CC BY-NC-ND
18
ABC Etymological Dictionary of Old Chinese By Axel Schuessler
Honolulu: University of Hawai'i Press. 2007. xx + 668 pp.
George Starostin (Russian State University for the Humanities)
Appearing somewhat surreptitiously and certainly lacking the proper publicity, Axel Schuessler's Etymological Dictionary of Old Chinese is, in fact, arguably the biggest thing that has so far happened to Chinese historical linguistics studies in the 21st century. Finally, the scholarly world has come into possession of a worthy reference source on the history and origins of a significant portion of Old Chinese vocabulary — the first such source ever published. Regardless, then, of whether Prof. Schuessler's work is flawless or not, it constitutes a major landmark in the history of Sinology, and will serve as the benchmark by which many, if not most, further works dealing with the history of the Chinese language will have to be judged.
With an immense baggage of knowledge on both the internal developments and the external connections of Chinese accumulated over the last century by dozens of specialists, Chinese and Western alike, the appearance of such a dictionary was, of course, a historical inevitability. Today, the Sino-Tibetan affiliation of Chinese is no longer disputed as it once used to be, and, with the acknowledged status of Austronesian, Austro-Asiatic, Hmong-Mien (Miao-Yao), and Tai-Kadai language families as forming strong areal rather than genetic connections with Chinese, it has become generally clear how to treat these connections from a methodological point of view. Moreover, the fact that the idea of finally compiling and publishing a proper etymological dictionary of Old Chinese was carried out by none other than Prof. Schuessler is quite auspicious, since he is one of the biggest Western authorities on Chinese historical linguistics; his earlier "Dictionary of Early Zhou Chinese" is easily the finest resource in existence on this archaic and, even today, least understood and well-studied (but most important for various philological studies) period of the Chinese language, and, over the past several decades, he has made numerous valuable contributions to the study of Chinese historical phonology, grammar, and external linguistic relations.
Several important and useful features of the dictionary should be stressed at once. First, it is preceded by a detailed, 150-page long introduction that carefully explains the structure of the work, the methodological principles underlying it, and the general linguistic situation in and around Ancient China as assumed by the author for the times when Old Chinese was still a spoken language. On an equally significant note, we get a thorough description of the author's view on the phonological and morphological structure of Old Chinese, as well as the origins of this structure, reconstructed through comparison with other branches of the Sino-Tibetan family. This is particularly meaningful inasmuch as no general consensus has so far been reached by linguists on many of the raised issues, and it is very important for specialists and casual readers alike to learn more about the motivation behind many (if not most) of Prof. Schuessler's etymologies offered in the main body of the work.
Concerning this main body, primary emphasis is placed on making the work a word-based rather than character-based dictionary — a decision with which most Sinologists, no doubt, will heartily concur. Relying too heavily on the graphic form of the etyma would make the reader miss many obvious (and some not so obvious) connections between words that, already from the earliest times, were written with different Chinese characters, yet were nevertheless etymologically related — or, vice versa, assume false connections between words of entirely different etymological origin, whose only similarity was in that the same character used for one word would also be used for a different one due to phonetic similarity between the two.
The dictionary is, therefore, structured according to the phonetic principle, with all the words arranged in the order of their modern pinyin readings. A separate character index would have been nice, but is not obligatory provided one knows the primary reading of the character; characters denoting derived words whose etymologies are inseparable from etymologies of other ('primary')
Journal of Language Relationship • Вопросы языкового родства • 1 (2009) • pp. 155 - 162
words simply refer to these words. E. g., a search for the etymology of bao 'leopard' will lead the reader to the entry bao ^ bo which should be read: "bao 'leopard' is originally derived from bo 'variegated', see under bo". This is a fairly reasonable approach.
The overall structure of the entry, besides the expected reconstruction for Middle Chinese and Old Chinese (usually several distinct stages of the latter) and the etymological section, usually includes other valuable information, such as attested dialect forms and/or transcriptions of the etymon in question in such medieval sources as the 'Phags-pa transcribed Menggu Ziyun ]. The
latter information, as the author correctly remarks, may occasionally shed light on the word's later development. However, it is provided somewhat inconsistently.
Obviously, no matter how ample our knowledge on Chinese etymology may be at present, multiple Old Chinese stems still lack any kind of historical explanation. Such words are simply omitted from the dictionary altogether, so the reader should not expect to open it in the hopes of finding at least something about any character ever recorded in attested Old Chinese sources.
Aside from a somewhat clumsy system of abbreviations (a problem exacerbated by the fact that different types of abbreviations are decoded in fairly distant parts of the book, so that it takes a fair amount of time to properly understand the contents of each entry), the general layout of the dictionary is, therefore, almost impeccable. Much more complicated, but, of course, also much more important is the issue of properly evaluating its scientific content.
We should begin with the obligatory remark that the creation of an etymological dictionary for Old Chinese runs into a set of problems quite different from the ones involved in the creation of, for instance, an etymological dictionary for any given Indo-European language. First, Old Chinese itself is, essentially, a reconstructed language, at least in its phonological (and partially morphological) aspects; therefore, before comparing Old Chinese forms to their hypothetical cognate forms in other Sino-Tibetan languages or their hypothetical ancestors in non-Sino-Tibetan languages from which they are supposed to have been borrowed, it is necessary to have a solid historical justification for the Chinese form itself.
Second, unlike Indo-European, Proto-Sino-Tibetan has yet to be properly reconstructed. Despite significant progress in this field, achieved through the work of scholars such as P. Benedict, R. Schafer, N. C. Bodman, W. S. Coblin, S. Starostin, I. Peiros and others, unresolved issues in Tibeto-Burman, let alone Sino-Tibetan, linguistics still abound, with several competing versions of Sino-Tibetan phonology, multiple languages and whole subgroups still linked to the main "bulk" of Sino-Tibetan only by a few suggested cognates, and no definitive corpus of Sino-Tibetan etymologies; the only extensive comparative dictionary of Sino-Tibetan languages in existence is I. Peiros' and S. Starostin's "Comparative Vocabulary of Five Sino-Tibetan Languages ", which is a very serious piece of work, but, with only five languages involved, is obviously not enough. Thus, any etymological dictionary of Old Chinese will, by definition, rest on a somewhat shaky foundation, since it will be attempting to link the Chinese forms to a family that has not yet been properly reconstructed itself.
Third, there is still nothing like a universally accepted scenario for both prehistorical and historical contacts between speakers of Old Chinese and speakers of other language families. At best, everyone recognizes the reality of such a contact between the Chinese (including "proto-Chinese") and various members of the hypothetical "Austric" macro-family, but the exact chronological, geographical, and quantitative scale of most of these contacts is far from being established. Numerous resemblances between Old Chinese and Tai-Kadai (or Hmong-Mien, or Mon-Khmer, &c. &c.) etyma have been spotted by researchers, and much of this evidence has been painstakingly documented in Prof. Schuessler's work, but for quite a few of them we still cannot definitely establish, for instance, the direction of borrowing, or even an approximate time and place in which the borrowing occurred (provided it is indeed a borrowing, and not a chance resemblance, or a trace of some older, deep-level relationship).
All of these complex factors have to be kept in mind by whoever wishes to use Prof. Schuessler's book as a reliable reference source. The biggest limitation (one for which Prof. Schuessler certainly cannot be held responsible), is that this is, indeed, the first etymological dictionary of Old Chinese — and, like all first etymological dictionaries of any language or language family, it is only the first step in the direction of producing a thorough deciphering of the prehistory of the Chinese lexicon. Most of the etymologies have to be taken with a grain of salt, not because they are unrealistic or seriously violate
any principles of comparative Sinology or comparative linguistics as a whole (such cases are quite rare), but simply because there is a high chance that better, more convincing etymologies will replace them in the future, as our knowledge of the linguistic situation in South-East Asia steadily increases.
In order to clarify this statement, it is perhaps reasonable to voice some specific concerns on each of the three issues raised above.
The basic foundation for all of the etymologies offered in the dictionary is the reconstructed Old Chinese (OC) form. At present, the two most comprehensive reconstructions for Old Chinese, available in published form, are Baxter ([1992]) and Starostin ([1989]); both are remarkably similar in many respects, especially concerning the treatment of OC finals, but significantly differ in others, such as proposed subsystems of OC initials. For the most part, Schuessler's version of OC is compatible with Baxter's, although he proposes quite a few of his own modifications to the system, at the same time keeping in touch with more recent hypotheses by L. Sagart ([1999]) and others. Starostin's work is mostly ignored (odd enough, Starostin ([1989]) is not even mentioned in the bibliography!), which is a pity, since many of the hypotheses presented there make a lot of sense from a general Sino-Tibetan perspective (e. g. the reconstruction of a phonological opposition between simple and palatalized affricates and fricatives, which, according to Starostin, is a good match for the same opposition in Tibetan and Lolo-Burmese).
An even worse problem is that, despite the length of the introduction, the reader does not get a very coherent picture of the phonological structure of Old Chinese as a whole. There is no general inventory table for OC, nor do we get a strict and compact set of rules to let us easily understand the transition from OC to Middle and Modern Chinese. In a rather chaotic way, chapters that deal with OC phonology take, as their point of departure, either OC itself ("Initial consonants", "Final consonants"), or Middle Chinese ("MC tones and their Old Chinese equivalents", "Initial and medial j and the MC divisions"), so that the relevant information sometimes gets doubled or even tripled in several locations. Discussion can easily branch from OC to either its descendants (MC), ancestors (Sino-Tibetan) or contacts (Austro-Asiatic, Tai, &c.); it would have probably been a better choice to treat all these connections separately and more systematically, especially considering that many of the statements made in these sections are debatable, and a clearer presentation of them is essential for further research.
The methodology of the reconstruction itself occasionally suffers from relying too heavily on isolated pieces of potential evidence that do not seem to form a coherent system. This is a very important point, since Old Chinese phonology has, for the most part, been established through a very specific brand of internal reconstruction, and one of the perils of internal reconstruction is that it generally relies much more on subjective judgement than external reconstruction, based on identifying regular phonetic correspondences between related languages. In reconstructing the phonology of OC, our most reliable sources are those whose statistic significance is unquestionable, namely, data gained from the analysis of OC rhymes and well-represented xiesheng (phonoideographic) series of Chinese characters. Unsurprisingly, it is the conclusions gained from these types of data that are among the less controversial among specialists on OC reconstruction.
Unfortunately, recent research on OC tends to focus on less reliable data — such as various isolated "oddities" of the xiesheng series; approximate phonetic glosses in Han-time dictionaries of Chinese characters; and, worst of all, on the analysis of so-called "word families" — etymological groupings of characters based on similarities in their pronunciation and meaning, even when these similarities do not seem to form easily explainable patterns. This dense mixture of subjective judgement on the part of both ancient Chinese and modern Western philologists inevitably leads to multiple complications of the existing reconstruction that do not have sufficient backing, and vary greatly from scholar to scholar, at the same time confusing the student and making him/her question the more solid parts of OC reconstruction as well, even up to rejecting the possibility of such a reconstruction altogether.
Fortunately for students and scholars alike, Prof. Schuessler's approach to OC is far more "moderate" than, e. g., Laurent Sagart's ([1999]), or, in certain respects, even that of B. Karlgren (good examples of Karlgren's exaggerated reliance on "word-families" are adduced on p. 8 of the dictionary). Nevertheless, I would still consider it way too dependent on subjective judgement. Thus, on the same p. 8, we meet with the following statement:
"For example, it seems obvious that the words jian ^ *krams 'look at' and Ian ff *ram? 'to see' are related, but what the difference in later tones and the presence / absence of an initial *k-
might have entailed is so far a matter of speculation. On the other hand, we can confidently state that zhi ^ *tskh or *tsks, literally 'something that has been woven', is a regular exopassive derivation from zhi ^ *tsk 'to weave'. We consider both jian and lan, and zhi and zhi to be allofams in their respective word families".
The last sentence, in my opinion, constitutes a grave methodological error. It is clear, both to the reader and to the author, that the two examples described above are of an entirely different nature — the latter being one example in a general pattern that can be illustrated by multiple additional examples, the former being merely a subjective similarity that does not form part of any subsystem; how much more 'obvious' does it really seem that jian and lan are related than, for example, English rod andp-rod (the former is, of course, something you prod with), or drain and s-train, two words whose meanings seriously overlap in some contexts, but which nevertheless have entirely different etymologies? We should also keep in mind that, although OC was unquestionably much richer in phonemic inventory than its modern descendants, and more permissible in its phonotactics, the number of allowed syllables was still limited, and the overall degree of homonymity in it (including homonyms with similar meanings) must have been greater than in the European languages we are used to.
From a methodological standpoint, then, it seems more reasonable for the moment to keep words like jian and lan apart rather than place them under the same etymon (as the author does on p. 305) — thus creating the illusion that there exists some kind of true scientific evidence to consider them "allofams". The same recommendation can be applied to dozens, if not hundreds, of etymologi-zations offered in the main body of the dictionary. An extreme example is KA 'solid, hard' (p. 261), which, according to the author, serves as the basis of a whole bunch of stems like ka-k, ka-ng, ka-r, ka-n, &c., with meanings like 'hard; solid; strong; dry; withered; freeze', scattered throughout Chinese, Tibeto-Burman, and even Mon-Khmer languages — but for most of the quoted examples, there is no real evidence other than a purely intuitive feeling that they really descend from a single ancestor *KA, and, as of now, I see no other possibility than to regard all of them as separate roots.
Sometimes this liberal policy on word-families leads to a fair amount of confusion in particular etymological entries. Thus, although there is a connection proposed between bing ^ (OC *prdij) 'ice' and ling ^ (OC id., with both placed under the same entry (p. 168), it is impossible to reconcile this connection with the proposed etymology: the author reasonably relates OC *prsij to PTB *pam 'snow, ice', but this means thatp- in OC *prsij is part of the root, not a prefix (contra [Sagart 1999]); what, then, does the ^p-less form *rsij represent? To answer this question, the author has to substitute his own reconstruction for W. Baxter's, respectively ^ *prjdij and ^ *b-rjdij; however, since Baxter's clusters like b-r-, &c., are also fairly controversial, he then reverts back to his own reconstruction, tentatively suggesting a semantically remote Austro-Asiatic origin for the forms and stating that 'the unusual initial alternationp ~ r also speaks for a non-ST origin'. So which etymology is preferable — the ST one or the Austro-Asiatic one? And if it is the latter, do the two forms represent independent borrowings from different Austro-Asiatic sources and / or into different OC dialects? Under these conditions, it would seem more reasonable to separate the two words.
Another problem is drawing questionable conclusions from several disconnected pieces of hypothetic evidence, which also happen to contradict other evidence. A case in point here is the author's reconstruction of an initial consonant cluster *ml- in OC (pp. 89 - 90). It has long been recognized that the initial lateral resonant in OC gives bifurcated reflexes in MC: OC *l- > MC j- in most cases, but also dz- in a smaller group of examples. Prof. Schuessler attempts to explain the difference by setting up an OC cluster *ml- (or, rather, *m-l-, which he separates from the unhyphenated *ml-; the phonetic difference between the two is left unexplained), for which MC dz- is the regular reflexation, and gives various examples of both proposed Tibeto-Burman cognates and Miao-Yao borrowing sources that seem to confirm this reconstruction.
The problem, of course, is that for most of these cognates, we are not really sure they have to be reconstructed with an initial *m- for Proto-Sino-Tibetan. Thus, ^ *m-tek 'eat' is compared with PTB *m-lyak 'to lick', but there is almost no evidence, except for a few scattered Kuki-Chin forms, that the PTB form indeed contained an initial *m-. The same goes for PTB *(m-)loy 'canoe' (the PTB reconstruction here and in most other places is taken directly from [Benedict 1972]) > Kuki-Naga *m-lauij,
but no m- is found in any other TB language. About the only good example in this group is PTB *m-lay 'tongue' (where sufficient evidence is found in more than one Tibeto-Burmese subgroup to assert that m- really reflects an old state of affairs rather than a local innovation) = OC ^ *m-le? 'to lick', but it goes without saying that a hypothesis cannot be solidly based on but one example of such a kind.
Two more examples show how MC dz- < OC *m-l- is connected with words with initial prenasalized consonants in Miao-Yao: MC ^ dzjuet 'glutinous millet' <> MY *nblut 'glutinous, sticky'; MC ^ dzjat 'tongue' < OC *m-let <> Yao *byet', Miao *nplaf id. However, the two Miao-Yao parallels adduced here are actually quite different. The former, according to [Peiros 1998: 141], is PMY *mblVt > PM *mblo-Q (Hmu ns, Hmong mplou), PY *mblut (Mien bjut, Kimmun blot). Borrowing from an early ancestor of PMY is likely here, with OC *lut or *(m-)lut having no better etymology. On the other hand, the PMY form for 'tongue', also according to Peiros, is rather *mbret > PM *mbls-Q (Hmu ni, Hmong mplai), but PY *mbjet (Mien bjet, Kimmun bjet). The two types of correspondences are not distinguished in [Wang & Mao: 1995] (in both cases, *mbl- is reconstructed for PMY), but they do exist, and a reconstruction like *mbret is in good agreement with other types of correspondences where PM *-l- or *-r- is the equivalent of PY *-j- (see [Peiros 1998: 136 - 137] for the full set). Furthermore, Prof. Schuessler seems to agree with this, as he quotes the PMY reconstruction for 'tongue' as *nbret rather than *nblet (p. 467).
Thus, if PMY 'tongue' is really *mbret (or *nbret), there is no solid ground for assuming that the word has anything to do with OC ^ *(m-)let 'tongue'; it is not clear why OC *ml- should have been borrowed as PMY *mbr- instead of *m(b)l- proper.
A further scrap of "evidence" for OC m-l- is allegedly found within Chinese itself: since OC ^ *lsy 'string' > MC dzjsy uses the character ^ mrsy? 'toad' > MC mey as phonetic, it has to be reconstructed as *m-lsy (and this is in agreement with MC dz- as a proper reflexation of OC m-l-). However, it is not mentioned that the same phonetic series also includes such a well-known character as MC ^ jisy 'fly', which, according to the author's own rules, could not have contained an initial *m- in OC, since it is MC jisy, not dzjsy. So this character has to be reconstructed as OC *lsy or *jsy (p. 576) — which makes the preceding argument hollow, since it proves that not every character in this phonetic series has to contain an initial m-, and that OC ^ *lsy rather than *m-lsy would be perfectly acceptable for it. In fact, the connection between *C-r- / *C-l- and *r- / *l-type syllables within the same phonetic series need not necessarily presuppose that the *r-/*l-type syllables necessarily or almost necessarily go back to *C-r-/*C-l-type syllables as well; this idea, upheld by W. Baxter ([1992]) and L. Sagart ([1999]), occasionally has some interesting applications, but cannot function as a universal rule.
Additional concerns: if it is indeed true that MC dz- < OC *m-l-, whereas MC j- < OC *l-, how do we explain numerous "doublets" attested in MC dictionaries which, for the exact same characters in the exact same meanings, give both j- and dz- (a good list can be found in [Starostin 1989: 514])? Do they all represent OC alternations between l- and m-l-? And also, what do we do with various potential TB cognates for OC words with *(m-)l- > MC dz- that have no traces of an initial nasal whatsoever? E. g., the author gives OC f *m-luk > MC dzjwok 'to ransom' = Written Tibetan blu-ba, blu-s 'to buy off, ransom', saying that "the WT forms can theoretically derive from earlier *mlu(t)"; but he does not mention the quite possibly related Kachin gs-lo 'to settle, pay an indemnity' or Lushai hlo? 'wages, salary', or Kiranti *lok 'lend', with no nasals anywhere; also, if Tibetan bl- < ml-, why doesn't any other example in Schuessler's group contain Tibetan bl- to confirm this? Finally, some very reasonable etymologies seem to be rejected simply because they do not fit in with the rule, e. g. OC ^ *lsy 'to mount, ascend' > MC dzjsy = Tibetan lay 'rise, get up', Kachin luy 'ascend' ([Peiros & Starostin 1996: III: 19], not quoted in the dictionary).
Overall, with virtually no internal evidence for OC *m-l-, and the external evidence being inconsistent, controversial, and contradictory, it feels more prudent to simply scrap the whole thing — for the moment, at least, until it becomes possible to present a more convincing structure. Currently, however, it is more reasonable to simply admit a dialectal variation reason for the bifurcation of OC *l- into dz- and j- in MC.
Speaking of dialectal variation, another minor flaw of the dictionary is that occasionally, theoretical statements are made that do not seem to have sufficient basis in fact and may give a somewhat misleading picture of the historical development of Chinese. Thus, on p. 7 the author introduces the concept of "rural dialects", saying that "words with rare and unusual features typically
have meanings with a rustic or vulgar flavour". He then lists the following words as "rural" or "vulgar" (identified on the basis of having initial x- < OC voiceless resonants instead of expected regular reflexes): "beard, to face / toward, ribs (of a horse), to know, to vomit, to rear animals, stupid, to roar, tiger, pig". While the general explanation of x-containing words as dialectal varieties may be true, it is absolutely not clear what exactly is so "rural" or "vulgar" about notions like "beard", "to face", "to know", or "tiger". Further below, we also learn that words like "old man", "to fall" (!), and "wrist" also "conjure up a rural sphere" (although other quoted words, like "farm" and "sickle", certainly could do that). It is simply hard to understand why it is necessary to set up and insist on the "rural" vs. "non-rural" (urban?) distinction when the data do not confirm it.
Already in our discussion of the hypothetical OC *m-l-, we have indicated certain problems with TB etymologies for supposed OC cognates that crop up in the dictionary. Fortunately, they are not ubiquitous, and in most cases, TB cognates are reasonable both from the phonetic and semantic sides. Not all the known etymologies are quoted; where formerly proposed etymologies contradict the author's ideas on phonetic correspondences between OC and TB, they are usually omitted. This is understandable, but, again, one should keep in mind that the development from PST to OC as traced (in a rather chaotic form) in the introduction is not the final word on Sino-Tibetan, and that various alternative hypotheses to some of the proposed solutions are quite probable. Thus, the idea that aspirated consonants in OC are of a secondary origin (and, therefore, we should not look for specific TB correspondences to OC aspirates) is mostly based on "word family" hypotheses and ideas of sound symbolism that are controversial and unprovable (pp. 58 - 61); Benedict's earlier explanation of OC unaspirated stops as reflecting traces of lost prefixes ([Benedict 1972: 165]) or Starostin's even more complex set of rules to explain the interaction between voiced, voiceless, and aspirated articulation in ST ([Starostin 1989: 136 - 146, &c.) both seem more preferable.
Concerning those etymologies of Chinese words for which an external, non-ST source has been suggested, some of the methodological problems that inevitably crop up here are already well described by the author himself on pp. 9 - 10 ("Identification of cognates") and need not be addressed further in this review. One thing, however, that must be mentioned is that the author's historical scenario for contacts between Chinese speakers and speakers of other language families seems to be seriously incomplete. Throughout the dictionary, he most prominently quotes such potential sources or targets of borrowing as Austro-Asiatic, Tai-Kadai, and Miao-Yao languages. These are, indeed, almost universally recognized as donors to / recipients from OC, although nobody has so far presented a coherent chronological model of these contacts. However, while almost overstating the importance of these contacts, he insists that "other language families" had "little contact with Chinese" (p. 133). This can hardly be so, at least in respect to two other prominent families:
(a) Austronesian. Numerous similarities between Austronesian and Chinese have been spotted by L. Sagart, who has even ventured so far as to postulate a genetic relationship between the two based on these connections (the so-called 'Sino-Austronesian' theory). In reality, most of the connections can be explained in contact and areal terms (with a few ones possibly indicating some ultra-deep level relation), but they are so numerous that ignoring them altogether is hardly a good solution. Thus, for OC ^ *tha? 'earth' we are told that "there is no obvious cognate and etymology" (p. 502), apart from a few phonetically dubious Austro-Asiatic parallels, but, in fact, there is a perfectly fine Austronesian connection: cf. PAN *-taq 'earth, soil' [Sagart 2002: 6]. Apart from Sa-gart's works, numerous examples on Chinese-Austronesian or, wider, Sino-Tibetan-Austronesian connections are given in [Peiros 1998] and other sources.
(b) Altaic. The idea that speakers of Altaic languages (or, if one does not subscribe to the Altaic theory, Turkic, Mongolic, Tungus-Manchu, and "pre-Korean" languages) had little or no contact with the earliest Chinese speakers is very odd, considering that nearly all the lands to the north of the Yellow River must have been populated with predominantly Altaic speakers long before the Chinese actually settled in its basin, and that epigraphic and historical records in China are full of references to contacts with nomadic tribes of the North that, in all probability, spoke Altaic languages.
It is true that the issue of Altaic-Chinese language contacts has been relatively unexplored, due to a variety of factors such as:
— typological dissimilarities between Altaic languages and Chinese, making such comparison less comfortable than the one between Chinese and more "compatible" languages such as Tai-Kadai or Mon-Khmer;
— emphasis on later, historical period contacts between Han-time Chinese and Turkic / Mon-golic tribes, obscuring earlier contacts;
— lack of direct prehistorical knowledge (earlier than the first centuries A. D.) about the ancestors of Turkic, Mongolic, &c., and the never-ending debate about the status of the Altaic connection.
Nevertheless, all of these problems are purely technical and psychological, and many of them can now be overcome with the appearance of An Etymological Dictionary of the Altaic Languages ([EDAL]); material from this source can be used for productive comparison with reconstructed OC forms even if one is skeptical about Altaic as a whole, since it also includes reconstructions for uncontested branches. A few examples of why exploring the Altaic connection is a must for further Sinologic studies, taken from [Starostin 2007], will suffice:
(a) For ^ tu 'hare' Schuessler offers an OC reconstruction *lhah, saying that "the OC initial *lh- is revealed by the graph's use as phonetic in a word for 'tiger' in Zuozhuan" (p. 502). However, there is no such character as ^ in attested versions of the Zuozhuan. The character in question is found in the Shuowen, where it is defined as part of the Chu word for 'tiger', ^The Zuozhuan quotes obviously the same word as whereas the Hanshu (not the Hdu Hanshu, as stated on p. 282 of the dictionary) quotes it as ^fp; for the latter Schuessler gives the modern reading yu-shi < MC ?jwo-sjak, but in this particular use, the Jiyun apparently gives the MC reading ?jwo-duo > modern yu-tu. What this all leads to is to state that there is no obvious connection between the regular OC form for tiger, ^ *hla?, and the above dialectal word for 'tiger', which could have easily been OC *?a-da(k) instead of the suggested *?a-lhak.
This, in turn, means that the word for 'hare', which Schuessler reconstructs as OC *lhah, may easily have been *thah or even an earlier *thaks. If so, then the connection with Middle Korean twos(ki) 'hare' is more complex than a borrowing 'from Chinese in connection with the animal cycle'. The Korean word finds a good Altaic parallel in Proto-Tungus-Manchu *tuksa- 'hare' ([EDAL: 1451]), which, in turn, is in that language derived from a verbal stem with the meaning 'to run'. Thus, it is, in fact, the Chinese word that is more likely to represent an old borrowing from an Altaic source rather than vice versa.
(b) OC ^ *mrsk 'wheat' is compared to Tibetan bra-bo 'buckwheat' and possibly also to Proto-Lolo-Burmese *g-ra id. (p. 374); this etymology, despite the obvious phonetic issues (final -k, m- / b- alternation, &c.), could be acceptable if no other were available, but cf. Tungus-Manchu *murgi 'barley', Middle Korean mir 'wheat', Old Japanese mugji 'wheat, barley' < Altaic *miurgu ([EDAL: 935]). With three consonantal segments matching, this seems a more likely solution.
(c) OC ^ *dek 'enemy' is compared to PTB *m-ta:y 'avenge, retaliate' (p. 209), but cf. rather Turkic *jagi 'enemy, war', Mongolic *dajin 'war' < Altaic *dagi ([EDAL: 457]).
Additionally, in [Starostin 2007] one can find reasonable Altaic etymologies for several important items in Chinese lexicon that are altogether lacking in Prof. Schuessler's dictionary — such as qin 'qin, musical instrument', que 'magpie', and even the famous "Chinese unicorn" ^^ qilin (traced back to an Altaic word for 'deer'). Not all the etymologies offered there will find acceptance, but cases like these, nevertheless, seem to disprove, once and for all, the idea that it is fruitless to include Altaic in the list of potential "donors" to the Chinese language.
Concluding this review, I would once again like to stress that all of the issues raised here do not in any way diminish the general importance of the work. On the contrary, even where one might feel the author has missed some valid data, or has jumped to unwarranted conclusions, this should still constitute a stimulus for discussion and improvement rather than a pretext to dismiss the work altogether. My only concern is that scholars and students that are not well-versed in historical Chinese and Sino-Tibetan studies may make the mistake of taking this dictionary for the definitive word on Chinese etymology, which it most certainly is not. I would not, therefore, recommend it for "blind" usage to any such student or scholar. However, if one approaches it with due caution, keeping in mind most of the caveats that have been voiced above, it will certainly be a valuable aid in any kind of research that has to do with the history of Chinese language, be it linguistics, philology, or history as such.
References
Baxter 1992 — William H. Baxter. A Handbook of Old Chinese Phonology. Mouton - de Gruyter, Berlin -New York, 1992.
Benedict 1972 — P. K. Benedict. Sino-Tibetan: A Conspectus. Cambridge University Printing House, 1972.
EDAL — S. Starostin, A. Dybo, O. Mudrak. An Etymological Dictionary of Altaic Languages. HdO Vol. 8-1, 8-2, 8-3. Leiden: Brill, 2003.
Peiros 1998 — I. Peiros. Comparative Linguistics in Southeast Asia II Pacific Linguistics Series, Vol. 142. Australian National University, Canberra, 1998.
Peiros & Starostin 1996 — I. Peiros, S. Starostin. A Comparative Vocabulary of Five Sino-Tibetan Languages (6 vols.). Melbourne, 1996.
Sagart 1999 — L. Sagart. The Roots of Old Chinese II Current Issues in Linguistic Theory, Vol. 184., Amsterdam - Philadelphia: John Benjamins Publishing Company.
Sagart 2002 — L. Sagart. Sino-Tibeto-Austronesian: an Updated and Improved Argument. Paper presented at Ninth International Conference on Austronesian Linguistics (ICAL9). 8 - 11 January 2002. Canberra, Australia.
Starostin 1989 — С. А. Старостин. Реконструкция древнекитайской фонологической системы [A Reconstruction of the Old Chinese Phonological System]. Moscow, "Nauka", 1989.
Starostin 2007 — S. A. Starostin. Altaic and Chinese II С. А. Старостин. Труды по языкознанию. М.: «Языки Славянской Культуры» [Works on Linguistics. Moscow: "Languages of Slavic Culture"]; pp. 850 - 853.
