Winter 2013-2014

Color Primitive

Hue, language, and semantic universals

Josh Berson

There could very easily be a tribe of people who are all colourblind and who nonetheless live very well; but would they have developed all our colour names, and how would their nomenclature correspond to ours? What would their natural language be like?? Do we know? Would they perhaps have three primary colours: blue, yellow and a third which takes the place of red and green?—What if we were to encounter such a tribe and wanted to learn their language? We would no doubt run into certain difficulties.1
—Ludwig Wittgenstein, Remarks on Colour, 1977

The striking contrast presented by the natives of Lifu [island] with those of the Torres Straits at once suggests that the existence of colour-blindness in a race might be of great importance as an ethnic character, and the other data also tend to show that colour-blindness may be a characteristic of certain races and the existence or absence of this defect may help us in the difficult task of deciding on ethnic affinities.2
— W. H. R. Rivers, Reports of the Cambridge Anthropological Expedition to Torres Straits, Volume II: Physiology and Psychology, 1901

When the informant appears reasonably at ease, please open the booklet to page 2 and proceed with the naming task. As you may be aware, many languages do not contain a word meaning “color.” Yours may well be one of these. Experience has shown, however, that it is always possible to find some verbal formula to elicit color responses. Sometimes these translate to, “How has it been dyed?” or “How does it strike the eye?” or “What is its appearance with respect to red, blue, etc.?” and so on. Probably none of these three is just the thing needed for your language, but with a little experimentation you should be able to find a question that elicits color words. You will be aided in this by the stimulus objects themselves, which differ from each other only with respect to color. 3
— Brent Berlin, Paul Kay, and William Merrifield, “World Color Term Survey: Instructions to Field Workers,” 1976

The concept of basic vocabulary is everywhere. No matter where you turn, linguists, sociologists, computer scientists, and even literary critics are paring back the dictionary, editing the lexicon, looking for a minimal set of words whose meanings can be relied upon to not vary too much from one occasion, person, text, culture, or era to another. The aim is a reinvention of the science of behavior, a data-driven exploration of the various strategies by which symbol-using creatures (that would be us) bind form to meaning and transmit meanings over space and time. Consider the Computational Story Lab at the University of Vermont, which has gathered “happiness scores” for the ten thousand most commonly occurring English words and word fragments on Twitter and used these to create a “Hedonometer,” a remote-sensing apparatus for human mood. The data are impressive, though some of the methodological assumptions—e.g., concerning the demographics of Twitter users and even the very reliability of word happiness scores—are open to question.4 Notable too are the collaborative experiments conducted by literary scholars at Stanford University and the University of Wisconsin to see if frequency counts for only a handful of commonly occurring words and punctuation marks could be just as useful as a functional typology of 200 million distinct English strings in determining the genre of nineteenth-century novels.5

Subscribe to access our entire archive.
Log In and read it now.