117

6 Bodily Sensations as an Obstacle for Representationism Ned Block Representationism,1 as I use the term, says that the phenomenal character of an experience just is its representational content, where that representational content can itself be understood and characterized without appeal to phenomenal character. Representationists seem to have a harder time handling pain than visual experience. (I say ‘seem’ because in my view, representationists cannot actually handle either type of experience successfully, but I will put that claim to one side here.) I will argue that Michael Tye’s heroic attempt (this volume) at a representationist theory of pain, although ingenious and enlightening, does not adequately come to terms with the root of this difference. Representationism is in part an attempt to make an account of phenomenal character comport with G. E. Moore’s diaphanousness intuition, the idea of which is that when I try to introspect my experience of the red tomato, I only succeed in attending to the color of the tomato itself, and not to any mental feature of the experience. The representationist thinks we can exploit this intuition to explain phenomenal character in nonphenomenal terms. To understand representationism, we need to know what to make of the phrase ‘representational content’ as applied to an experience. There is no clear pretheoretical notion of representational content as applied to an experience, certainly none that will be of use to the representationist. True, I can speak of seeing that and seeing as, and more generally of experiencing that and experiencing as. Looking at the gas gauge, I can say that I see that the tank is empty (Dretske 1995). And I can say that I experience my wound as a medical emergency. These (and other) pretheoretical ways of thinking of something that could be called the representational content of experience have little to do with phenomenology or with the kind of properties that the representationist takes the phenomenology to constitutively represent. Thus the representationist thesis involves a partially stipulated notion of representational content. This is not, in itself, a criticism, but as I shall argue, there is a problem about how the stipulation should go in the case of pain.138 Ned Block Thus in my view, the dispute between Tye and Colin McGinn over whether pain even has representational content is not a dispute about a matter of fact, but a dispute about how to talk. The same applies to Tye’s claim that a referred pain (e.g., a pain in the inside of the left arm caused by malfunction in the heart or a pain in the groin caused by malfunction in the kidney) is nonveridical. Pretheoretically, we might (might!) regard such a pain as misleading but not false or inaccurate or nonveridical. We are willing to allow hallucinations in which it seems to us that there is a colored surface in front of us but there is no colored surface that we are seeing, in front or elsewhere. However, we do not acknowledge pain hallucinations, cases where it seems that I have a pain but in fact there is no pain. Tye does not argue for pain hallucinations in which there seems to be a pain but there is no pain at all, but since he does say that referred pain is nonveridical, he must think that a referred pain in the arm is not actually in the arm. Where, then, is it? In the heart? It is not our practice to assign referred pain in this way, so such a claim is at best stipulative. In the case of representationism about some aspects of visual phenomenology, there is a fairly natural line of stipulation. My color experience represents colors, or colorlike properties. (In speaking of colorlike properties, I am alluding to Sydney Shoemaker’s “phenomenal properties” [1994, a,b] or “appearance properties” [2001] or Michael Thau’s [2002] nameless properties.) But, according to me, there is no obvious candidate for an objectively assessable property that bears to pain experience the same relation that color bears to color experience. But first, let us ask a prior question: what in the domain of pain corresponds to the tomato, namely, the thing that is red? Is it the chair leg on which I stub my toe (yet again), which could be said to have a painish or painy quality to it in virtue of its tendency to cause pain–experience in certain circumstances, just as the tomato causes the sensation of red in certain circumstances? Is it the stubbed toe itself, which we experience as aching, just as we experience the tomato as red? Or, given the fact of phantom-limb pain, is it the toeish part of the body image rather than the toe itself? None of these seems obviously better than the others. Once we have stipulated what we mean by the representational content of pain, it is a substantive and nonstipulative question whether the phenomenal character of pain is that stipulated representational content. The stipulative aspect of the issue is reflected in Tye’s presentation by the fact that two-thirds of the way through the paper, he has not yet quite stated what he intends to stipulate. He says “What, then, is the phenomenal character of pain?” and considers the possibility that one might say the representational content of pain is a matter of its representing subjective qualities of the bodily region in which the pain occurs. He rejects this proposal on the groundBodily Sensations 139 that the phenomenal character of a pain in the leg can be present even when there is no such bodily region (as in phantom-limb pain), suggesting instead that “the phenomenal character of pain is representational content of a certain sort, content into which the experienced qualities enter” (this volume, emphasis added). The “certain sort” alludes to his view that the relevant contents are nonconceptual, abstract, and poised. The problem that is worrying me is what these “subjective qualities” or “experienced qualities” are in terms of which Tye characterizes the representational contents of the phenomenal character of pain. (I will use the former phrase and indicate the problem of the obscurity typographically by talking of Subjective Qualities.) Examples of Subjective Qualities in Tye’s sense are what we speak of when we describe a pain as sharp, aching, throbbing, or burning. Here is the problem: why don’t these Subjective Qualities bring in the very unreduced phenomenality that the representationist is seeking to avoid? Let me explain via the comparison with Shoemaker’s (1994a,b; 2001) version of representationism mentioned above. Shoemaker honors the diaphanousness intuition without the reductionist aspect of representationism. He holds that when one looks at a red tomato, one’s experience has a phenomenal character that represents the tomato as having a certain appearance property and also as being red, the latter via the former. Each appearance property of an object can be defined in terms of production by it in certain circumstances of a certain phenomenal character of experience. The view is motivated in part by the possibility of an inverted spectrum. If Jack and Jill are spectrum inverted, Jack’s experience of the tomato represents it both as red and as having appearance property A (the former via the latter). Jill’s experience represents the tomato as red and as having appearance property A. (Jack’s experience represents grass as green and A, whereas Jill’s experience represents grass as green and A.) What determines that Jack’s experience represents appearance property A is that it has phenomenal character PC, and A gets its identity (with respect to Jack) from the production of PC in normal viewing conditions. Red can be identified with the production of PC in Jack, PC* in Jill, and so on. PC is metaphysically more basic than A since PC is what makes it the case that the experience represents A. But A is epistemically more basic than PC in that in perception of colors one is aware of A rather than PC. And in introspection, one is aware that one’s experience represents A. Awareness of PC, by contrast, is at least in part theoretical (which I see as a big problem with Shoemaker’s view). Shoemaker’s view of the relation between phenomenal character and appearance properties has been in flux, but what I think has been constant is something I can agree with, that PC and A are a pair such that each could be defined140 Ned Block in terms of the other taken as basic. Or, if the two are defined in terms of one another as a “package deal,” with neither as basic, the definition would not capture the difference between the ·PC, AÒ pair and the very different ·PC, AÒ pair. Shoemaker’s appearance properties are in that sense of a piece with phenomenal characters. Shoemaker’s (2001) view of pain is that pain experiences are perceptions that represent a part of the body as instantiating an appearance property. Such a view is not problematic for Shoemaker since if he is to be called a representationist, his representationism is nonreductionist: he is not attempting to explain phenomenal character in nonphenomenal terms. But if Tye’s Subjective Qualities are appearance properties, then Tye cannot be a representationist in the sense that he at least used to endorse, in which phenomenal character is supposed to be explained in nonphenomenal terms. Does Tye give us any reason to think that his Subjective Qualities are not appearance properties in a sense that undermines his (former?) project? Well, if he said that as a matter of empirical fact, these Subjective Qualities turn out to be (aspects of) tissue damage, then I think they could not be taken by him to be appearance properties. But Tye’s view is not that Subjective Qualities are features of tissue damage. Rather, what he says is something importantly different, namely that “pain” applies to tissue damage when it is within the content of a pain experience. And it is good that he does not identify Subjective Qualities with aspects of tissue damage, since that identification would be most implausible given that exactly the same tissue damage in the foot can give rise to a more intense pain—or one that is different in other ways—in me than in you because of differences between my fibers leading from the foot to the brain and yours. The representationist says that when I try to introspect my experience of the stubbed toe, I only succeed in attending to the Subjective Quality of the toe. My question to Tye has been: why think of the Subjective Quality of the toe as like the redness of the tomato rather than like an appearance property of the tomato? Of course, Tye is a representationist about visual experience as well as about pain, so presumably he will reject the question or regard it as a choice between a correct option (red) and a confused option (an appearance property). To see the difficulty in such a position, we have to recognize that colors are objective in a way that Subjective Qualities are not. There is an appearance–reality distinction for red but not for a Subjective Quality such as achiness. (Aydede [forthcoming] quotes a characterization of pain from the International Association for the Study of Pain that pretty much makes the point that there is no appearance–reality distinction for pain.) The tomato is red whether or not anyone is looking at it, but the achy toeBodily Sensations 141 cannot have its Subjective Quality if no one is having pain. That is, there can be unseen red but not unfelt achiness. Indeed, tomatoes would still be red even if there never had been any people or other creatures who could see them. But in a world without pain-feeling creatures, there would be no Subjective Qualities at all, no burning limbs or achy toes. In the case of color, a physicalist theory has some plausibility. For example, colors may be held to be sets of reflectances. This account fits with the idea that there could be colors in a world with no perceivers, since tomatoes could reflect light even if no one was there to see it. But a physicalist account of Subjective Qualities in terms of tissue damage is not remotely plausible, for the reason given above—the Subjective Qualities of a toe depend not only on the tissue damage but on the connection between tissue damage and the brain. Whether something is red can be an objective matter, but whether my toe aches is something others know about only because of my special privileged relation to it. Finally, as Shoemaker (2001) notes, there is a many–one relation between color–appearance properties and color. Looking around the room, I see all four walls as white, but the color–relevant appearance properties are nonetheless different because of differences in lighting. However, there is no corresponding distinction in the domain of pain. Every slight difference in appearance is a difference in Subjective Quality, indicating that Subjective Qualities are mentalistic in a way that colors are not. That is why bodily sensations have been a challenge for representationism. If the representationist proposes to explain phenomenal character in nonmentalistic and especially nonphenomenal terms, there must be something for the phenomenal character to (constitutively) represent that is not itself individuated with respect to phenomenal character. Color is a better bet to pass this test (even if it does not pass in the end) than are Subjective Qualities. Of course the view of color that I have been presupposing is itself controversial. It may be said that a physicalistic theory of color ignores the fact that what color something has is relative to the perceiver. Colored objects produce slightly different phenomenal characters in different normal observers in normal circumstances, because the various parts of the eye differ among normal perceivers—perceivers who can be assumed to perceive correctly—male versus female, young versus old, black versus white (Block 1999). Perhaps color is not objective after all. So perhaps we should say that in a world without perceivers, nothing has colors. Or perhaps we should say that they have all colors—each relative to a different possible but nonactual perceiver. And once we have gone that far, we might say instead that there are no colors even in the actual world, rather merely the projection of phenomenal characters onto objects (Boghossian and Velleman 1989, 1991). But these are all views of color that142 Ned Block would deprive representationism of its reductionist point. The challenge to Tye is to manage to assimilate Subjective Qualities to color as an objectivist would see color.2 Notes 1. Some say ‘representationalism,’ but I prefer ‘representationism.’ “Representationism” is shorter and ‘representationalism’ is ambiguous, being used also to mean the doctrine in epistemology that seeing is mediated by awareness of a representation, namely, indirect or representative realism. As Aydede (forthcoming) notes, representationism is more akin to direct rather than indirect realism, so the ambiguity is confusing. Since we still have a chance for the more rational use of terms, I hope readers will adopt ‘representationism.’ 2. I am grateful to Sydney Shoemaker for some comments on an earlier draft.