Explaining interactive alignment: A multimodal and multifactorial account

- Different factors predict alignment to occur at the lexical and gestural level.
- Cumulative priming is the key factor for lexical alignment.
- Gestural alignment depends on speakers' talkativeness and the gesture's timing.
- Lexical alignment does not co-occur with gestural alignment.
- Priming alone is not sufficient to account for the phenomenon of alignment.

A growing body of evidence shows that dialogue involves a process of synchronization across speakers at different semiotic levels. In this paper, we study which factors predict this synchronization process at the lexical and gestural level. A multifactorial analysis based on a video corpus of dyadic interactions reveals that cumulative priming can account for alignment at both levels. However, there is a crucial difference between the two modalities: at the lexical level cumulative priming is the only factor with explanatory power, whereas at the gestural level, alignment is best explained by how talkative speakers are, by whether or not two gestures overlap, and whether the gestures occur towards the end of the conversations. A comparison with related studies shows that high-level, referential synchronization and low-level, behavioural synchronization seem to be governed by different rules. Models of human interaction that focus on synchronization, should take both strands of research into account.