"Statistical analysis of cross-language color naming data."

International Computer Science Institute, Berkeley CA

Investigators

Abstract

With National Science Foundation support, Drs. Paul Kay and Terry Regier will build a web-accessible database to test the hypothesis of universal constraints on color naming across languages. Anthropologists and linguists used to think that the meanings of color words (e.g., "red", "black", etc.) differ arbitrarily from language to language. It was also thought that these words, rather than being constrained by biological constants of human color perception, vary arbitrarily across cultures and that the words cause speakers of different languages to see color differently. A 1969 study by Kay and colleagues challenged this view, maintaining that patterns in color naming across languages reveal certain constants. That study was followed up by the NSF-supported World Color Survey, in which data on color naming in 110 unwritten languages were gathered in the field by linguists highly familiar with the languages. The World Color Survey resulted in a number of publications, generally supporting the view of universal constraints on color naming. Because these data were collected in the early days of computers, they were not entered into a uniform database that would make them amenable to comprehensive statistical testing or to web publication. The current project will (1) convert the World Color Survey data to a uniform database, (2) make this database available to everyone via the web, and (3) perform comprehensive statistical testing on the data to test hypotheses about color naming. The question of universal constraints on systems of color naming has broad implications. Color words provide a natural laboratory in cross-language comparison of meanings because their meanings - colors - can be specified in a scientific system of description that is independent of any particular natural language. Color has thus been the vocabulary domain most often cited both by proponents of linguistic/cultural relativity and by their opponents, who tend to emphasize biologically and evolutionary based universals in human behavior. As a consequence, color words have come to occupy a special position in the very general nature versus nurture debate, a controversy within the biological and social sciences, as well as the general educated public and many issues of public policy, at least since Darwin. Increased scientific clarity with regard to universality versus arbitrariness in color naming is consequently of general interest.