Abstract
Studies have shown that learning abstract concepts (e.g., freedom compared to table) may be particularly challenging. To address concept learning, most studies use paradigms involving a single participant, such as lexical decision, match to sample, induction, and rating tasks. Here, we use an interactive task involving three family members—mother, father, and a 5-year-old child—and examine how concepts differing in abstractness are explained by the parents, either individually or together, and how the child learns and employs them. We investigate multimodal dynamics in interaction, studying gestures, movements of bodily parts, and movement coordination among interactors, in an effort to capture different embodied signatures characterizing concepts differing in abstractness. Results reveal that people gesture differently with abstract and concrete concepts, but in both cases tend to instantiate the conceptual referent, either metaphorically or iconically. Subtle differences in head movements and social touch suggest that people struggle more to understand the meanings of abstract concepts than concrete ones. Finally, with abstract concepts, coordination episodes are longer. Overall, the results provide support for the view that, with abstract concepts, people use more communicative signals and tend to converge more than with concrete concepts.