A systematic review and Bayesian meta-analysis of the development of turn taking in adult-child vocal interactions


Background: Turn taking appears to be an almost universal phenomenon in communicative behavior and requires tight coordination between interlocutors. However, there is no current consensus on how and when infants develop the ability to take turns in vocal interactions and which mechanisms are involved.

Aim: The current study aims at better understanding the current state of research on the development of turn-taking behaviors in human infants. In particular, we want to map the developmental trajectory of turn-taking abilities and identify key moderators affecting them.

Method: We performed a systematic review and Bayesian multilevel meta-analysis of studies reporting response latencies in infant-adult dyadic interactions.

Results: We found a poorly connected field (low rate of citations between papers), from which we identified 26 studies and 78 unique estimates of infant response latency. Infants display fast response latencies at an early point in development, which gradually increase up to 40 months. Infants’ responses also appear to be strongly related to the pause duration of their adult conversational partners, and are slower in groups with atypical development.

Conclusions: We identify current pitfalls and new directions of research. Specifically, we advocate for the development of shared longitudinal cross-linguistic corpora with turn-by-turn data and rich assessment of the infants' linguistic and social development. We also recommend more explicit definition and testing of computational models of the mechanisms underlying turn-taking.