07.4.14.16:50: OH DEER - BOO HOO!

Sorry, this isn't part of the Tangut "Oh Deer" series.

I looked up the 夫餘 puyO in the Korean Concise Britannica World Encyclopedia (콘사이스 브리태니커 世界 大百科事典) which gave two possible origins of the name.

The first one

... 평야를 의미하는 벌[伐・弗・火・夫里]에서 왔다 ...

... it came from pOl [伐・弗・火・夫里] meaning 'plain'

happened to more or less match the one I independently proposed in "The Seoul of Paekche":

The name 夫餘 may date back to the 2nd century BC, when it would have been pronounced as *pa la in Old Chinese. I have long wondered if this name was cognate to Old Japanese *para 'field' (and to Paekche 夫里 *por[ï] and/or Korean 벌 pOl 'field'?).

I forgot about another spelling of 夫餘:

扶餘

Old Chinese *ba la

(The earliest attestation I can find is in the Hou Han shu. In HHS times, it would have been pronounced as *buo yïO in Late Old Chinese.)

夫 was read as both *pa and as *ba in Old Chinese.  I chose the more common - and here inappropriate - reading *pa.  The addition of 扌 to 夫 results in 扶 OC *ba (which also has a less common reading *pa 'span of four fingers'). I assume that the choice of a normally *b- initial graph 扶 in the post-OC Hou Han shu spelling 扶餘 indicates that the 夫 in the earlier spelling 夫餘 was to be read with *b- rather than *p-.  I conclude that 夫餘 represented a foreign name like *bala with initial *b- and that 夫餘 was later changed to 扶餘 LOC *buo yïO which maintained the *b- but otherwise lost its resemblance to *bala. (The Chinese recycle transcriptions long after their 'phonological expiration date': i.e., the point at which they no longer sound like the originals.)

This revised reconstruction *bala does not necessarily invalidate the proposed cognates. Old Japanese *para may have come from an earlier *pàrà (with grave accents indicating low pitches).  This in turn may have come from an even earlier *bara, *baara, or *paara. (Voiced initials and long vowels are believed to be sources of low pitch.)  *bara and *baara are very close to 夫餘 *bala.

I discovered yet another spelling of 夫餘/扶餘 which may confirm an initial *b-:

鳧臾

Old Chinese *bo lo

But the earliest attestation I could find was in 李巡 Li Xun's mid-to-late 2nd century AD commentary on the 爾雅 Erya which lists the 九夷 nine kinds of barbarians. (The Tangut use of 'nine' to enumerate the types of barbarians surely originates from this Chinese term.)  If the transcription 鳧臾 were coined during that period, it might have represented a foreign name closer to Late Old Chinese *buo yuo, and the OC readings are irrelevant.

It's not clear whether 夫餘/扶餘 and 鳧臾 represented

- the same foreign name (= modern Korean puyO)

- in the same foreign language (= 'puyOese')

- at roughly the same point in time

In any case, it is unlikely that *bala became *boyo in 'puyOese', just as OC *ba la became LOC *buo yuo. Although 'puyOese' was in contact with Chinese, it probably would not have undergone the same sound changes (though then again, *a > *o and *l > *y are not exotic sound changes and could occur independently).

If 夫餘/扶餘 and 鳧臾 did represent variants of a single name (*bala ~ *bolo? if not a single *bOlO?), the final *-o of 鳧臾 is reminiscent of the rounded final vowel of Okinawan haru < *paru  (or *paro?)　'farm, tract of open country'. It is tempting to link 'puyOese' *bala ~ *bolo with the Japonic forms which also have final -a ~ rounded vowel variation.

(But note that the Japonic forms have the first vowel a, not o. One could try to maintain the puyO-Japonic link by claiming that

- 夫/扶 OC *ba and 鳧 *bo were attempts to write a syllable like *bO which did not exist in OC

- this *bO was borrowed into Japonic as *ba

- or proto-puyOese-Japonic [if such a thing existed] *bO became Japonic *ba)

This all implies that the puyO were the 'field' people.  (Cf. surnames like Field or Camp or Japanese 原 'field'.)  However, there is no ancient account stating that 夫餘/扶餘 or 鳧臾 meant 'field'. This gloss is based solely on the tenuous assumption that 夫餘/扶餘 and possibly 鳧臾 are linked to the Koreanic and Japonic terms for land.

That assumption, as weak as it may be, is still stronger than the second etymology provided by the Britannica:

이름의 유래에 대해서는 사슴을 뜻하는 만중어 'puhu' 와 연결시키는 견해도 있으나...

As for the meaning of the name [puyO], though there is the view that links it to Manchu puhu meaning 'deer' ...

There are two problems with this view.

First, there is no Manchu word puhu 'deer'. (There are very few Manchu words with initial p-  apart from onomatopoeia because earlier *p- became f- in Manchu.)  John Bentley pointed out that the Manchu word for 'deer' is buhû (IPA [buXU]*) with an initial b-.  That actually fits 夫餘/扶餘 and 鳧臾 better, since all three transcriptions were pronounced with initial *b- in Chinese.

Second and much more importantly, the -uhu of buhû comes from an earlier *-ukû [uqU] which does not match the  OC*-a la of 夫餘, the LOC *-uo yïO of 扶餘, or the LOC *-uo yuo of 鳧臾.  There is no other evidence suggesting that puyOese *a corresponds to Manchu u or that puyOese *l or *y corresponds to Manchu h [x] < *k. (The latter correspondence is bizarre. I know of no set of languages with such a correspondence.**)

Thus I don't think the name *bala had anything to do with Manchu buhû, though a connection with Koreanic and/or Japonic words may be possible.

*The IPA value of the Manchu letter transliterated as û is uncertain.  Roth Li (2000: 17) described it as a "[s]ound between o and u".  I suspect it was like the lowered allophone [U] of Arabic /u/ following 'emphatic(-like)' consonants (Kaye in Comrie [1987: 670]) since û appeared mostly after uvulars: cf. Arabic Qur'an [qUr'aan], sometimes transcribed as Koran with an -o- indicating a lowered first vowel.)

**Within Mandarin itself, one can find the triple readings gu [ku], yu [yü], and lu for 谷.  谷 appears as a phonetic in sinographs for both velar and y-initial syllables: e.g.,

Md gu: 唂

Md que < *kh-: 卻

Md yu: 欲慾浴裕峪鵒

Sagart (1999) would explain this variation by reconstructing an *l- with or without preceding phonemes:

OC *lok > 谷 Md yu (in the name 吐谷渾 Tuyuhun)

OC *klok > 谷 Md gu 'valley'

OC *kəlok > 谷 Md lu (in the Xiongnu title 谷蠡 Md Luli < OC *kəlok rey; 蠡 means 'worm-eaten', which shows how little the Chinese respected the Xiongnu barbarians, whose name they transcribed as 匈奴 'breast-slave')

One could try to link 夫餘 *bala with Manchu buhû < *bukû [buqU] by claiming that they both came from an earlier *bakla, but that would still not explain the vocalic mismatch.

07.4.12.6:40: THE SEOUL OF PAEKCHE?

(I'll write "The County Caper" in a couple of days and then return to Tangut.)

The Paekche place name element 夫里 is quite common. Using Ryu's (1983) numbering, it appears in:

302. 古良夫里

Late Old Chinese* *ko' lïang puo lï'

309. 毛良夫里

LOC *maw lïang puo lï'

314. 未冬夫里

LOC *muyh toung puo lï'

318. 波夫里

LOC *pa puo lï'

323. 所乙夫里

LOC *ShïO' 'ït puo lï'

336. 尔陵夫里 / 仁樹夫里 / 仁夫里

LOC *ñie' lïng puo lï' / *ñin juoh puo lï' / *ñin puo lï'

435. 古 莫夫里

LOC *ko' mak puo lï'

437. 古沙夫里 / 古沙夫

LOC *ko' Shæ puo lï' / *ko' Shæ puo

(Is the latter an error, or is 夫 short for 夫里? could 夫 and 夫里 both represent *por? Or was there dialect variation: *por ~ *porï? LOC had no final liquids, so a non-Chinese final *r could be written as zero or with a graph for a *lV syllable. What I write as *-r could have been *-l.)

450. 半奈夫里 / 半那夫里

LOC *panh nayh puo lï' / *panh na puo lï'

457. 所夫里 / 省津 / 泗沘 / 南夫餘

LOC *ShïO' puo lï' / *sieng' (or *Shieng') tsin / sih pi' (or *bi) / *nəm puo yïO

The last of these is the most interesting. #457 is modern 夫餘 (부여) puyO county (official site). Its capital 夫餘邑 (부여읍) puyO-Up (puYo town) was once 泗沘 (사비 'Sabi'), the last capital of Paekche. The modern Korean name Sabi is anachronistic and would have been unknown to the Paekche, who probably pronounced 泗 and 沘 as *si and *pi. (But that doesn't necessarily mean that the two-graph sequence 泗沘 represented *sipi. See below.)

所夫里 LOC *ShïO' puo lï' looks vaguely like Middle Korean syəvUr > modern Korean 서울 sOul 'capital' (i.e., Seoul). MK syəvUr may come from an earlier *sepUr (see my next post). This *sepUr also vaguely resembles 泗沘 ?*sipi and its first syllable is not unlike the first syllable of 省津 *sieng' (or *Shieng') tsin. But how can the vocalic differences be reconciled?

Could 所 LOC *ShïO' and 省 LOC *sieng' / *Shieng' have been attempts to write a Paekche *sö or *shö? This hypothetical *ö (not reconstructed by Bentley 2000) would have been rounded like LOC *O and front like LOC *i and *ie (and earlier Korean *e).

I think the second half of Paekche *s(h)ö-por(ï) is cognate to the *pUr of earlier Korean *sepUr, but I am not sure that the first halves are really cognate for reasons I will explain in tomorrow's post.

Earlier Korean *U in *pUr could be a reduction of a different earlier vowel (e.g., an *o which was still preserved in Paekche *por[ï]?). The real problem is trying to relate earlier Korean *pUr (< *porV?) and Paekche 夫里 *por(ï) to 沘 which as far as I know has never been pronounced with a rounded vowel in any form of Chinese. If 泗沘 really is related to earlier Korean *sepUr and Paekche *s(h)ö-por(ï), perhaps it represents a Paekche dialect which underwent these changes:

*ö > *ü > *i

*or > *oy > *uy > *wi (> *i?)

or *or > *oy > *ö > *ü > *i

or *orï > *oï > *oy > *uy > *wi (> *i)

or *orï > *oï > *oy > *ö > *ü > *i

It would be nice if I could find other Paekche names with correspondences similar to those in

所夫里 : 省津 : 泗沘

*s(h)ö-por(ï) : *s(h)ö : *s(h)ip(w)i

but I don't know of any.

夫餘 (부여) puyO is, of course, also the name of the ancient puyO state and its people. The name 夫餘 may date back to the 2nd century BC, when it would have been pronounced as *pa la in Old Chinese. I have long wondered if this name was cognate to Old Japanese *para 'field' (and to Paekche 夫里 *por[ï] and/or Korean 벌 pOl 'field'?). Perhaps the Paekche continued to pronounce this word as *pala or *para even though 夫餘 was pronounced as puo yïO in LOC. Therefore the Paekche pronunciation of 南夫餘 LOC *nəm puo yïO may have been whatever the Paekche word for 'south' was** plus *pala or *para.

(I now wonder if the OJ name 秦 *pata [written with the graph for the Qin Dynasty!] associated with peninsular immigrants is from a variant of *pala. Could *pala have come from an earlier *pata? Here's a more likely etymology of *pata.)

Next: The Chinese city that isn't.

*My ideas about Late Old Chinese are heavily influenced by Axel Schuessler and are constantly changing: e.g., 夫 was reconstructed as LOC*pwïa in the previous post but as LOC *puo here. This is because 夫 had lost its a-quality by the very end of the LOC period. In the poetry of 陶淵明 Tao Yuanming (365-427), 夫 rhymed with 徂 *dzo which definitely had a rounded vowel (Starostin 1989: 650). 夫 also had a rounded vowel in the poetry of 嵇康 Ji Kang (223-262), who rhymed it with 偶 *ngou' (Starostin 1989: 628).

**John Bentley (2000: 439) reconstructed the Paekche word for 'south; front' as *arIpIsI on the basis of a Japanese gloss aripisi in 日本書紀 Nihon shoki. The word is cognate to Korean 앞 aph- < *arpoh (Martin 1996) 'front'.

都守熙 (도수희) Toh Soo Hee (2005: 130) reconstructed another Paekche word for 'south':*nima.

***OJ *pata could simply be from a Paekche cognate of Middle Korean par^r < *pat^r 'ocean' and modern Korean 바다 pada 'ocean'. Also cf. Old Japanese *wata 'ocean' (a peninsular loan coexisiting alongside native *umi). Korean 바다 pada 'ocean' is not directly descended from MK par^r but seems to be from an earlier *pata without medial *-t- lenition. I don't know why the MK form has a final liquid corresponding to zero in modern 바다 pada.

07.4.11.1:41: A 布禮 PURE ETYMOLOGY?

Sven Osterkamp, author of the amazing Tangut software that made many of these posts possible, reminded me that (火+本*) ?*p^r resembles the Paekche word for 'village', reconstructed by John Bentley (2000: 436) as *purE (E = an uncertain e-like vowel) on the basis of a Japanese transcription**

布禮

Late Old Chinese *pOh ley'

Early Middle Chinese *poh ley' (?*lay' in the south [see Pulleyblank 1984: 199])

Go-on (Early Sino-Japanese)

fu rai (modern pron.)

< *pu rai (Old Japanese period pron.)

< ?*po rai (pre-Old Japanese period pron.)

Late Middle Chinese *po lyey (tone marks omitted)

Kan-on (Later Sino-Japanese)

ho rei (modern pron.)

< *po rei (Old Japanese period pron.)

(Kan-on was borrowed during the OJ period, so there is no pre-OJ pron.)

I do not know whether 布禮 was made up on the spot by a Japanese scribe who knew the Paekche word or reflected (perfectly or otherwise) a spelling used by Paekche speakers themselves.

I also do not know which readings of the characters underlay the transcription. 日本古典文學體系 Nihon koten bungaku taikei interpreted 布禮 as fure in modern pronunciation. 布 and 禮 usually represented the Old Japanese syllables *pu and *re which are now pronounced fu and re in Old Japanese.

In any case, 布 represented a Paekche syllable with an initial *p followed by a nonlow rounded vowel (*u? *o? *O?). Given that 火+本 sounded like the Pihwa Kaya and Shilla word(s) for 'fire',

if 布禮 'village' and 火+本 were cognate

then 布 禮 'village' must also have sounded like 'fire'

But the rounded vowel of 布禮 does not match the unrounded vowels of Koreanic words for 'fire':

Paekche *pər 'fire'*** (Bentley 2000: 428)

Middle Korean 블 pUr (U is IPA [ɯ]: an unrounded high back vowel.)

Therefore I don't think 布禮 'village' and 火+本 were related.

I do, however, think that 布禮 'village' might be related to the Paekche place name element

夫里

LOC*pwïa lïə'

Early Middle Chinese *puə lï'

if it had a rounded first vowel (i.e., if it were more like EMC *puə lï' than LOC *pwïa lïə'). 夫里 may have represented Paekche *purə and 布禮 may have represented a suffixed form *purə-i. Bentley (2000: 428) reconstructed 夫里 as *pworï 'land'. See "Origo ignis" for another interpretation of 夫里.

The resemblance between 夫里 and Indo-European words like Sanskrit pur (cf. Singapore 'lion city') and Greek polis is, of course, totally coincidental.

Next: The 郡 county caper.

*㶱 if you have a font that supports the 'CJK ideographs' (argh) extension A block of Unicode. Thanks to Sven for the codepoint.

My policy on this site is to stick to the CJKV sinographic core which is supported by Windows' default Asian fonts. I generally don't want visitors to have to see blank boxes or download fonts.

Yes, I did ask readers to download an Avestan font three years ago, but I'd like that to be a rare exception. I thought it was a reasonable imposition since

- there is no support for Avestan at all in standard Windows fonts (whereas I can type 9X% of the sinographs I need for this blog)

- an Avestan font is small and easy to install (unlike Mojikyo for Tangut)

**Strictly speaking, 布禮 was a gloss for the Chinese word 村邑 'village' in 日本書紀 Nihon shoki.

***Although I know of no direct evidence for the Paekche word for 'fire', the fact that 火 Late Old Chinese *xwa' 'fire' alternates with 伐 Late Old Chinese*bïat 'strike' in the spelling of a Paekche place name (Ryu 1983 #344)

比 自火

LOC *pi' dzih xwa'

比斯伐

LOC *pi' sie bïat

implies that the Paekche word for 'fire' (Bentley's *pər) sounded like *bïat. (Also cf. Go-on / early Sino-Japanese batsu < *mbatu for 伐. *mbatu - and early SJ as a whole - is probably based on Sino-Paekche.)

Also note that the Paekche place name (Ryu 1983 #325)

皆 火

LOC *kEi xwa'

was changed to

戒發

Shilla period Sino-Korean*?kyəy par (계발 kyebal in modern pron.)

under Silla rule. Presumably 發 *par 'emit' was supposed to sound like the Paekche word for 火 'fire' (my *p^r?).

皆火 may have represented Paekche *kep^r.

It's not clear whether Paekche had a distinction between an upper mid *e and a lower mid *E. For now I'll use the generic symbol *e. Bentley (2000: 434) reconstructed Paekche *e but not Paekche *E.

07.4.10.1:36: 火王 REX IGNIS

The name of the Korean county now known as 昌寧 (창녕) chhangnyOng 'Shining Peace' has been through a lot of changes over time (Ryu 1983: 477; LOC = Late Old Chinese):

3rd century, prior to Shilla takeover:

不 斯 (transcription from 魏志 Wei zhi in 三國志 Sanguozhi)

LOC *pwïə or *pwït sie

Early Shilla:

比 自(火+本)

LOC *pi' dzih ...

(火 + 本 had no Chn reading, but its right-hand phonetic element was 本 LOC *pən' or *p^n')

比自火

LOC *pi' dzih xwa'

比斯火

LOC *pi' sie xwa'

比自伐

LOC *pi' dzih bïat

比子伐

LOC *pi' tsïə' bïat

比斯伐

LOC *pi' sie bïat

(比 also had less common readings: *pih, *bi, *bih, and possibly*bit)

(斯 also had an uncommon reading *sieh)

非 火

LOC *pwïy xwa'

(in 非火伽倻, now called 비화가야 Pihwa Kaya in Korean)

Late Shilla:

火王

Late Middle Chinese *xwa wang (lit. 'fire king')

All but the last of these names had an initial *p- or *b- if pronounced in Chinese and 火 LMC *xwa 'fire' presumably stood for a Pihwa Kaya and/or Shilla word similar to Middle Korean 블 pUr 'fire'.

The vowel following the initial labial stop was high and unrounded (*ï?): cf. LOC *i and *ï and the U of the later Middle Korean word for 'fire'.

The second consonant was some kind of dental affricate (*ts?) cf. the LOC clusters *-t-s-, *-'-dz-, *-'-s-, *-'-ts-.

The variation may imply that the peninsular transcribers lacked a distinction between *ts and *dz in their language(s) or that */ts/ became *[dz] intervocalically and some wrote /ts/ phonemically and others wrote it phonetically.

(If 不斯 were pronounced *pwïə sie without a medial -t-, a Chinese transcriber may simply may misheard the foreign name.)

The variation among the second vowels (LOC *i, *ie, *ïə) made me wonder if they actually corresponded to zero in the original and were only there because LOC had no syllables ending in *-ts, so the transcribers had to pick graphs for LOC syllables of the shape *(t)sV to represent the coda of the Pihwa Kaya original.

Speaking of 非火 Pihwa, it has no trace of anything like *(t)sV. There are at least two explanations for this:

1. 非 LOC *pwïy was a monosyllabic abbreviation of a longer transcription 非X in which X would be a graph pronounced like LOC *(t)sV.

(But the trouble is that there is no known transcription like 非X.)

2. 非 LOC *pwïy was an attempt to write a non-Chinese *pïts (and the non-Chinese coda was completely ignored).

but if this were the case, why not transcribe the name with a final *-t or *-s graph like 必 LOC *pit, 閟 LOC *pwïs, or 沸 LOC *pwïs?

*pïts vaguely resembles Korean 빚 pich- 'debt' and 빛 pichh- < *pichVk 'light'. I doubt anyone would name a place 'debt'. *pichVk, on the other hand, is not far from the LOC transcriptions which have a common shape *pitsV('/h). (Maybe Koreanic *-k had been reduced to *-' or *-h in the Pihwa Kaya language.) If this place was called 'Light', that meaning may have been loosely reflected in the first elements of its later names 火王 'Fire King' and 昌寧 'Shining Peace'.

I have explained why I think 火 'fire' stood for a Koreanic word for 'field' in "Origo ignis".

伐 LOC *bïat appears to be a transcription of that Koreanic word for 'field'.

I don't know why the later names have 王 'king' and 寧 'peaceful' which have neither a semantic nor a phonetic connection with the earlier names.

Working with early Korean peninsular languages is in some ways much harder than working with Tangut. There are no dictionaries of these languages or even a single literary work in these languages apart from poetry in the Old Korean (Late Shilla) language. We can only infer their charateristics from the names that they left behind. Although we still do not fully understand the transcription systems they used, this peninsular adaptation of sinography is still less mysterious than tangraphy.

Next: E doesn't stand for easy.

07.4.9.1:56: 火本 ORIGO IGNIS

If you looked at the Dictionary of Chinese Character Variants entry for 火 'fire' + 本 'origin', you may have noticed that its Korean reading was listed as "Bal", which doesn't look like the ?*p^r 'field' that I reconstructed. Moreover, nowhere in the definitions is 'field' mentioned:

1. 百濟人名也。

'A Paekche person's name.'

2. 地名也。

'A place name.'

So how did I come up with ?*p^r 'field'?

"Bal" is equivalent to 발 pal in the romanization I use on this site. The Dictionary of Chinese Character Variants' romanization is based on modern Korean spelling which no longer uses the vowel letter arE a 'bottom a' (ㆍ= IPA [^]). arE a, as the name implies, has generally merged with regular a (ㅏ) in modern Korean. Hence 발 pal in modern spelling could be equivalent to an earlier pal or p^l. I have no way of knowing for sure which was the reading of 火+本, but I would guess that it was p^l, because it alternates with 火 'fire' in the pre-8th century 非火伽倻 (비화가야) Pihwa Kaya place name

比 自(火+本) ~ 比自火

(now 昌寧 [창녕] chhangnyOng 'Shining Peace')

and 火 'fire' was 블 pUr in Middle Korean (> modern Korean 불 pul). (Its Sino-Korean reading 화 hwa is irrelevant.) This alternation may suggest that 火+本 and the pre-8th century equivalent of 火 pUr were nearly homophonous. *p^r is closer to pUr than *par since *^ is a higher vowel than *a.

Moreover, *p^r is also closer to the vowel of the Late Old Chinese reading *pən' or *p^n' of the phonetic element 本. If 火+本 were supposed to represent *par, it should have had a Chinese phonetic element with the vowel *a: e.g., it could have been 火+般 with 般 Late Old Chinese *pan instead of 本 LOC *pən' or *p^n'.

Finally, 火+ 本 appears with the gloss po in the place name 比 自(火+本) in 日本書紀 Nihon shoki (Chronicles of Japan, 720 AD). If 火+本 were pronounced *par, its gloss would have been pa.

I think*p^r might have meant 'field' because it resembles Korean 벌 pOl 'field'. 火+本 *p^r alternated with 火 'fire' which appears at the end of many 新羅 Shilla place names: e.g. (using Ryu 1983's numbering)

659. 屈阿火, 670. 推火, 693. 于火, 699. 喟火, 700. 古里火

703. 骨火, 704. 仇火, 705. 只火, 706. 雉火, 711. 毛火, etc.

I presume that 火 stood for a Shilla period ancestor of Korean 벌 pOl 'field' (a probable near-homophone of the Shilla period ancestor of Middle Korean 블 pUr 'fire') and was not meant to be read literally as 'fire' or as a Shilla syllable resembling Late Old Chinese *xwa' 'fire'.

Although Korean 벌 pOl 'field' would have come from an earlier *pər rather than an earlier *p^r, Korean is a descendant of the Silla language whereas 火+本 was used to write a Paekche person's name and a Pihwa Kaya place name. Perhaps 'field' was *p^r in the Pihwa Kaya language and possibly also Paekche (but there is no guarantee that the Paekche name meant 'field'*).

Here is a wild scenario: If Japonic languages were spoken in Kaya (and maybe throughout the southern part of the peninsula?), and if the Proto-Japonic word for 'field' were *p^r, then

- the pelagic Japonic languages shifted *^ to *a: hence Old Japanese *para and Okinawan haru < *paru

- the peninsular Japonic language of Pihwa Kaya retained *^ in *p^r 'field'

- this word for 'field' survived in place names even after formerly Japonic-speaking areas became Koreanic (i.e., Shilla)-speaking (but why would the vowel shift from *^ to *ə? Did some Japonic languages have *ə instead of *^ in 'field'?)

- and Korean 벌 pOl 'field' is from the Japonic substratum whereas

펄 phOl 'wide expanse of land' < Middle Korean 퍼리 phəri < *kVpəri or *pVkəri

밭 pat 'field' < *pat(V)k(V)

are native Koreanic words.

Japanese 畑 hata(ke) < *pata(kai) 'field' is probably from a Koreanic word not unlike *pat(V)k(V).

I have also wondered if the Japonic words for 'field' with -r- (OJ *para, Okinawan haru < *paru) could be from the same source at a later date following -t- > -r- lenition, but intervocalic lenition may have occurred in Korean after the peak of peninsular-pelagic contact, judging from pre-modern Korean chh^ryəy 'order' < c. 8th c. Late Middle Chinese 次第 *chhï diəy 'id.'

(If lenition predated the 8th century, then the Korean word 'order', having been borrowed around the 8th century, would have a medial stop -t- [> modern -d-] instead of a medial liquid -r-. But lenition must have taken place after 'order' was borrowed around the 8th century. I suppose one could propose two waves of lenition, but I have no other reasons to assume that apart from the wish to derive the Japonic *parV words from a Koreanic *pat[V].)

Next: agniraajan: neither this, nor fire.

*According to Toh (2005), the Paekche word for 'field' may have been 夫里, pronounced*pïa lïə' in Late Old Chinese (representing a Paekche *p^r(ə)?). (里 was used to write pre-Old Japanese *rə and presumably had a similar sound value in Paekche, since the early Japanese writing system was probably a modification of the Paekche script.) I cannot find any direct evidence for 夫里 meaning 'field' (i.e., an alternation of 夫里 with 原 'field'), but the gloss seems like a reasonable guess given the similarity of 夫里 to Korean 벌 pOl 'field'.

The Paekche place name 比自火 (#344 in Ryu 1983; identical in spelling to the Pihwa Kaya place name but referring to modern 全州 (전주) chOnju) contains what may be an alternate (Shillafied?) spelling (or Shilla translation?) 火 for 夫里.

Another Paekche name for this place was 完山'Complete Mountain' which was changed to 全 州 'Complete Region' in the 8th century. This implies that the transcription 比自 'compare-self' (which I will discuss next time) meant something like 完/全 'complete'.

07.4.8.21:55: THE 本 ORIGIN OF 火 FIRE

Yesterday I discovered that the Taiwanese government's Dictionary of Chinese Character Variants has a section devoted to made-in-Korea characters (國字 국자 kukcha) such as the ones I mentioned in part 5 of "Oh Deer". Not all of the kukcha are of the type

CV sinograph + -C coda sinograph

e.g., 乫 갈 kal = 加 가 ka + 乙 을 Ul

The kukcha 火 'fire' + 本 'origin' (< 'root') (at the top of this page) appears to be a double phonetic kukcha for the 非火伽倻 (비 화가야) Pihwa Kaya place name element ?*p^r 'field' (cf. modern Korean 벌 pOl 'field', 펄 phOl 'wide expanse of land' < Middle Korean 퍼리 phəri [which may have Japonic relatives*])

火 Late Old Chinese xwa' 'fire' could be translated as 블 pUr in Middle Korean; presumably the Pihwa Kaya word for 'fire' was similar to MK 블 pUr and was (nearly?) homophonous with the Pihwa Kaya word ?*p^r for 'field'

本 Late Old Chinese pən' (phonetically closer to [p^n']?**) 'root' was similar to the Pihwa Kaya word ?*p^r for 'field'

A similar sinographic representation of the English word field might consist of

- a non-Sinitic left-hand phonetic 氈 Md zhan 'felt' (representing an English word sounding like field; the Md reading is irrelevant) +

- a Sinitic right-hand phonetic 非 Md fei 'not' (approximating the sound of Eng field; the Md meaning is irrelevant)

The resulting character (氈+非) would be rather cramped and not unlike a tangraph. Perhaps some tangraphs have a similar structure: i.e., they consist of combinations of Tangut A and Tangut B phonetic elements. If a tangraph contains a clear Tangut A phonetic element combined with an element of unknown function: e.g.,

TT4484 本 ORIGIN mə^r 2.76

mystery left side (Nishida radical 107) + Tangut A phonetic element mə^(r)

(see part 1 of "The Origin of Origin" for other tangraphs with this phonetic)

that other element may be a Tangut B phonetic element.

Next: Rex ignis.

*A peninsular language (Paekche?) word similar to Korean 벌 pOl 'field' and Pihwa Kaya ?*p^r may be the source of Jpn hara < *para 'field' and Okinawan haru < *paru 'farm, tract of open country'. -ra and -ru may have been attempts to imitate a foreign liquid coda.

I think Korean 펄 phOl 'wide expanse of land' < Middle Korean 퍼리 phəri might be an unrelated word. Korean ph- may be from an earlier *kVp- or *pVk-. Early Korean ?*kVpəri resembles 郡 koori < Old Japanese *kəpəri 'district'. However, the semantics leave something to be desired.

OJ *kəpəri 'district' has also been related to Korean 고을 koUl < *kop^r 'district'. The semantics match perfectly but the mismatch of vowels (OJ ə : Korean o) cannot be explained.

These words for 'district' have been linked to Late Old Chinese 郡 *gunh < Old Chinese*guns 'district'. (Starostin reconstructed Old Chinese *gurs, though I know of no Chinese-internal evidence for a final *-r.) However, there would have been no reason for Koreanic or Japonic speakers to insert a -p- into Late Old Chinese 郡 *gunh. Furthermore, the vowels do not match (LOC u : Kor o : OJ *ə).

Late Old Chinese 郡*gunh may be the source of Japanese kuni 'state' (probably via Paekche, since direct Chinese contact was very limited in early Japan).

**Starostin reconstructed the Old Chinese pronunciation of 本 'root' as*pəər'. For years, I suspected that the late Old Chinese dialect(s) known to the peoples of the Korean peninsula retained final *-r unlike mainstream late Old Chinese dialects. If that were correct, then perhaps the choice of 'origin' as a phonetic symbol for Pihwa Kaya ?*p^r 'field' reflected an eastern late Old Chinese reading *pər'. However, I know of no Chinese-internal evidence for reconstructing a final *-r instead of *-n in this word in Old Chinese.

Starostin linked his 本 *pəər' root' to Tibeto-Burman final -r/l words

Lushai bul 'root'

Garo bor 'firewood' (connecting this to 'root' seems like a stretch)

Andro phol '?'

Moshang phuul '?'

and even to Proto-Austronesian *bua'l 'uproot a tree'. (But CVVCC doesn't look like a normal Proto-Austronesian root structure, and there is nothing verbal about the Chinese or Tibeto-Burman words.)

Schuessler (2007: 160) has similar proposed cognates (note the differences from Starostin's data):

Lushai bul < *buul' 'beginning, origin, base, stump, lower part'

Northern Naga pul 'tree'

Garo bol 'root, stump'

Moshang puul 'root'

However, I am still reluctant to reconstruct a final *-l or *-r in Old Chinese 本 'root'. If most or all of these words (excluding the Proto-Austronesian word) are cognate, it is possible that Proto-Sino-Tibetan had *-l ~ *-n variation in this word and Old Chinese inherited only the *-n variant. That would explain why there is no Chinese-internal evidence for a final liquid: there never was one in Chinese.

Schuessler also proposed that Old Chinese *-n in 'root' may actually be from an earlier cluster of root final*-l plus a nominal *-n suffix. One could hypothesize that the late Old Chinese dialect(s) known to the peoples of the Korean peninsula lacked the *-n suffix in 'root', but I am hesistant to pile one speculation upon another.