Abstract
This paper concerns the question of how to specify what is to count as physical for the purposes of debates concerning either physicalism or the completeness of physics. I argue that what is needed from an account of the physical depends primarily on the particular issue at stake, and that the demand for a general a priori specification of the physical is misplaced. A number of attempts to say what should be counted as physical are defended from recent attacks by Chris Daly, and a specific proposal due to David Papineau developed and extended. I argue that this approach is more than suitable for the debates for which it is intended.