This dissertation develops a theory of contrastive topics (CTs)—what they mean, and how they are realized. I give a compositional semantics for CT constructions, built on the idea that CT marks anaphora to a complex question in the discourse. The account allows us to maintain an inclusive view of what counts as a contrastive topic, making reasonable predictions about sentences with CT phrases of difference types, in various combinations, and across various speech acts. Empirically, the dissertation focuses on contrastive topic marking in English and Mandarin Chinese. In English, CT phrases are typically realized with a “rising” prosody. I offer an explicit model that predicts the intonational features of English sentences containing contrastive topics. In Mandarin, sentences with CTs often exhibit the discourse particle -ne. I provide a detailed description of the particle’s distribution, and offer the first sustained argument that -ne is a CT marker.