This article navigates the findings of conversation analysis, interactional linguistics, and related multimodal studies to summarize what we know about the grammar-body interface. It shows how grammar is fitted to sequences and trajectories of embodied activities, as well as deployed interchangeably with bodily displays, resulting in truly multimodal patterns that emerge in real time. These findings problematize both the paradigmatic and syntagmatic structures documented in verbal-only linguistics. They call for a reconceptualization of grammar as an assembly of routinized methods for the organization of vocal conduct, capable of incorporating aspects of participants bodily behavior. Data are in Estonian, French, German, Italian, Japanese and Swedish.