Abstract

This paper explores the implications of some experimental data for views that identify colors with objective physical properties such as reflectance profiles. Those who reject objectivist views often argue from the existence of intersubjective differences in color categorization ; but objectivists have managed to stand their ground by identifying colors with sets or ranges of reflectances individuated by the ways in which they stimulate the visual system. In the interest of moving the debate forward, I provide a new kind of evidence against objectivism. Results of a psycholinguistic experiment reveal hysteresis and enhanced contrast in ordinary speakers’ applications of vague terms. These dynamical patterns are purely psychological and give rise to intra-subjective variation in subjects’ applications of vague predicates; in particular, in the case of color predicates, nothing in the stimulus configuration or the illuminant undergoes any change, the only variable being the order in which stimuli are judged. I hypothesize that these order effects are necessary if vague words are to be applicable to values on dimensions, like color, that admit of continuous change. To the extent that this hypothesis is correct, it suggests that if colors are the properties named by ordinary color predicates, and ordinary color predicates are vague, and the application of vague predicates exhibits the order effects found in the experiment, colours cannot be physical or otherwise objective in nature.