This chapter takes us farther down the road of the ontology behind social explanation. It would be nice if we could spare ourselves this discussion of basic entities and their interrelations. But those political theorists who take the easier way usually adopt an implicit ontology that they do not hesitate to use in rejecting views they disagree with. We saw a simple instance of this in E. P. Thompson's rejection of a materialist explanation of the arms race. His atomist ontology was needed if his criticism was to apply to materialism generally. In this chapter I propose an alternative to that atomist ontology.

A development in the theory of explanation

About the time I thought I was mastering the technicalities of the philosophy of explanation I came across an idea that made me think my efforts had been in vain. Much of the Anglo-American philosophy of the early 1950s was still under the spell of the great revolution in logic of the early part of the century. This meant that explanation was viewed as drawing the logical consequences of theoretical generalities. The ontological impact of this subsumptionist model was idealist in nature, since explanation started not with the world and its structures but with the rarefied realm of theoretical generalities. In a paper given in 1959, Wilfrid Sellars made the disturbing suggestion that all this was wrong.