Friday, May 21, 2010

The Varieties of Consciousness Conceptualism

This is Part 2 of the serialization of the long version of my paper, "Color-Consciousness Conceptualism," the short version of which appeared in the Second Annual Conference of Consciousness Online. This post contains sections 2 and 3 of the paper.

2. First-order and Higher-Order theories of Consciousness Conceptualism
Many debates surrounding conceptualism are cast in terms of a construal of conceptualism that we can characterize as the exhaustion thesis: conscious perceptual states have conceptual content, and the mental aspects distinguishing various perceptual states, aspects such as the phenomenal character or sensory qualities of the states, are exhausted by these conceptual contents. Focusing on conscious experience of color, the exhaustion thesis holds that the difference between a conscious experience of red and a conscious experience of blue just is the difference constituted by deploying the concept red in the one experience and the concept of blue in the other.[7]

Another kind of view that merits considering as a fellow traveler but is not as strong as the exhaustion thesis is what I’ll call the matching thesis. We might state the matching thesis as that there must be as many concepts deployed as there are mental aspects distinguishing perceptual states, but these concepts do not exhaust those aspects. The matching thesis thus allows that there may be nonconceptual aspects to conscious experience (Bengson et al., in press; Rosenthal, 2005). Put in terms of the conscious experience of color, the matching thesis has the same conceptual requirements as the exhaustion thesis: it is required by one experience’s being of blue and the other of red that there be a deployment of a concept of blue in the one and of red in the other. But unlike the exhaustion thesis, the matching thesis does not hold that the respective deployments of the concepts of red and of blue are the sole constituents of the mental differences between the two experiences. The matching thesis allows that there may additionally be nonconceptual differences, differences, for instance, at the level of sensations, impressions, or sensory quality.
Many arguments in the literature designed to attack conceptualism serve as attacks on both the exhaustion thesis and the matching thesis. Arguments hinging on experiential fineness of grain, for example, call into question whether there are as many concepts at the disposal of the perceiver as there are colors that the perceiver may consciously experience. Clearly such an argument is indifferent to the difference between the exhaustion and matching theses.

Though I don’t take the difference between exhaustion and matching to matter much for present purposes, I myself am inclined toward an exhaustion version of conceptualism and in what follows, I will frequently spell out various conceptualist points along exhaustion-thesis lines.[8]

3. Gareth Evans and fineness of grain
Much of the literature on color and conceptual content revolves around Gareth Evans’s rhetorical question: ‘‘Do we really understand the proposal that we have as many colour concepts as there are shades of colour that we can sensibly discriminate?’’ (Evans, 1982, p. 229). Setting aside the portion of the question regarding what it is that we understand, a portion seeming to invite viewing this as a matter of what we can conceive about the relations between concepts and sensible discrimination, Evans’s question seems to be a straightforwardly empirical question: does the number of discriminable colors exceed the number of color concepts? Further, if we make certain assumptions about what to count as discriminable colors and concepts for them, then the answer to this empirical question is already at hand.

Assume, on the experience side of things, that the number of colors discriminable in standard psychophysical tests is one and the same as the number of colors that may be consciously experienced. Assume, on the concept side of things, that by “colour concepts,” Evans intends what we can call lexical concepts of colors, concepts corresponding to individual color words, as opposed to phrasal concepts of colors, concepts corresponding to multi-word phrases. Further, assume that the number of lexical concepts a perceiver has is one and the same as the number of high-frequency (frequently used) individual color words in the perceiver’s native language. Such words are basic monolexemic color terms, and exclude terms such as “light blue,” which is not monolexemic, and “azure,” which is low-frequency. Such assumptions lead to experienced colors numbering around ten million and, for native speakers of English, the number of lexical concepts at a mere eleven.[9] The difference between the numbers is so vast that we can plausibly credit typical English-speaking perceivers with already having an intuitive grasp of such a difference. This intuitive grasp perhaps underlies so many philosophers finding intuitive the claim that we experience more colors than we have concepts for.

This difference in estimated numbers of color concepts and experienced colors depends on certain assumptions. However, if we change our assumptions on the concept side of things, then it is not at all clear that the gulf will be so vast between the numbers of concepts and the number of colors experienced. In particular, if we allow phrasal as well as lexical concepts, then we allow for a combinatoric conceptual wealth that may very well match the wealth of experienced colors. If the nonconceptualists are going to have a hope of defeating this combinatoric strategy on empirical grounds, they are going to need something much stronger than Evans’s rhetorical question. It is to such a stronger argument that I now turn.

[7] The differences relevant to the present example are intramodal differences, as would be expected in discussions of color experience. The larger issue of how and whether various conceptualists account for intermodal differences is too large for discussion in the present paper.

[8] In particular, I hold phenomenal character to be identical to a certain kind of conceptual content (Mandik, forthcoming).

[9] See (Hardin, 1988b, pp. 226-227), where he cites (Judd & Wyszecki, 1963, p. 359) regarding the number of discriminable colors and (Berlin & Kay, 1969) regarding the number of basic color terms in English.