Abstract
Visual preferences for artworks and scenes tend to align with regularities in spatial scene statistics of the natural world (Graham & Field, 2008, Spatial Vision, 21, 149-64). Similarly, preference is highest for color distributions which coincide with the chromatic variation present in natural scenes (Juricevic et al., 2010, Perception, 39, 884-99). These effects have been attributed to the efficiency of neural coding, or “cognitive fluency” (Reber et al., 2004, Pers. Soc. Psychol. Rev. 8, 364-82), such that the tuning of the visual system to the environment drives aesthetic preference. We compared aesthetic responses to color distributions for observers living in different geographical locations, and quantified the differences in the distributions of colors in their local environments. We measured aesthetic preference for Mondrians containing colors distributed along cardinal and oblique axes in the Macleod-Boynton chromaticity diagram for observers living at three locations – Brighton, UK; Quito, Ecuador; and rainforest in Ecuador. The preferences of UK observers replicated Juricevic et al. (2010): Mondrians with colors distributed along the yellow-blue (oblique) axis were most preferred. We found opposite preferences for participants in urban Ecuador (highest for purple-green distributions), while participants living in the Ecuadorian rainforest showed no consistent preferences. Using color-calibrated GoPro cameras, mounted to the heads of participants, we captured random samples of natural scenes typically encountered in the daily routines for people living in each of the three locations. By measuring the response functions of each camera’s RGB channels and recording images in RAW format, we reconstructed the MacLeod-Boynton chromatic scene statistics to which participants were exposed – the “chromatic diet”. The pattern of aesthetic preference is not straightforwardly predicted from the analysis of the chromatic diet at each location, challenging the notion that aesthetic preference is strongly driven by tuning to natural statistics.