Abstract
Across languages, verbal codes divide the color spectrum in different ways. Do linguistic codes affect color discrimination? In a 2AFC recognition memory task, it was shown that subjects (Ss) are less likely to false alarm to cross category (CC) than within category (WC) foils, but that this advantage disappears with a verbal interference task (Roberson & Davidoff). This effect might be explained by a verbal contribution to categorical perception, or by language acting as a secondary code to meet the demands of working memory. We asked whether Ss show a CC advantage in tasks with no memory component, whether this advantage is selectively reduced by verbal interference, and whether this advantage can be seen in a language group that has a verbal color boundary and not in one without the same boundary.
-
Ex 1: In a 2AFC task, Ss saw a blue or green sample square patch, together with a match and a foil patch. No patches abutted. The foil was either WC or CC, with distance in Munsell color space held constant. Ss showed a CC advantage. Crucially, a verbal but not a spatial interference task significantly reduced this advantage.
-
Ex 2: Ss were shown 2 adjacent square patches (blue or green), and made a same/different response, essentially reducing Ex 1 to an edge-detection task. We found a CC advantage, but there was no selective effect of verbal interference.
-
Ex 3: We repeated Ex 1 using only blue stimuli with Russian and English Ss. Russian has separate basic color words for light blue and dark blue. Preliminary results suggest that Russians have an advantage for CC over WC, and that this advantage is reduced by verbal interference. English speaking subjects, for whom all discriminations are WC, do not show this pattern.
We infer that there is an on-line effect of verbal coding on color discrimination (Ex 1), that if the task is made simple enough the effect of language is reduced (Ex 2), and that the effect may be specific to the coding system of the language (Ex 3).