Graphs, words, and meanings: three reference works for Shang oracle-bone studies, with an excursus on the religious role of the day or sun.

• Author: Keightley, David N.

The three works under review should do much to open up the field of Shang oracle-bone inscriptions to scholars with general sinological training. All those who, without immersing themselves in the challenges of epigraphic interpretation, wish to address particular questions to these earliest Chinese records will be greatly assisted by these new publications. The present review is intended as a guide to their use.(1)

The nature and virtues of Yinxu jiagu keci moshi zongji (Comprehensive Collection of Copies and Interpretations of Graphs Inscribed on the Oracle Bones from Yinxu [hereafter abbreviated as Moshi]) can be quickly described. It provides handcopies, in the numerical sequence in which they were published, of all the oracle-bone inscriptions reproduced in five major collections of rubbings: Heji (pp. 1-960), Tunnan (pp. 961-1049), Yingcang (pp. 1051-1105), Tokyo (pp. 1107-30), and White (pp. 1131-62).(2) I estimate that inscriptions on some 58,000 pieces of oracle bone are included in Moshi.(3) The handcopies transcribe each inscription twice, giving the jiagu [Chinese Text Omitted] forms in the top row of the register (hence the mo [Chinese Text Omitted], "copy," of the title) and their modern kaishu [Chinese Text Omitted] forms (hence the shi [Chinese Text Omitted], "interpretation") in the bottom row. Thus, to give a short example, one of nine inscriptions on HJ 272[f.sup.4] is reproduced at p. 10.2 (i.e., the second register on p. 10) in oracle-bone script as [Chinese Text Omitted] (top) and then, in modern script, as [Chinese Text Omitted] (below), which can be translated as, "Divined: 'To Ancestor Xin, perform an exorcism.'"

The value of Moshi is evident. Within the compass of its 1,162 pages one has ready access to most of the published oracle-bone inscriptions of scholarly worth excavated at Xiaotun. By providing a readily accessible, if preliminary, decipherment, in modern characters, for any particular inscription reproduced in Heji, Moshi permits the rapid testing of any scholarship that makes appeal to the Heji inscriptions. It also enables scholars to place a particular inscription in the context of other inscriptions on the same bone. If, for example, one wants to consider the spiritual climate in which the Shang were divining, as transcribed above, the exorcism to Ancestor Xin, the Moshi entries for HJ 272f and 272b transcribe twenty-five additional inscriptions on the same plastron fragment; these indicate, among other things, that the king had experienced a potentially ominous dream (but see too page 511, below), and that the Shang were also divining at this time about offerings to Ancestor Yi and Ancestor Mother Geng. A whole series of contextual questions, in fact, can now, through the pages of Moshi, be answered in this way. One could, to pick an example at random, determine the months (if recorded elsewhere on the bone) in which the Shang kings were most likely to divine about military campaigns, about mobilizing their labor gangs, about offering harvest prayers, or about suffering from nightmares. Or one could use Moshi to place individual divinations about ancestral cult in the context of other cult divinations on the same bone, and so on. Certain cautions, introduced below, must be borne in mind, but Moshi is a straightforward reference set that all scholars working with oracle-bone inscriptions will wish to consult.

The second work under review, Yinxu jiagu keci leizuan (A Concordance of Graphs Inscribed on the Oracle Bones from Yinxu) - hereafter abbreviated as Leizuan, when referring to the book, and as Y (for Yao Xiaosui, its chief editor) when used in a citation(5) - is a more ambitious and potentially more useful work. Its three volumes provide, with certain exceptions,(6) a concordance to the words and phrases found in the corpus of inscriptions transcribed in Moshi. The difference between the two reference works is that Moshi presents the inscriptions in the sequence in which they were published, so that it is essentially a bone-by-bone transcription of the contents of the five major collections. Leizuan, by contrast, arranges the inscriptions from these collections - once again transcribed in both jiagu and kaishu forms, with two registers to a page - under word-headings for all of the words (ideally) that compose each inscription. In theory, therefore, HJ 272f, [Chinese Text Omitted], the first inscription quoted above, would be transcribed five times under the heading of each of its five graphs. In fact, common Shang graphs - prepositions and negatives like [Chinese Text Omitted], [Chinese Text Omitted], [Chinese Text Omitted], [Chinese Text Omitted], and [Chinese Text Omitted] as well as common words like [Chinese Text Omitted] and [Chinese Text Omitted], and the gan [Chinese Text Omitted]

"stems" (in this case, the stem xin [Chinese Text Omitted]) and zhi [Chinese Text Omitted] "branches" - are not given separate entries ("Fanli" [Chinese Text Omitted], p. 16), so that this inscription is copied only twice, under the entries for [Chinese Text Omitted] (Y 1406.1) and [Chinese Text Omitted] (Y143.1). (The editors also inform us with a "1" at the foot of the kaishu transcription that this is, in their opinion, an inscription that can be dated to oracle-bone Period I.(7))

The omission of such common words in a concordance is perfectly acceptable, particularly for the purposes of historians.(8) What is unacceptable in Leizuan, however, is the incomplete concordancing of important words. The worth of any concordance is seriously impaired if its entries for key terms are frequently and capriciously incomplete. I regret to say that Leizuan is seriously compromised in this regard. That key words are not concordanced consistently means that the entries provide no basis for drawing reliable conclusions about the scope and frequency of particular oracle-bone terms. The result is that Leizuan is valuable for adding - although how completely one cannot, without much additional research and documentation (see note 3), be sure - the new oracle-bone collections listed above to Shima's concordance (cited in note 3), but its deficient and quirky handling of the traditional corpus of inscriptions assembled in Heji sharply reduces its value. Nobody should throw away their copy of Shima; indeed, I have frequently found it necessary to turn back to Shima in an attempt to make up for the lacunae in Leizuan.

I provide some examples. (In what follows, unless there is reason not to do so, I employ the kaishu transcriptions provided by the editors.)

At least three inscriptions (HJ 10976f, 14136, 14140f) containing the phrase [Chinese Text Omitted], "Di orders much rain," which are recorded at Y418.2-19.1 under the heading [Chinese Text Omitted], are not recorded, although they should have been, under the [Chinese Text Omitted] heading at Y450.2. Any scholar relying upon Y450.2 to determine the degree to which Di, the High God, was involved in divinations about "much rain," accordingly, will be misled.

The drawing, Xucun 2.187 (S513.2),(9) [Chinese Text Omitted] ". . . the one who goes to (?)(10) Shang Jia is Father Yi. . ." is not transcribed under the [Chinese Text Omitted] heading at Y1361.1-71.2; nor is it under the [Chinese Text Omitted] heading at Y320.1-21.1 or the [Chinese Text Omitted] entry at Y1450.2-55.2.(11) The inscription has, in fact, been published as HJ 39549, a drawing, but I discovered that only by accident. Leizuan was of no assistance because its editors appear to have omitted HJ 39549 entirely.

HJ 41339, a drawing of a period III inscription (originally published as Ninghu 3.228 [S456.2]), divined by He [Chinese Text Omitted], about [Chinese Text Omitted], "the king will not encounter rain," is reproduced, in sequence, in Moshi. It is not, however, recorded under the appropriate [Chinese Text Omitted] entry at Y1201.1, where the last Heji entry is 38187. I find no reference to the inscription elsewhere in Leizuan. Indeed, it is not even listed under the diviner He's inscriptions at Y1499, where the last Heji entry is 31916. One has the sense, in fact, that, in compiling the Leizuan entries, the editors tend to have omitted drawings (which run from HJ 39472 to 41956) rather than rubbings.

HJ 22646 (Y186.1), which contains the phrase, [Chinese Text Omitted], "From Shang Jia down to the many descendants," is not listed under the [Chinese Text Omitted] heading at Y1361.2-62.1; indeed, none of many similar inscriptions containing the phrase (or a minor variant) just quoted, such as HJ 22648, 22650, and 22652 (all found at Y186.1 under the [Chinese Text Omitted] heading), are recorded under [Chinese Text Omitted]. Conversely, HJ 22625 and 22647, which do contain the phrase [Chinese Text Omitted] (or a minor variant), and which are listed at Y1361.2, are not listed under the [Chinese Text Omitted] heading at Y185.2-86.1. It is almost as if the Leizuan editors had cavalierly decided to put half the inscriptions containing [Chinese Text Omitted] under one heading, half under the other, so that no comprehensive concordance entry is provided at either place. Nor is the division between the two headings absolute: HJ 22622 appears under both headings!

Under the heading of [Chinese Text Omitted], "the settlements report (?)," at Y119.2, the editors provide only two charges, both on HJ 2895. Yet at Y119.2, in the same register, but under the miscellaneous (qita [Chinese Text Omitted]) heading, they record a third, HJ 4467, [Chinese Text Omitted]: "Divined: 'The settlements will come to report.'" The editors, in short, are literal minded about the entries they include under particular headings; "exploded" versions of a phrase are not included.

The front of the large plastron, HJ 376 (Moshi 13.2), bears the inscription [Chinese Text Omitted], which one might translate, "Divined: 'Call out to raise the men whom Xiang sent in.'" This inscription is not recorded under the...