Multi-level Methodology and Multi-world Ontology: A Core Architecture of Realist Social Theory

It would not be exaggerated to say that the central problem of social theorizing1 lies in the relations or mediations between system (in particular, structure and function), agency (particularly, action and subject) and time (in particular, history and process). This problem arises due to a number of dichotomies. Historically, these various dichotomies can be grouped under four headings:
Local vs. Global, which refers to the dichotomy of the local scale, on which individuals interact, and the global scale, on which society as a system is identified; Static vs. Dynamic, which refers to the dichotomy of the static aspect, which is a feature of situations and structures, and the dynamic aspect, which is a characteristic of interactions and processes; Circular vs. Open, which refers to the dichotomy of the circular form of the generative mechanisms operating in focal complex of analysis - often called “recursivity” (e.g., Giddens 1984) or “self-reference” (e.g., Luhmann 1984) - and the open form of connections between levels of analysis;2 and Continuous vs. Discontinuous, which refers to the dichotomy of the assumed continuous nature of history, which is often asserted by “grand narratives” (Lyotard), and the discontinuous nature of distinctive episodes, according to which there is no such thing as general plot for a historical process.
All these dichotomies can be seen as variants of an overall micro-macro problem. Thus, the four dichotomies listed above can be termed first order, second order, third order and fourth order micro-macro problem respectively. The theme of an overall micro-macro problem may be a controversial and therefore interesting topic. However, I will not dwell here since my central concern in this paper is not the micro-macro problem itself but the architecture of Margaret Archer’s “realist social theory” (Archer 1995) that elegantly seeks to respond to this central challenge in sociological theorizing.

It is tempting to reconcile dichotomies in the form of reduction, or “conflation” in Archer’s terminology. The two main traditions in methodology of social sciences, namely, methodological individualism and functionalism, both tend to be reductionistic or, to say with Archer, conflationary in their own ways. Methodological individualism follows a kind of theorizing that Archer terms “upwards conflation,” while functionalism works with “downwards conflation.” Upwards conflation asserts the primacy of agency and views “structural properties as reducible to the effects of other actors, which are in their turn always recoverable by agency” (Archer 1995: 84). On the other hand, downwards conflation “cede(s) the explanatory rights of social theory to human biology, individual psychology, economic inevitability, evolutionary adaptation or simply to speculative metaphysics” (ibid.). In short, downwards conflation tries to establish the explanation of social phenomena at a level outside of the one of agency. Although pure reductionism is rare and many of sociological theories tries to search for linkage between the micro and the macro levels (Alexander et al. 1987), the majority of them remains reductionist in their core, for they locate their respective theoretical (or explanatory) primacy at a single level.