21.5.29.17:39: WHITE OX 4.18

𘬈𘭌𘲐𘲂𘭧𘰯𘬌𘲺

? uni ai ? sair par nyêm nyair

'white ox year, four month ten eight day'

1. Back in 1996, Alexander Vovin introduced me to Juha Janhunen's 1994 hypothesis of the Parhae script as the parent script of both the Khitan large script and the Jurchen (large) script.

In March of this year, I learned of Alexander Vovin's "Two Newly Found Xiōng-nú Inscriptions and Their Significance for the Early Linguistic History of Central Asia" (2020) which made me realize that there could be a 'Xiongnuic' or 'Greater Sinitic' family of northern Chinese-based scripts including the lost Northern Wei script as well as the barely attested Parhae script and the much better attested the Khitan large script and the Jurchen (large) script.

And today 戴忠沛 Tai Chung-pui brought to my attention this sample of what appears to be a heretofore unknown 吐谷渾 Tuyuhun script at the grave of 慕容智 Murong Zhi (650-691).

Maybe I am missing the obvious, but I don't see any exact matches with Parhae, Khitan, or Jurchen characters at first glance other than the fifth character in line 1, a lookalike of the Khitan large script phonogram <ha>.

I do, however, think the first two characters at the top right (which recur in the second line on the left) might be equivalent to Chinese 周國 'Zhou state' referring to 武則天 Wu Zetian's Zhou dynasty (690-705). The second of those characters is like the Khitan large script character <STATE> with an extra horizontal stroke.

The first character in the second line is similar to the Khitan large script phonogram <pa> and Jurchen <pa>.

The fourth character in the second line is similar to the Khitan large script phonogram , possibly <ri> or <li>, corresponding to 里 *li in Chinese transcription.

The ninth character in the second line is similar to the Khitan large script character  <gun> used to write the Sino-Khitan equivalent of Liao Chinese 軍 *gün 'army'.

2. How did I not see Arakawa Shintarō's translation of Viacheslav Zaytsev's landmark 2012 paper on Nova N176 until today?

21.5.28.23:59: WHITE OX 4.17

𘬈𘭌𘲐𘲂𘭧𘰯𘲧𘲺

? uni ai ? sair par ? nyair

'white ox year, four month ten seven day'

1. Are there no desiderative verb forms in the Digital Corpus of Sanskrit, or am I doing something wrong? I couldn't find any desiderative forms of √bādh 'drive off' (more on this below), √gam 'go', or √kr̥ 'do'. I also can't find the other secondary verb forms: causatives and intensives.

2. According to Goldman and Sutherland's Devavāṇīpraveśikā (1987: 305-306) the desiderative stem of Sanskrit ā-roots is formed as follows:

• partial reduplication of the root with i replacing ā

• no change in the root itself (apart from sandhi rules)

• suffixation of -sa- or -iṣa-
  

Examples:

• √jñā 'know' > ji-jñā-sa- 'want to know'

• √snā 'bathe' > si-ṣṇā-sa- 'want to bathe'

• s becomes ṣ after i, and n becomes ṇ after ṣ
  

• √khād 'chew' > ci-khād-iṣa- 'want to chew'

• velars reduplicate as palatals
  

So in theory

• √bādh 'drive off' > bi-bāt-sa- or bi-bādh-iṣa- 'want to drive off'

• dh devoices to t before voiceless s

But according to Whitney (1885: 106), the verb has two possible desideratives, the unsurprising bi-bādh-iṣa- and the surprising bī-bat-sa- with a long vowel in the reduplication and a short vowel in the root. Goldman and Sutherland (1987: 308) mention the derived noun bī-bat-sā 'loathing' without commenting on its unusual form.

Another stem with a long vowel in its reduplication is mī-māṁ-sa- 'investigate' (with an idiosyncratic meaning) from √man 'think'. √man has an irregular lengthening of its vowel in the desiderative like a few other an and am-roots, whereas as far as I know the shortening of a in bat is unique.

5.29.0:09: n becomes ṁ before s: hence man + -sa- = maṁ-sa-.

5.29.16:39: Monier-Williams says the desiderative of the desiderative of √man is mi-mā-m-iṣa- 'want to investigate' with a regular short vowel in mi- and the third (!) copy of the root reduced to -m-. How many other Sanskrit verbs have desideratives of desideratives? Is it possible to write rules for the formation of such tertiary forms?

21.5.27.23:57: WHITE OX 4.16

𘬈𘭌𘲐𘲂𘭧𘰯𘭬𘲺

? uni ai ? sair par ? nyair

'white ox year, four month ten six day'

Today I learned about the CJK TV series Strangers6. The Chinese Wikipedia has some strange katakana spellings for the names of Korean  actors:

5.29.16:20: I suspect that in a couple of cases, the katakana was based on English-based romanizations (EBR): e.g.,

• sŏn > EBR sun > スン sun

• su > EBR soo > ソ so

But I can't explain the others.

2. How did the Portuguese man o' war get its name?

21.5.26.23:19: WHITE OX 4.15

𘬈𘭌𘲐𘲂𘭧𘰯𘬦𘲺

? uni ai ? sair par tau nyair

'white ox year, four month ten five day'

Sorry, I fell asleep before I could blog last night.

Today I saw the Google Books preview of Eric C. Rath's Oishii: The History of Sushi (2021).

Rath says the kanji 寿司 for sushi "might mean 'felicitious rule' but instead are used solely for their sounds". I was initially surprised by 'felicitious', as 寿 normally represents Sino-Japanese ju 'long life', but I guess he got that from the native Japanese reading kotobuki 'congratulations'.

kotobuki (less commonly kotohogi) is from Old Japanese kətə-pok-i 'word-pray.for.good.outcome-INF'. Both modern forms are irregular. I would expect *kotohoki. Let's look at the irregularities:

• -b-: rendaku: voicing at a morphemic boundary

• -u-: possibly reflecting raising of pre-Old Japanese *o in medial position

• -g-: I can't explain this
  

I just learned from Wiktionary that Kotobuki as a surname can also be spelled 琴吹 <HARP BLOW>.

Back to sushi: The word also has two logographic spellings, and tonight I learned from Wikipedia that they are regional: 鮨 was the Edo spelling and 鮓 was the Osaka spelling. How many such regional spellings are there? (And has there been any investigation of regional patterns in nôm spelling?)

According to one hypothesis that dates to at least the end of the seventeenth century, the Japanese word 'sushi' was derived from the word sui, meaning 'sour tasting'.

I wonder how many readers would think that sui somehow got a -sh- inserted into it to become a noun. Of course that's not what actually happened. The reality is somewhat the other way:

• the adjective originally had an attributive form su-ki and a final predicative form su-shi (the source of the noun): su-ki X 'sour-tasting X' vs. X su-shi 'X is sour tasting'
  

• su-ki became su-i via lenition

• su-i came to be used for final predication as well as attribution

Rath ends that last sentence I quoted with a reference to a footnote that I can't see in the preview. Perhaps it explains that the noun sushi was from an old form su-shi of the modern adjective su-i.

The noun sushi is somewhat analogous to male -shi names derived from adjectives: e.g., Yasushi < yasu-shi 'at.ease-shi'.

-shi is usually regarded as a final predication suffix, but its early Old Japanese ancestor -si can also be an attributive suffix, and the attributive can function as a nominalizer: e.g.,

na-si-ni

not.be-ATTR-LOC

lit. 'in not being' = 'because there is no' (Kojiki song 23)

na-si there takes the locative as if it were a noun.

For further discussion of Old Japanese -si, see Vovin's grammar (2020: 406-411).

I regard names of the type Yasushi as vestiges of -si as a nominalizer.

Next: The accent of sushi.

21.5.24.22:48: WHITE OX 4.13

𘬈𘭌𘲐𘲂𘭧𘰯𘯙𘲺

? uni ai ? sair par ? nyair

'white ox year, four month ten three day'

Sorry, another interruption in my series on the Khitan small script character 𘲧 <SEVEN>: Last night I found John Kupchik's "Austronesian lights the the way: The origins of the words for 'sun' and other celestial vocabulary in Old Ryukyuan" (2021) which debunks something I believed in for a long time: the derivation of Proto-Ryukyuan *tenda 'sun' (> Okinawan tida) from premodern Sino-Japanese 天道 tendau 'id.'

I already knew about a phonetic problem with that etymology: the irregular, sui generis correspondence of SJ -au to PR *-a instead of PR *-au.

But Kupchik also notes a semantic problem: Sino-Japanese 天道 'heaven-road' did not shift in meaning to 'sun' until the late 16th century, long after Proto-Ryukyuan broke up. So the resemblance between *tenda and tendau is coincidental.

21.5.23.23:56: WHITE OX 4.12

𘬈𘭌𘲐𘲂𘭧𘰯𘲝𘲺

? uni ai ? sair par ? nyair

'white ox year, four month ten two day'

• d-: matches the d- of Khitan 𘳄𘮿 <da.313>  'seventh', Proto-Mongolic *dol/u.xa/n 'seven' and *dal.a/n 'seventy' (Janhunen 2003: 17) and Jurchen dalhon 'seventeen'. Also matches the unaspirated *t- of the Liao Chinese transcription 迪烈 *tiʔliêʔ for the name 𘲧𘰭 <SEVEN.n>. (Liao Chinese had no *d-, so *t- was the best approximation. It is even possible that Khitan d was phonetically [t]; if so, then the Liao Chinese transcription perfectly matched the Khitan initial.)

• -i-: matches the *i of the Liao Chinese transcription 迪烈 *tiʔliêʔ but I would expect -a- on the basis of the other evidence above. (Proto-Mongolic *o in *dol/u.xa/n is due to irregular labial assimilation with the following *u; contrast with *dal.a/n 'seventy' which retains the original *a and lacks a labial vowel triggering assimilation.)
  

• -r: does not match Proto-Mongolic *-l- or Jurchen -l-. Kiyose (1977: 133), however, reconstructs Jurchen 'seventeen' as darhon with *-r-. The Ming Chinese transcription of Jurchen 'seventeen' has 爾 <r> which could represent either syllable-final -l or -r since Ming Chinese had no *-l. See Kane (1989: 115) for examples of 爾 for a Jurchen liquid - presumably *-l - corresponding to Manchu -l.

Next: More on the problem of the vowel of Khitan 'seven'.