The characteristics of different classes

The four histograms in Figure 4-3 give a statistical overview of the collected data. The first top histogram presents the occurrence rate of each sub-class. The researcher noticed the landmark basic terms, namely the LB class members red, green, yellow and blue do not behave uniformly, thus need to be singled out and marked separately. The second histogram

Colorterms

37

shows the mean response time (ms) of the classes. Within the LB class green and red responses result in the shortest RTs, but the RTs for blue and yellow are longer. The RTs in LB are generally shorter than that in other classes. In the compound classes, the MB (modifier-basic) class shows a relatively short RTs, indicate the combination of the common modifier and basic terms are easier to retrieve than other forms of compound.

0 5 10 15 20 25

LB(G) LB(R) LB(B) LB(Y) OB S A BB SB MB MS C

0 1000 2000 3000 4000 5000 6000

LB(G) LB(R) LB(B) LB(Y) OB S A BB SB MB MS C

0 1 2 3 4

LB(G) LB(R) LB(B) LB(Y) OB S A BB SB MB MS C

Occurrence%

MRTs(ms)

MCR

38

Figure 4.3. From histogram of top to bottom: occurrence (%), mean response times, mean confidence rating, and numbers of Chinese characters of collected data. More details are given in the text.

The third histogram is the results of mean confidence rating (MCR), which is an evaluation score of each naming trial given by participants them self. RTs and CR are both sensitive indexes to task difficulty in color naming research. The MCR of the LB class is higher than the other complex forms of color terms, and red responses gained the highest MCR. The bottom histogram shows the frequency distribution of the color terms composed of one to more than four Chinese characters. The number of composing characters is a simple but effective index reflecting the complexity of color term peculiar to Chinese Mandarin.

From the current results, the two-character color terms are in majority.

The detailed descriptive statistics of the collected data are summarized in Table 4-1, including the examples of color terms and the mean response time (MRTs), mean confidence rating (MCR), and the occurrence rate of each class. The sub-classes of color terms in each row correspond to that in Figure 4-2. Only the color terms that were named more frequently (frequency count more than 87, i.e. over 2% occurrence to total amount) would appear as examples in the column of Mandarin and English translations. Notice that the column of MB (modifier-basic) class only listed the modifiers. The color terms shown in this table were reported with relatively higher consistency and frequency during the free naming experiment.

There is no color term surpassing 2% occurrence threshold (87 repetition times) in the MS (modifier-secondary) class as this form of color term is rarely used by Mandarin speakers.

0 10 20 30 40

1 2 3 4 >4

no.ofcharacters

39

Also the C (complex) class contains no item of high consensus. There are 39 color terms (for the MB class researcher counted the frequency of modifiers in order to bring out the modifier terms in common used) standing out when screened by the 2% occurrence consistency and frequency threshold. These terms provide valuable information about color idioms in current use.

Table 4.1. Statistic summary of collected terms (each occurrence percentage over 2%)

Class Mandarin Eng. Translation MRTs (SD) MCR (SD) %

Bright, dark, pale, powder, light,

deep, -ish, central/correct 3960 (1578.29) 3.60 (1.03) 24.9%

MS 5142 (1910.88) 2.73 (1.27) 1.8%

C 5406 (1259.93) 2.42 (1.09) 3.96%

There are diverse and larger amount of color terms recalled in previous free-recall survey, but in the task-orientated naming work only 39 color terms or modifiers are repeatedly used.

The pattern of these terms is described as following:

(a) The LB class: this class contains the perceptually unique hue, the most familiar and salient

40

basic color categories, red, green, blue and yellow. It thus results in relatively short RTs and high CR compared with other classes. The green and red responses are more confident and rapid than that of blue and yellow. The usage of the term blue results in the longest RTs, lowest CR and least consensus (as in the Figure 4.4 in next section), which is a point worthy of further study in itself.

(b) The OB class: contains terms representing the rest of B&K’s basic color categories except for black and white. The brightness of the stimuli was constrained at 50cd/m². The terms for describing color categories of purple and grey are࿫(Zi) and ۊ(hui) respectively.

However, similar to the phenomenon in free-recall survey, there are three different terms,

࠼೽(Ka-fei), ᓣ(He) and ཝ(Zong), were used for representing brown category. Likewise, the pink category was divided into ృદ(Fen-Hong)or ృ(Fen), and ௒દ(Tao-Hong), and orange category can be both ᖪ(Ju) or ᖨ(Cheng). Among these terms, ࠼೽, ృદ and ᖪ are the most frequently recalled and used terms corresponding to brown, pink and orange categories in both survey. However, these terms are translated inconsistently in Chinese-English dictionaries and in previous studies regarding Mandarin color terms (Lin, et al., 2001; Lu, 1997). It is therefore necessary to determine whether these terms are synonyms denoting the same color categories, or these are sub-categories used to represent different shades within or between major categories. The S class: There is a remarkable discrepancy of number of the collected secondary color terms in free-recall and free-naming works. There is only one term ᢎ(Ou, lotus root), a shade of pale grayish pink, surpassed 2% occurrence. Other terms overlapping with those in the free-recall task such asᗤ(Zhan, brick), ׬ࠡ(Ka-qi, khaki) or Ւ(Tu, soil) were sparsely reported. One possibility is that the stimuli for naming are limited, while many secondary terms cover the brightness range beyond the 50 cd/m² level , such as ᓅ(Fu, skin) or ۏ(Mi, rice) are used for representing very light warm colors. Another reason is the secondary color terms are rarely used alone in the task-oriented naming experiment. Many secondary terms were

41

used as hue-modifiers in front of basic colors terms, such as sea-blue, olive-green or grape-purple. This sort of color terms are classified as the SB class, which results in higher occurrence rate (17.1%) than S class (6.8%).

(c) The A class: contains color terms appeared in ancient classical Chinese literature or pigments of Chinese painting, such as ॹ (Ching), a widely applied color term representing the colors from green to blue, sometime even covers purple and black. This word was unsurprisingly mentioned for over 2% occurrence rate in the experiment, mainly used for naming stimuli of bluish green or light green. In contrast to the color terms in Chinese vernacular, a single antique color terms are rarely used in everyday color reference. But some of them served as hue-modifier in the naming task, such ڹદ (Zhu-Hong, bright red).

(d) The BB class: accounts for 19.6% occurrence, which is a very common pattern of compound color terms. This sort of color terms is composed of two basic color terms in order to describe the ambiguous shades locating between two color categories. However, due to the flexibility of Chinese grammar, the syntax of this type is somewhat controversial since the precedent basic term is difficult to tell its functional identity as a noun or a hue modifier. For example, ៴ጸ could means blue-green or bluish green in English. Currently the researcher treats these as simple compound nouns. The combinations of two basic terms seem to be constrained in some conventional rules and perceptually-logical order, instead of arbitrarily piecing any two terms together. One of the obvious rules is the composed color terms are neighboring in the hue circle, such blue-green and orange-red. Terms like yellow-red or green-purple were never found.

Secondly, there are idioms with specific combination order using for describing the color located between two color categories. For example, the color between yellow and orange would be always named as ᖪ ႓ (orange-yellow) but never named as ႓ ᖪ (yellow-orange). Generally, red, yellow green are mostly placed as the second noun.

42

(e) The SB class: contains the compound terms composed of a secondary color term or noun with a basic color terms. The frequently occurred terms including olive-green, grass-green, sky-blue, taros-purple, ink-green, soil-yellow and grape purple. It is remarkable that there is larger diversity in the usage of secondary terms or nouns for describing colors in green, blue and purple categories.

(f) The MB class: this sort of compound terms is composed of a tone modifier and a basic color term. The front modifier could modify the perceptual dimensions of hue, brightness or saturation of the main color term. The frequently used modifiers includingॽ(bright), ᄆ(dark), ෍(light), ෡(deep), ෉(pale), ృ(powder), ೣ(-ish) and إ(canonical/correct).

The first two correspond to the brightness of the color, while the third and the fourth are more related to the saturation level. The modifier ෉(pale) is usually used to describe a lighter tone than ෍(light), and ృ(powder/whitewash) is a unique Mandarin color modifier used to describe light, clean colors with opaque quality. The color terms modified by ృ are always associated with feminine or childishness. The modifier ೣ(-ish) is used to modify the hue of a color, such ጸೣ႓ or ೣ႓ጸ(yellowish green), which is in contrast to إ(canonical/correct). The modifier إ is always used as a stimulus approximates the focal color of a color categories, especially in the cases of landmark basic terms.

(g) The MS class: is a type of combination of modifier and a secondary color term. This sort of compound terms resulted in a low occurrence rate of 1.8%, which is lower than that in the English naming study(Guest & Van Laar, 2000).

(h) The C class: includes complex composition of more than two color terms and modifiers, such as ෍៴ጸ (light blue-green) or ೣ႓ᖬᨛጸ (yellowish olive green). There is no term surpassed 2% consensus level in this class. This sort of color terms appear more frequently than the Antique and MS classes (3.96%, 2.4% 1.8% respectively), and more than the rate (1.4%) of the same type of polylexemic terms observed in the English