Disjunctivism and Skepticism - Oxford Handbooks
Page of

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). (c) Oxford University Press, 2015. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Handbooks Online for personal use (for details see Privacy Policy).

Subscriber: null; date: 29 April 2016

Disjunctivism and Skepticism

Abstract and Keywords

This article considers the epistemological significance of disjunctivism and its bearing on philosophical skepticism. It explains that disjunctivism is a way of thinking about perceptual experience and perceptual knowledge and it is also the view that judgments that characterize how experience appears to a subject is as if things have disjunctive truth conditions. It discusses the relation of disjunctivism with the tradition of the conception of experience and the concept of recognitional abilities.

Keywords: disjunctivism, philosophical skepticism, perceptual experience, perceptual knowledge, truth conditions, recognitional abilities

1. Introduction

Disjunctivism is a way of thinking about perceptual experience and perceptual knowledge. In one central form, it is the view that judgments that characterize how it appears to a subject as if things are have disjunctive truth conditions. To appreciate what is distinctive about disjunctivist thinking and why it is interesting, it is necessary to consider a traditional approach in the theory of perception. This will provide us with the background needed to assess the epistemological significance of disjunctivism and its bearing on philosophical skepticism.

2. Disjunctivism and the Traditional Conception of Experience

When I see some physical object, say, a cat on a chair, I have various visual experiences. There is a traditional way of thinking according to which having such experiences always falls short of seeing a cat. Whether (phenomenally) visual experiences are the same or different is determined by whether they are the same or (p. 582) different with respect to how it looks to the subject as if things are. The experiences I have as I look at the cat need not differ in this respect from experiences I might have when hallucinating a cat. But assuming that seeing is a success concept, I do not see the cat unless the cat is there. It follows that merely having an experience such that it looks to me as if a cat is there before me does not suffice for seeing the cat, and even in a circumstance in which a cat is there, I still might not see it despite having an appropriate experience. If I have the experience because my brain has been cleverly manipulated by neurophysiologists, and the experience just happens to be such that it is as if a cat is there, then surely I do not see the cat. Under the conditions of the example, there is nothing to connect my having the experience with the presence of the cat. Some connection is required for perception. The response to this is to suppose that I see the cat only if

  1. (a) I have an appropriate experience,

  2. (b) a cat is there, and further,

  3. (c) my having the experience is causally dependent on there being this cat before me.1

What I shall call the traditional conception of experience is the conception of experience that forms part of this way of thinking. Some terminology will help sharpen the exposition. I shall call experiences perceptual when their subjects come to have them through the normal operation of the relevant sensory modality in an episode in which they perceive something. Experiences are phenomenally visual if they have the sort of phenomenal features that characterize visual perceptual experiences—the kind of features that perceivers can capture by describing how it looks to them as if things are. Experiences can be phenomenally visual by this account even if they are not perceptual. For traditionalists, the same experience I have when looking at the cat could be had by me when I was perfectly hallucinating. So whether an experience is perceptual or hallucinatory has nothing to do with its intrinsic character. The intrinsic character of an experience is its phenomenal character; whether the experience is hallucinatory or perceptual depends on its causation.

In an influential series of articles, Paul Snowdon, drawing upon earlier work by J. M. Hinton, presents a view of phenomenally visual experiences on which a difference in experience need not show up as a phenomenal difference.2 Even if it is conceded that how it looks to me as if things are in a situation in which I see something can be the same as how it would look to me as if things were in a situation in which I was hallucinating, it does not follow that the experiences in the two situations would be the same. Fundamental to the disjunctivists' way of thinking is a relational conception of perceptual experiences. If I see a lemon in a basket on the kitchen table, then under this conception I have an experience that is essentially relational because it is essentially a visual encounter with the lemon. If I were to have a hallucinatory counterpart of that experience—a perfectly hallucinatory experience that does not differ from the perceptual experience with respect to how it looks to me as if things are—this would be a different experience because it would not be essentially an encounter with the lemon. The difference between (p. 583) traditionalists and relationalists, as we might call them, is over the conditions under which experiences are the same or different.

Snowdon presents disjunctivism as a view about the truth conditions of sentences or judgments of the form “It looks to S as if p.” He writes

That “looks” sentences are true in hallucinations and in perceptions, and are not ambiguous, does not entail that they are made true by (or are true in virtue of*) exactly the same kind of occurrence in both cases.

(1990: 129)

He adds that such sentences, and the judgements they can be used to express

are made true by two types of occurrence: in hallucinations they are made true by some feature of a (non‐object‐involving) inner experience, whereas in perceptions they are made true by some feature of a certain relation to an object, a non‐inner experience, (which does not involve such an inner experience).

(1990: 130)

Arguably, the broader picture to which this kind of disjunctivism is linked is separable from any specific concern to provide truth conditions for looks judgments, if that involves supplying necessary and sufficient conditions for the truth of these judgments in other terms. One might be a relationalist about perceptual experience while not engaging in any such enterprise. What, I think, is at the heart of disjunctivism is a view about what it is for it to be the case that it appears to one as if such and such is there. In relation to visual experiences, it captures at least the following ideas:

  1. (a) Experiences that are such that it looks to the subject just as if an F is there include cases in which the subject sees an F and cases in which a subject merely hallucinates an F.

  2. (b) An experience of seeing an F is an essentially relational experience implicating an F. Any such experience differs in kind from any experience of hallucinating an F even if how it looks as if things are to the subject who sees the F is the same as how it looks as if things are to the subject who is hallucinating an F.

  3. (c) The experience of seeing an F does not break down into (1) the having of an experience that is not essentially relational (an inner experience, in Snowdon's terms) and (2) the satisfaction of further conditions.

3. Epistemological Disjunctivism

Snowdon's interest is not primarily epistemological. He is concerned with how we should conceive of perceptual experiences, and he takes it that a constraint on an adequate conception is that it should make sense of how experiences enable us to (p. 584) think about objects demonstratively. When we perceive objects, we are in a position to have demonstrative thoughts about them. Watching a tennis match, I can think, That return was good, where the reference of the demonstrative element of my thought, that return, is the very return I have just observed. Here I not only pick out perceptually a particular return, but I am also in a position to think of it, demonstratively, as that return. It is my having observed the return in question that enables me to do so. These considerations present a challenge to traditionalists about perceptual experience. If the experience that I have when I see an F is one I could have when not seeing an F, then it is not clear how the experience in the case in which I see an F enables me to think about the F in question demonstratively, never mind how it enables me to acquire knowledge about it. On the traditionalist's account, it is not intrinsic to the experience that it presents an object to the subject, so explanatory work is needed to show how it can contribute to making the object available as an object of demonstrative thought. The attraction of the idea that perceptual experience is essentially relational is that it avoids this problem. Just because the experience is essentially relational, the object is, so to speak, embraced by the experience and thus available to be thought about by suitably equipped subjects.3 That, at any rate, is the idea. It is clearly of interest for epistemology since any plausible view of perceptual knowledge must be consistent with facts about how perception makes demonstrative thoughts available to us.

In the work of John McDowell, we find a disjunctivist style of thinking recruited to shed light directly on epistemological matters. Although McDowell acknowledges a debt to Snowdon and Hinton, it is not immediately obvious how the disjunctivist strand in his thinking links up with their distinctive concerns about the nature of perceptual experience. I shall first describe a strand in McDowell's epistemological thinking in its own terms and then circle back later, in section 5, to pick up the question of how it relates to the theory of perceptual experience.

In “Criteria, Defeasibility, and Knowledge” (McDowell 1982), McDowell criticizes a way of thinking about knowledge that he thinks plays into the hands of skeptics. The skepticism that is the main target of the article is not “external‐world” skepticism but “other‐minds” skepticism. This is presented as relying on the following picture:

Judgements about other minds are, as a class, epistemologically problematic. Judgements about “behaviour” and “bodily” characteristics are, as a class, not epistemologically problematic; or at any rate, if they are, it is because of a different epistemological problem, which can be taken for these purposes to have been separately dealt with. The challenge is to explain how our unproblematic intake of “behavioural” and “bodily” information can adequately warrant our problematic judgements about other minds.

(1982: 382)

This picture incorporates an evidentialist model of knowledge of other minds. On this model, judgments about other minds can be justified and can constitute knowledge in virtue of being based on defeasible evidence relating to behavioral (p. 585) and bodily information that does not itself implicate mentalistic concepts. It is this evidentialist model that McDowell thinks plays into the hands of skeptics. For this purpose, skeptics may be represented as follows:

  1. (a) They accept that if there were knowledge of other minds, it would have to be on the lines of the evidentialist model.

  2. (b) They find the evidentialist model inadequate. Failing to appreciate that an alternative is available, they infer that we lack knowledge of other minds.

In McDowell's view, skeptics are right to find the evidentialist model inadequate. One reason for thinking it inadequate concerns our entitlement to take facts about bodily behavior, conceived in nonmentalistic terms, as evidence for the obtaining of psychological states that are not perceptually manifest. Another concern is focused on the very idea that an adequate view could “envisage ascribing knowledge on the strength of something compatible with the falsity of what is supposedly known” (1982: 372). This latter theme is not pursued in depth in “Criteria, Defeasibility, and Knowledge,” though it is always in the background. (I shall defer discussion of it until section 6.) The main strategy in that article has the following three stages:

  1. 1. The diagnosis: A version of the argument from illusion is viewed as motivating the evidentialist model of knowledge of other minds.

  2. 2. The way out: It is argued that the argument from illusion is not compelling. This is done with the help of a disjunctivist account of appearances and of a perceptualist model of some of our knowledge of other minds.

  3. 3. The response to the skeptic: The availability of a plausible perceptualist model is taken to show that the skeptic's argument is not compelling.

What is on offer, then, is an attempt to undermine the skeptic's argument. I turn now to a closer look at the diagnostic stage and the way out.

The argument from illusion, under McDowell's conception, runs as follows:

[S]ince 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—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 … in a non‐deceptive 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 deceptive and non‐deceptive cases alike, and hence as something that is a defeasible ground for the knowledge, though available with a certainty independent of whatever might put the knowledge in doubt. …

… In a deceptive case, what is embraced within the scope of experience is an appearance that such‐and‐such is the case, falling short of the fact: a mere (p. 586) appearance. So what is experienced in a non‐deceptive case is a mere appearance too.

(1982: 386)4

The application to other‐minds cases is straightforward. Take the nondeceptive case (the good case) to be one in which Bill is anxious and looks it, and the deceptive case (the bad case) to be one in which Bill looks just the same as he does in the other case but is not anxious. In the two cases, according to the argument, what the experience takes in must be the same—Bill's appearing a certain way that could be specified entirely in nonmentalistic terms. But if we assume that what experience takes in must supply the justification of a judgment that Bill is anxious, the justification must be the same in both cases. What experience takes in is the fact that Bill appears as he does. This is at best defeasible evidence that he is anxious. It is defeasible because its evidential value could be undermined by further information, for instance, information to the effect that Bill is fooling around or maliciously deceiving. This is all in keeping with the evidentialist model.

The crucial inference in the argument from illusion, as conceived by McDowell here, 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.” The good and the bad cases under consideration are experientially indistinguishable in the sense that Bill looks just the same in the two cases. The conclusion drawn is that what the subject takes in must be the same in these two cases. What is taken in is what is seen to be so. That, according to the argument, has to be the same in the two cases, which is why the only plausible model for knowledge of other minds is an evidentialist one. The basis for a judgment to the effect that Bill is anxious has to be something that is as much available to experience in the bad case as in the good case.5

Does McDowell deny the legitimacy of the crucial inference, or does he reject the assumption of the inference? On a certain understanding of “experiential indistinguishability,” the right answer seems to be that he challenges the legitimacy of the inference and does not reject the assumption. He thinks that if by “experiential indistinguishability” you just mean that there is no difference with respect to how things appear, then the good case and the bad case are experientially indistinguishable (see the middle paragraph of 1982: 389). But we are supposed to resist the conclusion that what the subject takes in is the same in both cases. This takes us to the stage of McDowell's strategy that constitutes what I called the way out.

The way out has two complementary strands. The first is a disjunctivist conception of what it is for it to appear to a subject that such and such is the case:

an appearance that such‐and‐such is the case can be either a mere appearance or the fact that such‐and‐such is the case making itself perceptually manifest to someone.

(1982: 386–87)

(p. 587)

Notice that here appearances seem to be subjective rather than worldly. The focus is not on how Bill looks, but on how it looks to a subject that Bill is. (I do not think that it distorts McDowell's position to think of subjective appearances in terms of how it appears to a subject as if things are.) We are to think of its looking to me as if Bill is anxious in these terms:

  1. (a) Either it is a case in which the fact that Bill is anxious makes itself visibly manifest to me or it is a case in which Bill merely looks to me as if he is anxious (though he is not anxious).

  2. (b) The case in which the fact that Bill is anxious makes itself visibly manifest does not break down into (1) my having visible evidence that would be as much available in a corresponding bad case and (2) making a judgment on the basis of that evidence.

It is noteworthy that thus far there is no reason to think that McDowell is committed to the kind of disjunctivism about perceptual experience considered by Snowdon. The rejection of a common factor in the good and bad cases is not presented as the rejection of the idea that there is an experience common to the two cases. It is presented as a rejection of the idea that what the experiences take in must be the same and, along with that, a rejection of the idea that there is a common evidential base in the two cases. The epistemological disjunctivism exemplified by (a) and (b) allows McDowell to concede that it can be true that it appears to me as if Bill is anxious in both the good case and the bad case and, indeed, that there is no difference with respect to how it appears as if Bill is in the two cases, even though what the experience takes in is different in the two cases. For in the good case, what is taken in is that Bill is anxious, but in the bad case, what is taken is at best merely that he has a certain look. But it is not clear at this stage whether the difference with respect to what the experiences take in requires that there be a difference in psychological kind between the experiences. It all depends on whether experiences are individuated in terms of what, if anything, they take in. (But see section 5.)

Taken on its own, this disjunctivist maneuver could easily seem like a sleight of hand, for it provides no explanation of how it is that what is taken in by an experience gained by looking can be different in the two cases. It should be noted, however, that the disjunctivist maneuver is not supposed to carry the whole weight of the case that is designed to undermine the evidentialist model. It is supplemented by a perceptualist model of some of our knowledge of other minds.

Initially it might not seem promising to treat any knowledge of other minds as perceptual. It might seem bizarre that anyone should suppose that we can see straight off that someone is anxious when it is agreed on all sides that a person can display the demeanor of one who is anxious yet not be anxious. What makes sense of the perceptualist model is a conception of perceptual knowledge as noninferential knowledge acquired by suitably equipped subjects from what they perceive. Perceptual knowledge under this conception is the kind of knowledge we acquire when we tell from its look that a bird is a magpie or that a flower is an (p. 588) orchid. It is noninferential knowledge in that it is not acquired by reasoning from prior assumptions. It is phenomenologically immediate in that what is known simply strikes the subject as being so. In order to be suitably equipped, subjects need more than the repertoire of concepts required in order to entertain the content of the knowledge. They need to have an appropriate sensibility that incorporates appropriate recognitional abilities. McDowell applies this general conception to cases that philosophical tradition has treated in evidentialist terms. For instance, he applies it to knowledge of what people are saying from hearing what they say. In this case, a prerequisite is obviously knowledge of the language of the speaker. It is entirely plausible that no inference, in any ordinary sense, is involved. Indeed, it looks wrong to suppose that the knowledge that the speaker says that p is in any sense based on evidence pertaining to the speaker's utterance that falls short of saying that p. Yet there is no doubt that knowledge of what people say is a species of knowledge of other minds.6

Bringing sensibilities into the picture helps explain why somebody with the appropriate sensibility can detect that Bill is anxious from his look when others lacking those abilities cannot. But what we are most concerned to understand is how somebody who does have the requisite abilities could be in a better epistemic position in the good case than in the corresponding bad case. To address this problem, we need a fuller account of recognitional abilities. McDowell himself does not provide such an account. In the next section, I sketch a way of thinking about the abilities that can be deployed to fill out the perceptualist model.7

4. Recognitional Abilities

I have the ability to tell by looking, in suitable conditions for observation, whether or not something is an orchid. That ability implicates modes of judgment formation whereby I am prepared to judge that something is or is not an orchid in response to suitable prompts. Suppose, for instance, that somebody wants to know whether a flower in a pot is an orchid. If, having looked in the right direction, I have appropriate visual experiences, then, absent countervailing factors, I shall judge that the thing in question is an orchid. It is not that I consider what kind of experiences I have, it is just that what I judge depends on the kinds of experiences I have. If I have a certain different range of experiences, then absent countervailing factors, I shall judge that it is not an orchid. If I do not have a good‐enough look, then I shall suspend judgment either way. In some few cases, I would suspend judgment if the thing in question were not close enough to visual type to tell either way. (It might be an orchid that has seen better days.) The modes of judgment formation that are implemented are such that they reliably yield true judgments. But having such modes in my repertoire does not suffice for my having the ability (p. 589) to tell by looking, under suitable conditions for observation, whether or not something is an orchid. The very same modes could come into operation in an unusual environment in which the visually orchidlike things one encounters are at least as likely to be skillfully made from synthetic materials. If I were in such an environment, I would not have the ability to tell of things seen in that environment that they were or were not orchids. To count as having this ability with respect to a given environment, that environment must be favorable—it must not be one in which things that look just like orchids could easily not be or things that are orchids could easily not look like orchids.

When we think of people as having the ability to tell by looking whether or not something is an F, we ascribe to them an ability of the sort I have described. But here we come to a couple of important twists. (1) Implementations of the relevant modes of judgment formation can yield false judgments even if the environment is one with respect to which the subject has an ability to tell whether or not it is an F that is present. The modes of judgment formation must yield true judgments in such an environment with a high degree of reliability, but that is compatible with their occasionally yielding a false judgment. (That there is the odd thing that is not an F but looks just like one does not render the environment unfavorable.) (2) The relevant recognitional ability will have been exercised only if the resulting judgment is true. The ability is an ability to tell, that is, to come to know, by looking whether or not something is an F. So if the subject does not come to know one way or the other, then the ability will not have been exercised.

Clearly we have recognitional abilities that enable us to tell by looking that something is so in some cases when it is, but that would not generally suffice for telling by looking whether it is so. My ability to tell from his look that Bill is anxious would be like that. We may suppose that Bill has a way of looking—a certain demeanor—such that when he displays that demeanor, I can tell just by looking at him that he is anxious. Even so, I may not in general be able to tell just by looking whether he is anxious. He might not always display his anxiety.

We can now explain why it should be thought that on encountering anxious‐looking Bill, I tell that he is anxious despite the fact that he could look just as he does and not be anxious. I tell because I have an appropriate recognitional ability, the environment is favorable to the exercise of that ability, and on the occasion in question, I do exercise it.

The conception of recognitional abilities in play here does not provide the materials for a reductive conceptual analysis of perceptually knowing‐that. To exercise such abilities is to come to know something by perceiving it to be so. I have suggested that these abilities implicate reliable modes of judgment formation, but it is not part of the suggested view that we can build up to a conceptual analysis of these abilities from a prior conception of these modes. It is our conceptions of the abilities that are in the driver's seat; our grip on the modes of judgment formation is in terms of those conceptions. For instance, the experiences that are of the right kind for judging that something is an F are just the ones that would lead someone with the appropriate ability to judge that it is an F. It is doubtful that the range of (p. 590) such experiences is specifiable independently of the concept of the ability in question. Note too that the fact that the abilities are characterized as abilities to know does not undermine their explanatory power. It is not as if they are conceived as powers, but goodness knows what powers, to acquire knowledge. They are powers of which we have commonsense conceptions, and which reflection reveals to implicate reliable modes of judgment formation and to be individuated, in part, in terms of environments. It is informative to know that an item of knowledge has been acquired by one such ability rather than another.

The unavailability of a conceptual analysis of recognitional abilities in nonepistemic terms is in keeping with the fact that our most basic engagement with the acquisition of perceptual knowledge—our own and that of others—is at the level of factive notions like telling that (*finding out that), by means of seeing that, hearing that, and the like. We seem to have learned to apply these notions in a largely satisfactory way without any guidance from reductive‐analytical accounts of them. It is arguably in favor of McDowell's way of thinking that these factive notions have this role.8

With this picture in place, we can make sense of the idea that I can take in that Bill is anxious by looking. At any rate, if taking this in is a matter of recognizing that Bill is anxious by looking, then I can do that. A crucial difference from standard approaches to epistemology is the rejection of the idea that I know that he is anxious on the basis of evidence that is common to the good and bad cases. My knowing turns on my having exercised an ability to tell that Bill is anxious from his look, but that he looks this way is not to be conceived as evidence on which I base the judgment that he is anxious. This is just as well, since I would be at a loss to sum up how Bill looks in the form of an assumption on the basis of which I judge that he is anxious, and there would be the further problem of explaining what entitles me to connect his look with his mental state. As things are, I simply tell by looking. This is in keeping with McDowell's way out. The other‐minds skeptic is to be rebuffed by introducing a plausible perceptualist model of some knowledge of other minds. The model helps make sense of McDowell's disjunctivism about appearances and, in particular, of his view that the experiential intake of the subject can be different in the good and bad cases even though how it appears as if things are and, indeed, how the things one sees appear are the same in the two cases.

5. The Character of Experience and What Experience Takes In

As I noted earlier, Snowdon, like Hinton, is primarily concerned with the character of perceptual experiences, not with issues about knowledge.9 Accepting the kind of epistemological disjunctivism that I have already identified in “Criteria, Defeasibility, and Knowledge” does not in any obvious way commit us to disjunctivism (p. 591) about perceptual experience. In the light of this, it might be wondered whether McDowell need take a stance on the intrinsic character of perceptual experience. Other‐minds cases might suggest otherwise. It is not obvious that the view of recognitional abilities I have presented commits us to supposing that the experience I gain when I tell that Bill is anxious is different, qua experience, from the experience I would gain if Bill were feigning anxiety. Nonetheless, McDowell does take a stance on the intrinsic character of perceptual experience. This stance is crucial for his view of what is needed to make sense of what he calls our having a vantage point on the external world. This vantage point is threatened, he suggests, “if subjectivity is confined to a tract of reality whose layout would be exactly as it is however things stood outside it” (1986: 241). The threat comes from a picture of subjectivity on which perception of the world is indirect, being mediated by denizens of inner space. McDowell not only rejects this view but also wishes to replace it with a picture on which experience is essentially world‐involving and thus essentially relational.10 It is at this point that it is natural to think of his epistemological concerns as linking up with Snowdon's concerns. The thought would be that no account of perceptual knowledge will be adequate unless it explains how a suitably equipped subject who has a perceptual experience is put in a position to have demonstrative thoughts about some object. By the light of relationalists about perceptual experience, traditionalists lack the resources to account for how experiences can make objects available for demonstrative thought. If the experience I have as I look at the lemon in the basket could have been had by me if I were perfectly hallucinating, then having the experience does not suffice to explain what makes the lemon available to me as an object of demonstrative thought. Notice that this view invokes a conception of demonstrative thoughts as essentially relational and has it that when such thoughts are made available by perception, the implicated experiences must also be essentially relational. It is a further step to suppose that experiences can be essentially takings‐in of facts, as McDowell often appears to think. Which facts are taken in is, plausibly, determined by which recognitional abilities one exercises. So there is space for a position according to which although perceptual experiences are essentially relational, it is not essential to any perceptual experience that any particular recognitional ability has been exercised and any particular fact taken in.

It is important to emphasize that traditionalists agree that perception is a subject‐world relation. They may agree that a theory of perceptual knowledge must account for the possibility of perceptual‐demonstrative thoughts—demonstrative thoughts about objects perceptually picked out. What they will certainly dispute is that the only plausible explanation will assume that the experiences implicated in perceptions are essentially relational. That leads us to the question whether traditionalists have the resources to account for the availability through perception of demonstrative thoughts.

One way to do so would be to view the sort of recognitional abilities of which I have been speaking as depending on abilities for perceptual discrimination. I have in mind here not the ability to discriminate one type of object from another at the level of judgment, but rather a lower level ability to pick out and track objects (p. 592) perceptually.11 Think of walking along a corridor and encountering various potential obstacles. You pick them out visually. What this amounts to is that your visual experiences and dispositions to behavior are shaped in a certain way. How you move is determined by sense‐perceptual and proprioceptive cues. Arguably, visual discrimination involves subdoxastic states that register such cues and prime subjects for behavior, including subintentional activity.12 We engage in such activity when we are catching balls, reaching out to pick up objects, playing shots at tennis, and much else besides. When you catch a ball in a game, you do so intentionally, but the systems that achieve hand‐eye coordination result in small adjustments to the trajectory of your arm that are not intentional (though neither are they unintentional). Whereas the carrying out of intentions is guided by beliefs, the kind of subintentional activity of which I am speaking seems to be guided by changes in subdoxastic states. The activity is responsive to cues that are picked up by sight but need not be registered at the level of belief. Through the operation of the visual system, changes in experience and changes in behavioral dispositions vary systematically with changes in the environment and changes in the orientation and position of the subject so as to enable the subject to move relatively smoothly in that environment. My picking out Bill by sight is a matter not merely of having a certain series of experiences, but of an interplay in which my experiences and my behavioral dispositions are shaped as a result of the visual impact of Bill's presence upon me. It is a relational state targeted on Bill because it is his presence, and possibly his changing location, that does the shaping and induces dispositions that enable me to be responsive to the relative positions of him and me in my egocentric space.

I see no reason why traditionalists about experience should not draw upon a view of visual discrimination along these lines in order to make sense of how perception makes objects available for demonstrative thought. Traditionalists may concede that an adequate account of perceptual knowledge must address the issue of how perception can enable us to have essentially relational demonstrative thoughts about physical objects, but it is not clear why they need take this to commit us to an essentially relational view of perceptual experience. Perceptual experience does enable us to pick out objects perceptually. But what the picking‐out consists in does not seem to require that the experiences themselves be essentially relational. Relationality comes in through the shaping by the object of the subject's experience and behavioral dispositions.

6. Knowledge, Justification, and Reasons

A central preoccupation of McDowell's Mind and World is with understanding how our thinking can be rationally constrained by perceptual experience—how it can be empirically grounded.13 His view is that to understand this, we need to (p. 593) understand how perceptual experience can bear on our justification for believing this and that. To do that, he thinks that we must settle on the right way to think about how experiences give us cognitive access to facts. This in itself hardly seems controversial, for no one is going to disagree that empirical knowledge depends on perceiving things to be thus and so. As McDowell sees it, though, the experiences themselves must be essentially relational. Not only are they essentially pickings‐out of objects, but they are also essentially takings‐in of facts about these objects. This is open to challenge even if we grant that perception furnishes us with knowledge. We surely do see, and thereby come to know, that this or that is so, and likewise for other modes of perception. But so far as I can see, here too, it is an open question whether such a view demands a conception of experience as essentially relational.

Be that as it may, McDowell has interesting and important things to say about knowledge and rationality that seem to me to be detachable from some of the details of his conception of perceptual experience. We can approach this strand in McDowell's thinking by turning to “Knowledge and the Internal” (McDowell 1995), a difficult article that makes penetrating criticisms of traditional epistemology. The target of this article is what McDowell calls the interiorization of the space of reasons.

Our thinking operates in the space of reasons in that it is responsive to what there is reason for us to think and do. The interiorization of this space that McDowell envisages is the upshot of a variant of the argument from illusion. Consider a case—a good case—in which I know because I see that, say, Mary has arrived. My knowing involves my taking it that Mary has arrived. By the argument, my taking this to be so is based on an experience such that it looks to me as if she has arrived. But it is possible, even if it is unlikely as things are, that I could have just such an experience and believe accordingly even though Mary has not arrived. This would be the corresponding bad case. In the bad case, by the argument, I have the same basis for thinking that Mary has arrived as in the good case. My standing in the space of reasons is therefore the same in the two cases. That standing has to do with what can be imputed to me as a rational thinker; it cannot turn on whether Mary has or has not arrived but depends on whether my belief is justified. Under the prevailing assumptions, I am justified and, indeed, have the same justification in both the good and bad cases. To take such a view is to interiorize the space of reasons.14 It straightforwardly follows from this view that whether one knows that something is so is not determined just by a standing in the space of reasons, since knowledge depends on the satisfaction of the truth requirement, and that is being conceived as external to the space of reasons. The upshot of this way of thinking is that knowledge is taken to be a hybrid involving both a suitable standing in the (interiorized) space of reasons and the satisfaction of the truth requirement.

On a variant hybrid approach, knowledge involves the operation of processes that reliably yield true beliefs plus the satisfaction of the truth requirement. A belief might result from a reliable process and yet be false, so the truth requirement is independent of the reliability requirement. On this type of hybrid view, the very idea of the space of reasons plays no role, and so it is even more obvious that the acquisition of knowledge cannot be viewed as an achievement of a rational subject (p. 594) that is secured through operations in that space. The same applies even on a view that retains a role for justified belief but also incorporates a reliability condition, for the satisfaction of the latter condition will be external to the space of reasons: it will depend on factors that are independent of the operations within that space. (For the themes of this paragraph, see McDowell 1995: 400–*402.)

It is bedrock for McDowell that knowledge itself is a satisfactory standing in the space of reasons. What is at issue is how this can be so. An important consideration for him is that it cannot be the case that the difference between a good and bad case, for instance, between the cases involving Mary just considered, consists simply in the fact that in the good case the truth requirement is satisfied and in the bad case it is not. He thinks that it makes a difference to our standing in the space of reasons whether or not we have embraced a fact. So it makes a difference to my standing in the space of reasons if I have seen that Mary has arrived, as opposed to having a nonveridical experience such that it looks to me as if she has arrived.

Suppose that McDowell is right to reject the view that the only difference between the good and bad cases is that in the latter the belief is false and in the former it is true. Traditionalists about knowledge who adopt an interiorized conception of the space of reasons might well agree, for they might concede that the view in question makes it hard to capture the idea that knowledge is a matter of having a cognitive purchase on, or contact with, a fact. (For this language, see McDowell 1995: 402.) For central cases of knowledge, one might think, there must be some connection between the subject's taking it that something is so and its being so. For cases of perceptual knowledge that p, a natural move is to require that the fact that p should have figured in the explanation of its looking to the subject as if p and thereby in the subject's coming to take it that p. It is open to defenders of the interiorized view of the space of reasons to take this plausible point on board. They might also observe that it is unsurprising that there should be parity with respect to justification in good and bad cases: if justification does not guarantee truth, there will pairs of good and bad cases that are on a par with respect to justification.

McDowell is committed to rejecting these ways of thinking about cognitive contact. The crucial consideration for him is that on the proposed account, the explanatory link is external to the space of reasons. It does not account for what it is to have cognitive purchase on a fact just because it does not display the posited explanatory link as relevant to accounting for the subject's achievement of an appropriate standing in the space of reasons. To explain that, according to McDowell, we have to reject the hybrid view.15

The concern to do justice to how cognitive purchase on a fact can be a satisfactory standing in the space of reasons might explain McDowell's resistance to the idea that we could have knowledge on the basis of something compatible with the falsity of what is supposedly known. (This takes us back to a point from “Criteria, Defeasibility, and Knowledge” to which I alluded in section 3. See McDowell 1982: 372.) If I judge correctly that Bill is anxious on the basis of behavioral (p. 595) evidence that might obtain even if he is not anxious, then under the hybrid conception, I might have done all that is to be expected of me within the space of reasons. Yet, as McDowell sees it, the fact that Bill is anxious would in that case lie beyond my cognitive reach. This is not an issue that I can pursue here. It raises difficult questions about how to treat evidence‐based knowledge within McDowell's framework.

Is it so clear that knowledge can be a satisfactory standing in the space of reasons? The view of perceptual‐recognitional abilities introduced earlier enabled us to make some sense of how visual‐perceptual knowledge about an object could be acquired on an occasion despite the fact that the object could look the same in a corresponding bad case. But nothing said in outlining the view made clear how the knowledge acquired can meet the requirements for being a standing in the space of reasons. To address this matter, we need to look more closely into the links between knowledge, justification, and reasons.

7. How Can Knowledge Be a Satisfactory Standing in the Space of Reasons?

There are two aspects to the standing in the space of reasons in terms of which McDowell conceives of perceptual knowledge. One is that knowing is a state in which what is known is available to the subject so that it can figure as a reason, or an ingredient of a reason, for believing other things. The other, more basic aspect is that emphasized in the following passage from “Knowledge by Hearsay”:

If knowledge is a standing in the space of reasons, someone whose taking things to be thus and so is a case of knowledge must have a reason (a justification) for taking things to be that way. But this is allowed for if [for instance] remembering that Clinton is President is itself the relevant standing in the space of reasons. Someone who remembers that things are a certain way, like someone who sees that things are a certain way, has an excellent reason for taking it that things are that way; the excellence comes out in the fact that from the premise that one remembers that things are thus and so, as from the premise that one sees that things are thus and so, it follows that things are thus and so.

(McDowell 1993: 427–28.)

In this passage, knowing that p through seeing that p is conceived as being in a state in which one has a reason to believe that p supplied by the fact that one sees that p. I shall call this the basic point.

To someone schooled in traditional approaches to epistemology, the idea that the basic point could explain how perception furnishes us with reasons, and thus (p. 596) justification, for beliefs could easily look like an abandonment of serious philosophical inquiry. It is hardly a matter of dispute that if I have cognitive access to the fact that I see that p, then this fact can supply me with a reason to believe that p. Having this pointed out might still leave us wondering how we can have cognitive access to such a fact. I am attempting to shed light on the character of perceptual‐cognitive access to facts and, in particular, on how such access can be an appropriate standing in the space of reasons. Addressing this issue by appeal to the availability of a fact, access to which seems to be equally problematic, does not initially look like a satisfying response to the problem. Further explanation is surely needed. Nonetheless, the sheer common sense of McDowell's thinking at this point should not be overlooked. In our ordinary thinking about knowledge and the possession of reasons, we regularly treat ourselves as having reasons to believe something because we have seen it to be so. Epistemological theorizing has made this seem problematic.

By reasoning similar to that which McDowell takes to be a species of the argument from illusion, many theorists will no doubt suppose that the only access I have to the fact that I see that p is via something that falls short of that fact—the fact that I have an experience such that it looks to me as if p. Because this is of a piece with the thinking that informs arguments from illusion, McDowell is committed to rejecting it. How, then, should we think of the access we have to facts as to what we see to be so?

A natural step is to introduce recognitional abilities again. We acquire the ability to tell from its appearance that an F is before us by learning when to apply the concept of an F in response to what we see. Given that we have such an ability, there is no great mystery to how we could have a second‐order ability to apply to oneself the concept of seeing that an F is before one. The very same experiences that enable us to pick out an F and recognize it to be an F enable us to tell that we see that an F is there provided we have the appropriate second‐order recognitional ability. These abilities are just as much tied to favorable environments as are the corresponding first‐order abilities. If the environment is not favorable, we may be primed to deploy the mode of judgment formation implicated by the relevant second‐order ability and yet lack the ability with respect to that environment. But the environment can be favorable. The environment in which we can tell by looking that an F is there can also be an environment in which we are in a position to know that we see that F is there.

With the notion of a second‐order recognitional ability to hand, the following looks like a plausible position:

  1. (a) We can know that an F is there in virtue of seeing (or having seen) an F and recognizing (or having recognized) it to be an F.

  2. (b) When we know that an F is there in this way, we have reason to believe that an F is there, constituted by the fact that we see (or saw) that an F is there.

  3. (c) We have the reason in question simply because the fact that constitutes it is available to us, thanks to our having an appropriate second‐order (p. 597) recognitional ability—an ability to tell whether or not we see (or saw) that an F is (was) there.

It is not a requirement of this view that the second‐order recognitional capacity should actually have been exercised on every occasion on which one knows. Even when it is not exercised, there is a sense in which the corresponding belief is justified. In the usual case, it will be justified in that the fact that provides the justification is readily accessible, and its accessibility has the potential to keep the belief in place.

I do not suggest that McDowell would endorse all the elements of the conception of recognitional abilities that I have been outlining. He would take issue with my view of how traditionalists about experience can do justice to the idea that perceptual knowledge involves cognitive contact with objects and facts (sections 4 and 5). I do think, however, that something like my conception of recognitional abilities, and of how their exercise links up with the possession of justified belief, is required to make sense of the idea that knowledge is itself a standing in the space of reasons. It is crucial to appreciate that as I have presented it, the view does not build up knowledge from justified belief. Rather, the exercise of a recognitional ability explains how we know that p on a given occasion, and since on such an occasion we can readily tell how we know—for instance, by seeing that p—our believing that p can be justified. For all that I have said so far by way of outlining the conception, there could be knowledge without justified belief. This marks a departure from McDowell's own views on the assumption that for him, knowledge is essentially a standing in the space of reasons. My conception of recognitional abilities and justified beliefs does not rule out the possibility that there could be creatures with knowledge who lack the second‐order recognitional abilities in virtue of which facts as to what we see to be so are available to us. But if knowledge is essentially a standing in the space of reasons, then creatures that lack the second‐order abilities do not have knowledge.

8. A Satisfying Response to Skepticism?

McDowell from time to time appears to be dismissive of skepticism. He sometimes gives the impression that once we reinstate a commonsense picture on which we are open to the world, directly embracing worldly facts, skepticism ceases to be of interest.16 However, we should not underestimate the amount of headway that McDowell makes against certain skeptical arguments and, in particular, those that rest on the style of thinking that he calls the argument from illusion.

Skeptics have models of knowledge—conceptions of what would be required for knowledge of various kinds—in the light of which they claim that the requirements (p. 598) for knowledge are not satisfied. The challenge to skeptics that is discernible in the work of McDowell can be seen as having two components.

The first component is to attack the models of knowledge that underpin the moves that the skeptic makes. Suppose that skeptics devise their models of knowledge, like everybody else, by considering cases that from a pretheoretical standpoint would be judged to be central cases of knowledge and reflecting on why they are counted as cases of knowledge. Then the battle is over whether skeptics are right to think that their models reflect the character of those cases. From McDowell's perspective, as I understand it, skeptics lose this battle so far as perceptual knowledge and knowledge of other minds are concerned because their models of knowledge misrepresent the character of the cases. The idea here is not that skeptics cannot do justice to the fact that the cases are cases of knowledge. (That would obviously beg the question.) It is that their models are not supported, and indeed are undermined, by reflection on how in practice we think of knowledge. (I take this to be the thrust of the attack on evidentialist models of knowledge of other minds and, later, on hybrid theories.) This is the first component. To make it convincing, we need alternative models of knowledge that are in keeping with commonsense thinking. McDowell supplies such models, and I have sought to provide additional materials in the form of a conception of recognitional abilities and how they relate to knowledge and justified belief. It is open to skeptics to argue that they do not base their models of knowledge on pretheoretical thinking about knowledge. They may profess to have some different methodology for arriving at models of knowledge. But then we need to know what that methodology is. Otherwise the skeptic stands charged with giving us models of knowledge from nowhere. It is not clear what skeptics can say by way of response.

The second component is simply to point out, from within the stream of life, as it were, that the requirements of knowledge are in fact often satisfied. To take this line is to assume something that the skeptic questions; it is to assume that we know this and that. But if we have the right models of knowledge, we shall still have the dialectical advantage. For if skeptics have lost the battle over models for knowledge, they are in no position to challenge the knowledge claims that in practice we take to be established in the ordinary run of things. That, in broad outline, is the kind of challenge that can be mounted from the resources discernible in McDowell's thinking. It does not set out to give any special philosophical vindication of the claim that there are things we know. It takes on board the task of seeking an understanding of how we know what we profess to know while showing that familiar truths about what we know, at least so far as we have been able to make sense of the knowledge in question, remain untouched by the moves made by skeptics.17

The perspective I have suggested on the dialectic between skeptics and antiskeptics might be questioned. In particular, it might be doubted that skeptics need make use of disputable models of knowledge rather than drawing attention to requirements implicit in our ordinary thinking about knowledge. There is a standard type of skeptical argument that might seem to do just that. The type is (p. 599) represented by the following schema, where not‐H is the negation of some skeptical hypothesis:

  1. (a) If I know that p, then I am in a position to know that not‐H.

  2. (b) I am not in a position to know that not‐H.

So,

  1. (c) I do not know that p.

Consider, for instance, a famous example discussed by Fred Dretske (1970). At the zoo, you see and recognize zebras in an enclosure. Consider, however, the skeptical hypothesis that the creatures in the enclosure are disguised mules. Do you have any reason to think that this false? On the assumption that they would look just the same if they were disguised mules, it might be thought that you have no reason to believe that they are not disguised mules. But if that is right, then, by the argument, you are not in a position to know that they are not disguised mules, and in that case you do not know that they are zebras. As is well known, some respond to this problem by denying that if you know that the creatures in the enclosure are zebras, then you are in a position to know that they are not disguised mules (Dretske 1970; Nozick 1981). I share the view that this is not a plausible option.

On the view that I outlined earlier, you may be able to know that the creatures are zebras because you have a suitable recognitional ability and have exercised it in an environment that is favorable to doing so. The environment, after all, could easily be one that is not liable to contain mules that are made to look like zebras. Moreover, you can have reason to think that the creatures are zebras because the fact that you have seen that they are zebras can be available to you, thanks to your having a suitable second‐order recognitional ability and to your being in an environment that is favorable to its exercise. Against this background, skeptical arguments of the type under consideration can be seen to rely on a contentious assumption that does not drop out of commonsense thinking—the assumption that we do not know that the skeptical hypothesis is false. In the particular case in hand, you can know that it is false that the animals are disguised mules because you can see that they are zebras.

Crispin Wright (2002) presents a challenge to this approach. Suppose, for the sake of argument, that you know that the creatures are zebras and that you are justified (Wright's term is “warranted”) in taking them to be. Does your justification transmit to the claim that they are not disguised mules? Surely not, Wright claims. Your justification, he thinks, is provided by the look of the beasts and has no bearing on the possibility that they are disguised mules. Moreover, if you did have justification for thinking that they are not disguised mules, this would need to be independent of your recognition that their being zebras entails that they are not disguised mules. That is because the justification you have for thinking that they are zebras does not transmit to the claim that they are not disguised mules.

Wright's challenge undoubtedly has some bite. At least when one is reflecting philosophically, it is not hard to put oneself in a frame of mind in which it seems (p. 600) puzzling that you can be justified in thinking that the beasts are not disguised mules when (a) all you have done is look at them and (b) they could look as they do and be disguised mules. Nonetheless, the challenge, as so far presented, begs the question against the kind of view I have drawn from McDowell. It assumes that the justification for thinking that the animals are zebras is provided by how they look and thus does not discriminate between the state of affairs in which they are zebras and that in which they are disguised mules. That is, it assumes that the justification would be the same even if the beasts had not been zebras. But on the view I have been taking seriously, the justification for thinking that the animals are zebras derives from your having seen that they are zebras. That does discriminate between the state of affairs in which they are zebras and that in which they are disguised mules, for if they had been disguised mules, you would not have seen that they are zebras and would therefore not have had the justification you do for thinking that they are zebras. The trouble now is that although Wright's challenge begs the question, the view challenged seems only to have pushed the problem back onto the question of what entitles you to think that you see that the animals are zebras. For (it might seem) the justification that you have for your thinking that you see that the animals are zebras must derive from or be accessible via how it appears to you as if things are, and that does not discriminate between the state of affairs in which you see that the animals are zebras and that in which you see animals that look just like zebras but are not. The point here, it should be noted, is independent of whether or not a disjunctive conception of appearances is correct. Even if the experience in virtue of which it appears to you just as if the animals are zebras is conceived as essentially one of embracing the fact that they are zebras, your justification for thinking that you are having that experience does not discriminate between the case in which you are having that experience and one in which you see animals that look just like zebras, but you are not having that experience. So the argument goes.

Understandable though it is, this last move still begs the question. Wright's challenge assumes that the good case in which you see zebras and the bad case in which you see disguised mules are on a par both with respect to justification of the belief that the animals are zebras and with respect to justification of the belief that you see that they are zebras. But on McDowell's view, there is no parity in either respect. You have a reason to believe that the animals are zebras in the good case that you do not have in the bad case. The reason in the good case is that you see that the animals are zebras. Given the conception of recognitional abilities I have outlined, we can explain your having that reason in terms of the idea that you are in a position to tell that you see that the animals are zebras in virtue of your having an appropriate second‐order recognitional ability. Note that thanks to this ability, it is possible for you to have access to your reason for thinking that the animals are zebras.18

The central issue, which Wright's discussion undoubtedly serves to bring out, is whether you have a reason to believe that the creatures are zebras in the good case that you do not have in the bad case. On the picture I envisage, you know that (p. 601) the creatures are zebras because you have exercised a first‐order ability to tell by looking that they are. What explains your acquiring knowledge on a given occasion is your having exercised such an ability. The ability is tied to the environment you are actually in and environments like it. You do not have this ability with respect to an unfavorable environment in which there is a decent chance that what look like zebras are not. On this picture, knowledge is not built up from justified true belief. But rational agents like ourselves, who have access to facts about what we think and what we see, have a reason to believe that p when we know that p through seeing that p. The reason consists in the fact that we see that p. This by itself will seem to skirt around the real issues if one insists that such facts are not readily available to us. But, arguably, they are readily available to us through the exercise of second‐order recognitional abilities. On this view, knowledge acquisition is possible only because various conditions are satisfied that we have done nothing to check out. That you know that the animals are zebras and know that you see that they are depends on your not being in a weird environment in which there is a decent chance that things looking like them are disguised mules. You need have done nothing to satisfy yourself that you are not in such an environment. For some, this latter claim will be the stumbling block. But it can reasonably seem to be a stumbling block only if it is assumed that knowing that the creatures are zebras would require independent assurance that one is not in a weird environment. That assumption is not obviously true. Indeed, if the model of perceptual knowledge in terms of recognitional abilities is correct, then it is false.

Does this general approach help at all with radical skepticism about the external world, induced by instances of the skeptical‐argument schema in which the skeptical hypothesis puts in question all knowledge of the external world? Arguably, it does. Radical skepticism, as much as skepticism linked to less radical skeptical hypotheses, rests on models of knowledge that are disputable since they impose requirements on knowing that it is not clear need be met. Because the models of knowledge are disputable, they do not subvert the assumption that the cases we count as central cases of knowledge are actual cases of knowledge. Better still, it is arguable that the model of perceptual knowledge that draws upon recognitional abilities more adequately fits our pretheoretical thinking.19

References

Brewer, B. (1999). Perception and Reason (Oxford: Clarendon Press).Find this resource:

    Campbell, J. (2002). Reference and Attention (Oxford: Clarendon Press).Find this resource:

      Child, W. (1994). Causality, Interpretation and the Mind (Oxford: Clarendon Press).Find this resource:

        Comesaña, J. (2005). “Justified vs. Warranted Belief: Resisting Disjunctivism.” Philosophy and Phenomenological Research 72: 367–83.Find this resource:

          Dretske, F. (1970). “Epistemic Operators.” Journal of Philosophy 67: 1007–23.Find this resource:

            Evans, G. (1982). The Varieties of Reference (Oxford: Clarendon Press).Find this resource:

              Greco, J. (2004). “Externalism and Scepticism.” In R. Shantz (ed.), The Externalist Challenge: New Studies on Cognition and Intentionality (New York: De Gruyter).Find this resource:

                Grice, H. P. (1961). “The Causal Theory of Perception.” Proceedings of the Aristotelian Society, supplementary volume 35: 121–52.Find this resource:

                  Haddock, A., and Macpherson, F. (eds.). (2008). Disjunctivism: Perception, Action, Knowledge (Oxford: Clarendon Press).Find this resource:

                    Hinton, J. (1973). Experiences (Oxford: Clarendon Press).Find this resource:

                      Macdonald, C., and Macdonald, G. (2006). McDowell and His Critics (Oxford: Blackwell).Find this resource:

                        Martin, M. G. F. (1997). “The Reality of Appearances.” In Mark Sainsbury (ed.), Thought and Ontology (Milan: FancoAngeli), 81–106.Find this resource:

                          ———. (2004). “The Limits of Self‐Awareness.” Philosophical Studies 120: 37–89.Find this resource:

                            ———. (2006). “On Being Alienated.” In Tamar Szabó Gendler and John Hawthorne (eds.), Perceptual Experience (Oxford: Clarendon Press).Find this resource:

                              McArthur, D. (2003). “McDowell, Scepticism, and the ‘Veil of Perception.’” Australasian Journal of Philosophy 81: 175–90.Find this resource:

                                McDowell, J. (1981). “Anti‐realism and the Epistemology of Understanding.” In Hermann Parret and Jacques Bouvresse (eds.), Meaning and Understanding (Berlin: Walter de Gruyter), 228–48. Reprinted in McDowell (1998), 314–43.Find this resource:

                                  ———. (1982). “Criteria, Defeasibility, and Knowledge.” Proceedings of the British Academy 68: 455–79. Reprinted in McDowell (1998), to which page numbers refer.Find this resource:

                                    ———. (1986). “Singular Thought and the Extent of Inner Space.” In Philip Pettit and Crispin Wright (eds.), Subject, Thought and Context (Oxford: Clarendon Press), 137–68. Reprinted in McDowell (1998), 228–59, to which page numbers refer.Find this resource:

                                      ———. (1993). “Knowledge by Hearsay.” In B. K. Matilal and A, Chakrabarti (eds.), Knowing from Words: Western and Indian Philosophical Analyses of Understanding and Testimony (Dordrecht: Kluwer), 195–224. Reprinted in McDowell (1998), 414–43, to which page numbers refer.Find this resource:

                                        ———. (1994). Mind and World (Cambridge, MA: Harvard University Press).Find this resource:

                                          ———. (1995). “Knowledge and the Internal.” Philosophy and Phenomenological Research 55: 877–93. Reprinted in McDowell (1998), 395–413, to which page numbers refer.Find this resource:

                                            ———. (1998). Meaning, Knowledge and Reality (Cambridge, MA: Harvard University Press).Find this resource:

                                              (p. 604) Millar, A. (2000). “The Scope of Perceptual Knowledge.” Philosophy 75: 73–88.Find this resource:

                                                ———. (2004). Understanding People: Normativity and Rationalizing Explanation (Oxford: Clarendon Press).Find this resource:

                                                  ———. (2007). “What the Disjunctivist Is Right About.” Philosophy and Phenomenological Research 74: 176–98.Find this resource:

                                                    ———. (2008). “Perceptual‐Recognitional Abilities and Perceptual Knowledge.” In Haddock and Macpherson (2008).Find this resource:

                                                      Neta, R., and Pritchard, D. (2007). “McDowell and the New Evil Genius.” Philosophy and Phenomenological Research 74, 381–96.Find this resource:

                                                        Nozick, R. (1981). Philosophical Explanations (Oxford: Oxford University Press).Find this resource:

                                                          Pritchard, D. (2003). “McDowell on Reasons, Externalism and Scepticism.” European Journal of Philosophy 11: 273–94.Find this resource:

                                                            Smith, N. H. (2002). Reading McDowell: On Mind and World (London: Routledge).Find this resource:

                                                              Snowdon, P. (1980–81). “Perception, Vision and Causation.” Proceedings of the Aristotelian Society 81: 175–92.Find this resource:

                                                                ———. (1990). “The Objects of Perceptual Experience.” Proceedings of the Aristotelian Society, supplementary volume 64: 121–50.Find this resource:

                                                                  ———. (2002). “What Is Realism?” Proceedings of the Aristotelian Society 102: 201–28.Find this resource:

                                                                    ———. (2005). “The Formulation of Disjunctivism: A Response to Fish.” Proceedings of the Aristotelian Society 105: 129–41.Find this resource:

                                                                      Strawson, P. F. (1959). Individuals (London: Methuen).Find this resource:

                                                                        Williamson, T. (2000). Knowledge and Its Limits (Oxford: Oxford University Press).Find this resource:

                                                                          Wright, C. (2002). “(Anti‐)Sceptics Simple and Subtle: G. E. Moore and John McDowell.” Philosophy and Phenomenological Research 65: 330–48.Find this resource:

                                                                            Notes:

                                                                            (1.) It is not easy to spell out the nature of the required causal dependence. The problem is noted in Grice 1961.

                                                                            (2.) My exposition is based principally on Snowdon 1980–81 and 1990. Snowdon has further commented on disjunctivism in his 2002 and 2005. For Hinton's ideas, see his 1973. The character of disjunctivist thinking about experience is lucidly explained and discussed in Child 1994. Haddock and Macpherson 2008 has articles on all aspects of disjunctivist thinking.

                                                                            (3.) See Snowdon 1990: sec. 6. That demonstrative thought is philosophically important is emphasized by Strawson 1959 and later in Evans 1982. The view is explored in McDowell 1986 and figures crucially in Brewer 1999 and Campbell 2002.

                                                                            (4.) In its classic form, the argument from illusion is not directly epistemological and is not about which facts experiences “take in.” It is about the objects of perception in the sense of “object” in which a thing is the object of perception when it is perceived.

                                                                            (5.) See Williamson 2000: chap. 8 for further pertinent reflections on evidence in relation to pairs of good and bad cases.

                                                                            (6.) See McDowell 1981. For a discussion of related issues concerning the scope of perceptual knowledge, see Millar 2000.

                                                                            (7.) I deploy the notion of a recognitional ability in Millar 2007 and 2008.

                                                                            (8.) The same applies to the thinking of Williamson 2000, which I have not been able to discuss here.

                                                                            (9.) The same may be said of Michael Martin, who has developed disjunctivist thinking in ways that are akin to, but differ from, that of Snowdon's disjunctivists. See Martin 1997, 2004, and 2006. Snowdon 2006 comments on the variety of disjunctivist thinking.

                                                                            (10.) This is the burden of the portions of McDowell 1986 that deal with experience. See especially sections 58. A similar theme is in play with the use of the metaphor of openness to the world in McDowell 1994, especially lectures 1, 2, and 6 and the afterword to part 1, 140–*41.

                                                                            (11.) I draw on such a view in Millar 2007.

                                                                            (12.) I put the notions of subintentional activity and subdoxastic states to work at various points in Millar 2004.

                                                                            (13.) McDowell 1994. For critical discussions of this book, see Smith 2002 and Macdonald and Macdonald 2006.

                                                                            (14.) I take the line of thought just set out to be the gist of the exposition of the argument from illusion in McDowell 1995.

                                                                            (15.) Juan Comesaña (2005) takes McDowell to task for assuming that his opponents are committed to holding that the truth‐value of a belief can be the only epistemically relevant difference between a good case and a corresponding bad case. I do not think that McDowell makes, or need rely on, any such assumption. It is open to him to concede that hybrid theorists may invoke differences other than truth‐value between good and bad cases. So far as he is concerned, that will not help if the differences concern factors external to the space of reasons.

                                                                            (16.) See, for instance, McDowell 1994: 113.

                                                                            (17.) John Greco (2004) criticizes McDowell for arguing, in effect, that content externalism is sufficient for rejection of skepticism. If the view I have been outlining is correct, content externalism is not the only weapon in McDowell's armory against skepticism. There are resources for a challenge to the skeptic's models of knowledge backed up by a plausible model of perceptual knowledge and justified belief. Duncan Pritchard (2003) and David McArthur (2003) also seem to me to underplay the dialectical value attaching to the case that can be made for thinking that McDowell finds fault with the skeptic's models and replaces their models with one that more adequately captures our pretheoretical thinking about knowledge. Pritchard raises interesting issues about the role of luck in McDowell's epistemology that I cannot pursue here.

                                                                            (18.) The relation between McDowell's views and internalist/externalist debates, and issues about reflective access, are explored in Neta and Pritchard 2007.

                                                                            (19.) For further critical discussions bearing on disjunctivism, see Haddock and Macpherson 2008. Work for this chapter contributed to a project on the value of knowledge generously funded by the Arts and Humanities Research Council. I am grateful to the council for its support and to coparticipants, Duncan Pritchard, and Adrian Haddock for regular discussion of the issues. I also thank John Greco for comments on an earlier version that helped improve this chapter.