07.1.13.23:59: MAN IN THE MIDDLE
What is he doing in
TT5232 MIDDLE gu 2.1?
He's 'standing' next to HORSE on the left and an unidentified element on the right (フ+キ). His presence differentiates MIDDLE from
TT5241 CENTER ka 1.17
MAN is an extremely high-frequency tangraphic element which occurs on the right-hand side of c. 350 tangraphs (Grinstead 1972: 137-142) and on the left-hand side of 606 tangraphs (Sofronov 1968 II.351-364). These two figures overlap somewhat due to some MAN ... MAN tangraphs like
TT3745 FORMERLY ryir 2.72
(how do two MEN signify FORMERLY? why doesn't this mean PAIR or COUPLE or even FOLLOW [cf. Chn 从 'follow', a drawing of one 人 'man' behind another.)
TT3357 FLOW/PASS/SIN rar 2.73
TT3466 PIERCE thew 1.43
(looks like BUDDHA with MAN on the right!)
TT3785 SUDDENLY lhywï 2.28
(why does WATER surrounded by two MEN signify SUDDENLY?)
TT3749 SIT dzuu 2.05
(two MEN atop BOTTOM; probably inspired by Chn 坐 'sit' with two 人 'men' atop 土 'earth'; the Tangut word may be borrowed from Chinese: cf. Old Chinese dzoy' 'sit')
The number of tangraphs with a MAN in the middle is unknown. TT5232 MIDDLE gu 2.1 is not alone. Others include
TT2882 YOU nya 2.17
TT3124 WIND (trigram; = Chn 巽) me 2.7
(shares its left side with FIVE, BRAHMA, SILVER-MEASURE, DUST, SCREEN, RUBBISH)
TT3751 SUDDENLY tyịy 1.61
TT4479 BLACK nya 1.21
TT4759 BLACKEN-THE-FACE tsur 1.75
TT0881 WOOD syi 1.11
(two MEN next to an unknown element [一+ヒ] beneath WOOD)
The total number of tangraphs containing MAN may approach 900 - roughly one out of seven tangraphs. None of these tangraphs, with the sole exception of
TT4105 MAN dzywo 2.44
was homophonous with
TT3344 MAN dzywo 2.44
(TT4105 and TT3344 form a homograph pair in the Homophones dictionary [32A72-32A73; Sofronov 1968: II.201])
so MAN could not have been phonetic in those ~900 tangraphs (out of a total of ~6,000). Or was it - if those tangraphs were read in Tangut B? Could it really have been semantic in roughly 15% of the Tangut writing system? For comparison, the Chinese semantic element 人 'person' appears in only 231 (4.24%) of the 5,446 characters in Nelson's Japanese-English Character Dictionary (the only Chinese character dictionary I have on hand with numbered entries). MAN was used 3.5 times more in Tangut than in Chinese. Why?
In Sanskrit, the vowel a is the most common vowel in Sanskrit, accounting for one-fifth of "an aggregate of 10,000 sounds of continuous text, in ten different passages, of 1,000 sounds each, selected from different epochs of the literature" (Whitney 1924: 73). That figure is not too far from the fraction of tangraphs containing MAN. I conclude that MAN in many (not all) cases was a phonetic symbol for a high-frequency vowel like a or schwa.
If the element HORSE was pronounced something like ba (cf. the shape and sound of 馬 mba, the northwestern Middle Chinese word for 'horse') and if (フ+キ) were pronounced as X, then
represented the same Tangut B root 'middle/center' (baX?) with and without a long vowel: baaX, baX. (An infixed vowel is less likely: ba-ə-X, baX). This interpretation suggests that vowel length had a morphological function in Tangut B: cf. Sanskrit pairs like
ashva 'horse' (cognate to equus)
aashva 'belonging to a horse' (< ashva)
maadhya 'middle' (< madhya)
Next: On the HORSE's ryir-ght.
07.1.14.1:14: Earlier tonight I briefly considered the possibility that MAN might be like the Japanese vowel length marker ー:
just as Jpn バー = ba + LONG = baa 'bar' (borrowed from English)
Tangut HORSE + MAN + (フ+キ) = ba + LONG + X = baaX 'middle'
However, Japanese ー never occurs word-initially since it cannot lengthen a zero vowel, whereas Tangut MAN is more common initially (i.e., as a left-hand element) than medially or finally (see above). Thus it could not have been a vowel lengthener in all cases. I considered the possibility that it was semantic in left-hand position and phonetic in other positions, but there are many left-hand-MAN tangraphs whose semantic relationship, if any, with MAN is not obvious: e.g.,
TT3509 FOUR lyïïr 1.92
TT3448 SIX chhyiw 1.46
TT3568 NINE gyïï 1.32
in addition to the examples already given above: FORMERLY, FLOW, PIERCE, SUDDENLY. Thus I believe that left-hand MAN can also be phonetic (from a Tangut B perspective only, of course). The question remains open whether MAN had the same Tangut B phonetic value in all positions.
07.1.11.23:55: IN THE HEART OF A HORSE-MAN
Just as INTESTINES apparently turned out to be INTERIOR, another recently cited graph isn't what it seems to be. It has the following analysis in the Precious Rhymes of the Tangraphic Sea:
TT5232 ? gu 2.1 =
[right of] TT1915 NAVEL?/HEART? tyiy 2.33 +
[left of] TT3366 Skt. g-?/MIDDLE? gyï 2.28 +
[right of] TT5241 CENTER ka 1.17
(I've temporarily hidden the gloss of TT5232, which you can find in the previous post.)
Grinstead (1972) and Shi et al. (2000) glossed TT1915 and TT3366 differently. I have given Grinstead's gloss first, followed by Shi et al.'s gloss.
The right side of NAVEL?/HEART? is, of all things, the element HORSE (Nishida 1966: 372) which appears to be derived from Chn 馬 'horse'. I suspect this may be a Tangut B phonogram for the syllable ma or ba (cf. mba, the northwestern Middle Chinese word for 'horse' < Late Old Chinese mæ'). Therefore the Tangut B readings for TT5232, NAVEL?/HEART?, and CENTER presumably contained the syllable ma or ba. HORSEs have nothing to do with NAVEL?/HEART? or CENTER.
The central element of TT5232 is PERSON. This element is so common that the reader has no way of guessing that it is an abbreviation of Skt. g-?/MIDDLE? instead of some other tangraph with PERSON.
The meaning of the element taken from TT5241 CENTER is unknown (at least to Nishida [1966: 365] and me). The reader would have no way of knowing that the mystery element was taken from CENTER as opposed to other tangraphs like
TT3084 TIGER ryir 2.72
(left-hand element is FEAR [Nishida 1966: 405]; cf. the right side of Chn 怖 'fear')
(borrowed from Old Chinese 虎 hra' 'tiger'?; some but not all -a shifted to -i in Tangut [see Gong in Matisoff 2003: 172] under unclear circumstances; for examples of a-retention, see Gong 1995)
TT1575 TIGER gyï 1.30
(what does the added vertical stroke signify? a Tangut B prefix?)
TT4802 TIGER le 2.07 (calendrical)
(with the mystery element doubled around ONE-OF-TWO)
(cf. Old Chinese 寅 ngələr 'tiger' [calendrical])
TT4303 BOAT jywï 1.69
(with the semantic element WATER on the left)
(cf. Middle Chinese 船 zhwyen 'boat'; the landlocked Tangut are unlikely to have had a native word for 'boat' [then again, they did have firsthand experience with inland bodies of water])
TT4201 CROOKED Gyu 1.3
(the function of the element between the halves of BOAT is unknown)
and others listed in Grinstead (1972: 77) like TAKE, RECEIVE, GRIEVE, RESPECT, HOLD, GRIEVANCE, DEBT, DIE, END, HOLD-IN-MOUTH, CROOKED, EXAMINE, CAST-ASIDE, ARROGANCE, INVESTIGATE, CONSIST-OF, and FEAR. (It is clear that some of these words form semantic groups, and this may indicate that they shared a common root in Tangut B represented in writing as フ+キ with different affixes. These morphological patterns of Tangut B would have no parallel in the tangraphs' Tangut A readings / translation equivalents.)
TT5232 ? gu 2.1 =
HORSE from NAVEL?/HEART? +
PERSON from MIDDLE +
フ+キ? from CENTER
but TT5232 was not necessarily related to HORSE-s or PERSON-s - it meant MIDDLE. I suspect that the Tangut B word for MIDDLE sounded like the Tangut B words for HORSE (ma? ba?) and PERSON combined with the last syllable of CENTER (which I'll call X). CENTER (Tangut A ka 1.17) would have been pronounced as ma or ba + X in Tangut B.
Next: Why the MAN in the MIDDLE? Dissecting the CENTER.
07.1.10.23:59: WHAT'S IN TESTINES?
If we interpret the Precious Rhymes of the Tangraphic Sea entry for IN (from "What Lies With-IN?") using Grinstead's (1972) glosses, we get this bizarre result:
IN = INTESTINES + ENCLOSURE
What do INTESTINES have to do with IN? One could say that intestines are in(side) the body, but so are a lot of other organs. And of course the similarity of in and intestines only works in English. ENCLOSURE might be defensible as a semantic element, but not INTESTINES.
Moreover, INTESTINES looks like a combination of IN + PERSON (no, not 'live' in the sense of 'live performance'!):
But in fact Tangraphic Sea analyzed it as
TT1460 INTESTINES dyi 1.11 =
all of TT1454 IN 'u 2.1 +
center of TT5232 MIDDLE gu 2.1 (Tib transcription bgu [Nevsky 1960 II.288])
Why would IN-MIDDLE indicate INTESTINES as opposed to some other central concept?
Finally, if tangraphy is semantocentric, why doesn't INTESTINES look anything like
TT1260 LARGE-INTESTINES kha 2.14 or
TT1263 SMALL-INTESTINES wyu 2.2
which, not surprisingly, look like each other?
(But they surprisingly contain the GRASS element in the middle! Nishida [1966: 242] identifies their shared left-hand component ㅑ as having to do with ... houses. Why HOUSE-GRASS in LARGE and SMALL INTESTINES? The right-hand side of LARGE-INTESTINES looks like part of these two tangraphs for BIG (TT5660 khwei 2.30 and TT0971 thạ 2.54 [< borrowed from NW Chn 大 tha 'big']), but the left-hand side of SMALL-INTESTINES looks like ... TEMPERAMENT [as identified by Kychanov; see Grinstead 1972: 15] or VOLITION [Nishida 1966: 244]! HOUSE-GRASS-BIG = LARGE-INTESTINES and HOUSE-GRASS-VOLITION = SMALL-INTESTINES!?)
There is a simple answer for these questions: INTESTINES apparently is not INTESTINES. Nevsky (1960 I.228) defines the INTESTINES tangraph as
[Writ.] Tib. nang "the space within a thing, the interior, the inside"
It is now obvious why it consists of IN + MIDDLE.
So where did Grinstead get the idea that it meant INTESTINES? Perhaps it does have that meaning in some text. Unfortunately, it does not appear in the Pearl in the Palm's glossary of human organ names. I wonder if the gloss has something to do with the puzzling two-tangraph word at the end of Nevsky's entry for IN(TESTINES):
[Writ.] Tib. nang grol; probably = nang-khrol - bowels, intestines, entrails [the name of the source text is unclear]
The second tangraph is TT1523 WEST lyi (rhyme unknown). Why would INTERIOR-WEST mean 'intestines'? The near-homophony of west and waste only works in English. Is this a euphemism? Did other organs have directional names? Was there an INTERIOR-EAST, etc.?
Next: IN the HEART of a HORSE-MAN.
07.1.9.23:59: WHAT LIES WITH-IN?
Abstract concepts like IN are difficult to draw. IN isn't associated with one particular object inside another.
The Chinese character for 'in'
in fact represents a single object - a crow (albeit distorted beyond recognition in its modern form). Why? Because the Old Chinese words for 'in' and 'crow' were nearly homophonous: 'a and 'a (phonetically [`a]?)
Similarly, one could draw an inn and use that as a representation of the English word in. (Of course, the pun would make no sense in other languages lacking that specific pattern of homophony.)
How did the Tangut tackle the problem of in-lustration? One tangraph for IN mentioned in the previous post does not look like a drawing of one thing inside another, but its Precious Rhymes of the Tangraphic Sea analysis contains a tangraph for INSIDE:
TT1454 IN 'u 2.1 =
[left of] TT1460 INSIDE dyi 1.11 +
[bottom right of] TT0924 HALL kụ 2.51
(Precious Rhymes of the Tangraphic Sea does not tell us exactly which parts of INSIDE and HALL were used for IN, but it's obvious. The Tangraphic Sea entries for this tangraph is presumably in the missing volume for second ['rising'] tone tangraphs.)
The glosses for the component graphs are taken from Shi et al. (2000: 199). Grinstead translated INSIDE and HALL as INTESTINES (well, it does have an in ... in English!) and ENCLOSURE. ENCLOSURE might have a semantic function, but the same cannot be said of INTESTINES which does not sound like IN (at least in 'Tangut A').
Next: What's in-testines?
The title refers to a link in a tangraphic chain.
Last time, I cited the Tangraphic Sea's derivation of BRIGHT from PERCEPTION + FACE.
The TS gives this somewhat circular derivation of PERCEPTION that fails to elucidate what the element (ソ+刂) represents:
TT2257 PERCEPTION byu 1.2 =
left of TT2245 BRIGHT swew 1.43 +
right of TT1369 UNDERSTAND mey 1.33
UNDERSTAND has the negative element 刂, so one might expect it to be composed of NOT (刂) plus part of IGNORANCE. In fact, its derivation takes us full circle:
TT1369 UNDERSTAND mey 1.33 =
left of TT1454 IN 'u 2.1 +
right of TT2257 PERCEPTION byu 1.2
It is an amazing coincidence that English in can be a negative prefix (cf. NOT [刂]!) as well as a preposition. Tangut IN is a postposition, not a preposition: Eng in X would be equivalent to Tangut X IN. (See Nevsky 1960 I.227-228 for examples.)
The phonetic resemblance of IN 'u 2.1 to the Classical Chinese locative prepositions (於 于 乎) pronounced as yu, yu, and hu in Mandarin is also coincidental, as those forms are from Old Chinese 'a (phonetically [`a]?), wa, and wa.
Next: What lies with-IN?
07.1.7.3:23: PERCEPTION + FACE =
Can you guess?
Some tangraphs look like obvious semantic compounds: e.g.,
TT4106 RIDE dzeey 2.34 =
bottom (sic!) of TT4105 PERSON dzywo 2.44 +
all of TT5233 HORSE ryiry 1.74
which could be interpreted as a man atop a horse. Without resorting to the Tangut B hypothesis, it is impossible to interpret RIDE as a phonetic compound since RIDE sounds nothing like PERSON or HORSE or a combination of the two (e.g., dzyiry 1.74 or rywo 2.44).
Most tangraphs are believed to be semantic compounds, though they are less obvious than RIDE because they consist of combinations of parts rather than wholes. So we'd expect
TT2245 ? ? ?.?, analyzed by the Tangraphic Sea as
left of TT2257 PERCEPTION byu 1.2 +
center of TT3370 FACE lhywịy 1.61
to have something to do with perceiving faces: RECOGNIZE, perhaps. We wouldn't expect TT2245 to sound anything like PERCEPTION or FACE (or a combination of the two) unless it were somehow derived from one of the words via affixation or vowel and/or tone change.
In fact, we have seen TT2245 before. It was pronounced swew 1.43 which is nothing like PERCEPTION byu 1.2 or FACE lhywịy 1.61. It meant BRIGHT, which does not immediately bring PERCEPTION or FACE to mind.
However, Nishida (1966: 361) translated TT2257 PERCEPTION as 聰い satoi 'clever'. The association of 'clever' and 'bright' works not only in English, but also in Chinese: e.g., Md 聰明 congming 'clever' (lit. 'intelligent' + 'bright'). If that Chinese compound existed a millennium ago when the Tangut script was invented, it could have influenced the structure of TT2245 CLEVER (?).
But even that contrived explanation doesn't account for the function of FACE. The element extracted from FACE has been interpreted as either FRONT (Nishida 1966: 244) or LIGHT (Kychanov cited in Grinstead 1972: 15). If BRIGHT is a semantic compound, it presumably consists of part of CLEVER plus the LIGHT (?) element. The author of Tangraphic Sea may have simply chosen FACE as an example of a tangraph containing the LIGHT element (meaning FRONT in FACE? - did it have two functions? - cf. the Chn element 月 'moon' / 'meat').
TT2245 BRIGHT swew 1.43
vaguely resembles its Chinese translation equivalent 明 'bright'. Rearranging the left-hand strokes results in 日 'sun' (the left side of 明 'bright') and with a little imagination one can see 月 'moon' in BRIGHT/FRONT. (And Chinese 前 'front' also contains the 月 element!) This probably means nothing because Tangut and Chinese use the same limited stroke repetoire, entailing vague, coincidental partial resemblances of elements.
Nishida has no gloss for the left-hand element of BRIGHT. Removing the ソ on top results in the negation element of
TT1491 NOT myi 1.11
TT1348 NOT-TO-BE nya 2.18
TT1469 NOT myịy 2.54 (presumably derived from the same root as myi 1.11)
(But not all negative characters have the negation element, which is not what we would expect from a semantocentric script.)
The function of ソ is unknown.