Biographie

Bevil Conway
Bevil Conway is Professor of Neuroscience at Wellesley College in the United States.
He is a visual neuroscientist and artist who studies the neural basis of color using physiological, behavioral, and modeling techniques. His laboratory uses a range of techniques, including fMRI-guided microelectrode recording and microstimulation in awake-behaving non-human primates trained to perform visual tasks, along with psychophysics and fMRI in humans, and computational modeling, to define and test hypotheses relating physiology and perception. In addition to maintaining an active studio practice, Prof. Conway is involved in ongoing projects at the interface of visual neuroscience, visual art, and the practice of making art. He teaches core courses in the Neuroscience program and an advanced interdisciplinary laboratory course, Vision and Art: Physics, Physiology, Perception, and Practice.
His artwork has been published in several books including Vision and Art (Abrams, 2002) and Brain and Visual Perception (Oxford University Press, 2004), and has been used by BOSE Wave Radio in advertising. A major solo show of his work, 'FACTS', took place at the Radcliffe Institute for Advanced Study in 2010/2011. The work is held in several private collections in Europe, Africa and North America and is in the public collection of the Fogg Museum and the Boston Public Library, and on semi-permanent exhibit at the N.I.H. He is currently working on a series of drawings and etchings, a 'Rake's Progress', exploring mark-making and movement, inspired by Mark Morris's Dancers.

Abstract: How Colour Language Reflects Usefulness of Colour

The question of color categorization has long been caught in the crossfire between universality and cultural relativism. Color categories across languages are somewhat consistent, suggesting a common basis in physiology. But the variability between languages, and the lack of physiological or other causal evidence, promote the alternative idea that languages are constitutive of color categories. I will attempt to reconcile these positions through an analysis of a new, extensive, well-controlled dataset on color-naming in three diverse groups: the Tsimane’, a hunter-gatherer isolate in the Amazon; Bolivian Spanish speakers; and English speakers near Boston. These data resolve widespread methodological concerns about the World Color Survey (WCS), and license an analysis of this impressive dataset covering 110 languages to determine how successfully people can communicate color. Despite a large range of average communicative efficiency across languages, we discovered remarkable within-language consistency: in all languages communication was best for warm colors (yellows/reds), rather than cool colors (blues/greens). We show that the communicative efficiency of a given color is predicted by the frequency with which salient objects have that color—in natural images, we found that objects tend to be warm-colored against cool-colored backgrounds; this relationship was true for natural objects as well as artificially colored objects. These results suggest that the color terms across languages are driven by the common goal to optimize communication about colors of likely relevance. In additional experiments, in Tsimane’, Bolivian Spanish, and English, we assessed memory color for familiar objects and labeling of real-world objects. The results support the idea that the overall communicative efficiency of a language is determined by the extent to which color is generally useful for the culture. Together, the results provide the first explanation of the most basic, universal color categorization, and show how this single mechanism—usefulness—can also account for the variance among color-naming systems.