18.104.22.168:59: ARAE A AFTER NONSIBILANTS IN OLD RYUKYUAN TRANSCRIPTION
The hangŭl letter arae a (ㆍ) appears almost exclusively after ts and s in the transcription of Old Ryukyuan (OR) in 海東諸國紀 Haedong chegukki (1471). This extremely skewed distribution implies that the OR segment represented by arae a was 'sibilant-friendly': e.g., it might have been a syllabic fricative [z̩] like the 'vowel' that only appears after Mandarin alveolar sibilants: zi [tsz̩], ci [tshz̩], si [sz̩]. Note that arae a also appears in the premodern Sino-Korean readings of characters now pronounced zi, ci, si in Mandarin: e.g.,
|Gloss||Sinograph||Premodern Sino-Korean||Modern Sino-Korean||Mandarin|
|child||子||ts + arae a||cha||zi [tsz̩]|
|this||此||tsh + arae a||chha||ci [tshz̩]|
|four||四||s + arae a||sa||si [sz̩]|
One cannot conclude that arae a in 15th century Korean was syllabic [z̩] since it frequently could also appear after nonsibilants: e.g., Middle Korean m + arae a + r 'horse' which was probably not [mz̩r].
Moreover, arae a appears in borrowings of Chinese morphemes that never had syllabic [z̩]: e.g., Middle Chinese 每 *məjʔ 'every' corresponds to Middle Korean m + arae a + i which could not have been [mz̩jʔ].Therefore the Middle Korean vowel written as arae a must have
1. sounded like an OR segment
- occurring only after sibilants (see below for exceptions)
- that later developed into Okinawan -i
2. sounded like the Chinese segment corresponding to later Mandarin [z̩]
3. been easily pronounceable after and even before nonsibilants in a single syllable - I have yet to see a language with syllables like [mz̩r].
Kim-Renaud's proposed value of *o for arae a fails criteria 1 and 2. There is no reason for an *o-like vowel to appear only after sibilants. It is unlikely that an *o-like vowel would develop into -i (but cf. Ukrainian in my next post). Finally, there is no evidence for an *o-like vowel in 子此四 at any time during the history of Chinese.
I have long favored an unrounded *ʌ for arae a which also fails criteria 1 and 2 for the same reasons.
Whatever arae a was, it might have sounded like the OR vowel in these two words without sibilants:
'older sister', transcribed as ar + arae a + i (cf. standard Jpn ane 'id.')
'white wine', transcribed as riŋk + arae a + na sakɯi (cf. standard Jpn nigori 'muddiness', sake 'wine'; Okinawan mingwi 'muddiness' [with m-!], saki 'wine')
Standard Japanese -e may come from *-ai, and the -i transcription for 'older sister' may imply a proto-Japonic *anai.
The Old Japanese root for 'turbid' was *niŋgər-. Arae a in the transcription riŋk + arae a + na of its PR cognate corresponds to OJ *ə and may represent a reduction of a PR *o that developed from schwa:
> Old Japanese *niŋgər- > standard Jpn nigor-
> Proto-Ryukyuan *miŋgor- ~ *niŋgor-
> Old Ryukyuan *niŋgVr- (transcribed as riŋk + arae a + na)
> earlier Okinawan *miŋgor-i > modern Okinawan mingwi
2.27.00:45: The Haedong chegukki transcription of the OR word for 'morning' implies that the OR vowel written as arae a was somehow minimal:(not directly descended from the OR form with *n-)
Other instances of OR syllables corresponding to Okinawan sibilant-i syllables and standard Japanese sibilant-u syllables were transcribed as sibilant + arae a in Haedong chegukki: e.g.,
Transcription: stomɯiti (with no vowel between s and t)
cf. Okinawan sutumiti, shitimiti, hitimiti and archaic mainland Japanese tsutomete < Early Middle Japanese tutomete
|Gloss||OR transcription||Okinawan||Standard Japanese|
|when||itts + arae a||ichi||itsu|
|first month||syooŋkwats + arae a
syaoŋkwats + arae a
|be at ease; flat||mas + arae a + ŋko||masshiigu||massugu 'straight'|
|summer||natts + arae a||nachi||natsu|
|vinegar||s + arae a + u||shii||su|
|ink slab||s + arae a + ts + arae a + ri||shijiri||suzuri|
|ink||s + arae a + mi||shimi||sumi|
|sheep||pits + arae a + tsya||hichiji||hitsuji|
|dragon||tats + arae a||tachi||tatsu|
The transcriber could have written 'morning' as s + arae a + tomɯiti, but chose to write no vowel between s and t. (st- was a permissible cluster in Middle Korean.) I conclude that sibilant + arae a sounded similar to sibilant + zero.