Download The Debate over Conceptual and Nonconceptual Content in Perception - Prof. Jeffrey Speaks and more Study notes Introduction to Philosophy in PDF only on Docsity!
Nonconceptual content
phil 93507
Jeff Speaks
December 8, 2009
1 Two meanings of ‘nonconceptual content’......................... 1 2 Arguments for absolutely nonconceptual content..................... 2 3 Arguments for relatively nonconceptual content...................... 3 3.1 ‘Possessing a concept’................................. 3 3.2 The argument from fineness of grain........................ 3 3.3 Martin’s memory argument............................. 4 3.4 Learning or possession-explanation arguments................... 5 4 A conjecture about perceptual content and thought availability............. 5
One of the central debates in recent discussions of the philosophy of perception has been the debate about whether the contents of experience are conceptual or nonconceptual.
1 Two meanings of ‘nonconceptual content’
Participants in this debate have spent too little time trying to make clear what it would mean for perceptual content to be nonconceptual. One can isolate in the literature the following two interpretations:
A mental state has absolutely nonconceptual content iff that mental state has a dif- ferent kind of content than do beliefs, thoughts, etc.
A mental state of an agent A (at a time t) has relatively nonconceptual content iff the content of that mental state includes contents not grasped (possessed) by A at t.
It should be obvious that to say that content is nonconceptual in one sense is not to say that it is nonconceptual in the other; it is also, I think, clear that there are no trivial entailment relations between the two.
A nice example of conflation of these two senses of ‘nonconceptual content’ is provided by Pea- cocke (2001):
“While being reluctant to attribute concepts to the lower animals, many of us would also want to insist that the property of (say) representing a flat brown surface as
being at a certain distance from one can be common to the perceptions of humans and of lower animals.... If the lower animals do not have states with conceptual content, but some of their states have contents in common with human perceptions, it follows that some perceptual representational content is nonconceptual.”
This argument seems to run as follows:
(1) Animals possess no concepts. (2) The contents of the perceptions of animals are nonconceptual. (1) (3) Animals and human beings are related to the same kind of content in perception. (C) The contents of human perceptions are nonconceptual. (2,3)
If we interpret ‘nonconceptual content’ to mean ‘absolutely nonconceptual content,’ (2) does not follow from (1).
But if we interpret ‘nonconceptual’ in this argument to mean ‘relatively nonconceptual’, (C) does not follow from (2) and (3). For, on this disambiguation, the argument from (2) and (3) to (C) would run as follows: non-human animals do not possess the contents of their experiences; the contents of animal experiences are the same kinds of things as the contents of human experi- ences; therefore humans do not possess the contents of their experiences. But this is not a valid argument, because nothing rules out the possibility that both human beings and animals are related to the same kinds of contents in perception, but that human beings, and not animals, must possess or grasp those contents.
Peacocke’s version of the argument from sameness of animal and human perceptual content is thus an excellent example of the problems caused by conflating absolute and relative noncon- ceptual content. Each step in his argument may be validated by one of the interpretations of ‘nonconceptual’, but neither interpretation makes both steps valid.
2 Arguments for absolutely nonconceptual content
Most arguments which are presented as arguments for absolutely nonconceptual content fall well short of their target. A good example is the argument from ‘fineness of grain,’ the intuition behind which is well-stated by Richard Heck:
“Consider your current perceptual state — and now imagine what a complete de- scription of the way the world appears to you at this moment might be like. Surely a thousand words would hardly begin to do the job.”
Suppose that this is right; suppose that the contents of an experience or perceptual state are far more detailed and full of information than could be captured in a single thought, or even in a lifetime of thoughts. On the face of it, this hardly shows that the information given in perception is of a different kind than the information about the world represented in a belief; it shows, at most, that there’s more of it in the case of perception.
The only plausible sorts of arguments for absolutely nonconceptual contents would be an argu- ment that show that, e.g., perceptual content is Russellian, and a separate argument to show that the contents of thoughts are Fregean. We’ve already seen one of the former, and there are many plausible examples of the latter. But these belong more to the philosophy of language than the philosophy of perception.
left. Suppose that we do this ten times, and that the subject answers ‘yes’ five times, and ‘no’ five times. Then the subject has failed the above possession condition for a demonstrative concept referring to the color of the chip originally presented on his left, and then presented ten times by itself: he cannot consistently re-identify the property. But he is clearly able to distinguish the property in experience, as evidenced by his proficiency in distinguishing the two color chips when presented together. Conclusion: the color was part of the content of his experience, but was not the content of any demonstrative concept he possessed.
It seems to me that the case Kelly describes is a clear reductio of the possession condition he defends, rather than a convincing argument against the conceptualist position.
Consider for a moment what is involved in denying that the subject can have demonstrative thoughts about the color of the chip originally presented on his left, and later presented by itself. We must say that, although he is looking directly at the color of the chip, the subject is unable to have any demonstrative thoughts about involving the color of the chip at all. But this seems excessively strong. It seems clear that when I am in direct perceptual contact with a color property, I am able to have thoughts about that property whatever happens when I am presented with the property for re-identification at a later time. If forced to choose between the claim that one can always have thoughts involving a color to which one is attending — whether this is the color of an object in one’s environment or merely the color that one perceives such an object as having — and the claim that the possession condition for demonstratives given above is correct, the choice seems clear.
One way to press this intuition is to imagine the subject uttering a demonstrative when presented with the color chip by itself. The subject might say, for example, “I’m not sure whether that color [while pointing at the chip] is the same as the color of the chip on the left earlier.” It is natural to think that the subject understands the sentence he has just uttered, and grasps the thought it expresses. But it is also natural to think that the thought expressed by the sentence has a constituent corresponding to the demonstrative phrase “that color”, and that this constituent is or refers to the color of the chip. But saying these two things commits us to saying that, contra the possession condition, the subject grasps a demonstrative concept which picks out the color of the chip. The only alternatives seem to be to say either that the subject fails to understand the sentence he has just uttered, or that “that color”, as it appears in this sentence, lacks a meaning. But neither of these moves seem particularly plausible.
Indeed, there is a sense in which the thought-experiment, construed as an argument that the subject does not possess the relevant color concepts, is self-refuting. For it is surely a part of the assumed background of the case that when the examiner asks the subject whether the color chip presented alone is of the same color as the chip originally presented on his left, or whether that color is the same as that of the chip originally on the left, the subject understands the question. But how could a subject do this without grasping a concept of the relevant color?
This exhibits a common failing of arguments for relatively nonconceptual content: they impose implausibly strong constraints on concept possession. Possible diagnosis: confusion about what ‘concept possession’ could mean in this context.
3.3 Martin’s memory argument
Martin (1992) describes a case in which an agent, Mary, is playing a game with dice, one of which is 8-sided and one of which is 12-sided. But Mary does not distinguish between the two dice; she treats all dice with more than six sides as the same. Martin claims that, when playing with the dice, Mary may well lack the concept of a dodecahedron; and this seems plausible. But, he says, Mary might later, after acquiring the concept of a dodecahedron, recall her experience playing
the game, and realize that one of the dice was a dodecahedron. This, Martin says, indicates that Mary’s original experience presented the die as a dodecahedron; but, if Mary did not possess the concept of a dodecahedron, this must mean that the content of her experience was nonconceptual.
The key step in this argument is evidently the use of a conditional of the following form:
(A can infer p from remembering an experience had at t) → (p was part of the content of A’s experience at t)
But this principle is open to clear counterexamples. Consider, for example, the following case: I remember seeing an inscription on a plaque in my school of the words “Ad majorem dei gloriam”; not knowing Latin, I did not know what these words meant. Later on, I learn a bit of Latin, and, recalling my perception of this inscription, come to judge that the plaque had an inscription which meant “For the greater honor and glory of God.” So we can infer that it was part of the content of my original experience that the plaque had an inscription which meant “For the greater honor and glory of God.”
This inference is clearly fallacious. My original perception did not have this as a part of its content; I am able to infer that the inscription had this meaning because, since perceiving the plaque, I have acquired the ability to understand certain sentences of Latin. This can be turned into an objection to Martin’s example: why should we not say that Mary’s original experience did not have the concept of a dodecahedron as part of its content, but that Mary was able to infer later that her childhood game was played using a dodecahedron due to her acquisition of conceptual capacities parallel to my acquisition of the ability to understand Latin?
3.4 Learning or possession-explanation arguments
Basic idea: we need to explain how we can come to possess concepts like the color red; the only plausible explanation is in terms of perceptual experience. But if perception were not relatively nonconceptual, one would have to already possess the concept of redness in order to have a perceptual experience which represents something as red. (See Heck (2000) and, for a clearer version, Roskies (2008).)
Why this seems to be a bad kind of argument: it seems that we can explain concept possession in terms of experience even if having the experience is metaphysically sufficient for possessing the concept.
4 A conjecture about perceptual content and thought availability
Let’s suppose that the thesis that perceptual content is relatively nonconceptual is false, so that having an experience involving content C is sufficient for having thoughts involving C.
Then we can ask whether something stronger is true: whether it is the case that for any experience e whose content involves C, then it is possible to acquire the ability to have thoughts involving C for the first time by having e.
This is the converse of the Content/Availability principle, which we have already discussed. If both are true, this gives us necessary and sufficient conditions for a perceptual experience to have a content involving a certain concept (propositional constituent):
