Egan - Appearance Properties

of 27
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Research paper
  NO ˆUS 40:3 (2006) 495–521 Appearance Properties?  1 A NDY  E GAN University of Michigan/Australian National University Introduction Intentionalism—the view that the phenomenal character of an experience iswholly determined by its representational content—is very attractive. Unfor-tunately, it’s in conflict with some quite robust intuitions about the possi-bility of phenomenal spectrum inversion without misrepresentation. Facedwith such a problem, there are the usual three options—reject intentionalism,discount the intuitions and deny that spectrum inversion without misrepre-sentation is possible, or find a way to reconcile the two by dissolving theapparent conflict.Sydney Shoemaker’s (1994) introduction of   appearance properties  is a par-ticularly ingenious way of pursuing the third strategy, by maintaining thatthere is a representational difference between the phenomenally spectrum-inverted subjects. 2 In introducing appearance properties, Shoemaker doestwo things: he identifies a theoretical role for some family of properties toplay, and he suggests a family of properties as candidates to play that role. I’llargue that his proposed candidates do not play the role as well as we wouldlike, suggest some new candidates, and argue that they do a better job of playing the role. The reason for the question mark in the title is that, if I’mright, it turns out that the best candidates to play the appearance propertyrole aren’t properties. 1. Intentionalism and Spectrum Inversion Intentionalists hold that the phenomenal character of an intentional mentalstate supervenes on its representational content. Visual experiences are inten-tional mental states. 3 So, if intentionalism is true, there can’t be two visual C  2006, Copyright the AuthorsJournal compilation  C  2006, Blackwell Publishing, Inc. 495  496  NO ˆ US  experiences that are alike in their representational content, but differ in theirphenomenal character.Suppose Ernie and Vert are phenomenally spectrum inverted with respectto each other. Ernie’s visual experiences when he looks at ripe tomatoes, fireengines, and cooked lobsters are phenomenologically just like ours. 4 Vert’svisual experiences are phenomenally inverted. His experiences when he looksat ripe tomatoes, fire engines, and cooked lobsters are phenomenally likeour (and Ernie’s) visual experiences when we look at unripe tomatoes, limes,and uncooked lobsters. We can put this in terms of the qualitative featuresof Ernie’s and Vert’s experiences by saying that Ernie’s visual experienceof Kermit is G, while Vert’s is R, where a G experience is one with thephenomenal character of our experiences of unripe tomatoes, etc., and an Rexperience is one with the phenomenal character of our experiences of ripetomatoes, etc.Suppose also that Ernie and Vert are (in relevant respects) historically andbehaviorally indistinguishable. They’ve both had relevantly similar sorts of causal intercourse with limes and lobsters, they both use color terms in thesame way, they both put things into the same piles when asked to sort themby color, and so on. For example, when Ernie and Vert look at Kermit, theyboth say “he’s green”, and they both stack Kermit in the pile with the unripetomatoes and uncooked lobsters when asked to sort things by color.A natural interpretation of Ernie and Vert’s situation is that it’s one inwhich Ernie and Vert are phenomenally spectrum inverted, but their visualexperiences agree on the colors of things. If their experiences  didn’t  agreeon the colors of things, one or the other of them would have to be gettingthe colors of things  wrong . And this seems implausible. Given the similarityof their histories, discriminatory abilities, stacking behavior, etc., we don’tseem to have any reasonable grounds for attributing the error to one of themrather than the other. (This is where the ‘without misrepresentation’ part of ‘spectrum inversion without misrepresentation’ comes in.) So we ought tosay that their visual experiences both represent Kermit as being green, eventhough their phenomenal character is different.Let’s suppose that this kind of case—in which two observers are phenome-nallyinverted,thoughtheirvisualexperienceshavethesamecolorcontent—ispossible. 5 Then we seem to have a counterexample to intentionalism. That is,we seem to have a difference in phenomenal character without a differencein representational content.However, an important part of the argument was suppressed in the pre-vious paragraphs. Why should we agree that Ernie’s and Vert’s visual expe-riences have the same (overall) representational content? Here is a plausibleargument: if Ernie’s and Vert’s experiences don’t differ with respect to  color content, they don’t differ with respect to representational content  at all   (orat least, not in any way that’s potentially relevant). The only available place  Appearance Properties  497 to locate a representational difference between Ernie’s and Vert’s experiencesis in which colors their visual experiences represent Kermit as having. Sosince Ernie’s and Vert’s experiences have the same color content (both rep-resent Kermit as being green), they have the same representational content simpliciter .Shoemaker’s strategy is to deny the claim that Ernie’s and Vert’s experi-ences differ with respect to representational content  only  if they differ withrespect to color content. Intentionalism is to be saved by supposing that,while Ernie’s and Vert’s experiences both represent Kermit as being  green ,there is a representational difference with respect to the  other  properties thatthey attribute to Kermit. 2. A Bunch of Distinctions, a Problem, and Another Distinction Before moving on to Shoemaker’s proposal, we should pause to head off analternative line of response to the conflict between intentionalism and thepossibility of spectrum inversion without error.Contentful experiences have  representational properties . One kind of rep-resentational property is the property of having a certain content. There area lot of competing ideas about what contents are, but all parties to the debatemust concede that representations have  at least  the following sort of content:they make a distinction between the possibilities in which things are as they’rerepresented to be, and the possibilities in which things are otherwise. Maybethere are other, finer-grained kinds of contents, too, but if there are, thosekinds of finer-grained contents will each at least  determine  a coarse-grainedpossibility-carving content of the sort described above.So one (fairly coarse-grained) class of representational property is the classof possibility-carving properties. Call these possibility-carving properties the  pure  representational properties. 6 (I will, from now on, stipulatively use ‘con-tent’ to mean ‘possibility-carving content’. I don’t mean to commit myself toany view about which notions of content are viable, theoretically useful, making this stipulation—I just need a less cumbersome expression than‘possibility-carving content’, and I won’t be employing any of the competingnotions in what follows.)When Grover believes, Oscar asserts, Miss Piggy fears, and Ernie and Vertsee that Kermit is a frog, Grover’s belief, Oscar’s assertion, Miss Piggy’s fear,and Ernie’s and Vert’s visual experiences all share a pure representationalproperty. They all have contents that separate the worlds in which Kermit isa frog from the rest.There are other kinds of representational properties. For example,  visu-ally representing that Kermit is a frog , or  representing Kermit, under mode of  presentation K, as a frog . Call these and their ilk  impure  representational prop-erties. We can think of the pure representational properties as very general  498  NO ˆ US  properties of   representing content C somehow-or-other , and the impure rep-resentational properties as properties of   representing content C   this  way . 7 The intentionalist slogan is, “same representational properties, same phe-nomenal character”. The distinction between pure and impure representa-tional properties lets us draw a distinction between two sorts of intention-alism.  Pure intentionalism  says that the phenomenal character of experiencesupervenes on the pure representational properties of experience.  Impure in-tentionalism  says only that phenomenal character supervenes on impure rep-resentational properties.Let’s also distinguish between  maximal   and  submaximal   representationalproperties of both kinds. Maximal representational properties specify thecontent (or the content plus the manner of representation) of an experienceexhaustively, while submaximal representational properties specify only partof the content of the experience. 8 So pure intentionalism says that maximalpure representational properties fix phenomenal character, while impure in-tentionalism makes the weaker claim that maximal  impure  representationalproperties fix phenomenal character.If it’s possible to have both an occurrent thought (or some other sort of cognitive experience) and a visual experience with the same possibility carv-ing content—that is, the same maximal pure representational property—thenpure intentionalism is false. (Since while the visual experience will have somedistinctive phenomenal character, having the thought either won’t have anyphenomenal character at all, or will have a very different one than that of thevisual experience with the same content). Since this is almost certainly possi-ble, pure intentionalism is almost certainly false. Call this the  problem of com-mon content . So it looks as if the intentionalist will have to appeal to impurerepresentational properties in order to make her thesis plausible, given thatthere are some maximal pure representational properties that can be sharedby experiences in different modalities (visual, cognitive, auditory, etc.). 9 So why be so worried about the sort of spectrum inversion case describedabove? What the case of Ernie and Vert is a clear counterexample to is  pure  in-tentionalism, and we already knew that that was false, because of the problemof common content. For all that’s been said so far, it could be that Ernie’sand Vert’s experiences have the same possibility-carving content, but theyeach represent it in different ways. They might, for example, each representthe property  being green  under a different mode of presentation. 10 So Ernie’sand Vert’s experiences would share all of their pure representational prop-erties, while differing in their impure representational properties. Since theintentionalist already needs to appeal to impure representational propertiesto handle the problem of common content, where’s the harm in appealing tothem here as well?Notice a distinction between two kinds of impure representationalproperties. First, there are impure representational properties like  visually
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks