Recently, John Kupchik published an extensive and impressive grammar on the languages collectively referred to as ‘Eastern Old Japanese.’ Being interested mainly in Korean, I reviewed the discussion on Korean loanwords in Chapter 4 and came across a note on OJ amo~omo ‘mother’ and LMK émi ‘id’ (Kupchik 2023: 120).
Due to the sound correspondences in the initial syllable, Vovin (2010: 234–235) proposes EOJ amo ‘mother’ is unrelated to OJ omo ‘mother’, though I disagree with that viewpoint (I reconstruct PJe *amə ‘mother’), since amo ‘mother’ is also phonographically attested in WOJ (in NK 82), and it is easy to explain the variant OJ form omo as the result of regressive vowel assimilation or initial vowel reduction of *a > /ə/. If the Korean comparison is too problematic, I believe it should be tossed out, rather than splitting up the Japonic forms into one ‘native’ and one ‘borrowed’, as Vovin does. Unfortunately, this word is not attested in Old Korean, but it is worth noting that the Middle Korean forms àpí ‘father’ and èpí ‘id.’ differ in their initial vowels, and it seems àpí is more archaic (Vovin 2010: 235).
That is, OJ amo~omo ‘mother’ seems to be a loanword from Korean, but the vowels of the first syllable do not correspond well.
I believe that the solution to this problem can be found in other Korean sources, outside of LMK. In this post, I aim to explore how to address the vowel mismatch issue posed by Kupchik (2023).
Reconciliation of OJ /a/ and LMK /e/
The word for ‘mother’ in Late Middle Korean was ‘émi’. However, its predecessor form is believed to have been *ámi. We can find direct evidences from 鷄林類事 Kyelim Yusa, a main source on Early Middle Korean.
母曰了彌
母 ‘mother’ is called ‘了彌 liau mi’姑曰漢了彌
姑 ‘mother-in-law’ is called ‘漢了彌 xan liau mi’
The character ‘了 liau’ here appears awkward in the context of Korean phonotactics, since word-initial rhotics are not allowed in every well-documented Korean variant. This is why all previous studies have regarded it as an error for ‘丫 a’ (Choi Youngsun 2015: 96). Consequently, the words above can be interpreted as EMK *ami and *hanami (<ha- ‘big, great’ + -n + *ami).
Some might be concerned that the transcription ‘丫彌 a mi’ may contain inaccuracies, as 鷄林類事 Kyelim Yusa is transcribed by a foreigner. In fact, the initial /a/ form for ‘mother’ is barely found in Hangul sources, in contrast to the /e/ form ‘émi’, which is amply attested. This led Jin Tae-Ha (2019: 235) to argue that ‘丫 a’ is the transcription of /e [ə]/.
However, the word *ámi is highly probable, since ámay [ámɛ] ‘mother’ is attested in North Hamkyeng dialects (Kwak Chunggu 2019). As far as I know, this word first appears in Азбука для Корейцевъ, a 19th-century Russian source on North Hamkyeng dialects.1
Amɛ! poo, nɛ sɛrɨ putturəsso. (Азбука 62)
Look, mother! I have caught a bird.
Modern Korean form emeni ‘mother (polite)’ is a descendant of LMK émánĭm ‘id’, which is the combination of LMK éma ‘mother’ and LMK nĭm ‘lord’ (Kim Moorim 2020). Similary, North Hamkyeng ámay ‘mother’ likely formed through the combination of *áma ‘mother’ and nim, followed by vowel nasalization of /ni/, monophthongization of /aĩ/, and the deletion of the final /m/.
Then, how did the initial /e/ in LMK émi ‘mother’ come about? I can’t provide a definitive explanation, as the research on the phonological changes of pre-LMK is limited, but I speculate that umlaut took place in *ámi ‘mother’. In Late Middle Korean, the semi-vowel /y/ was sporadically inserted in the preceding syllable of a syllable with form Ci or CyV (VCi > VyCi). If such insertion could occur before LMK, *ámi could have become *áymi. Since the earlier pronunciation of LMK /e [ə]/ was [e], adjacent assimilation would have led to *áymi>*éymi>émi.
Overlooked Issue ― the Second Vowel
So far, I have shown that the first vowel issue can be addressed using Korean sources. However, we have overlooked a more serious problem: the second vowel.
Vovin(2010: 234) argued that the /i/ in LMK émi is diminuive suffix -i, so éma in LMK émanĭm is more appropriate to be compared. However, I believe LMK éma is just a vocative form of émi, marked by LMK vocative marker -a. Hence, LMK éma does not seem like a suitable comparand for OJ amo~omo. Even if they can be compared, the vowels LMK /a/ and OJ /o/ does not appear to be correspond well, as Vovin(2010) has pointed out.
Then, should we reject this comparison just because the second vowel does not correspond well? It might be necessary to explore alternative sources. Though I am not generally fond of using Old Korean proper nouns due to the lack of systematic decipherment, it seems like I have no other choice at this point. Let’s examine the toponyms below from 三國史記 Samkwuk Saki.
雲峯縣, 本母山縣【或云阿英城, 或云阿莫城】 (Samkwuk Saki 34:11a)
雲峯縣 ‘Cloud summit’ district is originally 母山縣 ‘Mother mountain’ district. [Otherwise called 阿英城 a jaeng castle or 阿莫城 a mak castle]
We notice the correspondence between 母山 ‘mother mountain’ and 阿莫 a mak.2 Pak Pyong-chai(1968: 110) compared 阿莫 a mak to LMK ámh ‘female (of animals)’. Although Pak Pyong-chai(1968)’s study may require some revisions from today’s persepective, I find this comparison to be valid. Based on this toponym, I suggest to reconstruct OK *ámVk ‘mother’, which is evolved into LMK ámh ‘female (of animals)’.
One might ask as follows: “The form represented by 阿莫 a mak is obviously *amak, but why didn’t you specify the second vowel?” The reason is that it plays a crucial role in connecting OJ amo and LMK ámh. The change *ámVk>ámh implies the loss of the V, and it is not impossible that this V was /a/ and it vanished without any additional change. However, internal evidence from LMK suggests another possibility.
Late Middle Korean had two weak vowels /o [ʌ]/ and /u [ɨ]/, which were subject to synchronous vowel dropping in some phonetic environments, e.g., before a vowel or after a glide. Furthermore, they seems to be deleted before certain obstruents during the pre-LMK period. More precisely, unaccented weak vowels appear to have been dropped between a consonant and /k/ followed by a vowel (ò,ù>ø // C_kV). Words such as LMK namwò ‘tree’, colò ‘handle, azò ‘younger brother’ are thought to have undergone this change, as they alternate ‘nàmk, còlg, àzg’ before particles beginning with a vowel. This alternation has traditionally been explained by the reconstruction of pre-LMK forms *namòk, *colòk, *azòk (Lee Ki-Moon 1962, Lee & Ramsey 2011: 184-185). In constrast, there is no clear internal evidence within LMK suggesting the direct deletion of /a/ in the same phonetic context, as seen in LMK kalàk ‘strand’, isàk ‘ear of grain’, cwokàk ‘piece’.
Therefore, if we prioritize the internal evidence from LMK, the V in *ámVk could have been /o [ʌ]/ or /u [ɨ]/, as it needed to be dropped to become LMK ámh. Among these two options, /o [ʌ]/ is more suitable for our context because its pre-LMK pronunciation is [ə], which aligns well with *amə, a potential antecedent of OJ amo (Kupchik 2023: 120). The coda k in *ámVk might have already lenited to h by the time it was borrowed into OJ, or it might not have been reflected due to the syllable structure constraints of Japanese.
In summary, the Old Korean form borrowed into Old Japanese might have been *ámək ‘mother’. Although I have primarily relied on internal evidences from LMK in reconstructing V of *ámVk, it would not be impossible to consider the change *ámak>*ámək.
References
- Choi, Youngsun (2015), 「鷄林類事의 음운론적 연구」 [A phonological study on the Gyerim-Yusa]. Doctoral Dissertation, Cheonnam National University.
- Jin, Tae-Ha (2019), 『鷄林類事硏究』 [A Study of the Kyelim Yusa]. Seoul: Myengmundang.
- Kim, Moorim (2020). 『국어 어원사전』 [Korean Etymological Dictionary]. Seoul: Jisiggwagyojang.
- Kupchik, John. (2023), Azuma Old Japanese: A Comparative Grammar and Reconstruction. Walter de Gruyter GmbH.
- Kwak, Chunggu (1991), 「咸鏡北道 六鎭 方言의 音韻論」 [Phonology of Yukcin Dialect in Hamgyongbukdo]. Doctoral Dissertation, Seoul National University.
- Kwak, Chunggu (2019). 『두만강 유역의 조선어 방언 사전』 [A Dictionary of Korean Dialects in the Tumen River Basin]. Paju: Taehaksa.
- Lee, Ki-Moon (1962), 「中世國語의 特殊 語幹 交替에 대하여」 [On Some Special Stem Alternations in Middle Korean]. The Chin-Tan Society 23, 119-153.
- Lee, Ki-Moon & Ramsey, Samuel Robert (2011), A History of the Korean Language. Cambridge University Press.
- Pak, Pyong-chai (1968), 「古代 三國의 地名語彙攷 ― 三國史記 地理志의 複數地名을 中心으로」 [On Toponyms of Three Kingdoms of Korea ― Focusing on the Variant Transcriptions in the Placename glosses of the Samkwuk Saki]. The Paek-San Hakpo 5, 51-134.
- Vovin, Alexander (2010), Koreo-Japonica: A re-evaluation of a common genetic origin. University of Hawai’i Press.