#tangut

[thelovelygods]

[onion-souls]

The Tangut writing system is infuriating.

So in 1036AD or so, Yeli Renrong, a Chinese official of the Western Xia Dynasty was tasked with developing a writing system for the Tangut language.

He made the wrong decision at every single step in the process.

So most people attempting to create a writing system will default to an alphabet, a syllabary (based on syllables or mora, as with two of Japanese’s system), or an abjad (based on consonants, as with the Arabic and Hebrew scripts). But the hero/villain of our story chose to go with a system of logographic characters. And this is perfectly understandable, as he was a Chinese scholar and the Tangut language is a member of the Sino-Tibetan language family.

Using the Chinese ideogram script for Tangut would unify its literature and literate class with the wider Chinese sphere, allowing for readers of this minority language to easily adapt to the greatest body of literature in the world and heaven, albeit after learning the use of certain grammatical auxiliaries and syntax differences.

Yeli Renrong did not do this.

Instead, he developed an entirely new writing system, which looked like Chinese but fractal. Now, Chinese is famous for its complicated system of strokes, but actual characters are quite varied in their stroke count, with most sentences fluctuating in stroke count and form so it’s easier on the eyes and rather beautiful, particularly with the correct choice of calligraphy or font.

NOT TANGUT.

Tangut says fuck that noise and just jams about twenty strokes in every character, so every sentence is an assault on the senses.

There was no reason for this to happen. He could have literally just used the writing system that he knew

[yeli-renrong]

A phonological shared innovation between Tangut and Western Rgyalrongic is initial plosive lenition, most notably of *p- but also of *k- and *q- (and *t- seemingly only in ‘one’ - I haven’t noticed any other examples so far, at least), with AFAIK unknown conditioning.

Some examples: (hopefully this image will work?)

Wobzi, Dgebshes, Stau, and Nyagrong Minyag are Western Rgyalrongic; Japhug, Situ, and Tshobdun are Core Rgyalrong. RLD = Rgyalrongic Languages Database, which describes a slightly different [and tonal] variety of Nyagrong Minyag than Van Way’s dissertation; T1 = the first of two Tshobdun varieties in Hsiu’s spreadsheet. ‘Year’ here is the labial-initial word used in ‘last year’, ‘next year’, etc., not the velar-initial word used [outside Core Rgyalrong, which generalized the labial variant] with numerals; Jacques & Michaud 2011 cites this suppletion as evidence for Burmo-Qiangic.

Another interesting thing here is the vacillation between r- and l- in ‘axe’ - the Tangut form would seem to require *lvi, since *rvi should give **1vir3. Western Rgyalrongic is conventionally assumed to be divided into Khroskyabs (of which Wobzi is a dialect) and Stau-Dgebshes - it’s unclear where Nyagrong Minyag fits into this - so it’s possible that Tangut could be closer to Stau-Dgebshes than to Khroskyabs. I wouldn’t read too much into preinitial vacillation, though... but it’s also notable that the variety of Dgebshes described in Honkasalo 2019 has -əu as its only closing diphthong, which is similar to Tangut’s -ew and seemingly marginal -iw.

[thejackmeister-mongolhistory]

Today, August 18th, is the most commonly given date for the death of Chinggis Khan in 1227. Probably the most popular suggestion for his cause of death is complications from a fall while hunting, but other versions link it to illness or that murdered by a Tangut princess with a 'cleverly hidden' knife. The date as well is variously given over mid to late August 1227. By 1227, Chinggis was in his late sixties and had lived a long and hard life: internal complications from an injury like a fall are a reasonable explanation for his ultimate demise.

Part of the confusion stems from the fact that it seems there was an effort to keep his failing health and death a secret at the end. At that time, the Mongols were in the process of destroying the Tangut Kingdom (modern Gansu, China), and may have worried that news of Chinggis' death would have spurred a greater resistance from the Tangut.

Chinggis was probably buried on Mt. Burkhan Khaldun in Mongolia, according to the most popular traditions, but again, there are other versions. Such as that he was buried in Tangut, or the Ordos region, to a host of other locales. But as with so much of Chinggis Khan's life, it is hard to sift through some of these legends to the truth of history's greatest conqueror.

[ankahi20]

"Party is starting"

Noyan to Tangut, S2 Ertugrul

The Mongols are cray, but THESE SUBTITLES THO

[yeli-renrong]

There isn’t enough pronominal material in Tangut for a strict application of Wurm’s method for Papuan, but for identifying possible descendants of Tangut - and getting a preliminary idea of Tangut’s affiliation (which, given that both Qiangic and Rgyalrongic [Rgyalrongic + Horpa + Khroskyabs] have begun to look debatable, is even more open a question than before) - it may be possible to take a different sample. Numerals seem like a reasonable place to start.

I’d expect a descendant of Tangut to have: - a liquid initial in ‘one’ (not t-); cf. 1lew1 - a nonlow vowel in ‘five’; cf. 1ngwy1 - a nullish or rhotic initial in ‘eight’; cf. 1ar4 - a different tone (and no coronal affricate) in ‘ten’; cf. 2ghaq1

The problem is that there were multiple seemingly unrelated forms of many numerals. tangut.info lists L0332 1ngwyr1, L1423 1ngwyr1, L1347 1kaq1, and L4778 1shaq3 for ‘seven’. The first two could be purely graphic variation, but this still leaves us with three words. (What conditioned the variation?)

Many related languages preserve something t-like in ‘one’ - even as far out as Nuosu cyp, if that’s cognate - and Miyake says the l- initial is from lenition of earlier t-. So maybe numerals are stable enough to be useful.

In the Qiangic languages I’ve looked at so far, the only languages with a liquid initial in ‘one’ are Eastern Geshiza, Stau, and Khroskyabs; these first two fall under the umbrella of ‘Horpa’, and a Horpa-Khroskyabs clade seems to be uncontroversial. Eastern Geshiza is non-tonal, but otherwise fits all the diagnostic criteria above: rəu ‘one’, ŋuæ ‘five’, rjɛ ‘eight’, zɣa ‘ten’. (/æ/ being technically a low vowel isn’t concerning because it appears in three out of four numerals with Tangut y: 1ny'4 ~ wne 'two', 1ldyr'4 ~ wʑæ 'four', 1ngwy1 ~ ŋuæ 'five', 1gy'4 ~ ŋgæ 'nine'.)

Eastern Geshiza doesn’t look like a descendant of Tangut - it has extensive initial clusters and no obvious sign of Tangut-like grades - but could Tangut be ‘Horpaic’, or at least para-Horpaic? On the one hand, the numerals are closer; on the other hand, this seems to disagree with basically everyone. Hsiu in particular categorizes Tangut as a member of the ‘core Qiangic’ linkage of Burmo-Qiangic, alongside Qiang, Pumi, Guiqiong, and ‘Greater Choyo’ (incl. Minyak), even though he claims Horpa-Khroskyabs and Rgyalrong as entirely separate branches of Sino-Tibetan. But he doesn’t say much about Tangut, and it’s not clear to me why he put it there. Wurmian results are extremely preliminary, which is a polite way to say “unreliable”, but the numerals and directionals in Tanggu Mu-nya ( = Minyak) don’t look to me like they’re very close to the Tangut ones.

The Horpa directionals aren’t terribly close to the Tangut ones either, but they seem closer than the ones in Minyak.

Stau: (Jacques et al. 2016)

Eastern Geshiza (Honkasalo 2019):

rə-: away from the river, upward, “north” næ-: toward the river, downward, “south” gæ-: upriver, “west” wə-: downriver, “east”

(The absolute directions are in quotes because Honkasalo explicitly argues against a compass model of directionality in the language.)

Tangut (Jacques 2011):

The directional ablaut between -a/-ɨ and -ij looks similar to the ablaut in Stau, although Jacques doesn’t explain what the difference in use between the two series is.

Stau ‘north’ and ‘south’ seem to correspond to Eastern Geshiza ‘~west’ and ‘~east’, so we should expect a system of pairs of opposites rather than precise correspondence in meaning; and it looks like the vocalism has been leveled by analogy in both languages. This pretty much leaves us with initials - and the Tangut ‘closer/farther’ pair seems to match Stau ‘north/south’ and Eastern Geshiza ‘upriver/downriver’. Tangut ‘down’ also seems to correspond to Stau ‘down’ and Eastern Geshiza ‘toward the river’, but the other side of that pair in these two living Horpa languages seems closer to Tangut ‘neutral’ than to Tangut ‘up’, and there are no obvious signs in Horpa of anything corresponding to the Tangut translocative or cislocative. (Could Tangut ‘up’ be an innovation after a shift of earlier *‘up’ > ‘neutral’? Less plausibly, since the voicing doesn’t match, could Stau ‘no direction’ be related somehow to the ‘missing’ Tangut locative prefixes?)

The linked overview of Stau also points out some similarity to Tangut in the case system:

So the Wurmian case for Tangut as a member of ‘Horpaic’ seems decent. Hsiu disagrees - in his classification, Burmo-Qiangic, Rgyalrong, and “Horpa-Lavrung” (my ‘Horpaic’) are all distinct, and Tangut is a member of the “Core Qiangic” linkage within Burmo-Qiangic (alongside Qiang, Pumi, Guiqiong, nDrapa, Choyu, and Minyak) - but I can’t find anything on his reasoning for putting Tangut there. Jacques has some Tangut-Japhug comparative work which I haven’t read yet (I think it’s in French), and if the correspondences are sufficiently close and sufficiently numerous, something about this scenario (Horpaic as a distinct branch of ST that includes Tangut) may have to give.

[yeli-renrong]

(source)

cf. Miyake’s reconstruction of Pre-Tangut, where medial -r- conditions (IIRC) grade 2

[yeli-renrong]

Words with final -n in Tangut corresponding to -m elsewhere:

1dzon2 'bridge': Japhug dʑom, Geshiza dzo, Stau dzo, bTsanlha tazam 1son1 'clean, proper': Wobzi ɕɑrə́m, Japhug (kɯ) ɕo, Geshiza wɕu, Stau xtsoŋmɑ 2denq4 'cloud': Wobzi zdə̂m, Japhug zdɯm, Geshiza zdo-ma, Stau zdo, bTsanlha zdem 1shon3 'iron': Japhug ɕom 2lon1 'large': Wobzi lə́m, Japhug kɯ jom 2shon2 'otter': Wobzi sræ̂m, Japhug tɕhɯ ɕrɤm, Geshiza sræn (but cf. Tibetan sram) 1ghen4, 1ghon4 'pillow': Wobzi vɣə̂m, Japhug tɤ mkɯm, Geshiza rɣuen 1phon2 'white': Wobzi pʰrə̂m, Phosul pʰrəˠm, Japhug wɣrum, Geshiza pʰru, Stau pʰrupʰru (Lai's Proto-Khroskyabs: *qə̯-pʰˠri)

Words with final -n in Tangut not corresponding to -m elsewhere:

1jwon3 'bird': Geshiza wə-zə, bTsanlha choŋ 2ven2 'fart': gSar-rdzong tɯ phe 2jon'3 'flour': Wobzi jvɑ̂, Geshiza wdʑo (Lai's Proto-Khroskyabs *kə-C.kpæ... huh? possibly a spreadsheet error on my part, I haven’t checked) 1khon4 'give': Wobzi kʰɑ̂, Geshiza s/v-kʰo 1lwon1 'lazy, idle, slow': Japhug kɯ ɣɤ βlo

The spreadsheet is by no means complete, and it’s likely that some of these aren’t cognate. The comparisons without -m overall seem weaker than the comparisons with it. It’s also possible (esp. given ‘otter’) that there are two different lexical strata, one of which consists of loans from Tibetan into Rgyalrongic including Tangut...