11.4.23.23:50: UCHINAAGUCHI PART 30: U N I

sounds like 'you and I' and is a mnemonic device for these conversion rules:

Nu, Ni- > n-: e.g.,

nigami > njami 'bitterness'

midare > njari 'disorder

mukai > nkee 'opposite side'

uN, iN- > ʔnN-: e.g.,

umare > ʔnmari 'birth'

ine > ʔnni 'rice plant'

J imo : O ʔnmu 'sweet potato'

J mine : O nni 'peak'

J minna : O nnna 'all' (with three n's!)

J mune : O nni 'chest' (homophonous with 'peak')

simply nii in Nakijin (Curry 2004: 265)

J ume : O ʔnmi 'plum'

O niku : J niku 'meat' (< Chn 肉)

O nujumi : J nozomi 'wish'

O michi : J michi 'road'

O mura : J mura 'village'

O ʔimi : J imi 'meaning' (< Chn 意 + 味)

O ʔunu 'that' (no J cognate)

What might have blocked the 'U N I' rules from applying to those words?

11.4.22.23:11: UCHINAAGUCHI PART 29: THE BLUE BEAR'S BIG PICTURE

aoikuma.com (aoi kuma is Japanese for 'blue bear') has a one-page list of Okinawan-Japanese sound correspondences and more. Most should not surprise anyone who's read all the parts of this series so far.

Here are a couple I haven't mentioned yet:

1. d ~ r:

O dachi ~ rachi : J rachi 'limits' (< Chn 埒 'enclosure')

O danpu ~ ranpu : J ranpu 'lamp' (< Eng lamp)

Native Okinawan and Japanese words never begin with r-. I suspect Okinawans once pronounced foreign words with native d- to avoid r-, just as Koreans pronounced foreign words with n- to avoid r-: e.g., 남포 nampho 'lamp'  But now r- is permissible in both languages: hence O rampu and K 람프 ramphŭ 'lamp'.

2. Initial syllabic n-:

O njami : J nigami 'bitterness'

O njari : J midare 'disorder'

O nkee : J mukai 'opposite side'

O ʔnmari : J umare 'birth'

O ʔnni : J ine 'rice plant'

What determines when Okinawan has initial syllabic n-?

Can you predict the Okinawan words corresponding to these Japanese words?

J imo 'sweet potato'

J mine 'peak'

J minna 'all'

J mune 'chest'

J ume 'plum'

Answers tomorrow.

11.4.21.23:54: UCHINAAGUCHI PART 28: A HANDLE ON THE BIG PICTURE

Sven Osterkamp and I have been discussing w- and y- in Okinawan. Here's what he and I think is going on so far before front vowels:

*wi > yi (but J wi > i)

*we > wi (but J we > e)

An exception is

O yii ~ ʔii (not wi!) : J e < ye < we < Middle Chinese 繪 *ɣwajh 'picture'

which may have been borrowed when the Japanese word was ye.

Some words have a w- that eludes explanation: e.g.,

O wii ~ yii : J e < ye 'handle' (found by Sven)

O wiiri : J eri < yeri 'collar'

yeri 'collar' has a Middle Japanese homophone yer-i 'twines, and ...' whose root yer- is a variant of yor- 'to twine' which Sven identified as the Japanese cognate of O wiiruu ~ yiiruu 'cord'.

Could yiiruu be from a verb yiir- corresponding to MJ yer-?

I erroneously linked O wiiruu ~ yiiruu to J yu- < yup- 'to tie', but the real Okinawan cognate of that J verb is yuur- without the -ii- of wiiruu ~ yiiruu.

11.4.21.1:54: UCHINAAGUCHI PART 27: Y A W-ORD FOR CORD?

Continuing my survey of Okinawan alternations from part 26:

4. 'Cord' is both wiiruu and yiiruu, and Sakihara (2006) lists no obvious Japanese cognate to help us determine whether w- or y- is original. My initial guess was that the word is cognate to O yii ~ ʔii, yui(-maaruu) 'cooperative labor'. which is cognate in turn to J yui 'tie'.

But then what would -ruu be? The second half of yui-maaruu provides a clue. maaruu 'circle' is derived from *maar-u, an old adnominal form of the verb maar- 'to go round' that is cognate to J mawar- 'to go round'. Is yiiruu 'cord' derived from yiir-u, the adnominal form of a verb yiir- < yuir- 'to tie' (not listed in Sakihara 2006) that is cognate to J yu- < yup- 'to tie'? (No. There was no *yuir-. See Part 28.)

And what of the variant wiiruu? Is the w- [ɥ] a fusion of an earlier *yu-?

Could the w- of wii 'intoxication' also be from an earlier *yu- < *yo- (cf. its Japanese cognate yoi)?

Continuing my survey of Okinawan alternations from part 25:

3. Earlier Japanese wi, we may correspond to Okinawan yi ~ ʔi rather than O wi: e.g.,

O yii ~ ʔii : J i < wi 'boar' (of the twelve Earthly Branches)*

O yii ~ ʔii : J e < we < Middle Chinese 繪 *ɣwajh 'picture'

but

O wii : Old Japanese wep-i 'intoxication'

One might mechanically reconstruct a Proto-Japonic *ɥ- or even *ʔɥ- to account for 'boar' and 'picture'. But I don't think such exotic initials are necessary. According to Serafim (2008: 81), O wi and we were [ɥi] and [ɥe] in the "first half of the 20th century - now used in performance". Did

wi, we > wi > ɥi > yi > ʔi

in some but not all wi/we-words in Okinawan? Modern Japanese has no wi, ɥi, or yi. Could the ʔi-stage be due to Japanese influence?

*4.20.00:42: The regular, noncalendrical word for 'boar', 'pig', etc. is ʔwaa, from an earlier *ʔuwa transcribed in 琉球館譯語 Liuqiu guan yiyu (early 16th c. AD) as 烏哇 *ʔuʔwa.

11.4.18.23:59: UCHINAAGUCHI PART 25: STILL LOST IN SPACE

Here's the long-delayed follow-up to part 21 in which I listed examples of various correspondences and alternations involving Okinawan glides. Unexpected yet attested forms are in bold.

Tonight I'll get two easier types of alternations out of the way:

1. The wu ~ ʔu and yi ~ ʔi alternations involve glides preceding similar vowels. Japanese has no distinctions of the type

/wu/ : /u/

/yi/ : /i/

because it does not permit glides to precede similar vowels: wo and ye are now extinct, and wu and yi didn't even exist in Old Japanese.

Perhaps the Okinawan wu ~ ʔu and yi ~ ʔi alternations are due to Japanese influence and the ʔ-forms are innovations lacking the un-Japanese sequences wu and yi.

2. ʔii, ʔee, and ʔei instead of expected yii and yee may be due to Japanese influence since Japanese has ii and ee but not yi(i) or ye(e/i). Moreover, according to Sakihara (2006: x)*,

For many speakers, the sound we are writing as e (that is, ʔe) at the beginning of words has a subtle on-glide pronunciation that can soiund like ʔye.

(I have changed Sakihara's Okinawan orthography to match mine.)

ye and ʔye could easily be confused. The one and only minimal pair in Sakihara (2006) is

ʔ(Y)eema 'space' : Yeema 'Yaeyama'

but 'space' has the variant ʔweema (which I'll deal with later) and 'Yaeyama' has the variant ʔEema (homophonous with one version of 'space'). There is little motivation to maintain a rare distinction.

Variation could result from different layers of borrowing: e.g., Yeigo 'English' could be from a(n older?) variety of Japanese with ye and ʔEigo could be from a (newer?) variety of Japanese which had shifted ye to e.

*4.18.1:10: I feel odd referring to "Sakihara" because I suspect that someone else wrote the introduction to his dictionary. Sakihara was a professor of history, not historical linguistics. And not all entries in his dictionary were written by him. Some were added by others after his passing.

11.4.17.23:59: CAN PROTO-TAI LIVE WITHOUT *ʔY-(UU)?

In my last post, I derived Proto-Tai *ʔC- from an earlier *ʔVC-: e.g.,

Thai อยู่ <ʔyuu> yuu < *ʔyuu < *ʔV-yuu 'to stay'

This proposal could eliminate the third of Li Fang-Kuei's (1977: 58) four series of PT stops:

PT *ʔy- and the questionable *ʔw- are not stops, but I've put them into the table, indicating their nonstop status with parentheses.

The high frequency of PT *ʔb- and *ʔd- could be due to mergers:

*ʔV- + any labial > *ʔb-

*ʔV- + any dental > *ʔd-

But why didn't *ʔV- + any palatal lead to a lot of *ʔy-?

Recently, Pittayawat Pittayaporn derived the aspirated series of modern Tai languages from PT *r-clusters: *Cr- > *Ch-. The absence of an aspirated series in PT accounts for the early borrowing of Chinese aspirates as nonaspirates: e.g., Late Old Chinese 七 *tshit 'seven' as PT *cet (not *chit*).

If the aspirates are also secondary, then only two series of stops need to be reconstructed at an earlier stage:

Weera Ostapirat (2000: 207, 2004) reconstructed only two series of stops (voiceless and voiced) for Proto-Kra and Proto-Hlai, 'sisters' of Proto-Kam-Tai and 'aunts' of Proto-Tai:

Did Proto-Kra-Dai also have only two series?

*4.18.00:33: The mid vowel of PT *cet 'seven' is unexpected. I wonder if it reflects a dialect of Old Chinese with an *i lowered to *e by a preceding low presyllabic vowel that was later lost:

nonstandard OC *sʌ-snit > *sʌ-snɪt > *tshɪt > *tshet; borrowed into PT as *cet

standard OC *sʌ-snit > *s-snit > *tshit

(presyllable lost low vowel before it could condition lowering in the following syllable)

nonstandard OC *Cʌ-tshit > *Cʌ-tshɪt > *tshɪt > *tshet; borrowed into PT as *cet

(presyllable lost after its vowel conditioned lowering in the following syllable)

standard OC *(CV-)tshit > *tshit

(presyllable lost before it could condition lowering, was never present, or had a high vowel that didn't lower *i)