Color Names, Universal Categories, and Color Categorical Perception

Location: L2.5 Leech Building, Newcastle University
Time/Date: 3rd July 2012, 16:00 - 17:00

Color Names, Universal Categories, and Color Categorical Perception

To be presented jointly by Delwin Lindsey (Department of Pyschology) and & Angela M. Brown (College of Optometry), Ohio State

Hosted by the Institute of Neuroscience

Color naming is a classic topic of research on the relation between cognition and perception. The spectral composition of visible lights is continuously variable, yet in naming the colors, most world languages partition this continuum into distinct color categories. Is this partition strictly linguistic? Or is the perception of colors itself intrinsically categorical? For three quarters of a century, linguists, anthropologists and psychologists have debated these questions.

There is great variation across languages in the number of color terms in everyday use, yet Berlin & Kay (1969) proposed that “basic” color terms are universal, and occur only in certain combinations.  Our novel statistical analysis of 2616 informants from the World Color Survey (Kay et al., 2009), a database of color naming in 110 world languages spoken in pre-industrialized cultures) reveals 11 universal English-like color terms and about four universal color term combinations, which we call “motifs”.  Unexpectedly, the worldwide diversity in motifs is recapitulated within most of the individual WCS languages. We have also observed within-language diversity prospectively in a study of non-English speaking Somali immigrants to the US.  Critics of the universality hypothesis often attribute within-language diversity to informant incompetence, but we also find striking diversity in color naming even among educated native English speakers.
We have also tested the generality of the hypothesis that language affects color perception by studying visual search for colors, where target and distractor colors are drawn from the same or from different color categories. If color names and color perception are tightly linked, visual search should be disproportionately fast when the target and distractors are from different color categories. Contrary to that prediction, we find that search reaction times are generally insensitive to within- vs. cross- category conditions, when the stimuli are carefully controlled.  Instead, the results are generally well fit by a model based on color-opponent channel responses, irrespective of category membership.  Under some conditions, our simple model fails, but the model fits to the data are not improved if color category membership is taken into account.
We will discuss these results in the context of universal, neurobiological or environmental constraints on color vision.

Published: 27th June 2012