earticle

영어

This study investigates the distributional patterns of the English modal verbs can and could in native (L1) and non-native (L2) spoken discourse using a multifactorial corpus-based approach. Drawing on spoken data from the ICNALE Spoken Dialogue corpus, utterances containing either can or could were annotated for their morphosyntactic and pragmatic features. A multiple logistic regression analysis revealed that modality, sentence type, and subject person significantly predict their distributional patterns, with interrogative and epistemic contexts most strongly associated with could. Importantly, a significant interaction between language group and sentence type was observed, suggesting that L1 speakers frequently employ could in interrogatives to convey politeness or mitigate face-threatening acts, while the same does not hold for L2 learners. Contrary to prior research, L2 learners in this study did not show reduced usage of epistemic modals, indicating that spoken discourse may afford more opportunities for epistemic modals than previously reported. These findings shed light on how grammatical and pragmatic factors jointly shape L2 learners’ modal usage and highlight the importance of register in understanding L2 speakers’ acquisition of modal expressions.