Out of curiosity, I was trying to find the top half of 益 (from my previous post) in Unicode. Those five strokes resemble this Khitan small script character:


(7.31.1:25: Is the resemblance a coincidence? Could <i> be derived from Khitan period Chinese 益 *ʔi?)

The closest characters I could find in Unicode were

𠔅 𠔆 𠔇

The first two* look like 八 'eight' atop 尸 'corpse'* and are variants of 歹 'bone fragment'**. More 歹-variants here including one that looks like 上 'top' atop 夕 'evening' and another (䚟) that looks like 角 'horn' (which may have bone cores - semantic?) plus 㝵 'to get' (phonetic).

I can't find the meaning or reading of the third. I've never even seen the element below 八 before. 𠔇 isn't in any of the entries of the Taiwanese variant dictionary, but according to the Unihan Database, it's graph 3 on p. 242 of vol. 1 of 漢语大字典 which I don't have. Could someone check it for me? Thanks! (I often miss library access. There's only so much I can do from my secret hideout in the middle of nowhere.)

This "online 漢语大字典" at zd.eywedu.com doesn't have 𠔇 (and its search function doesn't even work in Chrome or Firefox or MSIE or Safari - good luck with Opera). What does zd.eywedu mean? zd might be 字典, but what's the rest?

7.31.1:41: Wiktionary lists the Vietnamese (!) readings of 𠔇 as sước and xước. But I still have no idea what the Chinese readings of 𠔇 are, though they must be similar to the Vietnamese readings (assuming the latter are correct). I can't find 𠔇 (or any graphs pronounced sước or xước) at nomfoundation.org.

7.31.1:49: I forgot about zdic.net, where I've found graphs not listed elsewhere. zdic.net equates 𠔇 with 與 Md yu which has multiple readings (with different tones) and meanings. I don't know if the two are fully interchangeable. zdic.net says 𠔇 is from 漢伯戚𥓓. Does anyone know what that is?

7.31.2:58: 與 is pronounced as dư, dữ, dư, trợ̉ in Vietnamese. Presumably its variant 𠔇 has Chinese readings like Mandarin yu which corresponds to the first three Chinese-based Vietnamese readings. (Chinese y- became Vietnamese d-. The fourth reading trợ̉ is a native Vietnamese word without a Chinese counterpart.) So where do the Vietnamese readings sước and xước in Wiktionary come from? See below.

7.31.3:01: I found 𠔇 under the radical 八 using the radical search feature in Andrew West's BabelMap freeware. But Unihan lists its radical as 辵/辶! And now I see that BabelMap also lists it under辵/辶. Aha, sước and xước are Vietnamese readings of 辵. A third reading is xích. Was 𠔇 used for 辵 in Vietnam?

*7.31.00:27: 𠔅 and 𠔆 are nearly identical in MingLiU_HKSCS-Ext B, MingLiU-Ext B, and PMingLiU-Ext B but clearly distinguished in SimSun-Ext B which has a bottom element ㄕ slightly different from 尸 'corpse' in𠔆.

**7.31.4:01: 歹 is unusual because it has three unrelated readings:

Md dai 'bad' < MC *təjʔ 'horn' (the source of a later meaning 'bad', or an unrelated homophone?)

Md e 'bone fragment' < MC *ŋat

Md jing 'excess of bone rot' < MC *kɨŋ

𠔅 and 𠔆 can substitute for the 'bone'-related words written with 歹. What looks like 八 eight 尸 corpses (𠔅) isn't 歹 'bad'.

㝵, the phonetic of 䚟, a variant of 歹, has two unrelated readings:

Md de < MC *tək 'to get' (I'm surprised 㝵 hasn't become an official simplification of its homophone 得.)

Md ai < MC * *ŋəjh 'to obstruct' (variant of 礙, now officially simplified as 碍; 石 is 'stone')
The first reading of 㝵 is from OC *tək which is close to OC *təʔ, the expected source of 䚟/歹 MC *təjʔ. However, as far as I know, only the  *ŋ-reading is attested in late Old Chinese. The Shuowen states that 歹 is to be read like 櫱 *ŋat 'tree stump'.

Usually even very different readings of a single character may share a common prototype: e.g., the two readings of 櫱 which have no phonemes in common in modern Mandarin:

Md e [ɤ] < MC *ŋat < OC *ŋat (with 'emphasis')

Md nie [njɛ] < MC *ŋɨet < OC *ŋat (without 'emphasis' - what a difference its absence makes!)

The phonetic of 櫱 is

辥 Md xie [ɕjɛ] < OC *sŋat 'to correct' A BAFFLING BOWL FULL OF *L-I-Q-UID (PART 1)

is what 益 'to increase, more, advantage, profitable' originally represented. 皿 at the bottom is a bowl and the rest on top is liquid. The current graph looks like liquid splashing out of a bowl. Increase the amount of liquid and it overflows.

益 and its derivatives have three kinds of initials in Middle Chinese:

*ʔ-:*ʔiek itself, 縊 *ʔieh, *ʔiejh 'to strangle' (from my previous post)

*ŋ-:*ŋek 'a kind of aquatic bird'

*j-:*jit 'to overflow' and 鎰 *jit 'an ancient weight'

Schuessler and Karlgren split this group into two:

a *-k-group: most 益-graphs (MC *-(j)h of 縊 can be from Old Chinese *-ks)

a *-t-group: 溢 and 鎰

On the other hand, I think there was only one 益-series. MC *-it can come from early OC *-ik, so perhaps the MC rhymes of all 益-graphs can be derived from OC *-Ik(-s).

益 MC *ʔiek < OC *-ek 'to increase'

鷁 MC *ŋek < OC *-ek 'a kind of aquatic bird'

溢 MC *jit < OC *-it < *-ik 'to overflow'

may be cognates with *e ~ *i ablaut. That resolves the problem of the finals for now. But what about the initials?

Years ago, I would have reconstructed the 益-series with *(C-)l-:

MC *ʔ- < OC *ʔ-l-
MC *ŋ- < OC ?*N-l- (cf. Sagart 1999's OC *N-l- which is borrowed into non-Chinese languages as a palatal nasal but becomes MC *j-)

MC *j- < OC *l-

(I'm going to ignore emphasis.)

However, Baxter and Sagart reconstructed the 益-series with uvulars:

MC *ʔ- < OC *q-

MC *j- < OC *N-q-, *ɢ-

They didn't reconstruct the OC initial of 鷁.*

Next: Which approach is correct?

*7.30.15:06: Baxter and Sagart reconstructed the following sources for MC *ŋ- in OC uvular series:

*-qh- *-qhˤ- *-qhʷˤ- *-ɢ- *-ɢʷ- *-ɢˤ-
minor syllable *m.-


prefix *m- *m-qh- *m.qhˤ- *m-qhʷˤ-

minor syllable *N.- *N.qh-

prefix *N- *N-qh- *N-qhˤ- (see below)
*N-ɢʷ- *N-ɢˤ-

The gaps may be accidental. OC *N-qhʷˤ- did occur, but did not develop into MC *ŋ-. See below.

Any of the emphatic sources (in bold) might have been an OC initial for 鷁 MC *ŋek in their system. My guess would be *m-ɢˤ- since *m- occurs in animal names.

Note the absence of unaspirated voiceless consonants in second position. *NQ-clusters became voiced nonnasals in their system: e.g.,

OC *m-q-, *N-q- > MC *j-

OC *m-qˤ- (no cases of OC *N-qˤ-) > MC *ɣ-

OC *m-qʷ-, N.qʷ-, *N-qʷ- > MC *w-

OC  *m.qʷˤ-, *m-qʷˁ- *N-qʷˤ- > MC *ɣw-

They reconstructed only one instance of OC *NQh- becoming a nonnasal.

華 OC *N-qhʷˤra- > MC *ɣwæ 'to flower, flowery'

The aspiration is carried over from the root noun

華 OC *qhʷˤra- > MC *xwæ 'flower'

I would expect OC *N-qhʷˤ- to become MC *ŋ-. I don't see any reason why it would develop differently from all other *N-Qh-clusters. There are no other OC words with *N-qhʷˤ-.

I think Baxter and Sagart reconstructed 華 as a uvular series because it contains MC *xw-. They no longer reconstruct *x- as an OC initial. They derive all MC *x(w)- from OC voiceless sonorants (e.g., *hm-, *hn-, *hŋ-, *hl-, *hr-) and voiceless aspirated uvulars (e.g., *qh-).

The only fricatives in Baxter and Sagart's reconstruction are *s- and *sˁ-. EXCEPTIONAL KINDNESS

can be a unexpected source of frustration.

恩 Middle Chinese *ʔən 'kindness' is the odd man out in the phonetic series of 因. All other members had front vowels (*i, *e) in Middle Chinese. If not for 恩, it would be tempting to reconstruct the series with Old Chinese *i:

Character Old Chinese Middle Chinese
因茵絪姻駰咽 *ʔin *ʔin
咽烟 *ʔin *ʔen
*ʔin-s *ʔenh
*ʔit *ʔet
*ʔʷin *ʔwen

Nonemphatic OC *i remained intact in MC, whereas emphatic OC *i lowered to MC *e.

The only two oddities in that 恩-less series are

*ʔit (with a stop-nasal alternation occasionally found in other series: e.g.,

螣 OC *ŋ 'flying snake'* ~ *k 'grain plant leaf-eating insect'.

盾 OC *lunʔ < ?*-nt 'shield' : 腯 *lut 'fat'

念 OC *nims 'to think of' : 敜 OC *nip 'to fill'

*ʔʷin with unexpected labialization of the initial - a trace of a lost preceding labial segment or presyllable?

*P(V)-ʔ- or *CU-ʔ- > *ʔʷ-

恩 adds a third complication that is much more difficult: a nonpalatal vowel. Although the phonetic 因 suggests a palatal vowel *i or *e, there is no other evidence to suggest that 恩 ever had such a vowel. Here are three ways to account for the anomaly of 恩.

1. Regard 因 as a *jə-series

This is a variation of what I proposed last night:

Character Early Old Chinese Late Old Chinese Middle Chinese
因茵絪姻駰咽 *ʔɯ-jən/ŋ *ʔin *ʔin
咽烟 *ʔʌ-jən/ŋ *ʔin *ʔen
*ʔʌ-jən/ŋ-s *ʔin-s *ʔenh
*ʔʌ-jən-t/ŋ-k *ʔit *ʔet
*ʔo-jən/ŋ *ʔʷin *ʔwen
*ʔʌ-ən *ʔən *ʔən

EOC *ʔ- could have been a *q- that was borrowed into Proto-Tai as *k-. This *q- is also close to the initial of Baxter and Sagart's 縊 *qˤik-s 'to strangle' which may be cognate to 咽 *ʔʌ-jəŋ-k 'to swallow'.

EOC *-j- could have been *-l- if 咽 was indeed the source of Siamese กลืน klɨɨn 'to swallow' and if its *-t reading is cognate to 縊 *-ik-s 'to strangle' whose initial may have contained *l (more on this next time).

EOC *jə became *ji before emphasis developed: e.g.,

*ʔɯ-jəŋ > *ʔɯ-jiŋ > *ʔjiŋ > *ʔin

*ʔʌ-jəŋ > *ʔʌ-jiŋ > *ʔʌ-jiŋ > *ʔjiŋ > *ʔin

LOC *-n after *i/*i could be from EOC *-ŋ as well as EOC *-n.

Similarly, LOC *-t after *i/*i could be from EOC *-k as well as EOC *-t.

One problem is that 恩 doesn't quite fit the *QV-jəN pattern either because it has no medial consonant.

2. Regard 因 as a *-əɲ series

Pulleyblank proposed OC *-əŋʲ instead of *-in. The *-ən of 恩 would have been close to his *-əŋʲ which may have been pronounced as *[əɲ].

3. Regard 因 as a semantically motivated imperfect phonetic element in 恩

恩 'kindness' could be what the 心 heart 因 relies on. 因 could be an imperfect phonetic chosen in spite of its vowel *i. 恩 probably had no homophones in OC, so there was no existing phonetic that would have been an exact match. For other possible examples of semantically motivated imperfect phonetic elements, see Schuessler (2009: 34-39).

7.29.6:59: At the moment, I think 3 is the most likely solution. Here is one more:

4. Regard the *ə-reading of 恩 as a loan from another dialect

In this scenario, the graph 恩 was devised for an OC syllable with *i, but in another dialect, *i became *ə, and MC *ʔən is a descendant of a loan from this dialect.

I don't want to posit an entire dialect simply to explain one word. Although an *i to *ə shift is plausible**, I don't know of any other evidence for a dialect with *ə where *i is expected. (Although 黽 below did have both *ə- and *i-readings, those readings may not have been cognates in different dialects.)

*7.29.7:05: Cognate to 蠅 OC *mV-ləŋ 'fly (insect)', whose phonetic 黽 had several OC readings:

*mrəŋ(ʔ) < ?*mʌ-rəŋʔ 'frog'

*mrəŋ < ?*mʌ-rəŋ (place name)

*menʔ < ?*mɯ-lenʔ (first half of place name 黽池)

*minʔ < ?*mliŋʔ (alternate reading of first half of place name 黽池)

*mrənʔ (first half of 黽勉 *mrənʔ-mranʔ 'work hard')

If 黽 had ~ *e ~ *i variation, then perhaps 因 could be a phonetic in *ə- and *i-syllables.

蠅 OC *mV-ləŋ 'fly' may have had a front vowel variant *mʌ-leŋ that was borrowed into Proto-Tai as *ma-lɛŋ > Siamese แมลง  ma-lɛɛŋ ~ แมง  mɛɛŋ 'insect'.

**Long after the OC period, *i lowered in closed syllables in the southern Middle Chinese dialect underlying Sino-Vietnamese and in Cantonese: e.g.,

因 MC *ʔin > SV nhân [ɲən] and Cantonese yan

7.29.7:17: Proto-Tai'o'Matic reconstructed Proto-Tai *klɛn 'to swallow' in Li Fang-Kuei's system. However, Li (1977: 265) himself wrote that the vowel in 'to swallow' may have been *ɨə. Perhaps PT *klɨən is from an Old Chinese *qlɨən or *ʔlɨən:

OC *qlən > *qlɨən > *ʔlɨən > *ʔjɨən > *ʔjiən > *ʔjin > *ʔin BIG IN A BOX OR A MAN ON A MAT

describes the Chinese character 因 which to modern eyes looks like 大 'big' in a 囗 box but originated as a drawing of "a 大 man resting on a 囗 mat" (Karlgren 1957: 106).

It is the phonetic of 咽 'swallow' (v.) from my last two entries. Below I list all other graphs with 因 as a phonetic from Karlgren (1957)* along with my reconstructions of their Old Chinese readings. I include Baxter-Sagart reconstructions for comparison whenever available. Glosses are not exhaustive.

Character Gloss My Old Chinese Baxter-Sagart's Old Chinese My Middle Chinese Mandarin Notes
rest upon, therefore *q(ɯ)lən *ʔin *ʔin yin
mat (?)
mat, bed, generative influence of Heaven and Earth
marriage alternate spelling 婣 with labialized phonetic implies *qʷl- < *qul-
gray-white horse
gullet *qʌlən > qlən *ʔˤin *ʔen yan
to swallow *qʌlən-s > *qlən-s *ʔˤin-s *ʔenh source of Siamese กลืน klɨɨn 'to swallow'?; but Siamese tone is irregular (tone A instead of expected tone B)
to choke with emotion (*ʔit) (?) *ʔet ye not attested in Old Chinese; probably same word as 噎 MC *ʔet with 壹 OC/MC *ʔit as phonetic?; newer spelling 咽 may date after merger of early OC *qlə and *ʔi into late OC *ʔe
drum beat *qolən > *qlən ~
*ʔin*ʔwen yin ~ yuan
smoke *qʌlən > *qlən *qˤin *ʔen yan also spelled 煙
kindness *Cʌqən > *qən *ʔˤən *ʔən en first borrowed into Vietnamese as ơn [ʔəən]


1. I reconstruct *q- instead of *ʔ- in the 因-series to

- account for Sino-Thai k- (Thai probably did not have *q- at the time of borrowing)

- avoid reconstructing *ʔʷ- found in only one OC word: 淵 'abyss' which shares a phonetic with 姻/婣; I reconstruct 淵 and 姻/婣 with uvulars instead

2. I reconstruct *-l- in the 因-series to account for Sino-Thai -l-.

3. I reconstruct in the 因-series to account for the nonpalatal vowels of Sino-Thai -ɨɨ- and the of 恩 MC *ʔən.

4. Schwa fronted after liquids in the 因-series:

OC *qlə > *qjə > *ʔjə > *ʔjɨə > *ʔjɨ > *ʔji > *ʔi

OC *qlə > *ʔlə > *ʔjə > *ʔje > *ʔe

The latter is improbable. OC *l became MC *j, but OC emphatic *l became *d. Why didn't OC *qd- become and then merge with *d-?

There was no liquid in 恩 *qən, so its schwa never fronted.

5. The emphasis in the 因-series (indicated with underlining) is due to earlier presyllables with nonhigh vowels. If the nonhigh vowels were lost before they conditioned emphasis, the syllable remains nonemphatic: e.g.,

*qolən > *qlən > *qjən > *ʔjən > *ʔjɨən > *ʔjɨn > *ʔjin > *ʔin 'drum beat'

and are cover symbols for unknown unstressed nonlow and low vowels in presyllables: e.g.,

*qɯlən could have been *[qilən], *[qələn], or *[qulən]

*qʌlən could have been *[qelən], *[qalən], or *[qolən]

6. The labiality in 咽 and implied in 姻/婣 is due to the labial vowel of earlier *q-presyllables.

7. I don't know why Baxter and Sagart reconstructed 烟 with *qˤ- while reconstructing all other 因-words with *ʔ(ˤ)-. 煙, the alternate spelling of 烟, does not contain an unambiguously uvular phonetic.

8. I am not comfortable with my reconstruction. Should a single possible Sino-Thai word justify projecting its features into an entire phonetic series? I doubt that Old Chinese had that many words with the structure *QVlən. Moreover, 恩 *Cʌqən 'kindness' isn't like *QVlən.

7.28.2:14: Baxter and Sagart reconstruct 噎, the older spelling of 咽 'to choke', as *qˤik even though they reconstruct its phonetic 壹 'one' as *ʔit, not *qik. I think this is because they regard their 噎 *qˤik as a cognate of 

*qik  'throat'

*qek 'flesh on the neck' (I don't understand why Baxter and Sagart don't reconstruct this word and 'throat' as homophones, as I know of no evidence for different vowels)

*qˤik-s 'to strangle'

If they are correct, perhaps the spelling 噎 dates from a period after *qˤik had merged with 壹 *ʔit.

The above words may be members of a *Q-K word family:

*qek 'flesh on the neck'
噎 ~ 咽 *qˤik 'to choke'

Bare ending *-s suffix
Stop ending *-k *qik 'throat' *qˤik-s 'to strangle'
Nasal ending  *-ŋ < *-k-N? *qˁin < *qˁiŋ 'gullet' *qˁin-s < *qˁiŋ-s 'to swallow'

This proposal is more or less within the Baxter-Sagart framework (but with my *q- for 咽). It clashes with my earlier *QVlən proposal. Not a problem since I don't really care for the latter.

7.28.2:28: Perhaps 'kindness' was once *lʌqən, and the spelling 恩 reflects a later metathesized variant *qlən:

*lʌqən > *lʌqən > *lqən > *qlən

But there is no evidence for an *l-presyllable or a medial *-l-. Those *l-s are merely my attempts to reconcile 恩 with the rest of the 因-series.

c. 7.28.3:00+: The 因-series may mix earlier velar codas with earlier dental codas if the additional changes in bold occurred:

OC *qləŋ > *qjəŋ > *ʔjəŋ > *ʔjɨəŋ > *ʔjɨŋ > *ʔjiŋ > *ʔiŋ > *ʔin

OC *qlək > *qjək > *ʔjək > *ʔjɨək > *ʔjɨk > *ʔjik > *ʔik > *ʔit

(The rest is identical to the sequence of changes in note 4 plus velar codas.)

However, no such changes could have occurred in emphatic syllables because emphasis would prevent from raising to an *i that conditions the fronting of velar codas.

*Other 因-characters (e.g., 氤銦洇栶秵筃裀鞇) were created later. These graphs cannot tell us anything about the earliest sound value of their phonetic 因. For instance, 銦 yin 'indium' (< 金 'metal' + 因 yin) was probably created some time after the discovery of indium in 1863. The structure of 銦 tells us that 銦 yin must have sounded like 因 yin within the last century or so, but we already knew that. A HARD-TO-SWALLOW ETYMOLOGY

Juha Janhunen (2003: 395) described the homophony of Khitan


(large script)/(small script)

*tau 'five'

and the first syllable of

, ,/

<tau.lia>, <tau.lia>, <tau.lia> / <tau.li.a>

(large script)/(small script)

*taulia 'hare'

as "the most crucial evidence of the Para-Mongolic identity of the underlying language" of the Khitan scripts.

Luc Kwanten has proposed that the underlying language of the Tangut script is 'Altaic' in spite of all evidence to the contrary. If that were true, I would expect homophones such as Old Turkic at 'name' / 'horse' to correspond to similar Tangut characters. But the Tangut characters for 'name' and 'horse' are clearly different and their readings have Sino-Tibetan etymologies

2639 2miee 'name'

cf. Old Chinese 名 *meŋ, Old Tibetan myiŋ 'name', Written Burmese maɲ 'be named'

0764 1rieʳ 'horse'

cf. Old Chinese 馬 *mraʔ, Old Tibetan rmaŋ < *m-r-, Written Burmese mraŋḥ 'horse'

This absence of homophony is not evidence for the absence of 'Altaic' elements in Tangut. It is possible that there are some undetected loanwords in Tangut, particularly in the 'ritual language' (and/or in 'Tangut B', if such a language ever existed). The Tangut word for 'Khitan'

4361 4389 1tʃhɨə 1tã

is an example of a known 'Altaic' loan. It is probably from a Khitan dialect with palatalization (a process unknown in Tangut or in the Chinese of the time).

Moreover, the presence of homophony may not be meaningful. If partial homophony is "crucial evidence" for the identification of Khitan as para-Mongolic (and I think it is), what are we to make of this total homophony that came to mind as I wrote my last entry?

Old Chinese 嚥 *ʔens 'swallow (verb)' : 燕 ʔens 'swallow (noun)'

English swallow (verb) : swallow (noun)

Suppose we knew nothing about the pronunciation of Chinese, but knew that similar-looking sinographs were likely to have similar readings. It would be reasonable (and in fact correct) to guess that 燕, a drawing of a swallow, sounded like 嚥 'to swallow' (with 口 'mouth' added on the left). Should we conclude that Chinese was related to English since 'swallow' and 'to swallow' are homophonous in both languages?

No. We have Chinese transcriptions of the Khitan words for 'five' and 'hare'


陶里 *thauli

that not only confirm the homophony but indicated that the shared syllable was *tau, close to Written Mongolian tabun 'five' and taulia 'hare'. Without transcriptions of 嚥 and 燕, we would not know whether those characters represented anything that was phonetically similar to English swallow.

The Khitan words for 'five' and 'hare' are unrelated. Similarly, Schuessler (2007: 552, 556) pointed out that

*ʔens 'swallow' (v.) is not related to 燕  *ʔens 'swallow' (n.)

*ʔins 'swallow' (v.) may have acquired a new spelling 嚥 with the phonetic 燕 *ʔens  after the lowering of emphatic *i to *e.

swallow (v.) < Old English swelgan is not related to swallow (n.) < Old English swealwe

Homophones aren't necessarily eternal. They may result from the convergence of two originally different words: *ʔins and *ʔens and swelgan and swealwe.

7.27.1:19: One could regard *ʔins 'swallow' (v.) and *ʔens 'swallow' (n.) (and 咽 *ʔin 'gullet') as members of a *ʔ-n word family with ablaut, but are there any other cases of i-verbs corresponding to e-nouns? WHAT LIES AT THE EDGE OF THE ABYSS?

In my last entry, I mentioned that there was only one word in the Baxter-Sagart reconstruction of Old Chinese  (BSOC) with the pharyngealized initial *ʔʷˤ-:

淵 BSOC *ʔʷˤin (my *ʔʷin *[ʕʷɪˁnˁ]) 'abyss'

There are no BSOC words with nonpharyngealized initial *ʔʷ-. The BSOC Excel file is by no means a complete vocabulary of Old Chinese (OC), not even at the root level. (It obviously excludes polysyllabic words.) So perhaps BSOC *ʔʷ- does exist in words outside this extensive sample. In any case, most words that Baxter and Sagart once would have reconstructed with labioglottals are now reconstructed with labiouvulars.

Can 淵 be reconstructed with a labiouvular instead of *ʔʷˤ-? 淵 belongs to a very small and unusual phonetic series with only two members*, itself and

婣 BSOC ?*ʔin 'marriage'

which has a variant 姻 with the phonetic 因 BSOC *ʔin

(Although there is no BSOC reconstruction for 婣/姻, they have been homophonous with 因 since Middle Chinese, and they may have been homophonous in Old Chinese as well.)

A phonetic series with C- ~ Cʷ-alternation is odd. I can't think of any other one.

淵 BSOC *ʔʷˤin became Middle Chinese *ʔwen, which could have also come from BSOC *qʷˤin. Such a labiouvular reconstruction makes the rare BSOC initial *ʔʷˤ- unnecessary. However, since a uvular ~ glottal alternation in a phonetic series is unusual, I would then want to reconstruct 婣/姻 with a uvular initial as *qin.

Sagart (1999: 109) pointed out that 咽 MC *ʔenh 'to swallow'

- corresponds to Siamese กลืน klɨɨn (a possible loanword from late OC) with an initial kl-

- may be cognate to 吞 BSOC *hlˁən 'to swallow'

suggesting a *Cl-cluster in Old Chinese. He reconstructed 咽 as OC *aq-lin-s in 1999. (But the current BSOC reconstruction of 咽 is *ʔˁin without a liquid! BSOC has no *ql-clusters, though it has *qr-clusters.)

Perhaps 婣/姻, sharing a phonetic with 咽, might have been *q(l)in with a uvular initial like my proposed 淵 *qʷˤin. A *ql- ~ *qʷ- alternation is phonetically plausible given the phonetic similarity of l and w: e.g., the cute variant wittle of English little and the [w] pronunciation of Polish ł < *l.

*I am not listing the many other variants of 淵: e.g., 渕 for 淵 with 关 'barrier' and 刀 'knife' in place of a drawing of a ditch or 囦 with 水 water in a box. Nor am I listing most of the variants of 婣: e.g., 因 'reason' + 生 'life' which is an interesting way to look at marriage. PROTO-INDO-EUROPEAN (AND OLD CHINESE?) *ʕʷ

Beekes (1995: 143, 148) reconstructed three laryngeals in Proto-Indo-European (PIE) which became vowels in Greek:

PIE > Gr e

PIE > Gr a

PIE *ʕʷ > Gr o

Why wasn't the third laryngeal reconstructed as ʔʷ which may be a little less exotic? The labialization -ʷ- accounts for the rounding of o, but I'm not sure why it's a pharyngeal instead of a glottal. My guess is that a pharyngeal accounts for the backness of o relative to e, just as it accounts for the backness of a relative to e.

Beekes (1995: 126) reconstructed the following changes between earlier and later PIE:

PIE *ʔe > *e

PIE *ʕe > *a

PIE *ʕʷe > *o

but note that PIE *ʔo, *ʕo, *ʕʷo all merged into o regardless of the preceding laryngeal.

Beekes' reconstruction of the laryngeals results in some complex clusters in PIE: e.g.,

PIE isʔros > Gr *iheros > hieros 'holy'

PIE pʕteer > Gr pateer 'father'

PIE dʕʷtos > Gr dotos 'given'

Are there any current languages with C-laryngeal-C clusters in initial position?

7.25.00:17: Khmer has seven Cʔ-sequences, but no CʔC-sequences:

pʔ-, mʔ-, tʔ-, lʔ-, sʔ-, chʔ-, kʔ-

(I don't know why chʔ- is the only cluster with an aspirate. Huffman (1970: 8) disregarded Khmer spelling and analyzed it as /cʔ/ without aspiration like /pʔ mʔ tʔ lʔ sʔ kʔ/.)

7.25.1:31: I have wondered if Old Chinese (OC) 'emphatic' *ʔ- and *ʔʷ- were phonetically *[ʕ] and *[ʕʷ]: e.g.,

*ʔan *[ʕɑˁnˁ] 'peace(ful)'

*ʔʷ *[ʕʷɑˁŋˁ] 'pool'

with nonemphatic phonetic 王 *w 'king'

also cf. 枉 *ʔʷaŋʔ 'bent' with a labioglottal sans emphasis

But perhaps there is no more need to reconstruct a labialized glottal stop or pharyngeal fricative in OC. In the new Baxter-Sagart OC reconstruction (BSOC; publicly quotable at last - hooray!), I can only find a single instance of an 'emphatic' labialized glottal stop:

淵 BSOC *ʔʷˤin (my *ʔʷin *[ʕʷɪˁnˁ]) 'abyss'

My *ʔʷ- and *ʔʷ- *[ʕʷ] otherwise correspond to their uvular *qʷ- and *qʷˁ-: e.g., they reconstruct

枉 BSOC *aŋʔ (my *ʔʷaŋʔ) 'bent'

汪 BSOC *qʷˁ (my *ʔʷ) 'pool'

with phonetic 王 BSOC *ɢʷ (my *waŋ) 'king'

A distinction between plain and pharyngealized (labialized) uvulars is very rare. My OC reconstruction has no such distinction. I can't find any language in UPSID with /qʷˁ/, though Rutul has a /qhʷ/ : /qhʷˁ/ distinction.

Rutul has eight voiceless uvular stops but lacks a simple q:

plain pharyngealized
aspirate qh qhˁ
labialized aspirate qhʷ qhʷˁ
ejective qˁʼ
labialized ejective qʷʼ qʷˁʼ

Rutul also lacks simple p, t, ts, k, kw. Its voiceless obstruents are either aspirates or ejectives.

Tangut fonts by Mojikyo.org
Tangut radical font by Andrew West
All other content copyright © 2002-2011 Amritavision