How can we come to grips with the theoretical tool kit of political science? The task is daunting, especially by reference to other fields, which it may be helpful to briefly survey. Economics has long assumed rational behavior (define it how you will), generally reduced to formal models (though this parsimonious view of the world is slowly breaking down in the face of challenges from psychology and behavioral economics). Sociology retains three historic macrotheoretical models of behavior, rooted in the traditions of Karl Marx, Max Weber, and Emile Durkheim (nowadays joined by rational models inherited from economics). Cultural anthropology assumes a Durkheimian model of behavior, with admixtures of Michel Foucault and poststructuralism. Biological anthropology assumes a Darwinian model. In short, each of the foregoing social science disciplines possesses one—or at most several—theoretical cores. Although these Lakatosian “research programs” are perhaps not as coherent as they were several decades ago, they are certainly recognizable.

The same cannot be said of political science. This discipline has no single founder, or even a recognized canon. Indeed, many of the most influential contemporary theorists in political science are not (or at least were not, in origin) political scientists. One thinks, for example, of Kenneth Arrow, Reinhard Bendix, James Coleman, Anthony Downs, Clifford Geertz, Albert Hirschman, Michael Mann, Barrington Moore, Gunnar Myrdal, Douglass North, Mancur Olson, Karl Polanyi, Joseph Schumpeter, Theda Skocpol, Charles Tilly, and Immanuel Wallerstein. There are strikingly few political science classics in political science in the modern era (i.e., after Aristotle, Machiavelli, and Hobbes).

Political science survives by stealing—from neighboring disciplines, from the hard sciences, from journalism, and, in an inductive fashion, from the facts of life under investigation. It is an opportunistic, syncretic field. It has, so far as I can see, no theoretical core. (Nor, for that matter, does it have a clearly recognizable subject matter. What, after all, is “politics”? Since this elision applies equally to the other social sciences, I let it slide.)

And yet, we must have some way of reducing this field of inquiry to manageable proportions. Theories are not infinite in variety, after all. And there is considerable redundancy across subfields of political science. The task undertaken by Craig Parsons in this ambitious work—to organize and systematize the theoretical tool kit of political science—is therefore absolutely essential. Yet at the same time, it is perhaps impossible to do thoroughly and neatly.

Parsons has written an interesting and engaging book that proposes a typology of arguments (theories) in political science. These, he argues (p. 12), are based on four logics of explanation. Structural claims “explain what people do as a function of their position vis-à-vis exogenously given ‘material’ structures like geography, a distribution of wealth, or a distribution of physical power.” Institutional claims “explain what people do as a function of their position within man-made organizations and rules (and within the path-dependent process implied by man-made constraints).” Ideational claims “explain what people do as a function of the cognitive and/or affective elements that organize their thinking.” And finally, psychological claims “explain what people do as a function of the cognitive, affective, or instinctual [but in any case hard-wired] elements that organize their thinking.”

Parsons is a clear and concise writer who confidently deploys an array of examples for each point that he makes—a remarkable array of works are cited in the course of this short and eminently readable work. It is an enjoyable jaunt for those entering the trade, pitched primarily to beginning students of political science (though the abstract lexicon will be daunting to some).

There are some apparent weaknesses in the proposed typology. First, the category “psychological” might be better termed genetic or cognitive, as Parsons does not mean to include individual psychological states—angry, depressed, psychotic—as a part of this explanatory trope. It refers only to genetically imprinted behavior. Second, the division between structural and institutional arguments is somewhat fuzzy. Note that most of the structures listed as part of the first category are man-made, which seems to qualify them as institutions as well. Third, the author associates rational (i.e., interest-based) explanations with structural factors, and critical-juncture/path-dependent explanations with institutional factors. It is not clear why this is so, since material structures can constitute critical junctures with path-dependent consequences, and both structural and institutional causes can be understood according to rationalist assumptions of human behavior. (The asteroid that hit the earth and destroyed the dinosaurs, according to one influential account, was surely a critical juncture. Likewise, the initiation—and the causal effect—of institutions like electoral laws may be understood within interest-based frameworks).

It becomes clear as the book progresses that Parsons is arguing for a reformation in the way that we currently use these terms. The pairing of institutionalist and path-dependent is an attempt to reformulate the concept of institutionalism so as to be identified exclusively with critical-juncture/path-dependent arguments. One wonders whether this neologism is likely to catch on, given the weight of established usage. Moreover, there may be some loss in utility in the concept if a single causal factor, for example, an electoral system, functions as either structural or institutionalist according to the kind of causal effect it appears to have (path-dependent or non-path-dependent). We already have a term designating path-dependent arguments—path dependency. Do we need another term to designate the same thing?

Granted, every typology is open to criticism, and Parsons's typology is especially ambitious and therefore sets a high mark for success. The four-part typology encompasses all work within what we have acknowledged is an extremely disparate discipline and attempts to elucidate distinct internal logics within each type. This is a tall order. Parsons is to be commended for his erudition. I doubt that anyone could do a better job with this fractious material. And I have no doubt that students will be educated—and stimulated (which is even better)—by the arguments.