Old Sinitic reconstructions and Tibeto-Burman cognates

« previous post | next post »

[The following is a guest post by Tsu-Lin Mei.]


The Old Chinese reconstruction of Gong Hwang-cherng and James Matisoff is not only internally consistent, but can be shown to have a Tibeto-Burman counterpart through Sino-Tibetan comparative studies.  Gong Hwang-cherng's Collected Papers on Sino-Tibetan Linguistics 龚煌城, Hàn-Zàngyǔ yánjiū lùnwén jí《汉藏语研究论文集》(2002) has about 300 cognate sets — involving Old Chinese, Written Tibetan, Written Burmese, and reconstructed Tangut. I am writing a paper whose purpose is to unite Gong's work with Zàng-Miǎn yǔzú yǔyán cíhuì《藏缅语族语言词汇》(Lexicon of Tibeto-Burman languages), edited by Huang Bufan 黄布凡 (1992). So far I have 142 cognate sets and can testify that Gong's cognate sets on the whole hold water.

Huang Bufan's volume brings together the lexicon of 40 Tibeto-Burman languages (all in China) with Written Tibetan and Written Burmese. The Written Tibetan forms are quite good, the same as in Jaeschke's Tibetan-English Dictionary.  The Written Burmese forms are in present-day Rangoon pronunciation, which is not satisfactory.  Gordon Luce and Nishi Yoshio have provided accurate transcriptions of the 12th-century Burmese inscriptions.

As mentioned above, I have now 142 cognate sets culled from Gong Hwang-cherng (2002).  I am presently trying to unite Gong's philological evidence with the descriptive data of 142 lexical items (e.g., sun, moon, eye, ear, breast, father, mother, grandmother, uncle, one, two, three, five, six, seven, eight, nine, hundred, horse, bee-fly, butterfly) in Zàng-Miǎn yǔzú yǔyán cíhuì《藏缅语族语言词汇》(Lexicon of Tibeto-Burman languages), item by item.

The work is quite interesting.  It involves the internal history of the Tibetan language, internal history of Tibeto-Burman, etc. Some of these areas have been covered by James Matisoff, and others are terra incognita.  My young colleague Jackson Sun at Academia Sinica has been working on comparative Tibetan dialects as well as Proto-rGyalrong 嘉戎。  I am making up the phonological history as I go along.  The final product will be 100 cognate sets, supported both by philological evidence and by evidence from living Tibeto-Burman languages.

As noted above, Gong Hwang-cherng has 300 cognate sets, involving Old Chinese, Written Tibetan, Written Burmese, and Tangut.  The work I am now doing is to unite the philological evidence of Gong and the lexicon of 40 living Tibeto-Burman languages via the WT and WB forms which are included in both works.  The result is a comparative word list of about 140 items.

Why am I doing this?  While there are many books on Sino-Tibetan comparative linguistics, there is no succinct account of the reasons why we believe Sino-Tibetan-Burman are genetically related.  The answer is quite simple. (1) There are between 140 to 300 cognate sets involving Old Chinese and Tibeto-Burman languages.  (2) Sino-Tibetan has a causative *s- and a nominalizing *-s.  Both (1) and (2) have been in the literature since 2000, but nobody took the trouble to give a short, easy- to-understand account.  This is what I am doing and I am writing in Chinese.

I have always thought that in Chinese lexicography there should be a section which tells the reader which Chinese words have Tibeto-Burman cognates and which do not.  The American Heritage Dictionary does that for English; for every English word, the Dictionary gives its Indo-European root, if any.  We should be able to do that for Chinese.


VHM:  This is the seventh post in a series on Eurasian connections and Old Sinitic reconstructions.  Previous posts include:

I am already planning at least one more post in this series.


  1. Guillaume Jacques said,

    April 18, 2016 @ 8:56 am

    A list of cognates between OC and Rgyalrongic languages is available since 2005, and (surprisingly) very few new cognates have been discovered since.

    I should say that the potentially reconstructible morphology common to OC and the rest of the family (for reasons explained here, I think that the notion of "Tibeto-Burman" as a clade including all languages apart from Chinese, should be abandoned) is not restricted to the handful of affixes mentioned by Mei Tsu-lin. Other affixes include at least:
    (1) The denominal *s-
    (2) The anticausative prenasalization (which Mei confuses with the causative; see the discussion in this article on the causative and that one on the anticausative, published in Folia Linguistica Historica and LTBA respectively).
    (3) Potentially various *-s and *-t suffixes, as suggested in this paper (submitted to the Bulletin of Chinese Linguistics).
    Chinese is one of the most innovative branch of the Sino-Tibetan family, and the remnants of morphology still present in OC are open to various interpretations, and difficult to analyse from a Chinese-only perspective. For this reason, it is advisable to follow more conservative languages, including Rgyalrongic and Kiranti, to build a reconstruction framework allowing to make sense of otherwise opaque alternations.

  2. Jason said,

    April 18, 2016 @ 10:02 am

    142 cognates or loan words?

  3. Chris Button said,

    April 18, 2016 @ 11:00 am

    One of the great things about Gong's Old Burmese reconstruction is that he does not try to squeeze "e" or "o" into the system. Matisoff's recent paper "On the Demise of the Proto-Tibeto-Burman Mid Vowels" (2015) happily also removes "e" and "o" from PTB.

    A few quibbles aside (e.g. no proper account for the distinction between the rhymes he treats as -i and -ij), the only major issues I have with Gong's Old Burmese is that it does not account for the curious distribution of medial -w- (a better treatment of which can account for Gong's "u") and, turning to inscriptional evidence, medial -j- as well (which can account for his "i"). Both of these point to a vertical schwa/a system. In this regard, it should of course be noted that Benedict's Sino-Tibetan Conspectus (1972:69-70) and Matisoff's Handbook of PTB (2003:159) both seem to suggest that the eventual solution is perhaps going to come from a vertical schwa/a system.

  4. Eidolon said,

    April 18, 2016 @ 12:21 pm

    I think a "simple" explanation of the genetic relationship between Sinitic and Tibeto-Burman should include information about the criteria linguists use to classify related languages. Cognate sets are, to my knowledge, not nearly enough for showing genetic relationship, as Vovin has, for example, shown in the case of "Altaic."

  5. David Marjanović said,

    April 18, 2016 @ 3:58 pm

    as Vovin has, for example, shown in the case of "Altaic."

    Ooh… you're opening a huge can of worms here… and oversimplifying: the modern incarnation of the Altaic hypothesis isn't just based on word lists, or on word lists and regular sound correspondences for that matter, and the controversy has a much longer history (it's not Vovin vs. everyone else, or anything like that).

  6. Anonymous Coward said,

    April 18, 2016 @ 11:35 pm

    If Altaic had a cognate set as impressive as Chinese – Other ST languages, no sane etymologist would show such a hostility to the proposal.

  7. AntC said,

    April 19, 2016 @ 2:12 am

    I am making up the phonological history as I go along. is a possibly unfortunate turn of phrase.

    Does this mean that the principles of phonological change are already understood, and Tsu-Lin Mei is applying them to particular cognates?

    Or are there no well-agreed principles, so this work is developing them? As a non-philologist looking on, I often wonder how such principles could be attested, in the absence of access to native speakers.

    Am I right in thinking that all the affixes mentioned with preceeding asterisk are inferred, not attested?

  8. Y said,

    April 19, 2016 @ 11:09 am

    So 'horse' is reconstructed back to Proto-Sino-Tibet-Burman?

  9. David B Solnit said,

    April 19, 2016 @ 1:47 pm

    Well, TB has *m-raŋ and Baxter & Sagart's Old Chinese has *mraʔ (with pharygealized initial). But I don't think 'horse' is your best bet to start for evidence of genetic relation.

  10. Greg Pandatshang said,

    April 19, 2016 @ 2:57 pm

    lots of apparently independent roots in Baxter & Sagart's wordlist for different types of horses and other concepts related to horses, not just 馬. Beckwith argues that 馬 mǎ is loan word from Indo-European (or a wanderwort?) cognate with English "mare". I couldn’t hazard a comment as how well supported that claim is.

  11. David B Solnit said,

    April 19, 2016 @ 5:39 pm

    Yes, that's what I mean; wanderwort is more like it. And there's also Manchu morin, Mongolian mor', Japanese uma, and so on. And maybe TB 'horse' isn't related to those at all but is indigenously derived from TB *m-raŋ 'high', as if 'the high (> noble, important, etc) one' (Paul Benedict's proposal). Probably worth a separate posting of its own. See Laurent Sagart's take on this on p.196 of his The Roots of Old Chinese (Chinese word unlikely to be from Indo-European, possibly an early borrowing from TB). Google "tibeto-burman horse high" and you'll get that page in Googlebooks.

  12. Y said,

    April 19, 2016 @ 9:35 pm

    That's what I figured, which makes me wonder how good is the control for Wanderwörter and other borrowings in this particular project.

  13. Randy LaPolla said,

    April 20, 2016 @ 12:57 am

    Hi Prof. Mei,
    Glad you are doing this. It would be great if such etymological information could be included in Chinese dictionaries, and you are the right one to make the IL at CASS and other dictionary makers pay attention! When we write dictionaries or glossaries of TB languages we generally do try to include such information.

    I think you might be being overly conservative in shooting for only 100+ cognate sets, but the problem is in working out the detailed correspondences between the forms (sometimes clear cognates don't look alike, e.g. 'dog' in Tibetan and Lahu), and also in looking for cognates that aren't lined up in the Lexicon of Tibeto-Burman Languages due to semantic shifts. You don't seem to be utilising the STEDT database, which has many more languages and forms than the Lexicon of Tibeto-Burman languages (and includes all of the forms in the latter), but from a methodological point of view I can understand the idea of working with a smaller number of languages so that you can work out the correspondences within that fixed set very clearly. That is sort of what Benedict did in the Conspectus with his small set of what he considered core languages and then added to it when possible for extra support.

    The extent to which the correspondences are clear and regular between the TB forms and Prof. Gong's reconstructions of Old Chinese forms will also be a validation of Prof. Gong's reconstructions.

RSS feed for comments on this post