The data problem for color objectivism.

Are colors objective or subjective? Are they properties, processes, or events of the physical world or, instead, of the perceiving subject? This question has been debated at least since the time of Galileo and remains unsettled to this day. Evidence from computational and psychophysical studies of vision has not decided the issue, with both objectivists and subjectivists claiming that the evidence to date is in their favor. In his article, ‘‘The Location Problem for Color Subjectivism,’’ Peter Ross proposes that color subjectivists make two mistakes, one logical and one empirical (Ross, 2001). The logical mistake is an unwitting commitment to a philosophical assumption he calls the ‘‘corresponding category constraint.’’ The empirical mistake is the failure of any subjectivist theory to properly account for recent data on sensed locations. Ross concludes that color subjectivism is untenable and proposes instead that disjunctive physicalism is the most viable remaining candidate. Here I argue that the data on sensed locations are different than Ross claims and that once the data are properly understood they pose no obstacle to adverbial-subjectivist theories. Then I argue that disjunctive physicalist accounts of color need the corresponding category constraint no less than subjectivist accounts or else they are devoid of empirical support. Finally I raise an empirical challenge for color subjectivists and a separate empirical challenge for disjunctive physicalists. Ross raises the problem of sensed locations as an empirical obstacle to acceptance of adverbial-subjectivist theories. According to such theories, sensing is not a relation between a perceiver and sense data or other objects, but rather a nonrelational way that a perceiver is. When a perceiver sees a red square he sees redly and squarely; when he sees a green circle he sees greenly and roundly. Seeing redly, greenly, squarely, and roundly describe kinds of mental processes or events of the perceiver. The problem with this theory, according to Ross, is that colors have sensed locations, and the theory cannot account for the empirical data on the binding of colors with locations in the visual field. We can, for instance, see a red circle inside a green square and distinguish this from a green square inside a red circle. The adverbialist must provide a nonrelational account of sensed locations to handle such cases, and the most straightforward way is to describe the visual field as an array of repeatable sensory events to which sensory adverbs, such as redly, can apply. Then he can describe one sensory event as redly and roundly and insidely and another as greenly