Abstract
The effects of color on visual performance have critically tested the Sapir-Whorf Hypothesis that language controls many aspects of cognition, including perception. Kay & Kempton (1984) investigated the salience of small color differences near the boundary between “green” and “blue” English color categories. Here, we examined the perceptual scaling of color differences in this boundary region: Is scaling monotonic at the boundary? Are there visual field differences? If so, effects restricted to the right visual field might suggest a Whorfian effect of language, as proposed by Gilbert et al. (2006).
In the first experiment, we used maximum likelihood difference scaling (MDLS; Maloney and Yang, 2003) to perceptually scale the angular interval (140° – 160°) spanned by isoluminant greens and blues falling on an arc (radius=43 ∆E) in CIELAB. If the blue-to-green boundary produced especially salient color differences, MLDS should reveal them. If the salience were related to language, the differences should be greater in the right visual field. We found neither of these predicted effects. Left and right visual field difference-scaling functions (difference-scale vs. CIELAB color angle) were monotonic and indicated that perceptual differences among “greens” were always significantly greater than differences among “blues”. There was usually a clear-cut discontinuity in these functions, indicating a transition between “green” and “blue” color appearance. But there was no violation of monotonicity at the discontinuity.
In a second series of experiments, we examined the transition region more closely by measuring the point of subjective equality (PSE) in color difference between a standard blue-green pair straddling each subject's blue-green boundary and variable-difference same-category color pairs. PSEs were consistent with the MLDS studies, indicating that difference scaling across the blue-green boundary is in no way enhanced relative to same-category differences.
Thus, we find no evidence for the cross-category privilege suggested by the Sapir-Whorf hypothesis.