Does Tangut

0281 1tsəɪʳ 'rabbit'

have any cognates?

Before I look for similar words in other languages, let me reconstruct its pre-Tangut form first.

1tsəɪʳ is a Grade I syllable with a 'depressed' vowel. (As far as I know, this interpretation of Grade I is unique to my reconstruction.*) This 'depression' was conditioned by a lost 'depressor' which may have been

- a presyllable with a low vowel: *Cʌ-

- a preinitial with a back consonant: e.g., *q-

The retroflexion of the vowel of 1tsəɪʳ may be from

- a preinitial or presyllabic *r-: *r(V)-

- a final *-r

(Medial *-r- conditions Grade II, not Grade I, unless it is combined with a preinitial/presyllabic *r(V)- and/or a final *-r: *rʌ-Cri(r), *Cʌ-Crir > Grade II Cɪʳ)

So 1tsəɪʳ 'rabbit' is from either

*rʌ-tsi (with a presyllable conditioning both 'depression' and retroflexion)
*Cʌ-tsir (with a presyllable conditionig retroflexion and a coda *-r conditioning retroflexion)

Neither resembles

y-words for 'rabbit':

Written Tibetan yos

Written Burmese yun

l-words for 'rabbit':

Proto-rGyalrong *qa-la (cf. forms in Jacques 2004: 209)

Old Chinese 兔 *hlas (< *qla-s?)

Qiang z-words for 'rabbit':

Mawo Qiang ʁdʒə < ?*q- (Sun 1981: 197)

Taoping Qiang ʑi (Sun 1981: 197)

Ronghong Qiang ʐə-tɕhɑq (Huang and LaPolla 1996: 335)

These z-words are not cognate to Written Tibetan ri-bong 'rabbit' < *ri-ɣ/k/gong (Hill 2006: 97). Matisoff (2004: 9) derived their initial from a palatalized lateral.

I conclude that 1tsəɪʳ is a Tangut innovation.

2.6.00:50: The l-words for 'rabbit' may be cognate to the first half of Tangut

2lɨ̣-2ʔo < *S-lɨH-ʔoH 'rabbit'
Tangut may be from a much earlier *-a (Matisoff 2004).

Li Fanwen (2008: 610) lists no examples of the first half without the second half.

However, the second half can occur without the first in the compound 2ɣew-2ʔo 'jade hare', implying that the second half was once an independent root.

So did early Tangut have three roots for 'rabbit', one with *ts-, another with *l-, and a third with *ʔ-?

Pre-Tangut *ʔoH has the same vowel as the second half of Proto-Tibetan *ri-ɣ/k/gong, but I doubt it's a borrowing from Tibetan since I would expect *ɣ/k/gong to develop into Tangut 1ɣõ with a fricative initial and nasal vowel.

More on this disyllabic Tangut word for 'rabbit' here and here.

*I regard 'depression' as the defining characteristic of all Grade I rhymes. Although Nishida and Gong did reconstruct Grade I i-class rhymes as and -e with 'depressed' vowels, their reconstructions of other Grade I rhymes do not have 'depressed' vowels: e.g., their Grade I u-class rhyme is -u, whereas I reconstruct -əʊ. YEAR OF THE DOTTED BEAST

Over a month ago, I asked,

[C]ould you explain the structure of the graph for 'rabbit'?

The answer according to the Tangraphic Sea is


0281 1tsəɪʳ 'rabbit' =

all of 0558 1neʳ 'wild animal' (semantic) +

right of 3851 1tsəɪʳ 'spot' (phonetic)

Is the right of 3851 really sufficient to indicate the pronunciation 1tsəɪʳ? There are 223 other characters with that right-hand component ('cok' in David Boxenhorn's alphacode) and most were not read 1tsəɪʳ. To put it another way, only 1% of cok-graphs were read 1tsəɪʳ. Moreover, there are only 7 characters pronounced 1tsəɪʳ and out of them only only three (0281, 2107, 3851) contain cok. 

The Tangraphic Sea analysis of 0558 is circular and improbable:


0558 1neʳ 'wild animal' =

left of 0281 1tsəɪʳ 'rabbit' (derived from 0558!) +

left of 3269 1vu (1st half of 1vu 1neʳ 'wild animal'; derived from 0558; see below!)


3269 1vu (1st half of 1vu 1neʳ 'wild animal') =

right of 0558 1neʳ (derived from 3269; see above!) +

all of 2952 1lwɔ̃ 'reckless, depraved, empty, unexpected' (all 'wild'? - cf. the semantic range of wild in English wild animal, wild guess, wild coincidence)

I think 3269 is from 2952 plus

'beast' (alphacode: gen)

which may or may not be from 0558. If 0558 is the source of gen, then the gen in 0558 must be from a source other than 3269 or has no source tangraph.

3269 and 0558 represent the disyllabic word

1vu-1neʳ 'wild animal'

Li Fanwen 2008 does not list any examples of 3269 outside this compound. I conclude that 3269 is a bound morpheme like cran- of cranberry. Like berry, 0558 can be an independent word. 0558 has a range of meanings (Li Fanwen 2008: 95):

1. 'wild animal'

2. 'wilderness' (cf. Chinese 野獸 'wilderness-beast' = 'wild animal')

3. 'hunt' (wild animals?)

4. 'enemy' (< person like a wild animal?)

They all share the common denominator 'wild'.

Li Fanwen also listed two other definitions for 0558, but I don't think they're necessary:

5. 'musk' in 0558-4135 1neʳ-1ʃɐ 'musk' which I regard as a compound 'wilderness' + 'musk' (loan from Chinese 麝 'musk')

6. 'monkey' in 0558-2174 1neʳ-1ba 'wild monkey' which I regard as a compound 'wilderness' + 'monkey'

Li Fanwen lists no examples of 1neʳ meaning 'musk' or 'monkey' in isolation.

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