Reconstruction of Middle Sinitic

« previous post | next post »

"What 'Ancient' Chinese Sounded Like – and how we know" (YouTube 7:56)


Based on the best modern scholarship and the most vital premodern primary materials, this short presentation is really quite impressive and has some cute touches, as when Bernhard Karlgren (1889-1978), the founder of Western studies on the scientific reconstruction of Middle Sinitic and Old Sinitic, blinks at us (5:45).

There is a general consensus on the sounds of "Middle Sinitic", which dates to roughly circa 600 AD, although we don't know for sure on which topolect it is based or whether it is a homogenized abstraction of a congeries of topolects.  "Old Sinitic", which supposedly dates to around 600 BC, is more chimerical still, with the most critical researchers having abandoned the quest to recover it.

On solider ground is Eastern / Later Han (25 AD-220 AD) Sinitic, since the leading scholar on the historical phonology of this period, W. South Coblin, has extensively utilized contemporary glosses, topolectal data, and Sanskrit and Prakrit transcriptions.

My own approach to the reconstruction of Old Sinitic is to emphasize words from other languages for which the pronunciation is better known that have been borrowed into Sinitic.  I have published many of these terms in Language Log posts (see the list of "Selected readings" below) and elsewhere.

Axel Schuessler's pathbreaking dictionary of Old Sinitic etymology is promising, especially since it systematically employs evidence from cognates in non-Sinitic languages of East Asia, South Asia, and Southeast Asia, and occasionally from elsewhere.  Such sources have enabled Schuessler to provide what he calls Minimal Old Chinese reconstructions.  Studies of the rhymes in the Poetry Classic (Shījīng 詩經) and other types of verse will continue to offer valuable evidence for the reconstruction of Old Sinitic.

So much for phonological reconstructions.  Another major endeavor is to distinguish between classical / literary Sinitic and vernacular Sinitic.  The largest project in this regard is the massive dictionary of Middle Vernacular Sinitic (MVS) which Zhu Qingzhi and I have been compiling for the last two decades and more.  Concentrating on the lexicon, morphology, and grammar of the middle period, our corpus is based on tens of thousands of quotations from texts that may be located precisely in time and space.  These data are potentially of tremendous value when used in conjunction with historical reconstructions arrived at by means of the methods described above.


Selected readings

[h.t. John Rohsenow]


  1. Antonio L. Banderas said,

    August 23, 2020 @ 1:37 pm


    A Phonological History of Chinese

    by Zhongwei Shen, University of Massachusetts, Amherst

    Publisher: Cambridge University Press
    Online publication date: June 2020
    Print publication year: 2020
    Online ISBN: 9781316476925

  2. Philip Taylor said,

    August 23, 2020 @ 1:59 pm

    Can anyone confirm that "Nativlang" is, in fact, Joshua Rudder ? I find myself frustrated when I do not know to whom I am listening, but despite my best efforts I can only say that I think that the narrator is Joshua Rudder — I cannot prove it.

  3. Alexander Browne said,

    August 23, 2020 @ 2:14 pm

    Not a big deal, but the embedded YouTube video doesn't start at the beginning.

    [VHM: fixed now]

  4. Philip Taylor said,

    August 23, 2020 @ 2:19 pm

    Agreed — the trailing "&t=302s" is the source of the problem. Alternative link here.

  5. John Laviolette said,

    August 23, 2020 @ 2:26 pm

    @Philip Taylor:

    I don't know who that is, but on the NativLang Patreon, he writes "I'm a language adventurer named Josh."

  6. John Laviolette said,

    August 23, 2020 @ 2:28 pm

    @Alexander, Philip:

    I'm pretty sure the YouTube link points to 7 min 56 seconds on purpose.

  7. Victor Mair said,

    August 23, 2020 @ 2:40 pm

    @John Laviolette

    And what might that purpose be? 7:56 is just the length of the video.

  8. Philip Taylor said,

    August 23, 2020 @ 3:04 pm

    302 seconds is 00:05:02, unless I miss my guess. 00:07:56 would be 476 seconds, would it not, John ?

  9. David Marjanović said,

    August 23, 2020 @ 4:11 pm

    Wikipedia: Reconstructions of Old Chinese, presenting and comparing the reconstruction systems of Karlgren (1940–1957), Yakhontov (1959–1965), Pulleyblank (1962), Lǐ (1971), Baxter (1992), Zhèng-Zhāng (1981–1995) and Baxter & Sagart (2014). Schuessler (1987) is said to have used Lǐ's system, Schuessler (2007) "a simplified version of Baxter's". The latest work cited as part of the ongoing discussion is this by Baxter & Sagart (2017, paywalled except for the first page) defending their 2014 system against a review by Schuessler (2015); in the process they briefly review the history of the field.

  10. David Marjanović said,

    August 23, 2020 @ 4:14 pm

    Oops, it's in the January 2017 issue which was only published in February 2018. I'm surprised it wasn't published online in time; I'm used to papers being printed a year or two after they're published, but issues coming out long after they're supposed to is so 1990s!

  11. David Marjanović said,

    August 23, 2020 @ 4:15 pm

    …"it" referring to a work cited in my previous comment which was shown to me as published but has now disappeared without a trace. What happened?

  12. John Laviolette said,

    August 23, 2020 @ 4:48 pm

    @Victor Mair:

    Sorry, I meant the starting point, whatever that might be. Seemed like the video started at a point just before the narrator introduced the term "Middle Sinitic", which is what you addressed first.

  13. David Marjanović said,

    August 23, 2020 @ 5:32 pm

    (My transiently disappeared comment has reappeared, nothing is missing anymore.)

  14. Chris Button said,

    August 23, 2020 @ 9:34 pm

    "Old Sinitic", which supposedly dates to around 600 BC, is more chimerical still, with the most critical researchers having abandoned the quest to recover it

    Wikipedia: Reconstructions of Old Chinese … Pulleyblank (1962) …

    Granted, Pulleyblank's work is dotted around in articles all over the place and written in a style that's hard to penetrate and thoroughly disorienting even for a general linguist, but almost half a century of outstanding (albeit not flawless) scholarship came after that that 1962 opus. It wasn't a one-hit wonder after all (admittedly a couple of other articles are briefly mentioned in the wikipedia entry).

    In my humble opinion, the only thing more amazing than the field not having caught up with Pulleyblank is just how far the field still has to go until it does.

  15. Chris Button said,

    August 23, 2020 @ 10:18 pm

    To be clear, my post above isn't intended to knock the many other truly great contributions by other scholars (Pulleyblank lent on not a few of them himself), but rather to make a more general comment about the state of the field and (historical/comparative) linguistics in general.

    Personally, I've always felt James Matisoff's brutally honest work on Proto-Tibeto-Burman (his allofams provide scant room for any cherry-picking of data or overwought phonological theory) has often provided a supportive counterpart to Pulleyblank's ruminations on Old Chinese. I don't think the field has ever quite realized that though.

  16. Chris Button said,

    August 23, 2020 @ 10:20 pm

    *sorry, "leant" (i.e., "leaned") not "lent"

  17. Keith said,

    August 24, 2020 @ 1:50 am

    I'm perplexed by his pronunciation of "comitative" as /kɒˈmaɪdədɪv/ and not /ˈkɒmɪtətɪv/as and (in "What English does – but most languages can't") of "kabyle" as /kəˈbaɪl/ and not /ka'bil/…

    I looked up Kabyle in Wiktionary and found that /kəˈbaɪl/ seems to be the accepted American pronunciation, but it still sounds odd to me.

  18. Chas Belov said,

    August 24, 2020 @ 1:57 am

    Interesting. I got an ad for, a language-learning site. The irony came when they referred to the beauty of Chinese characters, but were showing simplified characters. I realize it's an opinion, but it's the traditional characters I find beautiful.

  19. Philip Taylor said,

    August 24, 2020 @ 6:33 am

    " it's the traditional characters I find beautiful". And you are not the only one, Chas. Pinyin and the simplified script may well be of benefit to the masses, but if we (the world) were to lose the traditional Chinese script forever that would, to my mind, be an absolute disaster, on a par with the loss of the Library of Alexandria.

  20. Scott Mauldin said,

    August 25, 2020 @ 6:12 am

    @Philip Taylor
    But there's no way we would ever lose traditional Chinese script (barring some global informational catastrophe that would also lead to the loss of many other things); it's documented in millions of books, websites, engravings, dictionaries, etc. all over the world.

    Also, the Library of Alexandria is overly romanticized as the unmatched pinnacle of accumulated wisdom in the ancient world, but there were dozens of such large libraries throughout the ancient Mediterranean and Near East, including at Antioch, Rome, Ephesus, and Constantinople. There was a high degree of redundancy between all of them, and they would almost all have had their own copies of the most classic and significant works; plus there were things held in other libraries that were not held in Alexandria. Not even to mention that private collections were generally the way that most texts were kept, and those are uncountable.

  21. Twill said,

    August 25, 2020 @ 9:12 pm

    Not to mention the very one-dimensional traditional vs simplified debate ignores the millennia of writing before Kangxi's dictionary codified the arbitrary "traditional" standard, that the contemporary traditional standards deviate from that arbitrarily, that many simplified characters are ad(a/o)ptions of "unorthodox" or informal forms, that simplification was not a black and white political issue and indeed Japan equally maintains a simplified standard, etc. etc.

  22. alex said,

    August 29, 2020 @ 1:21 am

    I was wondering if there is any information on the tempo in which people read traditional Chinese Poems

  23. John Rohsenow said,

    September 1, 2020 @ 3:13 am

    Re:the tempo at which people read trad. Chinese poems:
    one is tempted to reply, it depends on whether they pronounced the traditional ru-sheng (stopped) endings or not.

  24. Jongseong Park said,

    September 4, 2020 @ 7:37 pm

    Regarding the pronunciation of Kabyle, the English pronunciation [kəˈbaɪl] is a better match for the Algerian Arabic قبايل‎ Qbayel corresponding to Standard Arabic قبائل Qabāʾil or for the Kabyle root in Iqbayliyen ("Kabyle people") and Taqbaylit ("Kabyle language").

    "Kabyle" was the spelling used by Englishman Thomas Shaw in his Travels in Barbary and the Levant (1738). The French [kabil] is probably a spelling pronunciation from the English-style rendering of the Arabic or Kabyle name.

RSS feed for comments on this post