We might think that humans understand one another because they share the same set of communicative signals such as words, gestures, and facial expressions. However, that intuition neglects the extreme flexibility with which we employ our utterances in everyday social interaction. For instance, we can readily use the word "bank" to refer to money, a basketball shot, or a place to fish. Our conversational partner can usually grasp our intended meaning at its first occurrence, thanks to its context of use in an ongoing interaction. Humans thus seem to be endowed with a remarkable ability to rapidly find relevant context for selecting and interpreting communicative signals. My lab aims to understand what counts as context and how that context determines the meaning of an utterance. To this end, we combine controlled studies of live communicative interaction with state-of-the-art neuroimaging and electrophysiology (dual-fMRI, dual-EEG, intracranial EEG). The overarching goal is to characterize a fundamental and evolutionarily unique ability of our species, opening the way for understanding and treating disorders of human social interaction such as seen in Autism Spectrum Disorder.