Date Published: November 28, 2018
Publisher: Public Library of Science
Author(s): Vincent C. Sun, Chien-Chung Chen, Zhiqiang Cai.
Basic color terms used in Mandarin Chinese have been controversial since first discussed by Berlin and Kay in 1969. Previous studies showed much inconsistency on what should be considered as basic color terms in Mandarin Chinese. In the present study, we investigated categories of color rather than merely the color terms used by Taiwanese native Mandarin speakers. Using samples conforming to the Berlin and Kay survey, various colors were chosen from a collection of Natural Color System (NCS) colored papers and mounted on a piece of neutral gray card. The card was then mounted on a touch-screen, under D65 illumination. Thirty-two single-character color related Mandarin terms were selected from a Chinese character database according to frequency of use. Participants were required to select the color sample that matched the term by pressing a virtual button on the touch screen. The results show that certain terms can be directly correlated to basic color terms in English, comparable with the results of Berlin and Kay’s original study and those that followed. However, some terms, such as Mo (墨 ink), Tie (鐵 iron), and Cai (菜vegetable), show a wide spread of term maps and inconsistent use among subjects. Principle component analysis (PCA) procedures were used to analysis the commodity of data among subjects. The findings suggest that the basic color categories among Mandarin Chinese speakers are similar to those found in the World Color Survey (WCS), but are represented by wide-spread and inconsistent color terms among speakers.
Modern studies on the basic color terms began with the seminal work of Berlin and Kay , who suggested that a basic color term should have four characteristics: (i) its meaning cannot be predicted from its parts; (ii) its meaning is not included in that of another term; (iii) it is not specific to a narrow class of objects; and (iv) it must be psychologically salient for the users. Berlin and Kay applied these criteria to determine the basic color terms in several languages [1, 2]. They reported that in the English language, there were eleven basic color terms: white, gray, black, blue, green, red, yellow, orange, brown, purple and pink. In the same study, Berlin and Kay  reported that Mandarin Chinese had six basic color terms.
Of the 63 participants who agreed to take part in the experiment, four did not finish the survey for various reasons. Data from the remaining 59 participants are reported here. For data analysis, we first pooled the responses from all participants together and counted the total number of times, or frequency, a color sample was selected for each color term. The acquired frequency matrix was analyzed in two ways. First, we mapped the frequency matrix to WCS palette. The frequencies for WCS colors not sampled in this study (see Fig 1) were filled with a bilinear interpolation from neighboring data. The result is the frequency distribution presented below. Second, we perform a principle component-based factor analysis the original frequency matrix. The rationale is that if two terms referred to the same color category, the participants’ response to them would be highly correlated. Thus analyzing the correlation structure of the frequency matrix should help us to identify the synonyms. The factor loadings of tested color terms on major factors are reported below.
In this study, we used a paradigm that is very different from the conventional methods used since Berlin and Kay’s research. That is, rather than showing color chips to an informant and inquiring about the corresponding color terms, we showed participants a color term and asked them to identify all the color samples in our stimulus set that matched the term. We hypothesized that much of the controversy about the basic color terms in Mandarin Chinese had to do with the large number of synonyms for each color category. As a result, previous surveys seemed to offer little consistency among respondents. We found there are many synonymous terms for color categories such as red, green, blue, yellow, brown, pink, orange, black, and gray, as listed in Table 2. Because these synonyms showed virtually an identical distribution on the WCS chart, we merged the frequency distribution of the synonyms to plot the boundaries of color categories on the WCS chart (Fig 13).
Biggam  suggested that there is a crucial difference between basic color terms and basic color categories. A color category as a concept belongs to the cognitive domain and color terms and words belong to the linguistic domain. Though the disparity between categories and terms was suggested, the phenomenon of multiple synonyms for a single category in a specific language was not discussed. The present study shows that this phenomenon is typical in Mandarin Chinese. Multiple color synonyms clouded previous searches for basic color terms but did not do so for basic color categories with the factor analysis methodology we used in the present study. These basic categories that can be translated into basic color terms in English have concentrated response maps comparable to the result of WCS.