Theorists have recently defended rival analyses of sound. The leading analyses reduce sound to sensations or mental representations, longitudinal compression waves, or sounding objects or events. Participants in the debate presuppose that because the features of the world targeted by these reductive strategies are distinct, at most one of the analyses is correct. In this article I argue that this presupposition is mistaken, endorsing a polysemy analysis of ‘sound’. Thus the ‘What is sound?’ debate is largely merely verbal, or so (...) I argue. All participants in this debate agree that there are the various reductions, they simply differ over which of them ‘is sound’. Yet there is no reason to think that, say, psychologists studying auditory sensations/representations, audio physicists studying sound waves, and anthropologists/ethnomusicologists studying sounding objects/events aren’t just studying different reductions of ‘sound’ despite the different explananda of their research. According to the polysemy theory of sound, we do not need to uniquely identify sound with one of these various explananda, outside of some context. (shrink)