Causes of War: Power and the Roots of Conflict.
Stephen Van Evera explicitly sets out to accomplish two tasks. The first is to present a set of five hypotheses on the causes of war grounded in "misperceptive fine-grained structural realism" (p. 11). He lists (1) false optimism about the outcome of a future war, (2) perceived first-mover advantages, (3) opening and closing windows of opportunity and vulnerability, (4) cumulativity of resources, and (5) beliefs about the offense-defense balance. He then develops 23 related hypotheses. The second task is to test some of the major hypotheses (the second, third, and fifth) against a small set of cases. He succeeds at the first task but is not so successful at the latter. He also briefly speculates on the effects of the "nuclear revolution."
The major contribution of Causes of War lies in Van Evera's demonstration of the similarity of what heretofore have been presented as competing research paradigms: realism and constructivism. Realists traditionally have focused on the material bases of the balance of power and its presumed effects on international relations; constructivists have emphasized the role of ideas and their socially constructed nature. Van Evera expands the explanatory and predictive power of realism by incorporating leaders' beliefs and misperceptions about "fine-grained power" relations between states (p. 7). In doing so, he demonstrates, however unwittingly, the underlying similarities of constructivism and realism. This is a particularly valuable contribution to international relations theory.
As Brooks and Wohlforth point out, "establishing a strong, independent role for ideas will be particularly difficult when material constraints are especially significant and/or when there is relatively little lag between material and policy changes" (Stephen Brooks and William C. Wohlforth, "Power, Globalization and the End of the Cold War: Reevaluating a Benchmark Case for Ideas," International Security 25 [Winter 2000-2001]). Van Evera manages this well, within the inferential constraints of his method. By noting that "realism thus is most powerful--if we repair it by shifting its focus ... from power itself to national perceptions of power" (p. 9), he blurs the distinction between realists and constructivists, whose key contribution has been to note the explanatory power of socially constructed national beliefs, or ideas about power and relations thereof (e.g., Alexander Wendt, Social Theory of International Politics, 1999).
There is much to like about the book, but there are problems. From an empirical perspective, several weaknesses stand out. Van Evera uses a method perhaps best termed historical constructivism, an empirical analog to the theorized social construction of ideas at the heart of the paradigm. The hypothesis tests consist of clever historical constructions based on carefully selected facts and recollections through which the author demonstrates that beliefs (also referred to as ideas or misperceptions) lie at the core of his microfoundational realist agenda. Van Evera commits four inferential sins that further obscure the eroding distinction between his brand of realism and social constructivism to the extent that, empirically, any remaining differences might be considered superfluous.
First, the observations from which the hypotheses are inferred are the same ones used to test them. For example, when testing his "jumping the gun" hypothesis (p. 63), Van Evera presents the historical evidence in support of the test in footnote 107. Rather than reveal an historical insight, however, this note simply refers the reader to the evidence presented in note 55 (p. 49) and notes 70 and 71 (p. 52), found in the section where the hypothesis was developed. Van Evera is well aware of this and other inferential problems but seems blissfully unconcerned. "The orthodox methodology creed ... requires that cases not be selected on the dependent variable.... It warns against testing theories with cases from which the theory was inferred.... It warns against selecting atypical cases loaded with the causal phenomenon.... I have never found these rules useful, and my case studies break them all. Readers can judge if my recalcitrance did any harm" (p. 12). What do we lose by adopting this historical constructivist method? Foremost is the ability to distinguish between cause and effect and the ability to falsify judgments thereof. From a methodological perspective, this is not Wendt's positive social constructivism but historical constructivism in the best tradition of Foucault at his most cynical (see Michel Foucault, Archaeology of Knowledge, 1982).
Second, both the tests and hypotheses rely completely on a review of the secondary literature on the origins of World War I and others. We become captive to the judgment of the historians Van Evera carefully selects to construct his argument. On World War I, for example, a more balanced review of the historiography reveals that the Russians' intentions and degree of military optimism were not as clear as Van Evera claims they were. As Goemans points out, "on the one hand, the secondary literature often asserts that like France and Germany, Russia expected the war to be short and victorious. On the other hand, closer study reveals that well-informed decision-makers were often much less sanguine about Russia's relative strength and the duration of a war" (Hein Goemans, War and Punishment: The Causes of War Termination and the First World War, 2000, p. 127, in particular note 9).
Third, Van Evera does not randomly sample the facts he cites from the population of states that might go to war. Rather, it appears the evidence has been selected both on the value of the dependent variable and also to bolster the author's argument, not necessarily to get to the truth of the matter (e.g., note 118, p. 66). Since we cannot identify the general population of crises at risk of escalation to major war, it becomes difficult to have confidence in the few policy recommendations based on the findings.
Fourth, the evidence typically consists of individuals' observations, but the stories being tested are state-level arguments. Why is this of concern? Early on, Van Evera makes clear what his brand of realism is all about: citing Robert Keohane, he summarizes: "1. States are the most important actors in world politics, 2. States are unitary rational actors, and 3. States seek power ... and they calculate their interests in terms of power" (note 11, p. 7). If states are the unitary rational actors Van Evera assumes them to be, then we should search for evidence at the state level or in observations about state characteristics. Instead, most of the evidence is individuals' recollections of how the tide of myopic decisions made by others drove events forward. We could just as easily construct a story using Van Evera's evidence that is more consistent with the psychological explanations of dispute escalation found in Richard Lebow's Between Peace and War (1984).
The historical constructivism in Causes of War will prove seductive to many because of the polished prose that leads the reader though what appear to be powerful tests of myriad hypotheses. Although Van Evera's construction of fine-grained realism succeeds on this level, it fails on a more rigorous epistemological level of positive social science. In the end, readers will likely take from the book what they bring to it. Realists may find the emphasis on ideas somewhat puzzling, and positivists will be deeply disappointed by the outright flaunting of scientific rules of inference and in the end will find little of systematic substance to take with them.
Allan C. Stare III, Dartmouth College
|Printer friendly Cite/link Email Feedback|
|Author:||Stam, Allan C. III|
|Publication:||American Political Science Review|
|Article Type:||Book Review|
|Date:||Mar 1, 2001|
|Previous Article:||The Moral Purpose of the State: Culture, Social Identity, and Institutional Rationality in International Relations.|
|Next Article:||Globalizing Concern for Women's Human Rights: The Failure of the American Model.|