The role of pragmatic competence is not often emphasized in research on the acquisition of syntactic structure and lexical semantics. But since children learn their first language in the context of communicative interactions (not from, say, a text corpus of child directed speech), they must constantly engage with input with multiple levels of meaning, both semantic and pragmatic. In this talk, I discuss two examples of how pragmatic factors affect learning: the acquisition of belief reports and speech reports. Belief reports pose a potential difficulty because the same word ('think') and structure are used to convey a variety of different speaker meanings, not necessarily emphasizing belief states. I argue that although that complication does affect children's understanding of belief reports--at least in experimental situations--children actually manage to acquire the literal meaning of belief reports much earlier than has been claimed previously. Speech reports are potentially problematic because of structural ambiguity: direct speech reports (Joan said, "I'm an idiot!") and indirect speech reports (Joan said I'm an idiot!) are often string identical. Some recent evidence suggests that children cannot distinguish the two structures until as late as 9-11 years old. I argue that pragmatic context plays an important role in licensing the two structures, and that children may take advantage of that to distinguish them from an earlier age.