² For further discussion on this aspect, see Zollinger, H., ‘Human color vision: an interdisciplinary research problem’. Palette, 40, 1972, 1–7.Google Scholar

³ Wittgentstein, L., Remarks on colour (Anscombe, G., ed., Oxford, 1978), I, 68.Google Scholar

⁴ For further philosophical discussion on Wittgenstein's theory of colour, see Westphal, J., Colour: some philosophical problems from Wittgenstein (Oxford, 1987).Google Scholar

⁵ See Harrison, B., Form and content (Oxford, 1973), 53–89Google Scholar; and Wyler, S., Colour and language: colour terms in English (Tübingen: Narr, 1992), esp. 22–28.Google Scholar

⁶ Lewis, L., Social anthropology in perspective: the relevance of social anthropology (Cambridge, 1976), 110.Google Scholar

⁷ Turner, V., The forest of symbols: aspects of Ndembu ritual (Ithaca and London, 1967), 69–71, 89–90.Google Scholar

⁸ See Sperber, D., Rethinking symbolism (Cambridge, 1975), esp. preface, 110–113.Google Scholar

⁹ Absolute chronologies reconstructed by scholars vary enormously; for a brief guide, see Keightley, D., Sources of Shang history: the oracle-bone inscriptions of Bronze Age China (Berkeley and Los Angeles, 1978, paperback ed., 1985), 227–228.Google Scholar

¹⁰ According to the Zhushi jinian : ‘From the time Pan Geng moved to Yin till the Zhou's annihilation, there were altogether 273 years; during this period the Shang did not again move their capital.’ However, the time varies in the different versions of the book; see Shiming, Fang and Wang Xiuling , ed. Guben zhushu jinian jizheng (Shanghai, 1981), 30Google Scholar. The name ‘Yinxu’ (Ruins of Yin) came into use soon after the Zhou conquered the Shang, and it is still conventional to use this name today. For an introduction to the place, see Zuobin, Dong, ‘Yinxu yanke ’. in Zhongyang ynajiuyan lishi yuyan yanjiusuo jikan (Bulletin of the Institute of History and Philology, Academia Sinica, abb. BIHP), no. 2, 1930, 224–240Google Scholar. (Repr. in Dong Zuobin xueshu lunzhu , Taibei, 1962, 199–215.)

¹¹ For a basic introduction to Shang archaeology, see Ji, Li, Anyang (Seattle, 1977)Google Scholar; Chang, K. C., Shang civilization (New Haven, 1980) and Archaeology of China (4th ed., New Haven and London, 1989), 295–367.Google Scholar

¹² The inscriptions on the ‘dragon bones’ first began to attract scholarly attention in about 1899, nearly 100 years ago, with the first collection of the bone inscriptions Tieyun cang gui compiled and published by Liu Er (1850–1909) in 1903. Within two years, Sun Yirang (1848–1908), an excellent scholar of the Chinese classics and ancient inscriptions, had written the first book on the decipherment of the inscriptions Qiwen juli . Although Sun's decipherment had many mistakes, his basic method became the starting-point for the study of oracle bone inscriptions by all later scholars. For almost a decade in the 1920s, before the scientific excavations at Yinxu started, Luo Zhenyu (1868–1940) and Wang Guowei (1877–1927) made remarkable progress in both the publishing and research of oracle bone inscriptions. Due to their efforts, the inscriptions became one of the most important sources of Shang history. For further information on the history of oracle bones studies, see Yuxin, Wang, Jiaguxue tonglun (Beijing, 1989), 320–367.Google Scholar

¹³ For a detailed introduction of the new method, cf. my thesis, ‘Colour symbolism’, 47a–56.

¹⁴ See Xigui, Qiu, Wenzixue gaiyao (Beijing, 1988), 151–156.Google Scholar

¹⁵ See Xiaoding, Li, ‘Cong liushu de guandian kan jiaguwenzi ’, Nanyang daxue xuebao , no. 2, 1968, 529–560Google Scholar. A table given by Li counts a total of 1226 oracle bone graphs, of which 27% are radical-phonetic compounds, 11% are loanwords, 23% are pictographs, 32% are compound ideographs, 2% are abstract symbols, and 6% are difficult to classify.

¹⁶ Xingwu, Yu, Jiagu wenzi shilin, (Beijing, 1979), 435–443Google Scholar. [This is a revised version of Shuangjianchi Yinqi pianzhi , 3 vols., Beijing, 1940–43.]

¹⁷ See Xigui, Qiu, ‘Hanzi xingcheng wenti de chubu tansuo ’, Zhongguo yuwen . 1978, no. 3, 168–169.Google Scholar

¹⁸ Berlin, B. and Kay, P., Basic color terms: their universality and evolution (Berkeley, 1969)Google Scholar. In their appendix, they have also provided an outline of previous studies.

¹⁹ ibid., esp. 134–51.

²⁰ In my thesis I examined the inscriptions containing evidence of the use of colour in Shang ritual and discussed the problems in interpreting the colour symbolism in relation to the later Wuxing theory; cf. Wang, T., ‘Colour symbolism’, in particular pp. 135–265.Google Scholar

²¹ Because of a semantic shift, this word usually means ‘naked’ in Modern Chinese (cf. Xiandai hanyu cidian , Beijing, 1988, 145)Google Scholar, but it also remains a colour term throughout the classical ages to the present.

²² cf. Jiaguwenbian (ed. Haibo, Sun, Beijing, 1965), 1238.Google Scholar

²³ cf. Jinwenbian (ed. Geng, Rong, Zhang Zhenlin and Ma Guoquan , Beijing, 1985), 1664.Google Scholar

²⁴ cf. Jiaguwen jishi (ed. Xiaoding, Li, Taibei. 1965Google Scholar, hereafter: Jishi), 3197. Jishi collected the decipherments by different scholars (before 1965) of each bone character and has served as a major reference for the study of OBI.

²⁵ Shuowen jiezi (Beijing, 1963, hereafter: Shuowen), 212.Google Scholar

²⁶ Several scholarly reconstructed systems of Old Chinese are available now, and among them, the system of Li Fankuei (Li Fanggui) is most widely accepted for its internal coherence. Li's system will be used throughout this study unless otherwise indicated. For Li's theory of Old Chinese, see ‘Shangguyin yanjiu ’, Qinghua xuebao , no. 9, 1–2, 1971, 1–61; and ‘Jige shanggu shengmu wenti ’. Zongtong Jianggong shishi jinian lunwenji (Taibei, 1976), 1143–50; and ‘Archaic Chinese’, in Origins of Chinese civilization (ed. Keightley, D., Berkeley, 1983), 393–408Google Scholar. The traditional method, that is, using Chinese characters to indicate the initial and final categories, will be used in this paper. For a brief introduction to the history of Chinese phonology, see Li, Wang, Hanyu yuyinshi (Beijing, 1985), esp. 17–81.Google Scholar

²⁷ Shuowen, 213.

²⁸ See Benedict, P., Sino-Tibetan: a conspectus (ed. by Matisoff, J., Cambridge, 1972)CrossRefGoogle Scholar, who has reconstructed the word ‘red’ in proto-Tibeto-Burman as *tsyak (no. 184, 46); see also Fagao's, Zhou review article in Zhongguo yinyunxue lunwen ji (Hong Kong, 1984), esp. Appendix I, 292Google Scholar, where Zhou provides his own reconstruction of Old Chinese for comparison.

²⁹ The meanings of the phrase qi li bu er in the inscription are not yet entirely clear, and here Yu Xingwu's interpretation has been followed and a tentative translation is provided; see Xingwu, Yu, Shilin, 328–329.Google Scholar

³⁰ It is written as a hewen , with an indication of the sex of the horse.

³¹ This character is found in the Shijing (Maoshi : 298), where Mao Heng's (c. first century B.C.) commentary says that it means ‘a sturdy-looking horse’; Maoshi zhengyi , juan 20.1 (Shisanjing zhushu , Beijing, 1980 [hereafter SSJZS], 610Google Scholar).

³² In Shang inscriptions, there are a number of words referring to the colour of horses' coats, such as bai mai (white horses), li ma (black horses) and bo ma (striped horses); for a further discussion, see Wang Yuxin , ‘Shangdai de ma he yangmaye ’, Zhongguoshi yanjiu , 1980, no. 1, 99–108. A later reference can also be found in the Shijing (Maoshi: 298), where we read that in Zhou royal horse farms there were many horses of various colours; cf. SSJZS, 609–10.

³³ A reference found in the Shanhaijing. Dahuang nanjing says that a ‘red horse’ lived in a mythical mountain; cf. Ke, Yuan, Shanhaijing jiao zhu (Shanghai, 1980), 384.Google Scholar

³⁴ This bone graph has multi-functions. In many inscriptions, it is transcribed as you (‘to have’), but in others it is broadly understood as you ‘to sacrifice’, ‘to offer’. For further discussions of the character, see Jishi, 2259–63, and D. Nivison, ‘The pronominal use of the verb yu (GIUG): ’. Early China, no. 3, 1977, 1–17; and K. Takashima, ‘Deciphment of the word yu in the Shang oracle bone inscriptions and in pre-classical Chinese’, Early China, no. 4, 1980, 19–29.

³⁵ For example, Li Pu discussed this inscription and claimed that chi is a method of hunting by setting fire to the forest to drive animals out; cf. Li, , Jiaguwen xuan zhu (Shanghai, 1989), 188–194.Google Scholar

³⁶ For a further discussion of this aspect, see Bingquan, Zhang, ‘Jiaguwen zhong suojian rendi tongming kao ’, Qingzhu Li Ji xiansheng qi shi sui lunwenji , vol. 2, (Taibei, 1967), 687–776Google Scholar. Chi is also to be seen in early Zhou bronze inscriptions as a personal name; see Xue zhong Chu fu ; cf. Chengyuan, Ma (ed. in chief), Shang Zhou qingtongqi mingwenxuan (Beijing, 1986–, hereafter Mingwenxuan), no. 823.Google Scholar

³⁷ These four names are mentioned in The Bamboo Annals, which record that the Nine Yi tribes came to the Xia dynasty for worship; cf. Fang's, and Wang's, Zhushu jinian, 9.Google Scholar

³⁸ Jiaguwenbian: 1155.

³⁹ Jishi, 3047.

⁴⁰ Maoshi zhengyi, juan. 22; (SSJZS, 615). However, in the Zhouli , a different usage ‘xing gang yong niu ’ is found, where the phrase xing gang probably refers not to animals but to a kind of reddish hard soil; cf. Yirang, Sun, Zhouli zhengyi , juan 30 (Beijing, 1987), 1184Google Scholar, where the xing, as Sun pointed out, is perhaps as a loan character for the original character with the ‘earth’ radical.

⁴¹ Maoshi zhengyi, juan 20.1; (SSJZS, 609–10).

⁴² Qiu Xigui has noted the possibility that xing might be read as a hewen in OBI; cf. Xigui, Qiu, Guwenzi lunji (Beijing, 1992), 84.Google Scholar

⁴³ This graph is understood as sheng (‘to ascend’), which probably refers to a sort of ritual; cf. Jishi, 4019–110. However, in his article ‘Shi jiaguwen li de “jiu” ’ (Zhongguo yuwen, 1985, no. 5, 384–8), Zhan Yinxin argues that the character shoud be read as jiu , meaning ‘to brand’.

⁴⁴ The decipherments of lao (with the ox-element, or with the sheep-element as ) are very different. The traditional explanation is that the former is tai lao ‘the combination of an ox, a sheep and a pig’; and the latter is shao lao ‘the combination of a sheep and a pig’. But, the majority of scholars now agree that they should be understood as ‘penned oxen’ and ‘penned sheep’. Cf. Jishi, 0313–6.

⁴⁵ Zhongshu, Xu (chief ed.), Jiaguwen zidian (Chengdu, 1988) lists this character as a variation of xing.Google Scholar

⁴⁶ This character is also used as a place-name in oracle-bone inscriptions; for example, Heji: 1141.

⁴⁷ There are sometimes exceptions; for example, Heji: 35986 is a very similar inscription to Heji: 36003 in which the ox-element and the sheep-element are, however, split up, as two independent characters.

⁴⁸ Pulleyblank, E.G., ‘The final consonants of Old Chinese’, Monumenta Serica, 33, 1977/1978, 183–187, 202–3.CrossRefGoogle Scholar

⁴⁹ This word probably also had a TB root; in reconstructed proto-TB, ‘red, crimson’ is *kyeng; cf. Benedict, P., Sino-Tibetan, no. 162, p. 45Google Scholar.; also see Fagao, Zhou, Yinynxue lunwenji, 1984, 293.Google Scholar

⁵⁰ Mingwenxuan: 395.

⁵¹ Shanxi wenwu gongzuo weiyuanhui , Houma mengshu (Beijing, 1976), no. 17:1, where the phrase should be read as .

⁵² Mingwenxuan: 534.

⁵³ Shuowen, 286

⁵⁴ ibid., 94.

⁵⁵ Maoshi zhengyi, juan 15.1; (SSJZS. 490). In the Shuowen quotation of the Shijing, the character xing is still written in the old form as . But in the received text of the Maoshi, the xing is written as , where the old phonetic element has been replaced by the new one. In fact, although the -element is employed as the phonetic, the phonetic value of the character xing is not the same as xin. They both share the same initial category, but xin belongs to the zhen final, and is reconstructed as <*hrjin. Phonologically, the zhen and geng final categories are close. Some evidence suggests that the change might have taken place during the Eastern Zhou period. Apart from the literary evidence in the Shijing and on the Qin stone drums, another piece of evidence has been found on a bronze jar of the Eastern Zhou period, the Cheng Xing hu , where the character xing is written as . It is noticeable here that the -element has replaced the -element as the phonetic sign. In this context, it is not entirely impossible that the phrase might be understood as an adjective describing the colour of the horns.

⁵⁶ cf. Moruo, Guo, ‘Shiguwen yanjiu ’, Guo Moruo quanji — kaogubian , vol. 6 (Beijing, 1982), 3–274Google Scholar; [first pub. Shanghai, 1939], in which a good rubbing of the inscription is provided; Guo's transcription is on p. 59. For a further discussion of the inscription, see Mattos, G., The stone drums of Ch'in (Monumenta Serica Monograph Series, no. 19, 1988), 144–145.Google Scholar

⁵⁷ Jiaguwenbian: 0984.

⁵⁸ Shuowen, 160.

⁵⁹ cf. Fagoa, Zhou (chief ed.), Jinwen gulin [hereafter Gulin] (Hong Kong, 1974/1975), 4929.Google Scholar

⁶⁰ Wieger, L., Chinese characters: their origin, etymology, history, classification and signification. A thorough study from Chinese documents (New York, 1965, first published in 1915), 223Google Scholar. This theory was based on the works of previous scholars, and evidence was found, as the sun and colour white seemed to be related in the early literature; see Junsheng, Zhu (1788–1858), Shuowen tongxun dingsheng (Beijing, 1984), 464–465Google Scholar; and Gulin, 4931–2, 4934–35.

⁶¹ Moruo, Guo, Jinwen congkao (Beijing, 1956), 181–182.Google Scholar

⁶² Cheng, Zhao, ‘Benzi tansuo ’, Gudai wenzi yinyun lunwenji (Beijing, 1991), 78–79.Google Scholar

⁶³ The character means ‘white silk’ in later literature, but it refers to a place-name in Shang inscriptions; for example, see Heji: 36842.

⁶⁴ cf. Jiaguwenbian: 0485.

⁶⁵ The graph is written as , probably depicting a castrated pig. See Jishi, 2985–6; and Yiduo, Wen, Wen Yiduo quanji (Shanghai, 1948), 539–544.Google Scholar

⁶⁶ An alternative transcription of the inscription is provided in Yinxu jiagu keci moshi zongji (Beijing, 1988), 440.

⁶⁷ The meaning of this graph is uncertain; it was probably a sort of sacrificial rite, which was often performed in hunting; see Cheng, Zhao, Jiaguwen jianming cidian (Beijing, 1988), 244.Google Scholar

⁶⁸ A number of inscriptions recorded that a larger wild animal, the was often chased and caught on hunting trips expeditions, and the colour of the animal was sometimes mentioned. The interpretation differ as to whether the animal was a wild buffalo or rhinoceros, see Lefeuvre, J. A., ‘Rhinoceros and wild buffaloes in north of the Yellow River at the end of the Shang dynasty’, Monumenta Serica, 39, 1990/1991, 131–157.CrossRefGoogle Scholar

⁶⁹ This graph is transcribed as , reading lin . In Chinese tradition, the lin or unicorn is an extremely rare beast and consequently carries great significance. See Zuobin, Dong, ‘Huo bailin jie ’ in Xueshu lunzhu, 217–271Google Scholar, in which Dong wrongly identified another graph si as lin; however, he has given a comprehensive discussion of the significance of the ‘unicorn’ in Chinese history. Some of the words in these inscriptions have not yet been fully deciphered, and even the understanding of the phrase bai lin sometimes differs; for instance, Hu Houxuan , in his article ‘Zhongguo nuli shehui de renxun he rensheng ’, Wenwu, 1974, no. 8, p. 63, read the word as a name, ‘Chief Lin’.

⁷⁰ See Xingwu, Yu, Shilin, 450.Google Scholar

⁷¹ Xiaosui, Yao, ‘Shangdai de fulu ’, Guwemi yanjiu , 6, 1981, 378.Google Scholar

⁷² Jiaguwenbian: 0083.

⁷³ cf. Jishi, 0317–18; also Guowei, Wang, Guantang jili (Beijing, 1959), 287Google Scholar; Wang did not recognize that the character could be an original hewen.

⁷⁴ cf. Jishi, 322; and a more lengthy discussion in Moruo, Guo, ‘Jiaguwen Yanjiu ’ Guo Moruo quanji: kaogubian, vol. 1, 83–92Google Scholar. [Originally Shanghai, 1931]. Later, in his Nulizhi shidai (Beijing, 1956), 7, Guo went on to use this linguistic evidence to illustrate an agricultural aspect, namely, that ploughing by buffalo was common in Shang times. However, oxen were usually used as sacrificial animals in Shang ritual, and there is as yet no evidence of buffalo being used for ploughing in Shang times. See Jinxiong, Xu. ‘Jiaguwen suo biaoxian de niugeng ’. Guwenzi yanjiu, no. 9, 1984, 53–74.Google Scholar

⁷⁵ Shuowen, 30.

⁷⁶ cf. Hanizi da idian (Chengdu, 1986–1990), 1805–6.Google Scholar

⁷⁷ Discounting all the literary records handed down through generations, the earliest example known of the character wu being used in the sense of ‘things’ is found in the inscription on a bronze vessel from Zhongshan State of the Warring States period. It reads: ‘jian yu tianxia zhi wu , Knowing of things under Heaven’ (Mingwenxuan: 880).

⁷⁸ cf. Jishi, 0318–9.

⁷⁹ Xigui, Qiu, ‘Shi “wu” “fa” ’, Guwenzi lunji, 70–74Google Scholar.

⁸⁰ Guoyu (Shanghai, 1988), juan 18, 565.

⁸¹ Zhouli zhengyi, juan 30, 1182–3.

⁸² ibid., juan 24, 937. The phrase wu se is still used in modern Chinese, but now only means ‘to choose’.

⁸³ ibid., juan 51.

⁸⁴ See Sun's commentary; ibid., 2124–5. The Shang already practised the divination method of observing the clouds; there are phrases such as ge yun (Heji: 10405, 10406, 21021, 21022), liu yun (Heji: 33273, Tunnan: 1062) and wu yun (Tunnan: 651) in OBI.

⁸⁵ ibid., juan 53, 2200, where Sun Yirang has also provided a very informative comment on the ‘Nine Flags’ which are distinguished by patterns and colours.

⁸⁶ A similar account is also found in the Shuowen. However, in Shuowen jiezi zhu (Shanghai, 1981), 453–4, Duan Yucai (1735–1815) noted that the character wu here should be written as ; and in Shuowen shili (Beijing, 1987), 446, Wang Jun (1784–1854) suggested that the phrase wu wu here should be read as cong cong .

⁸⁷ In the Shuowen, the ancient form of li is written as , (P. 91).

⁸⁸ ‘Qinshi’ (‘’), this particular document has been identified as a later forgery rather than original Zhou writing. See Shangshu zhengyi, , juan 11 (SSJZS, 181), where Kong Yingda noted that li might be better understood as ‘spotted’.

⁸⁹ Zhanguoce (Shanghai, 1985), juan 3, 85–6, where it is pointed out that the character li is interchangeable with . See also Yixing, Hao (1757–1825), Erya yishu (Beijing, 1982), ‘Shigu diyi –’. 23–24.Google Scholar

⁹⁰ Yinzhi, Wang, Jingyi shuwen (Shanghai, 1936), 266–267.Google Scholar

⁹¹ ‘Dongshanjing’ (‘’), see Ke, Yuan, Shanhaijing jiaozhu, 101Google Scholar. In Shanhaijing jian shu (Chengdu, 1985), juan 4, Hao Yixing cited another reference in his commentary, stating that the character li here means ‘multi-colour’, more precisely ‘a black pattern on a yellowish background’.

⁹² Lunyu zhu shu , juan 6 (SSJZS, 2478).

⁹³ Jin Xiangheng, ‘Shi wu ’, Zhongguo wenzi , 8, 1968, 2.

⁹⁴ cf. Jishi, 0321–23.

⁹⁵ See Xigui, Qiu, ‘Shi “wu” “fa”’, Guwenzi lunji, 70–74.Google Scholar

⁹⁶ This inscription was published by Zhiyu, Shen, ‘Jiagu buci xinhuo ’, Shanghai bowuguan jikan , 3, 1986, 161.Google Scholar

⁹⁷ Xigui, Qiu, ‚Jiaguwen zhong suojian de Shangdai nongye , Guwenzi lunji, 165Google Scholar, where Qiu argues that in OBI the character wu is often in contrast to the colour word xing , and it probably then refers to a darkish multi-coloured ox.

⁹⁸ Jiaguwenbian: 1511, where only the second type is printed.

⁹⁹ cf. Jishi, 0333.

¹⁰⁰ See Xingyan, Sun (1753–1818), Shangshu jinguwen zhushu (Beijing, 1986), 154–155.Google Scholar

¹⁰¹ Karlgren, B., ‘The Book of Documents’, Bulletin of the Museum of Far Eastern Antiquities (hereafter BMFEA), no. 22, 1950, 14Google Scholar.

¹⁰² Shangshu zhengyi, juan 6 (SSJZS, 148).

¹⁰³ Xianqian, Wang (1842–1917), Shiming shu zheng bu (Shanghai, 1984), 18Google Scholar.

¹⁰⁴ For example, see Mingwenxuan: 229 (Dou Bi gui ) and 252: (Mian gui ).

¹⁰⁵ For different opinions, see Gulin, 7003–18. In his Xi Zhou ceming zhidu yanjiu (Beijing, 1986), 226–228Google Scholar, Chen Hanping argues that the character zhi refers to a ‘yellow’ colour in the context.

¹⁰⁶ Mingwenxuan: 225.

¹⁰⁷ See Shaoming, Lian, ‘Shi Qiang pan yanjiu ’, Guwenzi yanjiu, no. 8, 1983, 35–36Google Scholar. E. L. Shaughnessy has annotated and translated this inscription into English and seems to follow Lian's interpretation of these two relevant sentences. The translation reads: ‘Even-horned and redly gleaming, appropriate were his sacrifice.’ Furthermore, Shaughnessy made the following observation of the graph, ‘ perhaps the two hands holding a basin beneath the etymonic compound are intended to imply a mirrorlike brightness’; cf. Sources of Western Zhou history: inscribed bronze vessels (Berkeley, Los Angeles and Oxford, 1991), 183–192, esp. 190Google Scholar. A number of articles have been written on the inscriptions of the Qian pan; they include Li Xueqin , ‘Lun “Shi Qiang pan” jiqi yiyi ’, Xinchu qingtongqi yanjiu (Beijing, 1990), 73–82; Xigui, Qiu, ‘“Shi Qiang pan” ming jieshi ’. Guwenzi lunji, 371–385Google Scholar; Xingwu, Yu, ‘“Qiang pan” mingwen zhier jie , Guwenzi yanjiu, no. 5, 1981; 1–16Google Scholar; and Cheng, Zhao, “Qiang pan” mingwen bushi ', Guwenzi yanjiu, 5, 1981, 17–26Google Scholar. Since scholars' reading of these sentences vary greatly, other possibilities cannot be ruled out.

¹⁰⁸ For example, Heji, 30036, 30037, 32775.

¹⁰⁹ For more examples, Heji: 5068, 5165, 1535, 16101, 16102, 16103, 16104, 16105. These are mainly from the early period, namely, the Bin, Li, and Zi diviners' groups.

¹¹⁰ Qiu Xigui, ‘Shuo jiagu buci zhong “zhi” zi de yizhong yongfa ’, Guwenzi luji, 111–116Google Scholar.

¹¹¹ Xingwu, Yu, Shilin, 182–184Google Scholar.

¹¹² There are several examples containing the phrase zhi and niu together, but unfortunately most of them are too fragmentary to be used as hard evidence. For example, Heji: 8969, 15761, 16229 and 23000, in which the character zhi is likely to be a name or a verb rather than an adjective.

¹¹³ The number of inscriptions including ri you zhi is great (for example, Heji: 33696–704, 27388, 29697–9, White: 1371); and the majority of the inscriptions are from the Li and Wuming groups, Tunnan: 726 reads: ‘yue you zhi’ in which it relates to the moon.

¹¹⁴ Moruo, Guo, Yinqi cuibian (Beijing, 1956), 13Google Scholar.

¹¹⁵ Mengjia, Chen, Yinxu buci zongshu (Beijing, 1956), 240Google Scholar.

¹¹⁶ See Shaofeng, Wen and Yan Tingdong , Yinxu buci yanjiu-kexue jishu pian (Chengdu, 1983), 29–31Google Scholar; and Liang, Zhang, ‘Xin chutu de “yue you zhi” bukao , Zhongguo tianwenshi yanjiu 1, 1984, 119–128Google Scholar.

¹¹⁷ Ito Michiharu, ‘Zhi zi kao ’ (unpublished paper presented at the Conference of the Xia-Shang Culture Studies, Luoyang, 1991).

¹¹⁸ As mentioned earlier, Luo Zhenyu was the first to decipher it as a colour term, and he transcribed the graph as .

¹¹⁹ There is an earlier Wuming-group inscription, Heji: 30718, in which the character zhi was used as an adjective, but the inscription is too fragmentary to be certain. Also, in the Huang-group, zhi is still used in a way referring to the ritual sacrifice; for example, Heji: 38115, but an epigraphic distinction has been made to it, written as .

¹²⁰ The translation of huang xiang as ‘a bronze vessel’ is very tentative; see Guo Moruo's ‘Zai Feng gu keci ’; cf. Guo Moruo quanji–kaogu bian, vol. 1, 405–10.

¹²¹ See Deming, Lu (c. 550–630), Jingdian shiwen (Shanghai, 1985), 683, 759, where he notes that zhi and te are interchangeable in the Liji textGoogle Scholar.

¹²² Zhouli zhengyi, juan 62, 2613.

¹²¹ Shuowen, 29.

¹²⁴ Yucai, Duan, Shuowen jiezi zhu, 50Google Scholar.

¹²⁵ Shangshu zhengyi, juan 3 (SSJZS, 127); see Kong Yingda's commentary.

¹²⁶ Chunqiu Zuozhuan zhengyi , juan 35 (SSJZS, 1974); see Du Yu (222–284)'s commentary.

¹²⁷ Guoyu, juan 8, 286; see Wei Zhao (204–273)'s commetary.

¹²⁸ ibid., juan 18, 564–5; Wei Zhao commentary.

¹²⁹ Liji zhengyi , juan 25 (SSJZS, 1444).

¹³⁰ See Tang Lan , ‘Maogong ding “zhu fu, congheng, yuhuan, yu tu” xinjie , Guanming ribao , 9 May, 1961, where he argues that the pictographic form of the character is perhaps a religious person such as a shaman.

¹³¹ See Xigui, Qiu, ‘Shuo buci de fen wu wang yu zuo tu long ’, Guwenzi lunji, 218Google Scholar.

¹³² Jiaguwenbian: 1606.

¹³³ Xinwu, Yu, Shilin, 227–230Google Scholar.

¹³⁴ This is probably a pictograph of a bier , but whether it is used here in its original meaning is uncertain; see Cheng, Zhao, Jiaguwen cidian, 224Google Scholar.

¹³⁵ See Yun, Yan, ‘Shangdai buci zhong de yezhu shiliao ’, Kaogu, 1973, no. 5, 299Google Scholar.

¹³⁶ Jinwenbian: 2207.

¹³⁷ Guo also argues that the graph depicts a man wearing an archaic jade pendant; cf. Jinwen congkao, 174–86.

¹³⁸ Mingwenxuan: 201, 471; Maoshi: 172, 246 and 302.

¹³⁹ See Yixing, Hao, Erya yishu, 23–25Google Scholar.

¹⁴⁰ Mingwenxuan: 301.

¹⁴¹ See Chusheng, Chen, Jinwen changyong cidian (Xi'an, 1987), 1107–1108Google Scholar. Other phrases such as Huang di (‘Yellow lord’) and huang zhong (‘Yellow tune’) can be found on the late Zhou bronzes.

¹⁴² Shuowen, 211.

¹⁴³ This form occurs in several early Zhou bronzes. As Chen Zhaorong argued recently, the bone graph cannot be linked directly to the later character hei and the characters may have different origins. Chen also discussed the problem of the confusion between huang and hei. See Zhaorong, Chen, ‘Guwenzi zhong de [ ] ji cong [ ] zhu zi , Hanxue yanjiu , 6.2, 1988, 135–173Google Scholar.

¹⁴⁴ For example, Heji: 10170, 10181, 10187, 10184. All these inscriptions bear the character hei, but are written differently. By not recognizing that this is no more than a variation, many scholars here misread this character; for example:

Cuibian: 551: use/jin-black/ox

Here, Guo Moruo read it as loanword of the character jin , meaning ‘red’; see Cuibian kaoshi, 20. Tsung-Tung Chang and Serruys read it literally as a pictograph ‘to burn a hunchback on fire’ or ‘to use (as victim) a hunchback’; see Chang, , Der Kult der Shang-Dynastie: im Spiegel der Orakelinschriften (Wiesbaden, 1970) 249Google Scholar; and Serruys's ‘Language of Shang oracle bone inscriptions’, Toung Pao 60, 1–3, 1974, 1050, n. 32.

¹⁴⁵ Jinwenbian: 1662.

¹⁴⁶ See Lan, Tang, ‘Shaanxisheng Qishan Dongjiacun xin chu Xi Zhou zhongyao tongqi mingci de shiwen he zhushi ’, Wenwu, 1976, no. 5, 63Google Scholar.

¹⁴⁷ In the Shuowen (p. 290), it is written as , and also as , meaning ‘a clay soil’; and its guwen form is written as . In Zhou bronze inscriptions, jin is sometimes used attributively to modify a jade tablet, e.g. jin zhang , in which the character jin is also understood as jin ‘a kind of precious stone’; see Mingwen xuan: 434 (Song ding ), 435 (Song gui ), 436 (Song hu ). But, it is not entirely impossible that the character jin there should be read as a term referring to dark or black colour. Many surviving archaic jade zhang-tablets are black.

¹⁴⁸ Lan, Tang, Yinxu wenzi ji (Beijing, 1981, first pub. 1934), 82–86Google Scholar.

¹⁴⁹ Xingwu, Yu, Shilin, 229–230Google Scholar.

¹⁵⁰ The later phonetic-compounds which bear the jin-element as their phonetic element also clearly divide into two different categories, for example: (a) and (b) . In the (a) category, the characters mostly have velar initials and wen <*-∂n or zhen <*-in final categories; the (b) category have various initials such as ni <*n-, xiao <*h- and tou <*th-, and their finals are mostly the same, yuan , <*-an. The evidence here seems to suggest that there was a movement from the central vowel -∂- to the low vowel -a- in that period; and that the palatization caused by the -j- and -r- may have been an important factor in such a transition. Tang Lan has once suggested that the zhen and yuan categories are interrelated in proto-Chinese; see Yinxu wenzi ji, 79–81, 86.

¹⁵¹ Jiaguwen bian: 0533.

¹⁵² Shuowen, 84.

¹⁵³ cf. Gulin, 2481–3.

¹⁵⁴ Shuowen, 84.

¹⁵⁵ See Hu Houxuan, ‘Shi “ziyong” “zi yu” , BIHP, 8.3, 1940, 467–84.

¹⁵⁶ It is sometimes interchangeable with the character zi ‘black’; cf. Shuowen, 84.

¹⁵⁷ In the inscription on the Zhongshanwang Cuo ding , you is written as a , this shows that the phonetic value of and are the same; cf. Mingwenxuan: 880.

¹⁵⁸ In his doctoral thesis, ‘Yin Zhou guwen tongyuan fenhua xianxiang tansuo ’ (Jilin Daxue , 1991), 180, Wang Yunzhi argues that these characters can be traced back stage by stage to the same Shang root <*?j∂u.

¹⁵⁹ For example, Mingwenxuan: 826: Zhugong Jing zhong , 827: Zhugong Hua zhong .

¹⁶⁰ ibid., 538: Wuwang Guang jian .

¹⁶¹ For example, see Mingwenxuan: 201: Shi Wangfu ding , 202: Shi Cai ding .

¹⁶² ibid., 416: Liu ding ; in this attributive phrase the character you is sometimes interchangeable with cong , for example, see also Maogong ding .

¹⁶³ Hanping, Chen, Ceming zhidu, 286–293Google Scholar.

¹⁶⁴ See Maoshi zhengyi, juan 15:2 and 15:3 (SSJZS, 495, 501).

¹⁶⁵ See Guying, Chen, Laozi zhu yi ji pingjie (Beijing, 1984), 85Google Scholar.

¹⁶⁶ cf. Hanyu da zidian, 280–81, 1094–5.

¹⁶⁷ For example, in the Zhaohun (‘’): ‘Souls, please return! Stop your descent to that Dark Capital’; in the same chapter, we read: ‘Flying fire is rising continually; the dark-red face turns steaming red.’ The youdu is known to be the underworld, and the phrase xuanyan ‘dark-red face’ refers, as Wang Yi (c. A.D. second century), pointed out, to heaven. See Xingzu, Hong (c. A.D. twelfth century), Chuci buzhu (Beijing, 1983), 201, 213–14Google Scholar.

¹⁶⁸ The Greek word for ‘fire’ (pyr) and the Latin word burrus meaning ‘fiery red’ derived from the same ancestral form. This led to Romance forms meaning ‘dark red’ (Provençal burel meant ‘brownish red’) and Old French buire (variant bure), meaning ‘dark brown’. Cf. Bolinger, D. and Sears, D., Aspects of language (New York, 1981), 266Google Scholar.

¹⁶⁹ For a further discussion on the attributive phrases in Shang inscriptions, see Takashima, K., ‘Noun phrases in the oracle-bone inscriptions’. Monumenta Serica, 36, 1984/1985, esp. 263–267CrossRefGoogle Scholar.

¹⁷⁰ Evans-Pritchard, E. E., The Nuer: a description of the modes of livehood and political institutions of a Nilotic people (Oxford, 1940), esp. 41–48Google Scholar.

¹⁷¹ Turton, D., ‘There's no such beast: cattle and colour naming among the Mursi’, Man, 15, 1980, 230–238CrossRefGoogle Scholar.

¹⁷² In the 1940s, Hu Puan tried to reconstruct the evolutionary history of Chinese colour terms. His starting-point was that the evolution of the colour vocabulary was associated with the evolution of human society; therefore his reconstruction is as follow: bai-‘white’ came first; it is a pictograph of a man's face, indicating the recognition of man himself; second is chi-‘red’, which is associated with ‘fire’; it probably emerged in the hunting stage; third huang-‘yellow’ derived from the ‘soil’, indicating the development of agriculture; and fourth hei-‘black’ came with a settled and civilized society, the form of the character representing cooking ‘smoke’. See Hu Puan, ‘Cong wenzixue shang kaojian gudai bianse benneng yu ranee jishu ’, Xuelin , 3, 1941, 53–67. However, while this kind of view reflects early academic interest in the study of colour terms, it has little value for a real understanding of colour terms and colour categorization, for it is too simplistic to account for any social or linguistic evolution.

¹⁷³ Berlin and Kay's theory since has been modified; see Kay, P. and McDaniel, C. K., ‘The linguistic significance of meanings of basic color terms’, Language 54, 1978, 610–646CrossRefGoogle Scholar.

¹⁷⁴ Baxter, W., ‘A look at the history of Chinese color terminology’, Journal of the Chinese Language Teachers Association, XVIII, 2, 1983, 1–25Google Scholar. Here, Baxter lists the character zhu as one of the early basic terms. Although the form can be found in OBI, it is never used as a colour term. It is, however, used as the colour term representing red in Zhou bronze inscriptions.

¹⁷⁵ See Baines, J., ‘Color terminology and color classification: ancient Egyptian color terminology and polychromy’, American Anthropologist, 87, 1985, 282–297CrossRefGoogle Scholar.

¹⁷⁶ Conklin, H., ‘Hanunoo color categories’. Southwestern Journal of Anthropology, 11/4, 1955, 339–344CrossRefGoogle Scholar.

¹⁷⁷ Shuowen, 291.

¹⁷⁸ In his study of the semantic system of Shang language, Zhao Cheng noted that the colour terms may cover domains much broader than their modern lexicography; he thinks that in OBI hei-‘black’ covers darkish grey, bai-‘white’ covers light grey and yellow, huang-‘yellow’ covers brown and chi-‘red’ covers bright yellow or orange. See Cheng, Zhao, ‘Jiaguwen ciyi xitong tansuo ’, Lunwenji, 101–102Google Scholar.

¹⁷⁹ For further discussion, see Conklin, H., ‘Color categorization’, American Anthropologist, 75, 1973, 931–942CrossRefGoogle Scholar.

¹⁸⁰ cf. Jishi, 1739, where Li Xiaoding argues that the character appears in one inscription in which it was probably used as qing-‘green’, but evidence shows that it should be read as nan ‘south’, rather than qing-‘green’.

¹⁸¹ Many scholars have read the qing as a loan character for jing ‘silent’, and translated the phrase as ‘the silent and mysterious ancestors’; for example, see Li Xueqin, ‘Lun “Shi Qiang pan” jiqi yiyi’ (see n. 167 above).

¹⁸² See Shimizi Shigeru , ‘Shuo qing ’, Wang Li xiansheng jinian wenji (ed. Xianggang yuwen xuehui , Hong Kong, 1987), 141–62.

¹⁸³ Maoshi zhengyi, juan 3.2: (SSJZS, 321).

¹⁸⁴ ibid., juan 15.3 (SSJZS, 501).

¹⁸⁵ ibid., juan 4.1; (SSJZS, 330).

¹⁸⁶ ibid., juan 6.4; (SSJZS, 373).

¹⁸⁷ Guo Qingfan (1844–1896), Zhuanzi jishi (Beijing, 1961), 4, 14.

¹⁸⁸ Baxter, W., ‘Chinese colour terminology’, 17–18Google Scholar.

¹⁸⁹ De-k'un, Cheng, ‘The T’u-lu colour-container of the Shang-Chou period’, BMFEA, 37, 1965, 239–250Google Scholar. The name ‘T’u-lu’ was taken from an inscription inscribed on one of the containers.

¹⁹⁰ The specimen was analysed by the Research Laboratory for Archaeology at the University of Oxford. A report was included in Cheng's paper.