Language is not one representation but many. A spoken utterance can be characterized as a string of phonemes, a nested set of prosodic phrases, a series of lexical items, a hierarchically-organized syntactic tree, a configuration of semantic relations, or the impetus for inferences about the speaker’s intentions. A fundamental challenge for the psychology of language is to understand the relations between these representations: the degree to which they are distinct, the ways in which they constrain one another, and the role that these connections play in language acquisition. My lab explores these questions with a primary focus on semantic representations and their relation to syntax and pragmatics. Semantic representations are central to cognitive science because they provide a window into our generative conceptual capacity. Language allows us to combine concepts from diverse cognitive domains. For example, “Two dogs ran across the road ” expresses a complex concept that incorporates a natural kind, a number, a geometrical relation and an artifact. It is unclear whether this is a unique property of linguistic representations or whether the semantics of natural language builds on a language of thought that is phylogenetically and ontogenetically prior to external language (contrast e.g., Fodor, 1975; Spelke, 2003). But on either theory, understanding how meaning is encoded in language is central to understanding conceptual combination.