Most of the debate for and against aesthetic formalism in the twentieth century has been little more than a sequence of assertions, on both sides. But there is one discussion that stands out for its argumentative subtlety and depth, and that is Kendall Walton’s paper ‘Categories of Art’.1 In what follows I shall defend a certain version of formalism against the antiformalist arguments which Walton deploys. I want to show that while Walton’s arguments do indeed create insurmountable difficulties for an (...) extreme version of formalism, he has not shown that a moderate version is problematic or inadequate. (I have no space to address anti-formalist arguments other than Walton’s.2) I shall defend moderate formalism rather than put forward positive considerations in its favour, although some of its attractions will become apparent as a side-effect. I pursue the positive case for moderate formalism elsewhere.3 I. FORMAL PROPERTIES AND FORMALISMS I.1. Walton begins his paper by raising an issue about whether those who make aesthetic judgements should only be concerned with what can be directly perceived in works of art. The issue of formalism has often been described in these terms. But Walton rightly distances himself from setting up a debate in that way. He moves on to take as his target the view that ‘Circumstances connected with a work’s origin ... have no essential bearing on an assessment of its aesthetic nature’ (p. 334). (shrink)