Constraints on the semantic extension of onomatopoeia

This article proposes three constraints on the complex polysemy patterns of onomatopoeia. Onomatopoeic forms for thirty types of sounds were collected from Japanese, Korean, Mandarin Chinese, and English, and examined in terms of their semantic extensibility and extension types. It was found that 1) Chinese onomatopoeia is generally resistant to semantic extension, 2) onomatopoeic forms for voice in Japanese and Korean are less likely to be polysemous than those for noise, many of which show metonymical extension, and 3) many onomatopoeic verbs in English have metaphorical meanings. These three patterns can be accounted for by generalizations that associate high semantic extensibility with referential specificity, event-structural complexity, and syntactic coreness, respectively. Each of these generalizations finds some independent support, and is compatible with or complementary to a frame-semantic approach to polysemous onomatopoeic forms, which has been taken to discuss existing, rather than non-existing, cases in each language. This study is, thus, an attempt to locate sound symbolism research in the center of cognitive studies of language.