According to the author, having separate words for singing and dancing is a relatively new phenomenon in linguistics, and the concept of a performer and an audience as a distinct separation is also relatively recent. He likens it to conversation - sure in any given instance there may be people more or less involved in the dialog of a conversation, but we would all think it very strange if someone said "I only listen to conversations, I don't talk in them" in the way someone today might say "I only listen to music, I don't sing/play/dance".