Investigating the Source of Idiom Transparency Intuitions

摘要

Two accounts have been given for the formation of idiom transparency intuitions. For most researchers, an idiom is transparent to the degree that a link can be found between its form and meaning. Cognitive linguists agree with the aforementioned view, but claim there is an additional source of transparency intuitions. They claim that transparency is partly the degree to which features inherent in an idiom (e.g., conceptual metaphors that are considered to underlie it) are seen as contributing to the idiom's meaning even before someone learns it. This article reports the results of an experiment which examines whether the cognitive linguistic claim about a hybrid source of idiom transparency intuitions is correct. Advanced second language learners of English guessed at the meaning of unknown idioms presented in or out of context. The results are congruent only with the hybrid view of idiom transparency. They indicate that idiom-inherent features contribute to the formation of transparency intuitions. In particular, although the idioms were unknown to participants, (1) high-transparency idioms received a significantly smaller number of interpretation types (not tokens) than low-transparency idioms in both context conditions, and (2) a significantly higher-than-chance number of correct idiom interpretations was given to high- rather than low- transparency idioms. However, the definition of transparency as the extent to which the meaning of an idiom can be guessed correctly, which is supported by some cognitive linguists, is not supported here: (a) the level of association between transparency and correct interpretations was low, and (b) when context was provided, the difference in the number of correct definitions between high- and low-transparency idioms was not significant.