Commentary on Alex Byrne

Abstract: 58 words
Main Text: 922 words
References: 49 words
Total Text: 1029 words

Perceptual Variation, Realism, and Relativization, Or: How I Learned To Stop Worrying And Love Variations in Color Vision

Jonathan Cohen
Department of Philosophy
University of California, San Diego
9500 Gilman Drive
La Jolla, CA 92093-0119


In many cases of variation in color vision, there is no non-arbitrary way of choosing between variants. Byrne and Hilbert insist that there is an unknown standard for choosing, while eliminativists claim that all the variants are erroneous. A better response relativizes colors to perceivers, thereby providing a color realism that avoids the need to choose between variants.

Byrne and Hilbert discuss variations in color vision in section 3.4 mainly in the context of blocking the eliminativist's argument from these phenomena to the conclusion that nothing is colored. I want to concede realism about color, but use the same phenomena to raise a distinct but related set of challenges for Byrne and Hilbert.

We can raise these challenges by reflecting on the case of variation Byrne and Hilbert adapt from (Hardin, 1993). You and your colleague view a range of Munsell chips under relevantly similar perceptual circumstances, but you disagree about which of the chips is unique green: chip C1 looks unique green to you, and fails to look unique green (say it looks bluish green) to your colleague, whereas chip C2 looks unique green to her but not to you. Now consider chip C1, and ask: is C1 unique green?

By hypothesis, consulting the perceivers (you and your colleague) will not provide a determinate answer, since they are divided over the question.1 What we need, then, is a standard for whether something is unique green that is independent of the reports of perceivers. Unfortunately, as Byrne and Hilbert point out, "in the color case, there is no such [independent] test" (section 3.4) to which we can turn.2

Byrne and Hilbert recognize that an independent standard is needed to answer our question about the chip, and that no such independent standard is available. But this does not alarm them; by way of analogy, they point out that even if we lack decisive evidence about the guilt of a particular suspect, this does not lead us to believe that no one murdered Professor Plum. Presumably, the intended force of the analogy is that, even if we lack an independent standard that would certify one of the ways C1 looks as the veridical representation of the chip's color, there may nonetheless be an unknown fact of the matter about which representation is veridical.

But the analogy is unconvincing because the background beliefs we bring to the inquiry are far less informative in the case at hand than they are in the case of Professor Plum. For our belief that Professor Plum has been murdered is sustained in the face of our lack of dispositive evidence about individual suspects only because of certain sorts of general background beliefs, and these are notably absent in the case of the missing standard for color perception.

In the case of Professor Plum, beliefs of two general kinds are relevant. 1. Our independently well-supported beliefs about how people move and behave imply that the good Professor would not have ended up in his present unhappy state (keeled over in the ballroom, knife protruding from his back) had some person or other not murdered him. 2. We have no trouble understanding how someone could be the murderer of Professor Plum without our having decisive evidence of his guilt. Here, belief 1 creates a presumptive prejudice to the effect that that Professor Plum was murdered, and belief 2 explains away potential counter-considerations engendered by our lack of evidence about specific individuals. This combination of beliefs, then, leads us to think that there is a fact of the matter about who murdered Plum, even if that fact of the matter is beyond our ken.

Contrast the case of the wanted independent standard for color perception. Here our general background beliefs both fail to establish a presumptive prejudice in favor of an independent but possibly unknown standard, and fail to override the counter-considerations engendered by our lack of evidence. Indeed, the failure of several hundred years of systematic efforts directed at articulating standards of this kind establishes a presumptive case against their existence.3 As such, Byrne's and Hilbert's view that there is an epistemically unavailable standard strikes me as a piece of unwarranted optimism.

Suppose that, as I suspect, there is no well-motivated independent standard to arbitrate between the two representations of C1's color. Must we, then, endorse color eliminativism? Like Byrne and Hilbert, I hope to avoid this outcome: eliminativism amounts to such a radical revision of our pre-theoretical views about the world that it should be regarded as a position of last resort.4

Luckily, there are non-eliminativist ways of accepting the absence of a perceiver-independent standard for C1's color. Namely, we can hold that the alternative representations of C1's color (the way it looks to you, the way it looks to your colleague) are both veridical. There are a number of ways of fleshing out this suggestion, but one of the most popular is to construe colors as relativized to perceivers.5 In the case at hand, this would amount to saying that C1 exemplifies both of these color properties: unique green to you, and bluish green to your colleague. This view both frees us from having to answer the otherwise pressing but apparently unanswerable question whether C1 is unique green or not, and explains why past efforts to answer it have failed (namely, on this view, there is no non-arbitrary reason for preferring either choice over the other). This is all to the good: hard cases make bad law.

The view I am recommending is a species of realism in that it insists that colors are real (not merely apparent) properties of objects.6 However, unlike Byrne's and Hilbert's preferred form of realism, it accomodates the data about perceptual variation without requiring either hard choices or unwarranted optimism. As such, I believe this view is a more attractive alternative for those in the market for a realist account of color.


1. The assumption that there are only two such observers is inessential; one can easily expand the number of perceivers without thereby generating a consensus about C1's color.

2. Byrne and Hilbert blame the lack of an independent test partly on "the fact that colors are not perceived by any other sensory modality". But I do not see why a further sense modality for perceiving colors would necessarily provide the sort of independent test we need. Suppose we had such an additional modality for perceiving colors. (i) That modality might not lend support to either of the conflicting visual representations of color; (ii) that modality might lend support to both of the conflicting representations. In either case, it would not arbitrate disputes among visual representations in the way that Byrne and Hilbert imagine.

3. The history of these efforts is recounted in (Hardin, 1993), 67-82; see also (Cohen, 2003).

4. As usual, Hume is eloquent on this point: "Philosophy scarce ever advances a greater paradox in the eyes of the people, than when it affirms that snow is neither cold nor white: fire hot nor red" (letter to Hugh Blair of 4 July 1762, printed in Mind, October 1986).

5. The dispositionalist view Byrne and Hilbert consider (and reject as unmotivated) in section 2.2 is one account of this type, although there are a number of others. Consequently, the point I am pressing is one way of providing the motivation for such views that Byrne and Hilbert think is lacking.

6. A number of authors have objected that such views unacceptably preclude erroneous color attributions (for example, see (Hilbert, 1987), p.88, (Matthen, 2001)). For a response to this objection, see (Cohen, 2000), (Cohen, 2003).


Cohen, J. (2003) Color properties and color ascriptions: A relationalist manifesto. Under review. (

Hardin, C. L. (1993) Color for philosophers: Unweaving the rainbow (expanded edition). Hackett.

Matthen, M. (2001) Our Knowledge of Colour, in Macintosh (ed.), Naturalism, Evolution, and Intentionality Canadian Journal of Philosophy, Supplementary volume 27, 2001. (