What do we mean, from a philosophical point of view, when we compare physics and art? This is, in its most general form, the issue which I shall address here. Clearly, contemporary philosophy gives much credit to the idea that science and art are not essentially different activities. For instance, for Goodman and Elgin the affinities between art, science and perception make their respective philosophies appear as different guises of a "general theory of knowledge" in which the concept of symbol plays a crucial part;¹ van Fraassen repeatedly hints at the similarities between the "joint enterprises of philosophy of art, of law, of religion and of science";² and Hacking has recently suggested that the idea of "styles of reasoning" can be developed in new ways in the philosophy of science.³ The idea of an essential affinity between physics and art is undoubtedly attractive, and closely involved with contemporary shifting conceptions of the nature of physics and the nature of art. Yet its presuppositions and implications are not obvious: What conception of knowledge do we implicitly have in mind when we make a comparison between art and science? More particularly, what conception of science and what of art?