[this Is a Comment on a Ms. by Stephen E. Palmer, Entitled Color, Consciousness and the Isomorphism Constraint, to Appear in Behavioral and Brain Sciences.]

PURPLE (RED-and-BLUE) is the most frequently occurring derived (binary) basic color term (BCT), but there is never a named composite BCT meaning RED-or-BLUE. GREEN-or-BLUE is the most frequently named composite color category, but there is never a BCT for the corresponding derived (binary) category CYAN (BLUE-and-GREEN). Why? Palmer notes that the naming of some but not all of the possible composite and derived basic color categories constitutes a puzzling set of asymmetries in color experience, assuming these categories are in fact reflections of color experience. He continues (p. 8 of ms.), "I am not aware, however, of any behvioral data that directly support these asymmetries for derived and composite color categories in color experience." I think such behavioral data exist, though they are admittedly few. I return to this point presently. Meanwhile, I would like to flesh out further the question of which possible composite and derived categories actually occur, what their relative frequencies are, and what significance may be attached to these relative frequencies. I will restrict my attention to the 110 languages of the World Color Survey (WCS. See Kay and Berlin 1997). Figure 1 summarizes much – not all – of the WCS data on composite and derived categories. It also displays the number of jnds separating each pair of non-opponent hue primaries (RED, YELLOW), (YELLOW, GREEN), GREEN, BLUE), and (BLUE, RED). Palmer raises the issue of the relevance of jnd separations: "For example, if for some reason there are more justnoticeable-differences (jnds) between unique red and unique yellow than between unique green and unique blue, the wider psychological gap might explain why there are BCTs for ORANGE in many languages, but not for CYAN (blue-green)" (p. 7). Although I am not certain that the number of jnds separating two unique hues measures a psychologically real distance, this information has been included in Figure 1 to allow its possible relation to variations in the popularity of corresponding composite and derived categories to be assessed.