Earticle

Jong Sup Jun. 2004. Enriched composition in conceptual semantics and the generative lexicon theory. Korean Journal of Linguistics, 29-3, 437-460. Compositionality, often called the Fregean principle of meaning, is an important assumption in many contemporary semantic theories. The traditional view of compositionality is to consider the minimal (morpho-)syntactic units as also the minimal units of semantic composition. This explains why an interpretive component goes hand in hand with PS-rules in traditional model-theoretic semantics. On the other hand, Jackendoff (1997, 2002) proposes an alternative view of semantic composition, namely enriched composition. In this view, semantics is an autonomous level of grammar like syntax; semantic decomposition provides a conceptual structure that is invisible from syntax; and meaning composition looks into the internal semantics of a word as well as the word-level composition that is transparent to syntax. In this paper, I present a number of Korean data that support the view of enriched composition. Data from Pustejovsky's (1995) generative lexicon (=GL) theory, the substantival nouns-hata construction, reference transfer, verbal aspectual coercion, binding, control, etc. all prove problematic without assuming enriched composition. To formalize enriched composition in the conceptual semantic framework (Jackendoff 1983, 1990), I adopt lambda extraction from Jackendoff (2002), which allows us to extract an argument, a variable, or property out of word-internal event descriptions like the qualia structure (Seoul National University).

This study aims to analyze the trends of cognitive semantics research conducted in Korea over the past decade and suggest future research directions based on this analysis. Focusing on the core theories of cognitive semantics? conceptual metaphor, conceptual metonymy, conceptual blending?the study analyzes academic papers published in Korea. Using generative AI tools like ChatGPT for large-scale data analysis, the study categorizes research outcomes and identifies trends in keywords and publication years. The results show that the theory of conceptual metaphor remains a primary research focus, expanding into various fields such as literature, politics, and education. While research on conceptual metonymy and conceptual blending is relatively sparse, the potential for expansion in these areas remains significant. This study proposes specific research directions for broadening the application of these theories, particularly in educational contexts and contrastive studies. Future research should emphasize the application of generative AI and large corpus-based methods to enhance the practical use and theoretical value of conceptual metaphors and metonymies.

长期以来, 在復合词结構的研究中, 句法构词理论和语义構词理论之间一直存在着争议和分歧｡ 但实际上, 这两种不同的研究角度有一个共同点, 即过去的復合词研究缺乏“復合词语义的底层是认知场景”的认识｡ 语义的底层是认知场景, 场景中存在多个语义角色, 角色之间存在着多种语义关系｡ 以往的復合词研究侧重于语素所表达的语义角色和语义关系, 对场景中未被现实化的语义角色和语义关系则疏于注意｡ 因此未能充分展开復合词语义的深层结構｡认知语言学的相关理论是本研究的一个出发点, 也是展开讨论的重要依据｡鉴于此, 本文从“復合词语义的底层是认知场景”的认识出发, 应用认知语言学的相关理论如图形-背景理论､ 场景理论､ 框架理论､ 论元结構理论等, 对復合词‘罢工’的生成过程进行描写｡ 而且从认知角度探索復合词语义框架的压模规律｡本文对语料作了框架压模情况的考察之后, 发现復合词的结构并非都是一般所说的那样“任意的”､ “意合的”, 而是有理可据的｡我们看到, 尽管復合词的语义结構类型和句法结構类型之间不存在一一对应关系, 但是每一种语义结構类型的大部分词还是对应于某一类句法结構类型｡之所以语义结構类型和句法结構类型之间能这样基本对应, 根本原因在于, 语义结構类型和形式结構类型是从语言的形式层和语义层两个不同层面对同一对象所作的分类, 语义结構是形式结構的基础, 形式结構是处于语言底层的语义结構在语言表层的映射, 所以两者的基本对应是必然的｡ 本文解释復合词的语义表达与形式表达之间的关联, 最终沟通復合词的形式层和语义层｡復合词内部的语义本身就是復杂多样､ 拐弯抹角的, 而这样的语义内容改写成一种简单明了､ 具有一定结構规律的形式语言是难度比较大的工作｡ 本文作为把认知语言学理论应用于復合词结構研究的一种尝试, 对语义構词的探索还需要做出更多的努力｡