Berys Gaut has recently articulated and defended a putatively anti-definitional ‘cluster’ theory of art. In the first part of this paper, I argue that Gaut's version of the cluster account is flawed. The key notion of ‘counting toward the application of a concept’ is formulated in such a way that a range of apparently irrelevant properties will count as criterial for the concept of art. Moreover, there does not appear to be any quick fix to this problem. I then turn to an exploration of the relationship between the cluster theory and the possibility of defining art. Gaut has claimed that the truth of the cluster account would provide an independent basis for rejecting the possibility of an adequate definition of art. I argue that this is not the case. But while the truth of the cluster account is consistent with definition, I also argue that it would not entail the definability of art in any ordinary sense.