Categorical color perception shown in a cross-lingual comparison of visual search

Abstract

Categorical perception (CP) for colors entails that hues within a category look more similar than would be predicted by their perceptual distance. We examined color CP in both a UK and a remote population (Himba) for newly acquired and long-established color terms. Previously, the Himba language used the same color term for blue and green but now they have labels that match the English terms. However, they still have no color terms for the purple areas of color space. Hence, we were able to investigate a color category boundary that exists in the Himba language but not in English as well as a boundary that is the same for both. CP was demonstrated for both populations in a visual search task for one different hue among 12 otherwise similar hues; a task that eliminated concerns of label matching. CP was found at the color-category boundaries that are specific to each language. Alternative explanations of our data are discussed and, in particular, that it is the task-dependent use of categorical rather than non-categorical (perceptual) color networks which produces CP. It is suggested that categorical networks for colors are bilaterally represented and are the default choice in a suprathreshold similarity judgment.