On the Misuse of Ancient Chinese Inscriptions: An Astronomical Fantasy

For the view that the marks on the potsherds from Pan-p'o and other Yang-shao sites were word-signs, indicating the genesis of an indigenous Chinese script during the fifth millenium bc, see Ho Ping-Ti , The cradle of the East: An inquiry into the indigenous origins of techniques and ideas of Neolithic and early historic China, 5000–1000 B.C. (Hong Kong and Chicago, 1975), 223–67, 393–405.

The Shang diviners cracked the bones, interpreted the cracks, and then carved the divination topic onto the bone. For the nature of these inscriptions, see Keightley David N. , Sources of Shang history: The oraclebone inscriptions of Bronze Age China (Berkeley, Los Angeles, and London, 1978). The date of the historical period for which we have inscriptions is not firmly established. In the view of Tung Tso-pin and others, it spans the period from 1398 to 1112 bc; in my view it spans the period from ca 1200 to ca 1050 bc; for an introduction to the issues see op. cit., Appendix 4, “Absolute chronology: A brief note”; but see too my review, to appear in Harvard journal of Asiatic studies, xxxvii (1977).

The most recent Chinese discussion of the solar and lunar eclipses recorded in the oracle bone inscriptions is P'ei-Yü Chang , “Chia-ku-wen jih-yüeh-shih chi-shih ti cheng-li yen chiu”, T'ien-wen hsüeh-pao, xvi (1975), 210–34; Chang's conclusions need to be treated with caution. For preliminary discussion of a supernova record, see Needham Joseph , with the collaboration of Ling Wang , Science and civilisation in China: Vol. iii: Mathematics and the sciences of the heavens and the earth (Cambridge, 1959), 424–5. I translate the inscription (Hou-pien 2.9.1) as: “On the seventh day [after the day of divination], in the period dividing the night of chi-ssu (cyclical day 6) and keng-wu (cyclical day 7) there was a new great star standing together with the fire-star.” The record verified the king's forecast of troubles also carved on the same bone; I will return to the significance of this below.

Muller P. M. and Stephenson S. R. , “The accelerations of the Earth and Moon from early astronomical observations”, in Growth, rhythms and the history of the Earth's rotation, edited by Rosenberg G. D. and Runcorn S. K. (London, New York, Sydney, Toronto, 1975), 489. A drawing of the inscription is reproduced as no. 209 in Fang Fa-han [Frank H. Chalfant] and Po Jui-hua [ Britton Roswell S. ], K'u-fang erh-shih-ts'ang chia-ku pu-tz'u (The Couling-Chalfant collection of inscribed oracle bone) (Shanghai, 1935; reprint, Taipei, 1966). The actual turtle shell fragment is preserved in the Royal Scottish Museum, Edinburgh.

Muller and Stephenson , “Accelerations”, 408; cf. 469.

ibid., 523.

See Berry James F. , “Appendix I: Identification of the inscribed turtle shells of Shang”, in Keightley , Sources.

The day number in the Chinese cycle (which ran continuously from 1 to 60) may be found by dividing the Julian Day number by 60 and adding 50 or subtracting 10 from the remainder. Julian Day number 0 would have been cyclical day 50 (kuei-ch'ou). The solar eclipse in question would have occurred on JDN 1235805; dividing by 60 gives a remainder of 45; subtracting 10 gives the cyclical day number 35 (wu-hsü). It is not certain that the Chou, who conquered the Shang in the eleventh century bc, used a cyclical day calendar that meshed exactly with that of the Shang; if one (or more) cyclical day(s) were “lost” at that time, then JDN 1235805 would be equivalent to cyclical day 36 (or some more distant day).

Even a cluster of four lunar eclipse inscriptions, which I would date, on the assumption of no calendrical displacement, to the period from 1198 to 1180 bc, cannot resolve the issue conclusively. The 1198–1180 span accords well with a chronology based on the average reign lengths of the Shang kings, but in astronomical terms alone, some other clusters, derived from a calendar displaced by 1 day, 2 days, etc., cannot be excluded. Chang And , “Chia-ku-wen”, 222, assuming zero displacement (and employing a lunar eclipse canon whose accuracy may be questioned), offers dates that are over a century earlier. The issues are sinological as well as astronomical and cannot be entered into here. See too, ref. 16 below.

These comments may be taken as applying to Chang , “Chia-ku-wen”. The solar eclipse inscriptions are transcribed by Kunio Shima , Inkyo bokuji sorui (2d rev. ed., Tokyo, 1971), 160.2. Fresh discoveries of inscribed oracle bone continue to be made; over 4,800 pieces were discovered in 1973, for example. It is to be hoped that such finds will include more eclipse records.

See ref. 13, below.

The statement that “Jao Tsung-i (1959) has identified an inscription which may relate to a large solar eclipse” ( Muller and Stephenson , “Accelerations”, 489) may mislead the reader. Jao (Yin-tai chen-pu jen-wu t'ung-k'ao (Hong Kong, 1959), 83) transcribes K'u-fang 209 in modern Chinese graphs; at no point does he suggest it is an eclipse inscription and his punctuation shows that he would reject such an interpretation and would translate the inscription in the way I do below. It is true that Liu Chao-yang (“Yin-mo Chou-ch'u jih-yüeh-shih ch'u-k'ao”, Chung-kuo wen-hua yen-chiu hui-k'an, iv (1944), 113–14) treated this inscription as a solar eclipse record, but his scholarship, crude even at that time, has been superseded; and, it is worth noting, he assumed that the “eclipse” record referred to day 34, not to day 35. More recently, Willy Hartner (“Some hsin-wen about Shang-Yin”, in XIVth International Congress of the History of Science: Proceedings, iv (Tokyo and Kyoto, 19–27 August 1974), 6) has also taken K'u-fang 209 to be a record of a solar eclipse, that of 4 March –1230. This, for reasons he does not elaborate, would have the eclipse occurring on the day the divination was performed (day 33), which is not the day to which the charge refers; his paper reveals no familiarity with recent scholarship on the subject. Chang , “Chia-ku-wen”, quite properly, does not even mention K'u-fang 209 in his recent discussion of the Shang eclipse records.

The graph refers to some atmospheric phenomenon—“fog”, “mist”, “cloud” have been proposed as translations—but the exact meaning is not yet known. For an introduction to the scholarship, see Hsiao-Ting Li , Chia-ku wen-tzu chi-shih (Nankang, Taiwan, 1965), 2539–43. Muller and Stephenson apparently translate the graph as “on the following day”; this is crucial to their dating of the eclipse record, but there is no justification for such a reading.

Scholars have traditionally taken the charges to be questions; the charge would thus be translated as, “If, on the next day, ting-yu, we offer wine libation and beheading sacrifice, will it be clear weather?” For reasons why I think this view is mistaken, see Keightley , Sources, ch. 2, n. 7. The mode of the charge has no bearing, in any event, on the point at issue.

For a fuller discussion of this point, see Keightley , “Legitimation in Shang China”, mimeographed (Conference on Legitimation of Chinese Imperial Regimes, 15–24 June 1975, Asilomar, California), 13–15.

For an introductory study of these lunar eclipse records in English, see Dubs Homer H. , “The date of the Shang Period”, T'oung Pao, xl (1951), 322–35; “Postscript”, ibid., xlii (1953), 101–5. I am currently engaged in studying the date of these inscriptions, using Newton R. R. , A canon of lunar eclipses for the years –1500 to –1000 (Research report CP 054, The Applied Physics Laboratory, Johns Hopkins University, Laurel, Md, February 1977); the initial results are referred to in ref. 9, above.

Muller and Stephenson mistakenly translate ming as “it was a clear day”. For occurrences of the time phrases, ta-shih, hsiao-shih, and ming, see Shima , Sorui, 162.3, 398.2; it is estimated that the time of the great and small meals was ca 9–10 a.m. and 4–5 p.m., respectively ( Tso-Pin Tung , “Pu-tz'u chung chih ta-hsiao ts'ai yü ta-hsiao shih shuo”, Ta-lu tsa-chih t'e-k'an, ii (1962), 412). Muller and Stephenson have punctuated the four graphs ta shih jih ch'i (great/eat[ing]/day [or sun]/clear) as ta shih jih; ch'i, “there was a great eclipse of the sun; it was a clear day”. The fact that ta-shih (“at the time of the great meal”) and jih ch'i (“the day cleared”) have to be taken as separate phrases is fully attested by the parallel inscriptions given by Shima , Sorui, 283.4–284.3, 398.2.

The classic introduction to the Shang calendar is Tso-Pin Tung , Yin-li-p'u (2 vols, Nan-ch'i, Szechwan, 1945). For Tung's views on Shang intercalation, not fully supported, in my view, by the evidence, see pt 2. ch. 1, pp. 4b-5a; ch. 3, 30a-30b; ch. 5, 1a-4b.