other minds, what threatens is solipsism.9 McDowell sees this threat as the result of two moves made by Descartes. The first of these is the move that extends the notions of truth and knowledge to the world of appearances. While this move is significant, McDowell thinks that, on its own, it is not sufficient to drive us to worry that we may be completely out of touch with external reality.10 This concern arises with Descartes’s second move, one which endorses ‘a picture of subjectivity as a region of reality whose layout is transparent—accessible through and through—to the capacity for knowledge that is newly recognized when appearances are brought within the range of truth and knowability’ (1986: 240).11 The important words here are: ‘accessible through and through’. What we get when we join this move with the first is that we have an inner realm to which the notions of truth and knowledge apply, and which realm is entirely transparent to the subject of it. We can now

370 Anita Avramides appreciate that what we have here is an inner realm that is being thought of as ‘self-standing, with everything arranged as it is independently of external circumstances’. McDowell labels this conception of the inner realm ‘autonomous’ (1986: 243).12 McDowell rejects this, Cartesian, conception of mind. In particular he rejects the picture of subjectivity as a region transparent through and through. He believes that rejecting this conception of mind liberates one to entertain Russell’s (adapted) idea, now in a non-Cartesian context. Once we drop the Cartesian picture of subjectivity, we are free to conceptualise not only our thought but also the experience that gives rise to it in another way. In the case of the latter, we can say either our experience is such as to yield us access to the world, or it is such as to yield only apparent access—and how things seem to one falls short of infallible knowledge as to which disjunct is in question (Ibid.: 240). The Russellian idea of allowing objects to figure in our experience together with the disjunctive move which tells us that our experience may be of such a type, gives us a picture of a subject and her world, of the inner and the outer, ‘as interpenetrating, not separated from one another by the characteristically Cartesian divide’ (Ibid.: 241). By employing this disjunctive move, McDowell thinks we can avoid Descartes’s problem of vulnerability to an evil demon who takes away access to the world.13 If we do adopt this way and avoid this problem, we need to understand the obvious fact that experiencing subjects are fallible. McDowell suggests that we locate a subject’s fallibility in her perceptual capacities: in any given case, these capacities are such as either to give her access to reality or to give her only apparent access. When our fallibility is understood in this way, we are in a position to allow that, where there is access, there we are in a position to have knowledge. McDowell’s embrace of fallibilism is part and parcel of his rejection of a Cartesian conception of mind as autonomous. At the heart of McDowell’s work here is a commitment to a thesis about perceptual capacities and the nature of the experience it may yield: my (fallible) perceptual capacities are such as to yield acquaintance with objects in the world; objects can figure in my perceptual experience. Once we see this, we can also see that a scepticism of the sort that plagued Descartes cannot arise. McDowell says in one place that his approach is to ‘reject the assumption that generates the sceptic’s problem’ (McDowell 1982: 469).14 What my brief summary of the way McDowell discusses disjunctivism in connection with Russell’s work reveals is that McDowell’s real preoccupation is to dislodge a Cartesian conception of mind, a conception that gives rise to a radical scepticism concerning the world of bodies and other minds. In connection with the disjunctive move with regards experience, McDowell acknowledges the work of J.M. Hinton and P. Snowdon.15 Snowdon’s work is largely concerned with the causal account of perception put forward by H.P. Grice and, in particular, with aspect of it that

Disjunctivism and Other Minds

371

attributes to a perceiver a state reportable in a sentence beginning ‘It looks to S as if . . .’. In opposition to the causal account, which is committed to a single interpretation of the looks statement in cases where there is a causal element and cases where there is none, Snowdon suggests that we interpret the looks statement as follows: either one is in an L-state that is ‘intrinsically independent of its surroundings’ or one is in an L-state that intrinsically ‘involves the surrounding objects’.16 Despite the fact that Hinton, Snowdon and McDowell all end up with a disjunctive account of perceptual experience, it is quite clear that the preoccupations that lead them to this account are very different. Snowdon and Hinton’s account emerges from their preoccupation with perception, while McDowell’s emerges from his preoccupation with a Cartesian conception of mind as outlined earlier.17 McDowell’s central concern is to break down the conceptual divide between the subjective and objective worlds that gives rise to a radical scepticism. Knowledge of the world and of other minds is problematic if one accepts a Cartesian conception of mind. If one starts where Descartes does, our access to the world looks shaky, that is vulnerable to sceptical attack. While McDowell’s concerns are epistemological, he is also concerned with the metaphysical picture of mind that gives rise to them. In aiming to grapple with Cartesian scepticism, McDowell is forced to consider matters both epistemological and metaphysical.18 His response to both issues is to try to persuade us of the merits of a disjunctive account of perceptual experience. A non-Cartesian conception of mind is one that allows for a certain sort of thought and knowledge where the objects of thought exist. What drives McDowell to experiential disjunctivism is not so much a concern with the philosophy of perception per se as with the scepticism that arises from adoption of a Cartesian conception of mind.

3. Introducing Other Minds Unlike most other disjunctivists, McDowell does not limit his discussion to the interface between a single subjectivity and the objective world, but includes the interface between subjectivities. This is an important dimension of McDowellian disjunctivism. If we concentrate on McDowell’s preoccupations with the Cartesian conception of mind, what we find is that that conception has repercussions not just for our knowledge of the world of bodies but also for our knowledge of other minds. The Cartesian picture threatens solipsism, and McDowell is also keen to counter this. But to do this, McDowell needs more than experiential disjunctivism, as I shall now explain. Experiential disjunctivism may be thought to be limited to the interface between a single subjectivity and the world. If we were also to be interested in the interface between one minded individual (or person) and another, it would be required that we examine our account of the

372 Anita Avramides relationship between behaviour and mental states. Another way to put the point is this: that my experience has the capacity to reach all the way to your body may not yet account for my access to and knowledge of your mental life without further consideration of the relationship of (your) mind to (your) behaviour. Experiential disjunctivism needs to be coupled with an account of the relationship of mind to behaviour in order for the former to yield access and knowledge that extends to other minds. Let us for the moment accept a disjunctive account of our knowledge of the world of bodies. How might we go from this to knowledge of another mind? We might consider the possibility that, while holding a disjunctive account of perceptual experience of another’s body, our knowledge of another’s mind must be the result of inference from what we can experience (the other’s body). Such a position, however, may be thought to be inconsistent with McDowell’s reasons for embracing a disjunctive account of experience. What McDowell objects to in a Cartesian conception of mind is not its commitment to an infallible subjective realm per se, but the problems this gives rise to when it comes to access beyond this realm. All knowledge beyond this realm must be taken to be insecure, in the sense of being liable to sceptical attack. The Cartesian sceptic takes the fact that our perceptions of the world are sometimes in error and pushes this to ask how we can know that any of our perceptions have a connection to a world ‘outside’ the subjective realm. McDowell thinks we can halt this slide. He wants an account of knowledge that is secure—by which he means knowledge that is such as not to allow for the Cartesian radical sceptical question—and he believes that one can get this if one is careful about the metaphysical picture that informs one’s epistemological position. A disjunctive account of experience allows for a metaphysical picture of mind in relation to a world that can yield knowledge of that world that is secure; epistemological security is the pay-off of adopting experiential disjunctivism.19 But McDowell does not take epistemological security to stop with knowledge of the world of objects; he believes that it can be extended to other minds. Furthermore, he does not believe that inferential accounts can provide this security. It is arguable that McDowell thinks that this epistemological security must extend to other minds. We can understand why this extension is required once we understand the concerns that drive McDowell to adopt experiential disjunctivism. McDowell presents an account of experience of the world that does not allow for a radical, Cartesian-style, mindbody divide. Prima facie, there would be something odd about insisting that there is no metaphysical division in place when we consider the way that minded beings experience the world, but allowing for such a division when considering the manner in which minded beings act in the world.20 Just as a Cartesian divide throws up seemingly insurmountable problems concerning our access to and knowledge of the world, it may be thought also to lead to insurmountable problems concerning our access

Disjunctivism and Other Minds

373

and knowledge of the mind of another. And just as McDowell identifies the Cartesian idea of an autonomous inner realm as the thing that stands in the way of our accepting a person’s inner and outer worlds as interpenetrating rather than as separate, so we can also see this idea of autonomy as standing in the way of an interpenetration of individual inner worlds. Cartesian autonomy in the case of other minds leads us to say, ‘Only I can really know what I am thinking and feeling; my mental life is hidden in a realm that is private to me’. The mind of the other is taken to lie behind or within the behaving body in such a way as make any access I can have to it insecure. The Cartesian gives us a picture of our relations to others that takes one’s experience of another as allowing for the possibility of radical deception, as allowing for the possibility that the other is a zombie (an individual who behaves like a minded individual, but who lacks mentality). The sceptic can ask how one knows, just on the evidence of another’s behaviour, that the other is minded at all. Paraphrasing what McDowell writes in connection with the external world, we can say that we have here another example of the characteristic Cartesian willingness to face up to losing others with one’s own inner life for consolation.21 What is needed, according to McDowell, is a conception of mind in relation to behaviour that does not allow for this sceptical doubt to arise. McDowell points out that lying at the heart of the Cartesian way of thinking about mind is the idea that what I can see—your behaviour—is the same whether you are minded or not. The behaving body is here thought of on the model of sense data—as that which may be conceived of in a uniform manner however else things are. And the pressure to think in this way in the case of other minds may also be thought to be similar to the pressure to think in terms of sense data in the case of our experience of the world: the possibility of deception.22 But McDowell does not believe that a retreat to a metaphysical divide is demanded in either case. In this connection we may recall an observation of Davidson’s: ‘if there is a logical or epistemic barrier between the mind and nature, it not only prevents us from seeing out; it also blocks a view from the outside in’ (1991: 154).23 The thought here is that if we allow for barriers between mind and nature, those barriers will have repercussions for our knowledge of other minds. The same barrier that stands in the way of my seeing out to the world of objects will prevent my seeing all the way ‘in’ to the mind of another. In the case of others, the barrier will present itself in the view we take of the behaviour of the other. We must be careful how we understand this behaviour if we are not to block a view ‘from the outside in’. That these barriers should have this double effect may be evident when one reflects on the fact that there are two main avenues for how mind relates to world: through perception and through action. That the barrier works in both directions is why a rejection of the barrier in one direction would naturally lead one to reject it in the other. If what interests one is the breaking down of logical and epistemic barriers between

374 Anita Avramides mind and nature, then one should be concerned with the way they affect the view ‘from the outside in’ just as much as the way they affect the view ‘from the inside out’. It is for this reason that I said at the outset that it is dubiously coherent that McDowellian disjunctivism be confined to consideration of the mind’s experience of the world of objects. So far I have concentrated on McDowell’s motivation for adopting a disjunctive account of a subject’s experience and knowledge of the world of bodies; I want now to consider how this might be extended to cover a subject’s experience and knowledge of another’s mind.

4. Extended Experiential Disjunctivism As I have explained, experiential disjunctivism alone is insufficient to yield knowledge of other minds. For an account of knowledge that extends to other minds, we need to understand how the mind of the other is related to the body of the other. McDowell believes that the key to understanding the latter relationship can be found in the Wittgensteinian notion of a criterion. The notion of a criterion, McDowell tells us is ‘thought to afford .  .  . a novel response to the traditional problem of other minds’ (1982: 455). The question is how to understand this notion and the ‘novel response’ it brings to this problem. We also need to understand how this novel response to the problem of other minds relates to a disjunctive account of experience. My suggestion is that we think of the notion of a criterion as playing a key role for McDowell in the extension or augmentation of the disjunctive account of experience in order that it may yield knowledge of other minds.24 Before I turn, in section (iii)(b), to explain how the notion of a criterion is meant to work in connection with other minds, I expand, in section (iii)(a), on how it is that McDowell thinks of knowledge. Just as with experiential disjunctivism, how McDowell thinks about knowledge is an important driver in his understanding of the Wittgensteinian notion of a criterion in connection with mental states. (a) Knowledge of the World and Other Minds Earlier I referred to McDowell as a fallibilist, but there are different ways of understanding a position that goes by this name. Simon Blackburn defines fallibilism thus: The doctrine due to Charles Sanders Pierce, that it is not necessary that beliefs be certain, or grounded on certainty. We may justifiably rest content with beliefs in circumstances in which further evidence, forcing us to revise our opinion, may yet come in. Indeed, since this is always our position, unless we settle for it we shall be driven to scepticism. (Blackburn 1994: 135)

Disjunctivism and Other Minds

375

The idea here is that our evidence for some belief or claim, while excellent, may still be impeachable. Let us for the moment consider a certain kind of evidence, perceptual evidence. Say someone asks you how you know, for example, that there is a cat before you and you reply (standing in front of the cat, and having good eyesight and a non-occluded line of sight), ‘I know because I can see that there is a cat before me’. How should we understand this appeal to perception here? A Piercean fallibilist might reply that we should understand our perceptual evidence as warranting a claim to know while allowing for the possibility that further evidence may force us to revise our claim. Of course, further evidence can always undermine earlier evidence. The question that one might press (and that McDowell does press)25 is this: in the light of the fact that later evidence may require that we revise our claim to know, how should we think of this earlier evidence? The Piercean fallibilist may reply that we should think of the earlier evidence as entitling our claim to know, while at the same time being defeasible. McDowell thinks that to understand perceptual evidence in this way is the result of thinking that how things are with us is one thing, and how things are in the world external to us is another. And to think in this way, McDowell points out, is to leave the door wide open for the sceptic to insert his worry: all we really are in position to know is how things are with us. Piercean fallibilism leaves room for radical scepticism (indeed, as Blackburn puts it in the quotation to begin this subsection, allowing for claims to knowledge in the face of fallible evidence is precisely the move we should make if we are not to be driven to scepticism). According to McDowell, to think of things along the lines of what I am calling Piercean fallibilism requires that we accept two ideas. The first is that when one claims to know (that, say, there is a cat before one) the evidence for one’s claim must be understood ‘as something one’s possession of which one can assure oneself of independently of the claim itself’ (1982: 470).26 I shall return to this idea later. The second idea is that of a Cartesian infallible inner realm. This is just the Cartesian idea of autonomy that, as we saw in section (i), McDowell rejects. The alternative he proposes is the idea of subjectivity not thought of as so arranged independently of external circumstance. If we accept this non-Cartesian idea, McDowell thinks that we can move away from Piercean fallibilism: where our perceptual capacities are such as to be in acquaintance with their objects, there our evidence is indefeasible; and where our evidence is indefeasible, it is not possible to count as the kind of evidence that it is while things not be as they seem to be. Indefeasible evidence yields knowledge that is secure from sceptical attack. What McDowell aims to rule out is a conception of knowledge whereby further evidence may force us to admit that, while allowing that the conditions for knowledge are fulfilled, knowledge does not result. According to McDowell, it is by dint of our perceptual capacities that either the evidence is such as to make the associated knowledge available, or it is not; either we have

376 Anita Avramides indefeasible reason to believe, or we do not. Where we do, the knowledge that results is secure in the sense that it is not susceptible to sceptical doubt of the sort Descartes introduces in his First Meditation. This is because, in those cases where experience delivers knowledge, the object of the experience must be understood to figure in that experience. When experience is understood in this way, Cartesian scepticism cannot get a grip. Knowledge, for McDowell, is sceptic proof.27 If we accept McDowell’s picture here, we must locate our fallibility elsewhere than did Pierce. Now we are in a position to understand how McDowellian fallibilism differs from Piercean, or standard, fallibilism. Recall the disjunctivist move: either our evidence is such as to make the associated knowledge available, or it is not; where it is not, we must reject the earlier evidence as evidence that yielded knowledge. This is, in effect, to say that not all evidence counts alike (this is where the move to talk of ‘criteria’ comes in, as I explain in section (iii)(b)). And the reason that not all evidence here counts alike is that our perceptual capacities are fallible. A fallible capacity can give rise to an experience which provides an indefeasible reason for believing, but we must firmly distinguish between a fallible capacity and the experience to which it gives rise. It is when we do this that we can see that we have a choice about where to locate our fallibility: the Piercean locates it in the epistemic significance of the experience; McDowell locates it in our perceptual capacities.28 The advantage of locating fallibility where McDowell does is this: acknowledging that one may, because of the fallibility of one’s perceptual capacities, be mistaken does not bring in its wake a need to acknowledge that, for all we know, all our experiences of the world could be in error. McDowellian fallibilism does not leave room for Cartesian scepticism. Let us now return to consider the first idea that McDowell associates with what I am calling Piercean fallibilism: the idea that one’s evidence for one’s claim to know is something one can assure oneself of in independence of the claim itself. Let us return to the situation in which you are asked how you know that there is a cat before you and you reply, ‘Because I can see that there is a cat before me’. If it is the case that your perceptual experience is such as to take in the world as it is before you, then you indeed know that there is a cat before you. That you indeed have such evidence for your claim is not something you can assure yourself of in independence of this claim and the circumstances that give rise to it. McDowell insists that to think that any independent assurance is required (that is, an assurance that can be given in independence of the claim) is to fall back into thinking in terms of Cartesian infallibility. In this connection, let us look at something else that McDowell writes: We can invert the order in which scepticism insists we should proceed, and say—as common sense would . . . —that our knowledge

Disjunctivism and Other Minds

377

that those supposed possibilities do not obtain is sustained by the fact that we know a great deal about our environment. (McDowell 2008: 370)29 The sceptic thinks that an account of knowledge must proceed by first proving that we are not being deceived by an evil demon. It is McDowell’s belief that, if we understand our common sense view of the world aright, then we need not be moved by such sceptical considerations. Rather than start with the possibility that we might be radically deceived in our world view, we should begin with confidence in our experience of it. What McDowell proposes is an account of our experience of the world that allows us to hold on to our common sense view of the world. Our understanding of these capacities and these experiences can also give rise to knowledge of the mind of another just so long as we also understand the way mind relates to behaviour aright. Just as with the world of bodies, McDowell believes that one can tell by experience whether another is, say, in pain. While the sceptic insists that we can only know another’s mind if we can show that the individual we encounter is not a zombie, McDowell insists—again in line with what he takes to be common sense—that I often do know the thoughts and feelings of others. McDowell reaffirms his commitment to common sense over scepticism in another place, this time explaining what stands in the way of a philosophical acceptance of the ‘face value’ of things. He writes: Only if the veil [of perception] is supposed in place can it seem that one would need to establish, or equip oneself with good reason to suppose, that one is not dreaming before one can be entitled to take one’s apparent perceptions at face value. Once the veil is lifted, we can say that things are the other way around; one’s good reason to believe that one is not dreaming, on the relevant occasions, can reside in all the knowledge of the environment that one’s senses are yielding one—something that does not happen when one is dreaming. (McDowell 1995: 408, fn. 19) A similar argument could be put forward to explain what stands in the way of a philosophical acceptance of the ‘face value’ of things when we consider our relationship to others.30 While the face value of one’s perceptions in connection with other minded individuals is that they often yield knowledge of what others think and feel, the sceptic insists that we are not entitled to take our experience at face value but must first show that solipsism is not the case.31 In response, it could be said—in parallel to the proposal just advanced to explain what stands in the way of taking our experience of objects at face value—that it is only if one accepts that a veil of behaviour is in place that one would feel any pressure to establish that one is not alone in the world before one is entitled to take

378 Anita Avramides one’s experience of another’s behaviour at face value. Just as we can take our experience of the world at face value once we lift the veil of perception, so we can take our perception of others at face value once we lift the veil of behaviour. Behaviour that operates like a veil is behaviour that is characterised in the same way whether or not the person is genuinely minded. Behaviour here is thought of on the model of sense data: it is the highest common factor as between the case of a body that is minded (or the body of a person) and one that is not. Thought of in this way, behaviour serves as a veil that cuts one person off from another. All that I can experience is your behaviour, neutrally conceived. Behaviour conceived of in this way leaves the way open for the question, how I do know that others are not zombies. The veil that cuts us off from the world of bodies is composed of sense data, while the veil that cuts us off from the mind of another is composed of behaviour conceived as divorceable from the mind of the other. Once either veil is in place, the knowledge we come by will be insecure, that is, subject to sceptical attack. The key to resisting both scepticisms is to lift both veils. One veil is lifted when we accept that our perceptual capacities are such as to be capable of yielding (indefeasible, secure) knowledge of the world of bodies. In the place of a conception of perceptual experience that places a veil of sense data in the way of such knowledge, McDowell proposes a disjunctive conception of experience. In order for us to achieve sceptic proof knowledge of another mind, yet another veil needs to be lifted, the behavioural veil. The placing of both veils is the legacy of a Cartesian conception of mind that takes the mind to be an autonomous region entirely transparent to its subject.32 McDowell identifies this conception and these veils in many standard views both of perception and of action. Once one appreciates that it is a rejection of a Cartesian conception of mind that drives McDowell to a disjunctive account of experience of the world, it follows quite naturally that he will also put forward an anti-Cartesian account of action or behaviour. To understand McDowell’s proposal for lifting this second veil, one needs to understand not so much his disjunctive account of experience as his understanding of the notion of criteria. It is easy to run these two accounts together as McDowell employs a disjunctive move both when proposing his conception of perceptual experience and when proposing his conception of behaviour. I turn now to McDowell’s understanding of the notion of a criterion. (b) Lifting the Behavioural Veil It is when urging a non-Cartesian conception of the relationship between mind and behaviour that McDowell appeals to the Wittgensteinian notion of a criterion. Criteria can be thought of as a kind of evidence, and in the case of another’s mental life, the evidence is their behaviour.

Disjunctivism and Other Minds

379

We might think of the behaviour of a person as evidence of her mental state in the same way as we think of a symptom (say, spots) as evidence of a disease (say, measles). Wittgenstein has suggested that we should not think of behaviour in this way; rather, we should think of behaviour as a criterion of a person’s mental state. What is the difference here, and how are we to understand the criterial relation? McDowell develops his understanding of the criterial relation by playing it off a rival understanding which he takes to be held by, inter alia, Crispin Wright. Wright characterises a criterion in such a way as to allow that a criterion can yield support for a claim that is defeasible; behaviour may serve as evidence that counts as a criterion of, say, a person’s pain even when it turns out that knowledge was not gained by appeal to it. McDowell thinks that Wright can say this because he holds that even on the occasions that seem most favourable for a claim to be able to see that someone else is in some ‘inner’ state, the reach of one’s experience falls short of that circumstance itself—not just in the sense that the person’s being in the ‘inner’ state is not itself embraced within the scope of one’s consciousness, but in the sense that what is available to one’s experience is something compatible with the person not being in the ‘inner’ state at all. (McDowell 1982: 457) In other words, McDowell finds in Wright’s work an understanding of criteria that leaves in place a veil that prevents a view from the outside in: I might experience the criteria and yet not come by knowledge your mental state.33 On all this, McDowell disagrees with Wright. As McDowell understands a criterion, if a person’s behaviour is genuinely a criterion of pain, then the connection here is indefeasible. He can say this because he holds that where behaviour serves as a criterion of another’s pain, the reach of experience does not fall short of the circumstance itself. This is why what is available to experience is not compatible with the person not being in that ‘inner’ mental state. But how can this be? How can your mental states be embraced within the scope of what I experience? McDowell has already explained how my experience can embrace a behaving body. But this is only an explanation of the mind’s interface with the world of bodies. He is now writing of a mind’s experience of something more, of another’s mental state. For experience to extend this far, we must also understand how an experience of a body can amount, in some cases at least, to something more.34 McDowell thinks experience of a body can amount to something more when the behaviour experienced is understood disjunctively: either behaviour may be genuinely expressive of another’s mental states, or it may be such as only to appear to be so.35 In what we might call the ‘good’ case, what

380 Anita Avramides we have is a criterion of, say, a person’s pain. Behaviour that serves as a criterion of a mental state is such as to be genuinely involving of that mental state, otherwise it is not a criterion but behaviour that only gives the appearance of that state. Thinking of behavioural criteria as McDowell does allows for a lifting of the veil that prevents a view ‘from the outside in’. This conception of behaviour is such as to allow that experience of it can yield access to other minds that is such as to put one in a position to have knowledge of them. What is experienced when one perceives the behaviour of another must not be compatible with the absence of the particular mental state experienced. The absence of the mental state would give rise in me to a quite different experience—and it would be a different experience because it would not be of behaviour which was serving as a criterion of the other’s mental state. Experience of such behaviour would not yield knowledge of another mind. Removing the veil of behaviour requires that we accept this McDowellian interpretation of criteria. Criteria must not be thought of as leaving room for knowledge which is insecure and so open to sceptical attack. In the good case, the criteria are satisfied, and experience of the criteria are such as to yield knowledge of the other’s mental life. Of course, our experience may not be of criteria, but only of bodily movement in which some particular mental state does not figure. As with perceptual experience itself, we are not in a position to know when we are encountering a good case of behaviour when we encounter others. Nonetheless, we can say that if it is a good case then what we experience does in fact yield knowledge—knowledge that is secure from sceptical attack. Thus we see that knowledge of other minds is the upshot both of a disjunctive account of experience and of a criterial account of the relationship between a mental state and behaviour.36 According to McDowell, it is only a particular understanding of criteria that is able to yield knowledge of another’s mental states that is secure from sceptical attack. That is, it can yield secure knowledge here when it is paired up with a disjunctive account of perceptual experience. Furthermore, given the dubious coherence of maintaining an anti-Cartesian conception of perceptual experience while at the same time accepting with a Cartesian conception of behaviour, such an understanding of the criterial account of behaviour in connection with mental states would seem required once one accepts a disjunctive account of experience.37 If one accepts both a disjunctive account of experience and a McDowellian understanding of a criterial account of behaviour (as, I have argued, one should if one’s acceptance of either is driven by a rejection of a Cartesian conception of mind), then one’s experience can reach out not just to the world, but also to the mind of another. A McDowellian understanding of a criterial account of behaviour ensures that what one experiences when one encounters another is not just a bodily state but is a state of the

Disjunctivism and Other Minds

381

other that can give one indefeasible warrant for the belief that the other is minded in certain ways. In other words, one’s experience is such as to allow for knowledge of another’s mental life. The result of employing the disjunctive move twice over—once in connection with our experience of the world of bodies and again in connection with another’s behaviour— is that we can now hold that when we experience others like ourselves, we do not experience them as mere moving bodies hidden behind which are other minds. It is the Cartesian sceptic who is committed to holding that what I encounter when I encounter another is a body about whom I hold ‘in suspense all attributions of psychological properties’ (McDowell 1982: 469). Contrary to this, McDowell claims to have given us a way of thinking about our experience of another and her behaviour which is not such as to leave us in suspense concerning her mental life. McDowell takes his response to the sceptic to be ‘to restore the concept of a human being to its proper place . . . a seamless whole of whose unity we ought not to have allowed ourselves to lose sight in the first place’ (McDowell 1982: 470).38 The employment of the disjunctive move gives us a very different way of conceiving of our minds in relation both to the world and to other minds. In my account of the application of the disjunctive move to behaviour, I helped myself to the Russellian idea that a person’s mental states ‘figure in’ their behaviour. When McDowell writes about the disjunct that allows for indefeasible warrant with regards another’s mental life, he tells us that the other’s mental life is ‘manifest’ to one, and that ‘the object of [one’s] experience does not fall short of the fact’. If this is indeed the way in which mental states figure in our experience, then one can conclude that, not only is our warrant indefeasible, but our knowledge here may be taken to be perceptual. We find McDowell writing in one place: we should not jib at, or interpret away, the common-sense thought that, on those occasions that are paradigmatically suited for training in the assertoric use of the relevant part of language, one can literally perceive, in another person’s facial expression or his behaviour, that he is in pain, and not just infer that he is in pain from what one perceives. (McDowell 1978: 305)39 Thus, one finds in McDowell’s work a defence of a perceptual account of our knowledge of other minds.40 Not only has McDowell shown that we can have knowledge of other minds, but he has also suggested an account of the source of this knowledge. It must be remembered, however, that what ultimately matters to McDowell’s anti-Cartesianism in connection with other minds is that the obtaining of the other’s mental life is ‘precisely not blankly external to [one’s] subjectivity’ (McDowell 1982: 476; cf. note 31 in this chapter).

382 Anita Avramides And McDowell considers at one point whether this requires that what is directly available to experience is the other’s mental state itself. He allows that what one encounters in experience is, at least in some cases, someone giving expression to an ‘inner’ state, which expression need not be understood as ‘itself actually being that “inner” state’ (McDowell 1982: 473). It may be thought that to count as a perceptual account of our knowledge of other minds, it is required that the other’s mental state is directly available in experience. By allowing for the possibility that a wedge may be driven between what is required for indefeasible knowledge and what is required for a perceptual account of our knowledge of other minds, McDowell here opens up the possibility that some sort of inferential account of our knowledge of other minds is possible.41 In any case, what is needed here is an argument for taking our experience in one way rather than another. Furthermore, and crucially, it must be shown that one can give up on the idea that one experiences another’s mental states directly while still maintaining that what one does experience is sufficient for indefeasible knowledge of another’s mind. Whether or not what one experiences suffices for us to say that we perceive the thoughts and feelings of others, it had better suffice for us to know others as persons.

5. Conclusion In this chapter I have aimed to show that a rejection of a Cartesian conception of mind lies at the heart of both McDowell’s understanding of perceptual experience and his account of behaviour in relation to mind. Descartes’s mind is designed to allow it to disconnect from its world and from others—albeit temporarily. But a conception that allows for even a temporary disconnect is one that is forever vulnerable to sceptical attack and hence one that makes our knowledge of the world and of others insecure. What drives McDowell’s rejection of the Cartesian conception of mind is its vulnerability to scepticism. McDowell believes that if we start with what we commonsensically take ourselves to know about the world and others, we can ‘ignore’ scepticism. We can take ourselves to be individuals capable of knowledge both of the world and of other minds. Our knowledge here is secure, despite our perceptual capacities being fallible. It is the fallibility of these capacities that gives rise to everyday error—I can be wrong about what I take myself to be experiencing. If we insist on understanding the error non-disjunctively—as the Cartesian does—then we will think that everyday error can be seen as a precursor to radical error. By rejecting the Cartesian conception of mind, McDowell aims to halt the slide into this radical, Cartesian, scepticism. As McDowell understand things, there can be no slide from the possibility of error of the everyday sort to error of this more radical sort. Our relations to the world in which we live and the others with whom we live can be

Disjunctivism and Other Minds

383

understood to be secure, and this security is the result of how we experience the world and others.42

Notes 1. Others whose work is closely associated with disjunctivism include J.M. Hinton (1967), P. Snowdon (2005) and M.G.F. Martin (2006). 2. It should be noted that, while Martin is interested in issues to do with perception, he also has a wider concern—naïve realism. For some thoughts on how one might think of the relationship between disjunctivism and naïve realism, see Snowdon (2005). 3. And also, in part, by the work of Fred Dretske (1969). Perception in connection with other minds was also championed by Nathalie Duddington (1918) in the earlier part of the 20th century. In recent times a wealth of literature has been devoted to this topic (see, for example, papers in Avramides and Parrott 2019). While the work of Dretske and McDowell is rooted in the analytic tradition in philosophy, the phenomenological tradition has also embraced perception in connection with other minds and for a somewhat longer period (cf Scheler 1954/2013). It is important to appreciate that different concerns drive different perceptual accounts of other minds. Some differences in concern between the perceptual accounts put forward by McDowell and Dretske will emerge in later notes to this chapter. There is also a question whether McDowell’s work on disjunctivism commits him to a perceptual account of our knowledge of other minds. I touch on this in section (iii)(b). 4. There may be arguments that may lead one to reject the extension here, but prima facie the extension is a natural one (in this connection see note 37 in this chapter). Unless indicated otherwise, all reference in this paper will be to McDowellian disjunctivism, by which I shall mean an experiential disjunctivism adopted for the reasons that inform McDowell’s work. 5. In what follows I outline the Russellian element in McDowell’s work and leave to one side the Fregean element. The latter is important to an overall understanding here, but the former is more directly related to the topic of this chapter, disjunctivism and other minds—or at least the aspect of it I am concerned with discussing here. 6. Here we find McDowell appealing to a move he had already discussed in McDowell (1982; 1978) in connection with, among other things, other minds. In McDowell (2010), McDowell notes that in McDowell (1986) he ‘does not use a disjunctive conception of singular thought to argue for a disjunctive conception of experience, but rather uses a disjunctive conception of experience to bring out features of singular thought’ (252). My interest here is not so much in McDowell’s disjunctive conception of singular thought as with the link between his disjunctive account of experience and reflections on one’s conception of mind, which link is made clear in McDowell’s discussion of Russell’s work. Towards the end of his 2010 paper, McDowell notes that it is when one takes into account considerations of singular thought that one can appreciate the wider significance of his account of perceptual experience—wider, that is, than the usually appreciated epistemological significance. 7. McDowell writes that in Russell’s work ‘acquaintance shapes itself . . . into a rejection of fallibility’ (1986: 236). 8. For McDowell’s discussion of these Cartesian ideas see McDowell (1986: sec. 6). This is a complex and intricate discussion, which I merely introduce quickly in what follows. One point to note: McDowell believes a subject can

384 Anita Avramides

9. 10.

11.

12.

13.

14.

15. 16.

17.

18.

19.

know about her own mental states in a way that is ‘special’ without having to buy into the Cartesian story about an autonomous inner realm. Taking away the world may be thought to give rise to two, related, problems: one is epistemological; the other is conceptual. In connection with the latter in relation to McDowellian disjunctivism, see Gomes (2011). Here McDowell suggests that we should remember the ancient sceptic option that it is not by knowing that we have our hold on the world (cf. McDowell 1986: sec. 5). For his interpretation of the ancient sceptics, McDowell refers us to Burnyeat (1982). In McDowell (2010), McDowell puts the point this way: ‘while we must accept that appearances are not open to question, Descartes extends the point to judgements about how things are to one and then adds that we can take this as constituting the whole truth about this inner state of affairs’ (244). McDowell 1986: 246) writes: ‘it is worth wondering whether it is the insistence on autonomy that is the real disease of thought, with the superficially striking peculiarities of Descartes’ own picture of mind no more than a symptom that something is amiss’. Note that when the sceptic takes away our world, she takes away access that allows us both to think about things as well as to know about them. It may be clearer how the disjunctive move brings about the former than it is how it brings about the latter. In connection the latter McDowell writes, ‘When someone has a fact made manifest to him, the obtaining of the fact contributes to his epistemic standing on the question’ (1986: 390–391). McDowell thus believes that there is an epistemic payoff from the metaphysical move. In McDowell (1994), he writes: ‘The aim here is not to answer the sceptical questions, but to begin to see how it might be intellectually respectable to ignore them, to treat them as unreal, in the way that common sense has always wanted to’ (113). Cf. also McDowell (1995: fn. 19). I return to this idea in section (iii)(a). The footnote occurs in McDowell (1982: 472). McDowell also acknowledges this work in his 2008 paper. Cf. Snowdon (1980–81). For a later clarification of his disjunctivism in response to W. Fish, see Snowdon (2005). Here Snowdon makes it clear that (his) disjunctivism is meant to be a claim about the nature of perceptual experience. In the case of Snowdon, one could argue that his concerns in introducing disjunctivism were rather minimal: he wanted simply to show that Grice’s causal account of perception is not required. For an account of Hinton’s position and how it stands relative to some of the other literature on this topic, see Snowdon (2008). In recent years philosophers have teased out different versions or varieties of disjunctivism. Adrian Haddock and Fiona Macpherson (2008), following in the footsteps of Alex Byrne and Heather Logue (2008), have suggested that we can separate out an epistemological version of disjunctivism from a metaphysical one. Byrne and Logue further suggest that while one can embrace metaphysical disjunctivism without embracing epistemological disjunctivism, the former ‘leads naturally if not inexorably’ to the latter (2008: 67). While it may be true of disjunctivism generally that one can embrace the epistemological version without embracing the metaphysical version, I take McDowellian disjunctivism to both metaphysical and epistemological in its scope (cf. Gomes 2011). This is due to the particular kind of epistemological work that McDowell thinks needs doing. See also note 20 in this chapter. I say more about this in section (iii)(a). It is worth emphasising that McDowell is not just looking to see how knowledge is possible; he is looking for

Disjunctivism and Other Minds

20.

21. 22.

23. 24. 25. 26. 27.

28. 29. 30. 31.

32. 33.

385

what I am calling ‘epistemological security’. This explains why McDowell is not enticed by other ways of arguing for knowledge here. It is only if one sees how one’s epistemology is bound up with one’s metaphysics that one will be able to ‘ignore’ the sceptic. It is interesting to consider that a reliabilist like Dretske also aims to ‘ignore’ the sceptic. However, the ways in which McDowell and Dretske ‘ignore’ scepticism is importantly different: while Dretske can only claim to turn a blind eye to scepticism, McDowell wants an account that allows no room for it. This is not to say that there might not be good arguments for accepting the latter division. My point here is that, in the absence of such an argument, accepting such a division looks odd (or as I said at the start of the chapter, ‘dubiously coherent’). See note 37 in this chapter. The passage being paraphrased is in McDowell (1986: 239). There are real questions about whether the way in which we can be deceived by the world does exactly parallel the way in which can be deceived by other minded individuals, but for the purposes of this chapter I am confining myself to explaining how I believe McDowell understands the two cases. Davidson is another philosopher who is concerned to undermine a Cartesian conception of mind. Rather than refer to the notion of a criterion, Haddock and Macpherson refer to this aspect of McDowell’s work as disjunctivism in the philosophy of action, or epistemological disjunctivism about bodily movement (2008: 18ff). See his response to Crispin Wright in McDowell (1982). Cf. also McDowell (2010: sec. 9). Which is not say that McDowell is proposing an argument against scepticism; cf. note 19 in this chapter. In another place McDowell adds: ‘if one’s experience makes an object present to one . . . one is thereby in a position to know that is how it is with one, and thus to recognize that one is conclusively warranted in believing there is something . . . in front of one’ (2010: sec. 9). Cf. McDowell (2010: sec. 9). The quotation continues: ‘which would not be the case if we were not perceptually in touch with the world in just about the way we ordinarily suppose we are’. Note the Moorean flavour of McDowell’s argument here. As far as I know, McDowell doesn’t explicitly put the argument in this way, but it may be thought to be implicit in what he writes. The sort of example usually given here includes, ‘I can just see that he is angry, or that she is in pain’, and the like. McDowell writes that ‘one can sometimes see what someone’s mental state is by (as we say) looking into her eyes’. McDowell is careful here to say that we don’t say such things in connection with all mental states and occurrences. By drawing attention to such locutions, McDowell is clear that he is wanting to avoid the idea ‘that the mental is withdrawn from direct engagement with the world’ (1995: 413). Whether avoiding this latter idea requires a perceptual account of our knowledge of other minds is something I touch on further on in the chapter. Cf. in this connection the discussion at the end of section (ii), of the quotation from Davidson where he writes of the logical and epistemic barriers which block our view both ‘from the inside out’ and ‘from the outside in’. One must be careful here. As I read his work, McDowell is not just interested in accounting for knowledge of another’s mental life. As in the case of knowledge of bodies, he wants an account that is secure from sceptical attack. This requires that the way in which we account for knowledge here will depend upon a metaphysical claim about the interdependence of mental states and behaviour, as I am about to explain. For the alternative understanding of criteria, see Wright (1982).

386 Anita Avramides 34. For this qualification (‘at least in some cases’), see McDowell (1982: 473, fn. 1). See also note 39 in this chapter. 35. In McDowell (2010), he writes that he has ‘recommended a conception of perceptual experience that is well expressed by a disjunction’ (244). He could equally have written that he also recommends a conception of behaviour that is well expressed by a disjunction. 36. One could also say: knowledge here is the upshot combining a disjunctive move in connection with perceptual experience and a disjunctive move in connection with behaviour. 37. That such an account of behaviour is ‘required’ may be thought to be too strong a statement. Perhaps it is. The thought is that it is required if one is to avoid the dubious coherence of maintaining an anti-Cartesian conception of perceptual experience while at the same time accepting with a Cartesian conception of behaviour. At the very least one could say that, once one accepts the disjunctive move in connection with perceptual experience of bodies, it requires an argument why would not accept the same move in connection with other minds, on pain of having to accept the sceptical possibility that others are zombies. 38. In this McDowell takes himself to be following Wittgenstein, and to be in line with some things that P.F. Strawson has written in connection with the idea of the concept of a person as ‘primitive’ (see Strawson 1959). 39. Here McDowell is mainly concerned to reply to Michael’s Dummett’s arguments against realism to the effect that if realism is the case, then one could not account for language learning; hence, the talk here about training in the relevant part of language. 40. Note that McDowell’s account is very different, because of its different epistemological concerns, from the perceptual account of other minds that one finds in Dretske’s work. 41. This is despite the fact that in much that he writes, McDowell proposes that his work entails a rejection of inferential accounts. Another possibility would be to continue to reject an inferential account and allow that one’s account of knowledge counts as perceptual just so long as one’s knowledge comes about as the result of one’s perceptual experience, whether or not that experience includes direct observation of another’s mental state. 42. I would like to thank Casey Doyle for much helpful discussion of this chapter. I would also like to thank Mike Martin, Anil Gomes, Nick Shea, and the other members of the Oxford Mind-Work-in-Progress Group for helping me to get clearer about disjunctivism and its commitments.

References Avramides, Anita and Parrott, Matthew, eds., 2019, Knowing and Understanding Others. Oxford: Oxford University Press. Blackburn, Simon, 1994, The Oxford Dictionary of Philosophy. Oxford: Oxford University Press. Burnyeat, Myles Fredric, 1982, Idealism and Greek philosophy: What Descartes saw and Berkeley missed, The Philosophical Review 21(1): 3–40. Byrne, Alex and Heather, Logue, 2008, Either/or, in Adrian Haddock and Fiona Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 57–95. Davidson, Donald, 1991, Three varieties of knowledge, in A. Phillips Griffiths, ed. A.J. Ayer: Memorial Essays. Cambridge: Cambridge University Press, pp. 153–167.

Disjunctivism and Other Minds

387

Dretske, Fred, 1969, Seeing and Knowing. London: Routledge & Kegan Paul. Duddington, Nathalie, 1918, Our knowledge of other minds, Proceedings of the Aristotelian Society 19: 147–178. Gomes, Anil, 2011, McDowell’s disjunctivism and other minds, Inquiry 54(3): 277–292. Haddock, Adrian and Macpherson, Fiona, 2008, Introduction: Varieties of disjunctivism, in Adrian Haddock and Fiona Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 1–25. Hinton, John Michael, 1967, Visual experiences, Mind 76: 212–227. Martin, Michael., 2006, On being alienated, in Tamar Szabo Gendler and John Hawthorne, eds. Perceptual Experience. Oxford: Oxford University Press, pp. 354–411. McDowell, John, 1978 [1998], On ‘on the reality of the past’, in his Meaning, Knowledge, & Reality. Cambridge, MA: Harvard University Press, pp. 295–314. ———, 1982, Criteria, defeasibility and knowledge, Proceedings of the British Academy LXVIII: 455–479. ———, 1986 [1998], Singular thought and the extent of inner space, in his Meaning, Knowledge, & Reality. Cambridge, MA: Harvard University Press, pp. 228–260. ———, 1994, Mind and World. Cambridge, MA: Harvard University Press. ———, 1995 [1998], Knowledge and the internal, in his Meaning, Knowledge, & Reality. Cambridge, MA: Harvard University Press, pp. 395–414. ———, 2008, The disjunctive conception of experience as material for a transcendental argument, in Adrian Haddock and Fiona Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 376–390. ———, 2010, Burge on disjunctivism, Philosophical Explorations 13(3): 243–255. Scheler, Max, 1954 [1913], The Nature of Sympathy. Trans. Peter Heath. London: Routledge and Kegan Paul. Snowdon, Paul, 1980–1981, Perception, vision, and causation, Proceeding of the Aristotelian Society Supplementary Volume 64: 121–150. ———, 2005, The foundations of disjunctivism: Reply to Fish, Proceedings of the Aristotelian Society 105(1): 129–141. ———, 2008, Hinton and the origins of disjunctivism, in Adrian Haddock and Fiona Macpherson, eds. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press, pp. 35–57. Strawson, Peter, 1959, Individuals. London: Methuen. Wright, Crispin, 1982, Anti-realist semantics: the role of criteria, Royal Institute of Philosophy Lectures, 13: 225–248.

Contributors

Anita Avramides is the Southover Manor Trust Fellow in Philosophy at St Hilda’s College and Reader in Philosophy of Mind at the University of Oxford. She is the author of Other Minds (Routledge) and Meaning and Mind: An Examination of a Gricean Account of Meaning (MIT Press) and has published articles in the philosophy of language and philosophy of mind. Most recently she has concentrated her work on questions relating to knowing and understanding others. Dorit Bar-On is a Professor of Philosophy at the University of Connecticut and the director of the Expression, Communication, and Origins of Meaning interdisciplinary research group. She specializes in philosophy of language, philosophy of mind, and epistemology. She also has research interests in metaethics. In her book Speaking My Mind: Expression and Self-Knowledge (Oxford, 2004), she develops a ‘neo-expressivist’ view of so-called first-person authority, drawing on insights from philosophy of language, mind, theory of action, and epistemology. She has published articles in the Journal of Philosophy, Mind and Language, Philosophy and Phenomenological Research, Nous, Synthese, Phil Studies, Inquiry, Philosophical Explorations (among other journals), and many volumes and special issues—on Quine, Davidson, Dummett, Grice, meaning and interpretation, semantic knowledge, conceptual relativism, deflationism, self-knowledge, introspection, and ethical expressivism. In recent years, she has been working on the topic of continuities between linguistic and non-linguistic communication and expressive behavior. She is currently working on a book manuscript titled Expression, Communication, and the Origins of Meaning, under contract with Oxford University Press. (She has another manuscript in progress with Keith Simmons, If Truth Be Told.) Casey Doyle is a Junior Research Fellow in Philosophy at St Hilda’s College, University of Oxford. His major research interests are in philosophy of mind and epistemology, especially self-knowledge.

Contributors

389

Craig French is an Assistant Professor of Philosophy at the University of Nottingham. His work is in the philosophy of mind and psychology, with a particular focus on the philosophy of perception. Adrian Haddock is a Senior Lecturer in Philosophy at the University of Stirling. He works mainly on the idea of self-consciousness and its significance for understanding theoretical and practical thought. From 2017 to 2019 he is an Alexander von Humboldt Fellow at Universität Leipzig. Drew Johnson is a PhD student in philosophy at the University of Connecticut and has MAs in philosophy from the University of Connecticut and Northern Illinois University. He has research interests in metaethics, epistemology, truth, and animal ethics, and is a member of the Expression, Communication, and Origins of Meaning research group at the University of Connecticut. In his dissertation, he develops a novel form of hybrid ethical expressivism. His “Hinge Epistemology, Radical Skepticism, and Domain Specific Skepticism” is forthcoming in The International Journal for the Study of Skepticism. Clayton Littlejohn is a Professor of Philosophy at King’s College London and the author of Justification and the Truth Connection (Cambridge). Thomas Lockhart is an Associate Professor of Philosophy at Auburn University. He is interested in early analytic philosophy, especially Frege, and epistemology, particularly the philosophy of perception. Guy Longworth is a Reader in Philosophy at the University of Warwick. He works mainly in the philosophy of mind and its history, construed broadly so as to incorporate epistemology, the philosophy of language, and their histories. John McDowell is a University Professor of Philosophy at the University of Pittsburgh. Before coming to Pittsburgh in 1986, he taught at University College, Oxford. He has held visiting appointments at Harvard University, the University of Michigan, UCLA, and Princeton University. He was the John Locke Lecturer at Oxford University in 1991. His major interests are Greek philosophy, philosophy of language, philosophy of mind, metaphysics and epistemology, and ethics. He is a fellow of the British Academy and a fellow of the American Academy of Arts and Sciences. Joe Milburn is a Visiting Professor at the University of Navarra in Pamplona. His research centers on epistemology and philosophy of religion. His recent publications deal with epistemic luck and with the philosophy of John Henry Newman. Veli Mitova is an Professor and Head of Philosophy at the University of Johannesburg. She is also co-founder of the African Centre for

390

Contributors

Epistemology and Philosophy of Science, and the South African team leader for the Geography of Philosophy Project. Veli obtained her PhD from Cambridge, and then taught and researched at Rhodes University, Universidad Nacional Autonoma de México, and Universitaet Wien. She is the author of Believable Evidence (Cambridge, 2017) and editor of The Factive Turn in Epistemology (Cambridge, 2018). She is currently working on epistemic injustice, under the auspices of a Newton Advanced Fellowship, Epistemic Injustice, Reasons, and Agency. Andrew Moon is an Assistant Professor of Philosophy at Virginia Commonwealth University. His area of specialty is epistemology, and he has interests in philosophy of mind and philosophy of religion. He has recently published on the natures of belief and confidence, religious epistemology, memory, the internalism/externalism debate, and evolutionary debunking arguments. He is also working on a book on the relationship between knowledge and the doxastic attitudes (belief, doubt, confidence, and certainty). Ram Neta is a Professor of Philosophy at the University of North Carolina, Chapel Hill. He is the author of dozens of articles, mostly in epistemology. His articles have appeared in The Philosophical Review, Nous, Philosophy and Phenomenological Research, the British Journal for the Philosophy of Science, and many other journals. Duncan Pritchard is the Chancellor’s Professor of Philosophy at the University of California, Irvine, and Professor of Philosophy at the University of Edinburgh. His monographs include Epistemic Luck (Oxford, 2005), The Nature and Value of Knowledge (co-authored, Oxford, 2010), Epistemological Disjunctivism (Oxford, 2012), and Epistemic Angst: Radical Skepticism and the Groundlessness of Our Believing (Princeton, 2015). Genia Schӧnbaumsfeld is a Professor of Philosophy at the University of Southampton. She is the author of The Illusion of Doubt (Oxford, 2016), A Confusion of the Spheres: Kierkegaard and Wittgenstein on Philosophy and Religion (Oxford, 2007), and Transzendentale Argumentation und Skeptizismus (Peter Lang, 2000). She is a Visiting Professor at the University of Regensburg and associate editor of Philosophical Investigations. Currently, she is exploring the connections between radical scepticism and epistemic vice. Sonia Sedivy is an Associate Professor in the Department of Philosophy, University of Toronto Scarborough and the Graduate Department of Philosophy, University of Toronto. Her research focuses on perception, aesthetics, and the later work of Wittgenstein. Her recent book, Beauty and the End of Art: Wittgenstein, Plurality and Perception, shows how all three areas can work together to offer new approaches to art and

Contributors

391

beauty. Her paper “Aesthetic Properties, History and Perception” argues that philosophical theories of perception and aesthetics mutually constrain one another. Currently, she is focusing on developing a conceptual realist account of perception that integrates considerations from aesthetics. Iakovos Vasiliou is a Professor of Philosophy at the Graduate Center of the City University of New York. His research focuses on ethics, moral psychology, philosophy of action, and philosophy of mind in Plato and Aristotle. He also works on contemporary neo-Aristotelian ethics. He is the editor of Moral Motivation (Oxford, 2016) and author of Aiming at Virtue in Plato (Cambridge, 2008). Stephen Wright is a Lecturer in Philosophy at Trinity College, Oxford. He is the author of Knowledge Transmission (Routledge) and, with Sanford Goldberg, a co-editor of the forthcoming Memory and Testimony: New Essays in Epistemology (Oxford). His other research interests include aesthetics, metaphysics, moral philosophy, and the philosophy of religion.

Author Index

Adler, J. 314 Alston, W. 177, 178, 183 Altschul, R. 339 Alvarez, M. 210, 211, 212 Annas, J. 83 Anscombe, G.E.M. 22, 262, 270–272 Asmis, E. 82 Austin, J. L. 14, 17–18, 131–150 Avramides, A. 12, 13, 14, 21, 25, 383, 386 Axtell, G. 54 Barnett, D. 314 Bar-On, D. 12, 21, 23–24, 325, 329–331, 339, 340, 341, 342, 347, 348, 352, 354, 362, 363 Bauer, N. 110 Bayne, T. 362 Bennett, J. 125 Bernecker, S. 289, 295 Bilgrami, A. 363 Blackburn, S. 374–375 Boghossian, P. 124, 128 Bowman, B. 110 Boyle, M. 326, 340, 347, 355, 362, 363 Brennan, T. 83 Brewer, B. 154, 167, 176, 179, 180, 191, 236, 238 Brittain, C. 85 Broadbent, A. 212 Broome, J. 210, 211 Brown, J. 297 Brunschwig, J. 82 Burge, T. 12, 23, 24, 25, 26, 54, 73, 112, 125, 128, 167, 168, 172, 176, 190, 191, 302–308, 314, 318, 319–322, 326, 328, 337, 338, 339, 340, 342, 343, 357, 363, 387 Burnyeat, M. 82, 384 Byrne, A. 10, 11, 12, 25, 110, 111, 196, 210, 213, 235, 238, 336,

338, 339, 343, 344, 347, 363, 384, 386 Campbell, J. 1, 154, 176, 191 Cassam, Q. 173, 174, 192, 339, 343, 347, 351, 362, 363 Castañeda, H. 277 Caston, V. 82 Code, L. 54 Coliva, A. 114, 124, 125, 126, 127, 128, 129, 339, 342, 343, 363 Conee, E. 125, 211, 219, 235 Cook Wilson, J. 14, 149, 362, 363 Crane, T. 1, 25, 172 Cunningham, J. J. 210, 213, 236, 238, 338, 343 Currie, G. 350 Dancy, J. 210, 211, 212 Davidson, D. 25, 61, 210, 373, 385 Descartes, R. 21, 61, 82, 126, 139, 218, 260–262, 266, 369–371, 376, 382, 384 Diamond, C. 125 Doyle, C. 21, 24–25, 341 Doyle, R. 277, 278 Drake, J. 212 Dretske, F. 124, 127, 128, 173, 174, 176, 192, 325, 339, 343, 383, 385, 386, 387 Duddington, N. 383, 387 Dummett, M. 124, 128, 314, 386 Dutant, J. 236 Evans, G. 272, 331, 340 Everson, S. 82 Falvey, K. 362 Fine, G. 82 Fish, W. 12, 26, 154, 167, 227, 230, 236, 238, 384, 387

Author Index Frede, M. 84 French, C. 18–19, 169, 174, 190, 191, 192, 196, 210, 211, 213, 236, 238, 239, 289, 290, 296, 297, 362, 364 Fricker, E. 302–304, 306–308, 338 Funkhouser, E. 350, 362 Gendler, T. 349–350 Gibbons, J. 212, 346 Ginsborg, H. 19, 177–189, 191, 192, 236, 239, 314, 315 Goldman, A. 211, 309 Gomes, A. 384 Greco, J. 52, 54, 82, 277 Grice, H. P. 370, 384 Haddock, A. 10, 11, 12, 21–22, 25, 26, 52, 53, 55, 56, 82, 83, 87, 130, 150, 169, 191, 192, 193, 210, 212, 213, 214, 238, 257, 343, 360, 361, 364, 384, 385, 386, 387 Hampshire, S. 352–354 Hankinson, R. 83, 84 Hanna, R. 110 Heal, J. 363 Hetherington, S. 52 Hieronymi, P. 210, 211 Hinton, J. M. 1, 14, 131, 134, 144–148, 314, 370, 371, 383, 384 Holton, R. 362 Hornsby, J. 212 Huemer, M. 207, 210 Hughes, N. 236 Johnson, D. 12, 21, 23–24, 347, 362, 363 Johnston, M. 186–188, 211, 257 Kalderon, M. 14, 26, 136, 149, 191, 192 Kant, I. 14, 16, 18, 89–112, 156, 167, 239, 277, 278 Kern, A. 24, 277, 278, 359, 360, 362, 363, 364 Kiesewetter, B. 237, 238 Kremer, M. 157 Lackey, J. 301, 304, 306, 310–311, 314 Land, T. 109, 110 Lee, M. 82 Lehrer, K. 53 Littlejohn, C. 20, 53, 55, 212, 213, 235, 236, 238, 239 Locke, D. 236

393

Lockhart, T. 12, 16, 18 Logue, H. 10, 11, 12, 25, 176, 191, 192, 196, 210, 213, 235, 236, 238, 239, 336, 338, 343, 344, 384, 386 Longworth, G. 17–18, 132, 135, 149, 236, 239 Lord, E. 237, 238 Luper, S. 52, 236 MacDonald, C. 338 Macpherson, F. 10, 11, 12, 25, 26, 55, 82, 83, 87, 130, 169, 191, 192, 210, 212, 213, 214, 238, 257, 343, 364, 384, 385, 386, 387 Maddy, P. 125 Madison, B. 210, 339 Malmgren, A. 314 Mantel, S. 212 Marcus, E. 347 Martin, M. G. F. 1, 10, 12, 26, 131, 132, 135, 142, 144, 148, 149, 154, 167, 176, 192, 314, 315, 337, 343, 363, 364, 383, 387 Matthews, G. 82 McDowell, J. 1, 2, 4–16, 18, 21, 23, 24, 25, 26, 53, 55, 61, 62, 63, 64, 72–82, 84, 85, 86, 87, 89–94, 97–109, 110, 111, 112, 125, 129, 130, 131, 150, 153–167, 167, 175–176, 190, 192, 194–197, 199, 210, 211, 212, 213, 216–228, 234–235, 236, 237, 239, 241–256, 257, 259–261, 267, 269, 277, 278, 281, 297, 298–314, 315, 320, 324, 338, 339, 341, 341, 343, 345–346, 352, 359, 360, 362, 363, 364, 366–383, 384, 385, 386, 387 McGinn, C. 125 McGinn, M. 235, 236 McLear, C. 110 McMyler, B. 314 Milburn, J. 21, 22, 23, 237 Millar, A. 12, 26, 52, 53, 56, 190, 193, 277, 279 Mitova, V. 12, 19, 56, 210, 211, 212, 213, 214, 258 Montmarquet, J. 54 Moon, A. 21, 22, 23, 295, 296, 297 Moore, G.E. 17, 115–116, 119–120, 123, 125 Moran, R. 314, 316, 347, 355, 364 Nagel, T. 125 Nawar, T. 83 Nelkin, D. 257

394 Author Index Neta, R. 11, 12, 20–21, 26, 53, 86, 87, 196, 210, 211, 214, 244, 248, 257, 258, 338, 339, 342, 343, 362, 364 Nozick, R. 127 O’Shaughnessy, B. 352 Owens, D. 314 Pappas, G. 324, 339 Peacocke, C. 27, 339, 343, 344, 351, 364, 365 Pears, D. 137, 150 Perin, C. 83 Pitt, D. 347, 351 Pollock J. 125 Price, H. H. 140, 147, 150 Prichard, H. A. 14 Pritchard, D. H. 1, 2–4, 10–15, 17, 22, 23, 26, 27, 41, 43, 45, 52, 53, 54, 55, 56, 70, 82, 83, 84, 85, 86, 87, 109, 110, 112, 122, 123, 125, 127, 129, 154, 168, 169, 170, 171, 173, 190, 193, 195, 196, 210, 211, 212, 214, 235, 236, 237, 238, 239, 241–242, 244, 247–248, 250, 258, 281–291, 295, 296, 297, 298, 300, 310, 312, 316, 317, 320–324, 338, 339, 343, 362, 364 Pryor, J. 85, 125, 126, 340 Putnam, H. 132, 135, 150 Reed, B. 83, 84 Reisner, A. 211 Roberts, R. 54 Rödl, S. 16, 24, 77, 78, 85, 88, 277, 347, 359, 360, 362, 364 Roessler, J. 347, 352, 353 Ross, A. 314 Rumfitt, I. 277–278 Sainsbury, R. 52 Scanlon, T. 210, 211 Scheler, M. 383, 387 Schellenberg, S. 236, 241, 243–244, 248, 250, 257 Schmidt, E. 235 Schönbaumsfeld, G. 14, 16–17, 110, 112, 129 Schwarz, A. 295 Sedivy, S. 12, 13, 16, 18, 167, 168 Sedley, D. 65, 82 Sellars, W. 6, 31, 33, 39, 40, 61, 73, 101–102, 110, 267, 368 Shoemaker, S. 331, 339, 340, 347, 362, 363 Siegel, S. 236

Smith, M. 257 Snowdon, P. 1, 10, 11, 12, 14, 26, 131, 133, 134, 135, 136, 142, 150, 314, 316, 337, 344, 370, 371, 383, 384, 387 Sosa, E. 52, 53, 236 Soteriou, M. 1, 10, 26, 142, 148, 150, 176, 191, 193, 210, 214 Srinivasan, A. 238 Strawson, P. F. 153, 168, 301, 316, 386, 387 Striker, G. 82, 83 Stroud, B. 125 Sutton, J. 236 Sylvan, K. 191 Thomas, A. 125 Travis, C. 14, 24, 26, 27, 140, 148, 150, 154, 168, 172, 193, 236, 239, 240, 347, 348, 350–361, 362, 363, 365 Turri, J. 211, 236 Unger, P. 222 Van Cleve, J. 236 Vasiliou, I. 15–16, 24, 82, 88 Velleman, J.D. 350 Vogt, K. 82, 85 Wallace, R.J. 205 Walton, K. 164 Way, J. 231 Weiner, M. 257 Weir, A. 172 Weisberg, J. 257 White, R. 125 Whiting, D. 231 Wiggins, D. 157 Williams, M. 52, 114, 118, 122, 129 Williamson, T. 57, 174, 193, 212, 214, 223, 236, 240, 241–244, 250–252, 256, 257, 258, 277, 279, 297, 337, 344 Wittgenstein, L. 1, 14, 16, 17, 52, 54, 61, 113–130, 168, 259, 267, 340, 342, 344, 355, 364, 365, 374, 378, 379, 386 Wright, C. J. G. 57, 114, 122, 124, 125, 126, 127, 129, 197, 325, 339, 340, 342, 343, 344, 352, 364, 365, 379, 385, 387 Wright, S. 11, 13, 21, 23, 27, 197, 214, 314, 316 Zagzebski, L. 52, 54 Zimmerman, A. 24, 27, 339, 340, 344, 347, 349–352, 361, 362, 365

Subject Index

ability conditions on knowledge 42–45, 49, 53 appearance(s) 4–6, 8–9, 11, 13, 15–16, 25, 64–83, 113–114, 119–120, 123, 125, 133, 147, 158, 164, 166, 195–197, 199–201, 203, 205, 207–208, 210–211, 221, 250, 252, 255, 274, 281, 296, 333, 334, 341, 346, 348–349, 352, 357, 368–369, 380, 384 cataleptic appearances 15–16, 66, 69–76, 78–83, 85; mere appearances 4, 68, 81, 114, 195–196, 199 conceptual capacities 16, 101–104, 107–109, 153–157, 162, 164, 166, 227, 230, 236 consciousness 24, 63, 114, 225, 350–352, 362, 379; self-consciousness 24, 33–36, 263, 346, 359–360 content 18–19, 65, 70–71, 83–84, 90, 95, 100–103, 105–108, 110–111, 121, 137, 153–167, 169, 172, 175–176, 178, 181, 190, 211–212, 236, 265, 276, 300, 306, 310, 331–332, 340, 350–351, 362 discriminate/discriminability 4, 76–77, 145–149, 288 distinguishability 4, 24, 146–149, 211, 243, 285, 287–288, 291, 346, 352–353, 355–357, 359–362 (epistemic) externalism 3, 6–7, 9, 15, 38, 45–46, 48–51, 54, 62, 64, 71, 73, 162, 209, 243–244, 249, 284–286, 290, 302, 308, 326, 338–339 (epistemic) internalism 3, 6, 15–16, 33–36, 38, 45–51, 53–54, 62, 73–74,

210, 221, 244, 251, 257, 284–286, 290, 302, 308, 317, 324–326, 338–339, 359 (epistemic) luck 7, 15, 41–45, 47, 52, 53, 222–224, 228, 236 epistemic twin earth 43, 45, 52, 53 evidence 1, 20–21, 66, 73, 81, 84, 114, 116, 119, 125, 134, 216, 223, 225, 233, 235, 236, 242–245, 248–250, 252–257, 296, 297, 314, 317, 325, 328, 332–334, 347, 350, 352–355, 361, 362, 373–376, 378–379 “external” world 4, 7, 13, 17, 21–22, 25, 66, 82, 97, 98, 113–115, 117–118, 120, 122–127, 215, 219–220, 223, 369, 373 fallibility/infallibility 5, 7, 9, 15, 18, 24–25, 37, 65–66, 74–78, 81, 85, 98, 110, 147, 149, 159–160, 202, 220, 259–260, 268–269, 276, 277, 294, 324, 332, 347–348, 360–361, 363, 369–370, 372, 375, 376, 382–383 Gettier case 44–45, 202, 209, 289 hallucination 1, 2, 8, 10, 11–12, 17–18, 20, 26, 63, 78, 84, 131, 133, 140–145, 148, 154, 158, 170–171, 194–196, 209–211, 218, 235, 241, 243–244, 246–249, 255, 285, 287–288, 296, 314, 337–338, 345, 348, 357, 359 highest common factor 8–9, 13, 17, 25, 63, 125, 345, 378 illusion(s)/illusory experience 2, 8, 10, 15, 17, 20, 63–64, 82, 110–111,

396

Subject Index

113, 115–116, 120, 123, 127, 132, 139–140, 145, 154, 158, 179, 181, 187–188, 195, 207–209, 241, 247, 285–288, 318–319, 334, 337–338, 362, 368 indiscriminability see discriminate/ discriminability indistinguishability see distinguishability introspection 2, 13, 18, 24, 132, 137, 145–149, 170, 225, 233, 243, 293, 322, 327, 332, 338–339, 347, 349, 356–357, 362 justification 6–8, 20, 22, 24, 45–46, 51, 82, 102, 116, 118–119, 196, 198–200, 203, 206, 209, 211, 216, 218–219, 230, 232, 235, 237, 317, 323, 325, 326, 334, 337, 339, 345, 360 metaphysical disjunctivism 1–2, 10–14, 17–20, 23, 83, 154, 158, 194–197, 199–201, 204, 207, 209–211, 215, 218, 221, 223, 231, 234–235, 237, 295, 314, 318–320, 338, 384 new evil demon 53, 210 other minds 2, 13, 21, 25, 260, 281, 345, 366–367, 369–374, 380, 381, 382, 383, 385, 386 perceptual knowledge 1–5, 9–10, 13, 14–15, 17, 19–25, 31–33, 35–40, 41, 46–47, 49, 53, 62, 77, 79–80, 85, 89–90, 113, 115, 169–174, 212, 215–218, 221, 223–225, 227, 230– 231, 236–237, 281–285, 287–291, 294–295, 298, 300–301, 310, 317–318, 320–322, 324, 327, 336, 341, 345–346, 356–357, 359–360, 362 probability 242, 249–250; indiscriminability (see discriminate/ discriminability); indistinguishability (see distinguishability) rationality 1, 14, 16, 31–34, 38–39, 73, 90–96, 100, 102–104, 108–110,

124–125, 127, 156, 181, 196, 238, 248–250, 347, 357 rational support 2–3, 6, 11, 18–19, 22, 46–48, 51, 154, 170–171, 173–177, 190–191, 195–196, 199, 206, 211–212, 218, 281, 283–287, 289–297, 320 realism 1, 14, 17, 18, 61, 113, 114, 124, 132, 133, 153, 154, 155, 161, 162, 163, 164, 167, 215, 383, 386; conceptual realism 18, 153–155, 163–164, 167; näive realism 1, 18, 61, 133, 153, 154, 155, 161–163, 215, 383 reasons: motivating reasons 12, 19, 194, 197–211; normative reasons 197–200, 202–204, 206, 211, 231–233, 235, 237; space of reasons 6–9, 31, 39, 64, 216, 217, 218, 219, 220, 223, 231, 234, 368 reasons for belief 1, 12, 19, 20, 177–178, 184, 186, 194, 196, 198–199, 201, 203–205, 207, 210, 326, 354, 356 recognition 137, 185, 331–334, 340 safety 47–49, 52, 236 skepticism (scepticism) 3, 5–9, 21, 25, 41, 51–52, 54, 64, 82, 84–86, 97–100, 110, 113, 115, 117, 121, 123, 127, 215, 218, 223, 227, 235, 259–261, 264–265, 269–270, 275–278, 284, 286, 295, 296, 370–371, 374–378, 382, 385 testimony 2, 8–9, 13, 21, 23, 203, 292–293, 298–314, 325, 337, 340, 345 veridical and non-veridical experience 1–2, 8, 10–12, 15, 17–18, 47, 63–65, 67–68, 74, 77, 79, 82, 84, 138–139, 141, 144, 146, 155, 157, 161–163, 170–171, 190, 194–195, 202, 207, 210, 248, 287–288, 295, 318–319, 321–324, 338, 341 warrant 7, 13, 16, 63–64, 69, 73–76, 79, 82, 85, 89–93, 96, 98, 101–102, 105–109, 111, 114, 120, 122, 153–154, 158–160, 162–163, 167, 175–176, 190, 199, 207, 317–318, 321–329, 333–339, 375, 381, 385