One of the tangraphs related to

məɨ 'heaven'

via the source of its left-hand side

diọ 'repair; build; compensate'

is the right-hand side of 'repair'

giụ 'post; pole; column; assistance; give resistance; oppose'

(The last two glosses are quite unexpected since the other glosses all involve support. Perhap one tangraph represents two homophonous but unrelated roots.)

(The right side of 'repair' and 'post' are different in my Mojikyo font but identical in the fonts used by Li Fanwen 1997 and Han Xiaomang 2004. However, Kychanov 2004 treats them as distinct elements B062 and B065!)

The tangraph for 'post' looks like it could be based on the right side of Chinese


diọ 'repair' and giụ 'post' do not sound alike, so I wonder if 'post' was chosen for 'repair' because

- repair is a form of assistance

- it vaguely resembles the right side 甫 of Chinese

'mend; patch' which are kinds of repairs

This right side 甫 is also shared with - and homophonous with - Chinese

'assist' (cf. Tangut giụ 'assistance')

which vaguely resembles the tangraph for 'repair'.

According to Tangraphic Sea, 'post' is the progeny of


the 'head' (top right) of diọ 'repair' +

the bottom of giụ 'position; post; pillar'

(why is 'person' on the left?)

but I assume the direction of derivation is the other way around. 'Post' is clearly phonetic in 'position' (which has a rising tone unlike 'post' with a level tone), which in turn is said to be the source for the bottom of

tʃɨu 'name of an ancestor; make; fulfill; Indian; Buddhist'

(presumably representing more than one unrelated homophonous root; 'Indian; Buddhist' is probably a loan from the second syllable of Chinese 天竺 'India', itself a transcription of early Iranian Hinduka 'India'; see page 8 of Wolfgang Behr's 2006 handout, a mind-blowing gateway to the universe of Chinese writing and Old Chinese phonology)

whose top is from

tʃɨu 'can only'

which is derived from


the top of tʃɨu 'name of an ancestor; make; fulfill; Indian; Buddhist' +

all of giụ 'post'

Which came first,

tʃɨu 'can only' or tʃɨu 'name of an ancestor' (identical to 'can only' except for the presence of 'person' at its bottom left)?

(4.12.2:08: Although the two tangraphs have different components in the Mojikyo font, Kychanov [2006: 174] classified before under radical B065. Li Fanwen 1997 and Han Xiaomang 2004 have the same component in both tangraphs.)

I am inclined to favor the simpler tangraph ('can only') as a phonetic in the more complex tangraph ('name of an ancestor' - hence the addition of 'person' if the ancestor is a person). THE GRAPHIC KIN OF HEAVEN

In the Tangraphic Sea dictionary, each Tangut character is derived from one or more other characters. Thousands of these derivations interlock to form a complex web. Here's part of the web radiating around 'heaven', the first character of airforcewife's Tangut name:

Arrows indicate that a graph is derived from another graph: e.g.,

'heaven' at the center

is derived from


'repair' and 'finish'

which are below it in the diagram. Eight derivatives of 'heaven' such as

'thunder' (from the Tangut name of Vajras)

are above it.

A bidirectional arrow indicates that two graphs were derived from each other: e.g.,


'can only' and 'name of an ancestor; make; fulfill; Indian; Buddhist' (what a range of meanings!)

Some of the graphs at the edges of this sample are 'parents' of still more graphs not included in this fragment of the web.. WIVES AT THE GATES

The third character of

məɨ gia miaa

'heaven army wife'

airforcewife of SpouseBUZZ

may be an incorrect translation because as far as I can tell, miaa 'wife' only occurs in the expression

zɨ̣ miaa

'husband (and) wife'

Modern Tangut dictionaries translate zɨ̣ as 'husband' and miaa as 'wife' since Tangutologists tend to assume that each Tangut character has its own meaning. However, I cannot find either half used in isolation, so I suspect that zɨ̣ miaa is an indivisible disyllabic word for 'married couple'.

Both halves of zɨ̣ miaa contain the element

Li Fanwen radical 278 'gate'

combined with the left-hand components

Li Fanwen radical 132 'head'

'head' + 'gate' = 'head of a gate (household)' > 'husband'?

Li Fanwen radical 120 (looks like radical 118 'person' but is distinct)

Replacing 'head' with

Li Fanwen radical 68 'leg'

in the second half of 'husband and wife' results in the second half of the disyllabic word

ɣa khiɨ

'wife' (Li Fanwen 1997: 738)

'threshold; doorstep' (Kychanov 2006: 356)

Wives are doorsteps? Ugh!

And what is

Li Fanwen radical 90 'tilt'

doing on the left side of the first half of ɣa khiɨ?

So if I chose the wrong Tangut character for 'wife' in my translation of 'airforcewife', what should the correct character be? I think I should use the second character from my translation of

ʃɨõ gi

'Guard Wife' (analysis of characters here)

gi 'wife' is attested in isolation in Cixiao.

Hence from now on, airforcewife will be

məɨ gia gi

'heaven army wife' THEY GATHER FEAR

The second character of

məɨ gia miaa

'heaven army wife'

airforcewife of SpouseBUZZ

(why 'heaven army' for 'air force'?*)

was analyzed in the Tangraphic Sea dictionary as


top of loo 'gather' + all of kiạ 'fear'

Nishida (1966: ) identified the top of 'army' and 'gather'

as an element meaning 'army'. It may be derived from Chinese 軍 'army'.

The bottom of 'gather' is

dzwio 'person'

which in turn was analzyed in Tangraphic Sea as


top of tshi 'serve' + all of dzwio 'person'

The structure of lo 'gather' may be influenced by Chinese


which was pronounced as *tshy (close to Tangut tshi 'serve') in the northwest dialect known to the Tangut.

*tshy 'gather' has 取 *tshy 'take' (phonetic) on the top and 人/亻 'person' (semantic) on the bottom. (If you can't see 人 'person' on the bottom of 聚 in your font, look at its variants.)

kiạ 'fear' is phonetic in gia 'army'. Although other characters could have been chosen as a phonetic element, kiạ 'fear' may also be partly semantic if it implies that armies are fearsome.

kiạ 'fear' appears to be a compound of


'fear' + dzwio 'person'

The element 'fear' is probably derived from the right side of Chinese 怖 'fear'.

The Tangraphic Sea analyzed kiạ 'fear' as


left of vɨe 'enemy' + right of dziɨ 'fear' (which in turn is a vertical line plus kiạ 'fear')

The order of elements happens to match Tangut object-verb word order: 'enemy fear' = 'fear (an) enemy'.

*4.9.0:36: My Tangut translation

məɨ gia

'heaven army'

of 'air force' is a calque (component-for-component translation) of the Chinese term for 'air force':




In honor of

məɨ gia miaa

'heaven army wife'

airforcewife of SpouseBUZZ

whose birthday was last week and who helped me find this store tonight, I'm going to look at the first Tangut character for her name

which combines the left side of

diọ 'repair'

with all of

to 'finish; go up a mountain'

(cognate to Old Chinese 冬 *tuŋ 'winter' and 終 *tuŋ 'end'?; could the character be a distortion of 冬 'end'?)

What do 'repair' and 'finish' have to do with 'heaven'?

'Finish' also appears in characters like

dii 'thunder'

dzwiu lhiạ 'lightning' (a disyllabic word; note how each half is the mirror image of the other)

and Grinstead (1972: 55) has glossed

as 'lightning', though it never means 'lightning' in isolation. Perhaps I could gloss it as 'atmospheric phenomena' and regard it as semantic in 'heaven'.

However, diọ 'repair' has no semantic or phonetic relationship to məɨ 'heaven'. Its left side

Li Fanwen radical 193

also appears on the left side of

L1932 ʃɛ 1.34 'victory' (borrowed from Tangut period northwestern Chinese 勝 *ʃɨ̃ 'id.'?; the vowel is unusual)

L1959 nie 1.36 'relative'; also written as L0213

L1977 məɨ 1.27 'supernatural being' (cf. its homophone 'heaven')

L2358 rieʳ 1.74 'run quickly'

L2391 məɨ 1.27 (insect name)

L2393 lie 2.35 'destroy; damage'

L2560 jiẽ 2.37 'tent; house'

L2571 kii 1.14 'grind; pestle'

L2716 rieʳ 1.74 'skillful; ingenious' (cf. its homophone 'run quickly' - 'quick-minded'?)

L2974 ŋwəu 2.1 'room'

L2978 tie 2.33 'good'

L3469 sie 2.33 'passion; know' (cognate to Written Tibetan shes-pa 'know'; also in Sarah's Tangut name)

L3470 məɨ 1.27 'sheep-guarding spirit' (cf. its homophones 'supernatural being' and 'heaven')

(L-numbers refer to Li Fanwen's 1997 dictionary. x.y numbers refer to tone and rhyme numbers: e.g., 1.27 = tone 1, rhyme 27.)

Radical 193 does not seem to appear on the right side. This fact may be significant.

Several of the above characters are pronounced . Hence I conclude that radical 193 is a phonetic element in məɨ 'heaven', though I don't understand why a məɨ-character like 'sheep-guarding spirit' wasn't chosen as a source of in the Tangraphic Sea analysis of 'heaven'. Nor do I understand what radical 193 is doing in characters that are not pronounced like məɨ. STONE BASIN SWELLING

TT1071 ɣɨ̣ R72 1.69 'stone'

and the source of its top component

TT1071 ləị R68 2.58 'burn; basin'

which has

TT0571 bi R11 1.11 'third of the ten Heavenly Stems' (associated with fire)

on the bottom are two of only three tangraphs with 工 on top. The third is


TT1072 lhwiuu R7 1.7 'swelling; submit; depressed'

(three unrelated homophones written with one tangraph?) =

top of TT1071 ɣɨ̣ R72 1.69 'stone' +

all of TT3140 liee R40 2.35 'devil; misfortune'

(were swellings demonic? depressed by demons?)

(looks like 'person' surrounded by two 'deaths')

I suppose that the lost analysis of 'burn; basin' is


with 工 taken from either 'stone' or 'swelling' and all of 'third Heavenly Stem' as a semantic element due to its association with fire.

The three 工 -tangraphs have nothing phonetic or semantic in common, so I have no idea what 工 represents. It cannot be related to Chinese 工 'work' which might have been pronounced as ?*kəũ in the Tangut period northwestern dialect. Could it be a Tangut B phonetic? STONE KING

Frequent contributor Mahaadaatṛ (Skt 'great-giver'; he has truly earned his mahaa- 'great') pointed out to me this morning that

ɣɨ̣ nie

'stone king'

for Skt vajra 'thunderbolt; diamond' is a calque of Written Tibetan rdo rje 'id.' (lit. 'stone king').

What other Tibetan calques in Tangut remain undiscovered?

I forgot to mention that Nevsky's dictionary (1960 I: 273) lists dgi, rgiH, rgi as Tibetan transcriptions of 'stone'. Tai (2008: 240) also lists a fourth Tibetan transcription rge.

The d- may signify the 'level' tone if Nie and Arakawa are correct.

The r- is harder to explain because ɣɨ̣ does not have a retroflex vowel. Could ɣ- have sounded like [ʁ]? Tai (2008: 195) only lists three other cases of rg- for ɣ- not following retroflex vowels:

rgu for LFW2763 ɣʊ R4 1.4 (twice); no other Tibetan transcription

rki for LFW2871 ɣɨ̣ R72 1.69 (once; homophone of 'stone'); also transcribed dgi in Tibetan

The most common Tibetan transcription of Tangut ɣ- is dg- (16 times in three handwritings); two other scribes preferred Hg-.

The Chinese transcriptions 乙 (Pearl 13.1.2) and (Pearl 21.3.1; the diacritic is in the upper left, not the bottom left, contra Nevsky and Nishida) indicate a non-stop initial since 乙 was probably *i in Tangut period northwestern Chinese. The box probably signifies 'something like a zero initial, but not zero': i.e., Tangut ɣ-.

I used to reconstruct the rhyme of 'stone' as R72 -ɨə̣/-iə̣ which accounted for Tibetan transcriptions with nonhigh vowels: -e and even -a (of syllables other than 'stone'; see Tai 2008: 224). Maybe I should again reconstruct a schwa in R72.

Nevsky listed rgi-ñe as a transcription of ɣɨ̣ nie 'vajra'. ñe for 'king' matches my nie but not Gong's njij (no mid vowel) or an Arakawa-style nee without a -i- or -j-.

Tangut fonts by Mojikyo.org
Tangut radical font by Andrew West
All other content copyright © 2002-2009 Amritavision