Abstract
We report a new, information theory based analysis of color naming. We compute mutual information (MI, in bits) in a simulated “game” involving a “sender” (S) who names out loud, based on his color idiolect, the colors of samples selected randomly (with replacement) from an array of N samples. A “receiver” (R) attempts to identify S’s selections from her duplicate array of color samples, based only on S’s color term message and her own color idiolect. MI measures how much S’s messages improve R’s chances of guessing S's selections correctly. Computing average MI for all pairwise permutations of informants of a language provides an estimate of shared color knowledge within the informants’ culture that takes into account the number of color samples tested, the number of color terms in informants’ idiolects, and the group consensus in color term deployment. The communication game reveals several interesting properties of worldwide color naming. For example, MI is quite variable among the 110 World Color Survey languages using a given number of high-frequency color terms, because the consensus for those terms is often highly variable. Moreover, 96% of the variance in WCS MI can be accounted for by informants’ responses to just 23 of 330 color samples tested in the WCS, suggesting a very efficient stimulus set for other color naming studies. Additionally, many informants of Hadzane, a language spoken by nomadic Tanzanian hunter-gatherers, profess no knowledge of names for many color samples, yet have surprisingly high group MI (relative to many WCS languages), because their limited lexicon is deployed with high consensus. Finally, we show that allowing informants to use two rather than one color name generally offers little improvement in group MI. Thus, information theory provides a powerful quantitative tool for studying human communication about color.
Meeting abstract presented at VSS 2015