The purpose of this study is to examine the effects of L1 similarity and semantic transparency on the meaning inference of L2 idiomatic expressions. Although these two factors have been widely regarded as important linguistic variables influencing the...
The purpose of this study is to examine the effects of L1 similarity and semantic transparency on the meaning inference of L2 idiomatic expressions. Although these two factors have been widely regarded as important linguistic variables influencing the comprehension and acquisition of L2 idioms, previous research has reported inconsistent findings regarding how their effects manifest in learners’ actual inference processes. In response, this study investigates in detail the influence of these two factors on the meaning inference of Korean idiomatic expressions by advanced Japanese learners of Korean.
Previous studies have proposed contrasting views on the relative importance of these factors, including the Idiom Diffusion Model (Liontas, 2002), which emphasizes greater inferential accuracy with higher L1 similarity, and the Dual Idiom Representation Model (Abel, 2003), which highlights semantic transparency based on the meanings of constituent elements. Meanwhile, heuristic approaches (Cooper, 1999) and the parasitic hypothesis (Cieślicka, 2015) conceptualize idiom meaning inference as an interactive process involving L1 similarity, semantic transparency, and contextual information. Against this background, the present study closely examines how these two factors are involved in the inference process by analyzing learners’ use of L1 idiomatic knowledge and literal-meaning cues.
Idiomatic expressions were categorized according to their degree of similarity to L1 into three types: partially form-similar and semantically equivalent idioms(Type 1), form-dissimilar but semantically equivalent idioms(Type 2), and idioms unique to Korean(Type 3). Each type was further divided based on semantic transparency into transparent(T) and opaque(O) subtypes, yielding six categories: 1-T, 1-O, 2-T, 2-O, 3-T, and 3-O. Fifty-three advanced Japanese learners of Korean were asked to infer the meanings of idiomatic expressions and to provide written explanations for their reasoning. The accuracy of meaning inference and learners’ use of L1 idiomatic cues and literal-meaning cues were analyzed by idiom type and item.
The results showed that accuracy was highest for Type 1-T idioms, which are both partially form-similar and semi-transparent, while Types 1-O and 2-T also demonstrated relatively high accuracy. In contrast, for Type 3 idioms, which are unique to Korean, no significant difference in accuracy was observed according to transparency. Moreover, under opaque conditions, no accuracy difference was found between Types 2-O and 3-O. These findings indicate that L1 similarity and semantic transparency do not operate independently but interactively influence inferential accuracy.
An analysis of cue usage revealed that learners appropriately utilized both literal-meaning cues and L1 idiomatic knowledge in Type 1-T, whereas in Types 2-O and 3-T, even available cues were not sufficiently exploited, resulting in inaccurate inference. This suggests that learners do not rely on a single cue, such as L1 knowledge or literal meaning, but instead integrate multiple available cues when inferring the meanings of L2 idiomatic expressions.
Based on these findings, this study suggests that idiom instruction should encourage learners to utilize both literal meaning and L1 idiomatic knowledge when encountering unfamiliar expressions. Furthermore, instructional approaches may be differentiated according to idiom types defined by L1 similarity and semantic transparency. For idioms that are relatively easy to infer, such as Types 1-T, 1-O, and 2-T, meaning inference can be facilitated by activating learners’ prior knowledge and providing scaffolding. In contrast, for idioms that are difficult to infer due to limited available cues, such as Types 2-O, 3-T, and 3-O, explicit instruction of meanings may enhance learning effectiveness. Overall, the findings of this study provide meaningful pedagogical implications for idiom instruction in Korean language education as well as for future research on L2 idiom acquisition.