A dialectal sound change in the Shi Jing: -it-s > -ut-s after labial consonant

[This post was revised March 12 and 15, 2020]

There are at least six examples in the Odes of OC *-it-s rhyming with *-ut-s (whether original OC *-ut-s, or *-ut-s from OC *-up-s). This happens only when *-it-s is preceded by a labial consonant. Most likely, a dialectal sound change has taken place, changing *-it-s to *-ut-s after a [+labial] consonant, and allowing words pronounced as *-it-s elsewhere to rhyme with *-ut-s:

*-it-s > *-ut-s /[+labial] __

The locations in the Odes where this rhyming can be seen are:

  • 58.5 *mi[t]-s with *sə-lu[t]-s,
  • 60.1 悸 *[g]ʷi[t]-s with *sə-lu[t]-s,
  • 65.2 穟 *s.[ɢ]ʷi[t]-s with 醉 *Cə.tsu[t]-s [added March 15]
  • 241.3 季 *kʷi[t]-s with 對 *[t]ˤ[u]p-s,
  • 241.4 季 *kʷi[t]-s with 類 *[r]u[t]-s,
  • 256.4 寐 *mi[t]-s with 內 *nˤ[u]p-s.

Presumably each of the twelve words above ended in [uts] in the pronunciation these odes were composed in, whence the particular rhyming pattern. Note that rhymes twice with *-ut-s in different stanzas of ode 241, each time with different words: this shows the phenomenon is not a fluke, but a dialectal characteristic.

This rhyming tends to make the OC vowel in *[t]ˤ[u]p-s and *nˤ[u]p-s more secure, as the alternative to *u in both cases is *ə, not *i. The brackets around [u] are no more needed.

Odes 241 and 256 are part of the Da Ya, a section believed to be associated with western Zhou and (north-)western China at the time. Odes 58, 60 and 65 are from the Guo Feng section (58 and 60 from 衛風, Hebei-Henan boundary;  65  from  王風 , We may be dealing with a northern/northwestern isogloss.

For the numbering of the Odes readers should refer to Appendix B in Baxter (1992), and to Mattis List’s Shijing Rhyme Browser (March 12, 2020: thanks, Mattis, for modifying  your browser to allow queries by OC rhyme! reader, to get the browser to list all Shi Jing occurrences of a certain OC rhyme, type the rhyme in B&S system, within double angle brackets, like this: «ək», in the query window).

A counterexample ?

There is in the entire Da Ya and Wei () Feng sections of the Shi Jing a single example of *-it-s after a labial NOT rhyming as *-ut-s: in 245.4 (Daya), ‘ear of grain’, OC *s.[ɢ]ʷi[t]-s (same word as , the older graph) rhymes with , MC bajH. This MC form could go back to either of *-at-s or (following a labial consonant) *-ots , but the fact that the phonetic series’s head-word is 巿 *put ‘knee covers’ suggests that the MC reading with -ajH is anomalous, even if from *-ot-s, and that Ode 245 read as *-ut-s.

Consequently there is no strong counterexample to the change *-it-s > *-ut-s /[+labial] __ in the Da Ya, Wei () Feng and Wang (王) Feng sections of the Odes, and we may regard this as a constant dialect feature of these sections of the Shi Jing.

Outside of the Daya, Wei () and Wang (王) Feng, the change does not occur at all. In the Guo Feng we have:

  • 30.3 *mi[t]-s with 嚏 *[t]ˤi[t]-s,
  • 53.1 *pi[t]-s with 四 *s.li[j]-s,
  • 110.2 *mi[t]-s and *kʷi[t]-s with *[kʰ]i[t]-s.

Odes 30, 53 and 110 belong, respectively, to Bei Feng 邶風 (Henan, between Zhengzhou and Anyang), Yong Feng 鄘風 (Henan, near Bei) and Wei Feng 魏風 (south Shanxi): a region on the middle course of the Yellow River valley.

In the Xiaoya there are no examples of the change, at all. Where *-it-s occurs after a labial consonant, it rhymes with *i words:

197.4 嘒 *qʷʰˤi[t] -s and 淠 *pʰˤi[t]-s and *mi[t]-s with 屆 *kˤr[i][t]-s,

222.2 qʷʰˤi[t] -s and 淠 *pʰˤi[t]-s with *s.li[j]-s and 屆 *kˤr[i][t]-s.

In these two instances, the presence of 屆 kˤr[i][t]-s and 駟 *s.li[j]-s in the rhyme sequences guarantees that 嘒* qʷʰˤi[t] -s, 淠 *pʰˤi[t]-s and 寐*mi[t]-s were pronounced with *i vowel.

[added Feb. 14, 2021: find a sequel to this post here]

Cite this post as: Laurent Sagart, "A dialectal sound change in the Shi Jing: -it-s > -ut-s after labial consonant," in Sino-Tibetan-Austronesian, 06/02/2020, https://stan.hypotheses.org/450.

Chinese-Tibetan vowel correspondences I

Hill (2019: §33, §224) reconstructs a vowel *e in the Sino-Tibetan parent language (Proto-Trans-Himalayan in his terminology; I will be using the term PCT ‘Proto-Chino-Tibetan’, meaning the language one would reconstruct if only Old Chinese an Tibetan were available). He argues that this vowel is preserved unchanged in OC, and that its reflexes in WT are conditioned by the ending: one has WT e before labials, i before velars, and a before dentals (including –r and –l). Below are examples of the correspondence OC *e to Written Tibetan a in words without a coda (1), with labial codas (2-3), velar codas (5-6), PCT *-q (7) and PCT *-ʔ (8-9)

OC *e : WT a

1. 溪 *kʰˤe > khej > xī ‘valley with stream in it’ : WT rka ‘small furrow conveying water from a conduit to trees and plants’
2. 占 *tem > tsyem > zhān ‘prognosticate’ : WT gtam ‘talk, discourse, speech’
3. 攝 *kə.n̥ep > syep > shè ‘catch, gather up’ : WT rnyab ‘seize or snatch together’
4. 貞 *treŋ > trjeng > zhēn ‘straight; correct’ : WT draŋ-po ‘straight’
5. 脛 *m-kʰˤeŋ-s > hengH > jìng ‘leg, shank’, 牼 *m-kʰˤeŋ > heang > kēng ‘shank bone’ : WT rkang ‘marrow, leg bones, stalk’
6. 清 *tsʰeŋ > tshjeng > qīng ‘clear (adj.)’, 淨 *m-tseŋ-s > dzjengH > jìng ‘cleanse (v.t.)’ : WT g-tsang-ba ‘to be clean, pure’
7. 舐 *Cə.leʔ > zyeX > shì ‘lick’ : WT ldag pa ‘to lick’
8. 庳 *N-peʔ > bjieX > bì ‘low, short’, 鼙 *[b]ˤe > bej > pí ‘small hand-drum’ : WT p’ra-ba ‘thin, fine, minute; little, small’
9. 諀 *pʰeʔ > phjieX > pǐ ‘slander’ : WT p’ra-ma ‘slander’

[added Oct 29, 2021]: two more examples: 葉  ‘leaf’ and  甲 ‘shell, buffcoat’ are discussed  here.

Similarly, the correspondence OC *e : Tibetan e is not limited to words with labial endings. Below are examples in words without a coda (1), ending in some kind of back-of-the-mouth fricative (2-3) and, in apparent violation of Dempsey’s law, in a velar (4-5):

OC *e : WT e

1. 佳 *[k]ˤre > kea > jia ‘good’ : WT dge ‘happiness, welfare, happy, propitious’
2. 秝 *[r]ˤek > lek > lì ‘successively, sequence’ : WT re ‘one at a time’, etc.
3. 譬 *pʰek-s > phjieH > pì ‘example’ : WT dpe ‘pattern, model; example, illustrative parable’
4. 積 *[ts]ek-s > tsjeH > jī ‘to hoard’ : WT rtseg-pa, pf. (b)rtsegs ‘to lay one thing on or over another, to pile up, stack up, build up’
5. 易 *lek-s > yeH > yì ‘easy’ : WT legs-pa ‘good, happy; neat, elegant, beautiful …’

These two correspondences of OC *e are not in complementary distribution.

Hill, Nathan W. 2019. The Historical Phonology of Tibetan, Burmese and Chinese. Cambridge: Cambridge University Press.

Cite this post as: Laurent Sagart, "Chinese-Tibetan vowel correspondences I," in Sino-Tibetan-Austronesian, 08/01/2020, https://stan.hypotheses.org/413.

Faulty etymologies in the STEDT II: ‘to love’

One of the weaker Sino-Tibetan comparisons in the STEDT can be found here:


The reconstructed form has detachable *ŋ- pre-initial (not a prefix), and a *w- onset marked as optional by means of parentheses.1 The possibility that forms with and without reflexes of these elements are unrelated is not given any consideration. In Matisoff’s Schrödinger-like version of comparative reconstruction, phonemes can be there, and not there.

At Sino-Tibetan level, ‘PTB *ŋ-(w)a:y’ is compared with Chinese 愛 ‘to love’, Middle Chinese  ‘ojH > modern ài.  The Chinese comparandum is vaunted as ‘excellent’ at the top of the STEDT page for this set.

愛 ài expresses a kind of feeling associated with the Confucian virtue of 仁 *niŋ > nyin > rén ‘kind(-ness), benevolence’.  Although its modern meaning is ‘to love’, and it is often translated by the verb ‘to love’, its more specific meaning in classical texts is  ‘to care for’,  as in this famous passage from the Confucian Analects 3.17/6/6, where in response to his disciple Zigong’s objection to the ritual sacrifice of a sheep, Confucius is quoted as saying “爾愛其羊,我愛其禮” “You love (=care for) the sheep, I love (=care  for) the rite”. Such is probably the original meaning.  The OC word for ‘to love’ was 字 *mə-dzə(ʔ)-s, correctly compared by Benedict to his PTB *m-dza ‘to love’.  In all Chinese dialects except (to my knowledge) the very archaic Waxiang and Caijia dialects, this word has been displaced by 愛 ài. ‘Love’ semantics, in the sense of a warm feeling one experiences for an individual person, are not old with 愛 ài.

A  minimal version of the comparison in STEDT had appeared in Benedict’s Conspectus (1972:192, fn.491): it related Proto-Karen *ʔai ‘love’ with 愛 ài. This would  be phonologically viable if the OC root ended in *-əj. In general, MC words with grave initials and rhyme -ojH (with ‘H’ for tone C, a.k.a qusheng) can go back to OC *-əj(ʔ)-s, *-ə(ʔ)-s, *-ək-s, *-ət-s or *-əp-s. Of these, only *-əj(ʔ)-s could potentially match forms in -a(:)j  outside of Sinitic. However, Baxter (1992) reconstructed *ʔɨt-s, with root-final *-t. Still, Zev J. Handel maintains at the bottom of the STEDT page for the item under review that *-j-s remains possible because OC rhyming does not provide direct evidence of contacts with *-t. While this is true, one should keep in mind that 愛 ài only rhymes once in the Shi Jing, and 僾 ài  ‘to pant, lose the breath’, whose phonetic is 愛 ài, also once: the opportunities for an unsuffixed final stop to surface in rhyming are thus quite limited. Absence of *-t (or *-p, for that matter) among the rhyme contacts of 愛 ài or 僾 ài  is not in itself evidence for a root ending *-j.

Word-families provides strong evidence for excluding *-j-s in 僾 ài ‘to pant, lose breath’, just cited: this word is the s-suffixed derivative of 唈 *qˤ[ə]p > ‘op > yì ‘short of breath’.  This is confirmed by the fact that  僾 ài  ‘to pant, lose the breath’ rhymes with 逮 modern dài < OC *m-rˤəp-s  ‘reach to’  in Ode 257. 逮 dài itself is the s-suffixed derivative of 眔 *m-rˤəp > dop > tà ‘reach to’.  There can be no question that 僾 ài and 逮 dài ended in *-p-s.

Baxter and Sagart (2014) accordingly reconstructed 僾 ài  as *qˤəp-s. This in turn cannot be reconciled with the view that 愛 ài ended in *-j-s: how could a word ending in *-j-s  be chosen as a phonetic for another ending in *-p-s ? evidently the root in 愛 ài  also ended in a stop. But which ?

While 愛 ài  ‘to care for’ was reconstructed with *-t-s in Baxter (1992), the OC contrast between *-p-s and *-t-s was lost at a late stage of OC: if the rhyming of 愛 ài   and 謂 *[ɢ]ʷə[t]-s > hjw+jH > wèi ‘say, tell, call’ in Ode 228 is based on a pronunciation in which the merger had taken place, 愛 ài may have had *-p-s earlier on, like 僾 ài. Accordingly Baxter and Sagart (2014) reconstructed 愛 ài as *[q]ˤə[p]-s, with square brackets around -[p] to convey the uncertainty between *-p and *-t.

Shortly after the publication of their book, and independently from it, Norquest (2015) arrived at the Proto-Hlai reconstruction *ʔə:p ‘to love’. He did not notice any connection to Chinese 愛 ài . Norquest’s *ʔə:p is not found in the other Kra-Dai branches,  in Austronesian or Austroasiatic. It is probably a Chinese loanword: this leaves no doubt that the Chinese word’s root ending was *-p.

Chinese root-final *-p is not reflected anywhere in the rest of Matisoff’s set: the likelihood that the Chinese form is related to the rest is nonexistent. The superficial resemblance between the Chinese and Karen forms is the fruit of convergence.


Baxter, W. H.  1992. A Handbook of Old Chinese phonology. Trends in Linguistics Studies and Monographs 64. Berlin: Mouton de Gruyter.

Baxter, William H. and Laurent Sagart. 2014. Old Chinese: a new reconstruction. New York: Oxford University Press.

Benedict, P.K. 1972. Sino-Tibetan: a Conspectus. Cambridge: University Printing House.

Norquest, Peter. 2015. A Phonological Reconstruction of Proto-Hlai. Brill.

Cite this post as: Laurent Sagart, "Faulty etymologies in the STEDT II: ‘to love’," in Sino-Tibetan-Austronesian, 21/07/2019, https://stan.hypotheses.org/365.

  1. There is nothing wrong in itself with parentheses. Baxter and Sagart (2014) use parentheses in their OC reconstruction to signal phonemes whose presence or absence in a protoform are both compatible with all the comparative evidence at hand: for instance they reconstruct medial-r- between parentheses in 夫 *p(r)a > pju > fū ‘man’  because given Middle Chinese pju it is impossible to decide whether Old Chinese had -r- or not. Matisoff’s parenthesized phonemes are different: they are used to make manifest Matisoff’s decision to overlook a phoneme in making cognate decisions. []

Conflated cognate sets in the STEDT I — ‘excrement’

I have elsewhere (Sagart 2006) pointed out that sound correspondences are a minor part of the grounds on which the cognate sets in Matisoff’s 2003 book were assembled, despite Matisoff’s anger (Matisoff 2007) when the point was made in print. The same applies to the cognate sets in STEDT, which are a more evolved state of those in the book. Granted, to some extent relying on educated guesses is unavoidable in cognatising a large number of related languages: however, the sound correspondences in Matisoff’s book, which are his most recent statement on ST phonological history, are not sufficient to distinguish between phonetically and semantically resemblant etyma in ST.

Consider the STEDT cognate set #572 *kləy ‘excrement’ (here). In several of the languages where this putative etymon is reflected, doublets appear. The most prominent type includes one form with a liquid in its onset (column I) and another without (column II). See table 1.

Kanauri (Sharma)s kli ‘urine’khə ‘shit’
Central Tsangla
le ‘intestines’khi chung ‘buttock’
Bodo (Bhat)bi klə́ ‘liver’bi kí ‘excrement of fish’
Mikir (Grüssner)mék-krí ‘tears’hī ‘feces / shit’
khrə¹¹ ‘body dirt’ci¹¹  ‘body dirt’
Newar (Genetti) ʈi (< kr-) ‘(ear)wax’khi  ‘shit’
Sunwar (Michailovsky)khriː ‘feces’kiː  ‘intestines’

Table 1: doublets in the STEDT cognate set #572 (accessed mid-may, 2019)

The differences in onsets and rhymes between the forms in the two columns are not explained anywhere in Matisoff (2003) or STEDT. They are not due to identified morphological processes applying on a single lexical root. We are dealing with two etyma: the first has an onset cluster, the second does not. Semantically the forms in the first column have more associations with ‘body dirt’ and those in the second column with ‘excrement’.

Taraon klɑi53, Proto-Northern Naga *C̥-kləy (French 1983), Written Tibetan lci < hlyi, Lepcha tə kli probably belong to the first etymon. Japhug Gyarong tɯ qe ‘excrement’ (Jacques 2015) probably belongs to the second: the Japhug onset cannot originate in a Cl-type cluster (Jacques, p.c.).

With the Chinese word 屎 MC syijX ‘excrement’, one source of MC sy- is OC *l̥-: if OC *l̥- itself can remount to PST *kl-, one seems to have a match to #kləy. However the Proto-Min initial for 屎 is *š-, which excludes OC *l̥-. An OC *l̥- would evolve to Proto-Min *tšh- (Baxter and Sagart 2014:93). Consequently, the Chinese word probably does not belong with the first etymon either. Baxter and Sagart reconstruct 屎 tentatively with a uvular initial, OC *[qʰ]ijʔ. None of the other sources of MC sy-: OC *s.t-, *n̥-, *ŋ̊-, *l̥-, are plausible in this word. Thus 屎 OC *[qʰ]ijʔ probably belongs to the second etymon, like Japhug Gyarong tɯ qe.

The first etymon, with the lateral cluster and ‘body dirt’ semantics, is best compared with Chinese 尸 *l̥əj ‘corpse’, via the notion of ‘carrion’.

The second etymon can be compared with Proto-Austronesian (Blust) *Caqi ‘excrement’. There are reasons within Austronesian to think that this word ended in a laryngeal or back-of-the-mouth fricative (here), although I am not sure anymore of the identity of that phoneme.

In my next post, I will discuss the STEDT *etymon #601 *m/s-tuːk ‘to spit’.


Baxter, William H. and Laurent Sagart. 2014. Old Chinese: a new reconstruction. New York: Oxford University Press.

French, Walter Thomas. 1983. Northern Naga: A Tibeto-Burman mesolanguage. New York, Univ., Diss. Ann Arbor : University Microfilms

Jacques, Guillaume. 2015. Dictionnaire Japhug-chinois-français 嘉绒-汉-法词典 Version 1.0.

Matisoff, J. A. 2003. Handbook of Proto-Tibeto-Burman. Berkeley, Los Angeles, London: University of California Press.

Matisoff, J. A. 2007. Response to Laurent Sagart’s review of Handbook of Proto-Tibeto-Burman: System and philosophy of Sino-Tibetan reconstruction. Diachronica 24,2: 435–444.

Sagart, L. 2006. Review: James A. Matisoff (2003) Handbook of Proto-Tibeto-Burman. System and philosophy of Sino-Tibeto-Burman Reconstruction. Diachronica 23,1: 206-223.

The origins of Kra-Dai tones, current status (I): tone C

October 8, 2019:  a paper on the origins of Kra-Dai tones is now  published (here). It represents my current understanding of the formation of Kra-Dai tones and supersedes the present three posts. There are more data in the paper, especially on the Austronesian final consonant I note as *X; there are also some significant differences between the present posts and the paper.  Thus, although it is clear that the word ‘excrement’, ususally reconstructed as *Caqi, ended in some kind of guttural fricative, it contains too many irregularities, possibly euphemistically motivated (as noted by Wolff), both in Austronesian and Kra-Dai, to identify that fricative. Therefore in the published paper I retreat from my suggestion that *Caqi ‘excrement’ was really *Caqix. I suggest instead that final *-x was the value of Tsuchida’s *-H2, thus converging with Blust’s proposal, but on completely different grounds (pages 7-9).

Like Old Vietnamese, Written Burmese and Proto-Hmong-Mien, the Kra-Dai languages exhibit a phonological typology strongly influenced by a form of Chinese which is later than Old Chinese and earlier than Early Middle Chinese: that is, a period extending beginning c. 200 BCE and covering the first half of the first millennium CE. Their tone system is structurally identical to that of Chinese of that time: three contrasting tones on words ending in sonorants (vowels, semi-vowels, nasals, liquids) while words ending in oral stops do not show any tonal contrast. The origin of the Vietnamese tone system was elucidated by Haudricourt (1954a): one tone originated in words ending in [h], another in [ʔ], a third in sonorants. Haudricourt (1954b) showed that the Chinese qùshēng 去聲 originated in an *-s suffix which evolved to [h] prior to becoming a tone. Mei (1970) completed Haudricourt’s picture by proposing that the Chinese shǎngshēng 上聲 goes back to words ending in a glottal stop. For details see Sagart (1999). The origins of the Burmese, Hmong-Mien and Kra-Dai tones have not been elucidated although some progress has been made for Kra-Dai by Ostapirat (2005) and Norquest (2013). Here I describe additional hypotheses which bring us closer to a solution.

The Kra-Dai tones are known as A, B, C and D. The core of these categories existed in Proto-Kra-Dai (Ostapirat 2000), though not necessarily as true phonetic tones having pitch as their main acoustic clue. When borrowed into Kra-Dai, Middle Chinese píngshēng 平聲 words (ending in sonorants in Old Chinese) typically have tone A; Middle Chinese shǎngshēng 上聲 words (ending in a glottal stop in Old Chinese) are treated by tone C; MC qùshēng 去聲 words (ending in *-h in late OC < OC *-s) show tone B. Chinese words ending in oral stops form the so-called rùshēng入聲 category. They do not contrast with the other Chinese tones. Kra-Dai ‘tone’ D is the corresponding category in Kra-Dai. Gedney (1978) suggests that the phonetic values of the Thai tones A, B and C are the same as the Old Chinese values of the Chinese píngshēng, qùshēng and shǎngshēng respectively, under the Haudricourt-Mei solution: sonorant endings, [h]-endings, [ʔ]-endings. Earlier attempts to solve the origin of Kra-Dai tones have assumed Gedney’s hypothesis.

Sagart (2004, 2005, 2008, 2011) argued that the Kra-Dai languages are a subgroup, not a sister, of the Austronesian language family. Here I attempt an account of the Kra-Dai tone categories from the viewpoint of Austronesian consonant endings. It should be borne in mind that the AN words in Kra-Dai, while very basic, are quantitatively limited: for that reason my proposals are based on small numbers of examples. This post deals with the Kra-Dai tone C. Two other posts deal with tones B (here) and A (here).

In examining the Kra-Dai tone categories in connexion with Austronesian, reference must be made to Tsuchida’s reconstruction of two final consonants (Tsuchida 1976): *H1 (-h in Pazeh, Saisiyat, Atayal, Sediq, Amis and Aklanon) and *H2 (-h in Takituduh and Aklanon). AN words ending in *H1 form the core of the Kra-Dai tone C:

PAN(Sa) Buyang (Li 1999) PHlai(Os) PTai(Pi)
to come, go’ *uwaH1 va11 < C2
head’ *quluH1 qa0 ðu11 < C2 *uRəu C *kraw C
shoot, outgrowth, flower’ *buŋaH1 ma0 ŋa11 < C2

PAN(Sa): Proto-Austronesian (Sagart, this post); PHlai(O): Proto-Hlai (Ostapirat 2004); PTai(Pi): Proto-Tai (Pittayaporn 2009)

Supporting evidence for these reconstructions:

*uwaH1 ‘to come, to go’: Atayal uah ‘come’, Mantauran Rukai oa ‘go’, Babuza m-oa ‘come’, Puyuma ua ‘go’;

*quluH1 ‘head’ : Paiwan qulu, Sediq (Taroko) qolox ‘skull’, Saisiyat ta-ʔœlœh, Aklanon úeo(h);

*buŋaH1 ‘shoot, outgrowth, flower’: Saisiyat poŋaeh ‘flower’ (p- unexplained; expect b-), Saaroa vuŋavuŋa ‘ear of foxtail’ (own fieldwork, 2014), Kanakanabu buŋabuŋa ‘flower’, Aklanon búːŋah ‘fruit’.

Some Kra-Dai C-category words cannot come from *H1. The word for ‘excrement’, Proto-Kra *kai C, Proto-Hlai *akaːi C, Proto-Tai *C̬.qɯj C, in all likelihood corresponds to a PAN word reconstructed as *C1aq3i (Tsuchida), Blust *Caqi (Blust), *taqí (Wolff). Under the generally accepted sound correspondences, these should yield sai in Pazeh and Kaxabu, two closely related languages of central-western Taiwan. If the word had ended in *-H1, saih would be expected in both. The attested forms are Pazeh saik, Kaxabu saix (Tun 2015). Recently, thanks to my friends Prof. Hsing Yue-ie and Hsu Tze-fu of the Institute of Microbial and Plant Biology, Academia Sinica, I was able to visit Mr Tun in Puli and to record his pronunciation of this word (here). Blust treats Pazeh saik as a metathesis, but that explanation cannot work for Kaxabu. Currently I do not know any other examples of the correspondence Pazeh –k to Kaxabu –x. One possibility is that the word ended in a rare consonantal ending PAN *-x, preserved unchanged in Kaxabu, and merged with *-k in Pazeh. The aberrant –l ending in Kavalan tal ‘excrement’ would be its indirect reflection, as would tone C in Kra-Dai. This is speculative, as no other example of the final-consonant correspondence shown by ‘excrement’ in Austronesian is currently known.

In two other posts I examine the origins of the Kra-Dai tone B (here). and A (here), B and C. A list of references is appended to the post on tone A.

Cite this post as: Laurent Sagart, "The origins of Kra-Dai tones, current status (I): tone C," in Sino-Tibetan-Austronesian, 20/11/2017, https://stan.hypotheses.org/123.

The origins of Kra-Dai tones, current status (II): tone B

October 8, 2019:  a paper on the origins of Kra-Dai tones is now  published (here). It represents my current understanding of the formation of Kra-Dai tones and supersedes the present three posts. There are more data in the paper, especially on the Austronesian final consonant I note as *X; there are also some significant differences between the present posts and the paper.  Thus, although it is clear that the word ‘excrement’, ususally reconstructed as *Caqi, ended in some kind of guttural fricative, it contains too many irregularities, possibly euphemistically motivated (as noted by Wolff), both in Austronesian and Kra-Dai, to identify that fricative. Therefore in the published paper I retreat from my suggestion that *Caqi ‘excrement’ was really *Caqix. I suggest instead that final *-x was the value of Tsuchida’s *-H2, thus converging with Blust’s proposal, but on completely different grounds (pages 7-9).

In my last post (here) I argued that the main origin of the Kra-Dai tone C category is Tsuchida’s * H1. Here I examine the origins of the Kra-Dai tone B. I will again argue that this tone category can be explain on the basis of observable Austronesian facts, although one of the sources I will propose has not so far been reconstructed.

It is becoming increasingly clear that PAN word-final *-R, a voiced uvular fricative [ʁ], is one of the sources of the Proto-Kra-Dai tone B. This was first pointed out by Norquest (2013), where the first two comparisons in the below table were noted.

PAN PMP(Sa) PHlai(No) Buyang (Li) PTai(Pi)
*tebuR ‘natural spring’ *tebuR —— —— *ɓoː B
*SulaR ‘snale’ —— *ljaːɦ —— ­——
*-kaR ‘dry’ *-kaR *kʰɯːɦ qha11< B *χaɰ B
—— *qi(d)zuR ‘saliva’ —— qa0 tu11 < B ‘sputum’ ——

Unaware of Norquest’s earlier observations, I noted the match between tone B and final *-R in ‘dry’ and ‘saliva/sputum’ while working on the Kra language Buyang in September-early October 2017. I became aware of Norquest’s proposal through his message of October 22, 2017 on an email discussion group strated by Cecil Brown, to which he had attached his 2013 paper in Mon-Khmer Studies.

Supporting material for root *-kaR ‘dry’: Siraya maskag ‘dry’, Puyuma (Nanwang) taŋkar ‘dry’ (fields), baʈəkar ‘dry, shallow’; Ilianen Manobo, Western Bukidnon Manobo kagkag ‘to dry’, Casiguran Dumagat kag’kag ‘to dry’, Sindangan Subanun mɨgɨngkag ‘to dry’, Siocon Subanon kumongkag ‘to dry’.

The Kra-Dai tone-B category also includes words with PAN or PMP cognates not ending in *-R:

Gloss AN Pre-PMP(Sa) Buyang (Li Jinfang 1999) P-Hlai (No) PTai (Pi) Aklanon
bran, chaff PMP(B) *qepah *qepaX ta0 pha11 < B —— —— opa(h)
knee PAN(W) *puqu *puquX qhu11 < B —— *χow B ——
shoulder PAN(B) *qabaRa *qabaRaX qa0 ʔba11 < B *C-ʋaːɦ *C̥.ba: B abága(h)
palm of hand PAN(B) *dapa *dapaX pa11 < B —— —— ——

It is noteworthy that Aklanon (Panay island, Philippines), already identified in Tsuchida 1976 as the only language outside of Taiwan with segmental reflexes (-h in both cases) of his *-H1 and *-H2, has –h for KD tone B. As argued in my preceding post, PAN *-H1 is the main source of the KD tone C; and as I will argue in my next post, *-H2 is one of the sources of the KD tone A. We must be dealing here with a final consonant distinct from both *-H1 and *-H2. This must be a consonant that (a) merges with *-R in PKD, and (b) evolves to -h in Aklanon. I propose the uvular fricative *-X [χ], with these reconstructions: ‘bran, chaff’ *qepaX; ‘knee’ *puquX; ‘shoulder *qabaRaX; ‘palm of hand’ *dapaX. By my proposal, the Kra-Dai B tone category originates in he merger of two AN uvular fricative endings, voiced and voiceless.

Here are some additional notes on the words in Table X:

Bran, chaff: the MP word has a probable Formosan cognate in Amis ʔəpah ‘wine’—bran can be an ingredient in millet or rice wine. Blust reconstructed PMP *-h based on Aklanon. In the current state of his ACD, final PAN/PMP *-h is similar to Tsuchida’s *-H1 (he has recently begun writing PAN * x for Tsuchida’s *-H2). Tsuchida’s *-H1 is reflected as -h in Amis. Amis ʔəpah ‘wine’ at first sight comforts his reconstruction of PMP final *-h: however my proposal is that PAN *-X is reflected as -h in both Amis and Aklanon, just like *-H1 (Pazeh, Kaxabu, Saisiyat, Atayal and Sediq distinguish them, though).

Shoulder is reconstructed by Blust as PAN *qabaRa. Unaware of final –h in Aklanon abága(h), Ostapirat (2005) and Norquest (2013) ascribe tone B in ‘shoulder’ to medial *-R-. For this they need a mechanism which allows the riming part of an AN word (last vowel plus any following consonant) to be lost in Kra-Dai, here: *qabaRa > qabaR. They do not explain what the conditioning is. This mechanism is overpowerful: it doubles the number of potential KD cognates for each AN word. Benedict (1975) put it to maximal use. Yet the quality of comparisons of this type tends to be low. Final –h in the Aklanon cognates of KD tone-B words like ‘shoulder’ supports a simpler and more constrained explanation: KD tone B reflects final *-X, not medial *-R-; in the evolution to KD, *-R- was lost when the -bR- cluster resulting from penultimate vowel syncope was reduced to -b-. By the correspondence rules described above one would expect final –h in the Amis word for ‘shoulder’, but the attested Amis form is afala. Lack of -h is unexplained, unless there has been dissimilation between the two uvular fricatives in the last syllable of *qabaRaX.

Palm of hand: *dapa ‘palm of hand’ was assigned to PMP by Blust until c. 2015, when he raised it to PAN following Sagart’s identification of an Atayal cognate.

Knee: reconstructed by Wolff (2010) as PAN *puqu on the basis of northern and western Formosan forms (Saisiyat, Favorlang, Thao), an impeccable PAN pedigree.

Cite this post as: Laurent Sagart, "The origins of Kra-Dai tones, current status (II): tone B," in Sino-Tibetan-Austronesian, 20/11/2017, https://stan.hypotheses.org/120.

The origins of Kra-Dai tones, current status III: tone A

October 8, 2019:  a paper on the origins of Kra-Dai tones is now  published (here). It represents my current understanding of the formation of Kra-Dai tones and supersedes the present three posts. There are more data in the paper, especially on the Austronesian final consonant I note as *X; there are also some significant differences between the present posts and the paper.  Thus, although it is clear that the word ‘excrement’, ususally reconstructed as *Caqi, ended in some kind of guttural fricative, it contains too many irregularities, possibly euphemistically motivated (as noted by Wolff), both in Austronesian and Kra-Dai, to identify that fricative. Therefore in the published paper I retreat from my suggestion that *Caqi ‘excrement’ was really *Caqix. I suggest instead that final *-x was the value of Tsuchida’s *-H2, thus converging with Blust’s proposal, but on completely different grounds (pages 7-9).

In two recent posts I discussed the origins of the Kra-Dai tones C (here) and B (here). Here I discuss tone A.

Kra-Dai A-category words correspond primarily to Austronesian words ending in sonorants (vowels, semi-vowels, nasals, liquids). This was already pointed out by Ostapirat (2005). Examples:


Buyang (Li 1999)






ma0 ta54 < A1

*ata A

*p.taː A




ma0 ðu312 < A2

*aRu A





tu312 < A2

*ʔtu A




bear (n.)

ta0312 < A2

*mui A

*ʰmwɯj A



to kill

ma054 < A1






pui54 < A1

*api A

*wɤj A




ʔdaːn31 < A1°

*ɲa:n A

*ɓlɯən A




nam54 < A1

*(ə)num A




to eat

ka:n54 < A1


*kɯɲ A



Note that the numerals in the Tai branch are Chinese loanwords.

A second source of Kra-Dai category-A words is in PAN *-H2 (see: LINK for Tsuchida’s *-H2)


Buyang (Li 1999)






qa0 tu54 < A1

*utu A

*traw A



ear ta0 ða312 < A2 *ilai A *krwɯː A *CaliŋaH2 or *CaŋilaH2 *taliŋah
three tu54 < A1

*utu C !


*tu-tuluH2 *tu-tuluh

Tone C in the Hlai word for ‘three’ may be the result of list analogy: it is part of a series of four consecutive numerals in tone C: *cɨ C ‘one’, *alau C ‘two’, *utu C ‘three’, *atəu C ‘four’.

To recapitulate, these sources are found for the Kra-Dai tone categories:

Tone A (this post): AN words ending in sonorants (vowels, semi-vowels, liquids, nasals); PAN words ending in *H2.

Tone B (here): PAN words ending in *-R [ʁ]; other PAN-related B-category words are believed to have had *-X [χ].

Tone C (here): PAN words ending in *H1 [h]; at least one other source, possibly *-x (‘excrement’).

The probable phonetic values of these categories in Proto-Kra-Dai wer:

Category A: ending in vowels, semi-vowels, liquids, nasals;

Category B: ending *-X;

Category C: ending in *-h.

By the time of contatct with Chinese, beginning c. 200 BCE, final *-h had changed to a glottal stop.


Benedict, Paul K. 1975. Austro-Thai: language and culture. New Haven: HRAF Press.

Gedney, William J. 1978. Speculations on early Tai tones. Paper presented at the 11th ICSTLL, U. of Arizona, Oct. 1978.

Haudricourt, André-Georges. 1954a. De l’origine des tons du vietnamien. Journal Asiatique 1954, 242: 69-82.

Haudricourt, André-Georges. 1954b. Comment reconstruire le chinois archaique. Word 10, 2-3: 351-64.

Li Jinfang. 1999. Buyang Yu Yanjiu. Beijing: Zhongyang Minzu Daxue.

Mei Tsu-lin. 1970. Tones and prosody in Middle Chinese and the origin of the Rising tone. Harvard Journal of Asian Studies 30, 86-110.

Norquest, Peter K. 2013. A revised inventory of Proto Austronesian consonants: Kra-Dai and Austroasiatic Evidence. Mon-Khmer Studies 42: 102-126.

Ostapirat, Weera. 2000. Proto-Kra. Linguistics of the Tibeto-Burman Area 23.1.

Ostapirat, W. 2004. Proto-Hlai sound system and lexicons. in: Ying-ching Lin, Fang-min Hsu, Chun-chih Lee, Jackson T.S. Sun, Hsiu-fang Yang and Dah-an Ho (eds), Studies in Sino-Tibetan languages, papers in honor of Professor Hwang-cherng Gong on his seventieth birthday, 121-175. Nankang: Institute of Linguistics, Academia Sinica.

Ostapirat, Weera. 2005. Kra-Dai and Austronesian: notes on phonological correspondences and vocabulary distribution. In: L. Sagart, R. Blench and A. Sanchez-Mazas (eds) The peopling of East Asia: Putting together Archaeology, Linguistics and Genetics, 109-133. London: RoutledgeCurzon.

Pittayaporn, Pittayawat. 2009. The phonology of proto-Tai. Unpublished Cornell U. dissertation.

Sagart, Laurent. 2004. The higher phylogeny of Austronesian and the position of Tai-Kadai. Oceanic Linguistics 43,2: 411-444.

Sagart, L. 2005. Tai-Kadai as a subgroup of Austronesian. In L. Sagart, R. Blench and A. Sanchez-Mazas (eds) The peopling of East Asia: Putting together Archaeology, Linguistics and Genetics 177-181. London: RoutledgeCurzon.

Sagart, Laurent. 2008. the expansion of setaria farmers in east asia: a linguistic and archaeological model. In Sanchez-Mazas A, Blench R, Ross M, Peiros I, Lin M, eds. Past human migrations in East Asia: matching archaeology, linguistics and genetics, 133-157. Routledge Studies in the Early History of Asia, London: Routledge.

Tsuchida, Shigeru. 1976. Reconstruction of Proto-Tsouic phonology. Tokyo: Study of languages and cultures of Asia and Africa monograph series N° 5.

Wolff, John U. 2010. Proto-Austronesian phonology with glossary. 2 vols. Ithaca: Cornell Southeast Asia program publications.

Cite this post as: Laurent Sagart, "The origins of Kra-Dai tones, current status III: tone A," in Sino-Tibetan-Austronesian, 20/11/2017, https://stan.hypotheses.org/117.

Middle Chinese y- opposite Written Tibetan g-

Baxter and Sagart (2014) reconstruct three uvular initial consonants in OC, *q-, *qʰ- and *ɢ-. Nonpharyngealized *ɢ- evolves to MC y-. Meanwhile, in Written Tibetan, OC uvular stops generally correspond to velar stops. Comparisons matching Tibetan initial *g- and MC y- should therefore exist. They are counter-intuitive to practitioners of ST comparison; to my knowledge, none has been presented before. Here are two:

1. “small of the back”

yín *[ɢ](r)ə[r] small of the back : WT sgal-pa “back of man, back of beast of burden, small of the back”

As part of a set of cognate words including the verb ‘gel/bkal/dgal/khol “to load, lay on a burden” and khal “burden, load”, WT sgal-pa “small of the back’ has been compared to *[g]ˤajʔ > haX > hè “carry” by Gong (System of Finals in Proto-Sino-Tibetan #165). A comparison between two words meaning “small of the back” is more specific than one between two words meaning “to carry”. The more specific comparison should be preferred. The small of the back—the narrower part of the back, in the lumbar region—is where pack animals are made to carry burdens. This implies that the Tibetan word-family (“to load/burden/small of the back”) is built around the body-part term. The nature of the s- element at the beginning of the word is uncertain.

2. “to pass”

The character has several pronunciations and meanings. In the pronunciation MC yen it writes a word meaning “pass, go beyond”, for which a WT comparison presents itself:

xiàn*[ɢ]a[n] “pass, go beyond” : WT rgal-pa to step, pass or climb over; to ford

The WT verb rgal/brgal/brgal/rgol has also been compared to *[C.g]ˤaj > ha > hé “river, especially the Yellow River”. This comparison is semantically not compelling since, especially where it crosses the early Chinese territory, the Yellow River certainly cannot be crossed on foot, in any season. This comparison seems to have been proposed primarily because of the phonological parallel it provides with the comparison for “small of the back”: i.e. OC gal(x) (in Gong’s system) to WT /gal/, with both Chinese words written by means of the phonetic .

In B&S reconstruction, in both comparisons, initial *[ɢ] is ambiguous for *N.q and *ɢ and the presence of medial *r cannot be excluded—although we omitted any explicit mention of it in our reconstruction *[ɢ]a[n]. Final *[n] is ambiguous for *-r. The vowel correspondences OC *a : WT a and OC * ə : WT a are regular, and so are the correspondences if codas, assuming final *[n] can be disambiguated to *-r in “pass, go beyond”.

Tibetan rtswa ‘grass’

In this paper, Guillaume Jacques proposes that the Old Tibetan semi-vowel –w– as part of a word onset is secondary, and that it has its origin in words ending in -u followed by -ba: he supposes the evolution Cu-ba > Cuwa > Cwa, for instance zwa ‘nettle’ < zu-ba, rwa ‘horn’ < ru-ba, grwa ‘corner’ < gru-ba. Another example is ‘grass’, OT rtswa, which must then come from an earlier rtsu-ba. Tibetan rtswa is compared to Chinese *[tsʰ]ˤuʔ > tshawX > cǎo ‘grass, plants’ by Matisoff (here, p. 177) under a reconstruction PST *r-tswa-n, as part of a list of mostly spurious comparisons. In the case of ‘grass’, Matisoff got lucky, but only Jacques’s proposal makes sense of the phonology of this comparison, since OT -a does not otherwise correspond to OC -u.

Cite this post as: Laurent Sagart, "Tibetan rtswa ‘grass’," in Sino-Tibetan-Austronesian, 21/06/2017, https://stan.hypotheses.org/60.

Tibetan sgro ‘large feather’

Written Tibetan (WT)  སྒྲོ sgro means ‘large feather, tail feather, quill’, as in སྒྲོ་མདོངས sgro-mdongs ‘peacock’s feather, as a badge of dignity’ (Jäschke). Benedict (1972) compared this word with Jingpo shagrau [ʃă31kʒau31] ‘outer skin’, apparently taking outer skin (of fruits etc.) and bird feathers to be different kinds of organic coverings. Coblin (1986) preferred comparing WT sgro with OC 羽 ‘feather’, OC *gwjagx in the reconstruction of Li. Coblin’s proposal seems to provide better semantics than Benedict’s: for this reason it has enjoyed broader support. A third interpretation is proposed below.

WT sgro སྒྲོ is a verb, meaning ‘to elevate, exalt, increase, exaggerate’. It is a homophone of the word for ‘large feather’.  Compared with normal featers, tail feathers are increased in size, strikingly so in the case of the pheasant or peacock: a derivation out of the verb is likely. Probable external cognates are Jingpo shagrau [ʃă31kʒau33] ‘to praise, extol’ and OC *[N-k](r)aw > gjew > qiáo ‘lift, elevated, high’. In the phonetic series of , one finds *[k](r)aw > kjew > jiāo ‘kind of pheasant’, said by 郭璞 Guō Pú (276-324 CE) to have a long tail and whose feathers were used as ornaments.

In yet another meaning, སྒྲོ sgro designates the bark of a species of willow; this, at least, is a good match for Benedict’s shagrau [ʃă31kʒau31] ‘outer skin’. There is no clear Chinese cognate.

Despite superficial resemblance, WT གྲོ་བ gro-ba or གྲོ་ག gro-ga ‘the thin bark of the birch tree’ is distinct from the preceding: it is reduced from grog-ba (Guillaume Jacques, p.c. 2016).





increase, elevate

སྒྲོ sgro ‘elevate, exalt, increase, exaggerate’

སྒྲོ sgro ‘large feather, tail feather’

*[N-k](r)aw‘lift, elevated, high’

*[k](r)aw ‘kind of pheasant’

ʃă31kʒau33 ‘to praise, extol’

outer skin, bark

སྒྲོ sgro



Cite this post as: Laurent Sagart, "Tibetan sgro ‘large feather’," in Sino-Tibetan-Austronesian, 19/06/2017, https://stan.hypotheses.org/45.

[July 10, 2021: see the sequel to this post and dicussion here]

Search OpenEdition Search

You will be redirected to OpenEdition Search