Abstract
Affective and aesthetic responses to images may be partly determined by spatiochromatic statistics, independently of semantic meaning or content (e.g. McAdams et al. 2023, J. Vis.). Such relationships are generally explored through positive rather than negative responses – e.g. liking or beauty ratings, not disliking or ugliness. Here we examine the extent to which chromatic properties alone may predict both perceived beauty and ugliness, in photographs. We further examine whether differences in colour distributions between beautiful and ugly photographs are predictable from abstract colour preferences; in Western culture, people of both sexes and different ages tend to prefer colours with cooler hues and higher saturation and lightness and to dislike dark yellow hues (browns). Methods: In an online experiment, 94 participants (94% of which were aged 16-25 years) each submitted four recently taken photographs, two that they considered ugly and two beautiful, with an indoor and outdoor scene for each aesthetic category. Photographs were restricted to scenes without people or other animate objects. In a subsidiary online experiment, 6 new participants independently ranked hue salience in each submitted image. Results: Analysis of the image statistics showed that across both scene types, colourfulness was significantly higher for beautiful vs. ugly pictures, and for outdoor images only, mean image chroma was significantly higher for beautiful pictures. Hue distributions - whether obtained by an automated colour category classification algorithm or from the independent human ratings - also differed between beautiful and ugly pictures. In outdoor pictures, the proportion of “blue” was significantly lower and “brown” and “grey” significantly higher in ugly vs. beautiful categories. The results suggest that aesthetic responses to pictures are partially predicted by affective responses to colour, both positive and negative. The results further suggest that perception of ugliness is not merely the opposite of beauty, but a distinct process.