Dispositions and the central problem of color

Philosophical Studies

Volumn 93, Issue 1, 1999, Pages 21-44

  Gold, I A N ^a  

^a AUSTRALIAN NATIONAL UNIVERSITY   (Australia)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 33751110646     PISSN: 00318116     EISSN: 15730883     Source Type: Journal    
DOI: 10.1023/A:1004215111799     Document Type: Article

References (25)

1. The phrase 'color as we see it' derives from John Mackie, Problems from Locke (Oxford: Clarendon, 1976).
2. Mark Johnston, "How to Speak of the Colors," Philosophical Studies 68 (1992): 221-263.
3. See Frank Jackson, "The Primary Quality View of Color," in James Tomberlin, ed., Philosophical Perspectives 10/11 (Atascadero: Ridgeview, forthcoming).
4. See, for example, David Hubert, Color and Color Perception (Stanford: Centre for the Study of Language and Information, 1987).
5. See Wilfrid Sellars, "Science and the Manifest Image of Man," in Sellars, Science, Perception, and Reality (New York: Routledge and Kegan Paul, 1963). Sellars thought that it was possible to defend color realism from this challenge, and he was himself an unusual sort of color realist.
6. See C. L. Hardin, Color for Philosophers (Indianapolis: Hackett, 1988).
7. See Leo M. Hurvich, Color Vision (Sunderland: Sinauer Associates, 1981).
8. The sense of 'appear' in this context is a technical one. In brief, it is possible to use green-colored light to "cancel" red-colored light. That is, the red quality of the appearance can be diminished by adding green-colored light and vice versa. See Leo Hurvich Color Vision.
9. See Colin McGinn, The Subjective View (Oxford: Clarendon, 1983).
10. For this definition see Frank Jackson, Elizabeth W. Prior and Robert Pargetter, "Three Theses About Dispositions,"American Philosophical Quarterly 19 (1982): 251-57. This matter and many others concerning dispositions are discussed in Elizabeth W. Prior, Dispositions (Aberdeen: Aberdeen University, 1982). There is an issue concerning Johnston's conception of dispositions which may be problematic for his view but which I cannot consider at length in this paper. In brief, the problem is that while Johnston is concerned with colors as constituted dispositions, he occasionally speaks of dispositions as relational properties rather than higherorder properties. For example, he claims that response-dispositionalism cannot save the core belief of revelation because colors fail to look like dispositions; that is, they "do not appear to be relational properties and hence do not appear to be dispositions to look colored" (p. 226). Johnston further claims that responsedispositionalism manages to achieve a kind of modified revelation in cases of highlight colors - the shimmering colors one sees, for example, on compact disks. He argues that because shimmering colors change with a change in relation to the observer or the light, they can be construed as colors appearing in their true guise as dispositions. But highlight colors can only support a form of revelation for dispositionalism if dispositions are relational properties. For colors to appear in their true guise as constituted dispositions, they would have to look like the properties of having some intrinsic properties which ... whatever that might mean. Even if dispositions were relations, however, the strategy of appealing to the "look" of some color manifestations to support a dispositional theory could work against dispositionalism in other cases. For example, consider the Hermann grid illusion: (Figure Presented) The Hermann grid illusion Illusory colored circles appear in the intersections of the grid except for the intersection one is looking at. (Here the circles are grey, but it is possible to produce chromatic versions of the illusion as well.) For this reason, as the eyes move to a new location on the grid the colored circle in the intersection at which one was previously looking does seem to "come on" in a way that makes the colors appear as a result of a relation between the grid and the position of the eyes. But it is precisely this relational feel that moves us to say that the colors are illusory. If colors are dispositions, however, one might think that they ought to seem more real, rather than less, when they appear in their true guise as dispositions. I owe this idea to Natalie Stoljar.
11. Martin developed the notion of finkish dispositions in 1957. For a recent discussion of his view see C.B. Martin, "Dispositions and Conditionals," Philosophical Quarterly 44 (1994): 1-8.
12. See also Robert Shope, "The Conditional Fallacy in Contemporary Philosophy," Journal of Philosophy 75 (1978): 397-413.
13. Strictly speaking, a film color is a color stimulus in which all information about the material composition and spatial location of the stimulus is eliminated. Any colored stimulus - including colored objects - can be viewed in the film color mode by observing the surface at a distance through a hole placed in a screen that is also at some distance from the viewer. See Jacob Beck, Surface Color Perception (Ithaca: Cornell University, 1972), pp. 17-18.
14. The concept of spectral radiant power distribution is also crucial to the definition of 'metamers' which plays an important role in the philosophy of color. Informally, two physically different objects or lights are metamers if they are indistinguishable in color to some observer in some set of circumstances. Metamerism is often thought to be caused by the fact that there are, according to Kurt Nassau, The Physics and Chemistry of Color (New York: Wiley, 1983), fifteen distinct microphysical causes of color (requiring five distinct branches of physics for explanation), but, in principle, not all cases of physically different objects that are color-indistinguishable are technically metameric. Strictly, two color stimuli are metamers just in case they have the same tristimulus values but different spectral radiant power distributions. (Tristimulus values refer to the quantities of the three reference lights used to match a given color stimulus and thus characterize it quantitatively.) Metamers are thus indistinguishable in color appearance but send different patterns of light to the eye. This difference in light pattern may come about in the case of two light emitters, or it may come about because of an appropriate combination of reflectance functions and incident light. Metamers can thus be either colored lights or colored objects. Two objects that have different microphysical properties (that are causally relevant to the way in which the object interacts with light) but produce identical perceived colors in some lighting conditions because they have identical reflectance functions do not count as metamers, therefore, because they produce the same spectral radiant power distribution in those conditions. On the other hand, objects with different microphysical properties as well as different reflectance functions that are judged identical with respect to color in some particular illumination do count as metamers. In short, then, identity of color appearance together with difference in microphysical constitution is not sufficient for metamerism.
15. See Gunter Wyszecki and W.S. Stiles, Color Science, second edition (New York: John Wiley and Sons, 1982), pp. 690-691 and 723-724.
16. Nor should we say that the blue color of the sky is really the color of the atmospheric particles that cause the refraction of the sunlight so that predominantly short wavelength light reaches the surface of the earth. The reason is this: consider a transparent material that refracts light in a manner similar to the particles in the atmosphere. This hypothetical material would look like glass but would have a different refractive index. Now if we were to direct the full spectral distribution of sunlight through the glass and then look at the light that came through, we would see a bluish color. But by hypothesis the refractive material is not blue but transparent. Since the particles in the atmosphere act like the refractive material does, we have no more reason to say that the particles are blue than we do to say that the refractive material is blue. I am grateful to John Bigelow and Phillip Gerrans for pointing out this issue to me.
17. Indeed this possibility (with a supernatural twist) generated considerable sceptical worries in some medieval philosophers.
18. Not quite: the number of color instances in the environment is twice the number we take there to be less the number of instances of film and, possibly, luminous colors.
19. What I say about these mixtures is not meant to be taken as fact.
20. Metaphysics IX: 6, 1048b.
21. Complete actions are those of which one can say "I have ø-ed" at every moment at which one ø-s. Thus, walking along the road between Canberra and Sydney is a complete action because I can say "I have walked along the road between Canberra and Sydney" at every moment that I walk along the road between Canberra and Sydney. But walking from Canberra to Sydney is an incomplete action because I cannot say "I have walked from Canberra to Sydney" until I arrive in Sydney.
22. Aristotle actually contrasts walking from/I to B with seeing, but I do not want to confuse the issue with a visual example.
23. Johnston rejects the idea that any theory of color can preserve the core belief of revelation, but in claiming that our beliefs about the relations among the colors are incorrigible he is, in effect, endorsing a restricted form of revelation, namely, one that applies to color relations. The unity relations must be fully revealed in color experience, according to response-dispositionalism, which is why we cannot be mistaken about them.
24. Hewitt D. Crane and Thomas P. Piantanida, "On Seeing Reddish Green and Yellowish Blue," Science 221 (1983): 1078-1080.
25. i is a shade of green. (3) Suppose Y is maximally similar to YG. (4) Then Y is a shade of green. That is, when we form a never before seen mixture, we might revise our categorization of shade Y (The fact that colors are perceived categorically - that is, that we have a tendency to lump together a large group of colors into classes - would presumably prevent a slippery slope in which it might be possible to reclassify every shade so that each shade turns out to be a shade of, say, green. [See Marc H. Bornstein, "Perceptual Categories in Vision and Audition," in Stevan Harnad, ed., Categorical Perception (Cambridge: Cambridge University, 1987), pp. 287-300.] Now if we replace the shade G with a shade of blue in the argument above, as the Crane and Piantanida experiment suggests is possible, then Y could turn out to be a shade of blue. This of course seems unlikely in the case of canary yellow because it is such a "paradigmatic" shade of yellow that we wouldn't revise. But the fact that any yellow could be reclassified as a shade of blue is a radical break with common sense but something that science (in the form of the Crane and Piantanida experiment) may have discovered. I am grateful to Steve Gardener, David Lewis, and George Molnar for helping me to understand this issue better. 24 Johnston himself argues that the realist knows that canary yellow could not be a shade of blue only in virtue of the claims of opponent color theory. But, Johnston argues, it seems absurd to say that this fact about color can only be known via a theory and not simply in virtue of color experience.