*La-st, er, last night I rediscovered Alexander Vovin's (2000) article "Pre-Hankul Materials, Koreo-Japonic, and Altaic" which dealt with Old Korean locatives in section 11. He did not include the locative 惡中 that I've been writing about but did cover some others. I've rethought my reconstruction and have revised it to incorporate some of Vovin's ideas:

Old Korean spelling Old Chinese readings Middle Chinese readings Premodern Sino-Korean readings Idu reading (Yu 1964) Vovin's (2000) Old Korean My Old Korean until now My latest Old Korean
良中 *raŋ truŋ *lɨaŋ ʈuŋ ryang tyung ahʌy, ahay, aəy, aɯy, əɯy *lang *(l)akʌy *(l)akʌy
也中 *ljajʔ truŋ *jæʔ ʈuŋ ya tyung (none) (none) *lakʌy
*raŋ *lɨaŋ ryang (r)a *(l)a *la
*raj *la ra (none) *la (none)
惡中 *ʔak truŋ *ʔak ak tyung (none) *akʌy *akʌy
*truŋ * ʈuŋ tyung aɯy
惡之 *ʔak *ʔak tɕɨ ak ci (none) (none)
惡希 *ʔak xəj *ʔak xɯj ak hɯi *axʌy
阿希 *ʔaj xəj *ʔa xɯj a hɯi *ahʌy

Vovin's *lang has nothing to do with the later idu readings, whereas I and others assume that the idu V(h)Vy readings are somehow related to the OK readings.

I agree with Vovin that 也 might have retained an archaic *l- in a Chinese dialect known to early Koreanic speakers. The original *l- of Old Chinese 舌 *m-let 'tongue' is still intact in Cantonese 脷 lei < *let-s today, roughly two millennia after it was lost in mainstream Late Old Chinese.

The spellings 阿希 and 惡希 indicate that intervocalic *-k- might have already begun to lenite in Old Korean. 希 may have been chosen for an archaic OC-like reading *xəj that was closer to OK *xʌy ~ *hʌy than the newer Sino-Korean reading hɯi. *akʌy, *axʌy, and *ahʌy exemplify three stages of lenition. Kyunyŏ (10 c. AD), author of nearly half the known OK corpus, used the spellings 惡中 and

The 惡-spellings are absent from idu, which may indicate that *-k-lenition was complete in the dialect(s) underlying idu, so writing the locative with 惡 *ak no longer made sense.

The Late Middle Korean locative -əy/-ay may be a contraction of OK *ahʌy < *axʌy < *akʌy. LMK -əy should come from an OK *-əkɯy, but the absence of such an allomorph suggests that -əy, the higher vowel counterpart of -ay, and the LMK vowel harmony system as a whole may not be old: i.e., it may postdate OK poetry.

The spelling 惡之 <ak.GENITIVE> suggests that OK *akʌy may in turn be a compound of *ak 'locative' (?) plus *ʌy 'genitive'.

請轉法論歌 3.1-5 has the interesting phrase


<LAW.WORLD ak.GENITIVE s> *pəpkyəy akʌy s

which combines the locative akʌy with the other OK genitive morpheme s. IN THE MIDDLE OF GOOD AND EVIL

I just realized that two spellings for the Old Korean locative postposition are

良中 <a.MIDDLE> with 良 'good'

惡中 <ak.MIDDLE> with 惡 'evil'

I didn't notice this earlier because 'good and evil' is 善惡 with a different morpheme 善 for 'good', not 良惡.

*ak 'evil' was borrowed from Chinese, though it's purely phonetic in OK 惡中 *akʌy.

I've been assuming that 良 *a was once read *la, a simplification of its Middle Chinese reading *lɨaŋ. But perhaps a native OK word a- 'good' underlay the use of 良 'good' as a phonogram for MK *a. However, there are no known descendants of such a word. The current native Korean gloss for 良 is ŏjil 'benevolent, virtuous'. Neither it nor Middle Korean tyoh- 'good' sound like a-.

Unlike 惡, 良 could stand by itself as a locative postposition: e.g., after 'branch' in

枝良 (祭亡妹歌 7.4-5)

<BRANCH ?> 'on a branch'

Kim Wancin (1980: 124) read this as *kaci ra, but Lee and Ramsey (2011: 70) imply that the OK reading of 良 was cognate to the Middle Korean locative postposition ay/əy. If 良中/惡中 was *akʌy, then perhaps 良 alone was *a. And Kim in fact interpreted locative 良 as *a just two lines later:

彌陁剎良 (祭亡妹歌 9.3-6)

<mi.tha.char a> *mithachar a 'in the land of Amitabha'

(陁 = 陀; 彌陁剎 < 阿彌陀刹 Amitaabha-kṣetra 'Amitabha land')

Was the locative postposition *a after consonants (e.g., *mithachar) but *ra after vowels (e.g., *kaci)? Could this alternation be extended to 良中: *akʌy ~ *rakʌy? If OK 'sea' ended in a consonant like the later MK words parʌr and patah, then it must have been followed by *akʌy:

海惡中 (普皆廻向歌 3.4-6)

<SEA ak.MIDDLE> *pa(t)tVC akʌy 'in the sea'

The spelling 惡中 <ak.MIDDLE> was chosen to unambiguously indicate the *a-variant of the postposition after the consonant-final word *pa(t)tVC 'sea'. A C-L-AS-H OF CODAS (PART 4: A MOMENTARY MISTAKE)

In part 3, I analyzed Old Korean

海惡中 (普皆廻向歌 3.4-5)

as 海惡 *pa(t)tak 'sea' + 中 akʌy. I was excited to see a spelling that seemed to reflect a final *-k corresponding to the -h of Middle Korean patah 'sea'. However, I overlooked another instance of 惡中 in 稱讚如來歌 right after an instance of the other OK spelling of 'sea' that I mentioned:





一念惡中 is 'at (that) one moment'.

一念 'one moment' is probably *irnyəm, a borrowing from Late Middle Chinese *ʔir niem, a translation of Sanskrit kṣaṇa 'moment'. There is no later native Korean word ending in -ak meaning 'one moment', so it's unlikely that 一念惡 should be interpreted as a phonogram-final sequence <ONE.MOMENT.ak> *...ak. Hence I regard 惡中 as a locative postposition *akʌy in both 海惡中 <SEA ak.MIDDLE> and 一念惡中 <ONE.MOMENT ak.MIDDLE>. A C-L-AS-H OF CODAS (PART 3: SINISTER SEAS)

I should have mentioned these Old Korean spellings of 'sea' much earlier in this series:

海等 (稱讚如來歌 3.6-7, 普皆廻向歌 5.4-5)

海惡 (! - 普皆廻向歌 3.4-5; stay tuned if this surprises you)

海 is a logogram for 'sea' that tells us nothing about the OK word(s) it represents. It's the following characters that are interesting.

Many Old Korean words are written with logogram-phonogram sequences: e.g.,

道尸 <ROAD.hli> *kil 'road'

夜音 <NIGHT.ɯm> *pam 'night'

二肹 <TWO.hɯr> *tuɣɯr 'two'

海等 is also such a sequence. 等 is a phonogram, but was it read as
Sino-Korean *tɯŋ 'class, grade'

or the OK ancestor of its loose Middle Korean translation equivalent tʌrh 'group, plural suffix'

Either reading began with a *t that corresponds to the r of MK parʌr 'sea' and the t of MK patah 'sea'. So 海等 may have represented *pat ... But what followed the t? I'm guessing that 等 was read as something like MK tʌrh. Did

海等 <SEA.?tʌrh>

represent OK *pa(t)tʌrh, with a final cluster -rh that was simplified differently in two MK dialects?

Proto-Koreanic? Old Korean Middle Korean
*pattarh dialect A: 海等 *pa(t)tʌrh parʌr (with *-tt- > *-t- > *-r-)
(dialect B form unattested) patah (with *-tt- > *-t-)

Or was 'sea' simply OK *pa(t)tʌr? MK also had -rh words: e.g., tʌrh itself and hanʌr(h) 'heaven'. Some of these -rh words had variants ending in -r, but patah would be the only -h variant of a -rh word.

海惡 <SEA.ak>

with the phonogram 惡 *ak 'evil' looked like a spelling of OK *pa(t)tak with a final -k corresponding to the -h of MK patah. However, 惡 may not be part of the noun, as it's followed by 中:

海惡中 <SEA ak.MIDDLE> *pa(t) ...? akʌy

惡中 may be an alternate spelling of the OK locative postposition 良中 *akʌy (later read as ahʌy, ahay, ahəy, aɯy, əɯy with two degrees of -k- lenition in idu texts).

Then again, Yu (1964: 782) lists aɯy as an idu reading of 中 sans 良*, so

海惡中 <SEA.ak MIDDLE> *pa(t)tak akʌy

could also be possible if the one-character spelling 中 for the locative can be projected back into OK. Lee and Ramsey (2011: 70) glossed OK 良, 中, 良中 as 'locative'. Some early Korean peninsular texts predating OK poetry have this un-Chinese usage of 中 as a case marker: e.g., 三月中 <THREE MONTH MIDDLE> = 'in the third month' in a box "believed to have been crafted in Koguryŏ in 451" (Lee and Ramsey 2011: 55).

This usage of 中 even spread to Japan. The Inariyama burial mound sword inscription from 471 or 531 has 七月中 <SEVEN MONTH MIDDLE> = 'in the seventh month'.

*The use of 良 Middle Chinese *lɨaŋ 'good' as a phonogram for early Korean a is difficult to explain. Perhaps this a was originally *la with a liquid that was later lost. A C-L-AS-H OF CODAS (PART 2)

I am not at all convinced that *r ever became Korean h as I proposed in part 1. Two examples in two different positions (onset and coda) are insufficient evidence. Moreover, all other evidence links h to k and ng: e.g.,

Late Middle Korean səyh : Old Japanese saki- 'three', Late Middle Korean k 'three' (allomorph before t-, c-: sək cah 'three feet')

Late Middle Korean cah : Early Middle Chinese 尺 *tɕhɨak 'foot'

Late Middle Korean tyəh : Late Middle Chinese 笛 *tɦiek or Early Middle Chinese *dek 'flute'

Late Middle Korean syoh : Late Middle Chinese 俗 *sɦyok 'vulgar'

Late Middle Korean zyoh : Late Middle Chinese 褥 *ɲʑyok 'mattress'

(1.17.00:24: These four borrowings may reflect a Chinese *-k that had lenited to *-ɣ just as Chinese *-t had lenited to *-r. See Lee and Ramsey 2011: 86.)

Early Middle Korean 亇支 *maki : Late Middle Korean mah 'yam' (Lee and Ramsey 2011: 87, 93; I can't find this word in Yu Chhang-don's 1964 dictionary)

Late Middle Korean stah : Modern Korean ttang 'earth' (Matisoff's rhinoglottophilia comes to mind.)

How can I reconcile these correspondences with my derivation of h from *r? Perhaps the Vietnamese word for 'carbon' may provide a clue. French carbone [kaʀbɔn] was borrowed into Vietnamese as các-bon [kaakɓɔn]. Did Korean

*r > > > > *x > *k

harden in certain environments: e.g., before *t and *c (and sporadically intervocalically*)? If so, why weren't all *r in those positions affected? Even Old Korean had words with final *r transcribed with 乙 Late Middle Chinese *ʔɨt.

*1.17.1:13: Old Japanese saki- 'three' may be borrowed from an early Koreanic *sahi with an intervocalic fricative. Perhaps

*səri > *səhi > səyh

in the Koreanic branch leading to Late Middle Korean. (The OJ loan may reflect a different branch with *a as the first vowel of 'three'.) Modern Korean rŭn 'thirty' still has an -r- before -ŭn 'ten'.

LMK 'river' may have a similar derivation

*nari > *nahi > nayh

whose earliest proto-form resembles the Paekche word transcribed as

那利 Early Middle Chinese *na lih

katakana glosses nari ~ nare (Bentley 2000: 425)

There seem to have been at least two kinds of liquids in earlier Koreanic. One may have remained a liquid r/l in later Korean, whereas the other may have become h or even k depending on circumstances.

1.17.1:37: Two liquids contrasted in final position in Old Korean (Lee and Ramsey 2011: 66):

乙 Late Middle Chinese *ʔɨt < Early Middle Chinese *ʔɨət

尸 (Late/Early Middle Chinese *ɕi)

The interpretation of the latter is uncertain. I've assumed that it was chosen for a reading like Old Chinese *l˳i that survived as an archaism in the peninsula, but it could be an abbreviation of 卢 < 盧 Late/Early Middle Chinese *lo. Could 尸 have represented a voiceless liquid or a voiceless lateral fricative similar to Old Chinese *l˳i? An early Koreanic *l˳ or could have sometimes lost its lateral quality and become *h.

The Old Korean prospective modifier suffix -尸 corresponds to Late Middle Korean -rʔ. Could the glottal stop be a remnant of an earlier *h?

*-l˳ > -lh (reanalyzed as a cluster?) > -rh > -rʔ

Old Korean is the language of Shilla. The prestige dialect of Late Middle Korean was from the capital region which was in former Paekche and Koguryo territory. Did non-Shilla Koreanic speakers pronounce Shilla *l˳ as *lh or *rh or even *h when they shifted to the Shilla language?

If 尸 represented a lateral, Old Korean 日尸 <DAY.l> 'day' might have been *nal (cf. Late Middle Korean nar). If early Koreanic *r and *l correspond to *r and *l in other Altaic languages, I would expect *nal to correspond to a non-Koreanic *-l word, but the matches that come to mind

Khitan *neir (transcribed in Chinese as 捏咿兒 *nie i r) 'day'

Written Mongolian nara(n) 'sun'

have r,  not l! A C-L-AS-H OF CODAS (PART 1)

So far, I've tried to explain the t : r correspondence in two Middle Korean words for 'sea' by reconstructing a geminate *tt as their common source:

patah < *-tt- (from a dialect in which geminates simplified after lenition)

parʌr < *-d- < *-t- < *-tt- (from a dialect in which geminates simplified before lenition)

MK ʌ could be a reduction of *a, so the ʌ : a correspondence is not a problem. However, I couldn't initially think of any other cases of an h : r correpsondence. Then I remembered the Old Korean accusative particles

乙 ~ 肹 (for *ɯr ~ *hɯr?; cf. their Middle Chinese readings *ʔɨt and *xɨt)

corresponding to Middle Korean (r)ʌr ~ (r)ɯr*. Could OK *h- correspond to MK r-?

OK *hɯr does not occur exactly where I'd expect it to: i.e., only after vowel-final nouns:

吾肹 'I-ACC' (獻花歌 3.1-2); cf. MK na rʌr

but 花肹 'FLOWER-ACC' (獻花歌 4.1-2); cf. MK koc ʌr (not *koc-rʌr)

Here's an instance in which the *h of 肹 may actually be at the end of the preceding noun:

地肹 'EARTH-ACC' (安民歌 7.2-3); cf. MK stah ʌr (not *stah-rʌr)

The problem is that we cannot be sure how the logograms 吾, 花, and 地 were read in OK. They could have represented

words lacking final consonants which were unrelated to MK na, koc, stah

or the vowel-final ancestors of those MK words: e.g., OK *na, *kocV, *stahV

But let's suppose *hɯr is cognate to MK rVr. Could the h of OK *hɯr and MK patah reflect dialects in which */r/ was pronounced *[ʁ], a fricative that was both h- and r-like? Brazilian Portuguese has similar pronunciations of *ʁ:

a voiceless velar fricative [x], voiceless uvular fricative [χ], or a voiceless glottal fricative [h].

MK as a whole is not descended from such a dialect. Do any such dialects exist today that have h instead of r as in MK and  r/l as in standard Korean?

Next: The Các-Bon Conundrum

*1.16.2:20: The height of the vowel of the MK particle depended on the vowel of the preceding noun:

'yin' (higher) vowels: ɯ, ə, u:  (r)ɯr

'yang' (lower) vowels: ʌ, a, o: (r)ʌr

neutral vowel: i: either (r)ɯr or (r)ʌr

In the 朝鮮館譯語 (c. 1400), the Korean phoneme soon to be written in hangul as -ㄹ was transcribed in Chinese as 二 *r, so I interpret MK -ㄹ as [r], not [l]. In modern Korean, -ㄹ is [l].

Tangut fonts by Mojikyo.org
Tangut radical and Khitan fonts by Andrew West
Jurchen font by Jason Glavy
All other content copyright © 2002-2012 Amritavision