ABSTRACT
The concept of a general theory of politics emanates from diverse perspectives. One distinct source has been, and continues to be, political science’s dream in its boldest moments, such as the height of the behavioral era or more recent hopes associated with rational choice analysis, of finding one dominant unifying conceptual structure for explaining political phenomena. Such a conceptual structure would, in turn, it was claimed, yield nomothetic empirical knowledge of politics. In more modest moods, the goal of the discipline has been to achieve a measure of theoretical unity either through plural, but converging, analytical frameworks or through cumulative empirical generalizations. Equally persistent, however, has been a quite different, and often oppositional, quest for a universal theory of politics or “the political.” Hannah Arendt, Leo Strauss, Sheldon Wolin, and others were committed to recovering and articulating a more qualitative sense of the essence and preeminence of political life. My concern is neither to revisit these arguments nor to examine in detail
contemporary varieties of these general positions. Rather, I will challenge the basic assumption that there can be a general theory of politics. Although I will attempt to unpack some of the complexities inherent in this argument, the core claim is that politics is a particular historical configuration of conventional or symbolic phenomena and cannot, in itself, be the subject of theoretical statements – either empirical or normative. Conventions are manifest in and constitutive of instances of action and speech, including practices such as politics, but only conventions as such can be a theoretical object. An elaboration of a general theory of conventions and human action is beyond the scope of this chapter.1 My principal concern is to clarify what we can, and should, mean when we talk about theory in political inquiry. I do, however, discuss one principal source of the uneasiness that attends the idea of accepting politics as merely a realm of historical particularities or conventional tokens. This is the problem of the practical relationship between social science and politics. My basic thesis is closely tied to two other propositions: that there is a
logical symmetry between theory and fact, and that there is a logical asymmetry between social and natural science. Stated abstractly, these are hardly
novel claims, particularly in the intellectual ambience of post-positivism, but in the course of explicating them, I will attempt both to add weight to the general claims and to distinguish sharply my formulation from the manner in which they have most often been framed and defended. As an entry into this discussion, I begin with a selective reprise of a certain line of argument in the philosophy of social science. I focus on the work of Peter Winch in part because his position is still often not clearly understood and in part because his work exemplifies many of the issues and problems that I wish to confront.
More than a generation ago, Winch argued that there could not be theories of social phenomena. His point was that only natural phenomena, with their inherent regularities, were susceptible to general causal law-like explanations. When social science is conceived in terms of the methodology of natural science, it is, he claimed, “misbegotten.” Social science properly understood, he suggested, is really a mode of philosophy or fundamentally like philosophy. Winch defined philosophy as concerned with conceptual analysis, and he argued that the task of social science is, similarly, to understand the concepts that inform conventional or “rule-governed” and “meaningful” social action. Furthermore, since social science must also specify “what is involved in the concept of a social phenomenon” and must be concerned with “giving an account of social phenomena in general,” it could be said that “many of the more important theoretical issues which have been raised in those studies belong to philosophy rather than to science and are, therefore, to be settled by a priori conceptual analysis rather than empirical research.”2
