One of the Tangut characters for GO in "What's w-ng with Tangut?" looks like it contains LITERATURE plus the mysterious element ヒ which appears on the right side of about 500 tangraphs in Grinstead (1972: 99-106):

< +

TT1957 GO (Gong shywï 1.29)

< TT1955 LITERATURE (Gong 'ywïr 2.77) + 匕

One might think that LITERATURE was phonetic in GO in spite of their mismatched initials (sh- and '- [but d- in the Tibetan transcription dwir of LITERATURE (Nevsky 1926: 15)]) and vowels (LITERATURE has a retroflex vowel and GO doesn't).

However, Tangraphic Sea gives an analysis of GO which does not involve LITERATURE or clearly account for its right-hand element:

< top of (and perhaps its right, though this is not stated? - why not 'frame' which would imply the top and the right?) REGION (Gong kiẹy 2.53) +

'frame' (implying more than its right, though this makes no sense) of CROSS (Gong dzyịy 1.61)

At least Tangraphic Sea confirms that LITERATURE is an element in

TT3125 SCREEN (Gong 'ywïï 1.32)

LITERATURE 'ywïr 2.77 is surely phonetic in SCREEN 'ywïï 1.32, but the function of the left side of SCREEN - shared with TT3119 FIVE ngwə 1.27 and said to be from

TT3124 TRIGRAM (associated with wind; = Chn 巽; Gong me 2.7)

- is unknown.

There are no other 'literary' characters.

TT1956 POND (Gong jiəy 2.36)

looks like TT1957 GO shywï 1.29 but of course doesn't sound or mean anything like it. And strangely it doesn't contain the semantic element for WATER - just like TT5280 SEA in the title

TT1955-5280 LITERARY [i.e., Tangraphic] SEA (Gong 'ywïr 2.77 ngyow 2.48)


TT1179 RIVER/LAKE (Gong shywa 1.19)

analyzed as left of COLORED + [bottom] right of RIVER + right of HAVING. One might expect its center element from

TT4969 RIVER (Gong mya 1.20)

to mean WATER, but it means PERSON. The top element (meaning unknown) of TT4969 RIVER is said to be from the upper right of

TT4282 WATER (Gong zyïïr 2.85)

whose left element is the semantic element 'water'! GONE W-NG

I think I recently saw someone claim that OC 往 wang' 'go' (from my recent zero-grade series) and its phonetic OC 王 wang 'king' shared a common root. (So were 'kings' those who 'went around'?*) I thought it was too improbable to mention two weeks ago. It's now stuck in my head, though the reference isn't. I can't find the article or book in which I saw this.

While looking around for the two wang-s that don't make a right, I rediscovered yet another unlikely etymology for OC wang' 'go'. EG Pulleyblank (1996: 181) thought that it was cognate to the Proto-Indo-European root gwem 'go'. He worked around the mismatching initials by deriving PIE gw- < earlier w-. But PIE also had a w-, so why wasn't that the initial of gwem 'go'? Because PIE w- < earlier xw- (1996: 167). There is no IE-internal evidence for either of these shifts which are necessary to link OC wang to PIE gwem. Pulleyblank interprets the vowel traditionally reconstructed as PIE e as a, so it matches OC -a-. He offered no explanation for the mismatch of OC -ng : PIE -m, though he could have posited a labiovelar ngw: earlier √?w-ngw > OC wang', but PIE gwem. It might be sort of convincing if PIE had a final laryngeal corresponding to the final glottal stop -' of OC, but it doesn't.

Although Gong Hwang-cherng, unlike Pulleyblank, is not an advocate of a Sino-Tibetan-Indo-European relationship, his OC reconstruction of 'go' as gwryang (1996: 58) has a gw- that matches the initial of the PIE root (though the remaining -ryang still doesn't match). Gong reconstructed OC -ngw but believed that this became -n, not -m (1996: 59), though this seems phonetically improbable to me.

So far, the only likely non-Sinitic cognates of OC wang' 'go' are in Tibeto-Burman: e.g., Written Tibetan Hong 'come', song-ba 'went' (past of Hgro 'go'), and perhaps yong 'come' (all from a root ?√wang; earlier -wa- > WT -o-). (Cf. how Eng come [but not go!**] is related to Skt gam 'go'.)

However, borrowing of such a basic verb cannot be ruled out: e.g., Vietnamese lại 'come' appears to be a loanword from an otherwise unattested Middle Chinese lay' 'come'. (The attested MC word 來 lay 'come' [borrowed into Vietnamese as lai] lacks the final glottal stop or creaky voice indicated by the final -'. This -' is reflected in modern Vietnamese as the nặng tone indicated by a subscript dot.)

Next: What more could Hgro w-ng?

*Here's a similar pseudoetymology of Greek basileus 'king' (real etymology unknown; borrowing from a pre-IE language?): it's 'from' Greek basis < PIE gwm-ti (zero-grade root 'go'; cf. Skt gati 'going').

**The actual Sanskrit cognate of Eng go is haa (< PIE ghee) 'leave', not gam 'go'. English g- corresponds to Sanskrit (g)h-, not g-. For the semantics, cf. 去 OC khaps 'leave' > Md qu 'go'. FAILING TO MAKE THE GRADE: -N-OTHER -N-ALYSIS

The title refers to one of four possible solutions to Old Chinese (OC) non-nasal-final syllable ~ nasal-final syllable sets. (I asked readers to devise two solutions in addition to the zero grade solution. I've added Gong's for completeness.)

OCZero-grade analysisSuffixal analysisDialectal analysisGong (1996)-style reconstruction
-a (~ -am)< earlier zero grade
(no vowel + final nasal of root pronounced as syllabic nasal)
root-final vowel -a< earlier nasal vowel from earlier vowel-nasal sequence:
-am, -an, -ang > -ã; later denasalized to -a
-ag (not -ab, which he might reconstruct for other purposes)
-a (~ -an)-ag (not -ad, which he reconstructs for other purposes)
-a (~ -ang)-ag
-am (~ -a)< earlier non-zero grade
(vowel + final nasal of root)
-a + suffix -mvowel-nasal sequence that did not shift to a nasal vowel-am
-an (~ -a)-a + suffix -n-an
-ang (~ -a)-a + suffix -ng-ang

The zero-grade analysis is modelled after the development of Proto-Indo-European syllabic m and n in Sanskrit: zero-grade m and n became Skt a, but m and n preceded by vowels were retained. This analysis has many problems, including the fact that OC non-nasal-final syllables with vowels other than a are also associated with nasal-final syllables (see here for examples with -m).

The suffixal analysis treats the non-nasal-final forms as basic and the nasal-final forms as derived. However, it is not clear what the functions of the nasal suffixes are. (The zero-grade analysis avoids this problem by assuming that the nasals are integral parts of the root, though the rhymes of zero-grade forms have lost their nasality.)

The dialectal analysis assumes that the nasal-final words were original and that the non-nasal-final words were borrowed from dialects which had developed nasal vowels from earlier vowel-nasal sequences (e.g., like French or, more relevantly, modern Taiwanese). Although dialect borrowing occurs (explaining f ~ v in fox [m.] ~ vixen [f.]), there is no evidence to indicate that, for example,

OC 語 nga' 'talk' (< from a dialect with ngã' < earlier ngan'?)

OC 于 wa 'go' (< from a dialect with < earlier wang?)

were from one dialect and

OC 言 ngan 'words'

OC 往 wang' 'go'

were from another.

Gong (1996) has revised Li Fang-kuei's (1971) OC reconstruction, retaining its total absence of open syllables. All OC syllables in Gong's reconstruction end in stops (nasal or oral), liquids, or -s. (I am ignoring the final -x which is his notation for later 'rising tone' syllables. I do not know how he interprets that symbol phonetically.) My OC -a corresponds to his OC -ag. He reconstructs my -a ~ -ang as -ag ~ -ang. One might then expect -b and -d in his system for sets with -m and -n. However, he uses -d for other purposes, so my OC -a ~ -an sets are equivalent to his -ag ~ -an (< -ag-n?) sets. I know of no definite -a ~ -am sets, and I do not know if Gong reconstructs -ab in OC. He does interpret what I see as OC -u ~ -əm sets (examples here) as OC -əb ~ -əm sets. He might explain an -a ~ -am set in my reconstruction as an -ag ~ -am (< -ag-m?) set in his reconstruction. I do not know if he has proposed nasal suffixes or rules to simply final stop-nasal clusters.

Here are the four words from the previous post in terms of the four analyses:


Zero-grade analysis

Suffixal analysis

Dialectal analysis

Gong (1996)-style reconstruction

wa 'go'

< w-Ø-ng

< wa

< < wang




< w-a-ng-'

< wa-ng-'

< wang-'


(< gwryag-n-x?)

wən 'say'

< w-ə-n

< wən or wə-n

same as suffixal?


wat 'say'

< w-Ø-n-t*

or w-a-n-t

< wan-t or wa-t

same as suffixal,

or < wãt < wa(-)n-t


(< gwryan-t?)

rootsw-ng 'go', √w-n 'say'wa 'go', √w- 'say'; or w- for both (if Pulleyblank [2000: 41] is right to link them in spite of their semantics)?wang 'go', √w-n 'say'gwryag 'go', √gwry-n 'say'

Next: Why do the Gong-style reconstructions have gwry- corresponding to my w-? What could Hgro wang?

* wat < w-Ø-n-t has parallels in Sanskrit:

Skt gata- 'gone' < Proto-Indo-European gw-Ø-m-to- and

Skt hata- 'killed' < Proto-Indo-European gwh-Ø-n-to-.