The Origin and Early Development of the Chinese Writing System.

The two essential stages marking the "linguistic history" of the Chinese writing system govern the two parts of this book: "The Shang Formation" in which the author explains the development of the system, and "The Ch'in-Han Reformation," in which he deals with the normalization of writing or what he calls the "reaffirmation of that original formation."

One of the first intentions of this book is to correct wrong ideas about the Chinese script. Even though Peter du Ponceau as early as 1838 had presented the true nature of the Chinese writing system (A Dissertation on the Nature and Character of the Chinese System of Writing [Philadelphia: American Philosophical Society]), it is regrettable that his voice was not widely heard and that the myth of the "ideographic" Chinese script still persists. As Boltz shows in the first part of his book (pp. 31-126), the Chinese writing system followed the general path and basic principles that governed other archaic writing systems - mainly those of Mesopotamia and Egypt in the ancient world (with occasional reference even to the Mayan system of writing in the New World). The three systems indeed have pictographic origins,(1) but this does not exclude signs with more "abstract" origins. They commonly went through a stage called "graphic multivalence," where graphs could represent homophones (rebus writing or paronomastic writing) as well as homeosemes (parasemantic use of a graph). The use of determinatives - semantic or phonetic - was then provoked by the emergence of semantic ambiguity caused by the multivalent use of graphs.

According to Boltz (p. 17), writing is "the graphic representation of speech," and a writing system is defined as "any graphic means for the systematic representation of speech." Later (p. 19) he says that "the essential and indispensable feature that must be present for a graph or a system of graphs to qualify as writing is phonetic representation." Thus, according to this definition, all graphs that are not associated with a pronunciation are excluded from "writing." It has to be said that in Chinese the term wen-tzu means "graph" as well as "writing system," without distinction. Although in the context of a general theory of writing it is necessary to distinguish these meanings clearly, and Boltz is right to do so, for a better understanding of the development of writing it would probably be more helpful to take an evolutionist (or historical) point of view. This would include, in a broad definition of writing, the signs that are given in Boltz's representation as [-P(honetic), +S(emantic)], that is, which show an association with meaning but not with pronunciation - provided of course that they belong to a codified system, which is not the case with the Chinese neolithic signs. Indeed, if one takes into account the other archaic systems of writing, one can see that things are not as clear-cut as Boltz describes them. As he puts it, with excellent reasoning (esp. pp. 38, 44, 46), it is clear that the signs appearing independently on pottery in the Chinese neolithic era are neither proto-writing nor the ancestors of the Shang script. But in Mesopotamia, as early as 3000 or 3200 B.C., a total of fewer than one thousand signs can be found on clay tablets. This supposes a systematic choice of the signs, whose shapes were already standardized. These signs are organized to record information, which is not the case with the neolithic signs of China. But because, in this first stage, the script can only, for the most part, recall known information, these tablets do not record all linguistic features and must essentially be already known if they are to be understood. Nevertheless, these signs are bound to the world-view as well as the language of their creators.(2) They are symbols representing images or symbols associated with different possible pronunciations: in Boltz's designation, these signs should be [-P, +S]. But, in order to understand the messages in a precise context, the "reader" is obliged to choose a single pronunciation of the signs. Which, in turn, is equivalent to saying that even though their pronunciation may be variable, pronunciation is nonetheless present though not yet codified. (Assyriologists really are right to call this "writing.")

Indeed, in its fully realized phase, writing is bound to phoneticism and can truly represent speech. But as we can see in the earliest documents from Mesopotamia, representing speech is not the motivation of writing at its beginning. It is a possibility that rebus writing was only realized through the exploitation of homophony, one or two centuries after writing came into use. The first known example of homophony in Mesopotamia goes back to the times of Uruk III, around 3000 B.C.(3) And it is only many centuries later that writing became a fully graphic representation of speech.

An original point of this book is that Boltz, following Boodberg,(4) insists on the polyphonic aspect of graphs as early as the "graphic multivalent" stage. Polyphony played an important role in the development of Mesopotamian and Egyptian writing, but its importance has been largely ignored in Chinese. Since signs expressing different meanings through semantic associations were known to exist, one may ask why polyphony has not been recognized? Yet, it is far from being proved that polyphony was as well developed as Boltz assumes (see the doubts below about his Set 6). Boltz distinguishes two types of hsieh-sheng ("shared phonetic") series: one is called "transparent" hsieh-sheng series and the other is called "elusive" hsieh-sheng series (p. 106). Necessary conditions for their existence are established (p. 90), with some restrictions (p. 104) for the "elusive" series. William Baxter, in his Handbook of Old Chinese Phonology (Berlin: Mouton de Gruyter, 1992), has shown that some of the reconstructed rhymes under a group of traditional rhymes may not have rhymed together. This is the case, for example, with the yuan rhyme group, for which Baxter reconstructs three different rhymes in Old Chinese: *-en, *-an, *-on. Boltz suggests the existence of an elusive hsieh-sheng series, based on the graph 'eye', read kians or mjians, for all the characters in his Set 6 (p. 105), belonging to the yuan rhyme group. In Baxter's reconstructions of Old Chinese, the pronunciation of these words should be:

jian < kenH < *kens mian < mjienH < *miens huan < hwaenH < *(g)wr(e,a)ns xian < xjonH < *xjons man < mjwonH < *mjwans

Accordingly, the first two pronunciations could rhyme together, but they could not do so with the last two.(5) This shows that Boltz's hypothesis of an elusive hsieh-sheng series based on a "second" reading of "eye" (as proposed by Boodberg in 1937) lacks evidence. Such elusive hsieh-sheng series based on a "second" reading of a graph are difficult to find, because they depend on phonological reconstructions of Old Chinese, about which specialists are not yet in agreement.

In the second part of the book, devoted to the normalization of writing at the time of the Ch'in-Han dynasties, Boltz recalls the legends of Ts'ang Chieh and Fu Hsi concerning the beginning of Chinese writing; the distinction between wen and tzu which, according to him, does not precede the Ch'in-Han period; and also the liu shu or "six [forms of] script."

The weakness of this book is the treatment given to "compound characters that do not have a phonophoric element within their graphic structure," that is, the so-called hui-i traditionally referred to as "ideograms." Boltz rejects the use of the latter term, because it has given rise to many misunderstandings about the nature of the Chinese script. But he is too categorical when he entirely denies the existence of such compounds (pp. 72, 149). One of the two examples of hui-i given by Hsu Shen in his Shuo-wen chieh-tzu of the first century A.D. is problematic, and Boltz is right when he says that hsin (PY: xin) has a phonophoric element in its graphic structure. But, according to Baxter's reconstructions as well as to Qui Xigui (cited at p. 149, n. 20), it is not the graphic element yen (PY: yan) but jen (PY: ren) that is phonophoric:

xin < sinH < *snjins, 'trust' ren < nyin < *njin 'man' yan < ngjon < *ngjan 'speech'

Most certainly, at the time of Hsu Shen the kinship between the pronunciation of xin and ren was no longer visible; the prefix s- gave rise to a voiceless fricative in the case of xin, while the initial remained nasal in the case of ren. This is probably the reason why Hsu Shen considered xin to be a true hui-i, that is, a graph based on semantic construction, rather than a hsing-sheng.

On p. 148, Boltz raises three doubts about the existence of hui-i characters, but these can be easily refuted. First, if Boltz did not incorrectly translate the i of hui-i as "meaning" but rendered it properly as "intention, idea," he would not be able to say that the hsiang-i of Pan Ku contradicts the hui-i of Hsu Shen, because "to represent an idea" (hsiang-i) or "to combine ideas" (hui-i) is exactly what these compound characters without phonophoric element are meant to do. Second, we may wonder why what Boltz calls "zodiograms" with no phonophoric element would be a more "viable process for generating characters" than the "combination of two or more graphic elements to make a new character based on the meaning of the parts." Because, third, unlike what Boltz assumes, examples of such graphs do exist in Mesopotamia as well as in Egypt, and played an important role in the development of writing.(6) Examples in Mesopotamian writing (Labat, Manuel d'epigraphie akkadienne [Paris, 1976], nos. 36, 558, 214) are:

'eat' ku: 'mouth' ka + 'bread' ninda 'female slave' geme: 'women' mi + 'mountain' kur 'bier' kash: 'vase' dug + 'grain' she

Examples from Egyptian (Gardiner, Egyptian Grammar [London, 1964], 451:9, 472:41, 485:4) are:

'cry': 'eye' + 'tears' 'throw': 'stick' + 'pintail duck' 'rain': 'sky' + 'rain'

After Boltz's demonstration of the basic principles governing the development of the three archaic writing systems, it would be difficult to explain why the Chinese script, with substantial pictographic origins, would be the only one to avoid such a fundamental stage as this. Moreover, it is noteworthy that Boltz says nothing about the second example of hui-i given by Hsu Shen, i.e., wu 'warrior'. According to Baxter, the two elements of which the graph is composed (ge < kwa < *k[a,o]j 'halberd' and zhi < tsyi < *tjt? 'foot') show no phonophoric role in this graph: wu < mjuX < *Np(r)ja? (? = glottal stop). Consequently, Hsu Shen was right to give the example of wu in his definition of hui-i. Another example of a graph for which the pronunciation cannot be deduced from any of its graphic elements is chu (PY: zhu) < drjuwk < *rjiwk 'follow' with the two elements zhi < tsyi < *tjt? 'foot' and shi < syjeX < *hlje? 'pig'.(7) One might suppose, as Boltz does, that another pronunciation of 'pig': chu (which is that of another graph for 'pig') can be related to the graph 'follow'. But in the Shang inscriptions this graph 'follow' can also be written with the element for lu 'deer' as well as those of other animals instead of 'pig'. So this does not allow us to say that the graphic element for 'pig'(8) was phonophoric in.

In short, I see no good reason to reject hui-i or "compound characters that do not have a phonophoric element within their graphic structure." If one does so, one falls into the opposite trap from that of the partisan of pure ideography: phoneticism at any price. The Chinese graphic system is neither one nor the other, it is a subtle and nuanced combination of the two.

The book concludes with a very important chapter, "The Impact of the Chinese World-View" (pp. 156-77), in which Boltz presents a thesis about the potentiality of the development of an asemantic writing system in China, a development that was prevented by a world-view that definitively fixed Chinese script in its logographic nature by "reinforcing the bonds of the graphic-phonetic-semantic trinity" (p. 176). The author here presents evidence from four Ma-wang-tui manuscripts found in 1973 near Changsha, two of which were probably written prior to 187 B.C. and the other two before 195 B.C. These texts, written with a large number of variant characters, described by Boltz (pp. 158-63), are interesting because they apparently do not conform to the reformation imposed by Li Ssu in 213 B.C. They consequently give a good idea of the script before the time of standardization. What is noteworthy in the list of variant characters from these manuscripts (pp. 164-66) is the existence of unknown graphs structured like hsieh-sheng, that is, combining a phonetic with a semantic element. This shows that there was not only a tendency toward simplification, which according to Boltz could have given rise to an asemantic writing system, but also always a need for disambiguation.

Finally, I would criticize the author's use of the term "classifier" (pp. 163-70). This term, which as Boltz properly says (pp. 67-68) only refers to the classification of graphs,(9) is not to be used in the context of their graphic structure. The example of le and li (p. 165), under "Subtype (c): variation between presence and absence of classifier," shows that the term is incorrect, because the real classifier of these two graphs is li.

Typographical errors are few. I have found only four: on p. 18, n. 3, the reference to "note 4 of the Introduction" should read "note 5"; on p. 40, line 17, for "facililtant" read "facilitant"; on p. 149, n. 20, for "Qui" read "Qiu"; on p. 155, line 23, for "Shou wen" read "Shuo wen." The author has noted a dropped line of text on p. 48; the sentence beginning on line 16 should read as follows: "In the terms we introduced above, these graphs are unquestionably [+S], that is, they have meaning, but we cannot say with certainty that they are [+P], that is, characterized by a conventionally associated pronunciation."

I highly recommend this very lucid and richly documented book, which also includes a useful glossary of technical terms (pp. 179-83), in addition to a bibliography, index of Chinese characters, and general index. The author has set forth an excellent and precise notion of the Chinese writing system in a manner that will stimulate profound reflection about the very idea of writing.

FRANCOISE BOTTERO ECOLE DES HAUTES ETUDES EN SCIENCES SOCIALES, PARIS

1 Here, following Peter A. Boodberg, "The Chinese Script: An Essay on Nomenclature (The First Hecaton)" Bulletin of the Institute of History and Philology (Academia Sinica) 29 (1957): 113-20, Boltz introduces a distinction between "pictographs" which represent a thing and "zodiographs" which represent a word.

2 The Mesopotamian sign for "female slave" (see below), composed of the elements 'women' and 'mountain,' is a good example of such a world-view.

3 See Jean Bottero, Mesopotamia: L'ecriture, la raison et les dieux (Paris: Gallimard, 1987), 103 et seq.

4 See Boodberg, "Some Proleptical Remarks on the Evolution of Archaic Chinese," Harvard Journal of Asiatic Studies 2 (1937): 329-72.

5 Note that the case of huan < hwaenH < *(g)wr(e,a)ns is uncertain, since it is not sure whether this was pronounced *(g)wrens or *(g)wrans.

6 See. J. Bottero, 99 et seq.

7 See Xu Zhongshu, Jiaguwen zidian (Chengdu: Sichuan chubanshe, 1990), 158.

8 One may suppose that 'foot' had another reading and was phonophoric in chu (PY: zhu), but that would need to be proven, and it would also need to be proven that 'foot' is phonophoric in wu 'warrior' as well as in most of the graphs in which it appears. That would attribute to 'foot' so many different readings that one might doubt the effectiveness of such a writing system, as against the simpler expedient of using the semantic value of graphic elements.

9 See Francoise Bottero, Semantisme et classification dans l'ecriture chinoise: Etude du systeme de classement des caracteres par cles du Shuowen jiezi au Kangxi zidian (Paris: Institut des hautes etudes chinoises, College de France, forthcoming).

Article Details

Printer friendly Cite/link Email Feedback
Author:	Bottero, Francoise
Publication:	The Journal of the American Oriental Society
Date:	Jul 1, 1996
Words:	2649
Previous Article:	Text and Context; Studies in the American New Testament: Papers Presented to the Conference on the Armenian New Testament, May 22-28, 1992.
Next Article:	Learning from Mount Hua: A Chinese Physician's Illustrated Travel Record and Painting Theory.
Topics:	Book reviews Books Book reviews