Shape-color associations are mediated by stereotypicality

In 1912, Kandinsky proposed a fundamental link between shapes and colors, where circles are associated with blue, triangles with yellow, and squares with red. These associations could not however be replicated in the general population, nor did yield any priming effect (Kharkhurin, 2012). Instead, other tendencies emerged; with the circles being yellow, triangles red, and squares blue (Jacobsen, 2002). Albertazzi et al. (2013), widened up the number of color answers from 3 to 40 color chips, and found that Italian observers would tend to associate the circle with red and yellow, the triangle with yellow, and the square with both red and blue. Using a similar paradigm, comparable tendencies appeared in Japanese observers (red circles, yellow triangles and blue square; Chen et al., 2015, 2017). While Jacobson (2005) hypothesized these associations are due to environmental influences (yellow sun or triangular red traffic sign), Albertazzi and colleagues (2013) and Chen and colleagues (2015) explored whether those shape-color associations were mediated by perceived temperature in shapes and colors. We investigate shape-color association in a novel population (i.e., Danish individuals) using a full color wheel, allowing free associations from observers. Our results reveal a significant association of the circle to red, yellow, and white; the triangle to green and yellow; the square to blue; and both the pentagon and hexagon to pink and purple. Children and teenagers seem to also associate the square to red. Overall, the most stereotypical shapes (circle, triangle and square) are associated to stereotypical colors (red, blue, green, yellow), while the non-stereotypical shapes are associated to non-stereotypical colors. Shape-color associations seem therefore mediate at the conceptual level, through stereotypicality. Our data suggest that this mediation process in already present at an early age.