In this article, we examine the ways in which an ESL instructor constructs contextually relevant meanings through the synchronization of speech and gesture during unplanned vocabulary explanations. Video recorded data are analysed, with focus on an in-class homework review in which students demonstrated difficulty in comprehending several key vocabulary items, which prompted the explanations. Drawing on McNeill's growth point hypothesis and the concept of thinking-for-teaching, we document how the precision timing of the teacher's gestures and their synchronization with her speech combine to make the meanings of the unfamiliar words transparent to her students. We discuss our analysis in relation to the link between internal mental functioning and social-interactive discourse, and we sketch out a number of implications for teaching, teacher education and future research.