Abstract

Are color categories the evolutionary product of their usefulness in communication, or is this an accidental benefit they give us? It is argued here that embodiment constraints on color categorization suggest that communication is an add-on at best. Thus, the Steels & Belpaeme (S&B) model may be important in explaining coordination, but only at the margin. Furthermore, the concentration on discrimination is questionable: coclassification is at least as important.