Table Of Content

THE AQUINAS LECTURE, 2011 PERCEPTION AS A CAPACITY FOR KNOWLEDGE JOHN MCDOWELL MARQUETTE UNIVERSITY PRESS © 2011 Marquette University Press T PREFATORY Milwaukee, Wisconsin 53201-3141 All rights reserved. he Wisconsin-Alpha Chapter of Phi Sigma www.marquette.edu/mupress/ Tau, the International Honor Society for Philosophy at Marquette University, each Under the auspices of the year invites a scholar to deliver a lecture in honor of Wisconsin-Alpha Chapter of Phi Sigma Tau St. Thomas Aquinas. The 2011 Aquinas Lecture, LIBRARY OF CONGRESS CATALOGING-IN-PUBLICATION DATA Perception as a Capacity for Knowledge, was delivered McDowell, John Henry. on Sunday, February 27, 2011, by John McDowell. Perception as a capacity for knowledge / by John Prof McDowell is a Distinguished University Pro McDowell. fessor of Philosophy at the University of Pittsburgh. p. cm. - (Aquinas lecture; no. 75) Includes bibliographical references. ISBN-13: 978-0-87462-179-2 (hardcover: alk. JOHN MCDOWELL paper) ISBN-I0: 0-87462-179-8 (hardcover: alk. pa Prof McDowell studied at the University College per) of Rhodesia and Nyasaland and at New College, 1. Perception (Philosophy) 2. Knowledge, Theo , I ry of I. Title. B828.45.M43 2011 121:34-dc22 2010051475 Printed in the United States of America. @The paper used in this publication meers the minimum requirements of the American National Standard for Information Sciences Permanence of Paper for Primed Library Marerials. ANSI Z39.48-1992. ~ Association of Amerocan 1(1~ University Presses MARQUETTE UNIVERSITY PRESS ~lLWAlJI<f:1' The A.ssodation of J~suit LJnivp-rsily PreSl;H~l' 6 JOHN MCDOWELL Perception as a Capacity for Knowledge 7 Oxford, and taught at University College, Oxford reception of premodern philosophy, of a conception from 1966-1986. He is an Honorary Fellow of Uni of nature that reflects an understanding, in itself versity College. He joined the Philosophy Depart perfectly correct, of the proper goals of the natural ment at the University of Pittsburgh in 1986. Prof. sciences. In a number of contexts, he has argued McDowell's research and teaching fields include that we can free ourselves from characteristic sorts Greek Philosophy, Philosophy of Language, Phi of philosophical anxiety by recalling the possibility losophy of Mind, Metaphysics, Epistemology, and of a less restrictive conception of what it takes for Ethics. Prof McDowell translated Plato's Theaetetus something to be natural. for the Clarendon Plato series (Oxford University Press, 1973). He is author of Mind and World (Har To Prof McDowell's distinguished list of publica vard University Press., 1994), and his essays have tions, Phi Sigma Tau is pleased to add: Perception been collected in several volumes: Mind, Value, and as a Capacity for Knowledge. Reality (Harvard University Press, 1998), Meaning, Knowledge, and Reality (Harvard University Press, 1998), Having the World in View: Essays on Kant, Hegel, and Sellars (Harvard University Press, 2009), and The Engaged Intellect: Philosophical Essays (Har vard University Press, 2009). Prof. McDowell has held visiting appointments at Harvard University, the University of Michigan, UCLA, and Princeton University. He was invited to give the John Locke lectures at Oxford University in 1991 and the Woodbridge Lectures at Columbia University in 1997. He is a Fellow of the British Academy and a Fellow of the American Academy of Arts and Sciences. He is an Honorary Doctor of Humane Letters of the University of Chicago. A central theme in much of his work is the harm ful effects, in modern philosophy and in modern PERCEPTION AS A CAPACITY FOR KNOWLEDGE JOHN MCDOWELL 1. My topic is perception as a capacity for knowl edge. But there are different things one could mean by describing perception in those terms. I want to consider perception as a capacity for knowledge of a distinctive sort. At one point in his classic investigation o£ among other things, the epistemology of knowledge through perception, "Empiricism and the Philoso phy of Mind': Wilfrid Sellars famously says this: [IJn characterizing an episode or a state as that of knowing, we are not giving an empirical description of that episode or state; we are placing it in the logi cal space of reasons, of justifying and being able to justify what one says.l In this remark Sellars is implicitly equating, on the one hand, reason, the faculty or family of facul ties that, in an ancient tradition, distinguishes ra tional animals (animals that can occupy positions in "the logical space of reasons") from the rest of 1 "Empiricism and the Philosophy of Mind'; in Sel lars, Science, Perception and Reality (Atascadero, Cali fornia: Ridgeview, 1991), §36. \ 10 JOHN MCDOWELL Perception as a Capacity for Knowledge 11 the animal kingdom, and, on the other hand, mas This applies to knowledge in general something tery of language, and in particular language with Sellars has been claiming, in the immediate con which one can give expression to one's credentials text, about observational knowledge in particular. for saying things ("being able to justify what one And that brings us to my topic, perception as a says"). That is, he is suggesting that to be a rational capacity for knowledge. Observational knowledge animal is to be a language-using animal. We might is knowledge one has through the operation of frame the suggestion like this: human beings are perceptual capacities. And Sellars considers obser born with a potential for rationality, and that po vational knowledge as a kind of knowledge whose tential is brought into first actuality, in the guise of instances are self-conscious rationality at work. a capacity available for exercise in second actuali I will not go into detail about Sellars's treatment ties that we might call "acts of reason'; by initiation of observational knowledge. But I can sketch its into a linguistic practice. basic shape quite quickly. Against that background, Sellars is indicating Given that he is considering observational that the topic of his epistemological reflections is knowledge as a form of the kind of knowledge that knowledge of a sort that is distinctive of rational is self-conscious rationality at work, and given the animals-that is, given his assumption, knowledge connection he assumes between rationality and of a sort that is distinctive of language-using ani language, it is unsurprising that he approaches mals. An instance of knowledge, of the sort Sellars observational knowledge by considering linguis is concerned with, is an act of reason. And part of tic moves that express it, speech acts that he calls what that amounts to is that someone who has a "reports". He puts two conditions on observational bit of knowledge of the sort Sellars is concerned reports: first, a report must have an authority that with is self-conscious about its credentials: in Sel consists in its issuing from a reliable capacity to lars's terms, she occupies a position in the space of speak the truth, in appropriate circumstances, being able to justify what one says. If someone has about the matters it concerns (consider, for in a bit of knowledge of the sort Sellars is concerned stance, the capacity to know the colours of things with, she can state not only what she knowledge by seeing them, an example Sellars makes much ably believes, but also how her believing it is ratio of); and, second, that authority "must in some sense nally grounded in a way that shows the belief to be be recognized by the person whose report it is" knowledgeable. 12 JOHN MCDOWELL Perception as a Capacity for Knowledge 13 (§35).2 Someone who knows something by obser is green, or perhaps from her being inclined to say vation, according to Sellars's conception, must be that it is green, with the cogency of the inference self-consciously aware of the authority with which underwritten by the reliability she needs to be able she speaks if she expresses what she knows, an au to ascribe to herself according to the second of Sel thority her position has by virtue of its being the lars's two conditions. upshot of a reliable capacity to be right about the But it seems bizarre to suggest that the thought relevant kind of thing. Spelled out in connection that something she sees is green is warranted like with the capacity to know the colours of things by that for the subject herself. Moreover, the idea is looking at them, the requirement is that someone out of line with one of Sellars's fundamental con who gives expression to such knowledge by saying, victions, that knowledge through perception is not for instance, "That's green" must be able to vindi inferential. cate the authority with which she speaks by saying 1 do not believe Sellars is even flirting with such something like "I can tell a green thing when 1 see an idea when he sets out the two conditions. No IL one". doubt someone else could exploit the reliability the This might make it seem as if Sellars is flirting subject ascribes to hersel£ to license an inference with the idea that the warrant by virtue of which from her saying that something is green to the con an utterance of "That's green" counts as expressing clusion that it is green. But Sellars is not implying knowledge has an inferential structure, like this: that the subject's own warrant for her belief that what warrants the subject in believing that the thing the thing is green consists in such an inference. she is talking about is green is the cogency of an He requires that she be able to claim reliability for inference to its being green from her saying that it hersel£ but he does ~~eive her claim of reli ability as formulating the principle of an inference 2 The distancing note sounded by Sellars's qualifi that captures her warrant for believing the thing is cation "in some sense" is there because he thinks the green. Her warrant for believing the thing is green second condition is difficult to accommodate; he is that she can see that it is. The reliability she must goes on to make heavy weather over the credentials be able to claim is a condition for her to be able to a reporter can have for counting as knowing that she have experiences in which she sees that things are has that authority. I think the difficulty Sellars makes green; it is such experiences that warrant her in for himself over this is misplaced, and I am going to , believing, and saying, that things are green. ignore the qualification. • I 14 JOHN MCDOWELL Perception as a Capacity for Knowledge 15 2. Sellars says knowledge is a position in the space lives of human children in whom the potential for of reasons. And he connects the ability to occupy rationality has not yet been brought into first actu positions in the space of reasons with the ability to ality. Sellars has no reason to deprecate the use of use language. One might be tempted to think this epistemic vocabulary in connection with subjects expresses a kind of human chauvinism, a position that do not count as rational animals. He has no that would refuse to attribute knowledge in any reason to claim, for instance, that any knowledge proper sense to non-human animals. But I think such beings are credited with is only loosely so that would be point-missing. called.3 In the terms I have suggested, his point is Certainly philosophers can be tempted into that just that the knowledge of such beings does not sort of position. Consider the Cartesian idea that belong to the species of knowledge he is concerned non-human animals are automata, beings that do with, knowledge that is rationality in operation. not have mental states; on such a view, non-hu And that applies to Sellars's treatment of per man animals do not have beliefs, and if we think of ception in particular. Directing our attention to knowledge as belief with appropriate credentials, perception as a capacity for a ~J._~i~c:ri~JdE-4. of we have to say, a fortiori, that they do not have knowledge, knowledge that is an act of reason, anything genuinely recognizable as knowledge. need not be prejudicial to the possibility of ac But as I understand him, Sellars is not making knowledging that perception is, on some suitable a pronouncement about how we may properly use understanding, a cognitive capacity in many kinds the word "know" and its cognates. He is noting of non-human animals, and in pre-rational (pre that his topic is knowledge as an act of reason in linguistic) human children, also. a sense that he connects with language and self In restricting his topic to knowledge of a kind consciousness. But that need not be prejudicial to whose instances are reason in operation, Sellars is a more liberal application of epistemic concepts, making a move that is not special to him, and it for purposes other than his. I do not believe we is not arbitrary or unmotivated. In the Critique of miss anything Sellars has reason to insist on if we Pure Reason, Kant makes this well-known remark: take him to conceive his topic as a species of a ge nus, which, for all he cares, can be recognized as 3 Compare, for instance, Robert Brandom, "Knowl edge and the Social Articulation of the Space of Rea being instantiated also in the lives of at least some sons'; Philosophy and Phenomenological Research vol. 55 non-human animals, and, we might add, in the no. 4 (1995),895-908, at 899-900, n. 3. 16 JOHN MCDOWELL Perception as a Capacity for Knowledge 17 All the interests of my reason, speculative as well as instance of the knowledge I am concerned with as practical, combine in the three following questions: a potentially self-conscious standing in the space 1. What can I know? of reasons. 2. What ought I to do? 3. It might seem that Sellars's focus on knowledge 3. What may I hoper as self-conscious rationality at work goes badly The first of those questions defines what we might with an interest in perception, in particular, as a ca call "philosophical epistemology': The question pacity for knowledge. is posed in the first person, by a self-consciously The conception of knowledge Sellars expresses rational subject, one who asks the question as in the remark I began with is epistemologically encapsulating one of the interests of his reason. internalist, in this sense: the warrant by virtue of Sellars is more explicit than Kant in linking self which a belief counts as knowledgeable is acces conscious rationality with the capacity to use lan sible to the knower; it is at least potentially known guage. But Kant too is tacitly presupposing the ca by her. Someone who has a bit of knowledge of pacity to use language, at least in the shape of the the relevant sort is self-conscious about the cre capacity to ask questions like these. And Sellars's dentials of her knowledge. As Sellars puts it, she remark about what we are doing when we classify occupies a position in the space of being able to something as a case of knowledge, of the distinc justify what one says. tive sort he is interested in, lines up well with the Now it seems obvious enough that some of the implicit definition of philosophical epistemology knowledge had by a rational subject is grounded put in place by Kant's question. If! ask"What can like that, grounded in a way that is accessible to the I know?': I must be conceiving any instance of the subject. Suppose I have a bit of knowledge whose knowledge I am asking about as an achievement warrant consists in the cogency of an inference to whose credentials to count as such, which must what I know from something else, my warrant for warrant the relevant beliefs by the lights of my which is for present purposes not under question. reason, I can recognize. I must be conceiving anx. I must be able to say not only what it is that I know in that way, but also why it counts as knowledge; 4 Critique of Pure Reason, translated by Norman I must be able to cite, and if necessary defend, the Kemp Smith (London: Macmillan, 1929), A804-5/ inferential basis on which I know it. "My neighbour B832-3. 18 JOHN MCDOWELL Perception as a Capacity for Knowledge 19 is at home'; I say. "How do you know?'; you ask. "entitlement" for a species of warrant that, as he I must be able to make some such reply as this: explains, "His car is in his driveway:' And if my interlocutor need not be fully conceptually accessible, even on responds "Maybe he went to work on the bus'; I reflection, to the warranted individual. The indi must be able to recognize the relevance of what he vidual need not have the concepts necessary to says. Perhaps I should then withdraw the claim to think the propositional content that formulates know that was implicit in my first assertion. the warrant.6 But the credentials of perceptual knowledge And he maintains that the warrant perceptual are surely not like the credentials of knowledge states provide for beliefs about the environment self-consciously grounded in inferential connec is entitlement in that sense. Even for a subject for tions between propositions. And it can seem that whom a perceptual warrant is conceptually acces if we give an internalist account of warrants for sible-a subject who can justify a belief she holds perceptual knowledge, the effect is that we cannot by citing a perceptual state she is in-the warrant accommodate that dissimilarity. that the perceptual state itself affords to the belie£ For reasons of this sort, Tyler Burge advocates as Burge sees things, is warrant of the same sort an epistemological externalism, in a sense that is that perceptual states provide for subjects who a counterpart to internalism as I have found it in cannot conceptualize that warrant. So we cannot Sellars, about the warrant that perceptual states give an acceptable account of that sort of warrant provide for knowledge.s Burge adopts the term by saying something that is meant to be specific to the perceptual beliefs of rational subjects. 5 Burge applies the idea of inference to "truth-and Part of what is motivating Burge here is a reason-preserving propositional transitions among thought I have already considered, that perceptual propositional attitudes" even in animals that are not able to give expression to the warrants-which, on knowledge can b~ had by non-human animals and this conception, may be inferential-for their beliefs. See "Perceptual Entitlement'; Philosophy and Phe can capture the objection by saying that as Burge nomenological Research vol. 67 no. 3 (2003), 503-48, understands him, Sellars groups perceptual knowledge with knowledge whose title to count as knowledge at 519. Burge's objection to Sellars ian internalism consists in an inferential warrant that is accessible to about perceptual knowledge cannot be put by saying the subject. Sellars groups perceptual knowledge with inferential knowledge on that undemanding conception. But we 6 "Perceptual Entitlement'; 504. 20 JOHN MCDOWELL Perception as a Capacity for Knowledge 21 . ~Il) ;~ pre-rational (pre-linguistic) human children as must be wrong in an account that is supposed to 1,' well as by mature human beings, subjects in whom fit perceptual knowledge in general. But that does (' a potential for reason has been brought into first not sh o w. t hat 'm ternal I' sm cannot be n'gh t ab o ut a I ': , fJ'il,l l· ''f,, f' actuality. Burge is calling for a general account of distinctive kind of warrant by virtue of which the \'/ J -rI. perceptual warrant as such, an account suitable for perceptual knowledge of rational subjects counts i ,f: all cases of perceptual knowledge. He writes: as the knowledge it is. If our concern is with a spe- I ~I ! I " cies, we do not have to restrict ourselves to things I A viable conception of warrant and knowledge must J include both primitive and sophisticated types. A ~h~t are tr~e of all instances of the genus of which condition of viability is that such a conception apply It IS a specIes. to animal and child perceptual belief, and knowl edge, as well as to mature human instances of belief 4. Such considerations perhaps undermine some and knowledge.7 of Burge's rhetoric. But they would not be enough As it stands, this seems to express a version of to disarm his resistance to a Sellarsian internalism the thought I tried to discourage when I argued about perceptual knowledge, even if we stress that that Sellars should not be accused of human chau the internalism is meant to apply only to the war vinism. As I urged in that context, giving a special rant for a special kind of perceptual knowledge, account of the perceptual knowledge of rational distinctive of rational subjects. Burge has an ob animals is consistent with regarding perceptual jection that such considerations do not address. If knowledge in rational animals as a sophisticated Burge's objection works, it tells against a Sellarsian species of a genus that is also instantiated more internalism about perceptual knowledge even if primitively in non-rational animals and pre-ratio the internalism is explicitly restricted to a species nal human children. of perceptual knowledge that is held to be distinc / To exploit the idea of genus and species like that tive of rational subjects. I is to accept what Burge insists on, that non-rational Burge actually considers two sorts of approach, subjects, subjects incapable of getting their minds supposedly inspired by Sellars, to perceptual around the warrant for their beliefs, can have knowledge. perceptual knowledge. It follows that internalism According to the first, there is nothing of epi stemological significance between circumstances in a perceiver's environment and the formation of 7 "Perceptual Entitlement'; 505.