theories of color
 
 
Theories of color make proposals about the constituting nature of the colors that we attribute to physical objects in visual perception. The most common proposals are that these colors are mental properties of visual states (subjectivism), they are physical properties of physical objects (physicalism), or they are dispositions of physical objects to produce visual states of color (dispositionalism). See theories of color perception.
 

Details:
 
Introduction
 
The problem of the constituting nature of color is typically put in terms of the following question about the intentional content of visual states of color: what sorts of properties are the colors that we attribute to physical objects in virtue of our visual states of color? Are attributed colors, for example, physical properties of physical objects? Or are they other sorts of properties, such as mental properties of visual states themselves, or dispositions of physical objects to produce visual states of color?
 
The problem of the constituting nature of color has been largely motivated by concerns about how color fits into a scientific description of the world. In the seventeenth century, Galileo and Newton undertook a comprehensive description of nature in terms of mathematical physics. Physics requires that if colors are properties of physical objects, then they are describable in physical terms.
 
However, due to problems with understanding how colors can be describable in physical terms, many philosophers and scientists since the time of Galileo have come to reject the claim that colors are physical properties of physical objects. Most theorists assume that if we reject this claim, then the nature of colors must be characterized--at least in part--in terms of our mental nature.
 
 
Visual states of color
 
In addressing the problem of the nature of the color we attribute to physical objects in visual states we must consider the following question: how do we characterize the nature of these visual states (also commonly called visual experiences).
 
Some theorists claim visual experiences of color have both intentional properties and mental qualitative properties which I'll call mental colors; other theorists claim they have intentional properties but don't have mental colors.
 
Those who claim that color experiences have both intentional properties and mental colors include some physicalists (Lewis, Shoemaker), most dispositionalists (currently, Peacocke and McDowell; formerly, McGinn and Johnston), and all subjectivists.
 
Those who claim that visual experiences of color have intentional properties but do not have mental colors include some dispositionalists (Smart [1961]) and most physicalists (Smart [1975], Armstrong, Hilbert [1987 and 1992], and Byrne and Hilbert).
 
 
Mental color
 
Mental colors are mental qualitative properties of color experiences. Amongst those who characterize color experiences as having mental colors, there's a controversy as to how to describe mental colors. The main divide is between those who claim that mental colors are what it's like for a perceiver to be conscious of color (Peacocke, McGinn, Shoemaker, Block), and those who describe mental colors in terms of causal relations, for example, in terms of functionally described processes of our visual systems (Hardin, Clark, McGilvray, Lewis).
 
Proponents of the claim that mental colors are what it's like for a perceiver to be conscious of color hold that mental colors cannot be described in terms of causal relations. Mental colors described as what it's like to be conscious of color are typically called qualia.
 
 
Subjectivism
 
According to subjectivism, the colors that we attribute to physical objects in color experiences are mental colors, which are mental qualitative properties of visual states themselves. Thus subjectivism claims that physical objects are colorless. (Because subjectivism denies the existence of colored physical objects, it is sometimes called color eliminativism.)
 
Two kinds of argument are offered in support of subjectivism. One kind stresses epistemological and phenomenological considerations; the other is founded on evidence from color science, in particular, psychophysics and neurophysiology.
 
Boghossian and Velleman provide arguments of the first kind. In their 1991, they claim against physicalism that qualitative similarity relations among colors, such as that orange is more similar to red than it is to blue, specify essential features of determinate colors. Furthermore, they assert, ordinary experience provides access to these essential features. However, if physicalism were correct, these essential features would be relations among physical properties. They conclude that physicalism is false because ordinary experience doesn't provide access to relations among the relevant physical properties (namely, relations among surface reflectances). Furthermore, in their 1989 they claim against dispositionalism that colors cannot be dispositions to produce visual states of color, for colors don't look like dispositions in our ordinary experience.
 
However, disputable epistemological assumptions underlie both of these arguments. Against dispositionalism they assume the very controversial claim that ordinary experience provides access to the constituting nature of color, a claim that Johnston calls Revelation. Against physicalism they assume a claim that physicalists deny, namely, that the qualitative similarity relations among colors specify essential features of determinate colors.
 
Hardin and McGilvray provide arguments for subjectivism founded on evidence from psychophysics and neurophysiology. Psychophysics indicates that objects with different surface reflectances can produce perceptions of the same determinate color (such as teal); such physically distinct objects are called metamers. Metamerism is explained in terms of neurophysiology, and this neurophysiological explanation shows that for each determinate color, indefinitely many different surface reflectances look that color.
 
Also, functionally described processes of our visual systems called opponent processes explain the qualitative similarity relations among perceived colors (such as that orange is qualitatively more similar to red than it is to blue) (see Hurvich). Furthermore, there is no range of physical properties of objects the members of which are related intrinsically in ways that correspond with the qualitative similarity relations. (Members of a range of properties are related intrinsically so long as these relations are don't hold in virtue of relations between members of the range and members of a distinct range of properties.) Rather, physical properties of objects are related extrinsically in these ways, in virtue of relations between these physical properties and properties of the human visual system.
 
Thus the qualitative identity and similarity relations among perceived colors are explained in terms of neurophysiology. Because Hardin and McGilvray assume that on any tenable proposal of the constituting nature of color, this constituting nature must explain these qualitative relations, they conclude color science shows that perceived colors aren't physical properties of physical objects. Rather, Hardin and McGilvray claim colors are neural events of our visual systems.
 
However, physicalists object to Hardin's and McGilvray's assumption. Physicalists propose that we distinguish between the neural properties that explain the qualitative relations among perceived colors and the perceived colors themselves, which are physical properties of objects.
 
 
Physicalism
 
Physicalism claims that colors we attribute to physical objects in color experiences are physical properties of those objects, or as I'll call them physical colors. There are several versions of physicalism.
 
According to one type of physicalism held by Hilbert (1987 and 1992), physical colors are surface reflectances, and are individuated as finely as surface reflectances. However, objects with different surface reflectances can be perceived as the same determinate color (such physically distinct objects are called metamers). Thus perceived colors are indeterminate with respect to physical color. And because color science shows that metamerism is explained in terms of the neurophysiology of the human visual system, Hilbert calls perceived colors anthropocentric colors. Nevertheless, Hilbert claims that perceived colors are objective, nonrelational properties of physical objects.
 
On another type of physicalism held by Smart, Armstrong, and Lewis, no distinction is drawn between physical colors and perceived colors. Physical colors just are the colors perceived by standard human color perceivers in standard viewing conditions. Since the neurophysiological explanation of metamerism shows that for each determinate perceived color indefinitely many different surface reflectances look that color, physical colors are indefinitely large disjunctions of surface reflectances.
 
Shoemaker proposes a third type of physicalism. He draws a distinction between the intentional contents and the qualitative contents of color experiences. He claims that physical colors are included in intentional contents. But, he holds, these physical properties aren't perceived colors--the colors included in qualitative contents. Rather, perceived colors are relations between physical colors and visual experiences with color qualia (which are mental qualitative properties which determine qualitative contents). Of course, Shoemaker's proposal depends on the controversial claim that visual states have qualia.
 
Hilbert (1987 and 1992) argues for physicalism largely on the basis of the visual effect called color constancy. Color constancy is an effect in which the color a physical object looks remains fairly constant despite changes in the light illuminating the object. As it turns out, the constant colors that we perceive physical surfaces as having are correlated with surface reflectance. Hilbert (1992) argues that color constancy shows that the biological function of color vision is to detect physical object surfaces by way of surface reflectance, and that, therefore, the colors we attribute to physical objects in visual states of color are surface reflectances. However, some dispute Hilbert's characterization of the biological function of color vision in physical terms, and rather characterize this function in ecological terms. For example, Thompson and Hatfield object to Hilbert's physicalism from the standpoint of an ecological approach to color perception, and claim that perceived colors are relations between objects and color experiences.
 
The Smart/Armstrong/Lewis view holds that descriptions of relations between physical objects and color experiences of standard perceivers in standard viewing conditions merely serve to fix the reference of color terms (for a similar claim, see Kripke). However, the claim that perceived colors are disjunctions of physical properties is controversial. But crucial to a defense of the this view is a distinction between properties and universals. Whereas Armstrong denies that universals can be disjunctive, he allows that properties can be disjunctive, where disjunctive properties can be explained in terms of ranges of universals. And perceived colors can be explained in terms of psychophysical laws that quantify the relations between ranges of physical properties of objects and visual states of color.
 
Since visual science indicates that the qualitative identity and similarity relations among perceived colors are explained in terms of neurophysiology, independently of the physical colors themselves, physicalists claim that these qualitative relations don't specify essential features of determinate perceived colors. However, this claim is controversial (Boghossian and Velleman [1991], as well as Hardin and McGilvray assume that it's false; Johnston argues against it).
 
 
Dispositionalism
 
According to dispositionalism, the colors we attribute to physical objects in color experiences are dispositions of physical objects to produce perceptual responses. Different versions of dispositionalism characterize perceptual responses differently. Smart (1961) characterizes perceptual responses nonqualitatively, namely, in terms of the discriminatory behavior of perceivers. But the most common version of dispositionalism characterizes perceptual responses in terms of visual experiences with color qualia, in particular such visual experiences of standard perceivers in standard viewing conditions (currently, Peacocke and McDowell; formerly, McGinn and Johnston).
 
If the colors we attribute to objects are dispositions to produce visual experiences with color qualia, then these colors are constituted by relations between some physical property or other of objects and visual experiences with color qualia. According to current versions of this view, even though the colors we attribute to objects are in part constituted by color qualia, the colors we attribute to objects aren't themselves color qualia. Rather, color qualia determine what it's like to experience physical properties of physical objects as colors, and colors we attribute to physical objects are dispositional properties of physical objects.
 
Proponents of this version of dispositionalism argue that the colors we attribute to objects are in part constituted by color qualia on epistemological grounds, in particular on the basis of the claim that ordinary experience provides access to essential features of color. For example, Johnston claims that qualitative similarity relations among color specify essential features of determinate colors to which ordinary experience provides access, and supports this claim with considerations about skepticism. However, physicalists reject this argument, and it remains controversial as to whether qualitative similarity relations specify essential features of determinate colors at all.
 
 
Peter Ross