According to Tangraphic Sea, the left side of

TT4614 biu R3 1.3 'rely on; according to; rhyme'

from my last two posts is from the right side of


TT0420 ʃwii R14 2.12

(I can't find the right-hand version of 'language' in my font, so I've substituted the left-hand version.)

which looks like 'bird' + 'person' + 'language'. Can you guess its meaning without looking back at my entry from Thursday?

TT0420 does not sound like dʒwɨõ 'bird' or dzwio 'person' or any Tangut word for 'language' that I know of.

If the components of TT0420 are not phonetic, could they be semantic? The center and right components form

TT3845 gwi R11 1.11 'words; say; speech; speak'

so one might think TT0420 meant 'bird language' (bird calls? warbling? speech of a parrot?), but Kychanov (2006: 275) defined it as 'correspond to; meet the requirement of something'.

Could TT0420 be a combination of


left of TT0422 phioo R55 2.46 'pair'

(which apparently consists of 'bird' + 'language'; why?)

center and right of TT4640 ŋwe R34 2.30 'consent; agree'

or some other tangraph containing 'consent; agree' such as

TT2066 ŋwe R34 2.30 'by mutual agreement; harmony'

(same morpheme as TT4640; did these two tangraphs represent cognate Tangut B words that were homophonous in Tangut A; if so, then the 'horned hat' top represents a Tangut B prefix)


TT5169 too R54 2.45 'reconciliation; be reconciled with; peaceful'

(with 'military' on top!)

A pair consists of two things that correspond to each other, and meeting requirements is a kind of agreement. In modern Chinese, 對應 'corresponding' consists of 對 'pair' plus 應 'respond'.

'Pair' combines with 'correspond' to form the word

phioo ʃwii 'coordinate; combine; melt'

It also forms the first half of

phioo ŋwe 'join; connect; unite'

with 'harmony' and

phioo sa 'join'

with 'whisper; tell' (why!?). How could 'pair' + 'whisper' = 'join'?

2.1.00:10: 'Whisper' consists of 'language' + (meaning unknown) + 'temperament'. VITAL, YET HIGHLY ORNAMENTAL

The phonetic of

TT4614 biu R3 1.3 'rely on; according to; rhyme'


TT4069 biu R3 1.3 'spine; back of a human body'

The three non-left hand components


vertical line + 'ornament' + 'high'

must stand for biu, though it's not clear why. There is no independent tangraph consisting of these components, and this trio only appears in TT4614 'according to' and TT4069 'spine'. 'Ornament' and 'high' have no obvious semantic relevance to either word.

The Tangraphic Sea analysis of TT4069 'spine' is not helpful:


left of TT4131 ʔəụ R61 2.51 'carry on back; bear burden' +

right of TT4614 biu R3 1.3 'rely on; according to; rhyme'

One might guess that

Li Fanwen radical 005

meant 'back'. Nishida (1966) has no definition for it. Li (1997) lists 58 tangraphs with it on the left side. Do they have any semantic overlap with 'spine' or 'carry on back'? Here are Kychanov's (2006) glosses for the first 25 in Andrew West's index of Li (1997):

LFW2046 raʳ R85 1.80 'old'

LFW2327 biu R3 1.3 'name of a constellation' (homophonous with 'according to' and 'spine')

LFW2334 biu R3 1.3 'womb; baby; infant' (homophonous with 'according to' and 'spine')

LFW2527 nee R38 2.34 'inner imperial city' (homophonous with LFW3729)

LFW2529 piụ R62 2.52 'hat; cap' (homophonous with LFW3406)

LFW2537 riʳ R84 2.72 'stop; rest'

LFW2736 bææ R23 2.20 'the surname Biaa; hemp (< Chn 麻); swear (< Chn 罵); barbarian (< Chn 蠻)' (homophonous with LFW3022)

LFW3022 bææ R23 2.20 'black' (homophonous with LFW2736)

LFW3086 ŋwəəu R5 1.5 'submit to' (cf. 'according to')

LFW3402 ʒɪɪ R13 1.13 'display; think over a matter carefully'

LFW3404 ʃɨii R14 1.14 'doubt'

LFW3406 piụ R62 2.52 'palace' (cf. LFW3422 'headquarters; residence'; (homophonous with LFW2529)

LFW3407 dʒɨõ R58 2.48 'official document; report; dossier'

LFW3409 miu R3 2.3 (a surname; cf. the pronunciation of LFW3412)

LFW3410 lị R70 2.60 'take a rest; idle'

LFW3412 miew R46 2.40 (a name; cf. the pronunciation of LFW3409)

LFW3414 dʒɛ̃ R42 2.36 'be pleased; glad; enjoy'

LFW3416 məəi R12 2.11 'suitable land; vast land'

LFW3422 diə R31 1.30 'headquarters; residence' (cf. LFW3406 'palace')

LFW3554 lɨõ R58 1.56 (surname of the Tangut empress; Chn 梁)

LFW3729 nee R38 2.34 'deaf' (homophonous with LFW2527)

LFW3730 R28 1.27 'chest; trunk' (homophonous with LFW3732 and LFW3739)

LFW3732 R28 1.27 'puppy' (homophonous with LFW3730 and LFW3739)

LFW3736 ʔo R51 2.42 'pleasant'

LFW3739 R28 1.27 'itch; mange; scab' (homophonous with LFW3730 and LFW3732)

Two are homophones of 'according to' and 'spine', but most do not sound like biu (though piụ and miu are close) and almost none have anything to do with 'according to' or 'spine'. What could radical 005 signify in those 25 tangraphs?

1.31.1:48: I have added readings for the 25 tangraphs. Radical 005 is associated with the following recurring syllable types:

1. labial + -iu, -iew

2. rVʳ (iʳ is a brightened version of aʳ)

3. nee

4. bææ

5. dʒṼ


Such a highly polyphonous radical would not be very useful to the reader, particularly since not all readings of radical 005 tangraphs fall into those six categories: e.g., ŋwəəu is the only radical 005 tangraph with initial ŋ-. A HIGHLY ORNAMENTAL WORD (OR, A RHYME TO RELY ON)

TT4614 biu R3 1.3 'rely on; according to; rhyme'

from the last post looks like a combination of


'word' + vertical line + 'ornament' (Grinstead 1972: 28) + 'high'

(note that in Tangut, adjectives follow nouns)

but was analyzed in Tangraphic Sea as


right of TT0420 ʃwii R14 2.12 'correspond to; meet the requirement of something' +

right of TT4069 biu R3 1.3 'spine; back of a human body'

The first seems vaguely semantically relevant, but the second doesn't, at least at first. The second is homophonous with TT4614 biu 'rely on' and hence must be phonetic.

Kychanov (2006: 741) glossed TT4614 biu in Russian as

опираться 'rely on'

полагаться 'rely on'

исходить из 'come from'

(исходить < ис- < из 'from' + ходить 'go')

as well as

в соответствии с... 'in accordance with'

(соответствии 'accordance' < со- 'with' + ответ 'answer')

рифма 'rhyme'

(cf. rhythm < Greek ῥυθμὸς; Russian ф corresponds to Greek θ)

though I have not yet seen it as a verb.

Guillaume Jacques (2007: 115) glossed TT4614 biu as a "[p]ostposition instrumentale ou ablative". This usage presumably derives from a(n earlier?) verb:

'be in accordance with' > 'with' (instrumental)

'come from' > 'from' (ablative)

But why would TT4614 biu also mean 'rhyme'? Although one could regard TT4614 as representing two nonhomophonous words, biu 'with/from' and biu 'rhyme', I think 'rhyme' originated as 'that which [poems, songs, dictionaries] rely upon'.

biu 'spine' could also be cognate, since the rest of the body relies upon the spine.

1.30.2:22: Here are a few examples of TT4614 biu in context from Li Fanwen (1997: ) and Nevsky (1960 I: 559):

noun + biu

tsiiʳ biu

lit. 'law ...'

'in accordance with the law' (Pearl 29.2)

dziẽ biu

lit. 'time ...'

'timely' (in the title of the phrasebook Timely Pearl in the Palm)

phi biu

lit. 'intention ...'

'as intended' (Suvarṇa-prabhāsa-uttamarāja Sutra IX:292)

(Li Fanwen [1997: 143] regards this as an example of verb + biu, but he [1997: 156] and Kychanov [2006: 555] defined phi as a noun.)

thia biu

lit. 'it ...'

'because of it' (Pearl 27.5; cf. the use of the ablative in Sanskrit and Japanese for 'because')

verb + biu

tshiee biu

lit. 'explain ..'

'as explained' (Suvarṇa-prabhāsa-uttamarāja Sutra VIII: 208)

thiə lɨə biu

lit. 'this discuss(ion) ...'

'in accordance with the discussion of this' (Suvarṇa-prabhāsa-uttamarāja Sutra VIII: 210)

(Li Fanwen [1997: 143] regards this as an example of verb + biu, though I think it could also be noun + biu: 'in accordance with this discussion'.) A (B)-ƗU-NIQUE COMBINATION

In the past two entries, I discussed two very similar-sounding Tangut words for 'elephant':

TT0190 bɨu R2 1.2 'large powerful mammal (ox or elephant)'

TT5225 biu R3 2.3 'elephant'

They belong to different rhymes and are hence listed in different homophone groups in Precious Rhymes of the Tangraphic Sea and Homophones. (biu R3 2.3 'elephant' isn't in the homophone group for bɨu R2 1.2 in the level tone volume of Tangraphic Sea, so I presume it is in the missing rising tone volume.)

The Tibetan transcriptional evidence indicates that R2 and R3 were very similar if not identical in the dialect(s) known to the scribes:

Tangraph TT My reconstruction Rhyme Rhyme.tone Transcriptions (frequency) Gloss
2257 bɨu R2 1.2 Hbu (1), Hbyu (1) intelligent; bright
4614 biu R3 1.3 Hbu (18), dbuH (1) rely on; according to; rhyme

Neither word for 'elephant' is attested in Tibetan transcription, so I have cited near-homophones from the transcriptions. TT2257 is completely homophonous with bɨu R2 'elephant' and TT4614 is nearly homophonous with biu 'elephant', differing only in tone. Some R3 transcriptions have -y-, but not with initial (C)b-.

The different medials (-ɨ- and -i-) of R2 and R3 are byproducts of my four-grade hypothesis based on Gong's (1994) three-grade hypothesis. R2 is a Grade III rhyme and R3 is a Grade IV rhyme.

Labials do not normally precede Grade III rhymes, yet obviously bɨu R2 (Grade III) has a labial initial. Stranger still, R2 only combines with b-. The R2 syllables pɨu, phɨu, or mɨu do not exist. I would expect an odd combination to occur only once or twice - perhaps for Chinese and/or Sanskrit transcription - but Homophones lists 9 (!) bɨu R2 1.2 tangraphs and 1 bɨu R2 2.2 tangraph. None represent foreign syllables with un-Tangut phonetic characteristics.

R3, on the other hand, can follow all four labial initials: piu, phiu, biu, miu.

Is it just a coincidence that b- occurs ten times before R2 whereas all other labials never do? Or does this imply that b- was originally something else: e.g., v- which is associated with Grade III. Could b- before R2 originate from a cluster *C-v- that fused into a stop? *v- in turn could be from an earlier labial stop: *CV-P-, so the initials of the bɨu R2 syllables may have undergone lenition followed by fortition.

If b- in R2 really is from *C-v-, I would expect such a cluster to occur before other Grade III rhymes as well. Yet there is no bɨi R10, bɨa R19, bɨə R30 or bɨe R36. It seems very arbitrary to claim that *C-v- hardened before *-ɨu, but not before other rhymes. I remain puzzled. KHA-GDA

If I had time, I'd post more about Tangut syntax. Although I don't specialize in syntax, I would like non-Tangutologist reader to get a feel for the phrase structure of the language. That's why I provided a literal translation of the Tangraphic Sea entry for TT0190 bɨu R2 1.2 'large powerful mammal' in my last post. Here are a few notes mostly pointing out the obvious:

First half of the definition

'bɨu as-for reʳ is lhị̣̣ pure is khiã ŋwo is pə biu is'

tia 'as-for' is a topic marker that follows the topic:

topic (here, bɨu) + tia + comment (i.e., what a bɨu is: a reʳ, a pure lhị̣̣, a khiã ŋwo, a pə biu)

It looks suspiciously like the Old Chinese topic marker 者 *tjaʔ. Although 者 was pronounced *tɕjæʔ in Middle Chinese, I wonder if a more archaic form *tja persisted in colloquial early northwestern Middle Chinese. (Cf. the Mandarin possessive particle 的 de [tɤ] which more or less preserves the Old Chinese pronunciation of 之 *tə, now read as zhi.) I think borrowing is more likely than chance resemblance or inheritance.

Tangut word order is often the reverse of English. So 'is X' is expressed as X lɨə. (lɨə has the same initial as its Old Chinese translation equivalent 也 *ljajʔ, but the finals are quite different. Hence I am hesitant to link them.)

Tangut adjectives normally follow nouns: e.g., 'lhị pure' is translated as 'pure elephant'. However, 'big' precedes biu 'elephant'. I suspect that there was a distinction between

pə biu 'the great elephant of Tangut folklore' (not just any big elephant)

?biu pə '(any) big elephant'

I wonder if Adj-N order was influenced by calques from Chinese like liẹ so 'sun' (lit. 'great male-principle') which is structurally identical to Chn 太陽 'great male-principle'. I presume that ?so liẹ would mean 'great male principle' as opposed to liẹ so 'sun'. If English had a similar distinction, the beginning of the universe would be the Big Bang but a loud explosion would be a bang big. Perhaps I should write words like liẹ so 'sun' as liẹ-so or liẹso to indicate that they are distinct from phrases like ?so liẹ 'great male principle'.

Second half of the definition

'gəuʳ with category similar inside gəuʳ than power big is'

The last part can be translated more or less using what I already said: '[it] is [the case that its] power is big[ger] than [that of a] gəuʳ [= ox].' səu 'than' follows rather than precedes a noun:

gəuʳ su 'ox than' > 'than [an] ox'

Tangut adjectives do not have a comparative form. The presence of su earlier in the sentence is sufficient to indicate that khwe 'big' should be translated as 'bigger'.

The first part looks like '[its] category [is] similar to [that of a] gəuʳ [= ox]'. Note that the original has 'with ... similar' instead of 'similar ... to'. The function of kha 'inside' may be initially puzzling. It cannot be a preposition taking gəuʳ as an object since it is normally a postposition. But how can a verb be the object of a postposition? Perhaps one could regard the verb as nominalized, or regard kha as a verb ending originating from a postposition. Kepping (1985: 156) translated postverbal kha as когда 'when'. The meaning of kha became more abstract:

'inside (a physical object)' > 'inside (a time)' > 'when'


'When its category is similar to that of an ox, its power is bigger than that of an ox'

sounds odd, so I have loosely translated it as

'... similar to the category of oxen, but more powerful.'

linking it to the previous line.

The grammaticalization of kha 'inside' > 'when' is parallel to that of Chinese 中:

'center' > 'inside' (spatial) > 'during' (temporal)

In English, one uses under instead of inside: e.g., under construction. ARE TOP COWS ELEPHANTS?

In "The Year of the Ox *r-ox", I mentioned the Tangut character

TT0190 bɨu R2 1.2 '?'

which looks like a combination of 'top/female' + 'cow':


The English expression top cow comes to mind. But Kychanov (2006: 603) defined it as 'слон; elephant; 象'. How does the Tangraphic Sea dictionary define it?


'?' = top of 'calendrical word for ox' + all of 'regular word for ox'

A bɨu is a reʳ [= large ox], a pure lhị' [elephant], a khiã ŋwo [= large elephant], a pə biu [large elephant in Tangut folklore], similar to the category of gəuʳ [= oxen] [but] more powerful.

'bɨu as-for reʳ is lhị̣̣ pure is khiã ŋwo is pə biu is gəuʳ with category similar inside gəuʳ than power big is'

(Word-for-word translation added 1.27.1:01. Original Tangut texts have no punctuation.)

Maybe a better gloss might be 'large powerful mammal'.

I'm struck by the abundance of words for 'elephant' in the definition: bɨu, khiã ŋwo, and pə biu. None resemble any words for 'elephant' that I can think of. Note that the second syllable of pə biu (lit. 'large elephant') is not bɨu, though the two may be cognate. (The difference in vowels may reflect an earlier difference in affixation.) Many Tangut words with nasal vowels are loans from Chinese, but khiã ŋwo is native. As is often the case with Tangut, I'm left with more questions than answers. THE YEAR OF THE OX *R-OX

Tomorrow (already today by the time I post this) is Chinese (and Tangut) New Year. I've already talked about the Tangut calendrical term for 'ox', so I'm going to look at the regular word for 'ox' and its graph:

TT1389 gəuʳ R80 1.75

It looks like a combination of 刂 'not', an element believed to be derived from Chinese 羊 'sheep', and an E-shaped common yet mysterious right-hand component of unknown function:


The Tangraphic Sea dictionary analyzes 'ox' as


bottom of TT2168 reʳ R77 2.66 'ox'+

bottom of TT0190 bɨu R2 1.2 '?' (I'll discuss this tomorrow)

This is obviously circular (X = part of AX + part of BX) and probably does not reflect the actual structure of the character.

The function of the top element of reʳ 'ox' is unknown:

The top element of bɨu '?' can mean either 'female' or 'top':

The word gəuʳ 'ox' resembles Tangut period Chinese 牛 ?*ŋgɨu 'ox' and could be an loanword. The retroflexion of the Tangut diphthong -əuʳ has no counterpart in Chinese and may be a trace of an earlier Tangut prefix *r- or suffix *-r. The unexpected Grade I rhyme -əuʳ might be due to a low-vowelled prefix that conditioned the partial lowering of original *-ɨu:

*rʌ-ŋgɨu > *rʌ-ŋgəu > *r-ŋgəuʳ > *ŋgəuʳ > gəuʳ R80 1.75

or ?*Cʌ-ŋgɨu-r > *Cʌ-ŋgəur > *ŋgəur > gəuʳ R80 1.75

1.26.1:16: The word gəuʳ may not be a relatively recent Chinese borrowing if is cognate to these forms I found in Guillaume Jacques' (2004) dissertation:

kɤmɲɯ Japhug rGyalrong ŋgorli 'hornless ox'

gSar-dzong Japhug rGyalrong ŋgorlɪj 'hornless ox'

Cog-tse rGyalrong kə-rgú 'ox'

Similar forms throughout Sino-Tibetan may be borrowings from Old Chinese 牛 *ŋʷu (later *ŋʷə) or be genuine cognates of the OC form. Similar words are also found in southern (but not northern!) Tai languages: e.g., Siamese งัว ŋua ~ วัว wua 'cow' (< Proto-Tai *ŋʷue).

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