<![CDATA[This study investigates the use of the modal verbs can and could by Japanese and Indonesian learners of English in comparison with native English speakers using spoken and written data from the ICNALE corpus. Adopting a usage-based constructional framework (Goldberg, 2006; Bybee, 2010), the study examines recurrent subject–modal pairings and employs Distinctive Collexeme Analysis (DCA) to identify subject types that are differentially associated with can and could across learner groups and modalities. Subject tokens were manually classified into four categories: first-person, second-person, third-person human, and third-person non-human. The results reveal distinct constructional preferences across groups. Native speakers consistently associated second-person subjects with can and third-person non-human subjects with could, whereas learner groups exhibited different attraction patterns, particularly involving third-person human and non-human subjects. Significant associations were especially evident in Indonesian spoken production and Japanese written production. The findings suggest that modal usage is shaped not only by the choice of modal verb but also by preferences for recurrent subject–modal constructions. The study contributes to learner corpus research by providing a constructional perspective on modality and highlighting variation in subject–modal preferences across learner and native-speaker populations.]]>
Copyrights © 2026
