Structure Mapping in Second-Language Metaphor Processing

ABSTRACT This study investigated metaphor processing in a second language (L2) by considering both analogy and categorization. Previous studies found that forward metaphors (e.g., “Some babies are angels”) were judged as to be more comprehensible than reversed metaphors (e.g., “Some angels are babies”) only when the sentences were presented for a long time. Based on this finding, structure mapping theory (SMT) was proposed, which posits that metaphor processing in a first language (L1) requires both the analogical alignment of nouns and categorization of the second noun. To assess SMT in L2, we tested Japanese learners of English using the following three experiments: an on-line comprehensibility judgment task, an on-line self-paced reading task, and an off-line rating task. We replicated SMT by observing a significant interaction between the viewing duration and sentence type (forward and reversed). We also identified the possibility that bilingual speakers rely on analogical alignment more than native speakers.