몽골어 동사 'ge- (말하다)'의 문법화에 대하여

Abstract

The aim of this paper is to look at the grammaticalization of the Khalkha Mongolian verb ge-. When the verb ge- is used as a main verb, it means `to say`. It is often used as a quotation verb, introducing direct or indirect quotation. A connective ending form of ge-, ge-j `say-Conn , is, however, used as a complementizer of saying, cognition, and perception verbs. The same form ge-j can be used to introduce a purpose clause as well. The verb ge- is also used as an auxiliary in Khalkha Mongolian, expressing desire/intention or prediction, preceded by a verb ending in the Irrealis verbal noun suffix -x. This paper explains the various usage of ge- as grammaticalization through reanalyses. We argue that the complement construction with ge-j started from a two-independent-clause construction. Originally ge-j was a predicate of the preceding clause. When the following clause has a verb of speaking as its predicate, the meaning of ge- becomes redundant. Furthermore, when the following verb is transitive, it provides a context where the first clause can be reanalyzed as a complement of the following verb, and hence ge-j as a complementizer. Now that ge-j is reanlyzed as a complementizer of saying verbs, it generalizes to include verbs of cognition and perception. It is also argued that the auxiliary ge- started from a construction in which ge- was originally a matrix verb taking a complement clause of a verbal noun. It is the equi-subejct construction and the Irrealis verbal noun suffix -x that enable ge- to be reanalyzed as an auxiliary from a matrix verb. The verb ge- as a main verb needs a human subject. The auxiliary ge- of desire/intention also needs a human (at least animate) subject. But when it is used as an auxiliary of prediction, it loses this lexical restriction. The auxiliary ge- of prediction may take not only a human/animate subject but also an inanimate subject. This indicates that the auxiliary of prediction is more grammaticalized than that of desire/intention.