There is only one firm intuition here, and it is that work identity is preserved just so long as structural integrity is preserved. Indeed, so strong is that intuition that we do not even require absolute preservation of structure; that is to say, we only require that structural relations be preserved. . . Performing a Bach fugue with a choir of kazoos may, of itself (although not necessarily), make it a very bad performance; of that there can be no possible doubt. But it cannot, of itself, make the performance a performance of something else.
(Kivy 1988a : 45, 55)
Consisting essentially of nine repetitions of the same sinuous melody and countermelody, varied almost exclusively through changes in instrumentation, [Ravel's Bolero] would make no sense if rendered on, say, two pianos. . . Nor would it make the sense Ravel gave it if comparable instrumental variety was retained, but not the particular sequence of changes that Ravel prescribed. For example, nothing can substitute for the heightening of sultriness and sassiness Ravel achieves by introducing the tenor, soprano, and sopranino saxes as carriers of his countermelody about halfway through, after all the more reserved and conventional woodwinds have had their say.
(Levinson 1990a : 247)
In this chapter I discuss elements crucial to the identity of musical works. I begin with sound structures, because it is generally agreed that these are central to any piece's singularity. Later, I ask if its instrumentation also contributes essentially to a piece's being the one it is. Finally, I consider issues of ontological contextualism—that is, I enquire whether relations between a piece's raw musical content and the socio-musical setting in which it is created generate features vital to its identity. Before getting down to these tasks, it is important to draw attention to several methodological issues.
When we set out to characterize the nature of limestone, we search for properties displayed by all and only its instances. If not all bits of