I'm interested in the semantics and prosody of biased questions. The prosody of such questions may reflect subtle distinctions in speaker commitment. Bias can be manifold: the speaker who asks the question can be fully committed to one or more of the alternatives raised by the question, or they can be merely biased towards them, or not be biased at all. My models are informed by data collected in a systematic way, via in-lab production experiments and online perception experiments. I use commitment-based and inquisitive semantic frameworks to model the meaning of rhetorical questions, adversative questions, declarative questions and tag questions.