통사문형 선택과 의미정보 융합에서 어순과 격탈락 효과

다국어 초록 (Multilingual Abstract)

Production choices between active and passive structures in simple sentences vary with the order of the agent and patient nouns. Interestingly, however, there exists a strong tendency of structure choices in English relative clauses with respect to noun animacy. Given that animate nouns are more accessible than inanimate nouns, both the animacy of the head noun and of the agent of the action affect structure choice in relative clauses.

It is reported that 98% of the participants produces passive relative clauses in English when the head noun is animate and the percentage of passive relative clauses declines sharply to 61.2% when the head noun is inanimate. Japanese speakers are known to produce almost exclusively passive relative clauses (99.3%) when the head noun is animate, but the percentage drops to 30.1% for inanimate head noun. It is thus shown that even though English and Japanese differ in the proportions of active and passive relative clauses for inanimate head noun, they both clearly exhibit an overall animacy effect on structure choices in relative clauses.

A series of pilot experiments of Korean for this research show that for animate head noun, Korean speakers produce 67% of production in the form of passive relative clause, which is rather low rate, compared to English and Japanese, but still represents that passives are much preferable to actives for animate head noun in Korean. Like Japanese, however, Korean speakers do not show a preference of passive relative clause when the head noun is inanimate and the proportion of passives drops to 29%.

The questions that arise here are ‘why passives are rather rare overall in Korean’ and ‘what other factors than word order should be considered to account for the production difference of actives/passives between Korean and Japanese, even though both of them have the same SOV order.’ To address these questions, this research will explore the effects of the preference of accusative Case deletion over subject/dative Case deletion on the semantic integration and syntactic choice in Korean.

This research will perform five experiments with different goals and targets: the experiments for online response time, off-line production choice of active/passive clause, off-line preference of modification in the complex NP, and two types of corpus analysis of the various texts. The results will be further analyzed with respect to age to eliminate the possible biases of cognitive aging on any specific aspects of both language production and comprehension. The proposal that this research tries to make is that the semantic integration on syntactic structure choice varies among languages in a way that the semantically primed phrases or syntactic structure are more tightly linked at the conceptual level, where the flexibility of word order and the preference in modification of NPs, interpretation of thematic roles, and most importantly frequency of Case deletion in Korean are manipulated in the rather complex and systematic ways.

The effect of semantic integration has crucial implications for how the language production system coordinates syntactic planning processes. The research will suggest that the speakers of English, Japanese, and Korean all make similar structural choices with respect to semantic factor of animacy in the relative clause construction. It will further claim that the cross-linguistic differences among languages can be accounted for, depending on how much they have been exposed to the frequencies of the relevant constructions, including Case deletion. This conclusion therefore suggests that production choices reflect both cognitively motivated production demands such as noun accessibility as well as language specific constraints on frequency/experience faced by the speakers.