Frank Jackson brings out very clearly a serious difficulty for some possible broadly physicalist view of experience. The view is not mine, however, and I confess that I am perplexed as to how Jackson could have thought that it was mine, given some of the basic commitments of the approach I lay out in the book.
Phenomenal properties are properties of experiences. They are, I claim, representational properties of a certain sort. More precisely, since I accept a tracking model for representation (at the level of basic sensory experiences and feelings), I take phenomenal properties to be broadly physical or functional properties of the general type: having a property that causally covaries with such and such an external property in optimal conditions and that stands poised to trigger certain cognitive responses.(1) Jackson argues that this gets me into immediate trouble as follows:
The difficulty is in allowing the very fact of playing a
causal role that mediates between X and Y itself to play a
causal role that mediates between X and Y, be that role
characterised in terms of invariably mediating between them or
only sometimes doing so. Tye wants to hold that any given
phenomenal quality tracks certain features of the environment
in optimal conditions and is poised to cause certain cognitive
states, at the same time as holding that the phenomenal qualty
is the very same property that tracks certain features in the
environment in optimal conditions and is poised to cause
certain cognitive states. But what tracks the age of a tree
is not its having a property that tracks the age of the tree,
and what causes our judgements about the age of the tree is
not its having a property that tends to do so; what does the
tracking is the number of rings, and what causes the
judgements is also the number of rings.
My reply is that I do not want to hold -- indeed I could not want to hold -- that the phenomenal qualities themselves do the tracking. In my view, pains, hunger pangs, tickles, etc are representational entities. They have the structure of symbolic arrays (see Tye 1995, pp. 120-123). So, they have both syntactic and semantic features. When I say that pains track tissue damage (in optimal conditions), that hunger pangs track stomach contractions, and so on, I take them to accomplish this via intrinsic, syntactic features of the arrays. The relevant syntactic features do the tracking. But obviously, according to the PANIC theory, phenomenal qualities are not themselves syntactic items.
I do not deny, of course, that the identification of phenomenal qualities with higher-order, broadly physical or functional properties may lead some to question their causal efficacy on that ground alone. This is not Jackson's complaint in the present context, however. My reply to the more general worry is the usual one: the higher-level properties inherit their causal powers via the lower-level physical properties in virtue of which they are instantiated.(2)
Both Ned Block and Sydney Shoemaker focus on various inversion scenarios. I begin with Shoemaker and the case of Jack and Jill. Red things look to Jack the way green things look to Jill. If neither one is misperceiving, then, according to Shoemaker, the phenomenal character of their color experiences cannot be accounted for in terms of the colors their experiences represent.
Obvious reply: Where there is no misperception, things are as they look. So if neither Jack nor Jill is misperceiving, red things must look red to Jack and green things must look green to Jill. Representationally, then, Jack's experiences, as he views red things, must differ from those Jill undergoes, as she views green things. This difference matches the phenomenal difference.
I am prepared to concede, of course, that there can be differences in color "quality space" between individuals and hence that there can be phenomenal differences in individuals brought about by their seeing the same object in exactly the same circumstances. But I am not prepared to concede that, in such cases, the color experiences of the individuals are representationally identical. If I see the world in shades of grey and you have normal color vision, my color quality space is different from yours. But equally, our experiences do not represent the world as having the same colors.
Obviously, in this example, there are functional differences between the two of us. If the case is switched to one involving a full-blooded inversion scenario so that the individuals are functionally identical in all respects, then, in the case of color, there is, I maintain, strong reason to doubt that inverted qualia are possible.(3)
This brings me to the Inverted Earth example. Suppose Oscar travelled to Inverted Earth in his youth (with inverting lenses in place) and has spent 50 years there. According to Shoemaker, it must be allowed that, on Inverted Earth, blue things do not look phenomenally to Oscar just as they did on Earth. However, Oscar's visual experiences, as he views blue things, surely represent them as blue. The conjunction of these two claims evidently spells trouble for representationalist theories of phenomenal character.
The representationalist has two avenues of reply to this example: (1) deny that Oscar's blue-caused experiences do change phenomenally; (2) deny that Oscar's blue-caused experiences, on Inverted Earth, represent blue. Shoemaker says that (1) is hard to do, since "Oscar reports each day that the sky looks the same color as it did the day before, and by any normal criterion his memory mechanisms are working perfectly" (p. 000). This too is Block's view. According to Block, the traveller's visual states do not change phenomenally, since he can't tell that there is any difference between his earlier states and his later ones (p. 000).
Of the two replies, I favor reply (2). But I also find myself unpersuaded by the Block/Shoemaker objections to reply (1), as I shall now explain. Externalism about memory is, I believe, a plausible position that is capable of independent justification.(4) Memory externalists can allow that in travelling cases, memory is not defective, so long as the term 'defective' is understood to apply to the hardware. Internally, things are operating just as they should. Externally, however, the change in environment has an adverse effect. For the concepts the traveller deploys in the new setting get their contents fixed by it; and that fact generates mismemories, where the new and the old settings differ.(5)
If this seems strange, consider the case of a machine that has been built to detect bombs. In the context in which it is first used -- to single out bombs in luggage at airports before the advent of plastic explosives -- it works extremely well. But later on, as non-metallic bombs are placed in luggage, it becomes ineffective. The design of the machine is such that in one setting it functions well, but in another setting it fails. In a sense, it becomes defective: for it no longer succeeds in doing what it was designed to do. Still the hardware continues to operate smoothly. There are no mechanical breakdowns in the running of the machine. Given a suitcase with a metallic bomb in it, the machine responds to it effectively just as it did when it was first introduced. In a second sense of the term 'defective', then, the machine is not defective.
Those who say that after 50 years on Inverted Earth, Oscar's experiences of the clear sky come to represent it as yellow and who also accept representationalism for the phenomenal character of color experiences can plausibly claim that Oscar, in reporting that the sky looks the same color as it did the day before, goes wrong somewhere along the line. Oscar's memory leads him astray. Such a position in no way threatens privileged access. For insofar as we have privileged access to the contents of our mental states, it pertains only to our present states, not to our past ones.(6)
Block says that this appeal to the externalist character of memory content really clouds his main point, which is that it is a necessary feature of phenomenal character that big, fast changes in it can always be noticed. I agree with Block that this claim is true relative to a single context, a single external setting. But in switching cases, the context shifts without the subject noticing. And here it seems to me not obviously true that large changes in the phenomenal character of experiences are always accessible from the first-person perspective (any more than large changes in the contents of thoughts are always accessible in switching cases).
Shoemaker accepts a principle about phenomenal change that is very similar to Block's. He holds that it is constitutive of the notion of phenomenal character that one has introspective sensitivity to changes in it, whatever the source; indeed, by appealing to this principle (P), Shoemaker mounts a further argument against representationalism. However, (P), as stated, is vulnerable to the criticism that one might misremember an earlier phenomenal character --even one just a moment ago given an appropriate memory defect -- and hence one might not be aware introspectively that one's phenomenal state has changed. This difficulty can be handled by relativizing (P) to cases in which memory is not defective. But the cost of the relativization is that (P) cannot do the work Shoemaker wants it to do. For in travelling cases like that of Oscar, as already noted, it is not at all obvious that, in the externalist sense, memory is not defective.
So far, my concern has been simply to show that Block and Shoemaker have not refuted a version of representationalism about phenomenal character under which Oscar's sensory states for color change their representational contents with the move to Inverted Earth. This version of representationism is not my own, however. In my view, Oscar's blue-caused experiences on Inverted Earth do not represent blue, no matter how long he stays there. Shoemaker wonders how I can hold such a position, given my adherence to the causal covariation model of representation for sensory experiences. He also wonders how, consistent with my position on Oscar, I can permit Swampman to be the subject of sensory experiences. The explanation is as follows.
Anyone who thinks that representational content is ever determined (fully or partly) by causal covariation or tracking needs to index the covariation or tracking to some particular range of conditions. What matters intuitively is the tracking that would obtain were conditions normal or optimal. The key question thus becomes: what are optimal conditions?
Consider the nomic generalization that mixing an acid with a base produces a salt. This generalization has an implicit ceteris paribus clause that insulates it from a range of counter-examples (e.g., the fact that no salt is produced if certain other chemicals are added too or if the mixing beaker is smashed as the acid is combined with the base). In such cases, ceteris is not paribus. External factors interfere. Conditions are not optimal.
Understanding optimal conditions in this way, it seems clear that optimal conditions do not obtain for Oscar on Inverted Earth; for the Inverted Earth story essentially involves an artificial intervention in the operation of certain transducers. Inverting lenses are placed in the eyes of the traveller.(7) These lenses reverse the way in which the light input is processed. Intuitively, the lenses deceive the traveller (in Block's 1990 version of the story) so that when he first arrives, he has false beliefs on the basis of the phenomenal character of his visual experiences. He believes that the clear sky is blue, when really it is yellow. Of course, through time the traveller's beliefs adjust. But no matter how long he stays, it remains the case that the scientists from Inverted Earth have tampered with his visual transducers. Their operation is altered by the insertion of the lenses and, at no later time, is the system restored to its initial, natural state. The insertion of the lenses interferes with the operation of the sensory transducers. Accordingly, the transduction process is not in itself normal or optimal.
This is true not just for Oscar, where the insertion of the lenses prevents his visual transducers from functioning as they were designed to do, but also for his swamp duplicate. In the latter case, there is still outside interference. Admittedly, Swamp Oscar functions well after the interference in his new environment, but intuitively the lenses, considered in themselves, distort his color experiences. This distortion is masked by the color reversal on Inverted Earth, so that he cannot tell that he has been shifted to a very different environment, color-wise. But it is surely pre-theoretically correct to say that his transducers have been meddled with.
Intuitively, then, it is true of the traveller's sensory state, as he looks at the clear sky on Inverted Earth (after however many years), that had there been no interference, that phenomenal state would have been causally correlated (in him) with blue things. Accordingly, by the above causal covariation proposal, the traveller's sensory state continues to represent the clear sky as blue.(8)
Block suggests that, given my views, I cannot plausibly claim that the swamp traveller's visual experiences do not change their (nonconceptual) content with his immigration to Inverted Earth. For, according to Block, there is no relevant difference between the case of the swamp immigrant and the case of the swamp creature who is a native of Inverted Earth. This is not so, however. Where the swamp creature is a traveller, there is interference. Ex hypothesi, inverting lenses are placed in his eyes. So, ceteris is not paribus. Where the swamp creature is a native, assuming that he is not wearing lenses (or any other devices that function analogously), there is no interference. He is in optimal conditions. Thus, his state, when he looks at the clear sky represents yellow.(9)
There are other possible accounts of optimal conditions. One of these is teleological. Here optimal conditions are just design conditions. Where design conditions fail to obtain, then the setting is abnormal, no matter how long it obtains. For a speedometer used in a car with tires of the wrong size -- tires other than those it was designed to be used with -- the position of the pointer misrepresents the speed of the car, even if the speedometer is never hooked up to tires of the right size.
This account, when combined with representationalism about phenomenal character, entails that no swamp duplicate can ever have any experiences.(10) Accordingly, it is not a view I accept.
A third possible account of optimal conditions appeals to conditions of well-functioning. Intuitively, something can function well even if no design is involved. Take again the case of Swampman. Although Swampman is not human, he is self-sustaining, energy-using, capable of reproduction. By any reasonable standard, he is alive, notwithstanding the absence of an evolutionary history. Moreover, there are conditions under which he will flourish, and there are conditions under which he will not. If objects in the external environment trigger internal states in Swampman that elicit behavior inappropriate to those objects -- if, say, light rays bend in peculiar ways, thereby causing Swampman to misidentify very badly the shapes and sizes of things -- then he isn't going to last long. His needs won't be met; he won't easily survive the predations of others.(11)
Since this account (in combination with representationalism) entails that even the basic visual experiences of human-beings switch phenomenally with the move to Inverted Earth, it is also not one I accept.
A fourth position involves a combination of the last two. Optimal conditions become the relevant optimal conditions, and these will vary with the kind of creature or system we are dealing with. Where there is a design, optimal conditions are ones in which the creature or system was designed to operate.(12) Where there is no design, optimal conditions are, more broadly, ones in which the creature or system happens to be located or settled, if it is functioning well in that environment.
On this view, Oscar's basic visual experiences remain unaltered, but those of his swamp duplicate switch. If the lack of symmetry here strikes you as ill-advised, consider what you would say about the representational accuracy of an earth speedometer, taken to Mars, and used there on a Martian vehicle with a wheel-size very different from any found on earth, versus a swamp duplicate of such a speedometer that also happens to duplicate physically a Martian speedometer and is again taken to Mars. My intuition is that design wins out, where there is a design; but where no design is involved, setting and usage are the crucial factors. Thus the earth speedometer remains just that ('50' on the dial still means 50 mph, for example). In this case, Martian vehicle-speed is misrepresented. Not so, however, for the swamp-device. Now '50' on the dial represents whatever '50' represents on the accurate, native, Martian devices.
Of the four accounts, I prefer the first, since it accords with what is a strong intuition for many philosophers, namely that travelling, in itself, cannot change phenomenal character. However, I think that the fourth position is defensible too, once it is combined with a proper defence of externalism for memory to cover the swamp traveller's assertion that phenomenally nothing has changed.(13) It should be noted that the first view (like the fourth) is still externalist, since it allows the possibility of microphysical duplicates in different natural habitats having sensory states that track different items in those habitats under optimal conditions, thereby differing in the representational contents of their sensory states.
Shoemaker says that he finds the line I take with respect to Oscar -- that his color experiences on Inverted Earth will misrepresent the colors of things -- "utterly implausible". He comments (p. 000):
If addressed to the case of total, behaviorally undetectable
inversion, it has the implausible consequence that someone
could be misperceiving the colors over an entire lifetime even
if throughout that period the person makes correct color
judgements and has the normal ability to discriminate colors.
Leaving to one side the question of whether a total, undetectable, color qualia inversion really is possible, what perplexes me in this remark is why Shoemaker finds the view so implausible. After all, there are plenty of community-wide misperceptions or sensory illusions (e.g., the Muller-Lyer, the Hering Figure, Benham's Top).(14) Why, then, couldn't someone be misperceiving in the full-blooded inversion scenario? The difference is simply that here the illusion is so complete that it is not reflected in the person's judgements and behavior. Once sensory representations are sharply distinguished from conceptual representations, as on my account, there seems to me no obvious reason why there could not be cases in which the two are systematically mismatched.
Enough said on inversions. I come finally to the issue of transparency raised by Shoemaker. In my view, one who has a visual experience of red can be aware introspectively of the property of being red and thereby of the fact that she is having an experience with the property of being 'directed upon' or of red. I claim that intentional properties of this sort individuate finely. Just as one can have the concept of water without having the concept of H2O, so too one can have a sensory representation of red without having a sensory representation of such-and-such a complex spectral reflectance property (even if red is such a reflectance). For something might look red to one without looking disposed to reflect such-and-such a percentage of the incident light at such and such wavelengths. What distinguishes the one intentional property from the other, I suggest, is the manner of encoding of the relevant external property.
Whether or not this general suggestion about the root of the distinction between the above properties is correct, contra Shoemaker, there is no threat in my remarks here to the thesis of transparency. For if transparency holds, then the only properties we are aware of when we introspect our experiences are the external ones the experiences represent plus derivatively the intentional properties. If, for example, you have an experience of red and you turn your attention inwards, as you do so, you are just aware of the external quality, red, and of the fact that your state is 'directed' on red. You are not aware of the fact that your state is 'directed' on such-and-such a reflectance property; for there is no such fact. Your experience is not of a reflectance at all.
There is also no threat to externalism. States in our heads, by causally covarying with external properties and encoding those properties appropriately, acquire certain representational properties. And those representational properties, together with the role that their bearers play, endow the bearers with phenomenal character. Thus creatures that are identical inside the head can still differ representationally in the relevant way and thereby differ phenomenally.
Michael Tye
Temple University and King's College London
REFERENCES
Block, N. 1990 "Inverted Earth," Philosophical Perspectives, 4, J. Tomberlin, ed., 53-79.
Dretske, F. 1995 Naturalizing the Mind, Cambridge, Mass: The MIT Press, Bradford Books.
McLaughlin, B and Tye, M. forthcoming, "Is Content-Externalism Incompatible With Privileged Access?"
Tye, M. 1995 Ten Problems of Consciousness, Cambridge, Mass: the MIT Press, Bradford Books.
Tye, M. forthcoming (a) "Inverted Earth, Swampman, and Representationism," Philosophical Perspectives, 12, J. Tomberlin, ed.
Tye, M. forthcoming (b) "Externalism and Memory," Proceedings of the Aristotelian Society
Supplementary Volume
NOTES
1. 1. This is a little rough and ready. For example, I ignore certain complications that arise in connection with the individuational fine-grainedness of the relevant representational properties. For more here, see below, p. 000.
2. 2. For another related causal objection and reply, see Tye 1995, pp. 164-165.
3. 3. See Tye 1995, pp. 201-206.
4. For a general defense of externalism with respect to memory, see Tye forthcoming (b).
5. 5. I ignore here an important distinction between phenomenal and conceptual memory. For the full story, see Tye forthcoming (a).
6. For a discussion of privileged access and externalism, see McLaughlin and Tye forthcoming.
7. 7. Alternatively, wires are switched in Oscar's visual system by a transponder. In the version of the story in which the traveller is taken to a special room on Earth, inverting goggles are donned.
8. 8. Another tracking account of representational content that intuitively delivers the same result as the one above for the case of nonconceptual visual experiences and Inverted Earth is Jerry Fodor's asymmetric nomic dependence theory. I am indebted to Jesse Prinz for pointing this out to me. For more here, see Tye forthcoming (a).
9. 9. Of course, if the swamp native has been outfitted with lenses, then there is no difference from the case of the swamp immigrant.
10. 10. For a defense of such a position, see Dretske 1995.
11. 11. Block complains that on this account (given representationalism) Swampman's agony will be unreal in the case that he is located in a environment in which he fails to flourish. Not so. His states represent what they would track, were he in a setting in which he functioned well. The fact that his actual setting is one in which he functions badly is not to the point.
12. 12. What I say here is consistent with what I say about the case of tree-rings (see my book, pp. 100-101) even though trees are living systems with evolutionary histories and the biological function of tree rings is not to indicate tree age. For the number of tree rings tracks age (in deciduous trees) so long as the tree is in the environment in which nature designed it to operate and it is flourishing there in accordance with its natural design.
13. 13. I urged this in earlier correspondence with Block (as he notes). Both the first and the fourth line are explored further in Tye forthcoming (a). In the final version of that paper, I indicate a preference for the first alternative. In Tye 1995, I remain neutral between the two approaches.
14. 14. These are all cases in which our sensory systems are tricked. Going on the basis of what our senses tell us and nothing else, we arrive at false beliefs about what we are seeing.