한국어 상호성 관련 표현 ‘서로’, ‘상호’, ‘피차’의 분포적 특성: 정량적 분석을 중심으로

초록

This paper investigates the distributional properties of three Korean reciprocal expressions: selo, sangho, and phicha. Unlike previous qualitative studies focusing on selo, this study examines all three in the NIKL written corpus by combining five quantitative methods: frequency analysis, collocation analysis, word embedding, surprisal, and prompt engineering. The results are interpreted with reference to features Nedjalkov(2007) discusses as points of differentiation among coexisting reciprocal markers, such as frequency, productivity, stylistic differentiation, co-occurrence, and interchangeability. Selo is by far the most frequent and serves as the central expression with the broadest distributional range. Sangho shows productivity in forming Sino-Korean compound nouns in formal contexts, while phicha displays a distinctive literary tendency. Selo–sangho and selo–phicha co-occurrence is attested, whereas sangho–phicha co-occurrence is rare. Languagemodel- based methods converge in suggesting that sangho exhibits the highest context-dependency. The models further suggest a directional asymmetry centered on selo, which appears substitutable for phicha and may extend into sangho’s predicate-modifying contexts. These findings characterize the three expressions as forming a hierarchical division of labor within the semantic domain of reciprocity in Korean.

키워드