Jonathan Isacoff argues that American political scientists, and in particular international relations researchers, should be much more self-critical about their historiography. Political science, he claims, is afflicted by a stubborn “positivism” that leads case-study, qualitative researchers (like Kenneth Waltz) and large-n, quantitative researchers (like J. David Singer) alike to treat historical data as if they are value-free and unchanging. This historiography blinds researchers to the normative and theoretical biases of their theories—especially a bias toward system-structural explanations for international conflict—by screening out data that contradict those theories. Moreover, subsequent revisions to the historical record undermine their findings, leaving research that is empirically disengaged and irrelevant to political experience.

Instead of replacing positivism with a “postmodern,” thoroughgoing historical constructivism and relativism, Isacoff argues for a historiography informed by the pragmatism of John Dewey. This historiography would seek neither objective historical truth nor an ungrounded pluralism of equally valid, socially constructed historical narratives, but a reconstruction of history around contemporary problems. “Rather than thinking in terms of true/untrue or good/bad history,” Isacoff summarizes this position, “we should instead evaluate the degree to which historical programmes, or narratives, can resolve problematic situations,” and “provide a more complete package of answers to questions in which researchers and society more broadly are interested,” ones “most likely to resolve ‘real’ political problems” (p. 180).

Isacoff demonstrates the problems of positivism by showing how Israeli revisionist historians undermined the static and structuralist accounts that political science positivists offered of the 1956 Arab-Israeli War, and how historians of 1960s U.S. foreign policy more sensitive to domestic politics undermined similar accounts of the Vietnam War. Isacoff offers Gary King, Robert O. Keohane, and Sidney Verba's Designing Social Inquiry (1994) as an example of how historiographical positivism is built into political scientists' methodological training, and he concludes his book with four proposals for methodological reform: 1) Researchers should be explicit about their normative assumptions regarding their problem and its various historical treatments. 2) Researchers should make a complete review of the ongoing historiography of and debates over the problems, and 3) identify the likely ideological and theoretical commitments informing them. Finally, 4) especially quantitative IR researchers should never think of their data sets as complete, should never draw a line between historical fact and historical interpretation, and should make certain that their data sets reflect the important debates among historians they have identified.

These proposals are not bad ones, even where they are unclear (how would data sets incorporate historiographical debates?), and we should take seriously any attempt to, in James Farr's words, get some “Deweyan leverage” on political science (see Farr, “John Dewey and American Political Science,” American Journal of Political Science 43 [April 1999]: 520–41). Dewey identified many of the same problems with social inquiry that Isacoff does, and his own theory of inquiry-as-reconstruction deserves more investigation. But many serious problems prevent Isacoff's argument from yielding the leverage he seeks. The two most pressing ones concern his analytical framework and his specific criticisms of political science methodology.

Isacoff's framework for analysis is critically flawed both in how he characterizes the debate between positivists and postmoderns and how he depicts his Deweyan alternative. The positivists he describes share most centrally the beliefs in a world independent of our sense data and beliefs, and in something like a correspondence theory of truth. This sounds more like a flavor of realism than positivism. Dewey's pragmatism is incompatible with the latter, but perfectly compatible with the former, so long as we reject the idea of truth as correspondence, which does not follow necessarily from realism anyway. This makes it harder for Isacoff to offer Dewey's historiography as an “alternative” to positivism.

How Isacoff describes Deweyan historiography is also questionable. It is sometimes difficult to pin down Dewey's thought, but we can rule out some accounts, including especially Richard Rorty's, upon which Isacoff's argument relies. Rorty reads Dewey as an ironic philosophical gadabout who rejected epistemology and metaphysics, had no theory of inquiry, and reduced all questions of ethics and truth to a given claim's alignment with a community's shared beliefs and its practical consequences. But this is the same picture of Dewey, only differently valenced, that Dewey's critics (among them Randolph Bourne, Lewis Mumford, Bertrand Russell, Reinhold Niebuhr, Gabriel Kolko, Christopher Lasch, and John Diggins) present. It is also one rejected by rigorous and attentive readers of Dewey, like Richard Bernstein, Michael Eldridge, Peter Manicas, Alan Ryan, Ralph Sleeper, and Richard Westbrook, for among other reasons precisely because it misreads Dewey's pragmatism as the radical constructivism and relativism that Isacoff wants to reject.

Dewey's formulations of inquiry as problem solving, and of truth as “warranted assertibility,” appeal to a community's identity, customs, and perceived needs for the sake of the subject matter and the resources of inquiry, not (as such) for standards to evaluate findings. That the world is always changing signified to Dewey that we must justify our truth claims prospectively and test them constantly against experience, not that we must dispense with truth claims altogether. Thus, Dewey could never argue that revisions to history are “malicious or beneficial depending on the configuration of political positions among communities concerned with the revisionism in question,” or that the difference between revised histories of the conflicts between white settlers and aboriginal Americans and the revised histories of the Holocaust deniers is that the former promote a community's “common good” and the latter do not (p. 42). Dewey would not have needed to appeal to community politics or some common good to draw a distinction between narratives in pursuit of truth based on evidence and reliable testimony, and history in denial of truth based on fabricated evidence and outright lies.

Even a sound account of a Deweyan pragmatic alternative would offer little purchase on the research problems Isacoff presents, however. Isacoff's critique boils down to this: that positivist IR researchers “assume historians to be atheoretical” and “treat the narratives of historians as a cumulative sum of unproblematic data” (p. 95), and that they are prone to “stylized interpretation and unrefined culling of historical data” to suit their theoretical assumptions (p. 161). But the specific criticisms Isacoff presents—the failure to consider alternative theoretical explanations, the low intercoder reliability across sets of quantitative data, the lack of updates to data sets based on revised histories, the sampling of data on dependent variables, and so on—all just as credibly could have come from his positivists consistently applying their own standards. If there is a case to be made that the absence of Deweyan pragmatic historiography prevents them from doing so, it is not here. An approach to political inquiry based on the precepts of Designing Social Inquiry has its problems, but these are not among the sins we can lay at King, Keohane, and Verba's feet alone.

The disengagement and irrelevance of political science research is in great part due to the influence of positivism—or, more precisely, to the influence of the debate between what Richard Bernstein has called a “positivist temper” in political science and its unreflective opponents (see Bernstein, The Reconstruction of Social and Political Theory, 1976). Dewey's pragmatism can explain why this influence is so pernicious. But Deweyan pragmatism would also examine how the institutional and political contexts within which political science operates—the research university and American political liberalism—have encouraged positivism's influence on the way that political scientists craft their disciplinary identity around their status as “scientists” and their choice of methodological exemplars. Isacoff's account downplays or omits these factors. And pragmatism would inform a far more radical criticism than Isacoff's of how political scientists do political inquiry, starting with the very identification of social inquiry as an explanatory and predictive science. The author's accounts of positivism in Chapter 1 suggest that he finds an explanatory-predictive model problematic, but he does not pursue that line of criticism, not even where that model informs King, Keohane, and Verba's demotion of qualitative research to an intermediate stage of properly scientific quantitative research.

Despite my sympathy for Isaacoff's goals and my shared affinity for Deweyan pragmatism, I believe that these limitations severely weaken Isaacoff's argument. The problems that drive this argument are real and deserve our attention. But this book does not give them the critical attention they require.