January 28, 2007

Color names and color perception: the cartoon version

Two fascinating recent papers on the back-and-forth between language and perception in the domain of color, involving sometime Language Log poster Paul Kay: G. V. Drivonikou, P. Kay, T. Regier, R. B. Ivry, A. L. Gilbert, A. Franklin, and I. R. L. Davies, "Further evidence that Whorfian effects are stronger in the right visual field than in the left ", PNAS 104: 1097-1102 (2007); and Terry Regier, Paul Kay, and Naveen Khetarpal, "Color naming reflects optimal partitions of color space", PNAS 104: 1436-1441 (2007).

There's some background on earlier work by Gilbert, Regier, Kay and Ivry in some earlier LL posts -- "What would Whorf say?", 1/3/2006; "What Whorf would have said", 1/15/2006. The second one is by Paul Kay himself. A good summary of the recent work (as far as I understand things) can be found in "How grue is your valley?", The Economist, 1/18/2007.

More on this later. For today, I just wanted to note that today's Get Fuzzy strip is either a commentary on this research, or a big coincidence.

On reflection, I guess it's a coincidence.

But I gather it's true that dog's color vision is based on two photopigments rather than the three that humans have, and also that the proportion of retinal cones is much lower than in humans. Thus according to Neitz et al., "Color Vision in the Dog", Vis. Neurosci. 3(2) 119-25 (1989):

Measurements of increment-threshold spectral sensitivity functions and direct tests of color matching indicate that the dog retina contains two classes of cone photopigment. These two pigments are computed to have spectral peaks of about 429 nm and 555 nm. The results of the color vision tests are all consistent with the conclusion that dogs have dichromatic color vision.

So if dogs could actually talk, their color language would be different. And it's plausible for Satchel to be confused about what blue means.

Of course, there are plenty of human dichromats. Has anyone looked into their color language, or more specifically the effects of linguistic categories on their color perception?

Posted by Mark Liberman at January 28, 2007 03:06 PM