18.2.28.23:59: A SHARED *SHCH-IFT IN CHINESE AND RUSSIAN

Last Friday (yes, I'm behind), I saw

新商品 'new product', lit. 'new trade item'

on packaging.

In Old Chinese, 商 was

either *sɯ-taŋ (corresponding to Baxter and Sagart 2014's *s-taŋ)

or *sɯ-laŋ (corresponding to Schuessler 2009's *lhaŋ)

and in Middle Chinese, it was *ɕɨaŋ.

It occurred to me that the palatalization of *sɯ-t- to *ɕ-

*sɯ-t- > *sɯ-tɨ-  > *stɨ- > *stɕɨ- > *ɕtɕɨ- > *ɕːɨ- > *ɕɨ-

was like what I understand to be the palatalization of *stj- to [ɕː] in Russian:

*stj- > *stɕ- > *ɕtɕ- > щ [ɕː]

Above I presume there was an intermediate *ɕtɕɨ-stage at some point in Old Chinese resembling romanizations of Russian щ as šč or shch (e.g., Хрущёв Khrushchev), but without external evidence (e.g., Old Chinese transcriptions of a foreign word with šč-), it's impossible to say when that point was.

вместить 'to contain (perf.)' ~ вмещу 'I will contain'

can be internally reconstructed as

*vmestitĭ ~ *vmestju

to fit the pattern of

вменить 'to consider (perf.)' ~ вменю 'I will consider'

< *vmenitĭ ~ *vmenju

Ideally I'd like to find an example of initial щ- [ɕː] from *stj-, but I think initial щ [ɕː] is normally from *sk-. A possible exception I found in Preobrazhensky's Etymological Dictionary of the Russian Language is щегол 'goldfinch'; Duden says German Stieglitz 'goldfinch' is of Slavic origin.

Proto-Slavic *štjegŭlŭ? > *ščegŭlŭ

East Slavic:

Ukrainian щиголь <ščyhol'>, щоголь <ščohol'>, щоглих <ščohlix>

Belarusian щигель <ščihel'>, щиглик <ščiglik> (I have kept Preobrazhensky's spellings with щ and и instead of modern шч and і)

(why -ль as if from *-lĭ?)

(no South Slavic reflexes? I would expect Bulgarian initial щ- [št], Serbo-Croatian initial št-, and Slovene initial šč-)

West Slavic:

Czech stehlec, stehlík (with ste- rather than the regular ště- [ʃcɛ] - could this be a borrowing from some variety of German in which st- was [st] instead of [ʃt]?)

Polish szczygieł [ʂtʂɨɡʲɛw]

Upper Sorbian šćihlica [ʃtsʲihlitsa]

Lower Sorbian ščgeľc [ʂtʂgɛlts] (I have kept Preobrazhensky's spelling with ľ instead of modern l)

The reflexes of *stj- could have had parallels in Old Chinese at different stages and/or different places.

18.2.27.23:54: A LITTLE MISTAKE: ÍT ÓT TO BE THE PHONETIC

In my last post, I wrote that 乚 ất was the phonetic of the Vietnamese Chữ Nôm character 𡮒 ót 'a kind of fish'. After announcing that post on Twitter, I realized that the actual phonetic was 𠃝 which has two readings, ít 'little' and út 'youngest'. I didn't think of 𠃝 because 乙 appears as 乚  in 𡮒.

If the creator of 𡮒 had the reading út in mind for its phonetic 𠃝, the score of 𡮒 would be 2 + 3 + 2 + 2 = 9 - much higher than my original score of 6.

乙 is a 'Semitic phonetic': it can represent syllables with a wide range of vowels as long as those vowels are within the consonantal frame [ʔ-t]:

Neutral or achromatic vowels (neither palatal nor labial)

㲸 ướt [ʔɨət]

乙 ất [ʔət]

艺 ớt [ʔəːt]

𢖮 ắt [ʔat]

𢖮 át [ʔaːt]

Palatal vowels

𠃝 ít [ʔit]

𠮙 ét [ʔɛt]

Labial vowels

𠃝 út [ʔut]

𡮒 ót [ʔɔt]

All of those syllables have the sắc tone written with an acute accent. Syllables with initial glottal stops and final stops regularly develop that tone.

Such a range of vocalism for a phonetic is unusual in Chữ Nôm. In my 2003 book, I proposed that phonetics generally belong to three vowel classes: neutral, palatal, or labial.

'Semitic phonetics' are exceptions to that generalization: e.g., 曰 viết in

neutral: 曰 vất [vət], 抇 vớt [vəːt]

palatal: 𢪏 vít [vit], 𧿭 vết [vet], 𢪏 vét [vɛt]

labial: ⿰曰𡿨 vót [vɔt]

3.1.0:39: Compare the ranges of readings for 'Semitic phonetics' above with those for کت <kt> listed in Hayyim's  New Persian-English Dictionary:

neutral: kat

palatal: ket

labial: kot

(Of course, Persian is not a Semitic language, but it is written in a Semitic script.)

One difference is that all of those k-t readings have no tones, whereas all of the readings for Chữ Nôm characters with the two 'Semitic phonetics' above have the same tone. Perhaps the term 'Semitic phonetic' is a misnomer if the consonantal frames are actually consonant-and-tone frames.

甘 cam is a third 'Semitic phonetic' whose derivatives below have readings with three different tones (ngang, huyền, sắc) as well as three different vowel classes:

neutral: 坩柑泔 cam [kaːm], 紺 cám [kaːm], ⿰月甘 cằm [kam], 𩚵 [kəːm], 鉗 cườm [kɨəm]

palatal: 鉗 kìm [kim], ghìm [ɣim], kiềm [kiəm], kềm [kem], kèm [kɛm]

labial: 鉗 cùm [kum], 柑 cùm [kum]

Note, however, that all but one of the readings in that sample have either the ngang or huyền tones which are variants of the same proto-tone conditioned by voicing or its absence in proto-onsets. Also, only one of those characters is a made-in-Vietnam character (⿰月甘). 甘 was already a neutral and palatal phonetic in Middle Chinese because Old Chinese *a often had palatal reflexes after nonemphatic initials. An ideal example of a 'Semitic phonetic' would have many made-in-Vietnam derivatives with a wide range of vowels and tones. I should dig deeper to see if I can find one.

18.2.26.11:46: ÓT TO BE WRITTEN: FISHING FOR PHONETICS

The Vietnamese Chữ Nôm script represents Vietnamese syllables with existing and modified Chinese characters. The problem is that Vietnamese has many more syllables than Sino-Vietnamese, the subset of Vietnamese syllables that are Chinese character readings. For instance, Vietnamese has syllables ending in -ót, a rhyme absent from Sino-Vietnamese.

In my last two posts, I looked at Vietnamese solutions for writing the syllable lót.

I got curious about how other -ót syllables were written and found several strategies. My examples are not exhaustive, and I have omitted glosses in most cases since I am focusing on readings.

1. Overall match

⿰口脫 thót : 脫 thoát (score: 2 + 3 + 2 + 2 = 9; not a 10 only because the vowel heights don't match: o [ɔ] is higher than oa [wa], though I could be generous and say oa is like [o] + [a], and [ɔ] is between those two vowels in height)

𡮒 ót 'a kind of fish' : 乚 ất (the unwritten onset is [ʔ]; score: 2 + 0 + 2 + 2 = 6)

(This analysis and scoring are wrong. See my next post.)

㩢 mót : 蔑 miệt (score: 2 + 1 + 2 + 1 = 6; the only matching vowel quality is length*)

⿰曰𡿨 vót : 曰 viết (score: 2 + 1 + 2 + 2 = 7; the only matching vowel quality is length)

This is the consonantal skeleton or Semitic strategy. If English were written with such a strategy:

cat = drawing of a cat

Kate = <woman> + <cat>

kite = <wing> (representing flight) + <cat>

cut = <blade> + <cat>

coat = <clothes> + <cat>

coot = <bird> + <cat>

caught = <hand> + <cat>

Cf. the reverse Semitic strategy (5 below).

3. Matching the rhyme without much if any regard for the onset

𡁾 hót : 說 thuyết < *ɕ- (or *sʰ-?) (score: 0 or 1 + 2 + 2 + 2 = 6 or 7, depending on whether the aspiration of th- [tʰ] < *sʰ-? counts as a partial match for h-)

3b. *Palatal onset : nonpalatal phonetic

卒 chót with initial [c] : 卒 tốt < *(t)s- (score: 1 + 3 + 2 + 2 = 8)

埣 giót < *ɟ- < *CV-c- : 卒 tốt < *(t)s- (score: 0 or 1 + 3 + 2 + 2 = 7 or 8, depending on how close the initials were when 埣 was created: *CV-c- is not far from *(t)s-, whereas modern gi- [z] ~ [j] is far from t-)

悴 xót < *ɕ- < *cʰ- : 卒 tốt < *(t)s- (score: 0 or 1 + 3 + 2 + 2 = 7 or 8, depending on how close the initials were when 埣 was created: *cʰ- is not far from *(t)s-, whereas modern x- [s] is far from t-)

⿰律𡿨 xót < *ɕ- < *cʰ- : 律 luật (score: 0 + 2 + 2 + 1 = 5)

3c. *Retroflex onset : nonpalatal phonetic

卒 sót < *ʂ- < *Cr- : 卒 tốt < *(t)s- (score: 0 or 1 + 3 + 2 + 2 = 7 or 8, depending on whether the proto-onset was *sr- which isn't too far from *(t)s-; *(t)s- had hardened to t- by the time *Cr- fused into *ʂ-)

捽 rót < *r- or *CV-s- (proto-onset unknown) : 卒 tốt < *(t)s- (score: 0 or 1 + 3 + 2 + 2 = 7 or 8, depending on whether the proto-onset was *CV-s-)

3d. Palatal nasal onset nh- [ɲ] : oral onset phonetic

踤 nhót : 卒 tốt < *(t)s- (score: 0 + 3 + 2 + 2 = 7)

𦝬 nhót : 突 đột with initial [ɗ] < *t- (score: 0 + 3 + 2 + 1 = 6)

𣑵 nhót : 聿 duật with initial [z] ~ [j] < *dʲ- < *j- (score: 0 or 1 + 2 + 2 + 1 = 5 or 6, depending what the onset of 聿 was when 𣑵 was created)

3e. Lateral onset : nonlateral onset phonetic

⿰貝骨 lót : 骨 cốt (score: 0 + 3 + 2 + 2 = 7)

䘹 lót : 卒 tốt < *(t)s- (score: 1 + 3 + 2 + 2 = 8)

3g. Labial onset : nonlabial onset phonetic

𡁾 vót : 說 thuyết < *ɕ- (or *sʰ-?) (score: 0 + 2 + 2 + 2 = 6)

啐 vót : 卒 tốt < *(t)s- (score: 0 + 3 + 2 + 2 = 7)

This character could belong to 2 or 3a depending on which part is phonetic:

⿰孛乙 ót 'back of brain' : 孛 bột '' 'comet' + 乙 ất 'second Heavenly Stem' (score: 0 + 3 + 2 + 1 if 孛 is phonetic or 6 or 2 + 0 + 2  + 2 = 6 if 乙 is phonetic)

Neither part is obviously semantic. The absence of any component meaning 'brain' or even 'head' is puzzling. Could this be a double phonetic compound with 孛 approximating the vowel and 乙 the rest?

𠲿 thót : 束 thúc (score: 2 + 3 + 1 + 2 = 8)

I suspect 𠲿 was created by a speaker of a central or southern dialect in which *-t > [k]. If so, 𠲿 is really an example of strategy 1, and the score should be 9 (with a penalty solely for vowel height mismatch).

5. Approximating the vowel and tone without regard for the consonants

The reverse Semitic strategy (cf. 2 which is the Semitic strategy).

束 hót : 束 thúc (score: 0 + 3 + 1 + 2 = 6)

I suspect this usage of 束 started with a speaker of a central or southern dialect in which *-t > [k]. If so, 束 is really an example of strategy 3, and the score should be 7 (with penalties for the onset and vowel height mismatch). The score could be raised to 8 if the aspiration of th- [tʰ] counts as a partial match for h-.

No solution has a score of 4 for vowels simply because no phonetic has a Sino-Vietnamese reading with o [ɔ]. The maximum possible score for -ót syllables is 9 out of an ideal of 10 (= 2 + 4 + 2 + 2). The actual scores above range from 5 to 9. It is not possible to determine the median or the mode of scores for ót-characters from the data in this post because it is incomplete and only typologically rather than statistically represenative: e.g., I omitted all but one strategy 1 character with a score of 9 because near-exact matches are boring.

Until now Chữ Nôm characters and readings have been treated as a uniform, timeless body. The next phase of Chữ Nôm studies should take space and time into account: where and when do certain spellings arise, and what can they tell us about Vietnamese phonetics in a given place and period?

*I consider all Vietnamese vowels and diphthongs to be the same length for scoring purposes with the exceptions of the short vowels ă [a] and â [ə] which cannot appear in syllable-final position because all Vietnamese syllables must be bimoraic. Hypothetical *Că and *Câ-syllables would be monomoraic and therefore not permissible.