Two posts ago, I asked why a variant of 圮 Late Old Chinese *bɨəʔ (not *bɨəjʔ) 'collapse, overthrow, destroy'

䤏

consists of

According to Shuowen,

圮或从手从非配省聲

lit. '圮 or from 手 from 非 配 abbreviate phonetic'

'[The variant 䤏 of] 圮 may also be from 手 and the abbreviated phonetics 非 配.'

手 'hand' presumably represents the means of 圮 collapse and destruction.

The phonetics 非 *pɨəj and 配 *phəjh imply that 圮 LOC *bɨəʔ also once had a *-j. Unfortunately, I know of no Early Old Chinese rhyming evidence for a *-j in 圮. Wait, I think I see what's going on ...

I followed Schuessler (2009: 104) by reconstructing 圮 as LOC *bɨəʔ. However, I now realize this was a mistake. In Middle Chinese, 圮 was *bɨiʔ which has two possible sources, LOC *bɨəʔ and *bɨəjʔ. The second LOC reconstruction must be the correct one since it has the *-j found in 非 *pɨəj and 配 *phəjh (and 妃 LOC *pɨəj sharing the same phonetic). The development of

圮 Md pi < MC *bɨiʔ < LOC *bɨəjʔ < EOC *brəjʔ < ?*r-bəjʔ

would then be parallel to

悲 'sad' Md bi < MC *pɨi < LOC *pɨəj < EOC *prəj < ?*r-pəj (possibly cognate to its phonetic 非 *pəj 'is not' - is absence sad?)

which is less controversial.

I also discovered a synonym and probable cognate of 圮 whose phonetic 配 LOC *phəjh and readings mostly indicate *-j:

The *-t readings could reflect a suffix added to a root *phəj. However, if such a suffix existed, I would expect to find the following MC/LOC alternations, though I can't think of any examples:

*-V ~ *-Vt

*-Vw ~ *Vt

*-N-t > *-Nʔ

*-C-t > *-C (if *C = *p, *t, *k, *wk, *ʔ; also *-w?)

*-s-t > *-ts

In my last post, I asked,

How can 己 ['self'] stand for both K- and P-type syllables?

Here are three solutions:

Solution 1: P- is from Early Old Chinese (EOC) labiovelars:

Late Old Chiense *ph- < EOC *khw-

Late Old Chinese *b- < EOC *gw-

This is phonetically plausible because it has known parallels: e.g., in P-Celtic and Greek. Within Chinese itself, *khw- became f- in Cantonese: e.g., 快 Ct faai corresponding to Mandarin kuai [khwaj] 'fast'.

Problem: Labiovelars are already reconstructed for EOC, as sources of later velars (not labials). Moreover, EOC labiovelars and EOC velars are in different phonetic series.

Solution 2: P- is from EOC *PK-sequences:

LOC *ph- < EOC *pk-

LOC *b- < EOC *pg- or *Npk-

This is reminscent of how Korean ph- might have come from *pk-, as discussed in previous entries.

This solution accounts for why 己 never appears as a phonetic in LOC syllables with unaspirated *p-.

Problem: It does not account for the fact that the K-syllables and P-syllables belong to different rhyme categories that do not normally coexist within the same phonetic series. (Final glottals are ignored below.)

K-syllables: rhymes without *-j

己紀記忌: LOC *-ɨə < EOC *-ə

改 LOC *-ə < EOC *-ə

P-syllables: mostly rhymes with *-j:

配 LOC *-əj < EOC *-əj

妃 LOC *-ɨəj < EOC *-əj

but 圮 LOC *-ɨə < EOC *-ə (same category as the K-syllables!)

Solution 3: 己 'self' is not always itself. An early form of 配 has a completely different phonetic which is not 己. Early sinography had two phonetics, 己 for K-syllables and 㔾 for P-syllables. Over time, 㔾 came to be confused with the more common phonetic 己. Yet in later periods, the P-syllables still have variant sinographs with 㔾 which is never used for K-syllables.

(12.5.12:15: I was wrong about 㔾 not being used for K-syllables. Today I discovered 紀 with 㔾. But I couldn't find 㔾-variants of 己記忌改 in 異體字字典.)

It is still strange that 圮 < 土+㔾 has an Old Chinese non-*j reading unlike 配 and 妃.

Note that there is another unrelated 㔾 phonetic for *Pam-syllables:

Last night, I mentioned how

Li Fanwen radical 59

appeared in tangraphs with three types of readings.

Li Fanwen (1997) translated

TT1654 kõ R56 1.54 'collapse'

as 圮 in Chinese. 圮 itself has a phonetic 己 with two types of readings:

I'm out of time, so I'll leave you with a couple of questions for next time:

1. How can 己 stand for both K- and P-type syllables?

2. Why does a variant of 圮

䤏

consist of

䤏 is like a worst-case Tangut character. The functions of its components are opaque.

09.12.2.23:59: RIDDLE OF THE RAISED RIGHT HAND

Having just mentioned 提 raising, I'm puzzled by a rare tangraphic element that looks like a man raising his right hand. This element appears with only slight variations

1. on the left of

TT1654 kõ R56 1.54 'collapse'

2. on the right of

TT0499 ka R17 1.17 second half of gɨəə ka 'distribute food'

I suspect that the angle of the 'upper arms' is irrelevant.

What sound or meaning does 'raising right hand man' represent? The Tangraphic Sea analyzed them as

=+

TT1654 kõ R56 1.54 'collapse' =

left of TT5383 ka R17 1.17 'collapse' (which can form a redundant compound with kõ: kõ ka 'collapse') +

bottom of TT2929 ləu R2 2.1 second half of kə ləu 'luck'

=+

TT0499 ka R17 1.17 second half of gɨəə ka 'distribute food' =

left of TT0476 tị R70 1.67 'eat; food' _

right of TT5383 ka R17 1.17 'collapse'

These analyses imply that the mystery element

and its diagonal 'upper arm' variant are variants of

Li Fanwen radical 59

which is a phonetic for ka in TT0499 and TT5383 (but not in other tangraphs!*). This equation is confirmed by a variant of TT1654 with LFW radical 59 on the left. Why does LFW radical 59 have three variants? And what is LFW radical 59 doing in TT1654 kõ whose reading doesn't end in -a? Although the answer to the first question eludes me, I suspect that this element is in TT1654 kõ because it is also in the second half of

kõ ka 'collapse'

But that begs a third question: why don't both halves of kõ ka have the same version of Li Fanwen radical 59 on the same side?

12.3.0:45: The Mojikyo font has a tangraph

which resembles TT1654 but lacks a horizontal top right stroke. I cannot find this tangraph anywhere else.

Moreover, the above tangraph is numbered 574023 in Mojikyo as if it were from LFW4023 but in fact it corresponds to no tangraph in LFW 1997.

The actual tangraph for LFW4023 resembles TT1654 but with the basic form of

LFW radical 59

on the left.

LFW radical 59 resembles Chinese 巾 'kerchief' which would have been pronounced *kĩ in the northwestern dialect known to the Tangut. I doubt there is any relationship between LFW radical 59 and 巾 since their phonetic values only share a consonant in common. Nishida (1966) has no gloss for this radical. I wonder if it is a distortion of the left side of Chn 加 which was pronounced *ka in Tangut period northwestern Chinese.

*12.3.1:16: Other tangraphs with LFW radical 59 have readings wholly unlike ka:

59 on the left:

TT1652 vɨuu R7 2.6

TT1653 tʃɨa R19 2.16

TT1655 tʃɨa R19 2.16

TT1656 va R17 1.17

TT4060 biuu R7 2.6

̣̀59 on the right:

TT4000 biuu R7 2.6

TT4726 tʃwɨa R19 1.19

All readings of LFW radical 59 tangraphs other than TT1654 kõ and TT1656 va belong to three types:

1. the ka-type

2. the biuu-type

3. the tʃɨa-type

Which English word for a naval rank is of Arabic origin? Answer here.

In Japanese and Korean, this rank is called

提督

Jpn teitoku

Kor 제독 chedok

lit. 'raise oversee'.

Vietnamese has a similar term

都督

đô đốc

'capital oversee'

These terms were originally Chinese imperial military ranks. I don't know how these terms came to be associated with navies. Could that have something to do with the 水師提督 'Water Master Lifter-Overseers' of Fujian, Canton, and the Yangtze River?

I don't know of any Chinese term for 'admiral' that doesn't incorporate a term for 'general'. Japanese and Korean also have similar 'general'-based terms for 'admiral'.

After explaining the pronunciation of colonel, I was asked about the British pronunciation of lieutenant as 'left-tenant'. I initially guessed that it had nothing to do with left or leave.

I am wary of folk etymologies. If a modern word X looks like modern words A + B, but has a meaning that isn't the sum of its apparent parts, then it may actually be from an earlier A + C or C + B or even Y*.

It turns out that lieutenant really is from lieu 'place' + tenant 'holding' and has parallels in Arabic and Hebrew. Steadholder was once suggested as a native English replacement.

But how did lieu end up being pronounced as if it were lef?

My guess was that u > w > v > f before t. I was on the right track. I found an Old French variant spelling luef for lieu in Wikipedia and confirmed it in the Oxford English Dictionary. I presume that some Old French dialect had undergone a shift like u > w > v > f, though I don't know whether this shift was limited to lieu (unlikely) or to certain environments: e.g., in word-final position (cf. Russian final -v as [f]) or before voiceless consonants: voiceless f + voiceless t is easier to pronounce than voiced v + voiceless t).

The Oxford English Dictionary lists many obsolete spellings of the lieu- of lieutenant with f or even v:

leef-, leyf-, lyef-, lyffe-, lief-, lefe-, lyffe-, lieuf-, liefe-, leif-, life-, luf-, luff (with space before tenande (sic)), lef-

leve-, lyve-, live-, leive-, liev-

lu-, leu, leu-, leue-, leuȝ- (with a yogh), lyeu-, lyue-, liue-, lieue-, leaue-, lew-, leiu-, lew-

Although the lieutenant spelling became standard in both British and American English, only the f-pronunciation survived in Britain whereas only the f-less pronunciation survived in the US. The f-pronunciation did exist in the US in the 19th century. According to the OED,

A newspaper quot. of 1893 in Funk's Standard Dictionary says that [lɛfˈtɛnənt] is in the U.S. 'almost confined to the retired list of the navy'.

*This doesn't mean there are no compounds which really are from A + B but have unpredictable meanings. For instance, a redcoat is not a red piece of clothing, but a British soldier: i.e., someone who wore a red coat. This example is from Goldman and Sutherland's Sanskrit (!) textbook (1987: 214). Sanskrit has many similar compounds of the bahuvriihi class. Bahuvriihi itself is an example. It consists of bahu- 'much' + vriihi 'rice' but actually refers to a person who has rice or to bahuvriihi compounds.

Why are real etymologies (e.g., redcoat) better than folk etymologies? The former are (a) simpler and/or (b) take account of historical information absent from folk etymologies: e.g., C elements absent in the modern language. An example of a C element is the were- of werewolf which is from an obsolete Old English word wer 'man' (cognate to virile and Sanskrit viira 'hero') and has nothing to do with the verb were or any other w-r word in modern English other than world (< 'man-old'!).

09.11.29.23:59: SAMUEL E. MARTIN (1924-2009)

Tonight I learned that Samuel E. Martin, the founder of Japanese and Korean linguistics in America, passed away yesterday. His grammars of Japanese and Korean are standard references, his Japanese Language through Time is the best single-volume manual of Japanese language history, and his Korean-English dictionary is still unsurpassed after over forty years. In addition to his descriptive, historical, and lexicographical work, he also wrote textbooks for Japanese and Korean. And outside the JK field, he also worked on ancient Chinese and Dagur Mongolian.

Although I only met him three times (1995, 1998, 1999) and exchanged a few emails with him over the years, his impact on my work has been enormous.

I had the honor of studying under his former student Leon Serafim.

I am too shaken to write anything else ... I'm sorry.