Critical Realism, Historical Sociology, and Causation
Its advocates have argued that critical realism offers an especially good defense of the working assumptions and practices of historical sociologists (Gorski 2009; Steinmetz 1998), yet, to date, critical realism has not enjoyed that much influence on the explicit methodological frameworks available to historical sociology (including the manner in which we approach case selection, employ comparisons, and infer causes). While one need not be self-conscious about philosophy (of whatever kind) to do good empirical work, I believe there are certain ways in which critical realism can help historical sociologists better align their formal research designs with their actual research practices. These center on issues of causal complexity characterized by “multiple-conjunctural causation” (Ragin 1987) and space-time dependency. By this I mean two interrelated things. First, I believe that historical sociologists tend to view their outcomes of interest as determined by configurations of intersecting causes, rather than a single variable, factor, or the like. Second, historical sociologists generally allow that different such configurations may produce the same type of outcome. Both positions are well defended by a critical realist ontology recognizing the intrinsic complexity of open systems (Bhaskar 1998: 47; Steinmetz 1998: 173, 177).
The first implication of causal complexity concerns case selection. Taking multiple-conjunctural causation seriously, I believe, suggests strategies of case selection that are quite distinct from, and more nuanced than, the Millian methods of agreement and difference that long dominated methodological discussions in historical sociology (cf. Skocpol 1979). One possible strategy would involve selecting a case where an explanatory outcome occurred despite the absence of previously identified causes of that class of outcome—not so as to reject those causes but, rather, to expand the known range of relevant causes. Conversely, one might select a case where an outcome did not occur despite the presence of previously identified causal mechanisms, in the interest of specifying the conditions under which those mechanisms are suppressed—and thus, by implication, the conditions under which they are efficacious.
The second implication of causal complexity concerns what critical realists have called the time-space dependence of social structures or mechanisms (Bhaskar 1998: 42; Gorski 2009: 166). Consistent with a critical realist ontology of the social, working historical sociologists presume that mechanisms, rather than being universal or law-like, are historically emergent and thus bounded in space and time. Historical sociologists thus grasp the fundamental historicity of things (and their respective causal powers) in the social world (cf. Gorski 2009: 150).
I think that time-space dependence matters tremendously for how historical sociology justifies its place within the social scientific division of labor. Specifically, it suggests that one way historical sociology contributes to the accumulation of knowledge is by identifying the boundary conditions, or the conditions of possibility, for the mechanisms posited by other types of social science (cf. Hirschman and Reed 2014). This is an essential task, for absent historical sociology, such mechanisms are often misconstrued as universal, thus negating history and—by extension—naturalizing the present.
The Problem of Events
While the purpose of my presentation was to discuss some of the ways in which critical realism can “underlabor” for historical sociology, I also noted one point of genuine tension. This involves the category of the event. In the last couple of decades, historical sociologists have developed a concern for “eventful temporality,” a notion which, at its strongest, “sees the course of history as determined by a succession of largely contingent events” (Sewell 2005: 83). I wonder whether critical realism’s emphasis on causal mechanisms, by contrast, ontologically deflates events — such that events are viewed only as the surface effects of underlying mechanisms. Could this help explain the reluctance of historical sociologists to embrace a critical realist account of historical sociology?
It perhaps goes without saying that historical sociologists care about events, but they do so because they care about moments in history when things could have gone otherwise but did not (moments when plausible historical alternatives were foreclosed). Historical sociologists, I believe, often identify “events” as just such moments. Consequently, by a sort of non-formal logic of counterfactual dependence, these events become a crucial element in causal explanation itself (although they certainly do not exhaust causal explanation). Without said event, the relevant outcome would not have occurred.
My claim, however, is that for the working historical sociologist, the ontological question of whether such events have real causal powers (or, rather, are fully reducible to underlying mechanisms) simply does not matter. I would argue that an event is a legitimate part of a causal explanation in historical sociology as long as it is irreducible with respect to the specific mechanisms referenced by the explanation in question (otherwise mentioning it would be redundant). To that extent, the centrality of events in historical sociology rests more on pragmatic than ontological grounds.
It seems, then, that historical sociologists need to come to terms with a real tension. First, we should resist the causal deflation entailed by the Millian method—in favor of an approach that explicitly recognizes the multiple-conjunctural and time-space dependent (rather than universal or law-like) nature of causal processes. Here historical sociologists can fruitfully draw upon a realist understanding of mechanisms. But second, and in so doing, we should equally avoid reducing events to the surface effects of mechanisms—in favor of an approach that incorporates events and mechanisms into our explanations.
To the extent that this tension is central to any historical research, critical realism can be free to underlabor for historical sociologists without directly infringing the latter’s concern for “eventful” history.
About the Author: I am a PhD candidate in Sociology at the University of California-Berkeley. My work takes a comparative and historical approach to the interrelated processes of state formation, elite formation, and geopolitics, principally in the context of early modern Europe. I am currently working on a dissertation that analyzes the diplomatic administrations of the eighteenth-century British and French states and their connections to both the emerging European interstate system and changing elite and class structures. I also have an interest in the logic of historical and comparative research methods in sociology. It is in this context that I have become engaged with critical realism, primarily as a way of thinking through the ontological and epistemological assumptions underlying historical-comparative inquiry.
Bhaskar, Roy. 1998. The Possibility of Naturalism: A Philosophical Critique of the Contemporary Human Sciences. London: Routledge.
Gorski, Philip S. 2009. “Social ‘Mechanisms’ and Comparative-Historical Sociology: A Critical Realist Proposal.” Pp. 147–194 in Frontiers of Sociology, edited by Peter Hedström and Björn Wittrock. Leiden: Brill.
Hirschman, Daniel, and Isaac Ariail Reed. 2014. “Formation Stories and Causality in Sociology.” Sociological Theory 32(4): 259–282.
Ragin, Charles C. 1987. The Comparative Method: Moving Beyond Qualitative and Quantitative Strategies. Berkeley: University of California Press.
Sewell, William H., Jr. 2005. “Three Temporalities: Toward an Eventful Sociology.” Pp. 81–123 in his Logics of History: Social Theory and Social Transformation.
Skocpol, Theda. 1979. States and Social Revolutions: A Comparative Analysis of France, Russia, and China. Cambridge: Cambridge University Press.
Steinmetz, George. 1998. “Critical Realism and Historical Sociology: A Review Article.” Comparative Studies in Society and History 40(1):170–186.