08.11.8.15:50: THE OTHER O: LFW5958

The tangraph

TT5488 ʔo R51 1.49 (transcription of Sanskrit o?)

=+
TT5487 ʔo R51 2.42 (transcription of Sanskrit o?) =

left of TT5488 ʔo R51 1.49 (transcription of Sanskrit o?) +

left of TT5109 dʒɨo R1.51 'long' (< Chn 長 'long'?)

The graphic analysis suggests that TT5487 is the long version of TT5488 even though their fanqie indicate that they belong to the same rhyme (R51) if tone is ignored:

=+

TT5488 ʔo R51 1.49 (transcription of Sanskrit o?) =

initial of TT5479 ʔa ?R17 (rhyme unknown) 'one' +

final of TT1794 ʔo R51 1.49 'master, owner, possessor'

=+

TT5487 ʔo R51 2.42 (transcription of Sanskrit o?) =

initial of TT5479 ʔa ?R17 (rhyme unknown) 'one' +

final of TT3736 ʔo R51 2.42 'pleasure'

Were rising tone syllables pronounced with longer vowels? (This might imply the following scale of vowel length: o level < o rising < oo level < oo long.) Or were level tone syllables pronounced with final glottal stops?

ʔo 1.49 [ʔoʔ] : ʔo 2.42 [ʔoo]

This latter scenario is unlikely, since level tone is more common than rising tone.

Moreover, some reconstructions of Tangut such as mine have a short/long vowel contrast. If TT5487 were longer than TT5488, why wasn't it ʔoo R54 1.52 or 2.45 (a syllable that doesn't exist) with a long vowel?

Mixed Categories of the Tangraphic Sea defines both TT5487 and TT5488 as Sanskrit (transcription tangraphs). But Sanskrit o can only be long [oo]. Why would Tangut need short and long o tangraphs for Sanskrit? And what were these two tangraphs doing in Mixed Categories if they did not have the MC-only initials dz-, dʒ-, or lh-? Why weren't they placed in the level and rising tone volumes of Tangraphic Sea like most other glottal stop-initial tangraphs?

Grinstead (1972: 77, 87) defined TT5487 and TT5488 as "[ā]" and "[a]", implying they represented Sanskrit long aa and short a. But those Sanskrit vowels already have their own special tangraphs:

TT0544 ʔææ R23 2.20

with 'long' on the right plus 'mouth' from the left of its short counterpart

TT0489 ʔa ?R17 (rhyme unknown)

so why would Sanskrit aa and a need two more tangraphs with different Tangut vowels?

Direct: Skt a > Tangut ʔa

Indirect: Skt a > Late Old/Middle Chinese *ʔa > Tangut period NW Chinese *ʔo > Tangut ʔo

Direct: Skt amida > Tangut ʔamida

Indirect: Skt amida > Late Old/Middle Chinese *ʔamida > Tangut period NW Chinese *ʔombitho > Tangut ʔobitho

08.11.7.23:57: O(BAMA)

The tangraph

TT5488 ʔo R51 1.49 (transcription of Sanskrit o?)

isn't as iconic as Obama's symbol

O

Its analysis is identical to its fanqie:

=+

TT5488 ʔo R51 1.49 (transcription of Sanskrit o?) =

left / initial of TT5479 ʔa ?R17 (rhyme unknown) 'one' +

a distortion of Chinese 阿 *ʔa?

right / final of TT1794 ʔo R51 1.49 'master, owner, possessor'

written as Li Fanwen radical 087 + 'hand'

LFW radical 087 is presumably phonetic (and sometimes semantic) in

=+

TT1794 ʔo R51 1.49 'master, owner, possessor' =

right of TT2386 ʔo R51 1.49 'hang; join' +

(bottom) left of TT1035 phii R14 1.14 'send'

=+

TT2386 ʔo R51 1.49 'hang; join' =

left of TT2393 vəi R8 1.8 'hang' +

left of TT1794 ʔo R51 1.49 'master'

=+

TT4803 ʔo R51 1.49 'rule; own; arrange' =

(cognate to and) left of TT1794 ʔo R51 1.49 'master' +

left of TT2413 vɨi R10 1.10 'do; make'

I briefly wondered if LFWR087 could be derived from part of 於/扵 like the current standard kana for o (オ), but the reading *ʔo would have been uncommon if not extinct for 於/扵 in the northwestern Chinese dialect known to the Tangut. The reading that the Tangut would have known would be something like *ʔju or *ʔy.

11.8.00:18: LFWR087 could be a distortion of Chinese 主 'master'.

(11.8.00:30: Analyses of 'master', hang', and 'rule' added.)

08.11.6.23:54: LFW 5955 1637 3369 (ANSWER)

Congratulations to

Guillaume Jacques

the first (and so far only) person who submitted a correct answer to yesterday's question. If you haven't figured out the answer yet, see if you can get it by looking at the reconstructed readings and meanings of the eight characters in the question:

1.1. TT4348 ba R17 2.14 (first syllable of the Tangut surname babəi; transcription of Sanskrit bha, va)

1.2. TT2425 riaʳ R87 1.82 'permit; allow'

1.3. TT2692 kɨaa R21 1.21 (transcription of Sanskrit ka and ku; more here)

2.1. TT5488 ʔo R51 1.49 (transcription of Sanskrit o)

2.2. TT4348: see 1.1 above

2.3. TT5593 mia R20 1.20 'end (< Chn 末?), tail; east'; transcription of Sanskrit ma; first syllable of mia lhiụ 'banana'

3.1. TT0510 ŋwəʳ R90 1.84 '皇 emperor; heaven'

3.2. TT2125 dzwiə R31 1.30 '帝 emperor'

Here are the readings arranged in a grid:

Is the answer obvious now?

It should have been possible to figure out the answer with the clues provided, some logic, and a lot of luck.

Hint 1: 4 November 2008

... is the date of the US presidential election.

Hint 2: The character that appears twice

... and in two different columns: column 1, character 1 and column 2, character 2. The characters in those columns represent names. The first syllable of name 1 is the second syllable of name 2.

Hint 3: The last two characters technically shouldn't be there right now

... but should be after the inauguration.

Hint 4. The last two characters are a very loose equivalent of a concept that didn't exist in the Tangut Empire

ŋwəʳ dzwiə

is the disyllabic Tangut translation of Chinese 皇帝 'emperor'. There were no presidents in the 11th century.

I preferred to use an attested Tangut word ('emperor') rather than a Tangut neologism for Chinese 總統 'president'.

When I wrote my friend Sarah's name in Tangut as

sie raʳ 'knowledge fountain'

I deliberately chose characters with positive meanings.

However, I wanted a neutral transcription of the name of my new president, so I picked characters from Arakawa Shintarou's kana-to-Tangut conversion table at the very bottom of this PDF.

ばらか・おばま Baraka Obama* > Tangut bariaʳkɨaa ʔobamia

*In Japanese, Barack Obama is バラック・オバマ Barakku Obama in katakana, but I wrote it in hiragana to match Arakawa's hiragana-based table.

Although I could have rendered -kku as

tiə kʊ, lit. 'if therefore'

using Arakawa's table, no Tangut would render foreign -ck as tiə kʊ with tiə corresponding to the Japanese consonant repeat symbol ッ. Moreover, bariaʳkɨaa is close to an actual Arabic word برك baraka cognate to the name بارك Baarak.

08.11.5.23:59: LFW 5955 1637 3369 (QUESTION)

I think you can guess why I took yesterday off. If not, here's the answer (in three vertical columns from right to left):

Try to decipher it. It's possible even without knowing Tangut. Select the 'blank' space below to see a few hints:

1. 4 November 2008

2. The character that appears twice

3. The last two characters technically shouldn't be there right now

4. The last two characters are a very loose equivalent of a concept that didn't exist in the Tangut Empire

I'll post the answer tomorrow.

08.11.3.23:39: NON-UNIQUE VELAR INITIALS IN MIXED CATEGORIES OF THE TANGRAPHIC SEA (PART 2)

Sofronov and Gong's reconstructions of the Mixed Categories tangraph

TT1719 kʊ R4 1.4 'loose' (Sofronov and Gong: ku)

are homophonous with their reconstructions of the Tangraphic Sea level tone volume tangraphs

One might expect all five to be in the same homophone group in Homophones. In fact, in edition A of Homophones, those four are lumped together with three R4 2.4 (rising tone!) tangraphs but not 'loose':

R4 2.4-1.4 (all but 'loose')

This implies that these seven have more in common with each other (in spite of their tonal differences) than with 'loose' which was put into an entirely separate group with R4 2.4 'princess':

'Princess' is in no part of any version of Tangraphic Sea, so I do not know if it would have fit the unknown criteria for inclusion in Mixed Categories.

In edition B of Homophones, groups are split by rhymes. 'Loose' is in its own group while 'princess' is now in the R4 2.4 group:

R4 1.4 group 1 (without 'loose')

R4 1.4 group 2 ('loose' only)

R4 2.4 group (including 'princess')

I would have expected 'princess' to be in its own group like 'loose'. Tone aside, is 'princess' more like 'loose' or like the other R4 2.4 readings?

08.11.2.23:59: NON-UNIQUE VELAR INITIALS IN MIXED CATEGORIES OF THE TANGRAPHIC SEA (PART 1)

The initials dz-, dʒ-, lh- are apparently unique to Mixed Categories* for an unknown reason.

But not all Mixed Categories tangraphs have unique initials.

For example, the level tone

kõʳ ɣɪ̣ 'velar sounds'

section of Mixed Categories contains six tangraphs (MCTS02A31-52) with k- and ŋ- which can be found elsewhere in (Precious Rhymes of the) Tangraphic Sea. You can see the original page on the right side of this PDF.

Tangraph	Tangut Telecode	Reconstruction	Rhyme	Tone.rhyme	Gloss	Initial speller	Final speller	Notes
	5813	kɪ	R9	1.9	discipline	TT5608 kə̣u R61 1.58	TT3409 ɣɪ R9 1.9	borrowing from Chinese 戒
	5072				careful; examine
	2962	kɨaa	R21	1.21	transcription for Skt ka and ku	TT2644 ki R11 1.11	TT1995 khɨaa R21 1.21	usage similar to transcription sinographs 迦, 羯 for Skt ka and 鳩 for Skt ku
	1719	kʊ	R4	1.4	loose	TT4833 kə R28 1.27	TT4725 ɣʊ R4 1.4
	2447	ŋwə	R28	1.27	sleepy; drowsy	TT0525 ŋwə̣u R61 1.58	TT1909 kwə R28 1.27
	3092	ŋwiə	R31	1.30	frightened	missing (accidentally left out?; page does not appear to be damaged); Sofronov (1968 II: 55) and Li Fanwen (1997: 636) have the fanqie TT5398 ŋwəi R8 1.8 + TT5642 gwiə R31 1.30 - where is this from?)		has no homophone in the level tone volume of Tangraphic Sea in Gong's reconstruction

All of the fanqie speller tangraphs are in the level tone volume of Tangraphic Sea. This implies that there is nothing unique about the initials or finals of the readings of the six Mixed Categories level tone velar-initial tangraphs. Gong even reconstructs homophones for the first five which are in the level tone volume of Tangraphic Sea. Yet these homophones are placed into distinct categories in Homophones, implying that they are not homophones. What's going on?

*See Gong, "Phonological Alternations in Tangut", p. 797, on dz- and "The Phonological Reconstruction of Tangut", p. 9, on lh-. I have yet to find a dʒ- tangraph in (Precious Rhymes of the) Tangraphic Sea which is not in the Mixed Categories volume.