J Vis. 2023 Dec 1;23(15):80. doi: 10.1167/jov.23.15.80.
When a person views a color sample, they can usually provide a color term for it. But will that color term allow someone else to understand which sample was named? We examined color understanding using a Color Communication Game, in which one person (the “sender”) names 30 color samples as in any color-naming study, then another person (the “receiver”) chooses the sample they think the sender intended to communicate. The receiver cannot always guess the right sample, and no choice strategy will do better than randomly choosing among the samples the receiver called by that term. When 70 English-speaking dyads and 63 Somali-speaking dyads played the game, receivers did not perform randomly. Instead, they systematically chose “focal” samples near the centers of their color term distributions. When the senders’ named samples were compared directly to the receivers’ chosen samples, the systematic distribution of receiver choices revealed color categories, which appeared without any statistical analysis of the players’ terms. Simulation of receiver choices based on senders’ color names showed that both Somali-speaking and English-speaking participants knew more color terms than the ones they used in color-naming. Our Color Communication Game showed that color-naming experiments underestimate color understanding: people understand colors categorically, but they express colors using multiple synonymous color terms that are well-understood by others.