Stylistic Alignment and Sentiment in L1–L1 and L2–L2 English Task-Based Dialogues

Abstract

Dialogue constitutes a primary context in which interlocutors coordinate their language use through the interplay of cognitive and social factors. Stylistic alignment refers to the convergence of interlocutors’ language styles during interaction, facilitating communicative coordination and fostering rapport. However, the extent to which cognitive and social factors jointly influence stylistic alignment in naturalistic dialogue remains insufficiently understood. To address this gap, the present study examined how cognitive and social factors are associated with stylistic alignment in L1–L1 and L2–L2 dialogues. Stylistic alignment between interlocutors was assessed using language style matching (LSM) metrics across multiple linguistic features, and a social affective factor, operationalized as text sentiment, was evaluated using a BERT-based sentiment analyzer. Results indicated that both L1 and L2 speakers demonstrated robust stylistic alignment across most linguistic features, with alignment strength varying by feature. In addition, sentiment was associated with stylistic alignment only for a limited number of linguistic features. Specifically, bigram range, trigram association score, and function word dispersion showed correlations exclusively in L2, whereas Delta P was linked to stylistic alignment in both L1 and L2. These findings deepen our understanding of stylistic alignment and suggest potential implications for L2 interaction and pedagogy.