I have argued that there is a requirement that must be observed by any acceptable theory of musical value: the theory must respect the autonomy of music. From the point of view of the listener, the value of music is intrinsic, not merely instrumental: the listener values the experience of a musical work in itself; the experience he values can be specified only by reference to the music that is experienced in undergoing the experience; the experience is not replaceable by a different experience that offers the listener exactly what he values in the original experience—for it is the experience itself, not some separable component or effect of the experience, that is valued. It is the failure of the transmission form of the expression theory of music to meet this requirement that disqualifies the theory.
I have also claimed that it is mistaken to think of each valuable piece of music as owing its value as music to the fact that the music is expressive of emotion: this is sometimes, but not always, so. Accordingly, a viable theory of musical value must not only respect the autonomy of music, but must allow music to possess different kinds of value: in some cases, but not all, music is valuable partly because it stands in a certain relation to an emotion—the music is an expression of the emotion. I have argued that the nature of this relation cannot be clarified by basing the explanation on the model of the expression of emotion in the human voice.
If the nature of this relation cannot be elucidated by a comparison with the expression of emotion in the human voice, one reason that might be suggested as an explanation of this fact is that music can penetrate beneath the surface of emotion to its innermost core: music is not