Amaravati: Abode of Amritas

11.2.19.23:59: KIM-RENAUD'S MIDDLE KOREAN VOWELS (PART 3)

In part 2, I asked,

Which reconstruction of [the Middle Korean vowel] arae a best accounts for 15th century hangŭl transcriptions of Old Ryukyuan?

The three reconstructions mentioned in part 2 were [o] (Kim-Renaud), [ɔ] (Yi and Martin), and [ʌ] (Ramsey). Koono Rokurou proposed [ɐ] and Pulleyblank (1984) proposed [ə]. All five are nonhigh and nonfront, but otherwise vary:

	central	back unrounded	back rounded
upper mid	ə		o
lower mid	ɐ	ʌ	ɔ

I have not seen upper mid back unrounded [ɤ] proposed as a possible value of arae a.

In 海東諸國紀 Haedong chegukki (1471, shortly after the invention of hangŭl in 1443), arae a generally transcribes an Old Ryukyuan vowel corresponding to Old Japanese u: e.g.,

'month': Old Japanese tukɨ, Old Ryukyuan ts?ki (transcribed in hangŭl as ㅈㆍ기)

One might assume that the mystery OR vowel was o as reconstructed by Kim-Renaud. o is close to u. However, if one accepts K-R's Middle Korean vowel reconstruction as a whole, the correspondences between OR, Old Japanese, and modern Okinawan become very complex, whereas my reconstruction of Middle Korean vowels leads to a simpler set of correspondences:

Old Japanese (Miyake 2003)	OR vowel using K-R to interpret hangŭl transcriptions	My interpretation of the OR transcriptions	Modern Okinawan
a	ɔ	a	a
i	i	i	i
ɨ	i	i
e	i, jɑ	i, jə(j) (for OR [je]?)
əj	əi	ɨi, jəj (for OR [je]?)
u	o after sibilants	ʌ after sibilants
u	ɨ elsewhere	u elsewhere	u
o	ɨ, u	u, o
ə	ɨ, u	u, o

(The above table is not comprehensive and only illustrates the basic pattern of correspondences.)

OR interpreted through K-R's reconstruction

- has jɑ but no simple a-vowel, which is highly unlikely

- has o almost exclusively after sibilants (see my next post for exceptions)
- necessitates the bizarre sound change OR o > Oki i

e.g. 'month': OR tsoki > Oki chichi

(2.20.00:33: Ukrainian did undergo such a change, but I don't understand it. I wrote a post about this shift in Ukrainian, but I didn't finish it and I can't find it. I'll upload it if I find it.)

My interpretation of OR has two similar problems:

- ʌ is almost exclusively after sibilants (see my next post for exceptions)

- it necessitates the bizarre sound change OR ʌ > Oki i

If I had never seen Haedong chegukki, I would have proposed a more normal sound change

Proto-Japonic *u > OR ɨ > Oki i after dentals and alveolars

which is similar to the changes

PJ, OJ *u > modern standard Japanese [ɯ] (backer than [ɨ] but otherwise identical)

Late Middle Chinese *ɨ > i after sibilants in many modern Chinese languages (e.g., Cantonese)

But if OR had ɨ instead of a vowel resembling any of the proposed values of arae a, why wasn't OR ɨ transcribed with the hangŭl letter for a high nonfront unrounded vowel (K-R's ㅜ or my ㅡ)?