Book Review

The author has had a distinguished career as a comparative political sociologist at Brown University, where he was co-founder of the Center for the Comparative Study of Development (latterly the Watson Institute for International Studies) and is now emeritus professor. This book draws upon his very considerable experience of conducting comparative research, primarily from a sociological perspective, but also with a strong interdisciplinary engagement with economics and political science. The book is both a capstone to a distinguished career and the articulation of a powerful personal credo about the possibility of explanatory social science in the face of its perceived decline (within sociology, at least).

Rueschemeyer makes a clear distinction between theory designed to facilitate empirical research and theory engaged with normative, or interpretive, issues (that is, political and social theory, respectively). The latter is frequently organised around a canonical set of authors, with commentaries designed to set out continuities, discontinuities and influences. In contrast, the theory that Rueschemeyer is interested in is organised in terms of analytical concepts and levels of analysis. Hypotheses derived from such concepts are fallible and to be used pragmatically in the exploration of the conditions and limits of their applicability. This does not mean that such theories are not sometimes associated with particular authors and traditions of thought, but that the aim is to create assemblages of concepts with demonstrated empirical relevance that will belong to the collectivity of researchers, rather than to particular theoretical traditions.

In this way, Rueschemeyer is invoking a particular (foundational) moment in sociology associated with the arguments of Homans, Merton and Parsons, among others. In a famous essay on the history and systematics of sociological theory, Merton (1949), for example, looked forward to the codification of sociological concepts and their development in research. This would allow a proper historical understanding of antecedents clearly distinguished from the collective extension of antecedent theory, with positive consequences for both intellectual activities. Yet, despite the promise and the ‘call to arms’, sociological theory took a different turn. Merton’s hope for ‘strong sociology’ devolved into fragmented and ‘weak sociology’, with the re-emergence of what Wolfe (1992) calls ‘strong sociologists’ and individuated traditions of thought.

The story behind these developments is complex, but, for present purposes, can be quickly told. The expansion of sociology in the 1960s made it difficult to enforce disciplinary hegemony (Turner and Turner, 1990) and social movements outside the academy gave rise to challenges to the standard account of scientific epistemology, replacing it with various kinds of standpoint epistemologies. At the same time, the conceptual frameworks of social research were transformed by different ‘turns’ against positivist social science—the ‘linguistic turn’ and the ‘cultural turn’, for example.

Rueschemeyer pays these developments scant regard, except to declare his commitment to predictive scientific theory. This is expressed in positivist terms, in that he believes that “the fundamentals of valid explanation and conditional prediction are the same across all fields of inquiry” (p. 299). Ironically, this is a commitment which he grounds in common sense:

Rueschemeyer resolutely sets himself to his task, but, every so often, doubts do intrude. Problems arise around two key issues. The first concerns the extent to which a consensus on categories is required to advance scientific work. On the one hand, he refers to ‘theory frames’, suggesting that there are a plurality of competing theoretical claims. Perhaps these are to be resolved by the empirical checking of relevant facts? Something of this is evident in his allegiance to Mertonian middle-range theory and the stepping stones by which we might build theory. However, that carries the implication that the stepping stones are toward some kind of synthesis and resolution of competing claims. Rueschemeyer is sceptical of progress of this sort. His hopes have to be moderated by the fact that his book proceeded from (a largely unexplained) disappointment, namely “the disappointing achievement of empirical social and political theory when measured against the ambitions of a social science modelled after the natural sciences” (p. 299). The best that can be hoped for in the future is progress within different frames and not the integration of those frames (p. 300). Of course, where the condition of integration would be mutual consistency of claims, it is difficult to see what progress could mean in the absence of integration, as lack of integration must place a question mark over any particular claim, including that of progress. ¹

This is connected to the other key issue, namely the relationship among disciplines. The theory frames that Rueschemeyer addresses are discipline-dependent. When he embraces pluralism, he is taking Merton’s side against Parsons, who argued for a single frame of reference. The latter also used that frame of reference to organise the social science disciplines, their relations to theoretical (analytical) categories and their fields of application. On the matter of analytical theory, Rueschemeyer takes the side of Homans (interpreted via Weber). Thus, he endorses an action frame of reference as providing the core set of assumptions for explanatory social science. However, he is aware that these are available in different forms. One is the ‘rational calculus’ basis of an ‘economics’ of society, with the social sciences unified under the hegemony of economics. However, Rueschemeyer prefers a more social version of rationality, such as was provided by Homans (and further developed by Blau, Coleman or Elster).

The ‘rational calculus’ assumptions of a concept of rationality grounded in economics are to be understood as having a high degree of explanatory power, but weak empirical application. For Rueschemeyer, they should be understood as operating only under special conditions; the nature of those conditions can be specified by other theory frames using thicker, ‘social’ definitions of rationality. This kind of ‘move’ is indicated by endorsing Weber’s extended account of types of rationality. The primary development of subsidiary assumptions of action is derived from Homans in terms of ‘internal’ or ‘subjective’ elements of action that modify the ‘rational calculus’. Thus, there are chapters on cognition, norms, preferences and emotions before the treatment is extended to larger social structures, marking the transition from micro to macro levels of analysis.

However, even as Rueschemeyer sets it out, there are some paradoxical consequences. The special case of the ‘rational calculus’ is specified not by a wider theory frame that has greater explanatory power (for example, as would occur if Newtonian physics were presented as a special case of quantum physics), but by a theory frame with less predictive power but with greater empirical ‘fit’. Thus, he says, “a fully developed rational choice theory, then, must surround its core of a rational calculus model with a belt of subsidiary theoretical ideas” (p. 31). He goes on to say that “the subsidiary theories I advocate remain at present incomplete and fragmentary” (p. 32).

Whether intentionally, or not, Rueschemeyer has called to mind Lakatos’s (1970) account of the operation of a research programme, in terms of hard core and protective auxiliary assumptions. He has placed the progressive aspect of theory development in the elaboration of the subsidiary theoretical ideas, but, at the same time, he has attributed to them an ad hoc character, especially as the move is made from micro to macro analysis. Although this is presented in terms of a general action frame of reference that embraces both the ‘rational calculus’ aspect and wider ‘social’ aspects, the two (in fact, the first is singular and the latter is made up of multiple frames) are also competing. This is evident in two respects. Firstly, the core assumptions are themselves potentially subject to development—for example, in rational choice and game theory—thus changing their relation to the subsidiary theoretical ideas. Secondly, the mechanisms associated with the core assumptions lay claim to the same objects as that covered by the subsidiary theoretical ideas. This is evident in the way in which the chapters that contain the macrostructural development of the theoretical frames (for example in terms of collective action, power and co-operation, and institutions) are organised in terms of dualisms that reflect the competing claims of the rational ‘calculus’ and the wider ‘social’ theories (significantly, the rational calculus is evident across all fields, while its ‘social’ modification tends to be field-specific).

The consequence of this is a book that is both confident and tentative. The general social action framework is argued to have broad endorsement. However, that is only because endorsement includes advocates of any one of the plural theoretical frames that are aggregated together to constitute the general framework. Thus, he eschews Parsons’s (1937) claims for convergence on a single framework while invoking convergence on a general framework of plural and mutually competing elements.

What is missing is a discussion of the sociology of knowledge and the self-understanding of the social sciences. This was central to Merton’s own orientation, but is absent in Rueschemeyer’s elaboration of his project. Given that his book is predicated on a disappointment with predictive social science, this might have been the occasion for a more considered discussion of that disappointment. The few confrontations with it are quickly side-stepped. Thus, he mentions Stephan Fuchs’s (2001) argument against the individualistic assumptions that form the core of the general framework; namely, that the individual is a social construct and, thus, the individual cannot be the foundation of social theory. His response is to state: “following my resolve to keep things simple, I will again not engage in a metatheoretical discussion of these claims” (p. 34); however, the clincher is that his argument is aligned with “common sense” (p. 38) and fits with “a common sense-based methodological individualism of much actually-existing empirical research”(pp. 38–39). Fuchs is charged with ‘contempt’ for common sense but, in fact, is merely addressing the nature of the transformation of categories necessary to constitute scientific theory in the context of a similar perception of hitherto disappointing results.

Readers of this journal will easily spot that Rueschemeyer has veered across the distinction between ‘traditional’ and ‘critical’ theory first outlined by Horkheimer (2002 [1937]). The claim for ‘common sense’ is necessarily ideologically charged. Despite his protestations that embracing the ‘rational calculus’ as the theoretical core of the action frame does not commit its user to a philosophy of individualism, or a distancing from ideas of solidarity and equity derived, say, from the ontological views of the Durkheimean tradition (pp. 28–9), this simply shows an unwillingness to address the uses to which ‘usable’ theory will be put. ² Nor is it a matter of simply arguing that all explanatory theories have normative implications (and vice versa). The strong suggestion by Rueschemeyer is that explicitly normative theories tend to be weak in explanatory terms and that there is a single form to explanatory social science— that of an action frame of reference—such that all normative theories implicitly contain this framing in so far as they provide cognitive insights.

Yet, in the debates that are associated with these issues—which Rueschemeyer studiously avoids, despite them being part of his own career experience—the critique of common sense was part of the critique of systematic empiricism as a ‘pseudo-science’ (Willer and Willer 1973). This critique was also occasioned by a disappointment with scientific social science. At the same time, the ‘cultural’ force of the ‘rational calculus’ derives, at least in part, from its ideological functions. Whereas Rueschemeyer suggests that its lack of direct empirical purchase suggests that it requires a supplement, it is precisely that lack of purchase that gives it ideological power. For example, if the ‘rational calculus’ can be elaborated to provide statements of efficiency and optima provided in market competition, then ‘rational’ policies can be derived with the design of removing ‘social’ distortions. By presenting sociology as only the supplement of economics and never its critic, while allowing the autonomous development of the core, sociology is always under erasure. Or to put it in Parsonsian terms, it remains a discipline formed by its concern with residuals that it can never quite transform into positive terms.

Perhaps the political context in which I am writing this review has led me to be churlish. There is no doubt that it is an excellent and rewarding book and there is much to be learned from it. Nonetheless, I suggest reading it strategically. Firstly, read it alongside recent challenges to the idea of progressive social science, such as those of Abbott (2001a, 2001b) which are, nonetheless, committed to formal methods. Secondly, read it with an eye to the fact that the subsidiary sociological theories that are drawn into the action frame of reference as supplements to its ‘rational calculus’ core are frequently in competition with that core and derive from an alternative ontology which is its critique. An alternative scientific sociology which is not supplementary may not only be possible, it may be necessary.

Rueschemeyer’s career spans the rise and demise of a particular ‘university-knowledge regime’ associated with mass public higher education. That regime is now undergoing privatisation and marketisation, dramatically so in England. ‘Usable’ theory in the Parsons-Merton-Homans mode is no longer in the service of the knowledge regime that gave rise to it. When that regime was in its infancy, Robert Lynd (1945) posed the question: “knowledge for what?”. Under the conditions of a new privatised knowledge regime, we also need to ask, usable theory for what uses?