COLOR FICTIONALISM: COLOR DISCOURSE WITHOUT COLORS Dimitria Electra Gatzia

DISSERTATION ABSTRACT In this dissertation I defend an error theory of color, which I call prescriptive color fictionalism. I argue that although our color discourse is false, we should continue employing it as we have thus far but stop believing that color properties exist. I begin with an analysis of ordinary color concepts and I argue that they fail to denote physical properties (chapter 1). This claim is widely accepted by most color realists, who I call revisionists, since they attempt to revise our color concepts in order to succeed in denoting physical properties. I discuss three such views: the primary-quality account, relationalism, and disjunctive physicalism. The primary-quality theorist takes colors to be bases of dispositions, thereby violating the axiom of unity –the thesis that all, say, blue things, share the same color property. In an attempt to solve the problem of individual variations among normal observers, the relationalist takes colors to be relational properties. I argue that relationalism fails because it leaves no room for error in color perception. The disjunctive physicalist acknowledges that there are a variety of causes that are responsible for objects looking, say, blue, and argues that colors are disjunctive properties. I argue that this view fails for two reasons: disjunctive properties can be neither causes nor projectible (in Kim’s sense). I further consider two non-revisionist color theories, which I call strict and relaxed representationalism. The strict representationalist argues that colors are objective properties represented in our visual experience and that color properties are responsible for our color phenomenology. This view fails because it cannot account for, among other things, contrast colors and individual color variations among normal observers. Relaxed representationalists attempt to solve the problem of individual variations among normal observers by arguing that it is not the color properties that are responsible for our color phenomenology but rather some other (relational) properties. I argue that this

account renders colors inexplicable. Since both revisionist and non-revisionist accounts fail, I conclude that color realism is false (chapter 2). In chapter three, I try to motivate an error theory of color and argue that eliminativism, which is a variation of an error theory according to which color discourse should be eliminated precisely because it is false, is unmotivated. I chapter four, I take up color subjectivism, which says that colors are mental properties, and argue that it faces two insuperable problems. The first has to do with the fact that the color subjectivist cannot provide standards of correctness – it cannot tell us which experiences are correct and which are not; the second pertains to commonality of reference –the color subjectivist cannot say that our color terms have a common reference in the mouths of different observers. In chapters five and six, I offer a positive account. I first consider how fictionalist accounts work in various domains, including moral, modal, mathematical and scientific fictionalisms (chapter 5). I then propose a prescriptive fictionalist account of color (chapter 6). My account is prescriptive because it recommends that we change our practices in order to continue talking about colors (ordinarily understood); and it is a kind of fictionalism because to talk about colors is to pretend that colors exist. Finally, I consider some general worries about the view I propose and argue that none of them posses a real threat to it. By formulating a fictionalist theory of color and showing that we can save our color discourse without having to add color properties in our ontology, I provide an attractive alternative to color realism and offer a viable solution to the general problem of color.