10.10.16.22:44: PLAY-ING WITH KORO
It's been over a week since I first heard of Koro, apparently a distant relative of Tangut. Since then, I've only seen a few words and sentences in imprecise English-based transcriptions. Two words might have Tangut cognates:
|Gloss||Koro||Tangut||Written Tibetan||Old Chinese|
1lɨəəʳ < *rɯ-lə
|bzhi < *blyi||四 *s-hli(t)s|
2niəə < *Cɯ-nəə < *Cɯ-nəC? 'day, daylight'
I assume Koro first syllables are prefixes and Koro second
syllables are roots.
The Koro prefix (?) ko- is not in soo-fee 'six', the only other numeral in the available data.
The Koro prefix (?) may- also occurs in may-pah 'night'. Is it in other weather-related words?
I don't understand why Tangut sometimes has central vowels corresponding to front vowels in its relatives. Another example is 'two':
Could pre-Tangut *ə have been from an *i that harmonized with a lost preceding central vowel or is it a preservation of a distinction lost in the other languages?
1niəə < *nəə < *nəC?
cf. Written Tibetan gnyis, Written Burmese hnac < *s-nik, Old Chinese *nis
10.10.16.22:27: MUONG RHOTICS
Since I just mentioned Muong, I looked it up at the Starling database and noticed that the dialect there (hereafter simply 'Muong') had no initial voiced r-. Like ancient Greek, it only has voiceless hr-. (Greek ῥ [hr] is transliterated as rh-.) Usually a language that has a voiceless hC- also has a voiced C-. What happened to voiced r- in Greek and Muong? Just as all Greek *r- became rh-, I propose that all Muong *r- became *hr-.
Relatives of Muong still retain *r-: e.g.,
|wide||hroŋ||rộng||rooŋʔ||rooŋ, ruuŋ, ruoŋ|
*hr- cannot be original since it occurs in words with tones that developed from *voiced as well as *voiceless initials.
On the other hand, Muong *Cr- became kh- (which I presume to be an aspirated stop [kh] rather than a fricative [x] as in Vietnamese):
*Cr- > *C-hr- > *hr- > *x- > kh-
*hr- > *x- must have occurred before *r- > *hr-.
The merger of *x- into kh- also occurred in Thai.
I briefly thought that Muong *hr- remained intact and that *r- > kh-, but we have just seen that *r- > hr- and Muong kh- often corresponds to clusters or cluster descendants in its relatives: e.g., Muong khăw 'six' corresponds to Arem prawʔ.
Vietnamese currently has sáu [ʂaw] without a cluster, but the old nom spellings for 'six'
六 'six' + 老 lao
娄 lâu + 六 'six'
have liquid-initial phonetics. Modern Ruc forms like psaaw and šraaw might resemble unattested transitional forms between earlier Vietnamese *praw and modern Vietnamese [ʂaw].
Muong kh- sometimes corresponds to r- in other languages: e.g.,
|axe||khiw||rìw||riw||n/a||mriiw, ʔmriiw, mǝriiiw||n/a|
Muong may have added a prefix to 'galingale' that was never in Vietnamese.
Vietnamese lost a prefix, presyllable, or preintial before *Cr- became [ʂ] in 'axe' and 'forest'.
The Khen dialect of Muong in Thompson (1976) which I read back in the 90s (and should reread) has r- and th- (but no hr-) corresponding to Vietnamese r-. I initially expected Khen th- to correspond to hr- in the dialect in Starling, but Thompson reconstructed
Proto-Vietic *r- > Khen r-Khen has a fricative x- corresponding to Vietnamese s- [ʂ]. This x- matches the *x- I reconstructed as a stage prior to kh- in the Starling Muong dialect.
Proto-Vietic *tr- > Khen th-
10.10.14.23:59: WERE THE SHAN-O THE MONTAGNARDS OF THE TANGUT EMPIRE?
(This is a revision and recreation of a post that I wrote last night that was lost when my laptop failed to 'wake up' after hibernation.)
Last month, I wrote (link added),
I have no idea who the Shan-O 'mountain masters' are. The name reminds me of French Montagnard. Are the Shan-O an ethnic group distinct from the previous ones mentioned in the Golden Guide, or are they just people of one or more ethnicities who live in the mountains? Kychanov (2006: 171) defined Shan-O as a 'name of a tribe'. Do the Shan-O appear in other texts that clarify their ethnicity?
Andrew West emailed,
The Shan'e [the Mandarin reading of 山訛, the Chinese name of the Shan-O] are mentioned in the History of the Song and History of the Jin, where they are identified as a particular Tangut tribe from Hengshan in Shaanxi that formed a renowned and fearsome element of the Tangut army.
The Tangut name of the Shan-O is a Sino-Tangut hybrid:
1ʃæ̃-1ʔo (Shan-O in simplified transcription)
1ʔo is a native Tangut word for 'master'. Could it also have been a Shan-O word with a similar if not identical meaning?
As for the first half, Andrew speculated that
maybe ʃæ̃ was a Tangut (Shan'e?) abbreviated borrowing from Chinese for the name of their particular mountain, i.e. Hengshan - Wenhai [Tangraphic Sea] defines #3763 [Li Fanwen's number for 1ʃæ̃] as being used as a place name character, and there seem to be no examples in Li Fanwen's dictionary of it being used as a general term for mountain, which perhaps supports the idea that the Shan'e were not simply "mountain masters" but the "masters of *the* mountain" (i.e. masters of Hengshan).
Anyhow, it is very interesting to speculate on the relationship between the Shan'e and the "ordinary" Tangut -- did the Shan'e speak a mutually intelligible dialect of Tangut? or was it a separate but related language? and does written Tangut reflect a Shan'e substratum?
All good questions without answers. They inspired me to take my Montagnard comparison even further. The similarities might go beyond the 'mountains' in Shan-O and Montagnard. In Vietnamese, the Montagnards are called Người Thượng, literally 'person above', i.e., highlanders. Like Shan-O, Người Thượng is a hybrid of a native word người and a Chinese loanword thượng (from 上). The highlanders were not Sinicized like the lowland Vietnamese, and many highlanders speak Mon-Khmer languages related to Vietnamese.
The closest relatives of the Vietnamese are the Muong (whose name is actually Tai; cf. Thai เมือง mɨaŋ 'town'), highlanders not included under the umbrella term Người Thượng. The Muong language is what Vietnamese might have been like without heavy Chinese influence. Perhaps Shan-O is what Tangut might have been like without heavy Chinese influence.
10.15.0:30: Another term for Montagnard is Degar. This Degar site defines it as 'sons of the mountains'. I don't know which language it comes from.
10.10.12.23:59: 2SI THE 5TH PAGE
The Tangut rhyme dictionaries Tangraphic Sea (TS) and Precious Rhymes of the Tangraphic Sea (PRTS) are also homophone dictionaries, though their homophone groups do not precisely match those of Homophones (and Homophones editions also have variations in grouping).
Let's look at 2si in context again. The second / rising tone volume of the TS is missing, but the corresponding volume of PRTS is still with us. 2si belongs to rhyme -i 2.10 (second tone, tenth rhyme) whose section is on pp. 2-4 of this PDF.
The name of the rhyme is the second tangraph from the bottom of the third line from the left:
1526 2tshi 'to wait upon'
Above it are two triangles that separate the 2.10 section from the 2.9 section. Beneath it is a small
1084 2ɣạ 'ten'
Beneath that rhyme number are homophone groups generally separated by circles. An exception is the tangraph directly under 'ten' which has no homophones and doesn't have a circle after it:
1427 2phi 'to lose, give up'
Was the omission an accident, or was the PRTS dialect reading of 1427 2mi like the next three tangraphs which do have a circle after them?
to distinguish them from the third homophone group:
Small tangraphs beneath main entry tangraphs are glosses or analyses: e.g.,
(See this entry for more on this equation.)
Homophone groups are arranged according to initial class in an order similar to the order of initials in Homophones. -i and -wi are separated.
-i 2.10 (initial order identical to Homophones)-wi 2.10
- dentals (not before velars as in -i 2.10 above and in Homophones)
As expected, 2si is in the alveolar section of -i 2.10 (2.5.1404-1504). The reference number breaks down as follows:
The order of 2si tangraphs is similar to that of Homophones, except that the last two tangraphs are listed first:
2 = second/rising tone volume of PRTS
5 = page number
1 = first side of page
4 = column number from right to left
04 = fourth tangraph
(I have excluded the two 1si tangraphs at the end preceding the circle separator.)
10.13.7:39: Note the homophone group divider present in PRTS but not in Homophones. Was it accidentally left out of the latter, or did PRTS make a distinction absent from Homophones?
The nature of the distinction, if any, is unknown. It cannot be a difference in rhymes since both groups share rhyme -i 2.10. One of the two groups cannot be swi 2.10 which is in another group later on the same page (2.5.1703-1706):
Members of the two 2si groups are transcribed differently in the Pearl:
5273 2si as 息
4953 2si as 西
but 息 and 西 were probably both *si in late 12th century northwestern Chinese.
Unfortunately, there are no fanqie for either 2si group which would clarify the distinction between them. Could that distinction be parallel to the unknown distinction between the two 1si groups (17B32-17B61 and 17B23-17B31) that I looked at last night?
I assume the circle cannot be used to divide semantic groups since the two members of the first 2si group have unrelated meanings (5273 is 'liver' and 3072 is 'to die').
On Saturday, I gave an overview of all the entries on page 30A of Homophones. I chose that page because it contained the 2si homophone group (30A16-30A26) that I mentioned on Friday. That group in bold is followed by a pair of 1si tangraphs (30A27-30A28) with underlining below:
In Homophones, homophone groups are separated by circles. One might expect a circle between the 2si group (30A16-30A26) and the 1si pair (30A27-30A28), but in fact there is no circle dividing the two, implying that 30A16-30A28 constitute a single homophone group of 11 tangraphs in spite of their tones according to Tangraphic Sea (TS) and Precious Rhymes of the Tangraphic Sea (PRTS).
Did the dialect of the author of Homophones lack tonal distinctions? If that were the case, I would expect a more random distribution of TS/PRTS tones. How would a toneless speaker be able to generally group syllables according to tones he didn't know? It would be like expecting modern English speakers to know German noun genders. Moreover, a lack of tones would not explain the fact that other 1si are in a separate homophone group on the previous page (29B62-29B68; in bold). O indicates the end of a homophone group: e.g, 29B14 is the last member of a homophone group. (Click the links to see notes on specific oddities.)
|1sị||1səu O||1sõ||2so||1tsha O||1dziew O||1tshi O||-1|
|1sị O||1si||1sõ O||2so||2sa||2dziew||1tshi O||-4|
|2tsi||1si||2səu||2so O||1sa O||2tsha||1dziew||-7|
The 1si group (29B62-29B68) is identical to the homophone group in Tangraphic Sea (17B32-17B61).
The 1si pair (30A27-30A28) is in the precding homophone group in Tangraphic Sea (17B23-17B31).The two 1si groups have different fanqie for what seem to be the same syllable:
1siu + 1khi = 1si
2siə + 1ki = 1si
No reconstruction I know of has any way to distinguish between the two kinds of 1si. Both final spellers (1khi and 1ki) belong to the same Tangraphic Sea rhyme (1.11), so the difference between the two 1si must be in the initials and/or medials.
10.12.0:16: If one assumes the Homophones groups were completely homophonous in the dialect of their author, perhaps the readings of the tangraphs on page 29B were as follows in bold:
|1sị||2səu O||1sõ||2so||2tsha O||1dziew O||1tshi O||-1|
|1sị O||1si||1sõ O||2so||2sa||2dziew||1tshi O||-4|
|2tsi||1si||2səu||2so O||2sa O||2tsha||1dziew||-7|
Readings have been altered to match the most common tone/rhyme in their groups.
Notes on a few individual alterations:
29B13: I don't know why Sofronov and Gong reconstructed R70 -ị (in my reconstruction) instead of R11 -i like the others. I also don't know what the difference between 29B12-29B14 and 29B11 was. In Tangraphic Sea, 29B11 and 29B12 are in the same homophone group, though they are in distinct groups in Homophones.
29B26: The inclusion of a R25 -ã syllable in an R17 -a group implies the two rhymes were similar.
29B48: I am puzzled by why this isn't in the preceding group. Is the circle placement in error, or was this really pronounced as 1sõ in the Homophones dialect but as 2so in the Precious Rhymes of the Tangraphic Sea dialect? Does this imply that R56 -õ and R51 -o were similar? (Cf. the mixed nasal/oral group in the previous note.)