The history of empirical research on the Sapir-Whorf hypothesis is reviewed. A more sensitive test of the hypothesis is devised and a clear Whorfian effect is detected in the domain of color. A specific mechanism is proposed to account for this effect and a second experiment, designed to block the hypothesized mechanism, is performed. The effect disappears as predicted. The Sapir-Whorf hypothesis is reevaluated in the light of these results.
