The volume is the product of a 2002 conference at Yale hosted by Ian Shapiro and Rogers Smith, in which they challenged the discipline to become more problem focused and less method focused. However, as John Mearsheimer notes in the concluding chapter, the contributors “have barely addressed the issue that [the editors] laid out” (p. 389). The result is a collection that is almost entirely method focused, becoming an inadvertent but stimulating addition to the discipline's methodological/epistemological debate.

The first section concerns the challenge to become more problem focused, reflecting the editors' belief that the discipline's fundamental dichotomy is between problem-driven and method-driven research. This claim is dubious, at best. The problem-method dichotomy, if it is one, certainly does not subsume the epistemological aspects of the methodological debate, nor can it be more fundamental than the debate about the very nature (constructed vs. biologically based) of humans. Several of the contributors explicitly dispute the reality of the problem-method dichotomy, suggesting that the discipline cannot be so neatly divided (William Connolly, p. 332), that the dichotomy is “not quite usefully posed” (John Ferejohn, p. 144), or that it is simply a “false dichotomy” (Margaret Levi, p. 201). Gary Cox argues that there is no conflict between method and problem so long as the discipline remains both methodologically eclectic and collaborative (p. 182).

Smith also defines the particular problem upon which we should be focused, claiming that we “should give high (although certainly not exclusive) priority” to understanding identity formation (p. 42). This challenge, like the first, is disputed by the contributors. Anne Norton is particularly critical, calling the problem-driven scholar “the tamed domestic form of the ‘scholar activist’” (p. 68), a creature, she argues, with a dismal history. She pointedly notes that problem-driven research “serves established interests best” (p. 73), as they have the power to define what (and who) are problems, and she asks if Smith means that “each political scientist [is] to be chained to an oar in the ship of state” (p. 78).

Oddly, Shapiro and Smith themselves drift off-topic into methodological arguments. While Shapiro briefly suggests that any method is subject to dogmatic application, his only sustained critique is of rational choice theory, even though he never shows that rational choice theorists are more prone to this problem than any other methodologically committed scholars. Likewise, Smith devotes substantial space to explaining why rational choice theorists are not welcome to participate in the identity-formation research program, claiming that they have little to contribute, another dubious proposition. Despite their stated intentions, Shapiro and Smith engage in a perverse form of method drivenness, focusing not on the one method they insist on using but the one method they insist on excluding.

The second section focuses solely on rational choice methodology, but the title, “Redeeming Rational Choice Theory?” seems a disingenuous effort to plant preemptive doubt in readers' minds. These contributors clearly do not feel the need for redemption, but provide an impressive range of explanation of rational choice theory. Gary Cox helps to make sense of the approach by identifying various strains within the rational choice “paradigm,” while Bruce Bueno de Mesquita argues for using any method (including rational choice) that is amenable to mathematical modeling, because models are especially useful “for ensuring logical consistency or for uncovering inconsistencies in complex arguments” (p. 227). Ferejohn (whose chapter actually concludes the previous section, but fits as well here) considers rational choice a form of internal explanation, as it assumes that rational agents make choices that best “fit their beliefs and desires” (p. 158). Alan Ryan agrees that all human action is based on such beliefs and desires; therefore, he says, all explanation “is prima facie rational actor explanation” (p. 187). But he concludes that full explanation demands substantially more, since the really interesting questions are “why people have the weird beliefs they do … and how they come by the strange values they seem to hold” (pp. 196–97). While this section contributes little to the problem versus method debate, its breadth of explanation about rational choice theory—which is only hinted at here—makes it useful reading for anyone following the rational choice debate. Perhaps it is because Shapiro has been such a leading figure in the critique of rational choice that so many of the contributors seem to be responding to his prior arguments, rather than to the challenge he poses here.

The third section is a broad methodological discussion, aptly titled “Possibilities of Pluralism and Convergence.” The first two chapters are not sympathetic to pluralism. Alan Gerber, Donald Green, and Edward Kaplan—rejecting the call to be problem focused—suggest “a new research program for methodologists” (p. 269) that promotes the superiority of experimental research over observational studies. Observation, they argue, gives only an “illusion” of learning. Lisa Wedeen follows them with a sharp critique of the large-n methods they advocate. She identifies what large-n studies cannot tell us—such as how different identity groups within a society invest terms such as “democracy” with substantively different meanings—and argues the superiority of interpretive methods. But she fails to point out that interpretivism is also limited and cannot reveal the kind of information that large-n methods can show us. Wedeen would have done well to heed Shapiro's warning that “[o]ne of the worst features of methodological disagreement … is the propensity … to compare the inadequacies of one method with the adequacies of a second, and then declare the first to be wanting” (p. 35).

Fortunately, respect for methodological pluralism and some degree of convergence characterizes the succeeding three chapters by Rudra Sil, William Connolly, and Elisabeth Ellis. Connolly describes both methodological and substantive commitments as a type of existential faith, and he criticizes those who “feel in their bones that if only their method were triumphant they would be more secure in their faith and politics would become more explicable” (p. 345). Although we should continue to engage in efforts to convince each other, we should also “remain open to the probability that many will resist conversion” (p. 347), an attitude which he suggests will incline us toward an ecumenical generosity of spirit.

So far as we are content with pure pluralism, Connolly's advice is well taken. Ellis and Sil go farther, however, and encourage real methodological convergence. Ellis recommends a “provisional, rather than a conclusive, perspective [that makes use of] the historical, rational, and empirical modes” (p. 351), while Sil asks for a self-conscious methodological eclecticism that can produce a “constrained pluralism,” a goal “not likely to be met either by those who insist on the universality of a particular set of methodological principles and standards or by those who simply reject the very idea of discipline or rigor” (p. 327). Although purists on all sides are likely to disagree, there may be a substantial middle ground of scholars who find satisfaction in these arguments. Margaret Levi (whose chapter is actually in the second section) provides a case study of such convergence. Her approach, which she calls “rational choice analytic narratives,” employs the self-conscious eclecticism urged by Sil and seems to combine historical, rational, and empirical methods within one project, as Ellis urges.

It is ironic that despite the editors' plea for less focus on methods and more on problems, this volume is almost wholly focused on method. But the editors clearly cannot sustain their argument that the most basic disciplinary division is between method and problem. As a discipline we are most seriously divided by epistemic differences, and methodological debates, at their best, are really epistemological debates. Inadvertent or not, Problems and Methods in the Study of Politics is a useful addition to that discussion.