By definition, more abstract categories can only shed limited (if profound) light on the empirical. This much is accepted by both Smith and Cooke. It is what to do about it that is disputed. Smith is worried (likewise Harvey, 1987) that a currently popular solution is to simply throw out the highly abstract categories as worthless. We believe that the way to move the debate forward is to spell out precisely what the abstract/concrete distinction means. In particular we have to set aside the idea, which one can too easily acquire from Sayer (1984), that this is a rigid dichotomy rather than a continuum from the most highly abstract ideas at one end, to the most concrete ideas at the other. What we deal with in fact is not an abstract/concrete dichotomy, but rather, a hierarchy of levels of abstraction (as Sayer (1984: 129) himself illustrates).