Developmental Dynamics of Multimodal Synchronization in Children's Peer Conversations: Integrating Motion and Language

This study explores the emergence of conversational synchronization by examining both motion and language in a peer-talk corpus of 75 Hebrew-speaking children across five age groups (2;6-6;0). We use AI-driven tools to track children's physical proximity and their corresponding referential expressions. Quantitative analyses reveal that when children are closer together, they tend to use less explicit referential forms, whereas greater distance is associated with more explicit lexical forms. The relationship between distance and referential form usage is highly dynamic, with distinct patterns emerging at different developmental stages. The linear mixed-effects models reveal that children employ both stable, context-dependent usage patterns across all ages, and that some of these patterns undergo significant developmental change. This study demonstrates the power of integrating motion analysis with corpus-based linguistics to unravel the multimodal fabric of early peer interaction, illuminating how synchronized interaction evolves in parallel with communicative skills and underlying cognitive development.