Does proficiency influence structural alignment? A comparison of Italian native and L2 speakers in task-based dialogues

Does Proficiency Influence Structural Alignment? A Comparison of Italian Native and L2 Speakers in Task-Based Dialogues Extensive research on the psycholinguistics of dialogue has shown that conversation is a dynamic process, in which interlocutors progressively align their linguistic behaviors as communication unfolds. According to the interactive alignment model (Pickering and Garrod 2004, 2013), this convergence occurs across multiple levels—such as lexicon, syntax, and prosody—helping to reduce cognitive load and foster mutual understanding (Fusaroli & Tylén, 2016). In the field of second language (L2) acquisition, where peer interaction is considered key to language development (Mackey, 2007, 2012), a growing body of research has focused on interactive alignment in L2 dialogue and its contribution to language learning (McDonough & Mackey, 2008; Kim et al., 2019; Dao et al., 2018; Mifka-Profozic, 2023). Costa et al. (2008) argue that the basic alignment mechanisms of dialogues between native (L1) speakers should also function in dialogues involving L2 speakers. However, they note that the conscious monitoring of linguistic decisions, often heightened in L2 dialogues, can reduce alignment. This study investigates interactive alignment in L2, addressing the following research questions: (i) Do L2 speakers display more or less alignment than L1 speakers? And (ii) Does proficiency level influence alignment, with more proficient L2 speakers aligning more than less proficient ones? The data consists of Map-Task dialogues involving Belgian-Dutch-speaking students of Italian as L2 at two proficiency levels: intermediate (n=20) and advanced (n=20). These dialogues are compared with dialogues between Italian L1 speakers (n=30), sourced from a subset of the online corpus CLIPS that was collected using the same task. Following the methodology of Dao et al. (2018), the dialogues’ transcripts are coded for structural alignment, defined as the tendency for interlocutors to reuse a syntactic structure that has been previously produced in the interaction either by themselves (within-speaker alignment) or the interlocutor (between-speaker alignment). Alignment scores are thus calculated for each dialogue. Results indicate that structural alignment occurs in both L1-L1 and L2-L2 dialogues, confirming its pervasive nature in conversation (Costa et al., 2008). However, while L1s align similarly within- and between-speaker, L2s show stronger within-speaker alignment, consistent with previous findings (e.g., Dao et al., 2018), possibly reflecting task-specific constraints. Additionally, among L2 participants, higher proficiency is associated with greater alignment, suggesting that lower proficiency speakers may need to monitor their choices more closely than more proficient speakers, reducing alignment. The comparison of L2 dialogues and L1 dialogues is often missing in prior research, and it is thus a point of novelty of the study (cf. Collentine & Collentine, 2013; Dao et al., 2018; Mifka-Profozic, 2023, a.o.). Furthermore, the use of task-based dialogues allows an investigation of alignment in more naturalistic interactions compared to previous studies employing experimental priming manipulations (e.g., McDonough, 2014; Shin & Christianson, 2012; Jiang & Huang, 2015). By examining L2 dialogue across different proficiency levels, this research not only advances theoretical understanding of alignment processes but also has practical implications for enhancing language teaching and promoting effective communication in L2 contexts.