In this and the following chapter I intend to examine the various ontological issues raised in the agent–structure problem. Put simply, the ontological problems concern the nature of both agents and structures, and, since I argue they are mutually implicated, their interrelationship. My decision to deal with the ontological issues in advance of the epistemological and methodological questions reflects the scientific realist belief that the ontological questions are of a more fundamental nature than those of epistemology and/or methodology. These latter issues, however, are not to be relegated to a second-order significance. All substantive ontological claims require epistemological justification. Likewise, the delineation of a complex social ontology will imply various methodologies apropos its study. Thus, and to reiterate, ontological issues take analytical priority only because any discussion of epistemology and methodology in an ontological vacuum would be arbitrary.

For scientific realists, and contrary to positivists, theoretical terms may, and often do, refer to real entities. Thus, for any realist approach the importance of developing theoretical accounts that capture some sense of the objects they purport to describe is pivotal. But scientific realism is not ontologically dogmatic. It combines a robust depth realism with an equally robust epistemological relativism about existential claims made within a particular science. To be a realist in this sense is to take ontological commitments seriously, and to postulate theoretical entities with extreme caution; and importantly, to accept the need to defend them on ontological and not simply methodological, or pragmatic/instrumentalist, grounds.