Synaesthesia, Color Terms, and Color Space: Color Claims Came From Color Names in Beeli, Esslen, and Jäncke (2007)

Abstract

Grapheme-color synaesthesia is an unusual condition in which letters and digits generate involuntary experiences of color. In this article, we show that grapheme-color synaesthetes make nonrandom associations that are best described by an understanding of color language, and that different measures of such associations converge to illustrate a nonarbitrary relationship for all people between color names and color space. We focus on a recent study (Beeli, Esslen, & Jancke, 2007) that presented the ostensibly novel finding of a relationship between the colors that synaesthetes associate with graphemes and the frequencies of those graphemes. However, in a study published 2 years earlier (Simner et al., 2005), but not cited by Beeli et al., we had already established that grapheme frequencies are important. We showed that high-frequency graphemes tend to be paired with high-frequency color names (e.g., ared) in verbal reports of synaesthetic associations. Beeli et al. measured the hue, saturation, and luminance (HSL) of synaesthetic colors and showed {inter alia) that grapheme frequency was correlated with the colors' saturation. In this Commentary, we compare data from our previous study with those of Beeli et al. and empirically demonstrate that synaesthetic colors are best accounted for by color naming; we also suggest that the similarity across the two studies may be indicative of an independent relationship between color naming and color space for all people. We reanalyzed the data from Beeli et al. in order to demonstrate three facts: that, as we predicted (Simner et al., 2005), their synaesthetes shared certain grapheme-color associations (e.g., a was associated with red more often than predicted by chance); that, also as we predicted, these grapheme-color combinations reflect a positive correlation between grapheme frequency and color-name frequency; and finally, that certain aspects of the HSL color space (upon which Beeli et al. based their conclusions) may be predicted from color naming. We converted the synaesthetic color choices from the graphemecolor synaesthetes tested by Beeli et al. (their Fig. 1) into the 11 irreducible color terms from Berlin and Kay (1969; i.e., black, white, red, yellow, green, blue, brown, orange, purple, pink, and gray). Our coding was performed by two independent assessors, with any disagreements resolved by a third coder. We then performed upper-tail binomial analyses, which showed that certain grapheme-color combinations occurred significantly more often than would be predicted by chance. For example, Figure 1 illustrates the distribution of blue and white color associations