Experimentally Elicited Judgments of Color Harmony

Abstract

Eight variants of each of four different geometrical forms were constructed by varying the proportions of six different colors. All possible pairs of the resulting colored pictures were presented to subjects who were asked to determine 1) which of the two pictures is more harmonious, and 2) by how much one picture is more harmonious than the other. Scaling of both individual and group data shows that subjects can produce consistent judgments of harmony which are measurable by a ratio scale. Two models of color harmony based on the work of Munsell and Itten yield similar predictions of color proportions, which are not supported by the data. The failure of the models is attributed in part to their disregard of color composition.