A Herman Ottó Múzeum Évkönyve 32. Kunt Ernő emlékére. (1994)

TANULMÁNYOK - KI–MOON LEE: A koreai és az altáji nyelvek összehasonlító tanulmányozása (magyar és angol nyelven)

ges the reconstructed forms to cilajun < *tila-gun and cul < *tjal < *tila and specifies that *i underwent the so-called "breaking". He does not, however, deal with the o in Korean tol. Meanwhile, Street (1978), in a reconsideration of examples showing an opposition between a in the initial syllsbles of Turkic words and i in the initial syllables of Mongolian words, suggests that they go back to the proto-language form *ja. His discussion includes not only various lexical items of the Altaic languages and Korean, but also Japanese isi'stone.

11 The one thing lacking is a detailed explanation of how *]a changed into i or o in the above-mentioned languages. I discussed accepting the breaking of *i in Middle Korean torh 'stone' and Old Turkic tas in Lee (1959). I have never changed my mind on this matter. Regarding Tungus |olo, I believe it necessary to find additional examples before accepting alternation of word-initial *t and *d. I still have my doubts regarding comparison with Japanese isi. The second Korean example in Ramstedt (1924) is kolta 'half, which he compares with Mongolian kaltasun 'half, Tungus kalta, and Japanese kata 'one side'. First, it must be pointed out that this word has never existed in Korean. Ramstedt (1949) corrected the Korean word to kari-da 'to divide, to cut into parts', and the Mongolian word to qaltagai, kelte-gei 'a part, a half. With Poppe (1960), the Mongolian words in this comparison are correctly shown as qaltas Teil, Fragment', qalta-ci- 'zerbrechen', qaltajai 'half. The above mentioned Korean verb shows the alternate forms kar

A-/karj- in Middle Korean, and I hypothesize that this stem is a development from ancient *kar

A

<jf-. This comparison enables us to reconstruct Proto-Altaic *kalta- 'to break apart'. Here it should be noted that intervocalic -*lt- corresponds to Korean -r- (<*1). This seems to be the result of simplification of intervocalic consonant clusters (deletion of the second consonant) in Korean. A similar simplification process can be seen in the following example. Poppe (1960) compares Mongolian ba^ta- 'hineingehen, hineinpassen, genug Platz haben' with Old Turkic bat- < *bakta- 'sinken, versinken', while Tsintsius et al. (1975-1977) compares with Evenki bat-, Solon bakta-, Manchu bakta- ' BMeuiaTBC« '. I would like to add Middle Korean pak- to this comparison. This verb has the meanings 'to thrust (in, into), to put (into)' (transtive) and 'to get stuck, to be embedded' (intransitive).

12 Deletion of the second consonant in a medial consonant cluster in Korean can be seen in this example as well. This kind of phenomena is a desirable research topic for future comparison of Altaic languages with Korean. To this point I have briefly described examples from Ramstedt (1924) of lexical comparison between the Altaic languages and Korean. Although these are no more than two minor examples, they more than suffice to give one a feeling for the complexity of this kind of comparison. It is my belief that lexical comparison is the most urgent task facing researchers comparing the Altaic languages and Korean. It goes without saying that linguistic comparison must be based on precise rules of sound correspondence, and accuracy in rules of sound correspondence can only be achieved after extensive lexical comparison has been accomplished. 11 Martin (1966) compares Korean tol and Japanese isi. 12 Ramstedt (1949) compares Modern Korean pak- (a transitive verb) with Tungus baka- 'to find'. On the other hand, Tsintsius et al. (1975-1977) compares this Tungus verb with Modern Korean verb pat- 'to receive'. Neither comparison seems valid. 101