Gert, Joshua - What Colors Could Not Be: An Argument For Color Primitivism

09/26/2008

Journal of Philosophy, March 2008

This is a paper that suggests that any way of reducing colors to physical or dispositional properties will give more modal properties to color than they should have. Author begins the paper with an analogy to Benacerraf's work on numbers, arguing they are primitive and not reducible to sets. The problem with reduction that Benacerraf reveals is that numbers are reducible to more than one system of sets. To pick between one seems ad hoc, but regardless of this, once picking one there are certain weird questions about numbers that have weird answers e.g.: 'does "3" contain "1" and "2"?' Yes? No? Either answer is weird because the question is 'ill-formed'. However, reducing numbers to sets 'forces' an answer to such a question, thereby giving numbers more properties than they should have. The solution is to make numbers primitive. Author takes this method of argument as an analogy to color.

Author wants color to be fully explained from within the domain of color. It is certainly possible that there will be physical-reductive facts (about surface reflectance, visual perception) that co-instantiate with color, but author argues this is like the set-theoretic account of numbers; they don't reduce to them. (pg130-1) Author takes the fact that there are disputes between color realists as analogous to the different methods of reducing numbers to sets. (pg132-3)

Author's positive account of color primitivism is that color terms do not rigidly designate any properties, nor do they non-rigidly designate any properties either. Instead, color rigidly designates colors. (pg136) If color terms designated anything other than that, then there would be ill-posed modal questions about colors that have truth values, which author argues needs to be avoided. An objection to this view is why you can't extend it by analogy to all property terms, like e.g. shape. Author replies that there are other observational/physical conditions for most other properties can that yield facts-of-the-matter in modal settings (pg139-140).

In the next section author takes Mark Johnston's discussion about 5 core beliefs about color. Johnston argues that these 5 core beliefs are inconsistent and therefore no solution to the reduction of color should have to account for all 5. Author's approach is to suggest a solution to what color is that can accommodate all 5 core beliefs (after he alters one of them). This is meant to recommend color primitivism. (pg 141) Author weakens and re-writes one of the beliefs:

Revelation: The intrinsic nature of Canary Yellow (CY) is fully revealed in an experience as of something CY
Re-written/Modest Revelation: A complete understanding of what CY is can be given by an experience as of something CY, in someone with normal human vision who has already acquired color concepts. (pg142)

Initially, Revelation conflicted with Explanation: Sometimes something's being CY explains why it looks CY. Author argues that with color primitivism, 'Modest Revelation' does not conflict with 'Explanation' (pg144-5) Author then walks through the other accounts of color: that it's a disjunctive primary quality/property of objects (pg145-6), that it's a disposition of a certain kind (pg147-8), that it's a secondary property of reflection (pg148-151). It is here where the reductionist is given a choice of saying that the properties that give rise to color are either nomologically or metaphysically connected to color. Author argues that either choice yields silly truth-values to ill-formed modal questions, and that the color primitivist does not have to pick between these two options. (pg149-150)

Lastly, author takes on the objection of spectrum inversion and considers Wittgenstein's private language argument as the reply to it (pg152-3). Author also discusses the possibility of animal color spectrum being larger than ours is.