The art of arranging sounds in time so as to produce a continuous, unified, and evocative composition, as through melody, harmony, rhythm, and timbre.
Well, I've had more time to think about it, and the definition above is pretty close to speaking or at least reciting poetry. There are prosodic elements to even normal speech. In English tone is not used phonemically (e.g., to distinguish lexical items) but supersegmentally to add or affect the syntax of an utterance.
I think the evocative part applies to all art forms. Somehow consuming the work of art produces some kind of emotional response in the consumer. Continuous and unified are problematic, too. What is the longest rest that does not cut a piece into two parts? Also, different songs released on the same CD. They can be more closely unified (concept album, opera, etc.) than other pieces by the same composers/performers.
