Old Chinese.

Another amazing gift of the internet: Old Chinese: A New Reconstruction, by William H. Baxter and Laurent Sagart, costs over $60, but you can download its Old Chinese reconstructions for free here. Via Matt at No-sword, who quotes this entertaining footnote:

We adopt the term “Kra-Dai” proposed by Ostapirat (2000) in place of the traditional “Tai-Kadai,” since to Thai speakers, “Tai-Kadai” evidently sounds unintentionally funny, meaning something like “Tai, or whatever” (Montatip Krishnamra, p.c.)

He also mentions that it’s a crappy printing job, which is unforgivable in a book being sold for such a high price (I know, I know, many academic books are even pricier).


  1. This one is extremely interesting

    鮮卑 *s[a]r.pe Xiānbēi

    It is well known that ethnonym Serb is of Iranian origin (Sarmatian tribe Serboi mentioned by Pliny the Elder).

    Iranian influence and presence of Iranian speaking peoples is well attested in ancient Mongolia, so one can’t exclude that the Xianbei were, in fact, “Serbs”, ie originally Iranian ethnic group with Iranian ethnonym.

  2. And OC reconstruction of 單于 chányú: Xiōngnú ruler is even more interesting.

    單 is reconstructed as *dar and 于 as *ɦʷa

    which would give *darɦʷa.

    This looks like Mongolian “darga” (classical Mongolian “daruga”) which means boss, chief (from verb daraqu – to press)

  3. He reconstructs 烏桓 Wūhuán as *ʔˤaɦʷˤar, ie Avar!


  4. 单于 darga

    So for once the crackpot Zhu Xueyuan got it right!

  5. Trond Engen says

    *darɦʷa > chányú
    *ʔˤaɦʷˤar > wūhuán

    Purely superficially speaking, the difference in outcome of *ɦʷ is troublesome.

  6. David Marjanović says

    Well, it’s *ɦʷ vs. *ɦʷˤ, and the *r at the end must have had an influence, too.

    Relatedly: an analysis of the Modern Standard Mandarin sound system postulating just two vowels and syllabic consonants.

  7. Greg Pandatshang says


    Along the lines of what David is saying, I believe the ˤ (pharyngealisation) as in *ʔˤaɦʷˤar is reconstructed specifically because it is a non-palatalising environment.

    Zev Handel summarises (https://www.academia.edu/2347515/2008_What_is_Sino-Tibetan_Snapshot_of_a_Field_and_a_Language_Family_in_Flux):

    Old Chinese is known to have had two distinct syllable types, one of which conditioned phonological developments generally characterizable as palatalization. The precise nature of the distinction remains uncertain. Among the possibilities that have been proposed are the presence or absence of a segmental feature, a prosodic distinction, a vowel length distinction, a phonation type distinction, and a laryngealization or pharyngealization distinction.

    So it is not surprising that *ɦʷ vs. *ɦʷˤ would have different results. One notes that the *d in *darɦʷa also became palatalised.

  8. While this paper is full of interesting stuff, it suffers from the “not yet proven” mental block that David M frequently moans about in linguistics. The field is in flux, as the title says, because in science πάντα ῥεῖ.

  9. A bit late, but a commenter at my blog explained that the word that sounds like “kadai” and means “or whatever” is ก็ได้ (romanized as, for example, “ko dai”):
    So, not a case of reduplication. The more you know.

  10. John Cowan says

    In addition, Tai-Kadai incorrectly suggests that the family is divided broadly into Tai and Kadai top-level branches, where “Kadai” is a name for a grouping of certain Kra languages with Hlai that is no longer accepted. Kra-Dai is now generally broken into the subfamilies Kra, Dai, and Kam-Sui, the two internal isolates Be and Hlai, and several unclassified ones.

    By a similar argument, some people are now calling Sino-Tibetan “Trans-Himalayan”, because they think Sinitic is part of Tibeto-Burman (though there is no agreement on where it might reside in the tree).

  11. Given also the nascent renaming of Altaic as “Transeurasian”, or the portmanteauing of Afro-Asiatic and Eskimo-Aleut into “Afrasian” and “Eskaleut” — I’m kind of wondering when will a proposal like this be (again) advanced also for Indo-European.

    Though I’m sure the fact that at least Europe is obviously a geographic area and not a subfamily helps. Similarly no one seems to care about unhyphenificating e.g. Niger-Congo either.

  12. Well, clearly it should be Tai-Kra-Dai-Kam-Sui-Be-Hlai-Et-Al.

  13. Or TKDKSBHEA for short.

  14. David Marjanović says

    单于 darga

    So for once the crackpot Zhu Xueyuan got it right!

    No idea who that is, but this paper says 单于 is this guy.

    and the *r at the end must have had an influence, too

    That was a wrong guess on my part. Syllable-medial *-r- does interesting things to vowels, but syllable-final *-r just became *-n in some dialects and *-j in others, creating some interesting doublets, according to Baxter & Sagart.

    the nascent renaming of Altaic as “Transeurasian”

    Specifically, “Transeurasian” is “Macro-Altaic” – and in her latest papers Martine Robbeets has quietly reintroduced “Altaic” for “Micro-Altaic” (Turkic, Mongolic, Tungusic as opposed to Koreo-Japonic), though not used it much.

  15. Ostapirat (2000)

    His motivation for choosing Kra for his group is different. Mostly simply because it is not the same group.

Speak Your Mind