Angela Brown¹, Charles McDonald¹, Delwin Lindsey¹; ¹Ohio State University

The surfaces of environmental objects generally have both texture and color qualities. Here we compare the naming of color to the naming of texture, then we use sorting game to ask which surface quality (texture vs. color) is more effectively communicated using single non-composite terms. 99 samples of materials varied in color and texture. 11 adults freely provided one or two terms for each sample. Information theoretic considerations suggested that color terms might be less effective for communication compared to texture, because each of 12 distinct color terms named 19 samples, whereas each of 17 distinct texture terms named only 6.5 samples. Each of 29 “other” terms named 32 samples. However, we found no evidence that texture and color are differentially communicated when participants actually interact. 46 dyads of adult participants played a sorting game with the 100 new samples. The first player of each dyad divided the samples into 2, 3 or 5 groups, based on similar color or similar texture, then named each of the groups they had created, following identical instructions for the texture sorts and the color sorts. The second player viewed the terms provided by the first player and sorted the samples into groups corresponding to the first player’s terms. Mutual Information (MI) quantified agreement between the two players’ sorts. Across all numbers of categories, MI for color sorting and texture sorting agreed well. The players of these dyadic games provided the same number of distinct terms per sample in the color and texture games. Our results emphasize the importance of “closing the loop” in studying communication about surface properties: simple naming results often miss important pragmatic features of dyadic interaction. Naming alone misses a lot of what communicators know about the properties of object surfaces.