Abstract
The universality of color categories has been demonstrated by various studies. A series of analyses of categorical clusters of color samples used in World Color Survey (WCS) have revealed several 'motifs,' or color naming systems (Lindsey and Brown, 2006, 2009, and 2014), which present commonly across languages. Furthermore, several motifs co-exist within a language. The motifs differ mainly at the border between the blue and green categories. In Japanese, one of the basic color terms 'ao' (translated as blue in English) is occasionally used when naming leaves of trees, traffic lights, copper rust, and etc. We investigated the basic color categories of Japanese using cluster analysis to clarify whether the color categories in Japanese show an unclear border between the blue and green categories, i.e., the 'grue' color term or the corresponding motif. We conducted a k-means analysis, in which the optimal number of clusters was determined, using gap statistics to define the optimal number of categories. 57 Japanese-speaking participants named the WCS color samples using monolexemic color terms without modifiers or compound words. We found six chromatic categories in addition to the basic color terms in Berlin and Kay (1969). They were: 'mizu (water)', 'hada (skin)', 'matcha (powdered green tea)', 'oudo (sand/mud)', 'ai (indigo)', and 'enji (reddish brown)'. The terms 'mizu', 'hada' and 'kusa (grass)' were reported as subsidiary basic terms in a study of Japanese categorical color terms (Uchikawa and Boynton, 1987). The term 'kusa' was not used in the present study, but the term 'matcha' may be a substitute. The 'mizu' cluster was consistently located at high lightness, between blue and green, suggesting that blue and green are distinct categories. However, the 'mizu' cluster is fundamentally different from the 'turquoise/teal' cluster observed in English because 'mizu' also includes many samples that most English speakers call "light blue".
Meeting abstract presented at VSS 2016