Book announced at https://linguistlist.org/issues/34.3004

AUTHOR: Yuzhi Shi
TITLE: The Evolution of Chinese Grammar
PUBLISHER: Cambridge University Press
YEAR: 2023

REVIEWER: Geoffrey Sampson

SUMMARY

Chinese has the longest well-documented history (over three millennia) of any of the world’s languages except possibly Greek. (Chinese is also the language with by far the most speakers at the present day.) Yet, to my knowledge, the book under review is the first serious attempt at a history of its grammar in a Western language. This is an important book.
It is a translation of a Chinese-language book published in 2016, which won the Publishers Association of China’s award as the outstanding book of its year. The author’s name is given on the title page in the Western surname-last sequence, Yuzhi Shi; in Chinese his name is 石毓智. (Shi is the second most-cited living Chinese linguist, behind only William Wang 王士元 – who has had longer to accumulate citations.) The Chinese original has been reworked to make things as easy as possible for Western readers. For instance, Shi dates the various documents from which he draws examples in terms of BC and AD years, rounded to the nearest fifty, rather than by dynasty as would be normal in Chinese writing.

The lack of any comparable earlier publication in a European language may be explained by the fact that Chinese, having always typologically been close to the “isolating” extreme, is commonly perceived as not having any grammar. Edwin Pulleyblank (1995: xiii) described “a widespread belief that Chinese, especially the classical language, had no grammar and that the only way to learn it was by a kind of osmosis”. Westerners were of course aware that the modern vernacular is a very different language from the Classical Chinese which remained the standard written language of China until a hundred years ago (though it had long ceased to be a spoken language), but even when these are compared the grammatical differences, while large, tend to be overshadowed by the huge and salient differences in vocabulary. Western scholars from von der Gabelentz (1881) onwards had described the grammar of Classical Chinese, but usually as a fairly static written system. They had not done very much to expound the historical development of the classical language while it remained a spoken language, or to investigate how the grammar of modern Mandarin evolved out of Classical Chinese.

Yet, Shi argues, the history of Chinese grammar is a matter of great interest not only for specialists in the language but for general linguistics, because grammatical developments in a language can take long periods to work themselves out, and hence can only be adequately studied in languages that have long recorded histories. In his introductory chapter Shi gives the example of the development of reduplication as a Chinese grammatical mechanism, which already commonly occurred with adjectives and adverbs “in texts composed around the eleventh century BC, but the reduplicated forms of nouns first appeared in texts around the third century AD, those of classifiers emerged around the eighth century AD, and not until the fourteenth century AD did verb reduplication start to occur.” It is a Leitmotiv of Shi’s book that “The reason why a new form can emerge, gradually mature, and eventually replace an old one is usually that it is favored by the overall properties of the grammar at that stage” (p. 23), and “Any linguistic development is typically a chain of changes. An earlier change enables a later change to take place, and the earlier change is made possible by yet another preceding change” (p. 124). As his book proceeds Shi discusses a number of grammatical developments which each evolved over many centuries or a millennium-plus.

For this reason, the book is organized in an unexpected way. Rather than successive chapters examining successive periods of Chinese, each chapter examines a different area of grammar and its fate over the entire period. For instance Shi’s Chapter 2 is about how, beginning about the first century BC, equative expressions, equivalent to English “Charles is the King”, evolved from the earlier form ‘C K yě’ (yě written 也), to ‘C shì K’, with shì 是 (which previously meant “this”) as a copular verb. One recent publication (Xiangdong Shi 2015) claimed that the new use of shì was brought about through the influence of literally-translated Sanskrit texts when Buddhism came to China. To me it seemed implausible that religious writings might cause such a radical change in the language of a largely-illiterate population (a point Shi makes himself, p. 207, in connexion with a proposal that another grammatical development had a Sanskrit origin). Yuzhi Shi does not mention Xiangdong Shi’s theory, but his discussion suggests that it is unlikely to be right. The first century BC looks rather early for copular shì to be linked to Buddhism, but in any case Shi finds that “the proper context emerged for grammaticalizing the demonstrative shì into a copula” as early as the fifth century BC, long before Buddhism came to China.
Then, once shì became established as a copula, and in consequence of that, Shi sees the word as having gone on to develop a further role as a “focus marker”. In the modern Chinese version of a sentence like “A at time T Verbed B”, by prefixing shì to any of the constituents (without altering their order) one gets equivalents of “It was A who Verbed B at T”, “It was at T that A Verbed B”, or (with an extra complication) “It was B whom A Verbed at T”. And Shi also argues that copular shì was one factor triggering the change by which wh– question words ceased to shift to initial position, as they do in English and did in Old Chinese: modern Chinese says “A Verbed who?” rather than “Who did A Verb?” Fronting and adding shì were alternative ways of marking a wh– word as focus of a question. (But I shall return to this in the Evaluation section.)
Later chapters explore some twenty further areas of grammar, drawing on abundant examples from a wide range of sources. Until the twentieth century, obviously all Shi’s evidence on grammatical evolution has to come from written sources, but wherever possible he uses genres such as novels or drama where the prose is likely to be closer than other writing to the spoken vernacular of its period. (All Shi’s evidence is taken from literary writing; he does not use the “oracle-bone” inscriptions which are the earliest extant examples of written Chinese, so his data begin from about 1000 BC.)

Not all areas of grammar explored by Shi are obvious ones. Chapter 7 is about information structure, and according to Shi, even within China few historical linguists have examined developments there; yet, he says, “there were fundamental changes in two types of information structure: one is at the clause level, involving the arrangement of new and given information, and the other is at the predicate level, involving the arrangement of resultative and non-resultative (accompanying) information.”

Shi is alive to relevant findings of general linguistic theory. For instance, he often discusses how Chinese grammar at different periods relates to one or another of Joseph Greenberg’s universals of word-order. Shi’s chapter 21 begins by identifying the most exception-free of the 45 universals listed in Greenberg (1966) as: “if a language places object after verb, it will place relative clauses after their noun head”. Shi describes Old Chinese as having two relative clause strategies, the more general of which (p. 525) used the particle zhě 者 as relativizer and placed the relative clause after its head; an example from the Shi Ji (“Historical Records”, about 100 BC) is glossed “guest recent have from Shandong come zhě call Cai Ze”, meaning “the guest who recently came from Shandong is called Cai Ze”. Old Chinese was an SVO language, hence this is as it should be according to Greenberg. But “Roughly after the sixth century AD … the Chinese language underwent a typological change from ‘N REL’ to ‘REL N’” (p. 529), a change which Shi relates to other grammatical developments at the period. In modern Chinese the only way to express the relative clause would place it before “guest”, linked by the modern relativizer de̊ 的, along the lines “recently from Shandong come de̊ (that) guest”. Thus modern Chinese is an exception to the Greenbergian universal: Shi quotes Dryer (1992) as finding that it is the sole exception in a sample of 645 languages. Shi mentions an idea put forward by some linguists (e.g. Li and Thompson 1974) that Mandarin is currently, via its “bǎ Object Verb” construction, in the process of shifting from SVO to SOV order, which would remove the inconsistency; but he gives that short shrift. Shi sees the inconsistency as real, and goes on to discuss possible theoretical reasons why such an inconsistency might arise in a language.
(Fascinatingly, where Chinese is in contact with Altaic SOV languages, Shi finds Mandarin dialects with true SOV grammar, e.g. – Qinghai dialect, p. 241 – Wǒ kāishuǐ hā hē-le, ‘I boiled-water ACC drink-PERF’, i.e. “I have drunk the boiled water”, where hā is a suffixed object marker, utterly alien to Standard Chinese.)
Within the space constraints of a Linguist List review it is really impossible to do more than hint at the comprehensiveness and depth of Yuzhi Shi’s analysis of the evolution of Chinese grammar.

EVALUATION

From what I have already written it will be clear that my overall evaluation of Shi’s book is very positive. I am sure it is destined to be a classic of Chinese historical linguistics.
Inevitably, though, not all the news is good.
To get two relatively trivial points out of the way first: the book could have been better proof-read, and the index made more complete. I found quite a few example sentences where a word in the Chinese was missing from the line of English glosses, or vice versa, and many linguists cited in the book do not appear in the index.
More important: sometimes I feel unsure of what Shi is saying because of vagueness in his grammatical terms. He makes heavy use of the related concepts “subject”, “focus”, and “topic (as opposed to comment)”. In European languages the concept “subject” is rather clear: the subject of a finite clause is the NP with which the verb agrees. But, in an isolating language like Chinese, “agreement” is meaningless. Subjects also tend to be agents of verbs of action, and to be the first NP of their clause, or at least to appear immediately before the verb when the clause has a separate topic, as in English “Raspberries I like” – structures like that are common in Chinese, which is said to be a “topic-prominent language”. But these tendencies are not enough to identify subjects clearly in many instances. On p. 143 Shi tries to give tighter definitions of topic, subject, and focus, but these still leave me wondering how he is applying the concepts in particular cases.
Thus, p. 162 discusses an example from the opening columns of the novel Ru Lin Wai Shi (“Unofficial History of the Scholars”, AD 1750). In English the passage runs (my translation) “The man’s name was Wang Mian, and he lived in a country village in Zhuji County; when he was seven years old his father died …”. The Chinese wording corresponding to what follows the semicolon can be glossed “seven year-of-age time die PERFECTIVE father”; and Shi says of it “the subject is construed as a location, indicating that something has happened to someone.” I think Shi means that the understood subject is Wang Mian, and the reference to “location” is intended to suggest that Wang Mian “had his father die on him”, as it might be put figuratively in English. But why postulate an “understood” subject in the “die” clause? One might call “father” the subject, as it is in the English translation, though in the Chinese this word neither precedes the verb nor is an agent (since “die” is not a verb of action). Or one might say that the clause has no subject – subjectless sentences are perfectly normal in Chinese, and subjects are not always supplied by context. Even if Wang Mian were an understood subject here, it would seem more natural to take it as understood within the subordinate clause “(Wang Mian) seven year-of-age time” than as a constituent of the main clause. Without tighter definitions of the terminology it is sometimes hard to know what structures would count as examples of Shi’s claims about grammar (or what structures would disconfirm them).
It is very good that Shi relates his analyses of Chinese grammar to general linguistic findings, such as Greenberg’s, about what does and what does not tend to happen in the languages of the world; and Shi evidently keeps very up-to-date with Western linguistics, for instance he makes valuable use of Bybee and Beckner (2015), published the year before the Chinese original of Shi’s book. But sometimes Shi pays more respect to theorizing than it is really worth. For instance, in connexion with the Chinese change from “wh– fronting” to “wh– in situ” grammar, already mentioned, Shi has a long discussion on pp. 29–30 which says among other things:
“Many researchers believe that overt wh– movement relates to a Q-morpheme in C^0, a notion that was first posited by Katz and Postal (1964) … Chomsky (1995) considered the Q-feature to be a matter of degree, and only when it is strong enough can it trigger overt movement. This speculation was supported by scarcely any empirical evidence and hence was soon abandoned. … Aoun and Li (1993) argued that the relation between wh– words and a Q-operator is a binder–bindee relation. Thus the wh– words in Chinese cannot express a question by themselves but can do so only by means of the question particle ‘ne’. However, this observation is entirely incorrect for the following two reasons …”
Is it really worthwhile to grind through all this material that perhaps seemed important once (to some linguists, anyway) but surely no longer? Shi’s book is one written for the ages. It would have been better not to clutter it up with ephemeral theorizing.
A particular problem in some parts of Shi’s book stems from the fact that, while he knows an enormous amount about the history of Chinese grammar, he seems to have a blind spot when it comes to Chinese phonology and its history.
In modern Chinese, all syllables have simple structures (no consonant clusters) and represent single morphemes. But several linguists have reconstructed Old Chinese as containing more complex syllables, with some syllables including derivational affixes. Shi comments (p. 118) “all of these devices have left no traces in any of the numerous dialects of Contemporary Chinese; hence it is difficult to find empirical evidence for these suggestions.” This is a textbook example of what Edwin Pulleyblank (1973: 112n) described as “a fairly widespread attitude that cannot conceive of Chinese as ever having had features that are not found in modern dialects”, adding “if one takes seriously the probability of a genetic relationship between Chinese and Tibetan, it is obvious that Chinese must once have had such features.” The scholars who have reconstructed affixes in Old Chinese may be mistaken, but one cannot say that they had no empirical evidence, as if the affixes were just fantasies – though some of the evidence comes from other languages.
Shi notes, p. 52, that all the wh­– words of Classical Chinese other than shéi or shuí 誰 “who?” were replaced by new vocabulary in the early centuries of the Christian Era (p. 157 puts the development later, in the sixth to tenth centuries – this inconsistency is not resolved); he claims that the replacement was “closely related to the emergence of the focus marker shì”. According to Shi, many of the new wh– words are phonetic reductions of phrases beginning with shì, and this applies even to shéi “who?”, where the derivation from a two-word phrase is concealed by the accident that both words of the unreduced phrase began with the same /ʂ/ phoneme. In particular, Shi believes that shémme “what?” is a reduction of shì hé 是何 “FOCUS what?” Shi quotes as his authority a book by the distinguished linguist Lü Shuxiang.
With due respect to Lü, there are two problems here. The obvious problem is phonetic: shì hé contains no /m/-like sound, and never did (according to Axel Schuessler’s (2007) reconstruction of Chinese pronunciation at the period, the words were then pronounced /dźe gɑi/, ignoring tone – I quote Schuessler as a relatively uncontroversial version of Chinese historical phonology, but so far as I know nothing here or below hangs on choice among alternative reconstructions). So how could the phrase turn into shémme? And this etymology contradicts a well-established alternative. According to Zhang Huiying (1982), shémme is a reduction of the phrase shí wù 十物 “ten objects”, which would explain the /m/: Schuessler’s version of shí wù at the same period is /dźip mut/. From “ten objects” to “what?” is quite a semantic journey, but Zhang evidently makes a case for this etymology which is sufficiently convincing for e.g. Jerry Norman (1988: 119–20) and Schuessler (2007: 458) to see the issue as settled without need for further debate. (Unfortunately, I have not obtained access to Zhang 1982 or Lü 1985.) Neither point, either the lack of /m/ in shì hé or the “ten objects” etymology, is mentioned by Shi.
Shi tries to bolster his argument that most modern wh– words derive from shì phrases by saying that Mandarin 怎麼 “how?”, standardly romanized zěnme, is pronounced [ʐǝnmǝ], where [ʐ] differs only in voicing from the [ʂ] of shémme. This looks odd, because ‘z’ in pinyin romanization normally represents the affricate /ts/. And over the page, Shi’s Table 3.1 directly contradicts what he has just said, by showing the pronunciation of the modern word for “how?” as [tsǝnmǝ]. There is no hint of a resolution to this contradiction. I cannot accept Shi’s idea that modern wh– words incorporate a focus marker.
Passages like these create skepticism about other passages where Shi’s assertions are not contradictory, but which leave me wondering why I ought to believe Shi’s account of a phenomenon in preference to a more obvious and at least equally plausible account. Immediately before the passage just discussed, on p. 51, Shi has mentioned the very common Mandarin form de̊ 的, which we met earlier as a relativizer and which fulfills a wide range of grammatical functions including genitive marker. The superscript circle means that de̊ is toneless, and the vowel is a shwa rather than the vowel which pinyin ‘e’ usually represents, hence the form must be a phonetic reduction of something. As a grammatical element, 的 does not occur in Classical Chinese; many of its functions are fulfilled there by a word zhī 之, which also had the lexical meaning “go”, and which does not occur in modern Mandarin. The Old Chinese pronunciation of 之 as reconstructed by Schuessler was /tǝ/, very close to the Mandarin pronunciation of 的 (‘d’ in pinyin stands for the same sound as /t/ in Schuessler’s transcription). Old Chinese /tǝ/ became zhī through a sound-law which turned initial stops into fricatives or affricates in about half of all cases – what phonetic feature triggered application of this rule is heavily debated.[1] It is easy to suppose that, when 之 was used as an unstressed grammatical element, phonetic reduction led the form to lose whatever feature this was, so the rule did not apply and de̊ now appeared to be a separate word from zhī “go”, which was given a written form by borrowing the existing graph 的 for a similar-sounding word, dì “bullseye”.
But Shi states that de̊ arose when, about the eighth century AD, “the demonstrative dǐ grammaticalized into … the phonologically reduced form” de̊ – that is, Shi believes that modern de̊ 的 is unrelated to Classical 之. Present-day Mandarin has no “demonstrative dǐ”, but on p. 444ff. this word is described as having existed a thousand years ago as itself a grammaticalization of the word 底 “bottom”. Well, perhaps; but Shi gives me no confidence that this etymology should be preferred to the derivation from 之. Rather than postulating “grammaticalization” it seems simpler to suppose that the graphs for “bullseye” and “bottom”, words etymologically unrelated to de̊ though sounding similar, were borrowed as alternative writings for a grammatical word that appeared to be new and hence to have no graph of its own. Shi mentions others who have derived de̊ from zhī, p. 492, but his dismissal of them seems perfunctory. I feel convinced that his real reason is lack of interest in phonology.

Luckily, phonology is not relevant to too many of the issues that arise in studying the history of Chinese grammar. Shi has done an outstanding job of identifying these issues, and finding examples in texts from different periods to illustrate his account of them. If sometimes his conclusions seem disputable, that is the way of scholarship. Surely, others who pursue this area of enquiry will be treating Shi as a starting point, whether they agree with what he says or not, for very many years to come.
[1] See <en.wikipedia.org/wiki/Old_Chinese_phonology#Type_A_and_B_syllables>, accessed 29 Jan 2024.

REFERENCES

Bybee, J. and C. Beckner, 2015. “Usage-based theory”. In B. Heine and H. Narrog, eds, The Oxford Handbook of Linguistic Analysis. Oxford University Press.

Dryer, M.S., 1992. “The Greenbergian word order correlations”. Language 68.81–138.

von der Gabelentz, G., 1881. Chinesische Grammatik, mit Ausschluss des niederen Stiles und der heutigen Umgangssprache. T.O. Weigel (Leipzig).

Greenberg, J.H., 1966. “Some universals of grammar with particular reference to the order of meaningful elements”. In Greenberg, ed., Universals of Language, MIT Press.

Li, C.N. and S.A. Thompson, 1974. “Historical change of word order: a case study in Chinese and its implications”. In J.M. Anderson and C. Jones, eds, Historical Linguistics. North-Holland (Amsterdam).

Lü Shuxiang, 1985. Jindai Hanyu Zhidaici [Pronominal Words in Modern Chinese]. Xuelin Press (Shanghai).

Norman, J., 1988. Chinese. Cambridge University Press.

Pulleyblank, E.G., 1973. “Some new hypotheses concerning word families in Chinese”. Journal of Chinese Linguistics 1.111–25.

Pulleyblank, E.G., 1995. Outline of Classical Chinese Grammar. University of British Columbia Press (Vancouver).

Schuessler, A., 2007. ABC Etymological Dictionary of Old Chinese. University of Hawai’i Press (Honolulu).

Shi Xiangdong, 2015. “The influence of Buddhist Sanskrit on Chinese”. In W.S.-Y. Wang and Chaofen Sun, eds, The Oxford Handbook of Chinese Linguistics, Oxford University Press.

Zhang Huiying, 1982. “Shi ‘shenme’”. Zhongguo Yuwen 4.302–5.

ABOUT THE REVIEWER

Geoffrey Sampson graduated in Oriental Studies at Cambridge University in 1965, and studied Linguistics and Computer Science as a graduate student at Yale University before teaching at the universities of Oxford, LSE, Lancaster, Leeds, and Sussex. After retiring from his Computing chair at Sussex he spent several years as a research fellow in Linguistics at the University of South Africa. Sampson has published in most areas of Linguistics and on a number of other subjects. His recent books include "The Linguistics Delusion" (2017), "Voices from Early China" (2020), and "God Proofs" (2022).