Whatever happened to the Khitan (a.k.a., Kidan and Qidan)? If things had been different, we would never have heard of 'Mongolians', and the Khitan would still be the masters of the territory occupied by Mongolia and northern China on our world. But instead the Khitan vanished into obscurity, like so many others before them: the Xiongnu, the Xianbei, and many more mostly known by Mandarin exonyms which would have been alien to them.
Or so I thought ... until I discovered this article last night via Wikipedia (specifically, the link labelled "Zu Khitan und Daur"):
At some time in ancient Chinese history a powerful nation - the Qidan - in what is now Inner Mongolia - simply disappeared. Where did these people go? A question that has perplexed scholars for generations may now be solved thanks to DNA.
Researchers from the Chinese Academy of Social Sciences (CASS) and Chinese Academy of Medical Sciences (CAMS) and others from Inner Mongolia Autonomous Region have concluded that the modern-day Daur people have a genetic match to the Qidan of ancient China, making them possible descendants of the Qidan people.
I am not entirely surprised, since the Daur are Mongolic speakers, and the Khitan language was probably Mongolic. However, I am hesitant to conclude that on the basis of DNA from a single (?) corpse of the Khitan ruling clan. (The author of the article may have meant corpses.)
(07.5.19.00:19: 野利任榮 'Yeli Renrong', the inventor of tangraphy, may have been a member of the 耶律 'Yelü' clan [Grinstead 1972: 13]. The transcription 野利 'Yeli' might reflect a Tangutized pronunciation ?*yeri [Grinstead's "Iri"] of the Khitan name ?*yärür transcribed as 耶律 'Yelü'. Gong's reconstruction of Tangut has no vowel like *ü, so Tangut*i might have been an acceptable substitute.)
Interestingly, non-Mongolic speakers in Yunnan, far from the Khitan homeland,
... who previously had claimed themselves descendants of Qidan rather than the ethnic groups such as Blang [speakers of a Mon-Khmer language] and Yi [speakers of a Tibeto-Burman language once thought to be a close relative of Tangut] with which they were categorized on the founding of the People’s Republic of China more than 50 years ago ...
also "have similar patrilineal origins with Daur" according to DNA tests. However, I wouldn't necessarily conclude that they too are descendants of the Khitan, since the Khitan-Daur link is based on matrilineal descent.
I wonder if a lot of Khitan descendants simply regard themselves as 'Chinese'.
07.5.17.21:00: RIVER WEST?
Kepping (2003: 216) regarded Mongolian Qashin 'Tangut' as a "corruption" of Chinese 河西 (Md Hexi) 'River West' (i.e., west of the Yellow River). Is this etymology phonologically sound? Here are various reconstructed pronunciations of 河西 going back to Old Chinese:
*xO si (Old Mandarin A: Zhongyuan yinyun)
*Go si (Old Mandarin B: following Coblin's  interpretation of the readings in Hphags-pa)
*Ga syey ~ syen (Late Middle Chinese; cf. Tibetan transcriptions ga, ha for 河, sye, se, Hshe [sic!] for 西, Khotanese Brahmi sii for 西)
*Ga sey ~ sen (Early Middle Chinese)
*ga(y) sey ~ sen (Late Old Chinese; the -y drops from 'river' in very late OC)
*gay snəl ~ snər (Old Chinese)
How does qashin compare to those reconstructions?1. q: No close matches: i.e., no Chinese *k(h) or *k(h) [q(h)]. For comparison, the q- of qaGan 'khan' was transcribed with *kh: 可. 河 should ideally have been borrowed as Ga or xa.
(I am assuming this q is not from an earlier G. Cf. q in Middle Mongolian qar 'arm', corresponding to Written Mongolian Gar.)2. a: Rules out borrowing from Old Mandarin which shifted *a to *O.
3. sh: An attempt to imitate Late Middle Chinese *sy-?
4. i: Could this be from an earlier *ï? It matches the later vowel *i but not the earlier *e. Contrast with the other vowel a which matches the earlier vowel *a but not the later vowel *O. Presumably both syllables of Qashin were borrowed from a single Chinese original at one point in space and time. Although it is remotely possible that one or the other half of the word was 'updated', I don't believe in forcing etymologies at any cost.*
5. n: This matches an obscure reading of 西 'west' with *-n in Jiyun (1037 AD). The *-n reading goes back to late Old Chinese, as the Baihu tongyi (1st c. AD) paranomastically glosses 西 with 遷 (Coblin 1983: 155, #44) which undoubtedly ended in *-n (and still does in standard Mandarin today: qian). However, I know of no evidence for a final *-n in the compound 河西 'river-west'. There is no guarantee that the *-n reading of 西 'west' was applicable in all contexts.
I wonder if Qashin is actually a Khitan borrowing from Chinese that was later borrowed into Mongolian. Since little is known about Khitan phonology (or about Khitan, period), the unexpected initial voiceless stop might reflect some constraint in Khitan against initial *x- or *G- or a Khitan-internal sound change (*x- or *G- > q-?). The -n may be a Khitan suffix that has nothing to do with Chinese.
*Japanese does have words like 經濟 keizai 'economics' which combine different strata of Chinese borrowing: kei is newer than zai. (The meaning 'economics' is newer still, as the word originally was a classical abbreviation for 經世濟民 'manage [the] world [and] help people'.) The word should theoretically be keisei (both new stratum readings) or kyouzai (both old stratum readings). There are many other old+new strata examples with 經-: e.g., 經常 keijou 'current' and 經度 keido 'longitude'. However, this phenomenon cannot be compared to Qashin (presumably the only occurrence of Chinese 河 and 西 in Mongolian), since many Chinese morphemes in Japanese tend to be read with the reading of one stratum or another, regardless of the stratum of adjacent morphemes. There is no hard rule of 'stratal harmony', though there are tendencies toward it in certain domains: e.g., Buddhist vocabulary tends to consist of combinations of older stratum morphemes. Hence 經 is kyou (older) in Buddhist contexts (kyou means 'sutra') but kei (newer) in lay contexts (as in the words above).
07.5.14.23:59: BUDDHA KHAN?Last night, I asked,
Who was Burqan?
Burqan is the Mongolian word for 'Buddha'. You can see how it is written in the traditional Mongolian script* at the bottom of this page.
The first half bur- northern Middle Chinese 佛 *bvur 'Buddha' (cf. Tibetan transcriptions of northwestern late Middle Chinese 佛 'Buddha' as bur, Hbur, phur and prescriptive early Sino-Korean ppur' [burʔ], now 불 pul). The -r is from an earlier *-t. 佛 *but was an approximation of the Bud- of Buddha. (Tangut tha 'Buddha' appears to be an approximation of the second half!)
The second half -qan is 'khan'. (220.127.116.11:16: Revised the etymology of burqan.)
Apparently the Mongols considered the ruler of the Tangut to be a 'Burqan khan' (Kepping 2003: 219):
Strange as it may be, but even in serious literature on Chinggis Khan one finds a muddle of all known designations for the [Tangut] state. Thus, in the index to the "Chinggis Khan: The Golden History of the Mongols" (London, 1993) we find nearly all the possible names: Xi Xia Kingdom (p. 188), Tangqut clan (p. 187), Qashin, place (p. 186) without indication that all they stand for one and the same state. As a result, one gets an idea that Qashin was a Buddhist tribe whose ruler was Burqan Khan (Buddhist ruler) (p. 136) who had nothing to do with the Tangut state.
Things could be worse. Kepping pointed out that in the video game Age of Empires (presumably this entry in the series?), "Tangut and Xi Xia regrettably stand for different states".
(07.5.15.00:10: Argh, I found this quote in Yu. I. Drobyshev's "Funeral and Memorial Rituals of the Medieval Mongols and Their Underlying Worldview":
the devastated Tangut state Xi Xia
It's like saying "the Japanese state Nihon".)
Next: Why is Qashin the Mongolian name for the Tangut Empire?
*The Mongolian script proper and its Manchu adaptation have always looked like vertical Arabic to me. They are 'relatives' of Arabic, as they are descended from the Syriac alphabet [via the Sogdian alphabet used for Old Uighur].)
07.5.13.22:59: A TAUSEND QUESTIONS
The title exaggerates by ... 994.
Last night, I discovered the Hazara of Hazarajat / Hazaristan (now in Afghanistan).
1. Is the alternate name Barbaristan for Hazarajat real?
The lost Buddhas of Bamyan were in Hazarajat.
Their name sounds like Persian هزار hazaar 'thousand'*.
2. Is the resemblance coincidental?
In any case, there are certainly more than a thousand of them around today. Wikipedia gives a vague range of 3 to 7 million.
3. What is their origin?
Googling for an answer, I found this article providing a long explanation involving ... "[t]he Tang'ud leader Burqan"!? (Relax - the author wasn't claiming that the Hazara are descended from the Tangut.) "Burqan" doesn't look like a Tangut word.
Next: Who was Burqan?
*Cf. Avestan hazangra and Sanskrit sahasra, both 'thousand'. Avestan isn't the ancestor of Persian, but it is a 'grandaunt'.
4. Why does Avestan have -ng- (and elsewhere, -ngh-) corresponding to Sanskrit -s-?
This has bothered me ever since I first looked at Avestan 14 years ago. Back then and even now, I think this phenomenon has something to do with this Sanskrit sandhi rule:
-n before dental t, th, retroflex T, Th, palatal ch, chh > -M + assimilated sibilant (anusvaara followed by s, Sh, or sh): e.g. (examples from Macdonell 1927:18):
patan tarus > pataMs taruH 'falling tree'
chalan TiTTibhas > chalaMSh TiTTibhaH 'moving sandpiper'
hasan chakaara > hasaMsh chakaara 'he did it laughing' (lit. 'laughing did')
All of the final -n above derived from an earlier *-ns. I would be tempted to claim that patan, etc. are underlyingly /patans/, but s-'insertion' even occurs in cases where there was no original *-ns cluster: e.g. (example modified from Goldman and Sutherland 1987: 36),
kasmin + chit = kasmiMshchit 'in some ...'
(kasmin is not from an earlier *kasmins).
The expected Avestan equivalent of Sanskrit -Ms- is -ngh-.
Avestan ng is the closest approximation of vowel nasalization. Cf. the Japanese use of uvular N to approximate Portuguese and French nasal vowels:
Port pão > Jpn paN
Fr chanson > Jpn shansoN
Proto-Indo-Iranian *s generally became h in Avestan but was preserved in Sanskrit. This is why Sanskrit has initial s- in sahasra 'thousand' corresponding to h- in Av hazangra.
For a long time, I've wondered if pre-Avestan speakers somehow mixed up *ngs and *s (or *ngh and *h if the *s > *h had already occurred), just as Sanskrit speakers mixed up original *-ns and *-n.I'm sure that Indo-Europeanists solved this problem long, long ago. I just haven't been able to recreate the solution on my own. For now, I assume something like this happened:
Proto-Indo-European *gheslo- (cf. Old Greek forms with kh- + front vowel + l-, kilo-)
Proto-Indo-Iranian *sa-jhesra- (with *sa- < *sM- 'one' added) >
Pre-Avestan *saza(ng)sra >
Avestan hazangra (after ngsr > nghr >ngr?)
and in the non-Indo-Iranian (but still IE) language Armenian: hazar (borrowed from some Iranian language)
See Pokorny for details.
English thousand, German Tausend, and Dutch duizend - and Russian тысяча - are not cognates of Persian hazaar, etc.
Watkins derived the Germanic words from a form of a Proto-Indo-European root *teuə- 'swell' + 'hundred': a 'thousand' is a 'swollen hundred'. I don't know whether he thinks the Slavic and Baltic words for 'thousand' are cognate or not. At least their initials are what I would expect for cognates:
Eng th- : Dutch d- : Slavic t- : Baltic t-
Notice that I've left out German, which should have d- like Dutch.
5. Why does German Tausend have initial T-?
Is Tausend a loanword from a dialect in which *θ (< Proto-Indo-European *t) became t instead of d as in other standard German words: e.g.,
dann (cf. Eng then)
drei (cf. Eng three)
Ding (cf. Eng thing)
Donner (cf. Eng thunder)
du (cf. Eng thou)
6. Why does English sometimes have voiced ð instead of voiceless θ in initial position: e.g., thou [ðaw]? Is this also due to dialect mixture?