Millennial Reflections on International Studies 9780472023943, 0472023942

Citation preview

Page 1 →

THE ESSENCE OF MILLENNIAL REFLECTIONS ON INTERNATIONAL STUDIES Michael Brecher and Frank P. Harvey When Michael Brecher was introduced to international relations (IR) at Yale in 1946, the field comprised international politics, international law and organization, international economics, international (diplomatic) history, and a regional specialization. The hegemonic paradigm was realism, as expressed in the work of E. H. Carr, Arnold Wolfers, Nicholas Spykman, W. T. R. Fox, Hans Morgenthau, Bernard Brodie, and others.1 The unquestioned focus of attention was interstate war and peace. By the time the other editor of this collection, Frank Harvey, was initiated into international relations at McGill in the late 1980s, the preeminent paradigm was neorealism,2 but there were several competing claimants to the “true path”: institutional theory,3 cognitive psychology,4 and postmodernism.5 And by the time he received his doctoral degree, other competitors had emerged, notably, critical theory,6 constructivism,7 and feminism.8 The consequence, at the dawn of the new millennium, was a vigorous, still inconclusive, debate about the optimal path to knowledge about international studies (IS), most clearly expressed in the views that it is a discipline—international relations (IR) or world politics—like economics, sociology, anthropology, history, or that it is a multidisciplinary field of study; the “big tent” conception of the premier organization, the International Studies Association (ISA). It was in this context that the Millennial Reflections Project was conceived. The origin and rationale of the idea may be found in the central theme of Michael Brecher's presidential address to the ISA conference in Washington in February 1999: “International Studies in the Twentieth Century and Beyond: Flawed Dichotomies, Synthesis, Cumulation.” The next stage was the creation of a set of ten millennial reflections theme-panels by Michael Brecher, then ISA president, and Frank Harvey, the program chair for ISA 2000: these panels served as the highly successful centerpiece of the Los Angeles conference in March 2000. Soon after, we enlisted the enthusiastic support of the University of Michigan Press for the idea of publishing revised and enlarged versions of these conference papers. Most of the participants Page 2 → in the Los Angeles panels readily agreed to revise and enlarge their papers. A few other papers were invited. The result is this volume and the accompanying set of four shorter, segment-focused volumes, prepared for the benefit of teachers and students of IS in colleges and universities everywhere. Whether a discipline or a multidisciplinary field of study, IS has developed over the last half century with diverse philosophical underpinnings, frameworks of analysis, methodologies, and foci of attention. This diversity is evident in the papers that were presented at the panels at the 2000 Los Angeles conference and revised for publication in this state-of-the-art collection of essays on international studies at the dawn of the new millennium. In an attempt to capture the range, diversity, and complexity of IS, we decided to organize the forty-four “think piece” essays into eight clusters. The mainstream paradigms of realism and institutionalism constitute the first two concentrations; critical perspectives (including critical theory, postmodernism, and constructivism); feminism and gender perspectives; methodology (including quantitative, formal modeling, and qualitative); foreign policy analysis; international security, peace, and war; and international political economy make up the remaining six. The raison d'être of the Millennial Reflections Project was set out in the theme statement of the Los Angeles conference, titled “Reflection, Integration, Cumulation: International Studies Past and Future.” As we noted in that statement, the number and size of subfields and sections has grown steadily since the founding of the International Studies Association in 1959. This diversity, while enriching, has made increasingly difficult the crucial task of identifying intrasubfield, let alone intersubfield, consensus about important theoretical and empirical insights. Aside from focusing on a cluster of shared research questions related, for example, to globalization, gender and

international relations, critical theory, political economy, international institutions, global development, democracy and peace, foreign and security policy, and so on, there are still few clear signs of cumulation. If the maturity of an academic discipline is based not only on its capacity to expand but also on its capacity to select, the lack of agreement within these communities is particularly disquieting. Realists, for instance, cannot fully agree on their paradigm's core assumptions, central postulates, or the lessons learned from empirical research. Similarly, feminist epistemologies encompass an array of research programs and findings that are not easily grouped into a common set of beliefs, theories, or conclusions. If those who share common interests and perspectives have difficulty agreeing on what they have accomplished to date or do not concern themselves with the question of what has been achieved so far, how can they establish clear targets to facilitate creative dialogue across these diverse perspectives and sub-fields? With this in mind, our objective was to challenge proponents of specific paradigms, theories, approaches, and substantive issue areas to confront their own limitations by engaging in self-critical reflection within epistemologies Page 3 → and perspectives. The objective was to stimulate debates about successes and failures but to do so by avoiding the tendency to define accomplishments with reference to the failures and weaknesses of other perspectives. It is important to note that our call to assess the state of the art of international studies was not meant as a reaffirmation of the standard proposition that a rigorous process of theoretical cumulation is both possible and necessary. Not all perspectives and subfields of IS are directed to accomplishing cumulation in this sense. Some participants found the use of such words as synthesis and progress suspect, declaring in their original papers that they could not, or were not prepared to, address these social science-type questions. We nevertheless encouraged these individuals to define what they considered to be fair measures of success and failure in regard to their subfield, and we asked them to assess the extent to which core objectives (whatever they may be) have or have not been met, and why. Our intention obviously was not to tie individuals to a particular set of methodological tenets, standards, assumptions, or constraints. We simply wanted to encourage self-reflective discussion and debate about significant achievements and failures. Even where critiques of mainstream theory and methodology are part of a subfield's raison d'être, the lack of consensus is still apparent and relevant. As a community of scholars, we are rarely challenged to address the larger question of success and progress (however one chooses to define these terms), perhaps because there is so little agreement on the methods and standards we should use to identify and integrate important observations, arguments, and findings. To prevent intellectual diversity descending into intellectual anarchy, the editors set out “guidelines” for the contributors, in the form of six theme questions, or tasks. The panelists were requested to address one or more of these tasks in their essays: 1. Engage in self-critical, state-of-the-art reflection on accomplishments and failures, especially since the creation of the ISA more than forty years ago. 2. Assess where we stand on unresolved debates and why we have failed to resolve them. 3. Evaluate the intrasubfield standards we should use to assess the significance of theoretical insights. 4. Explore ways to achieve fruitful synthesis of approaches, both in terms of core research questions and appropriate methodologies. 5. Address the broader question of progress in international studies. 6. Select an agenda of topics and research questions that should guide the subfield during the coming decades. The result, as is evident in the pages to follow, is an array of thought-provoking think pieces that indicate shortcomings as well as achievements and specify the unfinished business of IS as a scholarly field in the next Page 4 → decade or more, with wide-ranging policy implications in the shared quest for world order. Readers will no doubt derive different conclusions from the various contributions. Some will observe that divisions within and

across subfields of international studies are so entrenched that constructive dialogue is virtually impossible. Others will conclude that there is much more consensus than might have been imagined. In either case, the need for selfcritical assessment among IS scholars is imperative as we enter the new millennium. The essence of the assessments of each essay in the eight clusters is summarized in this introductory chapter. At the end of the volume, the editors present findings on the six theme questions about IS paradigms, methodologies, and the three broad substantive research areas—foreign policy analysis; international security, peace, and war; and international political economy.

Realism The first cluster of millennial reflections on international studies comprises six papers by proponents, critics, and a revisionist of realism, the dominant paradigm in the field during most of the Westphalian era, 1648 to 1990: John J. Mearsheimer, Joseph M. Grieco, John Vasquez, Kalevi J. Holsti, Manus I. Midlarsky, and Patrick James. Mearsheimer In “Realism, the Real World, and the Academy” John Mearsheimer presents a forceful, unremitting defense of realism as the paradigm wave of the future, no less than in the past. Responding to optimists who consider realism passé—such as President Clinton's 1992 view that “the cynical calculus of pure power politics … is ill-suited to a new era” and John Mueller's Retreat from Doomsday: The Obsolescence of Major War and Francis Fukuyama's The End of History, both published in 1989—Mearsheimer declares: “The claim that realism has bitten the dust is wrong. In fact, realist theories are likely to dominate academic debates about international politics for the next century, much the way they have since at least the early days of the cold war.” Mearsheimer's core thesis is that “despite the end of the cold war, the basic structure of the international system remains largely unchanged.” And since that is so, “we should not expect state behavior in the new century to be much different from what it was in past centuries”; nor, indeed, has it changed. He cites an array of evidence of different types: U.S. involvement in two full-scale wars in the 1990s—Iraq in 1991 and Serbia-Kosovo in 1999—and it came “dangerously close” to war against North Korea in 1994; persistent security competition and war in South Asia, the Persian Gulf, and Africa—surprisingly, he does not mention the Arab-Israel domain; larger U.S. defense Page 5 → expenditures in 2000 than the next six states combined; and the threat of war in Northeast Asia and Europe, the two regions of great powers where critics claim that realist logic is no longer relevant. In the former, the United States maintains one hundred thousand troops, “reducing the need for arms buildups and deterring the rise of hegemonic forces,” according to Joseph Nye, a respected nonrealist scholar and an architect of U.S. policy in Northeast Asia. War is a distinct possibility in that region, argues Mearsheimer—the United States against North Korea or against China over Taiwan: “NEAsia is a dangerous place, where security competition is a central element of interstate relations.” Even in Europe, he argues, it is NATO, a military institution, not the European Union, that ensures the security and stability of its members. He cites the following: the remarks of Russian president Putin in a 2000 policy statement—“Military force and violence remain substantial aspects of international relations”; the presence of one hundred thousand U.S. troops in Europe; and the pervasive view among many European political and military officials interviewed by Robert Art that “if the Americans removed their security blanket from Europe … the Western European states could well return to the destructive power politics” of the past, a view echoed by Christoph Bertram, a leading German strategic thinker. “The bottom line is that security competition and war between the great powers have not been burned out of the system,” writes Mearsheimer. States will “continue competing among themselves for power,” making realism highly relevant for the international politics of the twenty-first century. Mearsheimer's paper concludes with a blunt attack on academic critics of realism: “The most serious threat to realism comes not from the real world, but from inside universities, where dislike of realism is widespread and often intense, especially among liberal international relations theorists.” In particular, he cites John Vasquez,

another contributor to these millennial reflections, for intolerance, hostility, and ad hominem attacks on realists. Grieco Joseph Grieco, a self-critical realist, provides a much less uncompromising defense of the longtime preeminent IR paradigm than John Mearsheimer in his “Modern Realist Theory and the Study of International Politics in the Twenty-first Century.” He notes three fundamental insights of modern realism: insight into “the impact of the anarchical structure of the international system on the preferences, strategies, interactions, and domestic institutions of states”; insight into “the impact of inequalities on international affairs, and in particular, inequalities in power”; and insight into “the importance of continuity in international affairs, but [realism] also alerts us to the pervasiveness of change.” Realism is not without flaws, however, according to Grieco, and must address four problems or challenges: first, the need for “greater acknowledgment Page 6 → that, however important realist-posited variables might be, there are other factors that, under certain circumstances, might vitiate their impact … [and] we may wish to expand our range of output variables to include the consequences of state behavior”; second, are states “security or power maximizers, and is there an observable difference between the two goals?”; third, “even if the liberal institutional perspective is incorrect … there is still much to be done by realist-informed scholars about the question of why many states appear to have a preference for international institutions”; and fourth, “there is much more that can be done to bring about … an integration of realist-posited and domestic structural variables” along with research into the possible linkage of democratic peace and democratic hegemons, notably the United Kingdom in the nineteenth century and the United States in the twentieth century—this “might yield a richer appreciation of both modern realism and domestic structural approaches to international politics.” For Grieco, as for Mearsheimer but not for Vasquez or K. J. Holsti, as the following makes evident, “realism will remain the core approach in the field of international relations.” However, his balanced, self-critical assessment displays both pride in realism's accomplishments and an awareness of its shortcomings and the need to address them. Vasquez In a forthright critique, “Realism and the Study of Peace and War,” John Vasquez reaffirms his judgment that realism is a “degenerative research program” and, therefore, unworthy of future work: “it is precisely at the heart of the paradigm—on the questions of peace and war—where [realism] has failed … to provide more accurate, more powerful, and more relevant answers” than the idealist paradigm. His criteria for assessment are explanatory power, empirical accuracy, and policy relevance: on all of these he finds realism deficient. Waltz (Man, the State, and War and Theory of International Politics) is criticized on two grounds: first, that global anarchy, his fundamental “permissive cause” of war, is a constant and, logically, a constant cannot explain a variable (war); and second, that “a permissive cause implies that there are other causes, sufficient conditions, or probabilistic factors that bring war about in any one specific instance. Waltz's explanation is inadequate because it fails to specify what these are.” Gilpin fares better, for he posits a variable—the decline of a hegemon and the rise of a challenger—to explain a variable, war; but “his explanation applies only to a small set of hegemonic wars.” Morgenthau's classical realism does best of all, argues Vasquez, because it “comes close to predicting that war is a constant among states as they struggle for power and what we have in international politics is constant warfare.” That is, for Morgenthau, states are constantly preparing for war, engaging in war, or recovering from war. But even this sweeping view of war as constant is rejected—on empirical grounds, specifically three pieces of evidence. Page 7 → One is “the presence of peaceful eras, especially among major states”—that is, 1816–48 (Concert of Europe), 1871–95 (Bismarck's European order), 1919–32 (League of Nations), and 1963–(détente). Another is the

“presence of peaceful dyads,” based on Correlates of War Project data—seventy-five interstate wars from 1816 to 1992, with four hundred pairs of states in 1816 and about eighteen thousand in 1992—and many democratic dyads that do not fight each other. Third is “the presence of zones of peace”—the Nordic peace system, North America after 1848, and most of the Western Hemisphere and sub-Saharan Africa. Vasquez also contends that realism fails to explain the reality of peace, much less than rules, norms, and institutions. Thus he concludes: “the core questions that the realist paradigm has set for the field—the causes of war and the struggle to achieve peace—have not been adequately answered by classical realism [Morgenthau], neorealism [Waltz, Gilpin], and offensive realism [Mearsheimer].” Holsti (K. J.) Kalevi J. Holsti, too, is highly critical of the realist paradigm in his millennial reflections paper, “Performance and Perils of Realism in the Study of International Politics.” He notes four functions of realism, as of any theory: realism is (1) “a compressed, or shorthand, description of the essential characteristics of international politics”; (2) “a diagnostic tool for analyzing the sources of international conflict and derivatively for identifying the conditions for peace and security”; (3) “a set of prescriptions”; and (4) “a normative theory that posits the survival and security of the political community as the ultimate political value.” While acknowledging that “as a descriptive enterprise, realism has offered a great deal about how some states behave some of the time,” Holsti poses a series of questions arising from patterns of diplomatic-military practice—for example, the ubiquity of alliances, the absence of security dilemmas in South America's IR, the lack of preparations by most states to defend borders, the absence of hegemons in South America or Southeast Asia, and the virtually nonexistent prospects of war in Western Europe since 1945. These lead him to the view that these are “important trends in the texture of international politics that challenge realist descriptions and explanations.” He terms “the most damaging lacuna in realism [as] its dismissal of interest in revolutions,” which “are based on ideas.” “Realism is also incomplete within its own frame of reference … offering only a limited roster of strategies for coping with the consequences of anarchy or human psychology.” Holsti also finds realism wanting as a diagnostic tool, except in special cases, notably nineteenth-century Latin America and the protracted conflicts/enduring rivalries between India and Pakistan, Greece and Turkey, and Israel and the Arabs. More generally, he notes the absence of major-power war since 1945, with the partial exception of Korea (1950–53), the lack of war in South America since 1942, and the sharp decline in territory-based conflicts since 1945. Page 8 → As for its prescriptive value, “the ultimate problem of realist prescription is that the very policies supposed to maintain international peace and security may actually lead to the unintended consequence of war.” Finally, realism is deemed more successful as a normative theory than in its other three functions, but more for settled (Euro-American) societies than revolutionary ones (Third World). The main culprit is the U.S. realists: “by divesting themselves of normative content, American versions of realism became mechanistic, deterministic, and hopelessly ahistorical.” Holsti is, however, much more generous than Vasquez toward realism, saying that “it has done what any good theory should do: it has stimulated further research and promoted improvement of guiding assumptions, main tenets, and shorthand characterizations.” Moreover, Holsti acknowledges realism's continuing, important role in the explanation of world politics: “any understanding of past, contemporary, or future international politics is incomplete without the problem selection, the insights, the generalizations, and the explanations of realism. But realism is a perspective on politics, not the whole story. Even as a perspective, it suffers from serious descriptive, explanatory, and therefore prescriptive problems. But to pretend that it has nothing to tell us would be an immense error.” In conclusion, he notes important competitors as a guide to international politics—the English School, liberal internationalism, constructivism, and Kantian perspectives on civil society.

Midlarsky Manus Midlarsky is a realist without pretense to paradigm omniscience. The title of his millennial reflections paper, “Realism and the Democratic Peace: The Primacy of State Security in New Democracies,” specifies its primary thesis. He notes, correctly, “the virtual absence of the international security setting” in the vast research done on the democratic peace. And he postulates that the terms of the debate over which has temporal priority, peace or democracy, shift to security as the prime condition of democracy. “Clearly,” he argues, “the international setting is the key variable driving the relevant decision makers.” Midlarsky notes, with approval, “the inverted democratic peace theory” that “turns the theory on its head,” arguing that “it is the condition of international peace that yields democracy,” rather than the reverse. He explores this alternative explanation through a series of intriguing figures, pointing to support of his view that the pervasive “domestic explanations of the democratic peace [are] misguided to the extent that [they do] not include explicitly international factors.” Instructive case studies of Czechoslovakia, Poland, the Baltics, Finland, Ireland, and Austria in the inter-world war period provide Midlarsky with empirical evidence for his thesis. In his view, the emphasis on economic development “does not provide as strong a distinction between successful and unsuccessful democracies as does state security.” Page 9 → His conclusion, which challenges the pervasive democratic peace thesis, is in the realist tradition: “Among the several possible causes of democracy's success in Europe [1919–39], the only one that consistently distinguishes between the successes and failures of new democracies is state security.” This, in turn, leads him to call for multiple strands of research—“the contributions of more than one paradigm as well as the domestic and international settings.” James The concluding paper in the realism cluster, “Systemism and International Relations: Toward a Reassessment of Realism,” focuses on what Patrick James contends are its neglected philosophical bases. To this end he introduces into IR theory the concept of systemism as the essential key to a fresh reassessment of realism as a paradigm. “Systemism means a commitment to understanding a system in terms of a comprehensive set of functional relationships…. the approach stands in opposition to either a ‘black box’ or some form of reductionism, which result, respectively, from holism and individualism.” Its main point is “that a theory should be expressed in terms of a full range of linkages at the unit and system level.” Translated into IR theory, systemism requires the incorporation of all four structural linkages—M-M (macromacro), M-m (macro-micro), m-M (micro-macro), and m-m (micro-micro). Viewed in these novel terms, the two major strands of realism are found wanting: the individualism of neo-traditional (or neoclassical) realism is confined to the micro-micro (m-m) nexus, and the holism of structural realism to the macro-macro (M-M) linkage. As such, realism's full potential has never been realized. Only the unifying approach of systemism, incorporating the findings from all four connections at both system and unit levels of analysis, concludes James, would generate the “vast, untapped potential” of realism as a paradigm for international relations. The millennial reflections papers on realism are full of insights and provocative arguments: a spirited, uncompromising defense of its continuing relevance to world politics at the dawn of the new millennium and a blunt reassertion of its claim to paradigm primacy (Mearsheimer); a self-critical assessment by an advocate of realism that highlights its achievements yet acknowledges shortcomings that need to be addressed (Grieco); a candid, unqualified criticism of its alleged fundamental flaws in logic, explanatory power, and empirical accuracy, which render it passé as a guide to IR research in the future (Vasquez); a historically grounded critique leading to a judgment that the assumptions, postulates, empirical adequacy, and claims of realism no longer accord with world politics since the end of World War II, tempered by an acknowledgment of its significant contributions to

the IS field (Holsti); a critical view of the democratic peace thesis that turns it “on its head,” with the claim that democratic peace depends on the international Page 10 → security setting more than the reverse (Midlarsky); and a call for revision that attributes the shortcomings of the two principal strands of realism to their philosophical underpinnings—neoclassical realism (individualism) and structural realism (holism)—and advocates systemism as the key to developing realism's full potential as a paradigm for IR (James).

Institutionalism The second cluster of reflections on paradigms comprises four papers on institutional theory by David Lake, Robert Keohane, Joseph Nye, and Oran Young. Lake “Progress in International Relations: Beyond Paradigms in the Study of Institutions” presents David Lake's wideranging reflections on IR as a whole and on international institutions. These take the form of addressing three broad questions: “What is the nature of our enterprise and what should our standards be? Where has there been significant progress in the field … and why? Why hasn't there been more progress?” In his discussion of standards of inquiry, his “preferred approach is that of scientific realism … and the deductive method.” Lake focuses on two, though implying there are more, “significant areas of progress”: “our understanding of trade policy … employing what is sometimes called open economy politics (OEP)” and the “socalled democratic peace.” He attributes the success in both areas to commonalities in these research programs: “First, both are disciplined by a strong empirical foundation…. Second, there is an emphasis on deductive rigor…. Finally, both literatures have drawn upon and integrated their analyses into broader bodies of theory.” There are, however, “impediments to progress,” of which he notes three: “we debate assumptions, not implications,” for example, the now-dormant absolute-relative gains debate; “we work on separate islands of theory” and do not build bridges between them, as in choice theory; and “third, we fail to operationalize our variables, especially our dependent variables,” as with “cooperation.” In that context, “We need to devote more time and resources to developing operational definitions and collecting systematic data.” The record in the study of international institutions Lake terms a “mixed success.” On the one hand, he notes “enormous progress, especially over the last half century”—specifically, “strong theories that explain the formation and maintenance of international institutions, … detailed studies of how and where they matter … [and] a new generation of innovative midlevel theories designed to account for variations in legalization and institutional design.” The success, he suggests, was due to the concept of regime, a “progressive ‘problem shift’” that vastly extended the range of the subfield, and the use of “emerging work elsewhere in the social sciences on the sources and roles of institutions.” Page 11 → At the same time, Lake is candid about the failure of institutionalism, “after a flowering in the early 1980s, … to meet the high expectations of many of its proponents.” Specifically, it “lacks a ready operationalization of its key variable, whether regime or institution”; it “lacks an accepted empirical pattern or puzzle that motivates inquiry”; its practitioners “have not expanded their horizons far enough, and in so doing have fallen into … the endogeneity trap”; finally, they “have adopted an over-narrow conception of institutions,” ignoring many types evident in the past and today, a lacuna that creates selection bias. The way forward, for Lake, is to shift the focus from institutions to “the study of governance … defined as the design, construction, and maintenance of mechanisms to reach collective decisions and to enforce agreed-upon bargains,” a change that would reduce the likelihood of selection bias. He concludes with a blunt exposure of “the insidious side effects of current practices” and a powerful call for remedial action: insisting that “scholars ask interesting empirical questions”; ensuring that “‘normal science’ has a

prominent place in our journals”; encouraging debunkers of received wisdom and contributors to incremental understanding of IR; rewarding those who devote enormous time and energy “to define concepts systematically, code cases, and build data sets”; and, finally, downgrading the “paradigm warriors” by asking them, “What have you helped us understand lately?” Keohane It was Keohane and Nye who mounted the first major challenge to realism as the preeminent paradigm in IS. Here, in a succinct paper, “Institutional Theory in International Relations,” Robert Keohane presents a candid assessment of institutionalism's achievements and some challenges it must address. As so often in the past three decades, he pays tribute to realism's insights, “particularly the powerful understanding of international structure, and its constraining effects, developed by Kenneth Waltz in Theory of International Politics.” He also reaffirms that “institutional theory accepts three basic realist assumptions: (1) states are the primary actors in world politics; (2) they can be analyzed as if they were rational; and (3) they are not altruistic but, rather, are broadly ‘self-interested.’” However, “states … construct institutions in an attempt not only to reduce transaction costs but to improve the information environment, thus enabling them to overcome some of the dilemmas and defensive stances identified by realism.” Keohane credits institutional theory with illuminating how international regimes work. “We have done rather well at extending [this] theory to other issue areas (trade, environment, human rights). The field has been pretty good at recognizing new trends—‘complex interdependence,’ increases in nongovernmental organizations (NGOs) and their networks, transgovernmental relations, the information revolution, globalization”—along with developing new categories and taxonomies; “and we are alert to how institutions can change information and strategies.” Page 12 → At the same time, he is refreshingly self-critical about institutionalism's shortcomings. Stated simply, “our record is better at description, and even descriptive inference, than at explanation.” He notes four specific problems: “(1) Distributive issues have been recognized as important, but institutional theory has been slow to integrate them into its analysis…. (2) We have been slow to develop systematic databases of institutions, which could be used for sophisticated quantitative work…. (3) Methodologically, we face the usual problem of strategic interaction: rarely are there unique equilibrium solutions…. (4) In particular, there is a serious problem of endogeneity, embedded in institutional theory itself … which needs to be solved—I expect, through a combination of qualitative and quantitative approaches—before we will be able scientifically to establish the effects of international institutions on world politics.” He urges his colleagues to learn “from basic scientific discoveries in other fields that seem at first esoteric and irrelevant,” including game theory, simulation, mathematical explorations, and quantitative methods. He also reminds them not to denigrate interpretation and description in their commitment to science. Finally, while viewing arrogance and humility as necessary to intellectual progress, he says, “We are better at [the former], so it is more important to work at the latter!” Nye Like Keohane, Joseph Nye has consistently expressed respect for, and criticism of, realism: in the early 1960s, as a graduate student, “I was impressed by the power of the realist model and remain so,” he remarks in his millennial reflections paper, “Transnational Relations, Interdependence, and Globalization.” However, “I was equally impressed by the notion that realism captured only part of what was important in world politics. Ideas, institutions, and economic interdependence were left out.” In that context, he pays tribute to regional integration theory, developed by Ernst Haas, Karl Deutsch, and others in the 1950s and 1960s, as “a healthy antidote to the single-mindedness of the realist model, and it laid the basis for future work on transnational relations [and] interdependence…. [In sum,] the consistent theme of our [Keohane and Nye] argument has been to combine the two great theoretical traditions [realism and liberalism].” He also recognizes the contribution of constructivism:

“Although their approach lacks theoretical elegance, I welcome the corrective provided by constructivists.” Interestingly, neither Keohane nor Nye refers to any of the other competing paradigms in their millennial reflections papers—critical theory, the English School, postmodernism, feminism. The Westphalian statecentric model, according to Nye, has been undermined by three trends of the last half century: the information revolution, globalization, and marketization. However, despite the declining role of military force, the absence of hierarchy among issues, and the growth of multiple channels of communication, the three conditions of “complex interdependence,” “the nation-state is not about to be replaced as the primary Page 13 → instrument of domestic and global governance…. but it is not the only important actor.” This notion of more complex governance, similar to Rosenau's “Two Worlds of World Politics,” is sketched in a nine-cell matrix comprising three levels of activity—supranational, national, and subnational—and three spatial sectors—transnational corporations (TNCs), intergovernmental organizations (IGOs), and NGOs. Nye, like Rosenau, is sensitive to the negative effects of globalization: he cites, approvingly, Karl Polanyi's “powerful argument” that the failure of states in the nineteenth century to cope with these helped create the disruptive movements of the twentieth, notably communism and fascism. However, since world government is unlikely, there is need for an alternative, less ambitious, constraining governance structure. A candidate for this role is the Keohane-Nye concept of “networked democracies,” which is sketched in this paper, along with a call for “more modest normative principles and practices to enhance transparency and accountability” of IGOs, NGOs, and corporations, the triad of global governance today. Unlike Keohane's “think piece,” Nye does not, in his paper, acknowledge shortcomings and problems in institutional theory. Rather, he concludes by expressing pride in the continuing relevance of the “tools and concepts” they developed in the 1970s, in their Power and Interdependence (1977), for world politics at the beginning of the new millennium. Young A pioneer in research on international systems, crises, and mediation, Oran Young has also been a prominent contributor to the literature on international regimes and institutions. The title of his paper states his puzzle and thesis clearly: “Are Institutions Intervening Variables or Basic Causal Forces? Causal Clusters versus Causal Chains in International Society.” Young notes—but dissents from—the mainstream instrumental perspective on the role of institutions, of neorealists like Susan Strange and Mearsheimer, and even neoliberals like Keohane: “The basic causal forces are factors such as power, interests, and knowledge, and the role of institutions is to intervene between these forces and collective outcomes in such a way as to prevent the occurrence of joint losses or to secure the achievement of joint gains.” His alternative perspective argues that “institutions often emerge as basic causal forces in their own right …, forces that play important roles in determining the values or shaping the operation of [populations, affluence, and technology].” Much of Young's think piece is devoted to proving this proposition. Thus, “the size, rates of growth, and distribution of human populations are all influenced … by … the status of women…. where prevailing institutions accord women a full complement of rights, fertility rights are strikingly lower.” Similarly, affluence is a function of economic growth, and institutions Page 14 → play a key role in economic development. So too with technology, leading Young to conclude that institutions are not intervening but predictor or causal variables. More generally, he argues that each of the forces regarded as playing a causal role in collective outcomes, namely, “power, interests, and knowledge … is affected … by underlying institutional arrangements”; for example, “rules of the game frequently affect the bargaining power of those who participate in institutionalized activities at the

international level,” as with the veto power of the Five in the UN Security Council. To the mainstream rejoinder that causal forces such as power precede each institutional link in the chain, Young adopts an agnostic view: “neither side has a compelling case for the distinction it seeks to establish between basic causal forces and intervening variables.” To resolve the puzzle Young proposes the use of multivariate analyses of collective outcomes as “products of clusters of variables that operate more or less simultaneously,” with an attempt to determine their relative causal weight, that is, a form of regression analysis using the logic of statistical inference. He also advocates two other analytical procedures—“genetic explanations or tendency finding analysis” and “dynamic systems modeling.” In conclusion, he designates this as “a model in which clusters or combinations of interactive variables operate together to determine the content of collective outcomes.” The millennial reflections papers on institutional theory reveal a much greater consensus than do those on realism. All four authors identify themselves as “scientific realists” and have been major contributors to this second mainstream IR paradigm. All provide insights, but there are differences in emphasis. Lake provides an overview of the entire IR field (with special attention to institutions), a candid assessment of limited progress—at both levels of analysis—and why there has not been more, a blunt criticism of “paradigm warriors,” and a clear prescription for the “way forward,” notably a forceful call for change in attitudes of scholars; in particular, the need to ask clear empirical questions and recognize dissent. Keohane gives us a no less candid appraisal of progress and unfulfilled promise in institutional theory—strong on description, weaker on explanation, an explicit specification of problems that remain unsolved, and a call for more modesty and less arrogance among adherents to this paradigm, but without Lake's call to move beyond paradigm debates. Nye offers a proud, positive assessment of institutional theory's achievements (but lacking in self-criticism) and an assertion of the paradigm's continuing relevance to contemporary world politics; in fact, a claim to have anticipated many transformational changes, notably globalization. Young's offering is a provocative thesis that elevates the role of institutions in world politics and economics far beyond the conventional view—intervening variables—to that of basic causal forces, that is, independent variables, with a profound influence on power, interests, population, affluence, and technology. Page 15 →

Alternative and Critical Perspectives A diverse group of critical perspectives on international studies is represented in this segment: an overview (Steve Smith); critical theory (Robert W. Cox); radical theory (Michael Cox); constructivism (Ernst B. Haas and Peter M. Haas); postmodernism (Yosef Lapid, R. J. Walker); and system change (James N. Rosenau). Smith The rationale of Steve Smith's “Alternative and Critical Perspectives” is the need to defend the legitimacy and affirm the importance of “nonmainstream theoretical work within IR.” This takes the form of an unremitting critique of the U.S.–centered, social science-oriented mainstream of international relations (realism and liberalism). The pervasive tone of this millennial reflections paper is evident in Smith's lengthy response to the editors' request for a “self-critical, state-of-the-art reflection on accomplishments and failures” in one's own field: it takes the form of a blunt attack on “the incredible power of the critiques of a number of writers who claim that alternative approaches are not part of the social science enterprise” and engage in “the delegitimization of reflectivist approaches.” Those cited are Katzenstein, Keohane, and Krasner in the 1998 special issue of International Organization; Keohane's presidential address to the ISA in 1988; and Steve Walt. The attack continues: “This does not mean that there have been no [non-mainstream] successes, only that these successes have never changed the nature of the mainstream….Why is this? … [because] the mainstream has tended to reconsolidate itself so as to continue the U.S. social science project”; and, in that context, he accuses

Keohane and the others of co-opting the Wendt version of social constructivism. Smith notes that critical perspectives reject the “four main tenets of U.S. social science”—the fact/value distinction, empiricism as the optimal epistemology, naturalism as possible, and the goal of uncovering regularities in the social world. He acknowledges the failure to establish an alternative epistemology: this he attributes partly to disagreement among themselves; and “secondly, the grip on the discipline of positivism U.S.-style is so great as to close down any space for alternatives. This is the nub of the issue.” He also quotes, approvingly, Ole Waever's criticism in the 1998 IO special issue of “the internal intellectual structure of American IR … [its] disciplinary elite … [and the claim that] American IR alone generates an apex that therefore comes to serve as the global core of the discipline.” On the second theme of the Millennial Reflections Project—assess where we stand on key debates and why we have failed to resolve them—Smith rejects the two underlying assumptions: first, “that the key debates are debates open to all … [and] second, that they can be resolved…. I do not think that the mainstream is willing to assess ‘alternative’ perspectives on Page 16 → any terms other than its own … [and the notion of “resolved”] is based on a positivist view of inquiry.” Smith's fundamental rejection of mainstream IR thought is also apparent in his response to the third theme—that is, evaluate what intrasubfield standards we should use to evaluate the significance of theoretical insights. He responds: “The ‘alternative and critical’ approaches tend to resist any attempt to make the kinds of assessments about ‘progress’ or ‘success’…. The reason is obvious: the language of ‘assessment,’ ‘progress,’ ‘success’ smack of a modernist notion of knowledge … [and they are] not … willing to buy into this model of assessment.” As for “the broader question of progress,” “frankly, I do not see much progress in IR.” In a more subdued tone, he concludes by expressing the “hope that theoretical work, of whatever persuasion, should be informed by, and in turn inform, empirical material … [and that] the discipline will foster intellectual pluralism and open rather than close down analytical space.” Cox (Robert) Robert Cox introduced critical theory to international studies in the early 1980s and has been its most prolific and influential advocate since then. His paper, “Universality in International Studies: A Historicist Approach,” is much narrower in scope than Smith's overview of critical perspectives and is very different in tone: it focuses on “the issue of universality versus relativism (which includes the issue of positivism versus historicism).” The critical theory approach is framed in terms of three questions: “(1) What is critical theory and how has it come into international studies? (2) What is the core philosophical issue inherent to critical theory? (3) What ontology can the critical approach propose as a framework for inquiry and action?” For Cox, critical theory is synonymous with problem-solving theory. “The existing world is a given (datum)…. The analyst is concerned with dysfunctions, with homeostasis, and the restoration of equilibrium…. Critical theory takes a diachronic perspective, one of history informed by sociological theory…. The analyst, the historically oriented social scientist, goes through a double process of historical thinking … self-consciousness of one's own historical time and place … [and] the effort to understand the historical dynamics that brought about the conditions in which these questions arise.” The issue of universality relates to norms and knowledge. Cox rejects the European Enlightenment consensus that both were universal in time and space, in favor of the “dissident perspective” of Vico, Kant, and, later, Dilthey, Croce, Collingwood, Heidegger, and Gramsci, who viewed the study of history as basically different from physical science. Cox claims that the gap between positivist natural science and intersubjective social science has narrowed as a result of a growing recognition in physics that its laws are more relative than absolute. Page 17 →

Cox extends this line of argument to prehistorical biological evolution and concludes that humanity is “a product of the very longue durée [Braudel's term for early modern history] of biological history,” which will end through self-destruction or a long evolution, a theme for which he discerns support in Carl Jung's concept of the collective unconscious. He also notes his preference for Weber's notion of an ethic of responsibility rather than an ethic of ultimate ends. Cox concludes his reflections by framing the critical research program in terms of his view of the primary purpose of international studies: it is “to prepare the ground for a pluralist world of diversity in social organization and ethical norms”—not dissimilar from Smith's call for pluralism in IS. Thus his IS research program would include “the study of civilizations, civil society and its complex context including the covert world, the maintenance of the biosphere, and the problem of social cohesion for the maintenance of collective entities and their institutions.” As with Smith, this is a very different conception than mainstream IR, both in the realism and institutionalism versions. Cox (Michael) Michael Cox begins his wide-ranging survey of radical, mostly Marxist—traditional, neo–, Gramscian and other—literature on IR, “The Continuing Story of Another Death Foretold: Radical Theory and the New International Relations,” with a scathing criticism of IR: “by any measure [notably its proven ability to anticipate future global trends] it has to be judged to have been one of the more dismal of the social sciences.” Among the “major academic trends” in IR, the one that “has suffered the greatest damage … as a result of what happened in 1989 [is] that loosely defined, deeply divided and heterogeneous current … known rather vaguely as ‘radical.’” Cox, however, suggests that “just as with realism, it might still be a little premature to declare it dead altogether.” He defines his task as summarizing “a large, very uneven, and deeply schismatic literature”—radical IR—focusing on four themes: how radicals came to terms with the death of socialism in the Soviet bloc and the end of the cold war; their intellectual response to globalization; the development of a radical theory of crisis; and how they cope with American hegemony, as well as with a “new” world that displays “as much, if not more, suffering than at any point over the last fifty years.” Radical analysts, notes Cox, have finessed the issue of the fall of official socialism in several ways: denying that the Soviet Union and Eastern European regimes were genuinely socialist; claiming that socialism's demise was not inevitable but, rather, that it was due to Gorbachev's “ill-fated reforms”; or attributing it to external factors, notably the Reagan-era arms expenditure by the United States that Moscow could not match. Moreover, radicals tried to compensate for their problem of coming to terms with the fall of the Soviet Union by dismissing the “new” world order as nothing new Page 18 → at all. Some even argued that the end of communism created space for a social democratic “third way.” On the concept of “globalization,” Cox notes four “quite reasonable questions” posed by radicals, each with dissenting answers. First, “what exactly is so new about the idea?” The response is nothing— that is, Marx, Lenin, Trotsky, and the theorists of dependency had acknowledged economic interdependence; for Cox, “it was only a recycled version of a very old idea.” Second, is globalization “an accurate description of the world economy in the late twentieth century”? Here too, radicals dissent; they assert that the state and politicians can still make a difference. A third criticism of globalization is “its appalling human consequences…. [It] has led to the most extreme forms of inequality and economic polarization.” Finally, while acknowledging that major-power war is unlikely, radicals anticipate much more intense economic competition than do advocates of globalization. In the tradition of Hobson and classical Marxism, radical analysts are profoundly skeptical about the long-term stability of world capitalism. One part of their theory of crisis is a belief in a “fundamental contradiction between the world economy's capacity to produce and the people's ability to consume.” Further, they contend that world capitalism is now dominated by finance capital, which is disinterested in full employment. And, finally, they argue that the end of the cold war has reduced the need and rationale for persistent large-scale military spending that sustained capitalist economies for almost half a century.

Three aspects of the U.S. behemoth have attracted the attention of radical IR: the use of American power; U.S. post-cold war foreign policy; and American hegemony in the global system. Most radical writings criticized U.S. intervention in Iraq in 1991. However, its intervention in ex-Yugoslavia created a dilemma: visceral antagonism to any U.S. involvement in Europe, indeed anywhere, and the awareness that only U.S. power would end the genocide in Bosnia and, later, Kosovo. The U.S. policy of “promoting democracy,” by contrast, has evoked unanimous criticism as “a sham or … a device designed to obscure America's economic objectives in the larger capitalist system.” The third aspect—is the United States in decline?—has evoked varying radical views, with reluctant acceptance that “the twenty- first century will be American.” Finally, radicals accept that prospects for political renewal on the Left are dim, at best, confined to resistance to globalization, promoting democracy, and moving environmental issues to the center of the international political agenda. Haas and Haas The main thrust of Ernst Haas and Peter Haas's millennial reflections paper, a forthright dissent from the “scientific realism” espoused by leading figures in institutional theory such as Keohane and Nye, of whom Ernst Haas was an early mentor, is captured by its unusual title, “How We Learned to Escape Physics Envy and to Love Pluralism and Complexity.” Their objectives are “to frame a pragmatic-constructivist approach for the study of international Page 19 → institutions, and of IR more generally … [and] to sidestep the philosophical chasms now dividing the discipline”—ideographic versus nomothetic studies, explanation versus understanding, and positivism versus reflectivism, “procrustean constraints on inquiry.” Much of their paper focuses on their view of the fundamental difference between the natural and social sciences regarding truth, positivism, progress, knowledge, and cumulation. They appear to adopt a middle ground on truth: it “is neither as absolute as positivists and scientific realists demand nor as biased, subject to someone's domination, or hegemonic as relativists proclaim.” However, their “consensus theory of truth” posits that “a given solution [is] true … only temporarily and for a restricted purpose.” The Haas and Haas espousal of constructivism is also evident in their discussion of positivism: while acknowledging its “very impressive achievements” in the natural sciences, they term its procedures “appropriate in the social sciences [only] when and where their epistemological requirements do not clash with the social construction of reality…. More often [than not] … social science inquiry is outside the positivist domain.” Moreover, “In the social sciences nothing is ‘discovered.’ Our understanding of social reality is subject to the reflexivity of our minds.” Similarly, they reject the claim to universal validity of findings: “It is difficult to formulate universal claims over time and across cultures because of the mutable nature of institutions and the potential role of free will.” In a lengthy discussion of progress, they ask: “Is there something wrong with the very notion of progress as applied to international studies? … Yes and no…. A theoretical contribution may be considered progressive if it changes the way most scholars think about the problem or puzzle … at a level below that of grand theory.” On the foundations of their philosophical disposition, they declare, “Pragmatists, including ourselves, … differ from positivists because we eschew any crisp notion of causality.” Further, “the world is not immediately accessible to actors…. Willful actors, through intersubjective activities, construct the world. Institutions are thus causes and effects of this process.” Haas and Haas distinguish three types of knowledge: brute facts, the traditional domain of natural science; hybrid facts, “much more common in the study of international institutions”; and social facts, “which exist ultimately by definition and have no meaning independently of the social world in which they exist.” “Knowledge cumulation is a social activity conducted by communities of scholars rather than by individuals working in isolation…. Pragmatists hold that truth is provisional. They prefer to think in terms of complex relations among variables, not simply ‘A is explained by B’…. Progress can be had if we lower our ambitions. There will be no consensual grand theory for us.”

They conclude with a call for tolerance of a plurality of approaches—tolerance being “an inevitable by-product of the double hermeneutic in which we, as pragmatist-constructivists, are stuck”—and, with that, a denigration of knowledge cumulation: “any cumulating that may be achieved is temporally and spatially limited…. But a few puzzles get solved along the way.” Page 20 → Lapid “En Route to Knowledge: Is There a ‘Third Path’ (in the Third Debate)?” presents Yosef Lapid's reflections on achievements and failures in the domain of IS critical perspectives. This is done “through the related concepts of pluralism, dialogue, and reflexivity … [the] common denominators … in the change-oriented disciplinary transition known as the third debate.” His objective is “to highlight, in particular, the wisdom of exploring more fully the promise of possible, but currently still submerged, ‘third way/middle ground’ positions.” “My main argument,” he declares, “is based on the premise (and hope) that this polarized situation [between mainstream and critical IR perspectives] may be finally changing now.” His method is to offer an update on the three responses by early advocates of critical IR—despair, celebration, and reconstruction—noted in his seminal 1989 article on the “third debate,” and then to try to demonstrate how IR can benefit from via media, middleground solutions to the perennial problem of generating and employing knowledge. Although all three critical responses are visible a decade later, they have undergone change. Despair has diminished somewhat because “the recent spectacular rise of constructivism” has made the IR mainstream “far less suspicious or paranoid about third debate-minted ideas”; but Lapid hastens to add that the mood of despair has not vanished. Lapid himself does not share this mood: “Ten years into the third debate, IR's journey to the land of better knowledge continues.” The celebratory response, by contrast, “has been constantly losing altitude”: its two agenda items, the achievement of “a more open, diverse, and interdisciplinary theoretical enterprise” and “the hope for a more dialogically oriented and communicatively engaged scholarly enterprise,” remain unfulfilled, in the view of its adherents, though Lapid discerns “significant and, most likely, irreversible achievements” in the first item. The most vibrant of the critical responses at the dawn of the new millennium, in Lapid's view, is reconstruction, that is, a widespread “call … for ‘bridge building’ across the paradigmatic divides, ‘dialogue’ and ‘interaction’ among contending schools.” He, too, cautiously favors the “third way” as a means of “reconfiguring pluralism, dialogue, and reflexivity in the IR domain,” but he sees “little evidence of a plausible and imaginative architectural design … [for] such a critical disciplinary project.” Lapid's highest-priority goal is to foster pluralism, but of what kind? Here he follows Zerubavel's profile of three intellectual mind-sets and their corresponding pluralist type: the rigid, leading to “fortresslike” pluralism; the fuzzy, which generates “flabby” pluralism; and the flexible mind-set, which produces a via media pluralism—the clear preference of Lapid, who quotes Zerubavel approvingly: “an intellectual environment that would allow for both order and creativity, structure and openmindedess, focus and change.” Such an environment would also be conducive to dialogue, a means of reconciling differences in approach to knowledge, as well as generating understanding, Page 21 → empathy, and, possibly, collaboration between mainstream and critical perspectives. And pluralism and dialogue, together, can facilitate reflexivity in IR. Lapid concludes by reaffirming the central theme of his millennial reflections paper, the desirability and feasibility of energizing his third path in the third debate, so that “coalitions of the willing” (to use Ruggie's phrase) would engage in more “conversations” and fewer “great debates,” to the benefit of all the diverse theoretical, epistemological, and ontological strands in international studies. Walker (R. B. J.)

The R. B. J. Walker millennial reflections paper, “Alternative, Critical, Political,” centers on “four among many possible areas of convergence among the heterogeneous literatures that are currently identified, and disciplined, as alternative and critical.” However, he begins with the “primary argument” that these literatures “are interesting primarily as attempts to politicize practices that have been profoundly depoliticized.” His “secondary argument” is that they “lead to engagements with the constitutive practices of a specifically modern politics, especially with the practices of sovereign authorization,” a pivotal concept in his paper. In that context, he reiterates his oft-expressed view that “one way of understanding many of the problems of contemporary politics is to think of the difficulties of imagining an outside to a politics that is already constituted as a structure of insides and outsides.” And he reaffirms the view of all advocates of critical IR perspectives that there is no subfield or community that can be identified as alternative or critical. He dwells at some length on the Kantian distinction between the critical and the dogmatic, which he regards as “an unavoidable point of orientation … crucial, even if difficult.” Surprisingly, he claims that the questions addressed by Hobbes, Machiavelli, Rousseau, Kant, and others are closer to those posed by critical theorists than to those addressed by the IR mainstream. And, in a gesture of tolerance and all-inclusiveness, he acknowledges that noncritical perspectives, including those from the mainstream, along with literary, cultural, and ecological theory, “also have much to say.” Turning to the main focus of his paper, Walker specifies “a shared concern [among alternative and critical perspectives] with four core themes. First, the tendency to interrogate the received assumptions about what the world is taken to be; … Second, the tendency to engage with sovereignty as an enormously complicated site of political practice, … Third, the tendency to get very annoyed when told that difficult questions about ontology and sovereignty ought to be adjudicated on the ground of epistemology and method. And fourth, the insistence that there are interesting parallels … between … [notions of IR] as a discipline … and … what it means to have alternative or critical perspectives on that discipline.” Walker devotes the rest of his paper to an elaboration of these themes. Although he acknowledges that “each of these four themes can be made to Page 22 → seem highly arcane, trivial, philosophical, and so on … [they] seem to get at political practices that are of considerable practical importance.” He challenges the priority given to epistemology and method among mainstream theorists on two grounds: that epistemology depends on ontology and that the mainstream priorities “served to mask a sustained silence about the ontological categories and assumptions that are thereby reified.” He adds, to clarify his view, “questions about how we know what we claim to know are dependent on what we assume we are trying to know.” Yet “epistemology remains crucially important.” But on social science: “Critique trumps social science, not the other way around. Or rather, there is no social or any other kind of science without critique.” He is also critical of social constructivism, “a term that makes little sense to me unless taken as some vintage opposition to a claim to nature, to essence, or a methodological injunction to reify.” And he concludes: “it is this fourth set of convergences that seems to me to be the most significant…. the problem of authorization (the Hobbesian problem of sovereignty …) [is crucial; but IR theory, like political theory, has refused] to take the problem of sovereignty seriously”; “there is certainly a lot of work to be done.” Rosenau In his reflections on the state of international relations, James Rosenau expresses “mixed feelings”: he has “no doubt” there has been “enormous growth and progress” in concepts and methodology since the 1950s; “on the other hand, … [he has] doubts about the field's capacity for adapting to the transformations at work in the world today.” The central thesis of “The Globalization of Globalization” is that “the preoccupation with the dynamics of globalization, both good and bad, has undergone globalization,” although in his view, political science and IR lag far behind other social sciences in this regard. His “out-of-the-mainstream perspectives” are articulated in terms of four major dimensions of globalization. The first is change—“a major conceptual challenge … how do we know change when we see it? How do we

differentiate between evolutionary and breakpoint change?” He finds our conceptual ability to assess the durability of change or its early indicators rudimentary, at best. The second dimension is “fragmegration,” a concept Rosenau introduced in 1983 that merges forces for centralization and integration with decentralization and fragmentation. For Rosenau, the contemporary era is one of fragmegration, not simply globalization. There are many sources of this hybrid process—the “skill revolution”; authority crises; bifurcation of global structures; the organizational explosion; mobility upheaval; microelectronic technologies; the weakening of territoriality, states, and sovereignty; and the globalization of economies; all of which Rosenau discusses at four levels of aggregation—micro, macro, macro-macro, micro-macro, a concept similar to, but not quite the same as, systemism as elucidated by James in his paper on realism. The third dimension is micro-macro links—Rosenau deplores the lack of a viable theory in this domain Page 23 → and the indifference of IR scholars to the puzzles raised by these links. The last dimension is methodology—he deems conventional linear methods inadequate and urges reliance on nonlinear procedures made possible by advances in computer technology, although these are not specified except for a reference to computer simulation. His hope for future advances in the IR field rests with the training of students in these hitherto ignored or underused methodologies now available to the IR community.

Feminist and Gender Perspectives Ling In “The Fish and the Turtle: Multiple Worlds as Method,” L. H. M. Ling uses a Vietnamese folktale to assert “the central problematique facing our discipline … today: multiple worlds and how to know them.” After many traditional-type questions from the fish, to which the turtle replies simply, “No,” causing the fish to swim away in anger, the turtle sighs, “How can one know something new when one's questions are based on the prejudices of the old ?” Ling's bête noire is realism: “In the case of IR, the fish is more like a whale … [who] rules the seas…. For too long, the whale of IR … has subjected our discipline to an illusory universalism propagated through notions of power and the state…. Feminists and other critical theorists have led the charge against such realist one-worldism … [which] is but one representation of the world…. Feminists have pointed out that what … [realism] has passed off as ‘universal’ or ‘human’ … really are the experiences, practices, norms, and institutions of a particularly small segment of the world's population: propertied, white males.” For Ling, the challenge facing IR feminists is, therefore, “further explorations into other worlds and their ways of world making.” Those worlds, other than the dominant one of “men, masculinity, and patriarchy,” are “a second, hidden one of women, femininity, and matriarchy … [and] a ‘queer,’ third world of blurred pairings between men /women, masculinity/femininity, patriarchy/matriarchy.” In that context, she criticizes liberal, white, Western feminist IR: by proclaiming “global sisterhood,” it “presumes that gender supersedes race, class, and culture, thereby allocating one categorical patriarchy to rule over all. Doing so effectively upholds a de facto hegemonic identity that is white, heterosexual, and North American/western European…. This pluralist logic … returns Third World women to a center/periphery dichotomy … [and] hegemonic self-involvement [by liberal feminists] prohibits learning about who and what we are.” How, then, to know the multiple worlds? asks Ling. She responds in terms of her initial folktale: “we need the fish and the turtle [realism and feminism], first and foremost, to continue their mutual inquiry [begun through an exchange of articles between Robert Keohane and Ann Tickner Page 24 → in 1997 and 1998]. But the turtle needs to speak up … to show the fish/whale how to ‘see’ the other, unfamiliar world through its relationship to the familiar one…. The fish/whale, in turn, should listen, patiently and openly, while developing a method or hermeneutic of asking questions…. [It] has a vested interest in forgoing age-old hierarchies of anarchy/order, center/periphery, Self/Other, or masculinity/femininity…. [It] must learn, finally, to forsake the illusion of power and control in exchange for substantive growth and learning.”

Peterson The two interrelated themes of V. Spike Peterson's “On the Cut(ting) Edge” are the strengths and “failures” of feminist IR scholarship—“what renders it ‘cutting edge’ … [and] what renders it ‘cut’—in relation to mainstream IR.” Like all feminist contributors to the Millennial Reflections Project, she emphasizes the “wide spectrum of analytical and political positions [in feminism] … a continuum that spans positivist and ‘alternative’ (postpositivist, poststructuralist) orientations”; further, she writes, “gender is not simply an empirical category … but an analytical category, pervasively shaping how we conceptualize, think, and ‘know.’” The first of Peterson's two main theses, namely, that feminism is at the cutting edge analytically and politically is clearly articulated: “the study of gender requires and produces transdisciplinary orientations…. [these] stretch or transgress familiar and often constraining boundaries”; ergo feminism is more open than mainstream IR to multiple methods and interdisciplinary debates. The result is that it “generates more wide-ranging and encompassing analyses of social relations.” She concludes a lengthy discussion of why feminisms are cutting edge, analytically, as follows: “in part because they are transdisciplinary, translevel, and multidimensional, in part because they pioneered studies of identity, and in part because they address complexity through innovative and relational analytics.” And, in this context, she notes her “alternative analytical framing, or ‘triad analytics’ … [which] posits identities, social practices/institutions, and meaning systems as three co-constituting dimensions of social reality,” a conception that can be related to some mainstream IR concerns. Feminism does not confine its attention to critical scholarship, writes Peterson. Rather, it has normative and political commitments, for several reasons: first, feminists are committed to improving the conditions of women everywhere and in every respect; second, feminists must be political to survive, let alone prosper, in the academy; third, their diversity in outlook compels critical reflection on the meaning of feminism; and fourth, cutting edge theory and practice are inextricably linked, a linkage that “also informs and enacts emancipatory political practice, ” which Peterson regards as “progress.” Peterson's second thesis derives inescapably from that progress: “what renders feminisms analytically cutting edge effectively cuts them out of the Page 25 → picture framed by dominant starting points in IR…. [They are] cut down as suspicious and inappropriate by a mainstream that claims its knowledge is apolitical, its privilege is either irrelevant or unproblematic, and its desire for change is limited.” Nevertheless, in her conclusion on “what is to be done,” she favors continuing to ask feminist questions in IR for two reasons: it is important to build on the differences/progress feminism has already made in IR, however limited they may be; and feminism has resources that can contribute much to enhancing IR's “analytical contributions to our understanding of the world(s) we live in … [for example, its] glaring failure to take racism (and other structural oppressions) seriously.” Pettman “Bringing It All Back Home?” begins with some reflections, or “claims,” by Jan Jindy Pettman on feminism in international studies: first, it was effectively excluded until the late 1980s—on the assumption that IS was “gender neutral … or that [IS] was … men's business”; second, it is still on the periphery, measured in terms of courses, articles, and textbook references, but it has “a remarkably robust, diverse, and productive presence in and around IS”; third, the “visibility of women and feminist-informed critique in the ‘real world’” is greater than in academe and the discipline of IS; and fourth, while women are frequently noticed in the Western media, they tend to be portrayed as “victims or as symbols of the costs of world politics” or as survivor-witnesses, as in the case of Japanese military sexual slavery during World War II. There is no single feminist view on IS, notes Pettman, as do all the other feminist contributors; but she discerns a wide-ranging impact on the field. Specifically, feminist scholars “(1) radically expand the kinds of topics, issues, and content in IS … [that] unsettle the discipline. They suggest new ways of thinking about core issues, including power, security, and the state, and introduce new ones”; “(2) ask different kinds of questions”; “(3) reveal the masculinist assumptions, investments, and strategies of mainstream IS”; “(4) in the process, widen definitions of

the political in the international too”; “(5) disrupt or transgress disciplinary boundaries”; “(6) internationalize the account—raising questions about the identity of the discipline, too”; “(7) gender the account…. the constitutive role of gender in world politics is now the subject of keen interrogation”; “(8) sex the account … identifying the roles that sex and sexuality, and heterosexism, play in international politics”; and “(9) [provide] the possibility of an embodied IS…. feminist writings … identify particular kinds of bodies in different international engagements and the privilege or penalty that can flow from being identified as male or female, inside or outside, exotic or ‘normal.’” Pettman cites two illustrations of feminist approaches to IS. Referring to war, feminism provides “more, and more useful, ways of interrogating the kinds of relations that make organized large-scale violence possible … [and] whether certain kinds of masculinity are more violent than others.” So too does a feminist lens shed light on the shifting sands of Asian “triumphalism” Page 26 → and its 1997 crisis. The former “claimed a newly dynamic, successful Asia had recovered from its colonial emasculation … [while] ‘the West’ was seen to be faltering, losing [its] manly qualities … [and] becoming decadent, immoral, obsessed with consumption and pleasure—in short, feminized.” And the 1997 Asian crisis “ushered in new configurations in gender representations of state power, national identity, and international relations…. Both Asian miracle and crisis were culturalized.” She concludes on an ambivalent note: “feminism is now well into critical mass in IS. Clearly, however, feminists are (not yet?) at home in IS…. Like the real world, it is messy, jumpy, sometimes cumulative, sometimes contradictory or doubling back.” As for prescription, “it is not possible … to propose a future research agenda.” Her own goal is “to internationalize the agenda, to generate more feminist accounts from outside the usual centers of power.” Sylvester Most of Christine Sylvester's millennial reflections paper, “‘Progress’ as Feminist International Relations,” addresses one of the project's six theme questions. Her introductory remarks set the tone: “‘Progress’ is an odd thing. It twists my head around…. Much of IR tries to reach ‘it’ cumulatively, using the right methods harnessed to the right research programs…. Indeed, progress is a cre(a)p and a leap, a jump, and a crawl. It strives for and against, thrusts, moves, makes a crumble, flows. It is a fact, a sorrow, a make-believe, a virtuality, a compromise, a cynicism. Progress advances. It stands us still. It writhes. It invites nostalgia. It is at once a common notion, easily grasped by the modern mind, and something difficult to understand, make happen, or repudiate. Progressive deterioration even bespeaks of progress…. The term eludes definitional clarity.” She hails two feminist scholars, Jean Bethke Elshtain and Cynthia Enloe, as “innovators inclined toward framing questions of international relations around feminist concerns rather than around IR agendas,” in Women and War and Bananas, Bases, and Beaches: Making Feminist Sense of International Relations, respectively. But Sylvester notes that not all feminists adhere to that path, citing Ann Tickner [another contributor to the Millennial Reflections Project] among those who prefer to bring to “a recognizable IR topic … new gender-highlighting questions, which lately telescope feminist angles on international political economy, foreign policy, war, peace, globalization, the state, nationalism, trade, and so forth.” In this context, Sylvester lays bare the fundamental contrast in the approaches of feminism and traditional IR: “Feminists begin with gender, bodies, sexuality, difference, voice, patriarchy, subjectivity, representation, and/ or travel theory…. conventional IR typically begins with states, war, trade, regimes, anarchy, and/or rationality. It maps relations of force, power, decision, threat, and agreement.” The conventional subject matter of IR is inadequate in Sylvester's view: “I recommend forays into novels, choreography, drama, poetry, music, and Page 27 → visual arts. These places of fiction, expression, and emotion assist those of us trained in IR to escape narrow vision in order to see other international actors, interactions, products, exchanges, and dramatizations.” She is also skeptical of the search for patterns, as in Enloe's work on militarism. And she wonders whether building feminist IR theory is archaic or perhaps premature. She also displays a touch of criticism in her distinctive phrasing: “Feminist IR, impressive as it is, has not capitalized

sufficiently on its progressive tendencies by theorizing the connective tissues that bind the genre.” At the same time, she is emphatic in asserting once more the difference between (her view of) feminism and IR: “Theory need not be law-like. It need not be problem solving.” Tickner In “Feminist Theory and Gender Studies: Reflections for the Millennium,” J. Ann Tickner presents a survey of part of the burgeoning literature on this long-neglected segment of international studies, along with an assessment of its contribution to the field. The dominant theme of her reflections—as of all feminist papers in the Millennial Reflections Project—is the “great divide” between feminist and mainstream thought in IR: the questions they pose, their methods, their conception of theory, cumulation, and the primary purpose of inquiry. Referring to the six theme questions that millennial reflections contributors were asked to address, Tickner speaks for all feminists with her skeptical rejection: “these questions are not ones that many feminists would ask when engaging in self-evaluation”; by contrast, “the choice of questions we ask is never neutral.” Moreover, such terms as state of the art and cumulation “are in tension with an approach that ‘celebrates diversity’ in terms of its subject matters, normative focus, and methodologies.” Tickner notes a variety of IR feminist approaches—“liberal, radical, psychoanalytic, socialist, postmodern, and postcolonial”—but only the first employs mainstream empiricist methodology. Most feminist scholars “do not use social scientific methodologies, preferring instead methodologies more compatible with hermeneutic, interpretive epistemological traditions”; in fact, “their methodologies tend to be eclectic.” With regard to theory, she notes Zalewski's three types as a means of conveying the “great divide”: “a tool for understanding the world; theory as critique, or understanding how the world got to be as it is so that it can be changed; and theory as practice in … everyday life.” The first of these conceptions of theory is positivist and mainstream, unacceptable to feminism in IR. The second and third are more congenial. The reason is closely tied to their basic difference on the purpose of IR research: “For many feminist theorists, knowledge construction is explicitly linked to emancipatory political practice, primarily that of transforming unequal power relationships between women and men.” This in turn explains why “it is unlikely that feminists would use ‘scientific’ criteria, such as integration and cumulation, to evaluate them.” Page 28 → A consequence of this multifaceted “great divide,” reinforced by the definition of gender “as a social construction of inequality,” is to make difficult an assessment of feminist IR scholarship in terms of “explanatory power.” Nonetheless, Tickner does reflect on its achievements and limitations. In general, “the most significant accomplishments have been the enormous productivity of a relatively small group of scholars and their success in creating a supportive space within the ISA.” Specifically, she evaluates feminist progress through the McIntosh lens of five stages, in ascending order: “a womanless world”; the absence of women, except for the “famous few” (e.g., Thatcher); women as victims; “women as valid human beings whose various life experiences have shaped the world in which we live”; and at the apex, stage five, the inclusion in IR subject matter of “the experiences of all individuals regardless of race, culture, class, and gender.” IR scholarship, in Tickner's view, remains at stages one and two, while feminist IR has moved on to three and four. In the rest of her paper Tickner surveys briefly several feminist writings other than Zalewski's three types of theory and McIntosh's five stages of progress: Moon on military prostitution; Chin on the important supportive economic role of domestic workers in Malaysia, using the method of narrativity; Mansbridge on feminist research questions; and Enloe on the relevance of women's roles to international politics. At the same time, she acknowledges the legitimacy of feminist research conducted from a social scientific perspective and urges an engagement “across these theoretical divides.”

Zalewski Marysia Zalewski's reflections are presented in the form of two “stories”: “Feminism and/in International Relations: An Exhausted Conversation? or Feminists Doing International Relations: The Cut(ting) Edge of Contemporary Critical Theory and Practice?” She begins by explaining why the Millennial Reflections Project's guideline questions are “problematic”: first, “a critical approach such as feminism will have an alternative sense of what counts as ‘reasonable’ in the realm of the intellectual”; second, feminist work is diverse, even contradictory, making generalizations very difficult; and third, conventional-type IR questions could make “vulnerable” any critical perspective that challenges the status quo. The “conversation” between feminism and IR is “exhausted,” in Zalewski's view, because the latter resists “a radical restructuring of many of IR's epistemological, ontological, and political beliefs,” that is, separating them from power. Stated differently, feminism and conventional IR are “mismatched.” Her second story is told in the form of a response to two questions that Zalewski poses: what did IR feminists expect to happen, and what does feminist theory contribute to IR? In the early days of “euphoric optimism” feminists expected “a more equitable inclusion of women” in the real world of international politics and in the academy, as well as legitimacy in the discipline. Page 29 → The first was partly achieved, but few women have risen to a high level in the hierarchies. At the same time, Zalewski notes several achievements. One is in the realm of teaching—students are keenly interested in feminism. Moreover, feminist studies are “cutting edge” because they “alter the very subject of knowledge by calling into question what is considered knowledge in any discipline.” And by “opening the doors for other deconstructive theories and practices such as critical theory,” feminism upsets the conventional fabric of IR. She is also proud of the new International Feminist Journal of Politics and the publication of “backpack” books on feminism, reaching a larger audience than university students. As for what feminist theory does and its impact, its role is, in large part, “blurring disciplinary boundaries,” which leads to “incremental resistance” by those who dominate the discipline. The dialogue between feminism and IR, Zalewski acknowledges, does not work well; and feminist work is still underrepresented in journals and textbooks. Moreover, feminist scholarship has not yet attained legitimacy, using standards of conventional IR. Nor has an “intrasubfield consensus” been reached by the feminist community—but she does not lament the failure to do so, for diversity and differences are of the essence of feminism. In short, the record of “progress” has been mixed. As for the future, Zalewski calls for a new “conversation,” centering on the feminist IR community, rather than as in the past an attempt at dialogue between two separate communities, feminism and IR.

Methodology The reflections on methodology comprise nine papers on formal modeling (Michael Nicholson, Harvey Starr, Bruce Bueno de Mesquita, Steven Brams), quantitative methods (Dina Zinnes, James Lee Ray, Russell Leng), and qualitative methods (Jack Levy, Zeev Maoz). Nicholson The purpose of Michael Nicholson's overview of “Formal Methods in International Relations” is “to discuss the principles behind formalization … [and] to justify its extensive use mainly by conceptual argument and by referring very briefly to some of the achievements … [which] are genuine,” as well as to spell out future directions. The tone is modest throughout: Nicholson does not claim that all aspects of IR, only “some very important problems in the discipline,” must be formalized to achieve progress. The core of his argument is stated concisely: natural languages are good at description but “awkward and clumsy” for deductive argument; formal languages, notably mathematics, are the reverse. Moreover, theories involve deductive arguments, ergo, they need mathematical methods. In that context, he defines formal modeling as “the use of mathematical, diagrammatic, and symbolic methods in the study of problems in international relations.”

Page 30 → He adds that because the ultimate purpose of theory is to describe the real world, formal and statistical methods are closely related. Nicholson surveys some notable applications of game theory to IR in the area of decisions and choice. He cites the pioneering contributions of Schelling, Boulding, and Rapoport, followed by a relative drought until the hostility models of Zinnes and her associates and Brams's theory of moves. He pays special tribute to the innovative role of Richardson in both formal modeling—notably models of arms races and hostility—and statistical methods from 1919 until the early 1950s. His rationale for using mathematical methods in IR is that they have proved very successful in other disciplines—notably physics, but also in economics—within the social sciences. And these methods became central in social scientific inquiry with Hempel's covering law of explanation. He rejects the counterargument that IR is different and is better served by natural languages, claiming that stating a problem formally may uncover important findings that would otherwise be concealed. As for the future of formal methods, Nicholson remains refreshingly modest: “All we can do is suggest some plausible lines of progress that may be shown in the future to have been good or bad guesses…. As theory advances, it is more likely to become more deductively complex than less…. simulation methods will be used more frequently, and … these will generate a whole set of quasi-experimental worlds.” Because this is so, he urges formal modelers to make a special effort to communicate their findings effectively to readers who lack the training or interest in these methods. He also reiterates the limited use of formal analysis. And he concludes with an important cautionary note to his colleagues: “We must, at all cost, avoid becoming a mathematical priesthood purporting to dispense knowledge from behind a secure barrier of incomprehensibility.” Starr There are two recurring themes in Harvey Starr's reflections on “Cumulation, Synthesis, and Research Design for the Post-Fourth Wave.” One is the importance of the “research triad”—theory, logic, and method—and the centrality of theory: it “must be at the forefront of our endeavors, whatever the subfield or … the methodology employed.” The second is his emphasis on “an agent-structure theoretical framework…. I continue to see that approach as the most useful way to synthesize across levels of analysis, disciplines, and different forms of causal relationships.” Starr also reaffirms the importance and relevance of his “opportunity-willingness” concept, based on the work of Harold and Margaret Sprout: he terms it “simply one version of a more general agent-structure approach.” That approach and concept leads him to renew the view that “the current treatment of the agent-structure problem in the international relations literature is problematic … [and to criticize] the confusion among ontological, epistemological, and methodological issues and analyses.” He also cites Page 31 → their “methodological individualist model,” which is congenial to models of choice, such as expected utility and prospect theory. In this context, he discusses three challenges. One is “due to the lack of variance within the opportunity structure,” that is, insufficient cases. To cope with this problem he urges the wider use of simulations because they “permit the control and experimentation necessary to look for causal relationships … [, to assess] a model for its internal consistency, and to examine the plausibility of its assumptions.” The second challenge is “how [to] construct research designs to deal with the logical differences between models of necessity and sufficiency.” The controversy over deterrence illustrates for Starr the problem in studying necessity: there is nothing to observe if deterrence is successful, but failed deterrence can be observed. Thus critics focus on deterrence failures and conclude they are solving the not-y problem; they also conclude that rational deterrence theory does not predict to successful deterrence—but their logic is flawed. The third challenge is how to investigate “substitutability,” which is closely related to opportunity and willingness: it “indicates that any single cause may have multiple effects, and that any single effect may have

multiple causes. Different problems may lead to similar responses … [and] a given problem may be dealt with in multiple ways…. Substitutability … thus lies at the heart of … agent-structure frameworks.” Starr concludes that “cumulation and ‘progress’ [in IR] will depend on the quality and rigor of our theories and our methods.” The latter must facilitate our entry into “the continuous feedback loops between agent and structure … [, enable us to] design studies of necessity … [and of] phenomena that do not occur … [and of] substitutability…. Formal models … will most certainly constitute the main analytical approach of the post-fourth wave [following Deutsch's four waves—international law, then diplomatic history and international organization, behavioral sciences, and analytic and quantitative research], as a natural means to synthesize agency and structure, and choice (willingness) given constraints (opportunity).” Bueno de Mesquita Bruce Bueno de Mesquita's “Accomplishments and Limitations of a Game- Theoretic Approach to International Relations” provides “one person's views … focusing … on broad patterns and generalizations.” The bulk of his paper is devoted to five contributions of game theory: “(1) the provision of a motivational foundation for action; (2) an explanation of strategic behavior; (3) an ability to integrate other theories and approaches; (4) cumulative knowledge; and (5) insights into methodological issues influencing analyses of decision making.” The merit of illuminating a motive for action is “that the link between correlation and causation is readily discerned.” As for “problems involving strategic choice—and this surely includes most interesting problems in international relations—game theory is significantly advantaged over other approaches.” In particular, “psychological theories lack an equivalent to Page 32 → game theory's notion of equilibrium … and are unable at this time to achieve the subtlety of game theoretic accounts when choices involve strategy or the transmission of private information.” At the same time, game theoretic models can incorporate “the concerns and assumptions of structural, behavioral, constructivist, and psychological theories … while retaining the logical rigor imposed by [game theory's] axiomatic foundations.” Bueno de Mesquita acknowledges that no methodology has a monopoly over cumulating knowledge about IR. At the same time, he claims an advantage for game theory: “explicit assumptions make it easy to determine what the logical consequences are of relaxing one or another condition, substituting alternative assumptions, or … adding new assumptions…. In this way, anomalous cases can be carefully explored, modified theories constructed … while also uncovering new facts.” The fifth contribution relates to methodological insights: “endogeneity, selection effects, independence between argument and evidence, and prediction represent areas where rational choice models are especially helpful in clarifying empirical problems that frequently arise in other modes of analysis.” In substantive terms, Bueno de Mesquita notes a game theory contribution to understanding many important topics in IR: alliance formation, reliability, and termination; the democratic peace; deterrence; economic sanctions; and paths to cooperation. He acknowledges that “game theory endures significant limitations”: some may be overcome with improved knowledge; others are “inherent to the approach.” However, his discussion of shortcomings is limited to two aspects—information and multiple equilibria. As for the first, “games assume that everyone knows what the prior beliefs held by others are, whether they agree with them or not”; but this information assumption is a constraint that may undermine the analysis of expected behavior. The multiple equilibria problem is more serious: “game theory can predict a probability distribution over actions but cannot say definitively when one action or another will be taken provided each such action is part of an equilibrium strategy.” On the whole, however, Bueno de Mesquita's assessment is markedly positive about the contribution of this methodology to IR. Brams “Game Theory in Practice: Problems and Prospects in Applying It to International Relations” comprises three

sections: a brief discussion of shortcomings in classical game theory; a presentation of Steven Brams's alternative approach, the theory of moves (TOM); and its application to President Sadat's peace initiative to Israel in 1977. The first of four “major pitfalls … [is] misspecifying the rules.” Using as a “reality check” the idea that “any game-theoretic models … should propose rules of play that reflect how players think and act in the strategic situation being modeled,” Brams asserts: “Players do not usually … choose Page 33 → strategies simultaneously, or independently of each other, as assumed in the normal or strategic form of a game that can be represented by a payoff matrix; [or] adhere to a specified sequence of choices, as assumed in the extensive form of a game that can be represented by a game tree.” The second problem is “confusing goals with rational choice…. Rationality does not concern the ends themselves, which are neither rational nor irrational.” The third is “arbitrarily reducing the multiplicity of equilibria … [for] many games have multiple Nash equilibria … [and] a game-theoretic model may not be helpful in narrowing down a plethora of stable outcomes to a few precise predictions.” And finally, he criticizes “forsaking backward induction,” noting that “real players do not painstakingly work backward from the endpoints of a game tree…. Rather, players develop heuristics and use rules of thumb that simplify this process.” The merits of his corrective approach, which is briefly summarized, are clearly stated by Brams: “In my view, the rules of TOM, which allow players to move and countermove within a payoff matrix, capture the changing strategic nature of situations that evolve over time. TOM's ability, as well, to incorporate differences in power—moving, order, or threat—reflects an asymmetry of player capabilities in many IR games.” In his application of TOM to Sadat's 1977 peace initiative, Brams argues that what led Sadat to so act was not merely the “psychological barrier” between Arabs and Israelis, which the Egyptian leader emphasized in his memoirs, but also a rational calculus of choice for both adversaries: to initiate or not to initiate, for Sadat, to cooperate or not to cooperate, for Begin. He spells out the four possible outcomes: (1) Israel cooperates by offering concessions before Egypt commits to a permanent end to the conflict (the worst payoff for Israel and the best for Egypt); (2) no peace initiative is offered, which means continued Israeli occupation of Sinai (the next worst payoff for Egypt and the next best for Israel because it gets continued security but at a high price); (3) Israel doesn't cooperate with Sadat's initiative (the worst outcome for Egypt and not very satisfactory for Israel because of the likely loss of U.S. support and the likelihood of another war); and (4) Israel responds positively to Sadat's initiative (the best outcome for Israel, that is, the likelihood of permanent peace with the most powerful Arab state, and the next best for Egypt, that is, peace and the end of the drain on Egypt's resources, but alienation of the Arab world). As became evident soon after the peace treaty was signed in 1979 and since then, Sadat's initiative and Begin's response were rational and beneficial acts by the two adversaries. Brams concludes: “What TOM offers that classical game theory does not is a simple and dynamic rationale for the players' making their moves and countermoves…. In specific games, it provides detailed prescriptions for optimal play … that could aid foreign policy decision makers, especially in crises.” But, Brams writes, “TOM is no panacea.” It works best in two-person games, but many IR situations, such as alliance behavior, Page 34 → require modeling as n-person games, that is, a more general dynamic theory. Zinnes In “Reflections on Quantitative International Politics,” Dina Zinnes is skeptical about the relevance of the millennial reflections theme questions to this methodological strand of IR. Instead, she presents “more of a personal statement about where we have been, where we seem to be going, and what we need to consider in the decades ahead.” To answer her first two questions she initiated a survey of six academic journals from 1990 to 1999 to discover whether there were differences in the use of statistics and mathematical modeling. The findings, presented in table 1, are instructive: first, except for the Journal of Peace Research, especially from 1990 to 1994, most of the published work in IR uses statistics, mathematical modeling, or both; second,

quantitative research increased or, in the International Studies Quarterly, remained unchanged; third, quantitative work took the form mostly of statistical analysis; fourth, game theory was the dominant form of mathematical modeling; and finally, few quantitative articles contained both strands. As for the future, quantitative international politics seems likely to move increasingly to statistical analysis, and modeling is expected to remain primarily game theoretic. The first anticipated trend Zinnes finds “disturbing” because she questions the adequacy of generating theory through a statistical analysis of data: “We have reduced what may have been a rich interpretation of the international world to an often uninteresting if-then statement…. we are replacing possible explanation with far less interesting statements about what is associated with what…. in actual fact it might reasonably be argued that we have become less theoretical.” This view leads her to make a strong plea for greater use of mathematical models in the years ahead. And, to that end, she maps out a precise three-step program: first, “move statistics courses out of the limelight of graduate programs” and get incoming students to think in terms of puzzles “and teach the student how to ‘tell stories’ that fit these anomalies into what we know about the way the world works”; second, teach students how to translate stories into a mathematical language; and only then turn to statistics to test the story (theory). Similarly, Zinnes questions the overwhelming reliance on game theory: while it “is an excellent modeling language for capturing conflictual decision-theoretic situations … not all stories involve rational actors.” Moreover, even if her view of theorizing, that is, “telling stories about puzzles,” and translating the stories into a mathematical language were achieved, this would not be sufficient. To close the circle, the mathematical model must be tested—that is, “story and data must meet.” Thus, even after decades in the forefront of quantitative international politics, Zinnes concludes, “we have a long way to go in making our ideas, our theories explicit and an even longer way before we understand how to test our models.” Page 35 → Ray James Lee Ray's “Reflections on Millennia, Old and New: The Evolution and Role of Quantitative Approaches to the Study of International Politics” focuses on two questions: “Where have we been?” and “Where are we (and where should we be) going?” His survey of the past draws on evidence from Rosenau's two important editions of International Politics and Foreign Policy, published in 1961 and 1969, to make several pertinent observations: first, the 1950s were dominated by the idealist-realist debate, while the 1960s shifted attention to the “behavioralist revolution”; second, this change has deeper roots—it is evident in IR works during the 1950s by Richard Snyder, North and his associates, Deutsch, Benson, and Rapoport advocating, respectively, game theory, content analysis, quantitative studies, computer simulations, and formal modeling; third, the theme of continuity is emphasized, that is, “the more things change, the more they remain the same,” notably the persistence of idealism versus realism, in a different garb, and the enduring debate over the merits of formal/quantitative and qualitative methods of analysis; and fourth, Ray notes the attention to economic issues and variables in IR from the 1970s onward, which he regards as “real, tangible progress.” Ray also discerns “important insights” into quantitative IR from research on the democratic peace since the 1960s, which he surveys briefly. He quotes, approvingly, Vasquez's view that it is a “source of inspiration for a ‘rebirth’ of … quantitative approaches.” Moreover, he attributes the enormous growth of this literature in the 1990s not only to the spread of democracy after the cold war but also to vast improvements in the quality of evidence, the coherence of its theoretical structure, and the availability of large computerized data sets. His most novel “insight” relates to the theoretical impact of research on the democratic peace, likely to be as important as the evidence on the absence of war between democratic adversaries. Specifically, he cites the view of Bueno de Mesquita and associates that state behavior in conflict is motivated more by the wish of leaders to stay in power than the search for power or security per se; and this, in turn, facilitates the integration of domestic and international factors in explaining foreign policy and international politics, as postulated in Putnam's “two-level

games” concept. Turning to the future, Ray welcomes the broadening of the IR agenda in the last two decades to include class, ethnicity, gender, and the Third World. At the same time, he is critical of their “miscommunication, misperceptions, and misunderstandings.” Quantitative IR, he contends, is not free from these flaws. Moreover, he confesses to being “confused about the ideal model or theory or product toward which large-N analysts are striving”: “[an] extra-large multivariate equation? A set of simultaneous equations? Acomputer simulation? What?” And he is critical of the claim to total objectivity by quantitative IR specialists, calling on them to state clearly the aim of their research. At the same time, he is skeptical of postmodernist attacks on the statecentric Page 36 → model of mainstream IR and democracy: he cites Booth's remark, “The enemy is us. Western consumerist [capitalist] democracy … is the problem,” and urges advocates of critical perspectives to clarify what they propose in its place. Leng Much of Russell Leng's “Quantitative International Politics and Its Critics: Then and Now” takes the form of a debate with one of the earliest and most visible IR antiscience classicists. In a celebrated critique thirty-five years ago, Hedley Bull dismissed quantitative IR and its findings as marginal and its challenge to the classical—historical and philosophical—approach as “positively harmful.” Leng notes that regrettably during the first two decades of what Karl Deutsch termed the “fourth wave” of IR research—that is, the application of quantitative methods—“the only cross-fertilization between the two fields was in the form of mudslinging.” Bull's criticism was threefold: quantitative international politics research questions are—and are destined to remain—trivial because they are con- fined to aspects of IR that are measurable; it fails to cumulate knowledge; and methodological training of students is at the expense of access to the insights of classical scholarship. Leng responds to each of these points. Citing the self-doubt expressed by Thucydides about the reliability of his sources on the Peloponnesian Wars, Leng comments, “Issues of descriptive validity and reliability continue to be a problem in classical historical research…. A major advantage of the scientific approach is its methodological explicitness, its replicability.” He acknowledges that scientific research on world politics (quantitative international politics) has been unduly influenced by data availability and, further, that some aspects, normative and legal issues, are not accessible to scientific research. However, he claims that qualitative judgments are aided by empirical findings. Similarly, while accepting the postmodernist claim that research can never be entirely free from preconceptions, the effort to narrow the gap “distinguishes scientific research from pamphleteering.” And he invokes Weber's view that “the first presupposition of science is that the scientific method is valid.” As for Bull's second criticism, Leng responds vigorously, “We now know a great deal about war per se”: when it has occurred; its magnitude, intensity, severity, evolution over time, and relationship to structural changes in the international system; the fact that democracies do not go to war against each other; that a few states account for a large proportion of wars; that disputes over territory are a very frequent cause of war; and so on. And these findings, argues Leng, are not trivial. Yet they are only “islands of findings,” in Guetzkow's language, still in quest of an explanatory theory of war. Leng finds “some validity” in Bull's third criticism—“the scientific study of world politics could lead to an emphasis on methods over content” and the danger that students will not acquire the “feel” for the political and historical content of the subject matter of world politics; and, he adds, there Page 37 → is “the danger of dogmatism regarding what qualifies as acceptable knowledge.” Leng makes an intriguing comparison between quantitative international politics findings on war and research findings on coronary artery disease, and he views the former as no less persuasive than the latter. However, its policy relevance, unlike medical findings, depends on the political agenda of states.

Like Ray, he discusses the democratic peace thesis and notes the criticisms of the alleged spurious correlation—that democracies often do not share borders, are wealthy and thus satisfied with the status quo, and are more likely to be members of international organizations. Acknowledging that definitive proof of the validity of either major explanation, cultural or structural, is not yet possible, he cites the quest for indirect indicators. And on this point he concludes: “The scientific frontier is a contentious place…. [In the Darwinian tradition] no new finding is accepted without challenge, and only the most robust survive.” His concluding remarks are optimistic: “over the past decade some of the competitiveness has lost its edge, and there is a growing tendency to combine classical and scientific approaches”; he anticipates an end to their cold war through “an unconscious integrative solution.” Levy Jack S. Levy's wide-ranging assessment of “Qualitative Methods in International Relations” begins by noting a marked change—for the better—during the last three decades. In 1970 qualitative analysis was “historically specific rather than theoretically driven”; and its “ideographic nature” created an image as “highly subjective, pliable in fitting facts to theoretical arguments, nonreplicable, and essentially nonfalsifiable…. By the early 1990s, the [new, scientific] methodology of comparative and case study analysis had a well-defined place in the fields of comparative and international politics.” The closely interrelated foci of Levy's paper are “the meaning of the case concept, the different kinds and purposes of case studies, the different types of case study designs for testing theories, and the distinctive contributions and limitations of case study methods.” A “case” for Levy and, in his view, for most scholars is “an instance of something else, of a theoretically defined class of events…. [Moreover, cases] are made, not found; invented, not discovered.” He discusses several typologies of case studies, notably Lijphart's atheoretical, interpretive, hypothesis-generating, theory-confirming, theory-infirming, and deviant case studies and Eckstein's configurative-ideographic, disciplined configurative, heuristic, and crucial case studies, as well as the additional category of plausibility probe. Levy aligns himself with Achen and Snidal's view that dependable theory requires a more deductive orientation than case studies can provide. However, he notes several ways in which case studies can contribute to theory: hypothesis-generating case studies help to refine existing hypotheses; plausibility probes link hypothesis construction and hypothesis testing; deviant Page 38 → case studies “can help to identify omitted variables, interaction effects, or causal paths,” as can the analysis of the dependent variable. In sum, case studies play “an essential role in the explanation of individual historical episodes and a contributory role in the generation of hypotheses … [as well as] to test hypotheses and theories.” Levy evaluates several varieties of case selection strategies: “most different” and “most similar” systems designs, which he considers comparable to John Stuart Mill's “method of agreement” and “method of difference,” respectively (like most analysts, he favors “a careful selection of nonrandom cases” for the analysis of a limited number of cases); “no variance” designs, which are especially limiting; the crucial case strategy advocated by Eckstein, which is linked to “most likely” or “least likely” research designs; and the “process-tracing” method advocated by George, which he regards as fruitful, but he notes the problem raised by Hume, that is, “the logical impossibility of establishing causality from empirical observation.” Along with his overall positive evaluation of case study methods, Levy notes “rather serious limitations…. One is the large number of variables relative to the small number of cases…. A related problem is that case study methods cannot easily get at ‘probabilistic’ theories … a major strength of statistical analysis and a serious limitation of small-n research…. Case study methods also have difficulty in assessing the relative causal weights of the various factors influencing a particular outcome.” Levy is also aware that “statistical methods have their own limitations” and notes that many advocates of case study and statistical methods call for combining the two methods of analysis, though he deplores the few efforts to elaborate how this can be accomplished effectively. Beyond this technical task, he concludes that the most

important challenge for qualitative researchers is to develop strong theory and “to think more carefully about the testable implications of existing theories, in terms of quality as well as quantity … at the juncture of theory and research design.” Maoz In an in-depth critique, “Case Study Methodology in International Studies: From Storytelling to Hypothesis Testing,” Zeev Maoz focuses “on three aspects … that are symptomatic of the misuse of this methodology … case selection, case design, and cross-case comparison” and offers suggestions for improving the case study approach. He begins by noting some advantages of case study as a research method. It is “cost-effective”—that is, it enables “exploratory analyses at relatively low cost of resources.” Moreover, “the actual ability to match a hypothetical story that contains an explicit process derived from the theory with an actual story [process tracing] is a unique feature of the case study method.” It also has “the ability to uncover nonevents and their characteristics.” And “systematic comparison of processes may reveal unspecified features of the theory, Page 39 → even theories that were supported by quantitative evidence.” At the same time, Maoz deplores the paucity of cumulation in case studies of politics generally. Maoz's discussion of case selection contains an unconcealed criticism of many case studies “for freeform research where everything goes”: authors do not specify their research method; the criteria for selection of a case or several cases; their choice, processing, and analysis of data; and how their sweeping generalizations and lessons were derived from the case study, what he terms “heroic inferences.” In sum, “case selection entails the risk of selection bias and thus of resultant inferential bias.” To correct this shortcoming, he offers some principles of a “rational case selection strategy in case study design”: “Understand your theory. Lay out the hypotheses and spell out their status”—x as a necessary or sufficient condition of y, or both. “Define what kind of variability is required to test your theory…. Identify potentially confounding variables…. Explicitly define your population…. Attempt, as far as possible, to determine how the population is distributed on the confounding variables.” Maoz cites case design as “the key difference between exploratory and confirmatory case studies.” He defines “the art of case study design [as]—to a large extent—the art of reconstruction.” And he emphasizes that in confirmatory research, where “the design of the case is theory driven … the design … must be based on both a priori specification of evidence for the theory and evidence against the theory,” as well as identify the areas of competing and converging expectations. Using deterrence as an illustration, his guide to “a good case study” requires specifying evidence that a challenger intended to alter the status quo by force; “identifying the deterrence effort and its operational manifestations”; identifying evidence that the challenger has received and absorbed the threat; identifying evidence that the threat was considered in the challenger's decision-making process; identifying evidence that the deterrent threat altered the calculus of decision; identifying that the challenger decided not to violate the status quo, thereby making this a “‘candidate’ case for deterrence success.” But he emphasizes “the significance of testing evidence for a theory against competing explanations.” In his methodological critique, he also attaches importance to cross-case comparison, noting as “a necessary condition for generalization … a systematic design of individual cases and the execution of case studies in a manner that allows a fair comparison.” To that end he specifies several “basic principles” or guidelines of comparative case design: recognize that all cases have common and unique elements that need to be sorted out; maintain transparency; devise a cross-comparative design; “define standards of inference from multiple cases in light of possible variability in results across cases”; and indicate the possibly confounding factors revealed by case studies. Maoz concludes his paper with a call for pluralism: “we must be open to using multiple and varied methodologies to test our theories.” Page 40 →

Foreign Policy Analysis

The cluster of millennial reflections on foreign policy analysis comprises papers by four authors: Yaacov Y. I. Vertzberger, Stephen G. Walker, Ole R. Holsti, and Jonathan Wilkenfeld. Vertzberger The essence of Yaacov Vertzberger's sweeping critique, “Foreign Policy Analysis: Steady Progress and a HalfEmpty Glass,” is evident in his introductory “panoramic reflective view of the FPA [foreign policy analysis] landscape since the 1950s…. There has obviously been steady progress … in … theoretical and conceptual development, in … empirical richness and availability of both case studies and aggregate data analysis, and in increasing methodological diversity, rigor, and sophistication. Yet … we are still facing a half-empty glass.” The reasons for “straitjacketed progress” are “overlearning by imitation, decontextualization, oversimplification, and biases toward precision, conservatism, and risk aversion,” shortcomings that “represent a general malaise” in IS and, more broadly, political science. In a discussion of “the nature of progress,” Vertzberger ranks theory as most important, followed by empirical data, with methodology a distant third place in the research enterprise, and the logic of discovery as far more important to social scientific innovation than the logic of verification. He notes two types of progress: horizontal, an expansion in the amount of knowledge; and vertical, a qualitative change that is “more likely to be transformational.” Progress takes five forms, which are, in ascending order of significance, expansion, verification /falsification, integration, contextualization, and at the apex, extrapolation. These “may take place serially or simultaneously.” In foreign policy analysis, progress has centered on the first two of these forms. One of Vertzberger's intellectual bêtes noires is the unthinking adoption of norms and standards from the natural sciences, notably decontextualization, parsimony, precision, and methods—what he terms “scientificity”: “Once adopted, they acquired a life of their own and became difficult to unlearn.” Another major weakness of foreign policy analysis, in his view, is “excessive conservatism,” that is, confining its scope to diplomacy and security. Areas where it could have made a contribution include: IPE, specifically, foreign economic behavior by decision makers for states; institutionalism/ neoinstitutionalism; the role of accountability and policy acceptability in decision making and implementation; public policy; extending foreign policy analysis from a two-level to a three-level game, that is, domestic, international, and supranational, as with the EU; and systematic prescription so as to make foreign policy analysis policy-relevant, assessing the success or failure of policymakers' judgments and decisions. For Vertzberger, the grave methodological error of decontextualization in foreign policy analysis leads to “cognitive and motivational biases [that] Page 41 → are built into the parsimony imperative.” They are limitations on memory (“we have turned our natural cognitive limits into a normative scientific merit, … a search for a ‘do-itall’ paradigm … such as neorealism, rational choice theory, and their derivations that respond well to our biocognitive limitations”); the aggregation and sample bias; multiple causations and causation loops; the actor's motive bias; and slow integration. Turning to methodological progress, Vertzberger acknowledges “steady but major improvements and increased diversity.” However, these have been offset by three other serious shortcomings: the overvaluation of numbers—“numbers don't lie”; the “precision imperative,” part of “the drive toward scientificity”; and the high price in cost-effectiveness, the flawed assumption that costly methodologies pay off. Formal modeling, too, is assessed as a promise unfulfilled, without any important contributions to foreign policy analysis. Vertzberger concludes his reflections with a skeptical cognitive psychologist's view of theory and science: “Theory building in the social sciences in general, and international politics in particular, is the art of the possible. It does not always converge with the pure truth; at best it represents a practical truth. Evangelical scientificism and methodologism only stand in the way of creativity.” He closes the circle by reaffirming his respect for “the immense amount of accumulated knowledge since the late 1950s.” But, on balance, the foreign policy analysis glass emerges as much less than half full.

Walker (Stephen) The core concept of “Beliefs and Foreign Policy Analysis in the New Millennium,” beliefs, is used by Stephen Walker “generically to refer to the various conceptualizations for the products of cognition in the human mind—for example, schemata, attitudes, perceptions, images … all part of the ‘cognitivist’ research program in world politics.” Walker surveys the literature on beliefs and their impact on foreign policy behavior starting with the “seminal” work by Snyder, Bruck, and Sapin, which he identifies as the first serious challenge by the decisionmaking approach to realism's explanation of state behavior. He poses the question “Are beliefs merely epiphenomenal?”—that is, “merely doctrinal expressions of more fundamental bureaucratic interests, organizational structure, or institutional history”? He sums up the arguments of the critics thus: first, the influence of beliefs on foreign policy behavior is important and independent “only under a restricted range of conditions”—external crises; second, they constitute only one variable in a larger explanatory constellation; third, their effects are “indirect and general rather than direct and specific; fourth, they serve as an intervening variable only; fifth, they are “extraneous variables in foreign policy and world politics and [are] best ignored except … [for] deviant cases.” This is a formidable critique, to which he replies at the end of his paper, beliefs “are pivotal variables with mapping, steering, and solution effects within a dynamic interactive Page 42 → causal model rather than being simply intervening or dependent variables.” On the follow-on question “Are beliefs important causal mechanisms?” Walker cites Goldstein and Keohane's affirmative view that there are three types of causal role for beliefs: as “worldviews” for leaders, exerting mapping effects; as “principled beliefs,” exerting normative influence on behavior; and as “causal beliefs” in problem representation. The reasons for the resurgence of interest in beliefs, Walker writes, were the failure of scholars to anticipate the end of the cold war and a growing dissent from rational choice theory. He notes that “more recent work by rational choice theorists … recognizes beliefs as important intervening variables” and cites the distinction between “thin” and “thick” models of rationality, in which beliefs are relatively unimportant and relatively important, respectively. On that basis, he sketches “a strategy for synthesis,” with two lines of inquiry—beliefs as micro-level and as macro-level mechanisms of choice. The second half of Walker's millennial reflections paper sets out “a research agenda for future progress”: it emphasizes the analysis of initiatives toward peace as well as toward war, more specifically, the de-escalation phase of an international crisis—for example, the Sadat initiative to Israel in 1977. Such a “reconceptualization of the conventional crisis model expands the universe of cases [and permits viewing] military initiatives and diplomatic initiatives within the same general analytic framework.” His “interaction game … integrates threats of conflict and opportunities for cooperation, thereby synthesizing a focus on [explanations by the realist and liberal schools of thought].” These are the benefits of what he terms “a bidirectional, strategic approach with a two-level, psychological approach.” Holsti (O. R.) The primary focus of Ole Holsti's millennial reflections paper, “Public Opinion and Foreign Policy Analysis: Where We Were, Are, and Should Strive to Be,” is “some main currents of theory and research.” He begins by noting the dominant view of public opinion held in the 1950s, the result of influential writings by realists and American scholars, reinforced by the early findings of scientific public opinion polls. In Holsti's words, “Illinformed, emotion-driven, short-sighted, and self-absorbed publics constitute a major obstacle that threatens to prevent the United States and other Western democracies from playing a constructive role in an increasingly dangerous international system.” It took several decades to dethrone that harshly critical assessment. The catalyst for change was the Vietnam War, which generated a plethora of studies, including large-scale survey projects on U.S. foreign policy, like the quadrennial surveys by the Chicago Council on Foreign Relations since 1974, along with works by U.S. scholars, many of whom are cited in Holsti's comprehensive survey. Page 43 →

The upshot was a successful challenge to the earlier negative view, with “some convergence on two points of relevance to foreign policy analysis: First, even… rather poorly informed … attitudes about foreign affairs are in fact structured in moderately coherent ways…. Second, there is growing evidence that the single internationalistto-isolationist dimension … is inadequate.” This was replaced by the Chicago Council's three perspectives—conservative internationalism, liberal internationalism, and noninternationalism—which Wittkopf transformed into “two faces of internationalism,” militant and cooperative, yielding four foreign policy belief systems: hard-liners, accommodationists, internationalists, and isolationists. Nonetheless, observes Holsti, there has been more progress on description of trends in public opinion than on analysis of its impact on the foreign policy process. Holsti also assesses the three methodologies that have been used in this field—quantitative-correlational analysis, case study, and interviews. The “impressive” findings from the first of these methods reveal very little difference in the extent of congruence between public opinion and foreign policy (62 percent) and the public opiniondomestic policy nexus (70 percent). The results from case studies are mixed, indicating that “the impact of public opinion varies according to the type of issue, the stage in the policy process, the decision context, and policymakers' beliefs about and sensitivity to public opinion…. Public opinion had a direct impact on getting the issue on the agenda and on ratification of the agreements, but only an indirect effect on negotiations and implementation.” Turning to the future, Holsti suggests three major areas of research to improve our understanding of the public opinion-foreign policy process linkage. The first takes the form of specifying eight questions that should guide future research—such as the indicators of public sentiment relied upon by leaders; the purpose for which this information was used (to gain public support or to manipulate/influence public attitudes); and whether the timing of decisions and the choice of means to implement them were influenced by beliefs about what the public would accept, not accept, or demand, among other pertinent questions. The second recommendation is to engage in cross-national studies, for, he notes, “even among democracies, the nature and impact of public opinion may vary across countries and political systems.” And the third is the need to standardize questions for polling on foreign policy issues. Holsti concludes that much progress has been made since the 1950s and that public opinion “remains a vibrant area of inquiry.” However, he expresses concern about the danger of leaders using increasing survey data to manipulate the public and the evidence of “persistent poverty of international knowledge, despite the information revolution.” Wilkenfeld Of the value of experimentation in foreign policy analysis in general Jonathan Wilkenfeld has no doubt: “Experimentation allows researchers to isolate a Page 44 → crucial independent variable by providing the means to control extraneous factors…. It is the most reliable means of examining and understanding the nature of a causal relationship, of demonstrating the internal validity of a theory in political science,” he writes in “Simulation and Experimentation in Foreign Policy Analysis: Some Personal Observations on Problems and Prospects.” The essay contains a brief survey of experimental work in IR, followed by a report on the Conflict and Negotiation (CAN) Research Group, which is designed to enhance understanding of how humans and machines negotiate in crises to attain mutually beneficial outcomes. Wilkenfeld notes three features of CAN research that distinguish it from other experimental work in foreign policy analysis: an exclusive focus on behavior in crisis; the reliance of its subjects on decision support systems (DSSs); and the conduct of all communications in controlled network environments. The project has addressed five broad research issues so far, namely, “the impact of the use of DSSs on the utility-maximizing behavior of crisis negotiators”; “the impact of the dynamics of crisis negotiations on their outcomes”; “the relationship between the level of cognitive complexity of crisis negotiators and the outcomes of crisis negotiations”; the impact of the same mix, with the addition of crisis mediators, along with the link between crisis negotiators' level of cognitive complexity and crisis outcomes; and “the impact of the mix of cognitive complexity levels among crisis

negotiators on the outcome of the crisis.” Wilkenfeld puts great emphasis on the analytical value of merging empirical and experimental approaches to foreign policy analysis. His “merged research design”—using data from a large-N empirical study, the International Crisis Behavior (ICB) project, and the findings from experimental work on mediation in international crises—is the focus of the rest of his paper. While a consensus definition of mediation remains elusive, there seems to be agreement on its three roles in negotiation—the mediator as communicator or facilitator, as substantive contributor, and as manipulator. Two issues were addressed using ICB data: the circumstances in which mediation is most likely to occur and the most likely types of outcome from mediation. Wilkenfeld reports both empirical and experimental findings on mediation in international crises. The former relate to threat, polarity, system level, ethnic origins of crisis, geographic proximity, and types of outcome—content, form, extent of satisfaction, and escalation or reduction of tension. On outcomes he notes: “Clearly, … mediation makes a difference”; it is more likely to lead to compromise outcomes, agreements, mutual satisfaction by the parties, and a reduction of tension over time. The experimental findings are summed up thus: “mediator as manipulator produced higher utility outcomes for both parties, while mediator as facilitator produced greater mutual benefit…. [the former led to] shorter elapsed time to outcome in general, [the latter to] shorter elapsed time to agreements…. all parties were dissatisfied with mediation…. [and] there were no discernible differences in satisfaction with outcome resulting from the two forms of mediation.” Page 45 → Wilkenfeld has no illusions about the prospects for experimental work in foreign policy analysis. This is partly because the transition from simulation to an experimental lab is not simple, partly because most subjects, students, are relatively uninformed and often indifferent or hostile to the rigorous task of experimentation, and partly because of the expense and time required for design and execution of experiments. Yet he views experimentation as a valuable supplement to cross-national and in-depth case studies; and to that end, he urges proper training of future scholars in foreign policy analysis.

International Security, Peace, and War Papers by Edward A. Kolodziej, Davis B. Bobrow, Louis Kriesberg, J. David Singer, and Linda B. Miller, make up this cluster of millennial reflections. Kolodziej Edward Kolodziej's comprehensive survey, “Security Theory: Six Paradigms Searching for Security,” is divided into two sections: a brief discussion of the concept of security and a much longer assessment of realism, neorealism, economic liberalism, institutionalism, behaviorism, and constructivism, noting key differences and their implications for security theory. He acknowledges the existence of diverse views on the meaning and scope of security, from “preservation of the state and the nation-state system” and “physical security of individuals within civil society” to these conceptions enlarged to include “economic welfare” and/or the avoidance of anxiety. As a starting point, he opts for the traditional view that identifies security with power and violence, deriving from Thucydides and sustained by Machiavelli, Hobbes, and the modern exemplars of realism. However, “security theory … [must] be principally concerned with explaining both peace and war”; that is, it “may begin with Hobbes, but it must end with Kant and the possibility of perpetual peace if it is to be comprehensive and coherent.” The bulk of Kolodziej's paper takes the form of an analysis of the central tenets of each of the six

competing IR paradigms and of their failure to anticipate the collapse of the Soviet Union and the concomitant peaceful end of the cold war. Realism's and neorealism's ex post facto explanation emphasizes the growing economic and technological disparity between the two superpowers; this compelled the USSR to adopt a (failed) domestic reform program and retrenchment abroad; and this in turn generated an increasing imbalance in military power. Under the impact of that growing pressure the Soviet Union could no longer compete with the United States, and its imperial domain, external and internal, crumbled. Economic liberalism, too, had an ex post facto explanation for the pressures leading to Soviet attempts at economic reform; but “neoclassical economic theory has no theory of power” and could not explain the two related transforming events in world politics at the end of the 1980s. Liberal institutionalism Page 46 → describes actor behavior more effectively than realism, argues Kolodziej; however, by sacrificing parsimony to complexity, it too failed to anticipate or explain either of these seminal developments. For behaviorism, the end of the cold war and the collapse of the Soviet Union were “instances of a larger set of actor actions and reactions relevant to a pattern of behavior of interest to the researcher”; this preoccupation with methodology prevented any contribution to an explanation of these events. As for constructivism: some adherents emphasized cognitive reformulation (learning) by Gorbachev and his colleagues, recognizing their inferiority in military power and technological and economic sources of growth vis-à-vis the West; others focused on structural constraints, like resurgence of nationalism and the delegitimization of communist ideology and regimes throughout eastern Europe. Like all the other paradigms, these were belated, after-the-fact attempts at explanation. The findings for security studies, notes Kolodziej, are both negative and positive. No single paradigm can explain security behavior fully. No single methodology can lead to a definitive theory of state behavior. No single policy prescription can overcome the security dilemma. And there is no consensus on which actors have a crucial impact on security. At the same time, each of these paradigms “provide useful, if not fully satisfactory, theoretical tools” in search of a theory. In conclusion, one might despair over an array of competing paradigms. However, “pluralism serves us better than we might believe—or wish.” Bobrow In “Security and Peace: Understanding, Production, and Work Style,” Davis Bobrow presents a wide-ranging, critical assessment of the state of this subfield of international studies. His reflections begin with “The ‘Good News,’” followed by “The ‘Darker Side,’” “Some Lessons (We Should Have) Learned,” and “Some Suggested Emphases.” The bulk of this millennial reflections paper takes the form of a positive and negative evaluation of six expansions in this domain. These are identified as (1) “a cornucopia of ‘prisms’ for considering security and peace” beyond realism and idealism, namely, rational choice, institutionalism, critical theory, gender, power transition, domestic politics, failed state, democratic peace, and ideationalism; (2) “issue domains” beyond the political-military, including economic, environmental, demographic, public health, sociocultural continuity, and fundamental political forms; (3) “generic types of actors inside and outside of governments,” comprising civil departments of central governments, subnational authorities, multinationals, NGOs, and IGOs; (4) “relevant specific actors with asymmetric agendas and assets,” notably lesser powers on the periphery of the dominant system; (5) the revolution in technology, leading to a proliferation of independent and intervening variables in the security-peace nexus; and (6) a vast increase in information resources. The positive aspect, notes Bobrow, is “a period of ‘riches’ with respect to concepts, phenomena, information, and human capital.” At the same time, Page 47 → there is a “darker side,” for two reasons: first, “the first five expansions involve resurrection and relabeling rather than genuine discovery and innovation”; and second, “the improvements may not have kept pace with the challenges posed by the … expansions.” Thus, for example, while more prisms are welcome, this does not ensure a wise synthesis, as is evident in two high-visibility current

theories, power transition and the democratic peace: the U.S. policy community seems to subscribe to both simultaneously, but “international audiences may instead see a realist draped in an idealist democratic cloak.” His reservations about the benefits of the other five expansions are outlined. “The more fundamental difficulty is that the truth often is hard to know about recent and current security and peace matters.” The general lesson for Bobrow is that “security and peace are and will remain contested or captured concepts,” without a consensus on content, sufficient conditions, and a reliable indicator as to when, and to what extent, they exist. The operational lesson is that the six expansions do not provide a reliable guide for policy analysis and prescription; rather, “we should accept a continuing need to adapt to them.” In his concluding reflections Bobrow urges the security-peace research community to display humility and modesty and notes the merit of pluralism—a recognition that each element of his noted prisms, issue domain, or generic type of actor is relevant, but none “tells the whole story.” He also recommends three practical changes: “in the education of security and peace specialists,” who must acquire literacy in science and technology; better use of available resources, including non-American peace and security analyses; and the adoption of Wohlstetter's applied systems analysis approach, both “opposed systems design” and “end-to-end analysis,” for explanation, prediction, and prescription, because they “are well-suited to accommodating and indeed being strengthened by the six widely recognized expansions.” Kriesberg A clearly articulated goal permeates Louis Kriesberg's “Convergences between International Security Studies and Peace Studies”: to enhance cooperation between these two subfields of international studies. The paper divides into three sections: “Earlier Relations,” “Contemporary Developments,” and “Future Developments.” In the past, notes Kriesberg, there were several significant differences between these subfields: their IR intellectual sources for one—realism for security studies, liberal-idealism for peace studies; the latter was also influenced by the feminist perspective in IR, the impact of religion and culture on war and peace, the role of transnational social movements, and the critical perspective on global political economy. In addition, security studies focused on one primary actor, the analysts' own state, military means, and the avoidance of war—that is, negative peace—while peace studies concentrated on global and systemic concerns, nonviolent methods, and justice and equity—that is, positive peace. Page 48 → Moreover, the former relied on foundation and government funding, while the latter concentrated on social movement organizations. These differences in theory, research, and practice were more evident in the United States than in Europe. There were, however, some bridge-building foci, notably, a shared interest in the causes of war, crisis management, and foreign policy decision making. During the four decades of the cold war, security analysts concentrated on the military dimension, including nuclear and general deterrence, while peace analysts focused on the peace movement and the military-industrial complex. Neither group anticipated the end of the East-West conflict, and each offered a different ex post facto explanation. According to the security analysts, the West's containment of the Soviet Union and the arms buildup forced Moscow to yield; the peace analysts placed credence in the Western policy of reassurance as embodied in the 1975 Helsinki Accords and the legitimacy given to post-World War II borders in Eastern Europe. Finally, the two groups tended to ignore each other's work. In short, they differed on the questions they posed, methodologies, the answers they provided, and their underlying values and goals. Rapid global changes after the end of the cold war—growing economic integration, vast improvements in communication, and the reduction of state sovereignty as a result of the increase in number, size, and resources of multinational corporations, transnational organizations, and intergovernmental institutions—moved the two subfields closer together.

Norms of international human rights, including the necessity of international military intervention in extreme cases, are increasingly shared. So, too, is a commitment to extend the ambit of international conflict resolution, the prevention of large-scale violence, and the building of stable peace. Turning to the future, Kriesberg anticipates—and welcomes—the emergence of new subfields focusing on substantive issues. One “is likely to relate increasingly to different conflict stages: prevention, limiting and interrupting violent conflict escalation, terminating violence, and postconflict peace building,” as well as extending the role of mediation and negotiation as instruments of conflict termination. Other new subfields will concentrate on early warning and preventive diplomacy, interrupting and stopping violence, peacekeeping, creating institutions to reduce intercommunal conflicts, extending democratic processes and institutions, and enlarging the scope of civil society to sustain them. Kriesberg also expects a convergence between peace and security studies in their common attention to the link between democratic polities and foreign intervention and to the task of achieving reconciliation between conflict adversaries. He concludes on a note of optimism about the continuing convergence of the two subfields. And he views this as being aided by the increasing involvement of women in both subfields, along with the growth of centers of conflict resolution in the Third World. But he also notes the risk of peace research being co-opted by political authorities to serve their end; and he expects the time frame of the two fields to continue to differ—short term for security studies, long term for peace studies: “Such differences in approach … are needed.” Page 49 → Singer The raison d'être of J. David Singer's millennial reflections paper is the need “to get more serious about trying to explain, and perhaps reduce the incidence of, that brutal, stupid, and destructive form of collective behavior known as war.” To that end he begins “Accounting for Interstate War: Progress and Cumulation” with a blunt critique of what he terms the illusory view that “such a process is already underway.” Singer's criticism is directed at many in the world politics community. Citing the optimism of Mueller and Levy, he contends that “the indicators to the contrary are difficult to ignore”—landmines in many parts of the world, “the steadily increasing worldwide trade in light weapons, the obscene U.S. military budget, and the presence of more than thirteen thousand strategic warheads on high alert among the nuclear powers.” He also dismisses the “Kantian peace” school and Axelrod's “evolution of cooperation” as “week reeds”: “there is not much hard evidence to support the alleged expansion of democracy”; and “how often in history have we seen most states … behaving in a cooperative fashion…. So much for peace on the cheap.” Singer's wrecking operation is also evident in his later discussion of “Premises: Epistemological and Empirical.” One object of disdain is “the posties”—postmodernism, poststructuralism, postbehaviorism, postpositivism, and constructivism—which, denying the existence of any empirical reality, consider a science of human behavior as “mission impossible.” Another is Alker who calls “for more hermeneutics and less exegesis.” Still another is followers of Bull, who “assure us that the classical methods of the historian will do just fine.” There is also a “variety of dubious epistemologies … [that] lead to … increasingly ambitious ‘theories’ that are, in truth, little more than speculative models that rest on problematic and far from operational and thus testable premises.” He cites two variants—the “analytic narrative” approach and “the so-called ‘structured, focused, comparison,’” whose individual case studies result “in something only weakly structured, focused, and comparative.” Singer also castigates the “doomed effort to ascertain/demonstrate ‘causality,’” as well as “evolutionary epistemology.” And he finds some major works in the philosophy of science—Hempel, Popper, Harre, and Lakatos—“not very helpful.” A negative tone also permeates Singer's extensive discussion of “semantic problems.” He rejects “international relations/IR” in favor of “world politics”—“international misses a lot.” He criticizes the “archaic” concept of nation-state—he prefers “territorial state”—and the misuse of “bipolarity” and “multipolarity”; the use of “collective defense” and “collective security” as synonyms; the facade of terming “military spending” a “defense budget”; and especially, the misunderstanding of “theory” and “data.” He also is harsh on “the pernicious

distinction between positive theory and normative theory.” In a more positive vein, Singer sets out his four-stage reading of the quest for an explanation of war: (1) preWorld War I—a literature that was “interesting, Page 50 → diverse, and insightful, but resting largely on assertion, speculation, and anecdote”; (2) 1920 to 1960, notably the work of Richardson and Wright, which “illustrated the possibility of utilizing scientific method in our quest”; (3) his Correlates of War Project, ongoing since the early 1960s, the “natural history” stage, “combining a rigorous methodology with an explicitly agnostic, multitheoretical orientation”, and (4) the current stage, just beginning, testing more formal models with the results of the Correlates of War's “systematic historical empiricism.” Much later in his paper Singer spells out the “empirical-epistemic assumptions” underlying his project. The “central axiom” is the presence of “highly regular and thus recurring patterns” in the global system, along with a recognition of “plenty of exceptions” and an inherent randomness, leading to “statistical distributions rather than deterministic and perfectly uniform law” and erratic advances in the search for large regularities. He concludes with a reprint of a summary of his 1975 article to highlight the continuing “bad habits” that thwart cumulative progress in 2000. Miller Linda Miller's reflections on international security take the form of personal “Notes from the Underground: A Tale of Three Perspectives”: the “underground” refers to her career as teacher of mostly female undergraduates at U.S. liberal arts colleges; and her “perspectives” are those of student, teacher-researcher, and editor. Her preferences are indicated forthrightly. She lauds Wolfers's notion that security is “an ambiguous symbol”; the “important role of history as the database for international relations rather than the hard sciences … a self-evident truth”; “the interplay of domestic and international factors in … analyses of security and insecurity”; and political theory as a “rich source of complex questions” relating to security. These are “bedrock views.” Miller is no less blunt in her criticism of much of the IR literature—theory, methodology, and research findings. Waltz's Theory of International Politics “threaten[s] to distort … [IR] by narrowing the agenda of acceptable questions to investigate more than twenty years after publication”; the democratic peace findings “might be limited to the North Atlantic-North American region, rather than constitute a law of global politics in the twentyfirst century”; the “self-styled ‘great debates’” in the security field often bury “the organic connection between theory and practice.” She also decries the application of “‘scientific’ methods” to the social sciences; the denigration of ideas in explanations of transforming events such as the end of the cold war; the “primacy of theory” (“we know that these intellectual constructs are just that. Theories do not and cannot on their own delineate the range of choice and consequence, let alone the sequence of cause and effect”); the hazards of relying on other disciplines, notably economics—“that unfortunate deformity”; the “infatuation with natural science models” and the clinging by political scientists “to obsolete Page 51 → hard science methods”; and “a professoriate that is increasingly specialized and doctrinaire.” All this earns the opprobrium, “this dreary trend.” As for “the way forward,” Miller strongly recommends multidisciplinary cooperative research: “groups of scholars brought together by a belief that each of their disciplines has something needed to complete an empirically rich picture of the phenomena.” She cites as a model the Borderline series of the University of Minnesota Press. She also lauds the growing interest in normative “theories” and the ethical dimension of IR, concluding, “our studies must try to repair the world.”

International Political Economy Three millennial reflections papers on international political economy (IPE) are offered—by Helen Milner, Robert Kudrle, and Lisa Martin.

Milner Helen Milner's “Reflections on the Field of International Political Economy” presents a sweeping survey of IPE since its emergence in the early 1970s under the impact of the first wave of oil shocks experienced by the developed parts of the world economy—Europe, North America, and Japan. “What is IPE?” Milner asks, and she first draws attention to the two competing views of its proper scope. One embraces everything in IR other than security studies; that is, it includes all aspects of international institutions, the environment, human rights, gender, minorities, and international cooperation writ large, as well as the economic domain of IR. In this context, Milner cites, but emphatically rejects, Susan Strange's extreme formulation: “Far from being a subdiscipline of international relations, IPE should claim that international relations are a subdiscipline of IPE.” The second view, to which Milner subscribes, limits IPE to the “interaction of politics and economics, or more narrowly markets and states…. Without an economic component, I would argue, … phenomena are not properly part of IPE.” And to those who propose expanding the field even beyond Strange, she responds: “the role of norms and values, ‘critical’ theory, reflectivity, methodology, epistemology, and ontology … are not intrinsically or largely the domain of IPE.” At the same time, any international issue or topic with an economic dimension, such as the economic causes of war and peace and the economic consequences of conflict, fall within the domain of IPE. Milner considers IPE not merely a subfield of IR but rather as extending beyond, with strong links to economics and comparative politics. Regarding the first of these disciplinary links, she notes that some IPE scholars rely heavily on economic models and tend to fuse the two fields, while others insist that economists should add political variables to their explanations. As for IPE and comparative politics, “Integrating the two fields seems to be an important avenue for future progress…. It is imperative to a better understanding of IPE topics … to use models of domestic politics that explain Page 52 → how policy is made…. It is difficult today to distinguish the fields of comparative and international political economy.” In contrast to security studies, where a consensus exists on the centrality of war, IPE lacks a single focus of attention. She notes three crucial foci: “economic development and differential growth rates”; “the impact of the international economy on domestic politics and vice versa”; and “conflict and cooperation among states in the international economy.” The IPE research agenda has changed over time, notes Milner. During the 1970s and early 1980s there were five key issues: economic sanctions versus military force as a means of achieving foreign policy goals; hegemonic stability theory; the role of domestic politics; dependency theory and development; and international institutions. By the late 1980s, two of these—hegemonic stability theory and dependency theory—had become passé, overtaken by events. Others had shifted their focus under the influence of globalization. In 2000, she specifies the main IPE topics as the effect of regime type, especially democratization, on development; globalization; and, with an altered focus, the role of international institutions, as well as the effects of domestic politics on economic integration. Milner cites two methodology debates in IPE: the merits of rational choice models and the value of large-N versus small-N studies. Formalism helps sometimes, she writes, if one wants to communicate with economists; but rational choice does not explain everything about IPE. As for aggregate data versus case studies, she shares the view expressed by many, namely, a preference for a combination of the two. She concludes by calling for a response to the increasing penetration of IPE by economists: either a drastic break with economics, that is, establishing the autonomy of IPE, or “taking on the economists on their own ground … by doing more economically sophisticated work in the field … [and fostering] an interdisciplinary brand of IPE.” Kudrle Robert Kudrle's “natural starting point” for “Some Thoughts on International Political Economy in the Context of Public Policy Education” is William Riker's definition of political economy as “the application of microeconomics to political phenomena.” In contrast to Milner, he stresses method, rather than substance, to explore the sources of

foreign economic policy. He treats this “as a species of public policy,” linked to comparative politics, though from an economist's perspective. Kudrle identifies three fundamental goals in international affairs—autonomy, security, and prosperity. The first is “akin to common definitions of nationalism.” The second is operationalized as “protection from foreign violence.” And the third is “the maximization of national wealth because modern nation states command redistribution policies.” A fourth goal, though less universally important, is cited: “standing,” that is, “international rank and influence upon, or regard from, those in other states”—“[but] the valuation [of these goals] cannot be directly observed.” Page 53 → In the section “Basic Approach,” Kudrle distinguishes between preferences and constraints: the latter include “the structure of the international system and the place of the state in it, the levels of income and technology of system members, and the degree of congruence of purpose in various dimensions of the state in question.” The presentation of his approach is followed by a lengthy survey of recent research relevant to IPE. This includes the role of ideas, game theory, “policy networks” in framing public policy, constructivist and rational choice literature, and works on the “new institutional economics.” And he expresses the novel view, for an economist, that “qualitative analysis remains essential in careful political analysis, including IPE … for the indefinite future.” In “Looking to the Future,” Kudrle focuses on globalization. He is skeptical about its threat to key areas of national policy, notably growth and taxation, though he acknowledges its increasing effect on attitudes and identities relevant to national goals. He cites approvingly Rodrik's notion of a political economy “trilemma”—of the nation-state, mass politics, and integrated national economies. And he notes some obstacles to further European integration, including the role of cultural differences in inhibiting labor mobility and looming fiscal tensions. He concludes with remarks about the three categories of international economic interaction: exchange, which has little effect; penetration, highlighted by the role of international corporations; and constraints on the operation of domestic policy by the forces of globalization. And he notes the increasing relevance of public finance, labor economics, and environmental economics in the study of policy aspects of IPE. Martin In “International Political Economy: From Paradigmatic Debates to Productive Disagreements,” Lisa L. Martin dissents from the editors' assumption that the task of identifying intrasubfield consensus is difficult: IPE, she asserts, “is today characterized by growing consensus on theories, methods, analytical frameworks, and important questions…. disagreement today generally takes the form of productive, theoretically and empirically motivated claims, rather than the paradigmatic clashes that characterized IPE as an early field of study.” In a brief discussion of the development of IPE as a field of study, Martin acknowledges Gilpin's contribution, which viewed IPE as divided into three paradigms: realism, liberalism, and Marxism. However, she notes “an alternative organizing device,” a dual focus on the explanatory role of interests or institutions, at the domestic or international level. This, in turn, after “a deep divide between IPE and international economics,” led to “intense interaction between economics and political scientists,” with a growing shift from an IR to a political economy orientation, especially in the area of international trade. Specifically, “Much of the debate today over interests on trade policy revolves around applications of Stolper-Samuelson versus the Page 54 → Ricardo-Viner model.” Similarly, the analysis of domestic institutions benefits from input from both disciplines. According to Martin, IPE is characterized by “productive research … in all four of the boxes” of the interests /institutions and domestic/international framework in IPE. Overall, “IPE is a mature, exciting field of inquiry…. Current research agendas provide the tools and point the way; but a multitude of challenges remain,” notably “breaking down the barriers between interests and institutions and between the domestic and international levels

of analysis … [in] an exciting and tractable research agenda for the future.”

Notes 1. E. H. Carr, The Twenty Years' Crisis, 1919–1939: An Introduction to the Study of International Relations (London: Macmillan, 1939); E. H. Carr, Conditions of Peace (London: Macmillan, 1942); Arnold Wolfers, Britain and France between Two Wars: Conflicting Strategies of Peace since Versailles (New York: Harcourt, Brace, 1940); Arnold Wolfers, Discord and Collaboration: Essays on International Politics (Baltimore: Johns Hopkins University Press, 1962); Nicholas J. Spykman, America's Strategy in World Politics: The United States and the Balance of Power (New York: Harcourt, Brace, 1941); W. T. R. Fox, The Super-Powers: The United States, Britain, and the Soviet Union—Their Responsibility for Peace (New York: Harcourt, Brace, 1944); Hans Morgenthau, Scientific Man versus Power Politics (Chicago: University of Chicago Press, 1946); Hans Morgenthau, Politics among Nations: The Struggle for Power and Peace (New York: Knopf, 1948); Bernard Brodie, ed., The Absolute Weapon: Atomic Power and World Order (New York: Harcourt, Brace, 1946). 2. Kenneth N. Waltz, Theory of International Politics (Reading, Mass.: Addison-Wesley, 1979); Robert Gilpin, War and Change in World Politics (Cambridge: Cambridge University Press, 1981). 3. E. B. Haas, The Uniting of Europe (Stanford, Calif.: Stanford University Press, 1958); E. B. Haas, Beyond the Nation-State: Functionalism and International Organization (Stanford, Calif.: Stanford University Press, 1964); Robert O. Keohane and Joseph S. Nye Jr., Power and Interdependence: World Politics in Transition (Boston: Little, Brown, 1977). 4. Robert Jervis, Perception and Misperception in International Politics (Princeton, N.J.: Princeton University Press, 1976); Robert Jervis, Richard Ned Lebow, and Janice Gross Stein, Psychology and Deterrence (Baltimore: Johns Hopkins University Press, 1985). 5. Richard K. Ashley, “The Poverty of Neorealism,” International Organization 38, no. 2 (1984): 225–86. 6. Robert W. Cox, “Social Forces, States, and World Orders: Beyond International Relations Theory,” Millennium 10, no. 2 (1981): 126–55; Robert W. Cox, Production, Power, and World Order: Social Forces in the Making of History (New York: Columbia University Press, 1987). 7. Alexander Wendt, “The Agent-Structure Problem in International Relations Theory,” International Organization 41, no. 3 (1987): 335–70; F. V. Kratochwil, Rules, Norms, and Decisions: On the Conditions of Practical and Legal Reasoning in International Relations and Domestic Affairs (Cambridge: Cambridge University Press, 1989). 8. Jean B. Elshtain, Women and War (New York: Basic Books, 1987); Cynthia Enloe, Bananas, Beaches, and Bases: Making Feminist Sense of International Politics (Berkeley: University of California Press, 1990); V. Spike Peterson, ed., Gendered States: (Re)Visions of International Relations Theory (Boulder, Colo.: Lynne Rienner, 1992); J. Ann Tickner, Gender in International Relations: Feminist Perspectives on Achieving Global Security (Minneapolis: University of Minnesota Press, 1992).

Page 55 →

REALISM

Page 56 → Page 57 →

REALISM, THE REAL WORLD, AND THE ACADEMY John J. Mearsheimer Realism, with its emphasis on security competition and war among the great powers, has dominated the study of international relations over the past fifty years. Hans Morgenthau and Kenneth Waltz, in particular, have towered over the field during that period. Years from now when the history of the discipline in the twentieth century is written, two books will stand out above the rest: Morgenthau's Politics among Nations and Waltz's Theory of International Politics. They have few serious competitors. Many Americans and Europeans, however, believe that realism has a dim future. With the end of the cold war, so the argument goes, international politics has changed in fundamental ways. The world has not simply moved from bipolarity to multipolarity, but instead we have entered an era where there is little prospect of security competition among the great powers, not to mention war, and where concepts such as polarity and the balance of power matter little for understanding international relations. Most states now view each other as members of an emerging “international community,” not as potential military rivals. Opportunities for cooperation are abundant in this new world, and the result is likely to be increased prosperity and peace for almost all the states in the system. Not surprisingly, proponents of this optimistic perspective argue that realism is old thinking, and largely irrelevant to the new realities of world politics. Realists have gone the way of the dinosaurs; they just don't realize it. The best that might be said for realism is that it was helpful for understanding how states interacted with each other before 1990, but it is largely useless now that the cold war is over. Therefore, we need new theories to help us make sense of international politics in the twenty-first century. President Clinton was a strong proponent of this view. For example, he declared in 1992 that “in a world where freedom, not tyranny, is on the march, the cynical calculus of pure power politics simply does not compute. It is ill-suited to a new era.” Five years later in 1997 he sounded the same theme in defense of NATO expansion. The president argued that the charge that this policy might isolate Russia was based on the belief “that the great Page 58 → power territorial politics of the 20th century will dominate the 21st century,” which he rejected. Instead, he emphasized that “enlightened self-interest, as well as shared values, will compel countries to define their greatness in more constructive ways … and will compel us to cooperate in more constructive ways.”1 This hopeful perspective is widely shared by academics. Consider two important statements that appeared in 1989, just as the cold war was coming to a peaceful conclusion. John Mueller claimed in Retreat from Doomsday that there is no longer a serious threat of war among the great powers, because wars of that kind have become too deadly (even without nuclear weapons) to have any political utility. “For the last two or three centuries,” he wrote, “major war—war among developed countries—has gradually moved towards terminal disrepute because of its perceived repulsiveness and futility.”2 Like dueling and slavery, great-power war no longer serves a useful social purpose. Francis Fukuyama made a similar claim in his famous article “The End of History?”3 He argued that “large-scale conflict must involve large states still caught in the grip of history, and they are what appear to be passing from the scene.” With the collapse of the Soviet Union, “there is no struggle or conflict over ‘large’ issues, and consequently no need for generals or statesmen; what remains is primarily economic activity.” The claim that realism has bitten the dust is wrong. In fact, realist theories are likely to dominate academic debates about international politics for the next century, much the way they have since at least the early days of the cold war.4 What the future has in store for realism will be largely a function of two considerations. First, what will it tell us about events in the real world? Because the study of international politics is an empirical science, the most important criterion for assessing the worth of any theory is how well it explains state behavior. Realism has long

been recognized as the dominant paradigm in international relations—even by scholars who dislike it—because it does a better job of explaining politics among states than any other body of theories. Second, what kind of respect will academia—one of the key institutions in any society for fostering thinking about world politics—accord realism? The potential problem for realism is that academics sometimes attempt to denigrate and silence ideas they dislike, even when those ideas shed light on important subjects. Realism appears to have a bright future in the twenty-first century. Unfortunately, we still live in a nasty and brutish world where the great powers compete with each other for power. The only possible threat to realism is likely to come from inside academia, where it is frequently reviled. But any attempt to silence realism within the academy is likely to fail, simply because it is so difficult to repress or exclude compelling arguments, especially in the United States.

Realism The realist paradigm comprises a body of theories that share a handful of core beliefs. Specifically, states are the principal actors in world politics, and Page 59 → no higher authority sits above them. This absence of hierarchy in the state system is commonly called anarchy, which does not mean chaos and violence, but simply that states are sovereign political entities. Furthermore, calculations about power dominate state thinking, and states compete for power among themselves. There is a zero-sum quality to that competition, which sometimes makes it intense and unforgiving. States cooperate with each other for sure, but at root they have conflicting interests, not a harmony of interests. Finally, war is a legitimate instrument of statecraft. To paraphrase Clausewitz, war is a continuation of politics by other means. These common tenets among realist theories notwithstanding, there are significant differences among them. For example, Morgenthau maintains that states are hardwired with an insatiable lust for power, which causes them to seek to maximize their share of world power. Waltz, on the other hand, emphasizes that the structure of the system causes states to compete for power but that states should not strive to maximize power. Instead, they should aim to control an “appropriate” amount of power.5 Furthermore, there are important differences between “defensive realists” (Robert Jervis, Jack Snyder, Stephen Van Evera) and “offensive realists” (myself and Randall Schweller).6 My goal here, however, is not to assess the relative merits of particular realist theories but rather to determine the future relevance of the more general realist paradigm.

The Real World Despite the end of the cold war, the basic structure of the international system remains largely unchanged. States are still the key actors in world politics, and they continue to operate in an anarchic system. It is difficult to find a serious scholar who argues that the United Nations or any other international institution has coercive leverage over the great powers or is likely to have it anytime soon. Moreover, not only is there no plausible replacement for the state on the horizon, but there is little interest anywhere in the world for doing away with the state and putting an alternative political arrangement in its place. Nothing is forever, but there is good reason to think that the sovereign state's time has not yet passed. If the basic structure of the system has not changed since 1990, we should not expect state behavior in the new century to be much different from what it was in past centuries. In fact, there is abundant evidence that states still care deeply about power and will compete for it among themselves in the foreseeable future. Furthermore, the danger still remains that security competition might lead to war, neither of which has gone away with the disappearance of the Soviet Union. To illustrate this point, consider that the United States has fought two wars since the end of the cold war—Iraq (1991) and Kosovo (1999)—and it came dangerously close to going to war against North Korea in 1994. Although the United States now spends more on defense than the next six countries combined, U.S. officials do not seem to think this is enough. Indeed, both candidates in the 2000 presidential campaign advocated spending Page 60 → even more money on the Pentagon. Thus, there is little reason to think that states no longer care about their

security. Furthermore, it is hard to imagine anyone arguing that security competition and war are outmoded in (1) South Asia, where India and Pakistan are bitter enemies, are armed with nuclear weapons, and are caught up in a raging dispute over Kashmir; (2) the Persian Gulf, where Iraq and Iran are bent on acquiring nuclear weapons and show no signs of becoming status quo powers; and (3) Africa, where interstate conflict appears to have increased since the end of the cold war. One might concede, however, that these regions remain mired in the old ways of doing business and argue instead that it is security competition and war among the system's great powers, not minor powers such as Pakistan and Iran, that is passé. Therefore, Europe and Northeast Asia (NEAsia), where there are clusters of great powers, are the places where realist logic no longer has much relevance. But this argument does not stand up to scrutiny. There is a large literature on security in NEAsia after the cold war, and almost every author recognizes that power politics is alive and well in the region and that there is good reason to worry about armed conflict.7 The thought of Japan seriously rearming strikes fear in the heart of virtually every country in Asia, and if China continues to grow economically and militarily over the next few decades, there is likely to be intense security competition between China and its neighbors, as well as the United States. According to one expert, China “may well be the high church of realpolitik in the post-Cold War world.”8 Apparently the Chinese have not gotten the word that realism has been relegated to the scrap heap of history. Furthermore, the United States is in a position today where it might find itself in a war with North Korea or with China over Taiwan. In short, NEAsia is a potentially dangerous place, where security competition is a central element of interstate relations. Possibly the best evidence that power politics is still relevant in NEAsia is that the United States maintains one hundred thousand troops in the region and plans to keep them there for a long time. If NEAsia were a zone of peace, those American forces would be unnecessary and they could be sent home and demobilized, saving the U.S. taxpayer an appreciable sum of money. Instead, they are kept in place to help pacify a potentially volatile region. Joseph Nye, one of the main architects of post-cold war American policy in NEAsia and a scholar with a wellestablished reputation as a liberal international relations theorist (not a realist), made just this point in a 1995 article in Foreign Affairs.9 “It has become fashionable,” he wrote, “to say that the world after the Cold War has moved beyond the age of power politics to the age of geoeconomics. Such clichés reflect narrow analysis. Politics and economics are connected. International economic systems rest upon international political order.” He then made the pacifier argument: “The U.S. presence [in Asia] is a force for stability, reducing the need for arms buildups and deterring the rise of hegemonic forces.” Not only do “forward-deployed forces in Asia ensure broad regional stability,” but they also “contribute to the tremendous political and economic advances made by the Page 61 → nations of the region.” In short, “the United States is the critical variable in the East Asia security equation.” What about Europe, which some writers believe is the best place to look for evidence that power politics is outmoded among the great powers? It was widely believed in the early 1990s that Russia had undergone a fundamental transformation in its thinking about international politics. Its leaders, many believed, understood that the pursuit of power was not likely to enhance Russian security and that the best way to achieve that end was to work with the West to create a peaceful order across all of Europe. Furthermore, some argued that in the wake of the cold war the European Union (EU) would provide the foundation for a stable political order in western Europe, and eventually across the entire continent. It seems clear, however, that things are not working out this way in Europe. NATO, not the EU, provides the basis of stability on the western half of the continent, and NATO is a military institution. Moreover, the expansion of NATO eastward has angered the Russians, who now appear to be thinking and acting like old-fashioned realists. Consider what President Vladimir Putin says in “The National Security Concept of the Russian Federation,” a seminal policy document that he signed on January 10, 2000. “The formation of international relations,” he writes, “is accompanied by competition and also by the aspiration of a number of states to strengthen their influence on global politics, including by creating weapons of mass destruction. Military force and violence remain substantial

aspects of international relations.”10 But as with NEAsia, probably the best evidence that power politics has not disappeared from Europe is that the United States maintains one hundred thousand troops in the region, and it places great importance on keeping NATO intact. If Europe were “primed for peace,” NATO could be disbanded and those American forces could be sent home and demobilized.11 Instead, they are kept in place, because there is potential for intense security competition in Europe, and the United States is determined to keep it from breaking out. Otherwise why would Washington spend tens of billions of dollars every year to maintain a large military presence in Europe? It appears that many Europeans believe that the United States is keeping a lid on security competition in their region. Between 1990 and 1994, Robert Art conducted more than one hundred interviews with European politicalmilitary elites. He found that most believed that “if the Americans removed their security blanket from Europe … the Western European states could well return to the destructive power politics that they had just spent the last forty-five years trying to banish from their part of the continent.”12 Presumably that perspective is even more tightly held today, given that the early 1990s was the heyday of optimism about the prospects for peace in Europe. For a closer look at how many Europeans think about the American presence in Europe, consider the views of Christoph Bertram, a former director of the International Institute for Strategic Studies in London and one of Germany's foremost strategic thinkers. He wrote in 1995, “To disband NATO now would throw Europe into deep insecurity…. It would be a strategic disaster.”13 He goes on to say that “if the United States turned its back on Page 62 → Europe, NATO would collapse and the European Union would be strained to the point of disintegration. Germany would stand out as the dominant power in the West of the continent, and Russia as the disturbing power in the East. The United States would lose much of its international authority as well as the means to help prevent European instability from igniting international conflict once again.” The bottom line is that security competition and war between the great powers have not been burned out of the system. States across the globe are likely to continue competing among themselves for power. And this means that realist theories are likely to have much to say about international politics in the twenty-first century.

The Academy The most serious threat to realism comes not from the real world but from inside universities, where dislike of realism is widespread and often intense, especially among liberal international relations theorists. A good example of this hostility is found in a recent article in the American Political Science Review by John Vasquez, a member of the millennial reflections panel on realism held at the International Studies Association's 2000 convention. After assessing some of the more important realist writings of the past twenty years, he concludes that realism is a degenerating paradigm (as opposed to a progressive paradigm). He is certainly entitled to advance this view, of course, but he goes on to suggest that research by realist scholars does not “deserve continued funding, publication, and so forth.”14 He probably also believes that political science departments should not hire realists. This statement is remarkable for its intolerance. Indeed, it is hard to understand how it got past the editors at the American Political Science Review. Surely if someone had called for running liberal theories or formal modeling out of the scholarly world, the editors would have told the author that ad hominem attacks of that sort have no place in the flagship journal of political science. Apparently standards are different when it comes to realism. This example is hardly unique, however, as any university-based realist knows all too well. Morgenthau, for example, speculated that he probably would not have received tenure at the University of Chicago in 1946 if his first book dealing with realist themes, Scientific Man vs. Power Politics, had been in print when the tenure decision was made.15 In a 1989 book of autobiographical essays by thirty-four international relations scholars, Morgenthau's name “is more frequently cited than any other name.” Yet, the editor notes, “many of the references in these pages are negative in tone. He seems to have inspired his critics even more than his supporters.”16 Kenneth Waltz, the other king of realist scholarship generates similar hostility. For example, Fareed Zakaria notes

that “it was a ritual, at almost every IR seminar I attended as a graduate student at Harvard, for the speaker to spend some time denouncing Waltz's work.”17 Page 63 → Parenthetically, it is worth noting that despite realism's widely acknowledged dominance of the intellectual agenda in international relations, Harvard's government department has not employed a realist theorist since Henry Kissinger left in 1969. Moreover, it made no effort to hire either Morgenthau or Waltz, the two most influential international relations scholars of the past fifty years. Harvard's apparent antipathy toward realism is not unique. Yale's political science department has not had a realist on its distinguished faculty since 1957, when Arnold Wolfers retired. Robert Gilpin, a realist who taught at Princeton until recently, maintains that “liberal intolerance” of realism, which has a long and rich history, has increased recently because realists refuse to “believe that, with the defeat of the Soviet Union and the end of the cold war, the liberal millennium of democracy, unfettered markets, and peace is upon us.”18 There is certainly much evidence of realism bashing in the aftermath of the cold war. To cite a few examples, a 1995 article in International Studies Quarterly not only predicted the imminent demise of realism but concluded that future research should “explore why for the past several decades the discipline of international politics remained mesmerized by a false theory”!19 A prominent British scholar argued in 1994 that “many of the central themes of realism appear as (domesticated) descendants of the militaristic and racist Social Darwinism of the late nineteenth and early twentieth century.”20 Finally, Stanley Hoffmann, a Harvard professor, told the New York Times in March 1993 that realism is “utter nonsense today.”21 There is little reason to think that such hostility toward realism will subside anytime soon. Nevertheless, this intolerance is unlikely to put an end to the realist research agenda, mainly because realism offers too many important insights about international politics to be silenced for long. Try as they may, realism's detractors will find that it is almost impossible to suppress sound arguments in the United States. The American university system is large and decentralized, and serious scholars with controversial ideas can always find a few institutions willing to support them. Realism might be abandoned if it ceased to say anything important about international politics, but as discussed, that is unlikely to be the case. As a result, people who care about the real world will continue to rely on the insights that realist theory provides, and students will continue to be impressed by realism's compelling explanations of state behavior. Realism will disappear only if there is a revolutionary change in the structure of the international system, but that is not likely to happen anytime soon.

Notes 1. William J. Clinton, “American Foreign Policy and the Democratic Ideal,” campaign speech, Pabst Theater, Milwaukee, Wisc., October 1, 1992; “In Clinton's Words: ‘Building Lines of Partnership and Bridges to the Future,’” New York Times, July 10, 1997. 2. John Mueller, Retreat from Doomsday: The Obsolescence of Major War (New York: Basic Books, 1989), 4.Page 64 → 3. Francis Fukuyama, “The End of History?” The National Interest, no. 16 (1989): 3–18. The quotations in this paragraph are from pp. 5 and 18. Also see Francis Fukuyama, The End of History and The Last Man (New York: Free Press, 1992). 4. For evidence that realism has always been the dominant paradigm in the discipline of international relations, see Brian C. Schmidt, The Political Discourse of Anarchy: A Disciplinary History of International

Relations (Albany: State University of New York Press, 1998). 5. Kenneth N. Waltz, “The Origins of War in Neorealist Theory,” in The Origin and Prevention of Major Wars, ed. Robert I. Rotberg and Theodore K. Rabb (New York: Cambridge University Press, 1989), 40. 6. See Sean M. Lynn-Jones and Steven E. Miller, preface to The Perils of Anarchy: Contemporary Realism and International Security, ed. Michael E. Brown, Sean M. Lynn-Jones, and Steven E. Miller (Cambridge: MIT Press, 1995), ix–xii; and Benjamin Frankel, “Restating the Realist Case: An Introduction,” Security Studies (special issue: “Realism: Restatements and Renewal”) 5, no. 3 (1996): xiv–xx. 7. See, for example, the many articles on Asian security published over the past decade in Foreign Affairs, International Security, and Survival. Some of the best pieces from International Security are published in Michael E. Brown, Sean M. Lynn-Jones, and Steven E. Miller, eds., East Asian Security (Cambridge: MIT Press, 1996). 8. Thomas J. Christensen, “Chinese Realpolitik,” Foreign Affairs 75, no. 5 (1996): 37. 9. Joseph S. Nye Jr., “East Asian Security: The Case for Deep Engagement,” Foreign Affairs 74, no. 4 (1995): 90–102. The quotations in this paragraph are from pages 90–91, 102. 10. The document was originally published in Nezavisimoye Voennoye Obozreniye on January 14, 2000. For key excerpts, from which this quote is taken, see “Russia's National Security Concept,” Arms Control Today 30, no. 1 (2000): 15–20. 11. The phrase “primed for peace” was coined by Stephen Van Evera to describe post-cold war Europe. See Stephen Van Evera, “Primed for Peace: Europe after the Cold War,” International Security 15, no. 3 (1990–91): 7–57. 12. Robert J. Art, “Why Western Europe Needs the United States and NATO,” Political Science Quarterly 111, no. 1 (1996): 5–6. 13. Christoph Bertram, Europe in the Balance: Securing the Peace Won in the Cold War (Washington, D.C.: Carnegie Endowment, 1995), 17–18, 85. Also see pp. 10–11. 14. John A. Vasquez, “The Realist Paradigm and Degenerative versus Progressive Research Programs: An Appraisal of Neotraditional Research on Waltz's Balancing Proposition,” American Political Science Review 91, no. 4 (1997): 900. 15. See “Bernard Johnson's Interview with Hans J. Morgenthau” in Truth and Tragedy: A Tribute to Hans J. Morgenthau, ed. Kenneth Thompson and Robert J. Myers (New Brunswick, N.J.: Transaction Books, 1984), 371. 16. Joseph Kruzel, “Reflections on the Journeys,” in Journeys through World Politics: Autobiographical Reflections of Thirty-four Academic Travelers, ed. Joseph Kruzel and James N. Rosenau (Lexington, Mass.: Lexington Books, 1989), 505. 17. Personal correspondence, October 7, 1999. 18. Robert G. Gilpin, “No One Loves a Political Realist,” Security Studies (special issue: “Realism: Restatements and Renewal”) 5, no. 3 (1996): 3. 19. Bahman Fozouni, “Confutation of Political Realism,” International Studies Quarterly 39, no. 4 (1995): 508. 20. Fred Halliday, Rethinking International Relations (Vancouver: University of British Columbia Press, 1994), 11. 21. Quoted in Thomas L. Friedman, “Friends Like Russia Make Diplomacy a Mess,” New York Times, March 28, 1993, sect. 4, p. 5.

Page 65 →

MODERN REALIST THEORY AND THE STUDY OF INTERNATIONAL POLITICS IN THE TWENTY-FIRST CENTURY Joseph M. Grieco Participants in the millennial reflections panels held under the auspices of the 2000 annual meeting of the International Studies Association were invited to offer self-critical reflections about the state of different elements of the field of international relations and to put forward recommendations on how these different elements might push forward most productively in the years ahead. Toward that end, I present ideas about some of the main contributions of modern realist theory to the study of world politics, a few of its main problems as an approach to international studies, and a sample of its most promising areas of future development.

Modern Realism's Contributions to International Relations Employing terms suggested by Quincy Wright, the field of international relations is centrally concerned with identifying and understanding the causes of war and the conditions of peace.1 Modern realist theory—the work of such scholars as Hans Morgenthau, E. H. Carr, Nicholas Spykman, Raymond Aron, Robert Gilpin, and Kenneth Waltz—has surely contributed vitally to international relations insofar as it casts important light on both the problem of war and the problem of peace among nations.2 It does so by providing us with three fundamental insights into world politics. 1. Modern realism helps us appreciate the impact of the anarchical structure of the international system on the preferences, strategies, interactions, and domestic institutions of states. Modern realist theory provides the strongest and clearest analysis of the peculiar (but not unique) setting in which states interact: the absence of a centralized authority reliably capable of providing protection and the redress of grievances. This political context, in turn, through processes of selection and socialization, has a constraining effect on the strategies that states are likely to pursue and even, at the limit, the preferences that they seek to satisfy in the international domain. In particular, modern realism provides us with the tools to appreciate that states that are co-located in a self-help system will be concerned about their Page 66 → security and their independence and, by consequence, the means by which they can retain some minimally acceptable level of both, or at least mitigate the harm that may be done to one or the other. In addition, modern realism helps us to apprehend the systemic sources of constraints on the amenability of national governments to work together even when they share common interests. Finally, realism sheds light on the observable behavior of many major states to resist rather than to acquiesce in the face of overwhelming power. Other systems-level theories of international relations—for example, neoliberal institutionalism and social constructivism—largely accept this modern realist depiction of world politics as the starting point for their own analyses.3 These approaches put themselves forward as fundamental alternatives to realist theory. However, in the main they do no more than suggest that not all of the inferences put forward by modern realism about world politics necessarily obtain at present, or they need not do so for all time. These alternative approaches are essentially “yes, but” commentaries on the realist understanding of world affairs.4 This is not the case, it should be noted, with domestic structuralist and individual-level theories that focus either on the cognitive dynamics of individual leaders or their pragmatic interest in achieving and retaining political office. Such approaches do indeed provide a very different understanding of world affairs from that put forward by realism.5 However, one of the key points made by realist theories and by other literatures whose main findings are highly consistent with realist expectations is that international power conditions may in fact influence the shape and operation of the domestic institutions of nations. In that sense, it might be suggested that the forces emphasized by realism—and in particular, the problem of security in an anarchical international environment—might have an important direct or indirect role in generating the domestic institutions that then have a subsequent impact on the foreign behavior of nations.6

2. Modern realism helps us appreciate the impact of inequalities on international affairs, and in particular inequalities in power. While modern realist theory emphasizes the similarity in structural circumstances in which states find themselves (the absence of a centralized world authority), it emphasizes as well that the degree to which different states will be able to ensure their individual security and independence in the risk-laden context of world politics, and the means by which they will pursue those and other ends, will differ greatly as a function of their power. International life is profoundly unfair: the range of choice that is available to the United States in international matters and even domestic matters is greater than it is for France. It may not be going too far to suggest that no other theory of international relations, whether it focuses on systemic or domestic-structural or individual-level factors, pays as much serious attention to the problem of power and inequality in power in the international domain as does realist international theory. Indeed, most alternatives to realist theory—most notably, neoliberal institutionalism—deal with the problem of power simply by defining it away or by ignoring its role in human and group interactions. Page 67 → It is, indeed, sometimes suggested that power is coming to play a smaller role in world politics. The behavior of two countries, Germany and Japan, is often cited as evidence for this diminished salience of power in the international system. Germany allows its smaller European neighbors to have disproportionate power in the framework of European Union (EU) institutions, and this willingness to so eschew German relative influence is “puzzling” unless one understands that the German public and especially its political class have become international in their “acquired collective identities.”7 It is surely true that the Germany of today and the foreseeable future is different from the Germany of 1914 or 1939. However, if Germany has become uninterested in possessing formal influence commensurate with its great and growing relative capabilities in Europe, how do we account for the facts that the German government sought unsuccessfully, as matters developed at the EU summit at Nice in December 2000, to increase Germany's votes in the Council of Ministers beyond that possessed by France, and did succeed, by virtue of pressing for changes in EU voting rules to include national population thresholds, to bring the EU to a new set of circumstances wherein Germany's relatively large population has now been translated into a level of formal German weighting in EU voting that is greater than that of France or Britain or Italy?8 In the case of Japan, there is no question but that that country pursued its interests during the second half of the twentieth century in a more peaceful and cooperative manner than it did during the first half of that century, and that there is an antimilitarist culture strongly in place in that country as well as Germany, as Berger has demonstrated.9 However, as Berger has also noted, Japan's cooperative, antimilitarist orientation in world affairs rests not just on important domestic institutional, political, and indeed cultural dynamics, but also on a very important international-political condition, namely, that the United States largely supplies Japan with its external security. As Berger also has noted, notwithstanding all the domestic changes that have occurred in Japan since World War II, there are circumstances in which many of the cooperative/antimilitarist assumptions and values and practices we have witnessed in postwar Japanese foreign policy would be severely tested, namely, those in which the U.S.-Japanese alliance came to an end and a serious new security threat came into being. Thus, the abjuring of traditional forms of power on the part of Japan, that most civilian of modern states, seems to rest to a large degree on the presence of the very traditional forms of power put forward in Asia by America. 3. Modern realism helps us appreciate the importance of continuity in international affairs, but it also alerts us to the pervasiveness of change. Modern realism also helps us understand both continuity and change in world affairs. In particular, it helps us understand why the interstate system remains anarchical and why no single world dictator or world government has emerged (the process of balancing); it helps us understand why cooperation among nations is hard but not impossible to achieve (the problem of relative gains and the problem of dependency, and the solution of hegemonic leadership); and it helps us understand why states that are culturally and Page 68 → economically very different nevertheless converge in certain respects in their military and diplomatic infrastructure, and even in their domestic institutions (the mechanics of selection and socialization).10 At the same time, modern realism pinpoints two sources of profound change in world affairs, namely, shifts in the distribution of capabilities across states and reactions by national actors to structural conditions and dynamics.11

Modern Realism's Problems Modern realism provides us with powerful insights into world affairs, and it is the starting point for all other important theories of international relations. However, as with every other approach, realism is flawed and in need of improvement. I note five problems or challenges before it, and how modern realism might address each. Realism and the Positing of Empirical Expectations Realists and their critics seem to concur that realism yields the view that states will invariably balance against a rising power rather than bandwagon with it, that states will invariably fail to cooperate rather than succeed in doing so if relative gains problems are in operation, and that states will always resist rather than embrace specialization in the face of the risk of dependency. These are useful and interesting hypotheses, but I wonder if both realist and nonrealist scholarship would benefit from two adjustments of these hypotheses. The first and very modest adjustment would involve greater acknowledgment that, however important realistposited variables might be, there are other factors that under certain circumstances might vitiate their impact. Hence, rather than suggest that states always balance rather than bandwagon, we might suggest that, other things being equal, states facing a rising power will be more likely to balance against rather than to bandwagon toward that state. This approach requires modesty in what we think realism (or any theory) can accomplish at present: even if international structure is quite important, we would need to acknowledge that the best we can do is to put forward probabilistic statements about its impact on real-world decisions and actions of states. Second, rather than looking only at expected state policy as the basis for generating and testing realist and nonrealist hypotheses, we may wish to expand our range of output variables to include the consequences of state behavior. Peter Feaver has suggested that such a focus on consequences and not just state actions might be very helpful in estimating the veracity of realist claims.12 His premise is that realism is both a deductive argument (with a cluster of resulting explicit empirical observable expectations) about how the world operates and a normative claim that states ought to adjust their policies according to systemic conditions or risk disaster. That realists seek to generate policy proposals on the basis of their research is evidence, Feaver suggests, that realism recognizes that some states will fail to act in Page 69 → accord with their true, “real” interests. Along these lines, Feaver also suggests, Waltz's core insight is not that states will always act in accord with systemic requirements, but that states that engage in “system-insensitive” behavior are more likely to suffer than are states that engage in “system-responsive” behavior. This general insight, Feaver then argues, can be applied to the specific issue of balancing: what may be termed the Feaver adverse consequences hypothesis about balancing is that states that fail to balance when systemic conditions call for balancing are more likely to suffer harm than those states that do elect to balance when circumstances call for that strategy. A problem that would still be in operation with the Feaver hypothesis is the specification of the precise systemic conditions that would lead realists to predict balancing rather than other strategies, such as bandwagoning or hiding. Assuming that this challenge could be overcome, the key question then would be, in general, and controlling for a range of other conditions that might affect success and failure by a state in the international arena, the following: do the states that fail to balance experience a higher probability of suffering adverse consequences at the hands of a rising power than do the states that elect to balance against that rising power? This path of research suggested by Feaver would not just be relevant to the question of realist arguments about balancing but would be useful in assessments of the whole range of realist arguments about world politics so long as such arguments suggest that states that are system-insensitive will pay a cost for such insensitivity. Realism and State Preferences for Power A second theoretical problem for realism can be phrased as a question: does the co-location of states in an anarchical political environment typically lead “normal” states to be security or power maximizers, and is there an observable difference between the two goals?13 The present writer, drawing upon Waltz,14 has suggested that states, from a realist viewpoint, are “defensive positionalists”—that is, they do not so much wish to dominate as to prevent their becoming dominated by others—and as such they will typically prefer in their foreign relations to

attain that level of capabilities and retain that margin of autonomy that are needed to maintain their relative power position and thereby prevent their becoming vulnerable to pressures or threats from other nations.15 However, an alternative view has been put forward with great vigor from within the general realist perspective: Schweller has argued that at least some states want to change their status in the system and therefore will want to advance their relative power position.16 Further, Mearsheimer has suggested that, as a general rule, “states in the international system aim to maximize their relative power position over other states.”17 The question of whether states seek to maximize power for its own sake or attempt to maximize security and therefore are power-maximizers is important for realism for at least two reasons. First, if realists do think that states value power even more than security, they would need to expect Page 70 → states to bandwagon rather than balance in those instances in which the former yielded greater power than the latter. This is a point that is made with great effect by Schweller in his analysis of revisionist states.18 Second, an assumption that states value power above all else would require realists to change their current specification of the problem of international cooperation. That is, they would need to argue that states seek not to avoid gaps in gains favoring partners but instead to maximize gaps in their favor. A strong desire to maximize gaps in gains would make cheating much more attractive to states, and therefore the fear of cheating would be much greater. States would also have a much more aggressive set of interests regarding the distribution of gains from cooperation. Realists, in specifying such a world, would be driven to argue that cooperation among states is essentially impossible to achieve. Yet this would cause them to face the unbearable burden of explaining why there are in fact substantial amounts of cooperation among states. Perhaps one way realist theory may be able to avoid the power-security maximization trap is to restrict, in the way suggested in general by Waltz, its expectations about the impact of anarchy on the interests of states regarding relative power. The argument would be that realism's focus on anarchy cannot readily lead it to expect more than that states are concerned primarily about their survival and security and that they seek to ensure both not by maximizing power to their advantage but by minimizing gaps in power that are likely to favor rivals or adversaries. This restricted understanding of the effects of anarchy on state preferences regarding power may itself be rather far-reaching, for it yields such important expectations as those regarding the tendency of states to choose balancing over bandwagoning, the tendency of states to fear functional differentiation, and the likely prevalence of the relative gains problem for cooperation. Of course, there are states that seek to maximize their power, and some do so to enhance their security. Yet realists may wish to acknowledge that a systemic-based theory may not be able to account for these instances of state behavior by itself. Instead, they may wish to argue that the factors leading states to so define their security problem in terms of the need for a favorable imbalance of power may be driven by such nonsystemic factors as military technology or the perceptions of national decision makers, although they could still argue that international anarchy exacerbates tensions or “tightens” the security dilemma.19 At the same time, realists would be able to argue that their expectations and analyses become highly relevant once a power-maximizing state comes into being and begins to operate in the international system—for example, one would expect to see balancing against that highly assertive state. Realism and International Institutions There is very little concrete or systematic evidence, as Mearsheimer has emphasized,20 that international arrangements have strong independent effects on the preferences, policies, or interactions of states. Indeed, in contrast Page 71 → to liberal institutionalist expectations, a recent statistical analysis by Mansfield and Bronson of the political sources of trade provides evidence that mutual membership in the General Agreement on Tariffs and Trade either has no significant effect on bilateral trade between countries or may actually be a negative predictor of such trade.21 In the same vein, a recent statistical analysis by Bennett and Stam suggests that mutual membership in international institutions does not appear to have a consistently depressive effect on the likelihood of militarized conflicts between countries, and it may be possible that mutual organizational membership is associated with an increase, not a decrease, in the probability of such conflicts; as they conclude, “We thus find some support for realism's contention that IGOs [intergovernmental organizations] will have little effect on the use

of force.”22 Yet even if the liberal institutional perspective is incorrect and international institutions do not have a systematic impact on state preferences and behavior, realism may still want to study international institutions. That is because states often seem to have a preference to construct and work through international institutions. Priess and Schweller have gone some way in this direction, as has Ikenberry.23 However, there is still much to be done by realist-informed scholars about the question of why many states appear to have a preference for international institutions. One possible line of analysis in this respect might be to suggest that relatively weaker states may choose to cooperate through an institution in order to attain “voice opportunities”24 with regard to their stronger partners.25 But that still leaves us with the stronger partners: why do they accept and even embrace institutionalization? One possible realist hypothesis is that such institutionalization allows the stronger state to exercise its hegemony in an efficient manner and under a cover of legal equality such that there is less resistance to its preferences and proposals. So, for example, the EU might be a vehicle by which Germany is able to employ its power discretely and legitimately and thereby dominate its neighbors without arousing substantial resistance on their part or even very much resentment. Of course, cutting against this function of international institutions as a basis for formal equality of power but real inequality of influence is the tendency for states often to wish at some point to have institutional arrangements operate and indeed be structured in such a way that there is no hiding of real power conditions in the arrangement: see, for example, the earlier discussion of Germany and its new preferences for voting rules in the EU. Realism and Interaction Effects Some of the very best realist-informed research done in recent years—for example, Snyder's work on state overextension, Walt's work on revolutionary states and the international system, and Schweller's work on revisionist states—seeks to integrate domestic structural and realist-posited systemic-level variables in a coherent argument about different elements of world politics.26 Much more can be done to bring about such an integration of realist-posited and domestic structural variables. Page 72 → For example, a growing body of research suggests that regime type interacts in important ways with systemic conditions and dynamics. The best developed of this literature concerns regime type and international conflict management—that is, the democratic peace literature. But there are other interesting areas of interaction between domestic regime type and international politics. For example, studies by Bruce Bueno de Mesquita and David Lalman and by David Rousseau, Christopher Gelpi, Dan Reiter, and Paul Huth have suggested that democracies are disproportionately likely to be the target of militarized challenges in world politics.27 This is an odd finding, since studies by Lake and by Stam and Reiter suggest that democracies are disproportionately likely to win the wars they fight, and Ajin Choi is completing a thesis at Duke that identifies as a key reason for this war-fighting superiority of democracies their superior capacity to work together as war-coalition partners.28 Kurt Gaubatz has found that democracies form more durable alliances, which would also seem to be a deterrent to challenges being launched against them.29 Why, then, do democracies appear disproportionately to attract trouble if they are such effective fighters? It would be very interesting to learn if realist-informed arguments could help address that question. Moreover, there is the question of whether realist theory can contribute to the new literature on how the preferences of leaders to remain in office interact with domestic institutional circumstances to affect leadership policy choices and even leadership performance in the political-military domain. Recent years have seen several interesting studies that focus on the impact on foreign policy that is induced by the interest of leaders to retain office, importantly mediated by the character of the domestic institutional setting in which such leaders find themselves, and it would be quite interesting to learn whether and in what manner these interactions might be influenced by differences in international political or military circumstances.30

In addition, and following the pathbreaking work on Chinese foreign policy by Johnston,31 there might be important and interesting interaction effects in operation between realist-posited variables and those emphasized by the social constructivists. The latter suggest that there may exist at any given time a body of beliefs commonly shared by national decision makers within or perhaps even across different countries, and these beliefs may act as an independent force on the positing by national leaders of state preferences and the selection by them of national strategies to satisfy those preferences. One potentially interesting line of research might inquire as to whether such structures of commonly held beliefs vary with recent and current international conditions. This line of inquiry might, in practical terms, seek to explore the effects of World War II, for example, on the beliefs of major-state leaders regarding the best paths to security. Realism and the Democratic Peace We often think that the democratic peace thesis raises serious problems for modern realist theory, as might the related vein of work constituting a “democratic Page 73 → efficacy” thesis, which suggests that democracies are disproportionately likely to prevail in international crises and wars.32 However, drawing upon the realist work discussed previously regarding the possible generative effects of the international system for domestic structures, as well as the work by Gilpin on hegemony and international economic cooperation,33 future research might investigate the possibility that the democratic peace is endogenous to particular international structures—specifically, the presence of a democratic hegemon, Britain in the nineteenth century and America in the twentieth—for these hegemons may have played a key role in the emergence and sustained viability of many democracies and thus of the democratic peace. This line of research would not in any way undermine the democratic peace finding or the different arguments about the manner in which democracies might in fact be better able to resolve disputes among themselves short of the use of military violence. However, if this line of research proved to be fruitful, it might yield a richer appreciation of both modern realism and domestic structural approaches to international politics.

Modern Realism's Future Realism will remain the core approach in the field of international relations. So long as the international system has certain key attributes—the presence of a number of powerful states that can do serious harm to one another and the absence of a reliable centralized source of authority that can settle conflicts among those states—then realist theory, based as it is on trying to understand the impact of those attributes, will have something serious to say about world politics. However, realism is likely to evolve in its own understanding of world politics and to have some interesting conversations with other perspectives. In my view, the Feaver emphasis on the adverse consequences of systeminsensitive state behavior offers the most promising pathway for realism's own development. In addition, my sense is that the two strongest candidates for useful dialogue with modern realism in the years ahead are social constructivism and, especially, as has always been true, domestic structuralism. Through such internal development and external dialogue with other approaches, modern realism will continue to provide the surest starting point for our understanding of the problem of war and peace among nations.

Notes 1. Quincy Wright, The Causes of War and the Conditions of Peace (London: Longmans, Green, 1935). 2. Hans Morgenthau, Politics among Nations: The Struggle for Power and Peace, 4th ed. (New York: Knopf, 1973, first published 1948); Edward H. Carr, The Twenty Years Crisis, 1919–1939: An Introduction to the Study of International Relations (London: Macmillan, 1939); Nicholas J. Spykman, America's Strategy in World Politics: The United States and the Balance of Power (New York: Harcourt, Brace, 1941); Raymond Aron, Peace and War: A Theory of International Relations, trans. Richard Howard and Annette Baker Fox (Garden City, N.Y.: Doubleday, 1966); Robert Gilpin, “The Politics of Transnational

Relations,” Page 74 → in Transnational Relations and World Politics, ed. Robert O. Keohane and Joseph S. Nye Jr. (Cambridge: Harvard University Press, 1972), 48–69; Robert Gilpin, U.S. Power and the Multinational Corporation: The Political Economy of Foreign Direct Investment (New York: Basic Books, 1975); Robert Gilpin, War and Change in World Politics (Cambridge: Cambridge University Press, 1981); Robert Gilpin, The Political Economy of International Relations (Princeton, N.J.: Princeton University Press, 1987); Kenneth N. Waltz, Man, the State, and War: A Theoretical Analysis (New York: Columbia University Press, 1959); Kenneth N. Waltz, Theory of International Politics (Reading, Mass.: AddisonWesley, 1979); Kenneth N. Waltz, “Structural Realism after the Cold War,” International Security 25 (summer 2000): 41. 3. On the development of neoliberal institutionalism, see Arthur A. Stein, “Coordination and Collaboration: Regimes in an Anarchic World,” in International Regimes, ed. Stephen D. Krasner (Ithaca, N.Y.: Cornell University Press, 1983), 115–40; Robert M. Axelrod, The Evolution of Cooperation (New York: Basic Books, 1984); Robert O. Keohane, “The Demand for International Regimes,” in Krasner, International Regimes, 141–71; Robert O. Keohane, After Hegemony: Cooperation and Discord in the World Political Economy (Princeton, N.J.: Princeton University Press, 1984); and David A. Baldwin, ed., Neorealism and Neoliberalism: The Contemporary Debate (New York: Columbia University Press, 1993). In a recent essay, Legro and Moravscik put forward the thesis that neorealism has converged with liberalism; actually, it may be more accurate to conclude that, in very broad terms, the opposite has occurred. Key constructivist works would include John G. Ruggie, “International Regimes, Transactions, and Change: Embedded Liberalism in the Postwar Economic Order,” in Krasner, International Regimes, 195–231; Friedrich Kratochwil and John G. Ruggie, “International Organization: A State of the Art on an Art of the State,” International Organization 40 (autumn 1986): 753–75; Jeffrey W. Legro, Cooperation under Fire: Anglo-German Restraint during World War II (Ithaca, N.Y.: Cornell University Press, 1995); Jeffrey W. Legro, “Whence American Internationalism,” International Organization 54 (spring 2000): 253–89; Peter J. Katzenstein, ed., The Culture of National Security: Norms and Identity in World Politics (New York: Columbia University Press, 1996); Martha Finne-more, National Interests in International Society (Ithaca, N.Y.: Cornell University Press, 1996); and Margaret Keck and Kathryn Sikkink, Activists beyond Borders: Advocacy Networks in International Politics (Ithaca, N.Y.: Cornell University Press, 1998). Constructivism has largely focused to date on how commonly held norms and beliefs promote peace and cooperation among nations. However, there has also been work by Johnston on how commonly held beliefs at the national level (in particular, China) might propel a state to be less rather than more amenable to international cooperation. Alastair Iain Johnston, Cultural Realism: Strategic Culture and Grand Strategy in Ming China (Princeton, N.J.: Princeton University Press, 1995); Alastair Iain Johnston, “Cultural Realism and Strategy in Maoist China,” in Katzenstein, The Culture of National Security ; Alastair Iain Johnston, “Realism(s) and Chinese Security Policy in the Post-Cold War Period,” in Unipolar Politics: Realism and State Strategies after the Cold War, ed. Ethan B. Kapstein and Michael Mastanduno (New York: Columbia University Press, 1999), 261–318. 4. One of the helpful characteristics of the new literature (for example, John Oneal and Bruce Russett, “Assessing the Liberal Peace with Alternative Specifications: Trade Still Reduces Conflict,” Journal of Peace Research 36 [July 1999], 423–42) on the potentially dampening effects of economic interdependence on the probability of military conflict between nations is that it does not seek to claim that any such effects obviate the importance of realist-emphasized variables, such as military power and alliances. Of course, it is not clear that economic interdependence does in fact generally reduce the risk of militarized conflicts: for the contrary view, see for example Waltz, Theory of International Politics, and Katherine Barbieri, “International Trade and Conflict: The Debatable Relationship” (paper presented at the Thirty-ninth Annual Convention of Page 75 → the International Studies Association, Minneapolis, March 17–21, 1998). It is possible that the effects of economic interdependence vary in response to other conditions, such as the relevant states' expectations about future market access or whether they are both democracies (in which case interdependence is likely to mitigate the likelihood of conflict) or are both nondemocracies or constitute a mixed pair (in which cases interdependence may exacerbate the risk of conflict). On the interaction between regime type and interdependence as it relates to the problem of war, see Paul A. Papayoanou, “Interdependence, Institutions, and the Balance of Power,” International Security 20 (spring 1996): 42–76; and Christopher Gelpi and Joseph M. Grieco, “Democracy, Interdependence, and the Liberal Peace” (Duke

University, March 31, 2000, unpublished). Finally, it should be noted that one of the more interesting developments in the recent literature on interdependence is its focus not just on trade integration but interdependence via foreign direct investments, which actually was of greater interest to one of the key earlier writers on interdependence, Norman Angell. See Richard Rosecrance, The Rise of the Virtual State: Wealth and Power in the Coming Century (New York: Basic Books, 1999); Norman Angell, The Great Illusion: A Study of the Relation of Military Power to National Advantage, 4th ed. (New York: Putnam, first published 1913). 5. Key examples of individual-level works that focus on cognitive or other psychological characteristics of leaders include Robert Jervis, Perception and Misperception in International Politics (Princeton, N.J.: Princeton University Press, 1976); Irving L. Janis, Groupthink, 2d ed. (Boston: Houghton Mifflin, 1980); Richard N. Lebow, Between Peace and War: The Nature of International Crises (Baltimore: Johns Hopkins University Press, 1981); and Robert Jervis, Richard Ned Lebow, and Janice Gross Stein, Psychology and Deterrence (Baltimore: Johns Hopkins University Press, 1985). An interesting discussion of socialpsychological dynamics that would seek to reinforce realist-specified dynamics in international relations is put forward by Mercer. Jonathan Mercer, “Anarchy and Identity,” International Organization 49 (spring 1995): 229–52. Other key works on the impact of different levels of domestic politics on foreign policy that emphasize the manner in which domestic dynamics may propel states to go in directions not anticipated by a focus solely on their systemic position would include Jack Snyder, Myths of Empire: Domestic Politics and International Ambition (Ithaca, N.Y.: Cornell University Press, 1991); and Charles Kupchan, The Vulnerability of Empire (Ithaca, N.Y.: Cornell University Press, 1994). 6. A good starting point for reading in this area is Otto Hintze, “The Formation of States and Constitutional Development” and “Military Organization and the Organization of the State,” both in The Historical Essays of Otto Hintze, ed. Felix Gilbert (Oxford: Oxford University Press, 1975), 157–215; Charles Tilly, Coercion, Capital, and European States, a.d. 900–1990 (Oxford: Blackwell, 1990); Theda Skocpol, “A Critical Review of Barrington Moore's Social Origins of Dictatorship and Democracy,” American Journal of Sociology 4 (fall 1973): 1–34; Aristide Zolberg, “Strategic Interactions and the Formation of Modern States: France and England,” International Social Science Journal 32, no. 4 (1980): 687–716; Barry Posen, “Nationalism, the Mass Army, and Military Power,” International Security 18 (autumn 1993): 80–124. 7. Peter J. Katzenstein, “United Germany in an Integrating Europe,” in Tamed Power: Germany in Europe, ed. Peter J. Katzenstein (Ithaca, N.Y.: Cornell University Press, 1997), 3. 8. On the difficult negotiations at Nice and the question of German efforts to increase its formal power in EU decision making, see “EU Strikes Reform Deal after Marathon,” BBC Online, December 11, 2000, . 9. Thomas V. Berger, “Norms, Identity, and National Security in Germany and Japan,” in The Culture of National Security: Norms and Identity in World Politics, ed. Peter J. Katzenstein (New York: Columbia University Press, 1996), 312–56. 10. The question of whether states balance, and if so if they balance against a state rising in material capabilities (Waltz, Theory of International Politics) or one Page 76 → that is becoming more threatening, which may be due to the possession of greater material resources of that state but also because it presents a threatening ideology or some other nonmaterial threat (Stephen Walt, The Origins of Alliances [Ithaca, N.Y.: Cornell University Press, 1987]), has been very controversial for neorealist theory. For a very useful discussion of the question of balancing in realist theory, see John A. Vasquez, “The Realist Paradigm and Degenerative versus Progressive Research Programs: An Appraisal of Neotraditional Research on Waltz's Balancing Proposition,” American Political Science Review 91 (December 1997): 899–912. The relative gains debate is pursued in, for example, Joseph M. Grieco, Cooperation among Nations: Europe, America, and Non-Tariff Barriers to Trade (Ithaca, N.Y.: Cornell University Press, 1990); a number of essays in Baldwin, Neorealism and Neoliberalism; and Emerson M. S. Niou and Peter C. Ordeshook, “Less Filling, Tastes Great: The Realist-Neoliberal Debate,” World Politics 46 (January 1994): 209–34. For recent realistinspired work on the question of systemically induced socialization of states, see Posen, “Nationalism, the Mass Army, and Military Power”; and Stephen Walt, Revolution and War (Ithaca, N.Y.: Cornell University Press, 1996). 11. A highly developed realist-inspired argument about international structural change and its consequences for state behavior concerns the role of hegemonic leadership for the international economy: on this theme

see Gilpin, U.S. Power and the Multinational Corporation; Stephen D. Krasner, “State Power and the Structure of International Trade,” World Politics 28 (April 1976): 317–47; and David A. Lake, Power, Protection, and Free Trade: International Sources of U.S. Commercial Strategy, 1887–1939 (Ithaca, N.Y.: Cornell University Press, 1987). A related vein of research emphasizes that a country's trading patterns are likely to be highly influenced by its security relations with different partners: on this thesis see Brian M. Pollins, “Does Trade Still Follow the Flag?” American Political Science Review 83 (June 1989): 465–80; Joanne S. Gowa, Allies, Adversaries, and International Trade (Princeton, N.J.: Princeton University Press, 1994); and Edward D. Mansfield and Rachel Bronson, “Alliances, Preferential Trading Arrangements, and International Trade,” American Political Science Review 91 (March 1997): 94–107. 12. See Peter D. Feaver, “Brother, Can You Spare a Paradigm? (Or Was Anyone Ever a Realist?),” correspondence section of International Security 25 (summer 2000): 165–69. 13. A superior discussion of this problem is presented in Arnold Wolfers, Discord and Collaboration: Essays on International Politics (Baltimore: Johns Hopkins University Press, 1962), 81–102. 14. Waltz, Theory of International Politics. 15. Grieco, Cooperation among Nations. 16. Wolfers, Discord and Collaboration, especially 85–88. 17. John J. Mearsheimer, “The False Promise of International Institutions,” International Security 19 (winter 1995–96): 11; also see Fareed Zakaria, “Realism and Domestic Politics: A Review Essay,” International Security 17 (summer 1992): 177–98. 18. Randall Schweller, “Bandwagoning for Profit: Bringing the Revisionist State Back,” International Security 19 (summer 1994): 72–107. 19. This is the path taken by Jack Snyder, Myths of Empire. 20. Mearsheimer, “The False Promise of International Institutions.” 21. Mansfield and Bronson, “Alliances, Preferential Trading Arrangements, and International Trade,” 100, 103. 22. D. Scott Bennett and Allan C. Stam, “Research Design and Estimator Choices in the Analysis of Interstate Dyads: When Decisions Matter,” Journal of Conflict Resolution 44 (October 2000): 667. For alternative results on the question of international institutions and the dampening of the risk of militarized conflicts, see Russett and Oneal, “Assessing the Liberal Peace with Alternative Specifications,” to which Bennett and Stam responded in their 2000 essay. It should be noted that Bennett and Stam emphasize that their results may differ from those reported by Russett and Oneal because while the former's data set Page 77 → ends in 1984, the latter's does so in 1992. For a quantitative analysis that does seem to support the proposition that international institutions facilitate cooperation (in the particular instance, the imposition of economic sanctions), see Lisa Martin, Coercive Cooperation: Explaining Multilateral Economic Sanctions (Princeton, N.J.: Princeton University Press, 1992). For a recent general argument that international institutions matter, see Lisa Martin and Beth Simmons, “Theories and Empirical Studies of International Institutions,” International Organization 52 (autumn 1998): 729–57. 23. Randall L. Schweller and David Priess, “A Tale of Two Realisms: Explaining the Institutions Debate,” Mershon International Studies Review, 41 (May 1992): 1–32; G. John Ikenberry, After Victory: Institutions, Strategic Restraint, and the Rebuilding of Order after Major Wars (Princeton, N.J.: Princeton University Press, 2000). 24. Joseph M. Grieco, “The Maastricht Treaty, Economic and Monetary Union, and the Neorealist Research Programme,” Review of International Studies 21 (January 1995): 21–40; Joseph M. Grieco, “State Interests and International Rule Trajectories: A Neorealist Interpretation of the Maastricht Treaty and European Economic and Monetary Union,” Security Studies 5 (spring 1996): 176–222. 25. This line of discussion follows the suggestions offered by Morgenthau regarding the European Coal and Steel Community in the second edition of Politics among Nations in 1958, and, beginning with the third edition in 1966, the European Economic Community; see, for example, Morgenthau, Politics among Nations, 3d ed., 531–34. 26. Snyder, Myths of Empire; Walt, Revolution and War; Randall Schweller, Deadly Imbalances: Tripolarity and Hitler's Strategy of World Conquest (New York: Columbia University Press, 1998). 27. Bruce Bueno de Mesquita and David Lalman, War and Reason: Domestic and International Imperatives (New Haven, Conn.: Yale University Press, 1992); David L. Rousseau, Christopher Gelpi, Dan Reiter, and

Paul K. Huth, “Assessing the Dyadic Nature of the Democratic Peace, 1918–1988,” American Political Science Review 90 (September 1996): 512–33. 28. David A. Lake, “Powerful Pacifists: Democratic States and War,” American Political Science Review 86 (March 1992): 24–37; Dan Reiter and Allan C. Stam III, “Democracy, War Initiation, and Victory,” American Political Science Review 92 (June 1998): 377–90; Dan Reiter and Allan C. Stam III, “Democracy and Battlefield Military Effectiveness,” Journal of Conflict Resolution 42 (June 1998): 259–77; Ajin Choi, “Cooperation for Victory: Democracy, International Partnerships, and War Performance, 1816–1992” (Ph.D. thesis, Duke University, 2001). 29. Kurt T. Gaubatz, “Democratic States and Commitment in International Relations,” International Organization 50 (winter 1996): 109–39. 30. For example, Fearon has explored the manner in which democracies, able to generate relatively greater domestic audience costs for leaders who escalate but lose crises short of war, are more likely to prevail in such crises; Bueno de Mesquita, Morrow, Siverson, and Smith have investigated the relatively greater interest of democratic as opposed to nondemocratic leaders to ensure victory by their nations in wars so as to ensure their political survival; and Goemans has shown that authoritarian leaders might find themselves trapped into continuing wars they cannot expect to win for fear of the especially dire consequences for them of losing such wars. James D. Fearon, “Domestic Political Audiences and the Escalation of International Disputes,” American Political Science Review 88 (September 1994): 577–92; Bruce Bueno de Mesquita, James D. Morrow, Randolph M. Siverson, and Alastair Smith, “Institutional Explanation of the Democratic Peace,” American Political Science Review 93 (December 1999): 791–807; Hein Goemans, War and Punishment: The Causes of War Termination and the First World War (Princeton, N.J.: Princeton University Press, 2000). 31. Johnston, Cultural Realism; Johnston, “Cultural Realism and Strategy in Maoist China”; Johnston, “Realism(s) and Chinese Security in the Post-Cold War Period.”Page 78 → 32. Among the many important readings in this line of analysis, see those by Michael Doyle, “Kant, Liberal Legacies, and Foreign Affairs, Part I,” Philosophy and Public Affairs 12 (summer 1983): 205–35; Michael Doyle, “Kant, Liberal Legacies, and Foreign Affairs, Part II,” Philosophy and Public Affairs 12 (fall 1983): 323–53; Michael Doyle, Ways of War and Peace: Realism, Liberalism, and Socialism (New York: Norton, 1997); Bruce Russett and Zeev Maoz, “Normative and Structural Causes of the Democratic Peace, 1946–1986,” American Political Science Review 87 (1993): 624–38; Bruce Russett, Grasping the Democratic Peace: Principles for a Post-Cold War World (Princeton, N.J.: Princeton University Press, 1993); John Owen, “How Liberalism Produces Democratic Peace,” International Security 19 (fall 1994): 87–125; and Bueno de Mesquita et al., “Institutional Explanation of the Democratic Peace.” The democratic peace thesis has been subjected to a number of analytical challenges and empirical tests, but thus far the thesis has remained highly robust. For criticisms of the thesis, see in particular those by Christopher Layne, “Kant or Cant: The Myth of the Democratic Peace,” International Security 19 (summer 1994): 5–49; David Spiro, “The Insignificance of the Liberal Peace,” International Security 19 (fall 1994): 50–86; Henry Farber and Joanne Gowa, “Polities and Peace,” International Security 20 (fall 1995): 123–46; Ido Oren, “The Subjectivity of the Democratic Peace,” International Security 20 (fall 1995): 147–84; William Thompson, “Democracy and Peace: Putting the Cart before the Horse?” International Organization 50 (winter 1996): 141–74; and Joanne S. Gowa, Ballots and Bullets: The Elusive Democratic Peace (Princeton, N.J.: Princeton University Press, 1999). For a highly effective response to most of the critiques of the democratic peace thesis, see Zeev Maoz, “The Controversy over the Democratic Peace: Rearguard Action or Cracks in the Wall?” International Security 22 (summer 1997): 162–98; and, in respect to Gowa's Ballots and Bullets, John Owen, “Review of Joanne Gowa's Ballots and Bullets,” American Political Science Review 94 (June 2000): 509–10; and Thomas Schwartz, “Review of Joanne Gowa's Ballots and Bullets,” Comparative Political Studies (June 2000): 690–93. An interesting qualification of the democratic peace thesis is found in the work of Mansfield and Snyder, in which they suggest that whatever might be the behavior of fullfledged democratic polities, democratizing states are peculiarly prone to become embroiled in military conflicts. Edward Mansfield and Jack Snyder, “Democratization and the Danger of War,” International Security 20 (summer 1995): 5–38. For the “democratic efficacy” literature, see Lake, “Powerful Pacificists”; Fearon, “Domestic Political Audiences and the Escalation of International Disputes”; Stam and Reiter, “Democracy, War Initiation, and Victory”; Bueno de Mesquita et al., “Institutional Explanation of the

Democratic Peace”; and Choi, “Cooperation for Victory.” Superior foreign decision making in security matters is also found to be associated with democracies in the work of Gaubautz, “Democratic States and Commitment in International Relations”; Snyder, Myths of Empire; and Goemans, War and Punishment. 33. Gilpin, U.S. Power and the Multinational Corporation.

Page 79 →

REALISM AND THE STUDY OF PEACE AND WAR John Vasquez Appraising realist theories or evaluating the realist paradigm from which they are derived is a very important topic; indeed it is one that has given rise to intense debate within the field from time to time. It is, however, too broad and complicated a topic for a short essay like this one. Therefore, I shall focus on realist explanations of two specific topics—peace and war—with an emphasis on the classical realism of Morgenthau, the neorealist work of Waltz and Gilpin, and the “offensive realism” of Mearsheimer.1 If appraising theory and evaluating paradigms is going to lead to fruitful debate between the adherents of competing approaches, there must be some shared understanding as to what are appropriate criteria of adequacy. Most of these are widely shared across disciplines, even between those as disparate as the physical and social sciences. These include such basic criteria as explanations should be empirically accurate and they should in principle be testable (i.e., there should be at least some body of evidence that would be accepted as refuting an explanation). In addition, it is generally accepted that theories that are more parsimonious are better than those that include all possible variables, and that theories that have the power (ability) to explain a variety of phenomena and events are better than those that can only explain the minuscule.2 Elsewhere, I have elaborated and justified various criteria that can be used to appraise theory and paradigms.3 Here, I will briefly apply the following to assess the adequacy of realist explanations of peace and of war: empirical accuracy, explanatory power, and policy relevance. Empirical accuracy means that propositions and hypotheses derived from the theory must be true, in that they do not contradict evidence marshaled to test or assess the explanation. Explanatory power means that a theory is able to provide a plausible answer to the major question(s) that a field of inquiry is trying to answer. Explanatory power is separate from empirical accuracy; in Page 80 → other words, a theory may have great explanatory power in that one can adopt the “logic” of a theory's or paradigm's perspective to derive a plausible account of how and why certain things occur in the world, but whether this account is true must await systematic investigation. Thus, astrology has great explanatory power in that it can explain and even claims to be able to predict a great number of events, but many think it is not accurate. Policy relevance is defined here as the ability of a theory to provide guidance to the most pressing practical problems of the day. Explanatory power and policy relevance are particularly important because these criteria are often seen as the strongest suits of realist analysis. There are also other criteria that can be employed, and one that has attracted widespread attention is Lakatos's notion of degenerative versus progressive research programs.4 I have applied that criterion to neotraditional research on Waltz's balancing-of-power proposition elsewhere.5 Some have maintained that this criterion is one of the more important, especially when part of the evidence is discrepant with a theory and gives rise to reformulations intended to save the theory. In the appraisal I made using Lakatos's criterion, I argued that there is sufficient evidence of degeneration in the neotraditional research program on balancing that a warning flag should be raised on future work in that area. By this I meant that serious scholarly work on the balance of power must show how, in principle, any reformulation that constructs a new theory is falsifiable and acknowledge that discrepant evidence falsified the previous formulations of the theory or show why it does not. Work that does not do that should not be labeled as progressive, but as failing to satisfy a crucial criterion in philosophy of science. I did not maintain, as Mearsheimer states, that the research of realist scholars does not “deserve continued funding, publication, and so forth.”6 The full context of that passage is given in note 7.7 What I did say, at the beginning of the article, was that a degenerative research program should not continue to command the intellectual energy and resources of the field. At the end of the article I said that continued theoretical reformulation of the balance of power needed special scrutiny because of this problem. I was careful to differentiate the failure of this one research program from other realist research programs, which I did not appraise in terms of this criterion, but I did say it raised serious questions about the realist paradigm as a whole. Paradigm

displacement, if it occurs, will come as a result of where individual scholars are willing to place their research bets. Open debate, agreement on criteria for appraisal, and careful rigorous research based on testable differences between competing theories will advance this process. I did not suggest that realist scholars should be denied access to journals, teaching positions, or grants simply because they are working within the realist tradition. Appraisal need not go hand in hand with the specter of the Spanish Inquisition, but it does mean that empirical claims cannot go on being made without substantiating evidence that meets rigorous standards. All of this, of course, is meant to apply only to empirical claims and theory. In the realm of normative claims, there is much less agreement on criteria of adequacy. More important, these questions are logically of a different Page 81 → order, since they deal with preferences. Here, realism still has much to offer, as I have stated elsewhere,8 especially in an age of “humanitarian interventionism.” Each area of realism, therefore, must be evaluated on its own terms. In this essay, I look at realist explanations of peace and war. I do not apply the criterion of degenerative versus progressive research programs, because there has not been a plethora of reformulating realist theory on this topic, so the question of degenerative versus progressive research does not even arise. Various elements of the realist paradigm can be traced back in the West at least to the time of Thucydides. In terms of the contemporary institutionalization of international relations inquiry as an academic discipline, which was founded at the University of Wales at Aberystwyth,9 the realist paradigm did not come to dominate the field until after the start of World War II with the idealist-realist debate taking place at the beginning of that war and immediately afterward.10 It is generally conceded that Hans J. Morgenthau's Politics among Nations 11 was the single work that transformed the field of international relations from “idealist advocacy to realist analysis,”12 even though his work is the cumulation of an entire generation's trying to formulate a realist alternative to the dominant idealism. Realist thinkers who helped lead the way for Morgenthau were Schuman, Carr, Niebuhr, Schwarzenberger, Wight, and Kennan among others.13 I argue that Morgenthau's work was of such importance in the post-World War II period that it can be regarded as an exemplar that gave rise to a paradigm for the field displacing the idealist paradigm.14 While his specific theory, which today we often refer to as classical realism, is different from other specific realist theories, it embodied certain fundamental assumptions about the world scholars were studying and said, in effect, that if one adopts these assumptions (as opposed to those of idealism or Marxism), then one would come to an understanding of the way the world really works. I identify the three most fundamental assumptions of the realist paradigm as: 1. Nation-states or their decision makers are the most important actors for understanding international relations. 2. There is a sharp distinction between domestic politics and international politics. 3. International relations is the struggle for power and peace. Understanding how and why that struggle occurs and suggesting ways for regulating it is the purpose of the discipline. All research that is not at least indirectly related to this purpose is trivial.15 Theories that share all three of these assumptions can be regarded as realist, those that reject one or more as nonrealist. Kuhn tells us that a paradigm does not provide all the answers; indeed, it may only provide scholars with the promise of answers.16 Theory construction, for Kuhn, is a process of paradigm articulation in which scholars use the assumptions of a paradigm to build theories that fulfill that promise. The heart of the realist paradigm and the very reason it was able to displace Page 82 → the idealist paradigm is that it promised to provide more accurate, more powerful, and more relevant answers to the questions of peace and war. I argue in this essay that it is precisely at the heart of the paradigm—on the questions of peace and war—where it has failed. The appraisal that is conducted here is based on explicit criteria, and it provides a strong test of realism, unlike the test Mearsheimer (herein) provides, which is too easy for realism to pass.

Mearsheimer's Irrelevant and Weak Test

Mearsheimer's (herein) main argument in his defense of realism is that realism is alive and well because states still engage in power politics and wars still occur. This is a weak defense of realism because it misses the deeper and more fundamental criticism that has been made of realism. I and other critics of realism, such as Burton,17 have not denied that power politics behavior exists or that it is about to be abolished forever in today's world. Power politics does exist, but my point is that it is but one form of relations, one form of doing politics. There are also other forms of relations, other ways of conducting politics. A good theory should explain when and why power politics behavior occurs, when and why other forms of relations occur, and what connections, if any, there are between the two. Long ago, I wrote: Power politics is not so much an explanation as a description of one type of behavior…. If this is correct, then power politics behavior itself must be explained; it does not explain…. An adequate theory of world politics would seek to discover when policy makers adopt a power politics image of the world, what kinds of behavior this image fosters, and when such behavior results in war…. [I]t should be possible to find non-power politics behavior and develop a theory of what conditions promote power politics and non-power politics behavior…. Such an approach … would not only explain everything the realist paradigm purported to explain but would also discover and explain a vast area of behavior that the realist paradigm purportedly ignored.18 The adequacy of realism, therefore, cannot be appraised simply by looking at the world and pointing to conflict and various forms of power politics practices. A good theory must explain when and why this behavior occurs, when it does not, why and how it leads to war, and whether it can be superceded to create peace. I have spent a great deal of my professional life dealing with these questions, and here I do not revisit that work. In this essay I examine a different aspect of realism—its explanations of peace and war. I examine the work of Waltz, Gilpin, Morgenthau, and Mearsheimer as it relates to this topic in light of the criteria of explanatory power and policy relevance. The work presented here is somewhat preliminary, because space does not permit a full textual analysis of all these thinkers. Nevertheless, after applying these criteria I reach several conclusions. I show that neo-realist Page 83 → explanations of war in many instances lack explanatory power because they are incomplete (or logically flawed, or both); they do not provide the minimal information we would expect from an adequate explanation. Since they fail, in principle, to provide what we want from an explanation, they do not provide much guidance for scholars trying to understand the causes of war. Classical realist and “offensive realist” explanations of war are logically superior to neorealist explanations, and while they too are somewhat incomplete, the main problem with them is that they are empirically inaccurate. After dealing with realists' explanations of war, I turn to their explanations of peace. Here, again, I find a paucity of explanations, so much so that I conclude that realism provides little guidance to a pressing issue of our time—constructing a peaceful post-cold war era. On this basis I conclude that realist explanations of peace have limited policy relevance for a key political issue of our day—the building of a structure of peace among major states.

The Realist Explanation of War This section begins by applying the criterion of explanatory power to realist explanations of the onset of war. It assesses realist explanations by asking whether they are logically sound and powerful explanations. It then turns to whether the explanations that appear the most powerful are empirically accurate. Although it may seem strange, neither Morgenthau nor Waltz provides a complete explanation of war. In his major statement on the subject Waltz divides explanations of war into three images: first, that human nature is the cause of war; second, that war results from the nature of states—typically their polity or economy; and third, that war results from global anarchy.19 Waltz maintains that war is a product of global anarchy. This he admits is only a “permissive cause,” in that war occurs because there is nothing to prevent it. The problem with this third image explanation is that anarchy is a constant. The criticism Waltz lays against the first image explanation is equally apropos here—namely, a constant cannot explain a variable.20

This puts Waltz in a bit of a dilemma because he can either predict constant warfare, which he obviously does not want to do, since this is not true, or he can introduce this unconventional idea of a “permissive cause.” The problem with the latter is that a permissive cause implies that there are other causes, sufficient conditions, or probabilistic factors that bring war about in any one specific instance. Waltz's explanation is inadequate because it fails to specify what these are. Presumably, because Waltz is a realist, we could assume that the causes of war have something to do with power. Nevertheless, the obvious ways in which power might affect war are not endorsed by Waltz. He refuses to say, for example, that balances of power are associated with peace; indeed he takes Gulick's position that they are not,21 since states may go to war to Page 84 → preserve the balance of power. Waltz's argument that balancing of power is a law means that anarchy produces balancing-of-power politics but not that balancing itself prevents war.22 Likewise, he does not attribute the causes of war to any other specific distribution of power or its disruption. While he maintains that a bipolar system is more stable than a multipolar system, this does not mean that wars will not occur under bipolarity. By maintaining that anarchy is a permissive cause, Waltz leaves the factors that actually bring about war unspecified, and thereby even implies that by default these factors may lie outside the system structure, for example, in the interactions of states. An adequate explanation of the onset of war should present an account of the factors that actually bring about wars. Relying on a constant, like anarchy, cannot logically make such a specification. Gilpin's explanation of war is better, in that he has a variable explaining a variable.23 For him, wars occur when the hegemon declines and a challenger arises. In his later treatment of the question of what causes war Gilpin says that the decline of a hegemon and the rise of a challenger is a necessary condition of war, not a sufficient condition.24 While this is an advance over Waltz, we still do not know what sufficient conditions are important, and without knowing the sufficient conditions of war, we cannot know why war occurs. Gilpin's explanation is also incomplete in another sense—his explanation applies only to a small set of hegemonic wars, that is, those between the hegemon and its challenger. It leaves a multitude of other interstate and imperial wars unexplained. The failure of both Gilpin and Waltz to specify sufficient conditions (or the factors that make war probable) means that the two major neorealist theories of war are still woefully incomplete and provide little, if any, guidance to the study of the causes of war. As such they can be seen as having limited explanatory power, at least to the question of the causes of war. Can the classical realism of Morgenthau do better? Classical realism may have an easier time than neorealism in its analysis of war because Morgenthau comes close to predicting that war is constant among states as they struggle for power and that what we have in international politics is constant warfare. In a famous quote he says: “All history shows that nations active in international politics are continuously preparing for, actively involved in, or recovering from organized violence in the form of war.”25 For Morgenthau, anarchy is not the source of war (indeed, unlike Carr26 he does not use this concept per se); rather war is a natural occurrence, a side effect if you will, of the struggle for power. The struggle for power Morgenthau attributes to a “will to power” inherent in human nature,27 but his analysis could just as easily be reformulated to have anarchy play this role. With this proposition, the debate over realism shifts from a logical problem with the explanatory power of realism and the question of “what a good explanation should look like” to a question of the empirical accuracy of the explanation that is offered. Page 85 → Applying the criterion of empirical accuracy to this explanation means examining the evidence provided in the historical record in as rigorous a way as possible to determine if warfare is fairly constant. Is it true that active states are continuously preparing for, involved in, or recovering from war? So long as defenders and critics of

realism do not try to define “preparing,” “active,” and “recovering” so broadly as to make Morgenthau's proposition nonfalsifiable, it should be possible to answer the question. Three pieces of evidence in the historical record suggest that the proposition is incorrect: (1) the presence of peaceful eras, especially among major states; (2) the presence of peaceful dyads; and (3) the presence of zones of peace. None of these three things should occur if war is relatively constant. For example, if Morgenthau and the thrust of realism, including Mearsheimer's offensive realism, are correct, we would not expect that the major states are always going to be at war with one another and that peace among them is an idea whose time will never come. This, in part, is what Mearsheimer (herein) is saying—namely, that the post-cold war period is not peaceful, that there are still wars and states engaged in power politics behavior. While the latter, as stated in the previous section, is not a strong defense of realism, it does illustrate the tendency of realism to deny that peaceful eras have actually occurred in the past. In other words, Morgenthau and Mearsheimer seem to agree that the struggle for power and war are relatively constant. Given this emphasis, it is not unreasonable to expect that war—or preparing for and recovering from it—is the condition that states find themselves in. The evidence since 1815, however, portrays a different picture. Of course, wars and power politics, especially among the European major states, are fairly frequent, but they are not constant. Peter Wallensteen finds that periods where major states attempt to construct rules of the game to regulate their interactions, even if these are not entirely successful, are associated with the absence of any wars among major states and a reduction in the number of militarized interstate disputes among them.28 The periods of peace he identifies are not unknown to historians: Concert of Europe, 1816–48; Bismarck's order, 1871–95; League of Nations, 1919–32; and détente, 1963–.29 Contrary to realist expectations, peace not only is possible but has actually occurred in the international system, and among the most war-prone states in the system, the major states.30 Even more damaging to the realist portrayal of the world is to examine all states in the post-Napoleonic period. If we look at how many years since 1815 there has not been an interstate war under way, as Geller and Singer do for the 1816–1992 period,31 we find that there are 81 peaceful years out of 180. In addition, they find that there are 150 states that never fight a war and an additional 49 who experience only one or two; conversely, a comparatively few states account for many of the wars that break out. These states are France, Britain, Germany, Italy, Russia, Greece, Egypt, and Turkey, all of which have more than ten wars, and in the post-1945 period—Israel, India, Pakistan, and the United States. Page 86 → The concentration of war among certain states means that there are many pairs of states that are peaceful dyads; that is, they do not fight each other. There are seventy-five interstate wars from 1816 to 1992, according to Geller and Singer.32 In this period, they point out that there are about 400 pairs of legally recognized states in 1816 and about 18,000 in 1992. Many of these pairs of states do not fight each other because they cannot reach each other; still the potential for war far outstrips the actual occurrence of war. If one looks just at neighbors, which have a higher war rate, the potential still outstrips the actual in that there are 40 bordering states in 1816 (with two interstate wars up to 1848) and 317 bordering states in 1992 (with five interstate wars from 1980 to 1991).33 War, according to Geller and Singer, is relatively rare and is concentrated in certain dyads. Most pairs of states, even if they are neighbors, are peaceful dyads for much of their history. This, of course, does not mean that there are not certain regions or neighbors that have the kind of continuous war preparation/war fighting/war recovery Morgenthau describes. One thinks of India and Pakistan, Israel and her Arab neighbors, and in an earlier era, Russia and the Ottoman Empire. The point is not that Morgenthau never applies but that the domain of the theory is less than universal and maybe even confined to certain periods and zones. The most obvious class of peaceful dyads that has received attention in recent years is the democratic dyad. There is considerable evidence that democratic states do not fight each other.34 Even if one wants to challenge the robustness of this finding,35 it remains the case that there are large numbers of dyads consisting of democracies that do not fight each other. This is troubling for realism, because it is not clear why a class of dyads whose common characteristic is their type of government should not fight each other. The peacefulness of the democratic

dyads is an anomaly for realism that is not easily explained away (see Midlarsky, herein). The weakness of realism on this point is underlined by the fact that this pattern was predicted by the rival liberal paradigm some time ago,36 and that a leading realist, Waltz in Man, the State and War, denied that the causes of war lay within the characteristics of states and specifically that democracy was a factor. The last piece of evidence that contradicts realist claims about the pervasiveness of war is the presence of zones of peace. The democratic peace is one such zone, but there are others. Some of these are geographic such as the Nordic peace system37 and North America after 1848 (excluding the Caribbean) as well as the general paucity of interstate warfare in the Western Hemisphere and sub-Saharan Africa (compared with Europe). Some regions of the world are less prone to interstate wars than others. Why is that? Just as realism in its various forms does not expect war to vary by time or type of government, so too it does not expect it to vary by geographic location. In this regard the west European peace since 1945 is an especially difficult anomaly for realism to explain away. The creation of the Schuman Plan and the evolution of the Common Market was based not on a realist theory of international politics but on the kind of nonrealist and often liberal theory of international relations exemplified by Mitrany and by Deutsch et al.38 Deutsch's idea of creating a security community where states do not even Page 87 → think of fighting each other is a very difficult test for nonrealist theory to pass, and a relatively easy one for realism to pass, yet in western Europe, the most war-torn region in the world from 1816 to 1945, and for a good time before that, peace was established. Nor can these states be seen as even preparing for war against each other. The armaments they have are not directed toward one another. But disarmament is not needed to support the point that peace is possible. States may still have arms in case peace decays, but the presence of arms does not mean that the armaments have caused the (west European) peace that exists. Explaining away this west European peace is critical for realists, and realists like Mearsheimer have recognized this, by saying that it was not the kinds of factors that Deutsch and others identify that produced this peace but the fear of the Soviet Union's power and the threat it posed. Thus, Mearsheimer predicts that with the end of the cold war, this peace will begin to fall apart and that European and democratic states will begin to turn on each other.39 This is an important prediction that provides a testable difference between realism and its critics, just as Waltz's prediction—that states will balance against the United States in the coming years—does.40 While the future will be the best arbiter of this disagreement, it remains the case that integration theory was used to build the European Union and that it was explicitly stated beforehand that such integration would eliminate the threat of another European war. It was the latter, not the Soviet Union, that was seen as the immediate threat by Schuman and the other architects of the European Coal and Steel Community. Another zone of peace that contradicts realist theory is the territorial zone of peace that exists across many dyads in the world.41 According to the territorial explanation of war,42 power politics and the struggle for power are most characteristic of neighbors that do not have legitimately recognized or established borders with each other. This class of dyads exhibits many of the behaviors that realists emphasize, but once borders are accepted, the territorial explanation maintains that peace can reign across the dyad even if other major nonterritorial issues arise and even if there is a shift in the capability of these states. Such a theory helps explain what may be underlying the relations of so many of the peaceful dyads that exist in the world—the settlement of their borders. The settlement of border disputes may not always be peaceful, but so long as it is accepted as a legitimate border (as the border between the United States and Mexico is, which was established through war), then according to this theory the probability of war should go down. Stephan Kocs provides some empirical evidence to support this claim.43 These examples also illustrate that while states may have fought each other in the past, they can reach such a state of peace that they do not seriously prepare for war against each other. This is certainly true of the United States and Canada despite the existence of military contingency plans each has against each other buried deep in their file drawers. Examining the question of whether peace is possible presents three anomalies for realism: the presence of peaceful eras, peaceful dyads, and zones of peace. None of these is really expected by classical and offensive realism. If Page 88 → one asks how these cases of peace were found, then one finds that in each instance it was the result of scholars being guided by nonrealist theory. The sudden “discovery” that democratic dyads do not fight each other

was produced this way, as was Wallensteen's findings. The findings unrelated to the democratic peace have not been as extensive or as closely scrutinized; nevertheless, they point to important theoretical differences between a realist and nonrealist approach and also serve as a basis for questioning the relevance of realism to deal with the most pressing problems of our own time—the establishment of a more peaceful order and one that would extend the current peace among major states far into the future.

The Relevance of Realist Prescriptions for Building a Structure of Peace The fact that both classical realism and offensive realism do not think a permanent peace is possible raises questions about how much their empirical theory can tell us about peace, and how it should be constructed. One of the criteria for appraising the ability of a theory to guide practice is the criterion of relevance. This criterion maintains that a good practical theory be able to provide sound guidance on the most difficult policy problems of the day.44 The most pressing problem major states face in the post-cold war era is how to build a peaceful order amongst themselves. How much guidance does realism provide for this major issue of our time? If one is interested in building a contemporary peaceful order or extending the current peace among major states far into the future, realism does not tell one how to build that order or why any given peaceful order ought to work. If one wants to know how to avoid specific wars, realism will provide prescriptions on how to avoid attack by building up power—peace through strength. I have argued that if you follow these prescriptions, generally, you will increase the probability of war, not avoid it.45 Today, and for long periods within international relations theory, many nonrealists have argued that constructing a global order through the creation of a common set of rules of the game, the promulgation of norms (and a body of international law) to regulate behavior, and the erection of a set of international institutions is critical for building and maintaining peace. Mearsheimer has referred to this as “the false promise of international institutions.” He argues that there is little evidence that institutions can preserve a peace, let alone create it.46 Mearsheimer ignores, however, an entire body of evidence produced by quantitative peace research that has analyzed data on the post-1815 period. This research has found evidence that rules, norms, and institutions do matter. As with any set of limited and continuing studies within social science, the evidence on this question is not definitive, but it is still a more reliable guide than argumentation based on anecdotal evidence. Regardless of one's position on quantitative analysis, this is a body of evidence that should not be ignored. What does that evidence say? Page 89 → In “Universalism versus Particularism: On the Limits of Major Power Order,” Wallensteen attempted to measure, through the use of historical judgment, those periods from 1815 to 1976 in which the major states attempted to establish rules of the game for managing their relations. He found that during these periods there were no wars among major states, although there were other types of wars, especially imperial wars. His analysis has been criticized for relying on historical judgment rather than on more precise operational indicators. This concern is met by Kegley and Raymond,47 who in a series of studies find that in the half decades when the pacta sunt servanda tradition of international law (which maintains that treaties are binding regardless of circumstances) is dominant there are fewer wars and militarized disputes than when the rebus sic stantibus tradition of international law (which says that treaties are binding only so long as the matters stand) is dominant. Raymond in a recent review of this literature argues that this evidence strongly suggests that restrictive normative orders are more peaceful than permissive normative orders.48 Norms that rein in unilateral behavior reduce the incidences of the use of force, including war. Why this helps produce peace can be explained by the findings of a very early study by Wallace,49 who documented a path to peace separate from a path to war for the 1820–1964 period. He found that when there is status inconsistency among major states, they tend to engage in alliance making and then arms racing, which he

found to be highly correlated with war. Conversely, when there is no status inconsistency, states do not engage in this sort of power politics behavior. Instead, the absence of status inconsistency is correlated with the creation of intergovernmental organizations (IGOs), which in turn is negatively correlated with the presence of arms racing. This can be seen as a path to peace, whereas the other can be seen as a path to war. Faber and Weaver provide additional support for these results in their study of European relations from 1816 to 1915.50 They find that the more conferences and treaties signed, the less likely war. They also find that settling certain kinds of issues, namely territorial issues, is negatively correlated with war. Faber and Weaver provide important quantitative evidence that the Concert of Europe did work. Their study, along with Wallace's, suggests that IGOs and conference diplomacy can provide a functional equivalent to war, by providing procedures for making binding decisions. Since Mearsheimer wrote “The False Promise of International Institutions” in 1994–95, there has been other research on the importance of institutions. Russett, Oneal, and Davis find that states that have high shared memberships in international organizations are less likely to fight each other.51 I have found that in Wallensteen's (“universalist”) periods, where norms are present, the threat or use of force to settle territorial issues is less frequent and arms racing among major states is almost nonexistent.52 Those are two important clues as to why norms are associated with peaceful eras. The use of force on territorial issues greatly increases the probability of war,53 as does the presence of arms racing in the context of crisis interactions,54 yet in Wallensteen's universalist periods these phenomena go down. Page 90 → The thrust of this body of research suggests that in the international relations of major states there are two separate worlds. One is the world of power politics, which often degenerates into war. The other is a world of rules of the game, norms, and international organizations that is negatively correlated with the power politics practices of alliance making, coercive diplomacy, and arms racing. The limited evidence produced by peace research implies that the presence of norms, rules, and international organizations reflects a different way of conducting political business. To avoid war it is necessary to create nonviolent ways of contending on political issues that are as effective as war. It is also necessary to control and manage the kinds of issues that are on the political agenda so that the most war-prone issues, which are territorial disputes, are kept off the agenda.

Conclusion This analysis maintains that the core questions that the realist paradigm has set for the field—the causes of war and the struggle to achieve peace—have not been adequately answered by classical realism, neorealism, and offensive realism. These three branches of the realist paradigm, in terms of the study of peace and war, have been found deficient on two of the criteria on which realism has historically scored high—explanatory power and policy relevance. These three branches of realism lack a complete or full explanation of war. They do not specify the sufficient conditions of war and thus leave scholars with little guidance as to what factors bring about war. While Gilpin is better than Waltz and Morgenthau in that he is more precise, he confines his analysis to only the necessary conditions of war, and to only the very small set of hegemonic wars between the two strongest states in the system, which leaves most wars in history unexplained and unexamined. Morgenthau may provide the least flawed analysis if he is seen as employing a constant (the will to power) to explain a constant (war). Such an explanation, however, seems empirically inaccurate because war is comparatively rare within the system, at least since 1815. Because realism tends to see war as fairly frequent, it has not given much attention to an explanation of peace. If permanent peace is not possible, what is there to explain? This means that realism provides little guidance to building a structure of peace in the current era.55 The denial that peace is possible has produced three anomalies for realism—the democratic peace, the presence of peaceful eras, and the presence of other zones of peace. None of these is expected by realism, especially peace among major states. Because peace is not expected, realists have spent little time trying to explain it. It is dismissed as either a respite between periods of warfare or the result of overwhelming power.

Understanding politics requires more complicated models. The perspective of power politics prevents us from seeing what are the more important factors in international relations by focusing our attention on that which, at best, applies only to a very limited domain of the historical record. A theory that can explain that limited domain plus all the rest of world politics Page 91 → is, by definition, a better theory and deserves to displace realism. What that theory and paradigm will be is still an open question, but it is apt to be a synthesis of midrange theories that will explain not only the onset of war but also the building of peace. Much of this problem results from too great a focus on the role of power and an insufficient attention to other variables. This means that those forms of realism that have broken with this focus to incorporate the importance of institutions,56 domestic politics,57 cognitive psychology and perceptions,58 or the security dilemma59 hold more promise. Nevertheless, a complete break with the realist paradigm will probably produce a better guide to research, as evinced by the few studies reviewed here, as well as more theoretical insights. There is no need for the field as a whole to choose one research strategy over the other; indeed, there is less collective risk in a theoretical pluralism in the short run. However, there is a need for systematic and rigorous theory appraisal using a common set of criteria if the long-term risk of devolving into nonfalsifiability and dogma is to be avoided. War is brought about by more than just differences in capability. Peace cannot be preserved and built simply because one side is more powerful than another. Politics, like everyday life, does not consist solely of intimidation and coercion. Until the field has a paradigm that examines more than just the single factor of power, it will not understand what causes war or how peace can be built.

Notes My thanks to Michael Brecher and Marie Henehan for valuable comments. The views expressed here, however, should not be attributed to either of them. 1. Hans J. Morgenthau, Politics among Nations: The Struggle for Power and Peace, 1st and 5th eds. (New York: Knopf, 1948, 1978); Kenneth N. Waltz, Theory of International Politics (Reading, Mass.: AddisonWesley, 1979); Robert Gilpin, War and Change in World Politics (Cambridge: Cambridge University Press, 1981); John J. Mearsheimer, “Back to the Future: Instability in Europe after the Cold War,” International Security 15 (summer 1990): 5–56; John J. Mearsheimer, “The False Promise of International Institutions,” International Security 19 (winter 1994–95): 5–49. 2. Thomas S. Kuhn, The Structure of Scientific Revolutions, 2d ed., enlarged (Chicago: University of Chicago Press, 1970). 3. John A. Vasquez, “World Politics Theory,” in Encyclopedia of Government and Politics, ed. M. Hawkesworth and M. Kogan (London: Routledge, 1992), 836–61; John A. Vasquez, The Power of Power Politics: From Classical Realism to Neotraditionalism (Cambridge: Cambridge University Press, 1998), ch. 10. 4. John A. Vasquez and Colin Elman, eds., Realism and the Balance of Power: A New Debate (Upper Saddle River, N.J.: Prentice Hall, forthcoming). 5. John A. Vasquez, “The Realist Paradigm and Degenerative versus Progressive Research Programs: An Appraisal of Neotraditional Research on Waltz's Balancing Proposition,” American Political Science Review 91 (December 1997): 899–912. 6. John J. Mearsheimer, “Realism, the Real World, and the Academy,” in this volume. 7. “The case that any given research program is degenerating (or progressive) cannot be logically proven…. A more reasonable stance is that exemplified by the trade-off between type 1 and type 2 errors in deciding to accept or reject the null hypothesis. Deciding whether a research program is degenerating involves many individual decisions about where scholars are willing to place their research Page 92 → bets, as well as collective decisions as to which research programs deserve continued funding, publication, and so forth. Some individuals will be willing to take more risks than others. This analysis seeks to present evidence that

is relevant to the making of such decisions.” Vasquez, “The Realist Paradigm,” 900. 8. John A. Vasquez, “The Enduring Contributions of Hans J. Morgenthau's Politics among Nations,” International Studies Notes 24, no. 1 (1999): 5–9. 9. Brian Porter, ed., The Aberystwyth Papers: International Politics, 1919–1969 (London: Oxford University Press, 1972). 10. John A. Vasquez, The Power of Power Politics: A Critique (New Brunswick, N.J.: Rutgers University Press, 1983); Vasquez, The Power of Power Politics: From Classical Realism to Neotraditionalism, ch. 2. 11. Morgenthau, Politics among Nations, 1st ed. 12. William C. Olson, “The Growth of a Discipline,” in Porter, The Aberystwyth Papers, 19–20. 13. Frederick L. Schuman, International Politics (New York: McGraw-Hill, 1933); E. H. Carr, The Twenty Years' Crisis (New York: Harper and Row, 1939); Reinhold Niebuhr, Christianity and Power Politics (New York: Scribner's, 1940); Georg Schwarzenberger, Power Politics (New York: Praeger, 1941); Martin Wight, Power Politics: “Looking Forward,” pamphlet no. 8 (London: Royal Institute of International Affairs, 1946); and George Kennan, American Diplomacy (Chicago: University of Chicago Press, 1951). 14. Vasquez, The Power of Power Politics: A Critique, 35–39. 15. Vasquez, The Power of Power Politics: From Classical Realism to Neotraditionalism, 37. 16. Kuhn, The Structure of Scientific Revolutions, 23–24. 17. John Burton, Global Conflict (Brighton, England: Wheatsheaf, 1984). 18. Vasquez, The Power of Power Politics: A Critique, 216; reprinted in Vasquez, The Power of Power Politics: From Classical Realism to Neotraditionalism, 167–68. 19. Kenneth N. Waltz, Man, the State, and War (New York: Columbia University Press, 1959). 20. Waltz, Man, the State, and War. Hidemi Suganami, On the Causes of War (Oxford: Clarendon Press, 1996), 24–25, makes a similar point. 21. Edward V. Gulick, Europe's Classical Balance of Power (New York: Norton, 1955). 22. Waltz, Theory of International Politics, 121. 23. Gilpin, War and Change in World Politics. 24. Robert Gilpin, “The Theory of Hegemonic War,” in The Origin and Prevention of Major Wars, ed. R. Rotberg and T. Rabb (Cambridge: Cambridge University Press, 1989), 17. 25. Morgenthau, Politics among Nations, 5th ed., 42. 26. Carr, The Twenty Years' Crisis. 27. Morgenthau, Politics among Nations, 5th ed., 36–37. 28. Peter Wallensteen, “Universalism vs. Particularism: On the Limits of Major Power Order,” Journal of Peace Research 21, no. 3 (1984): 243–57. 29. Wallensteen's analysis in ibid. ends in 1976, but I have assumed that his classification of this period as universalist would extend to the present. 30. On the war proneness of major states in comparison with other types of states, see Stuart Bremer, “National Capabilities and War Proneness,” in The Correlates of War, vol. 2, ed. J. David Singer (New York: Free Press, 1980), 57–82; and Stuart Bremer, “Who Fights Whom, When, Where, and Why,” in What Do We Know about War? ed. John A. Vasquez (Lanham, Md.: Rowman and Little-field, 2000): 23–36. 31. Daniel S. Geller and J. David Singer, Nations at War: A Scientific Study of International Conflict (Cambridge: Cambridge University Press, 1998), 1. 32. Ibid.Page 93 → 33. The number of wars is calculated from Meredith Reid Sarkees, “The Correlates of War Data on War: An Update to 1997,” Conflict Management and Peace Science 18 (fall 2000): 123–44, especially app. A. 34. Bruce Russett, Grasping the Democratic Peace (Princeton, N.J.: Princeton University Press, 1993); Bruce Russett and John Oneal, Triangulating Peace (New York: Norton, 2001); and James Lee Ray, Democracy and International Conflict (Columbia: University of South Carolina Press, 1995). 35. Joanne Gowa, Ballots and Bullets: The Elusive Democratic Peace (Princeton, N.J.: Princeton University Press, 1999); for specific cases see Miriam Fendius Elman, ed., Paths to Peace: Is Democracy the Answer? (Cambridge: MIT Press, 1997). 36. Immanuel Kant, Perpetual Peace, 1795; Rudolph J. Rummel, “Libertarianism and International Violence,” Journal of Conflict Resolution 27 (March 1983): 27–32; Michael Doyle, “Liberalism and World Politics,” American Political Science Review 80 (December 1986): 1151–69.

37. Nazli Choucri, “In Search of Peace Systems: Scandinavia and the Netherlands, 1870–1970,” in Peace, War, and Numbers, ed. B. M. Russett (Beverly Hills, Calif.: Sage, 1972), 239–99. 38. David Mitrany, A Working Peace System (Chicago: Quadrangle Books, 1943); Karl W. Deutsch et al., Political Community and the North Atlantic Area (Princeton, N.J.: Princeton University Press, 1957). 39. Mearsheimer, “Back to the Future,” 46–48. 40. Kenneth N. Waltz, “The Emerging Structure of International Politics,” International Security 18 (fall 1993): 44–79. 41. Arie M. Kacowicz, “Explaining Zones of Peace: Democracies as Satisfied Powers?” Journal of Peace Research 32 (August 1995): 265–76. 42. John A. Vasquez, The War Puzzle (Cambridge: Cambridge University Press, 1993), ch. 4. 43. See Stephan Kocs, “Territorial Disputes and Interstate War, 1945–1987,” Journal of Politics 57, no. 1 (1995): 159–75; Kacowicz, “Explaining Zones of Peace”; and Mark Zacher, “The Territorial Integrity Norm: International Boundaries and the Use of Force,” International Organization 55 (spring 2001): 215–50. 44. Vasquez, The Power of Power Politics: From Classical Realism to Neotraditionalism, 234–35. 45. Vasquez, The War Puzzle, chs. 3, 5. 46. Mearsheimer, “The False Promise of International Institutions,” 84. 47. Charles W. Kegley Jr. and Gregory A. Raymond, “Alliance Norms and War: A New Piece in an Old Puzzle,” International Studies Quarterly 26 (December 1982): 572–95; Charles W. Kegley Jr. and Gregory A. Raymond, “Normative Constraints on the Use of Force Short of War,” Journal of Peace Research 23 (September 1986): 213–27; Charles W. Kegley Jr. and Gregory A. Raymond, When Trust Breaks Down (Columbia: University of South Carolina Press, 1990). 48. Gregory A. Raymond, “International Norms: Normative Orders and Peace,” in Vasquez, What Do We Know about War? 233–36. 49. Michael D. Wallace, “Status, Formal Organization, and Arms Levels as Factors Leading to the Onset of War, 1820–1964,” in Russett, Peace, War, and Numbers, 49–69. 50. Jan Faber and R. Weaver, “Participation in Conferences, Treaties, and Warfare in the European System, 1816–1915,” Journal of Conflict Resolution 28 (September 1984): 522–34. 51. Bruce Russett, John Oneal, and David R. Davis, “The Third Leg of the Kantian Tripod for Peace: International Organization and Militarized Disputes, 1950–1985,” International Organization 52, no. 3 (1998): 441–68. 52. John A. Vasquez, “Mapping the Probability of War and Analyzing the Possibility of Peace,” presidential address to the Peace Science Society (International), Conflict Management and Peace Science (forthcoming).Page 94 → 53. Paul Hensel, “Charting a Course to Conflict: Territorial Issues and Interstate Conflict, 1816–1992,” Conflict Management and Peace Science 15, no. 1 (1996): 43–73; Paul D. Senese, “Geographical Proximity and Issue Salience: Their Effects on the Escalation of Militarized Interstate Conflict,” Conflict Management and Peace Science 15, no. 2 (1996): 133–61; John A. Vasquez and Marie T. Henehan, “Territorial Disputes and the Probability of War, 1815–1992,” Journal of Peace Research (forthcoming). 54. Michael D. Wallace, “Armaments and Escalation: Two Competing Hypotheses,” International Studies Quarterly 26 (March 1982): 37–51; Susan G. Sample, “Arms Races and Dispute Escalations: Resolving the Debate,” Journal of Peace Research 34 (February 1997): 7–22. 55. Does this mean that realism has nothing to say or offer, or that it has not provided an important contribution to international relations theory? No, see Vasquez, “The Enduring Contributions of Hans J. Morgenthau's Politics among Nations,” where I delimit some of Morgenthau's lasting contributions, including realism's role as an important antidote to liberalism's penchant for messianism. 56. Randall L. Schweller and David Priess, “A Tale of Two Realisms: Expanding the Institutions Debate,” Mershon International Studies Review 41 (May 1997): 1–32. 57. Jack Snyder, Myths of Empire: Domestic Politics and International Ambition (Ithaca, N.Y.: Cornell University Press, 1991); Jennifer Sterling-Folker, “Realist Environment, Liberal Process, and DomesticLevel Variables,” International Studies Quarterly 41 (March 1997): 1–25. 58. Jack Snyder, The Ideology of the Offensive (Ithaca, N.Y.: Cornell University Press, 1984); William Curti Wohlforth, The Elusive Balance: Power and Perceptions in the Cold War (Ithaca, N.Y.: Cornell University

Press, 1993). 59. Charles L. Glaser, “Realists as Optimists: Cooperation as Self-Help,” International Security 19 (winter 1994–95): 50–90.

Page 95 →

PERFORMANCE AND PERILS OF REALISM IN THE STUDY OF INTERNATIONAL POLITICS K. J. Holsti A long tradition of research, interpretation, and speculation about international politics goes under the name of realism. Although some dispute the pedigree as simply a “construction” of contemporary analysts, in fact a number of writers whose roster includes Machiavelli, Rousseau, Meinecke, von Gentz, Morgenthau, and Waltz, to name just a few, share a number of common themes such as their images of the world, their understanding of historical processes, the dilemmas and paradoxes that are peculiar to states that exist in a condition of anarchy, and the normative primacy of security in such an environment. They each made unique contributions to our interpretation and understanding of international politics, but they also shared many perspectives. Hence we can say that realism constitutes a distinct tradition in the field. Like all theories of international politics, realism has four functions: (1) it is a compressed, or shorthand, description of the essential characteristics of international politics; (2) it is a diagnostic tool for analyzing the sources of international conflict and derivatively for identifying the conditions for peace and security; (3) it is a set of prescriptions that derives from the diagnostic exercise; and (4) it is a normative theory that posits the survival and security of the political community as the ultimate political value. Any evaluation of the tradition must distinguish between these four functions.

The Essential Characteristics of International Politics As a descriptive enterprise, realism has offered a great deal about how some states behave some of the time. It has generalized and thus helped create understanding in an area that would otherwise be dominated solely by historical narratives organized around frameworks of time, place, and personality. Without generalization there would be no distinct field of study called international politics, but only diplomatic or military histories. Morgenthau claimed more than a half century ago that realism offers a “portrait” of a phenomenon, not an exact replica. It highlights essences at the cost of empirical Page 96 → precision. This is the necessary cost of generalization. Thus, the charge that realism both over-and underdetermines is in part beside the point. No competing theoretical rendering of the field offers generalizations that are less deficient from an empirical perspective. Some—one thinks of the theories of globalization or constructivist approaches, for example—have yet to meet even the most basic requirements of authoritative social science, nor have they developed a tradition of speculation and analysis of continuing ethical problems and dilemmas. The realist belief in the famous Waltz proposition that the quality of international life does not change in anarchical systems overlooks a mounting body of evidence that matters are otherwise today. The texture of international politics differs in significant ways from that of the eighteenth or nineteenth century, or even of the first ninety years of the twentieth century. In their descriptions of the essential characteristics of international politics, realists have typically made two important assumptions: (1) these characteristics are universal: we will find similar behaviors wherever there are independent political communities and the consequent condition of anarchy; and (2) these characteristics will not change through historical processes so long as the condition of universal anarchy exists. In brief, the realist descriptions of the essential characteristics of international politics transcend time, place, and personality. The universalist assumption is no longer warranted. Realism has not been able completely to transcend time, place, and personality. Consider some questions that arise from recent patterns and anomalies in diplomaticmilitary practice: Why are alliances—a ubiquitous phenomenon of eighteenth- or nineteenth-century Europe—not to be found in twentieth-century South America, or in contemporary Africa or Southeast Asia? In fact, why are alliances

today so rare throughout the world except in Europe? Why are there no security dilemmas in South America today even though it is a complete subsystem, with a nineteenth-century history characterized by frequent wars, bids for hegemony, and balancing behavior? Why are so many countries in the world reconfiguring their armed forces away from defense-related functions to peacekeeping and internal policing (including repression of the state's population)? If the world is characterized by anarchy and the security dilemma—threats from neighbors—why is it that today most states do not make preparations to defend their borders against external aggression? The Maginot Line is an archaeological relic, and one finds few counterparts elsewhere. The Franco-German border is today a weapons-free zone, as is the border between Norway and Sweden, Canada and the United States, Mozambique and Malawi, the Ivory Coast and Ghana, and many others. Why are there no hegemons or pretenders to such role or status in South America or Southeast Asia and perhaps in some other areas of the world?Page 97 → Why have the vast majority of armed conflicts occurred within states since 1945? The problem par excellence that drives realist analysis is the relations between states, yet the incidence of interstate war has declined precipitously in the last half century.1 In this period, far more people have died or been otherwise threatened by their own governments, secessionist movements, protostates, or warlords than by foreign armies. Why are the prospects of interstate war in some regions in the world almost nonexistent? What does the dramatic decline in the probability of war between states in Western Europe in the past fifty-five years tell us about the contemporary texture of international politics? Why are the relations between states in Western Europe so fundamentally different than they were just one generation ago, not to mention three hundred years ago? Do these questions suggest mere anomalies? Or are there important trends in the texture of international politics that challenge realist descriptions and explanations? John Mearsheimer argues in his contribution to this volume that realist-type behavior is “alive and well.” Indeed it is—at least in some areas of the world. But can one productively employ traditional strategic concepts that may possibly be relevant to Asia also to the processes of European unification?2 Numerous studies have demonstrated beyond doubt that a realist-type analysis of the relations between the European Union members would be incomplete at best and totally misleading at worst. Other approaches, including institutionalism and constructivism, do a better job of describing and explaining the political dynamics of this region. So, contrary to realist characterizations of international politics, much has indeed changed in the quality of international life. Realist-inspired literature neither acknowledges it nor attempts to provide explanations. Its universalist pretensions are open to serious scrutiny. Many trenchant critiques of realism and the problem of change are available. My fellow authors John Vasquez and Joseph Grieco have addressed this issue, so I do not need to cover the terrain again. However, I would add the point that at the most fundamental level, all theories of international politics are basically arguments about change: of what kinds, when, and whether it is possible, given certain psychological or structural characteristics. The problem is that few theorists specify what they mean by change, how they identify it, and how to assess its consequences. Change can mean anything from added complexity, through transformation, to obsolescence. New phenomena are also a form of change. By failing to address these critical distinctions, most theories or approaches to the field are deficient. But realism is particularly flawed and narrow-sighted, for the only change it acknowledges is in the number and cast of great-power characters and in the distribution of capabilities between them. Major international institutions such as colonialism have come and gone, hardly noticed in the realist literature. Major innovations such as peacekeeping do not fit well with the great-power bias of realism. There appear to be significant changes in international Page 98 → norms and how they operate to constrain government choices, and yet, as Stephen Krasner has recently argued,3 power and interest generally trump norms. The evidence partly supports this view, but recent constructivist and norm-related literatures show convincingly that norms often modify interests, indeed that norms and interests often become inseparable.4 The list of changes that have significant consequences on the quality of international life is long but largely ignored in the realist-inspired literature.

Perhaps the most damaging lacuna in realism is its dismissal of or lack of interest in revolutions. The main variables in realist analyses are distributions of power or other structural characteristics. Since states are vitally interested in power either as an end or a means, it follows that conflict is an inevitable consequence of power differentials. In terms of historical change, then, war and its outcomes have been the most significant consequences of the search for power and security. War, in this view, is the major motor of historical change. Yet, as Fred Halliday has convincingly argued, revolutions have been no less significant engines of international change.5 They have been both major sources and major consequences of wars. While balances and imbalances of power, alliances, and concentrations of power have all been correlated with the incidence of war, realist-inspired studies have not examined the significant causal relationships between revolutions and war. To leave out one of the most important “international” events—revolutions—from analysis is a major shortcoming of realism's theoretical lens and its capacity to generate convincing general characterizations of the world of international politics. Revolutions are based on ideas. Ideas, coupled with political mobilization, are a main source of international change. One can hardly understand comprehensively the main features of nineteenth-or twentieth-century international politics while ignoring ideas. But most realists are intellectual materialists. The only significant source of change in the distribution of capabilities is the law of uneven economic growth between states.6 The rise and decline of hegemons, balances, imbalances, degrees of polarity, and the like are explained in realist lore through changes in capabilities. There is sparse room in structural and other versions of realism for the driving force of ideas.7 As constructivist critics term it, interests are exogenously given in realism. Yet one would get a highly distorted portrait of the essential characteristics of international politics during the last two centuries without acknowledging the fundamental importance of ideas as sources of policy, crises, wars, and revolutions. Realism is also incomplete within its own frame of reference. It is incomplete, for example, in offering only a limited roster of strategies for coping with the consequences of anarchy or human psychology. It emphasizes balancing as the only option available to threats posed by potential or actual hegemons. The historical record shows, as the diplomatic historian Paul Schroeder8 has demonstrated, that other choices such as “indifference” (Latin America during World War II), “hiding” (neutrality or isolationism), subversion (undermining the adversary), or integration are available and have been made. The realist strategic menu also underlines that bandwagoning Page 99 → is not an option for dealing with the hegemony problem, but I know of no other way to describe the 1936 anti-Commintern Pact, the May 1939 “Pact of Steel,” or the Soviet-Nazi arrangements of September 1939. In fact, the responses to Germany's threat to traditional Westphalian norms of territorial independence in the late 1930s ranged from complete indifference, through hiding and balancing, all the way to bandwagoning. Such a wide range of responses in just one diplomatic scenario necessarily raises serious doubts about the descriptive and predictive capacity of realism's balance-of-power theory. Indeed, that theory is mostly a muddle, for it has been used to explain both war and peace. Even if the real purpose of balancing is to prevent hegemony—as it did several times in Europe's modern history—then why is there no countercoalition to today's hegemon? One possible answer to such empirically-based criticisms is that realism was never a complete project and has done what any good theory should do: it has stimulated further research and promoted improvement of guiding assumptions, main tenets, and shorthand characterizations. Refinements, amendments, and reconfigurations have indeed appeared in works by Robert Art, Robert Gilpin, Jack Levy, Michael Barnett, and many others. But from an empirical point of view, even if the research program is not entirely “degenerative,” as John Vasquez has argued, there seems to be an increasing gap between what realism—even its amended versions—offers as the essential characteristics of international politics and what we observe in everyday events and trends. Take the matter of conflict, the normative core of realism. Both the structuralist and psychology versions of realism propose that conflict is an inevitable or at least highly probable consequence of anarchy. But as Karl Deutsch and Alexander Wendt have demonstrated in their research programs, “anarchy is what states make of it.” Karl Deutsch's great contribution to the theory of international politics was to turn Rousseau and Hobbes on their heads: instead of insecurity as an inevitable consequence of human nature or the condition of anarchy, he demonstrated empirically that we can have states (anarchy) and predictable peace between them too.9 The vast

literature on the democratic peace has largely confirmed Deutsch's original formulation of the concept of the pluralistic security community. If we con- fine our studies only to power rivalries and to those relationships that do in fact conform to realist characterizations, we would have to ignore the relations of most states to each other most of the time. The truth is that for them, insecurity, war, and conflict no longer command the relevant parts of their diplomatic environments that they once did. If the realist explanation for this is a successful hegemonic Pax Americana, then what happened to the balance of power that was supposed to prevent the rise of a hegemon?

Realism as a Diagnostic Tool Theorists also seek explanations. They want to discover the sources of the essential characteristics they describe. Both Rousseau and Waltz have contributed brilliant explanations for some critical questions in the field, such Page 100 → as why wars and balances of power recur, why the state system reproduces itself, why the price of sovereignty is (often) insecurity, why states resist integration and specialization, and why, despite the obvious increments to security and economic welfare through cooperation, they avoid dependence and frequently fail to cooperate. Structural explanations of such phenomena are powerful and parsimonious, both virtues of any theoretical statement. They help us understand a particular syndrome that accompanies major rivalries and numerous international conflicts. The essential features of the India-Pakistan rivalry, pre-earthquake GreekTurkish relations, Latin America in the nineteenth century, or the Middle East are nicely encapsulated by such theoretical explanations. Realism was the predominant academic version of international politics during most of the cold war precisely because there was a reasonable fit between what analysts observed in East-West relations and what realist hypotheses predict. In comparison, liberalism, Marxism, or “world society” explanations or theories could not deliver many goods in that era. But realist diagnostics, particularly those of the Hobbesian tradition, have had a difficult time establishing any sort of lasting explanatory reputation. Any explanation that rests upon a fixed list of personality attributes such as the search for power, the search for glory, and diffidence falls flat when lifted to the level of states. This version of realism simply cannot explain the huge variations in state behavior when they are aggregated. Immense systemic variations such as the absence of great-power war (Korea as a partial exception) since 1945, terminating a pattern of almost constant great-power warfare since the sixteenth century; the lack of war in South America since 1942; the precipitous decline of territorially-based conflicts since 1945; and a host of other trends, generalizations, and anomalies escape detection when viewed from the Hobbesian lens of international politics. Structural explanations do not fare much better, but it can be argued, as Rousseau and Waltz suggest, that they were never designed to explain particularities, but only generalizations of a very broad order. Yet students of the field are not comfortable only with broad generalizations. There is much more to the field, they suggest, than propositions such as “the price of sovereignty is international insecurity” or “states are compelled to seek relative rather than absolute gains.”

Prescription It is difficult to evaluate the prescriptive lessons that derive from diagnostic exercises of a very general order. Some, like Morgenthau, insist that prudence and maintaining a balance of power are sufficient to create some semblance of order and security in the system. But as Morgenthau himself learned during his strident efforts to change American policy in Vietnam in the 1970s, his notions of prudence or balance were not shared by Lyndon Johnson. Johnson's policies and the intellectual analyses upon which they were based were at least consistent with realist policy tenets, but in hindsight they led to folly. The prescriptive dimension of realism is by no means Page 101 → self-evident, even when analysts share its general analytical perspectives. Their prescriptions are, moreover, undermined by structuralist arguments to the effect that balances form naturally irrespective of human intentions. Maybe the only significant prescriptions that derive logically from the explanations are that the world is a dangerous place and that one must prepare for the eventuality of malefactors. This leaves immense latitude of choice, notwithstanding structural arguments that system characteristics compel policymakers to behave in certain ways irrespective of their personal preferences. Moreover, there is not much evidence that policymakers are familiar with, or guided by, prescriptions deriving

from academic theories of international politics. True, the early prototypes of the French diplomatic academy in the eighteenth century had in their compulsory curriculum texts by Machiavelli, but they also included Pufendorf, Grotius, and (later) Vattel. Some have argued that most American foreign service officers during the cold war were familiar with Morgenthau's great work, but it is unlikely that they took it much to heart as virtually the whole of American policy during the cold war, and particularly in Vietnam, violated almost every prescription advocated by the professor. Indeed, the two most vocal critics of American policy during the cold war have both been characterized as realists: Hans Morgenthau and George Kennan. Thus, either this pigeonholing is invalid or American policy during the cold war was not based on realist tenets. But it could be argued that the influence of academic theories on the public or policymakers, or on both, is more subtle. For example, realism may legitimize a climate of fear. The notion that the inevitable consequence of anarchy is insecurity, for example, may end up being a self-fulfilling prophecy. Certainly during the cold war the mood of fear that served as the permissive background for spending $5.8 trillion in perfecting the system of mutual assured destruction was sustained if not openly advocated by realist-type analyses. The United States spent vast sums in the forty-five years of the cold war to counter a threat that was blown out of proportion—sometimes deliberately for domestic political reasons. That a gullible public could swallow the story of the Soviet Union as a mighty power out to conquer the world through military means might have been in part due to academic theories that emphasize the nastiest side of international politics. I doubt it was a necessary condition, however, because despite the recent decline of realism as the predominant academic version of international politics, such official thinking persists. For example, we used to be told that the only way to deal with the communist threat was to build mighty second-strike retaliatory forces. But now we are told that such forces cannot deter “rogue” states and that therefore we need an antimissile defense system. The theory of deterrence was supposed to be foolproof, or at least the best possible way to prevent communist aggression. Now we learn that it was wrong or insufficient. And so American taxpayers by implication wasted an estimated $5.8 trillion on a faulty system of mutual assured destruction and will spend billions more to build a supposedly foolproof national missile defense system.10 The drive to create a national missile defense system actually has little to do with external threats and everything to do with bureaucratic and Page 102 → scientific aspirations, prestige, and status. Could it be that the constant harping about the world as a dangerous place, which is the leitmotif of realism, conditions people to accept fictions as fact and thus to endure having their pockets picked for worthless projects? But the problem is that American/Western behavior would probably have been similar no matter what types of academic theories were in vogue. Such prescriptive problems can be linked to the most obvious and often-noted omission in realist analyses, the lack of concern with domestic politics. The rational unitary actor serves as a useful assumption for some kinds of studies, particularly those dealing with bargaining and overall government strategies, but it eliminates sensitivity to the domestic sources of foreign policy. Even during the height of the cold war those were never negligible, and today one could make the case that, at least for the United States, they have become commanding. There is little in recent decisions to stay out of the Comprehensive Nuclear Test Ban Treaty, to promote policies that antagonize most American trade partners, to stall on paying United Nations assessments, or to revise the Antiballistic Missile Treaty that smacks of national interest, rationality, or prudence. These decisions derive from American domestic politics. That is why so few governments outside the United States can do anything about them. There is no prescription deriving from realist diagnostics that raises domestically inspired aspirations, quirks, or follies to a diplomatic necessity, prudence, or virtue. Structural versions of realism simply close out anything that occurs within black boxes and contains a sole prescription, which is to balance rather than to bandwagon, but since balancing in the system occurs automatically, we need not be bothered with actual policy choices. The ultimate problem of realist prescription is that the very policies supposed to maintain international peace and security may actually lead to the unintended consequence of war. This theme is developed further in John Vasquez's contribution to this volume. Overall, then, the realist prescriptions that follow diagnosis are inadequate on numerous grounds: they are too vague and inherently subject to extremes of interpretation; they ignore the demands of domestic politics; they are overdetermined; and they are often counterproductive.

Yet we know on the basis of historical evidence that alternative prescriptive frameworks may offer little more. One thinks of the 1930s, when not one of the great powers of the era attempted to deter Hitler. Until the Polish crisis in the summer of 1939, none of Hitler's ultimate victims issued any threats; appeasement (France and England) and bandwagoning (the Soviet Union and Italy) were the preferred strategies, and they also proved to be utter failures.

Realism as a Normative Theory Throughout its history as a reasonably coherent body of thought about the problems and perplexities of international politics, realism has been at heart a normative exercise. It contains four propositions. First, it insists that the Page 103 → problems of security, war, and peace have a greater claim to examination, study, and speculation than do other problems. Why is this so? Because, second, realists believe that the ultimate political good is the safety and survival of independent political communities. And why is this the case? It is, third, as my colleague Robert Jackson has insisted, because states provide the “good life” for their citizens.11 Transnational corporations cannot do it; the Catholic Church cannot do it; the International Monetary Fund cannot do it; nor can the drug cartels. They may assist or hinder, but they cannot do the things that states do, however poorly some of them do it. If we acknowledge these as acceptable moral propositions, then it follows, fourth, that the ultimate political virtue is prudence, the quality or strategy that sustains a reasonable balance between the national interest in safety and survival on the one hand and the means by which those purposes are promoted on the other. Moralism, legalism, and ideologically driven policies, devoid of an understanding of the national interest, can only lead to the modern-day equivalent of holy wars, to a lack of tolerance for diversity in the world, and to the Munichs, Vietnams, and Afghanistans of our era. As a normative theory, realism has been on the whole more successful than as a descriptive, diagnostic, or prescriptive enterprise. It has a surface appeal that effectively transcends time, place, and personality. Its main tenets are as applicable to the diplomatic situation of 1756 or 1854 as they are today. Statements about the demise of the state notwithstanding, the independence of political communities, however defined, remains one of the most important values in contemporary civilization. The principles of self-rule and self-determination have a profound impact on the texture and practices of international politics. They are the Grundnorms of the United Nations and the basic component of that organization's commitment to “international peace and security.” Realism posits independence of political communities as the ultimate moral good, the foundation of all national interests. This position is ineffectively challenged in the postmodern and other literatures. Yet it avoids any examination of alternative types of political communities. Realism is a normative theory relevant to settled societies, a guide to those communities that are committed to political independence and that value, tolerate, and acknowledge the independence of other political communities. But how does it deal with those societies that are committed to transforming others into their own image? In the late eighteenth century, Europeans and Americans were divided between those, like Burke, who saw the French Revolution as a threat to the continued independence and distinct political arrangements of states and others, like Thomas Paine, who regarded the state as a base for promoting revolution abroad. The Bolsheviks presented the same problem, and as we see in contemporary American foreign policy, the issue has not gone away. If realism is simply a fig leaf for American ideological proselytization and lack of tolerance for political diversity, then it can have no claim to universal validity. As a moral discourse, realism assumes a value in domestic diversity, selfdetermination, and self-rule. Hegemonic, imperial, and domineering practices are to be resisted through balancing. Yet many people today claim that the values or ideologies Page 104 → of democratization, human rights, and free markets are more important than political diversity and pluralism. Indeed, a whole tradition of thinking about, for example, American foreign policy has wavered between a tolerant, pragmatic approach that takes the world as it is and a missionary zeal to convert others to America's own image. The roster of regimes toppled and statesmen assassinated or indirectly killed by the United States in its zeal for democratization includes Mosaddeq in Iran (1951), Arbenz in Guatemala (1954), Patrice Lumumba in the Congo (1961), Ngo Dinh Diem in Vietnam (1963), Salvador Allende in Chile (1973), the radical, unstable regime in Grenada (1983), Manuel Noriega in Panama (1989), and the Sandinista regime in Nicaragua (1980s). Attempts to assassinate Fidel Castro and Muammar alQaddafi and to destroy Saddam Hussein and Slobodan Milosevic through bombing have failed. This record of

intolerance stands in stark contrast to realism's cautionary prescriptions concerning prudence, the value of self-rule and self-determination, and the idea that a political community should enjoy the right to work out its political problems without external interference. On the other hand, realism also emphasizes the importance of protecting strategic interests, particularly those of the great powers. Realism has never really resolved the tension between these two types of political values when they come into conflict. Realists also acknowledge and examine other normative aspects of international politics. Hans Morgenthau in his famous text did not include chapters on international law, organization, and sovereignty just to strike them down as so many instrumental stratagems for advancing state interests. While not giving them a determinative role in international politics, Morgenthau and most “rationalists” of the so-called English School have always recognized that the human conduct we call international politics is ultimately a form of normative politics. It is an activity and discourse that involves notions of obligations, rights, legal status, and procedural and substantive norms that establish the limits of legitimate behavior. Structural realist analyses miss all of this, as do some of the cruder implicit and explicit scientific models of international politics. Earlier commentators on European diplomacy, such as Meinecke, von Gentz, Vattel, and others, have always understood that the game of international politics is a combination of self-interested and other-regarding activities. But in the last half of the twentieth century, particularly in the American academy, international law and international politics separated, much to the disadvantage of both. By divesting themselves of normative content, American versions of realism became mechanistic, deterministic, and hopelessly ahistorical.12

What Does This Leave? These comments have been highly critical of realism, perhaps leaving the impression that as a descriptive, diagnostic, and prescriptive enterprise it is empirically inadequate, logically inconsistent, and hopelessly vague. After all, the main ideas of realism were developed in Europe during the eighteenth century, and, it could be claimed, too much has changed in the texture Page 105 → of international life to warrant spending more time and money on elaboration or on extensive research programs. I cannot share this conclusion because some of the elements of realism help us understand the dynamics of some types of relationships in some areas of the world. We continue to have enduring rivalries, some states do experience significant security dilemmas, and the military dimension of interstate relations has not disappeared. There is no valid reason to abandon studies on crisis behavior, on the characteristics of rivalries, or on the theories of conflict resolution just because we have elements of globalism or the development of an increasingly networked global civil society. Despite the growing chorus of claims that we have a borderless world, that the days of the state are numbered, or that all military spending is a waste of resources, the relations between some states are still characterized by underlying fear, the compulsion to build protection by military means, and the search for outside support. Studies of alliances, the requirements of effective deterrence, or arms control, while commanding less attention in a world that seems more benign today than it did two decades ago, should not disappear, even if they should not hold the place in the field that they once did. Any course curriculum that emphasizes the latest fad to the exclusion of the realist tradition would be incomplete at best and intellectually irresponsible at worst. We know on the basis of relatively sound empirical research a great deal more (in the probabilistic sense) about the sources and dynamics of rivalries, arms races, threat perception, international crises, and war decisions than we did just one generation ago. Much of this valuable work was inspired by realist theories and hypotheses, as well as by the commanding problems and fears generated by the cold war.13 I conclude with the proposition that any understanding of past, contemporary, or future international politics is incomplete without the problem selection, the insights, the generalizations, and the explanations of realism. But any understanding of these based solely on realism will be insufficient and incomplete. In his famous analysis of world order, Hedley Bull insisted that at any given time the international system has simultaneously Hobbesian, Grotian, and Kantian characteristics.14 Realism is a perspective on politics, not the whole story. Even as a perspective, it suffers from serious descriptive, explanatory, and therefore prescriptive problems. But to pretend that it has nothing to tell us would be an immense error. In an age of academic fads, where it seems to be de rigueur to begin theoretical discourse by listing the sins and omissions of realism, and when many theorists have not even bothered to read the main texts carefully (and in some cases not read them at all), there is intellectual

peril in ignoring that which still has some important things to tell us. On the other hand, those who assume that realism is the only story in town and who insist that nothing has changed since 1756, 1914, or 1939 face the prospect of becoming sectarians rather than scientists. There are now important competitors in the field today, perhaps best expressed in the English School, in liberal institutionalism, and in constructivism. They also have important things to say, so our students need to study them. Dialogue between faiths or perspectives has been a Page 106 → characteristic of the field since its early days. Realism was in some periods a commanding perspective, but others always offered alternative visions and diagnoses. Today, the balance has shifted to those just because the texture of international politics has changed. Pessimists may look forward to the return of realist hegemony or pine for the good old days, but others, reflecting a greater mood of optimism at the turn of the millennium, hope that we have overcome at least part of the realist tradition and that the next task is to describe and explain the new textures of international politics.

Notes 1. For evidence, see K. J. Holsti, The State, War, and the State of War (Cambridge: Cambridge University Press, 1996), esp. ch. 2. 2. A realist interpretation of Asian international politics is by no means the only valid approach. Liberal institutionalist and constructivist approaches help to assess and explain a good deal of the behavior that one sees in this area. For a comparative theoretical analysis of Asian security problems, see Thomas Berger, “Set for Stability? Prospects for Conflict and Cooperation in East Asia,” Review of International Studies 26, no. 3 (2000): 405–28. 3. Stephen Krasner, Sovereignty: Organized Hypocrisy (Princeton, N.J.: Princeton University Press, 1999). 4. Cf. Ian Hurd, “Legitimacy and Authority in International Politics,” International Organization 53, no. 2 (1999): 379–408; and Martha Finnemore, National Interests in International Society (Ithaca, N.Y.: Cornell University Press, 1996). 5. Fred Halliday, The Sixth Great Power: Revolution and International Relations (Durham, N.C.: Duke University Press, 1999). 6. Robert Gilpin, War and Change in International Politics (Cambridge: Cambridge University Press, 1981). 7. Important exceptions include Richard Rosecrance, Action and Reaction in World Politics: International Systems in Perspective (Boston: Little, Brown, 1963); and Stephen Walt, Revolution and War (Ithaca, N.Y.: Cornell University Press, 1996). 8. Paul W. Schroeder, “The Nineteenth-Century System: Balance of Power or Political Equilibrium?” Review of International Studies 15, no. 2 (1989): 135–53. 9. Deutsch's seminal work is in Political Community at the International Level: Problems of Definition and Measurement (Garden City, N.Y.: Doubleday, 1954); and Karl Deutsch et al., Political Community and the North Atlantic Area (Princeton, N.J.: Princeton University Press, 1957). 10. The costs of mutual assured destruction are estimated in Stephen I. Schwartz, ed., Atomic Audit: The Costs and Consequences of U.S. Nuclear Weapons since 1940 (Washington, D.C.: Brookings Institution Press, 1998). 11. See Robert Jackson, “Martin Wight, International Theory, and the Good Life,” Millennium 19, no. 2 (1990): 261–71. 12. This theme is elaborated at great length in Robert Jackson, The Global Covenant: Human Conduct in a World of States (Oxford: Oxford University Press, 2000). 13. For an extended discussion about the connections between realist theory, the cold war, and empirical research, see K. J. Holsti, “Scholarship in an Era of Anxiety: The Study of International Politics during the Cold War,” in The Eighty Years' Crisis: International Relations, 1919–1999, ed. Tim Dunne, Michael Cox, and Ken Booth (Cambridge: Cambridge University Press, 1998), 17–46. 14. Hedley Bull, The Anarchical Society: A Study of Order in World Politics (London: Macmillan, 1977).

Page 107 →

REALISM AND THE DEMOCRATIC PEACE The Primacy of State Security in New Democracies Manus I. Midlarsky For all of the now vast amount of research done on the democratic peace, there is still a chasm in that research that has only been very partially filled. Simply put, it is the virtual absence of the international security setting. This is all the more surprising because of the growth of international security as a field and the rather obvious concern that scholars in the field should have for the international in international relations. Perhaps it is a reaction to the emphasis on polarity during the long cold war period when structural relations dominated much of the thinking in international relations. More recent emphases on rational choice of course have reinforced this reaction,1 for typically the units of analysis have at most consisted of dyads. Indeed, the dyad as a unit of analysis has virtually supplanted larger units in the investigation of international conflict. The democratic peace also is a useful “laboratory” for investigating our failure to resolve key debates in international relations theory. The democratic peace is an extremely prominent theory drawn from the liberal tradition, one that has received considerable empirical validation, perhaps more than any other.2 Yet the level of explanation has not kept pace with the success of the empirical research program. Arguably, we know little more about the actual workings of the democratic peace now than we knew several years ago.3 I contend that one reason for that failure is the larger failure to resolve our theoretical debates through syncretism. Realism has a time-honored history; one has only to mention the names of Thucydides, E. H. Carr, Hans Morgenthau, and Kenneth Waltz. Similarly, the long liberal tradition has deep roots in the writings of Immanuel Kant and other eighteenth-century European and American thinkers. Given this historical depth, there must be considerable merit in both intellectual traditions. Page 108 → When there is a failure to move forward either in resolving larger debates over the suitability of paradigms, or in furthering our understanding of major processes such as the democratic peace, there most likely is an absence of syncretism. Although the democratic peace emerges from liberal thought, I contend that realism has an important role to play in understanding the persistence of democracy, clearly the salient constituent element in the emergence of a widespread democratic peace. And the analysis goes further in indicating the primacy of geopolitical security in choices between national or state security on the one hand and highly desirable modes of governance such as democracy on the other. As a result of the analysis, the terms of the debate over which has temporal priority, peace or democracy, are shifted from an emphasis on peace in the facilitation of democracy to that of security. Following a graphical representation of the democratic peace and power transition theories in relation to realism and liberalism, an empirical test is carried out demonstrating the importance of security issues in the preservation of democracy. A variable suggested by the graphical representation, number of alliances, proves important in understanding the successes and failures of new democracies. A focus on realism and the international setting gives rise to the use of this variable; here the importance of the paradigm in selecting our variables is emphasized. Equally important is the realization that very few of our theories, if they are to be empirically valid over a significant domain,4 can be based solely on one paradigm and one setting, either domestic or international. There is another approach also suggesting that our heavy emphasis on domestic politics in matters pertaining to the democratic peace is, if not misguided, then at least incomplete. And as our theories of the peace between democracies (or republics) have found at least a partial foundation in the thinking of philosophers such as Immanuel Kant, the present emphasis on security can be found in the theoretical constructs of John Stuart Mill. As he put it, “The moral rules which forbid mankind to hurt one another (in which we must never forget to include wrongful interference with each other's freedom) are more vital to human well-being than any maxims, however important, which only point out the best mode of managing some department of human affairs.”5 According to the political philosopher John Gray, “For Mill, justice was a set of practices protecting the human interest in security,

not a theory that prescribed the structure of society.” In other words, human security, or in the international realm, national or state security,6 comes first before any claims to procedural primacy. To be sure, Mill clearly understood the importance of protecting basic freedoms, but their protection was part of the security net. It is this philosophical tradition, along with realism, that informs the present analysis.

Realism and Liberalism as Continua A source of the fairly widespread de-emphasis on the international security setting is its tendency to be associated with realism.7 Systemic structures, Page 109 → polarity, and balance of power all are uniquely international theories concerned with security; at the same time they lie at the heart of realism.8 And certainly the place of realism as the reigning paradigm of international politics has been under attack for some time. Recent titles of articles published in prestigious journals tell the story best. Legro and Moravcsik ask, “Is Anybody Still a Realist? ”, while Vasquez titles his article “The Realist Paradigm and Degenerative versus Progressive Research Programs.”9 Realists of various schools have felt obliged recently to define, modify, or defend their positions.10 Liberalism, on the other hand, has experienced a major florescence. The democratic peace has emerged as virtually emblematic of liberalism's triumph in the debate with realism. Proceeding from Kantian liberal assumptions about human needs and desires, this research program has demonstrated a spectacular growth in recent years, eclipsing virtually all others.11 In its primary concern for the welfare of the individual, liberal democracy almost tautologically embodies the precepts of liberalism. Human wants, needs, and preferences are at the heart of liberalism.12 But then there also are instances when security for the state is paramount. I identify at least three such circumstances: (1) the decision to go to war; (2) making peace; and (3) alliance formation. The first can be seen as residing in the domain of balancing threat,13 the second in the domain of the balance of power, and the third, a mixture of the two. And it is in the last of these that we will see a role for realism in understanding the workings of the democratic peace. Alliance formation is a time-honored consequence of attempts to alter or restore a balance. The formation of NATO and the Warsaw Pact in the post-World War II era is a case in point, as are the Dual Alliance and Triple Entente following one upon the other in an earlier period. The contemporary Israeli-Turkish alliance, identifying Syria, Iraq, and Iran as potential common enemies, of course is another classic instance. When national survival is at stake, “the enemy of my enemy is my friend” is a time-honored principle that even operated during World War II in the alliance between the Soviet Union and the West. Clearly, upon reviewing those scenarios, the international setting is the key variable driving the relevant decision makers. Earlier alliances influence the formation of later ones, as an emergent communist bloc in eastern Europe under Soviet rule led to a cohering Western one headed by the U.S. Realpolitik is very much driven by the structure of the international system as Waltz clearly saw.14 At the same time, liberalism, neglecting for the moment neoliberal institutions, is very much a property of domestic politics.15 The democratic peace is a centerpiece of liberal theory and theoretically emerges almost entirely from domestic political considerations. International and domestic sources of salient processes can be seen as falling on a continuum defined by the purely domestic at one end and the international at the other, with most important events falling somewhere between the two poles. Liberalism with its emphasis on individual need and realism with its focus on national security (these are my intended meanings here) also can Page 110 → be seen as occupying opposite poles on a continuum. One has at its core the welfare of the individual; the other concerns itself almost exclusively with the welfare of the collectivity, in most cases the state. Between the two poles can be found theories that are not exclusively based on one or the other paradigm but some mix of the two as we shall see shortly. Now one can plot these two axes that comprise what I call a theory space in figure 1. Theories of international politics can be located in this space, giving us a pictorial representation of our contemporary theoretical concerns. Further, controversies surrounding these theories also can be so mapped.

The Democratic Peace Consider the democratic peace. It follows from the basic tenets of liberalism and also is almost entirely domestic in its emphases. As the theory has been constituted for the most part, it relies on a particular regime, democratic in nature, not willing to wage war against another of a similar constitution. But there exists an emergent competing view that turns the theory on its head.16 According to the inverted democratic peace theory, instead of the internal political condition of democracy yielding peace between democracies, it is the condition of international peace that yields democracy. A long period of peace can effectively allow for democracy to be nurtured, essentially because of the absence of need for mobilizing leaders that eventually could institutionalize autocracies. How would a theory space help us locate these theories? figure 1 plots the democracy⇉peace (D⇉P) and peace⇉democracy (P⇉D) theories. As can be seen, the former lies within Page 111 → the liberal-domestic quadrant whereas the latter is found in the realist-international quadrant. This reversal of positions is to be expected. Now consider another theory, the power transition that in its initial formulation emphasized politics internal to the state,17 but is principally in the realist camp in the sense that it is centered on national security. Economic development internal to the state was said to be the primary cause of major-power war in contrast to the emphasis on alliances and balancing of power among states in balance-of-power theory. The two theories with their differing emphases are plotted in figure 2. Despite being in the realist camp (although elements of the power transition have liberal roots, hence it is positioned slightly to the left in the figure),18 they are located in two separate quadrants as the result of their principal loci. But subsequent variants of the theory19 needed to include alliance formation to achieve a satisfactory fit between theory and data for major-power war. More recently, Kim's treatment of power transitions between alliances even more explicitly introduces the element of alliance.20 In other words, the explicitly international element needed to be accounted for, leading to a shift in the theory's emphasis from the domestic to some mix of the domestic and international. Can the same be true for the democratic peace? Is our current near-total emphasis on domestic explanations of the democratic peace misguided to the extent that it does not include explicitly international factors? One of these emerges from the peace-democracy literature. It was hypothesized Page 112 → that the minimization of the threat of war is associated with democracy, under the expectation that the absence of necessity for a mobilizing leader21 to face the threat of war minimizes the probability of an emerging autocracy. Countries with more sea borders as a measure of the minimization of the threat of war were found to be associated with greater democracy using several measures of democracy in two studies.22 Now water barriers are not the only way that minimization of the threat of war can be achieved. Mountains as in the case of Switzerland or, more to the point here, alliances with apparent deterrent value also can serve that purpose. A more thorough understanding of the democratic peace may reside precisely in the way that alliances among democracies have minimized the threat of war thus making democracy more secure in democratizing states. In other words, peace itself may not be the variable necessary for democracy's perpetuation, but the concept of security may be paramount. (Neither the War of 1812, the American Civil War, the Finnish Civil War of 1918, the Polish-Soviet War of 1919–20, nor the several Indo-Pakistani and Indo- Chinese conflicts were sufficient to end the relevant experiments in democracy.) The experience of success in war conducted by a democracy may be proof against future successful antidemocratic coups.23 Indeed, in some circumstances, when faced with an implacable and unrelenting foe (e.g., Nazi Germany), the only means of achieving state security is the defeat of the enemy in war. The threat of war and especially a threat of the loss of newly achieved sovereignty may be far more consequential for a democracy's demise than the experience of war itself. Further, as we shall see in the “Conclusion,” democracies are especially vulnerable to a rechallenge after an earlier crisis. If the democracies are new and highly insecure because of prior military occupation, they may be especially susceptible to successful military coups. Introduction of an international security element, alliances, in both the democratic peace and power transition theories reduces the distance between competing variants as can be seen in figure 3. Security (S) replaces peace (P) at the onset of the sequence, emphasizing the salience of national or state security instead of

peace.

The Role of Alliances The Fashoda Crisis, the Spanish-American War, and the U.S. Civil War suggest that this is a productive avenue of investigation. These cases represent the modern “near misses” of democratic peace theory. All three, as well as other candidates, such as the Anglo-American dispute over Venezuela in 1895–96, were not in any way muted by the existence of alliance bonds. These countries were free to engage in military confrontations without the presence of allies of either one or the other protagonist. The function of alliance among democracies may be to magnify the trust and habits of cooperation that initially, in a small way, characterize relations between democracies. In other words the initial, fairly small tendencies toward cooperation may become magnified over time, as the protective alliance umbrella Page 113 → allows both for the maturation of democracy within the allied societies and the habits of cooperation between them. The European Union even in its earlier incarnations operated in this fashion. The case of Greece, for example, is an outstanding illustration of the choice of an international framework, the European Economic Community (EEC), to provide a protective umbrella from an aggressive Turkey that had just invaded Cyprus in 1974 and was threatening Greece on other Aegean issues. The reversion to democracy immediately thereafter under conditions of military failure and the projected future failure to protect the state can only be understood within the context of the need to have a democratic government in order to join the EEC. (Greece's application had been suspended during the 1967–74 period of military rule.) NATO already had been ruled out as a protective framework because of the deep divisions between the United States and Greece during this period. Thus, only the EEC could provide a bulwark against Turkey, and to join, Greece had to be a democracy as the most powerful military rulers themselves clearly saw, as they called for a return to civilian rule.24 Here, national security is a priority, with democracy chosen to maximize security. As one set of observers put it, “Another more pressing reason why Greece joined the community was the perceived Turkish threat especially following Turkey's invasion of Cyprus in 1974. Since neither NATO nor the US prevented the invasion, Greece had Page 114 → to seek a new framework for protecting itself from Turkey's military aggressiveness.”25 This perspective on security emerging from realism helps explain certain anomalies from the cold war period. India's perceived need for security to counter the direct Chinese threat as well as support for her arch rival Pakistan led to an ongoing informal alliance with the Soviet Union. Thus, an autocratic state was preferred as an ally to a democratic one such as the United States because of the U.S. “tilt” toward Pakistan. As an explanation of military dispute behavior, an affinity index based on UN voting, revelatory of preferences (security and otherwise), was superior in explanatory power to joint democracy.26 The affinity index likely includes security preferences as well as democratic affinities, thus providing a more complete explanation than joint democracy alone.

Democracy and Security: An Empirical Test Thus, alliance behavior is suggested by our theoretical concerns. But theoretically, it is the perception of security by alliance signatories that may be the critical variable relevant to the democratic peace. For based on substitutability criteria,27 alliances and other forms of security such as sea borders may be different expressions of national security. If one form is absent, another may be substituted for it. Here, the initial interpretation of sea borders minimizing the threat of war is expanded to encompass the concept of security. The threat of war brings with it the possibility of the loss of sovereignty in the extreme instance, especially for new states, or unfavorable boundary changes in less serious cases. In either event, national security is threatened. Geography can maximize such security, or in its absence, alliances can provide a substitute. Two specific propositions can now be investigated: (1) the greater the extent of perceived national security, the more likely the maintenance of democracy; and (2) the earlier the perceived threat to national security, the earlier the demise of democracy. Thus the existence of democracy in the first proposition and the timing of democratic collapse are both seen as functions of perceived national security. Alliances and even surrounding water may provide an illusion of

security, not its actuality. Nevertheless, given the human choices of democracy or autocracy in response to perceived threat, it is the perception that counts. In addition to the case of Greece noted earlier, can we examine other instances of the relationship between democracy and national security? The interwar period in Europe provides an excellent setting for this sort of analysis. As a result of the partial or complete breakup of empires, a number of new democracies came into existence at the end of World War I. Some survived as democracies, others failed. Hence, we have considerable variation in outcomes that allows for a comparative analysis. We also can investigate how variation in critical international conditions, namely national or state security, affected these outcomes in the success or failure of maintaining democracy. Although the term national security is commonly used, as it is in this essay, we will see later that a more precise referent is that of state Page 115 → security. How this conclusion is reached will itself be instructive. Table 1 presents all new European electoral democracies existing between 1918 and 1934 divided by success and failure. Only new states are chosen for analysis; those emanating directly from older polities without substantial boundary changes such as Germany are excluded. The focus here is on the first fifteen years or so of the new democracy's existence. After that time, the habits of democracy and indeed its matured organizational ability to defend itself may have successfully weathered serious threats to its security. Hungary and Yugoslavia (politically, an expanded Serbia) are not eligible for inclusion. The two countries were never democracies during this time period. Hungary was a regency in the expectation of the reconstitution of monarchy, and Yugoslavia was a monarchy. The year 1934 is chosen as an end point because after this time, as Germany rapidly increases its force capabilities, the Nazi-driven European political scene is vastly different from the preceding post—World War I period.28 The first data column lists the number of sea borders (minimum 0, maximum 4) per country, followed by the number of borders with major powers in the second column. The sea borders variable is composed of approximate numbers based on a four-sided model, which distinguishes between countries that can be invaded easily from land and those with greater protection from water. Numbers of alliances generally, and specifically with major powers, follow. If relevant, the date of the democracy's demise follows. Lastly, real product per capita is given as a basis for evaluating a competing explanation. Because of the small number of cases, large-N statistics are inappropriate as a form of inferential analysis. Nevertheless, inferences from a comparative analysis can be drawn as to the most important influences on the success of democracy during this period. These variables allow a preliminary test of the two propositions. It is immediately evident that among the successful democracies, a degree of substitutability is found. Ireland especially, and Finland somewhat less so, are blessed with sea borders that remove them at least partially from harm's way. Czechoslovakia, with no sea borders, simultaneously has by far the largest number of alliances generally and with major powers in particular. Indeed the top half of table 1 reveals the inverse relationship between alliances and sea borders—the one substituting for the other—among the successful democracies, a relationship not found among the unsuccessful democracies in the bottom portion of the table. Landlocked Austria, and Latvia, Lithuania, and Poland with only one sea border each, and Poland with two bordering major powers—Germany and the Soviet Union—appear to have less national security and at the same time do not compensate for that deficit with alliance behavior anywhere near comparable to that of Czechoslovakia. Estonia will be discussed later. Regarding the second proposition, the most threatened countries, Poland and Lithuania (threatened not only by Germany and the Soviet Union but also by Poland), experience the earliest failures of democracy, as can be seen on the right-hand side of the table. But these data, although generally confirmatory, do not especially allow for an analysis in depth. In order to find these propositions credible, we must Page 116 → Page 117 → also have narrative evidence consistent with these findings. In other words, to begin with, did Czechoslovakia consider her alliances to be a part of her security program, and, at the same time, was there an awareness of the possibility of the rise of an antidemocratic mobilizing leader to compensate for the absence of external security? Was there, in other words, an awareness of the implicit trade-off between external security in the form of alliances or geographical advantage on the one hand and the advantages for national security (albeit perhaps temporary) of a mobilizing leader on the other? In the absence of the external sources of security, an internal one must be chosen.

Czechoslovakia It is clear that Thomas Masaryk, the first president of Czechoslovakia who was fully committed to democracy,29 and the man who along with the first foreign minister, Edvard Benesš, set the pattern for Czechoslovak foreign policy, was indeed so aware. Regarding the latter concern about mobilizing leaders, Masaryk stated that “the early forms of society were largely fashioned by the despotic force of strong and capable leaders, so that States [sic] took on a military character and relied on the army.”30 In the context of the Czechoslovak legions of World War I that Masaryk led, he said that “the legionaries proclaimed me dictator, because in the rush of wartime it enables quick decisions to be made. The Romans always did the same. And I had the full confidence of my soldiers.”31 Masaryk understood the necessity for quick decisions in time of war or the threat of war. Such a threatening international environment can easily lead to demands for a mobilizing leader who could end a democracy. Indeed, an argument of this nature was used to end Estonia's democracy in 1934 in favor of a relatively benign autocracy in order to stave off fascist-style military dictatorship,32 as we shall see shortly. Beneš, in his 1937 funeral oration for Masaryk, stated that his philosophy and practice “in strikingly expressive manner contrasted in all their historical greatness two great moral worlds that oppose one another: Jesus—not Caesar. ”33 Caesarism with its implications of war, if not perpetual war, is to be strongly opposed. To this end, the system of alliances listed in table 1 was constructed. The basic policy of Masaryk and Beneš was to support the League of Nations in the interests of collective security. Beneš was six times chairman of the League Council, and one time president of the Assembly.34 Collective security was to be a policy bulwark of a small landlocked state, yet in early recognizing the limitations of the League, appropriate alliances were eagerly sought. In 1924 a treaty with France was signed that provided for military cooperation; in 1925 it was elevated to a full defensive alliance.35 The same year, an entente was signed with Italy. Earlier, in 1922, although not an alliance, a treaty with the Soviet Union was signed, recognizing that country.36 The Little Entente uniting Czechoslovakia with Romania and Yugoslavia became a small-power bulwark to supplement the guarantees of great powers such as France, with its implications of ultimate British support. Page 118 → At Locarno in October 1925, the first signs of insecurity for the new democracies made themselves felt. The European great powers, Great Britain, France, Germany, and Italy, guaranteed the borders of Germany with Belgium and France emanating from the Treaty of Versailles, but not the eastern ones, especially those with Poland and Czechoslovakia. As these boundaries were extremely unpopular within Germany because of the loss of formerly German territory to Poland and the large number of Sudeten Germans in Czechoslovakia, the insecurity of these two new democracies was greatly magnified. For the first time now, a threat was directed at the fledgling Czech democracy and from a formerly military, now fascist source. The chief of staff of the Czechoslovak army, General Rudolph Gajda, a commander of the Czechoslovak legion in Russia during World War I and the favorite of the extreme right, was scheduled to seize power in Prague in July of 1926. The “Falcon” (Sokol) gymnastic organization was expected to attract hundreds of thousands of people to the capital that summer. Alerted to the possibility of such a coup, Masaryk released Gajda from his duties on July 2, and the demise of Czech democracy was averted.37 Poland The other democracy threatened by Locarno was Poland, and here the outcome was very different. Instead of the continuation of democracy, there occurred on May 12–14, 1926, the seizure of power by Marshal Józef Pilsudski, the hero of the 1919–20 war with the Bolsheviks.38 His was a coup against the very system that he himself had initiated. To what can we ascribe these different outcomes in Czechoslovakia and Poland? Certainly, domestic political considerations played a role in both cases, but the differential effect of Locarno was paramount. The evidence is to be found in Pilsudski's own statements and motivations. Initially Pilsudski had no intention of governing directly. To Kraków conservatives in the winter of 1914, he stated, “I leave politics to you, and I keep the sword.”39 Not only did Pilsudski not change his mind upon

establishing the Polish state in 1920, but even after the coup of 1926, he took the office of prime minister only from October 1926 to June 1928 and again from August to December 1930; from the outset, he never relinquished control of the War Ministry.40 “His [Pilsudski's] main interests were in the army and in foreign policy.”41 To observers more interested in the domestic scene,42 it would appear as if the unstable coalitions and seeming powerlessness of the democratically elected government were sufficient to motivate the coup. But Pilsudski's relative lack of interest in domestic politics would belie this conclusion. Only the incapacity to act in time of international threat would motivate Pilsudski, and in this respect, the coalition governments were not reassuring to him. The international setting is made even more salient by the Soviet-German Neutrality Pact of April 1926.43 Not only did Locarno give the Germans a virtual green light to seek changes in Germany's eastern border with Poland, but the treaty between Germany Page 119 → and the Soviet Union now reinforced this possibility. With Soviet agreement, German troops were already training on Soviet soil in order to avoid the strict arms limitations imposed by the Treaty of Versailles. One month later, Pilsudski staged his coup. After the coup, Pilsudski, in one of his few stated aims, indicated that the principle of “balance between east and west”44 was at the center of his foreign policy. By this he meant of course that Poland had to steer a course between the two great powers to avoid “the jaws of the two colossal powers which by closing them could destroy her.”45 Put another way, Pilsudski assumed the role of a mobilizing leader who left control over much of the political scene to civilian politicians but maintained control over virtually all important aspects of the armed forces. Now we see the fundamental difference in geopolitical circumstances that led to these decidedly different outcomes. Beneš could describe the Locarno Pact as “a tremendous advance on anything that has as yet been accomplished”46 because despite the pact's limitations, he believed that Franco-German reconciliation was beneficial for Czechoslovakia, especially as Britain was now part of the agreements. Possible German expansion appeared to threaten Poland, not Czechoslovakia, because, as yet, Germany had no irredentist claims against Czechoslovakia, only Poland. (The Sudetenland had been part of Austria-Hungary, not Germany.) Although not quite as secure as she had been prior to Locarno, Czechoslovakia still did not feel especially threatened; hence the perceived need for a mobilizing leader and support for a coup would be correspondingly less than that found in Poland, sharing two borders with cooperating major powers and holding the former territory claimed by one of them.47 Clearly the Polish international position was substantially more threatened than that of Czechoslovakia in the immediate post-Locarno and post-Soviet-German rapprochement periods. The Baltics and Finland Consider now the impact of these events on smaller neighboring countries. Lithuania, geographically closest to Poland, had seen her past and future capital, Vilnius, forcibly annexed to Poland, and could not remain unaffected. Very soon after the Polish coup of May 1926, Lithuania too experienced the overthrow of her democratically elected government in December 1926. From the perspective of aggrieved and tiny Lithuania, Poland was the equivalent of a great power, or at least she had acted like one in the military takeover of Vilnius. Lithuania's other border was with the Soviet Union, which had troops fighting Lithuanian nationalists during the Russian Civil War period. A military government in Warsaw must have been seen as terribly threatening to Lithuania, and the response was identical to that of Poland when faced with a similar threat from Germany. A government capable of rapid mobilization in such a threatening environment was preferable to the continuation of a potentially weak and vacillating democracy. Security trumps democracy as we saw in the case of Poland and earlier in Page 120 → the instance of Greece. Both Eidintas and Vardys and Sedaitis acknowledge the importance of the Polish model for the Lithuanian coup.48 The remaining Baltic states present a somewhat different although still threatening scenario. Hitler's rise to power had now distinctly made itself felt in eastern Europe, although his intentions were not yet clear. Instead of antagonizing Hitler, in January 1934 the Pilsudski government signed a nonaggression pact with Nazi Germany. French, and by implication, British influence would be correspondingly diminished. Indeed, future cooperation between Germany and Poland could yield a corridor for German troops to once again invade and occupy the Baltic region, as they had during World War I. And Estonia had a model for a mobilizing response in the Lapua

Movement in Finland. The origins of this Finnish threat to democracy are to be found in the appearance in the fall of 1929 of red-shirted and red-scarfed communist youth of the Young Communist League sporting Russian symbols in the town of Lapua. The extreme right-wing response to this local phenomenon is known as the Lapua Movement. Despite the reduction of the Soviet armed forces from a high of five million during the Russian Civil War to less than eight hundred thousand by 1922, Soviet Russia was still seen as a threat to Finnish independence, albeit the only such threat.49 Local reaction to this meeting of the Young Communist League was violent even to the point of nearly wrecking the building in which the meeting took place. Leaders of the anticommunist movement such as General Wallenius, Captain Kalsta, and several Lutheran pastors broadened their objectives to stamp out communism, secularism, atheism, and other “ungodly” faiths throughout Finland.50 In 1930, this organization was put on a firm footing, beginning now a virtual crusade against communism. Members of the government began to cooperate with the movement, perhaps largely because of substantial peasant support for the goals of the movement. Communist members of parliament were arrested by then President Svinhufvud, and legislation was contemplated to further limit civil liberties of the left. But it was only when the Lapua Movement began to organize military units headed by General Wallenius that the government reacted against the Lapua. In 1932, the government ordered disarmament of these units, and General Wallenius was placed under house arrest, thus ending the period of danger. Anatole Mazour, the esteemed historian of Russian as well as northern and eastern Europe, cites “the influence of ‘the neighbor from the east’ in producing this sudden rash on the healthy body of the Finnish public.”51 But, in point of fact, as we know, the Soviet Union was not a threat to the Finns at this time. In addition to the drastic reduction in size of the Soviet armed forces, the Soviet Union was engaged in an effort to normalize relations with the noncommunist world, culminating in admission to the League of Nations in 1934. The 1932 nonaggression treaty between the Soviet Union and Finland is to be seen in this light. Thus from every realpolitik perspective, the Soviet Union could not be considered threatening at this time; it was this reality that allowed the government to move against elements Page 121 → of the military who at the moment were not necessary for national defense and at the same time might have initiated a military coup. The outcome for Estonia was different, principally as the result of the German threat. Whereas Finland had good relations with Germany partly as a consequence of German military intervention on behalf of the government against the Reds in the Finnish Civil War, Estonia had recent memories of German occupation forces on her soil during World War I, as well as the centuries-long domination of German landowners. There existed a very real fear of a renewed German occupation and potential colonization. The rise of an aggressive Nazi Germany in 1933 was very threatening indeed. The whole issue of irredentism could be raised again as it once was as early as the 1871 unification of Germany.52 The 1934 German-Polish treaty made such aggressive ventures more likely, in the Estonian view. Thus when a Lapua-like League of Veterans movement entered the political arena, the result was not a continuation of democracy as in Finland, but a jettisoning of Estonian democracy. After the veterans became increasingly powerful, the president, Konstantin Päts, declared a state of martial law on March 12, 1934, for the next six months and appointed a war hero, General Laidoner, as commander in chief of the armed forces. Päts shut down the League of Veterans organizations and arrested four hundred of their leading members. In this fashion, a fascist threat to the state was countered, but the “cumbersome” Estonian democracy was abandoned in favor of a dictatorship, which, although benevolent in many ways, was autocratic and militarized. As Raun put it, “With the rise of totalitarian powers—an expansionist Nazi Germany and Stalinist Russia—on either side of Estonia, it appears that much of the Estonian population countenanced a departure from democracy in order to insure domestic stability and ward off the threat of foreign intervention.” 53 Soon after this coup, in May 1934 Latvia too experienced a coup stimulated by many of the same fears of Germany and perhaps a resuscitated Russia, as well as modeling after the Estonian prototype.54 As Kirby put it, “Whereas Finland lay on the periphery—albeit an important one—of the Russian security zone, the Baltic states were right in the centre, between Russia and Germany. Furthermore, the southernmost state, Lithuania, was in conflict with Poland over Vilna.”55

Here we see the critical distinction between Finland and the Baltic states. Moreover, in addition to this geopolitical disadvantage for the Baltics, an important difference between the successful democracies in table 1 and the unsuccessful ones is, with only one exception to be discussed later, that in all remaining cases, the latter countries had experienced invading or occupying German and Russian troops on their soil in recent years. Under these circumstances a newly aggressive Germany constituted an immediate and salient threat requiring some form of mobilized response. This condition is absent in all instances of the successful democracies. No other factor so clearly distinguishes between the successful and unsuccessful democracies in table 1. Page 122 → Ireland Turning now to a seemingly “easy” case, that of Ireland, the reality is more complex than it first appears. Despite the heritage of British parliamentarism, there were serious sources of autocratic government. The “hard men,” or those who had fought at the head of the most combat-hardened units, refused to countenance the peace treaty with Great Britain. According to Garvin, “The hard core of the old IRA had a strong collective loyalty and thought of themselves as sovereign. They had organised themselves and armed themselves and paid their own way. They saw themselves as having created the Republic, and none had the right to give it away, democratically or otherwise. The superiority of the brave guerrilla soldier over the average citizen was a key theme in anti-Treaty mentality.”56 Thus, strong antidemocratic sentiments pervaded the ranks of the best fighters, yet Irish democracy survived. Even more so than in Finland, we can see that geopolitical circumstances played a major role. Had there been the need for large-scale defense forces, the outcome might have been different. But “the underlying assumption of all these governments was that Ireland could not and need not defend itself, so that its army was a mere symbol of sovereignty. The threat from Britain was discounted, and Ireland's geographical position ensured that no other power could threaten it.”57 As a consequence, there was never any need for a large military presence potentially dominated by the antidemocratic leaders of the old IRA. Even during World War II, when Ireland maintained a strict neutrality for fear of antagonizing Nazi Germany, she could still make use of the essentially good relations with Britain. Ireland was entirely dependent on her larger neighbor for air and sea defense; the extent of covert cooperation between the secret services of the two countries was considerable.58 In other words, the number of sea borders and presence of a protecting major power obviated the need for many alliances to protect the nascent democracy. Austria Let us now consider the one apparent exception, Austria, to the finding here of invading or occupying German and Russian troops in recent memory on the territories of the unsuccessful democracies. This exception will lead us now to the distinction between state and society in matters of national security. Indeed, the appropriate referent for this analysis may not be the nation as in national security, but the state as in state security. Although an electoral democracy since the breakup of the old Austro-Hungarian Empire and the promulgation of its constitution in November 1918, by 1933 Austria was a dictatorship governed by Engelbert Dollfuss. This transition reveals the importance of the state. Initially, few German speakers within the territory soon to be the republic of Austria wanted to have a separate republic. Indeed, Article 2 of the 1918 constitution stated explicitly that German Austria is part of the German republic. “This declaration of loyalty to a German identity and to the union with the German Page 123 → republic was common to all political persuasions.”59 But the Allies at Versailles sought to limit German expansion not endorse it, and so they insisted on the permanence of the Austrian state despite the wishes of the vast majority of its population. The state-society discontinuity therefore was present from the outset. With the rise of Hitler, pan-Germanism received a new impetus. Nazi sentiment or at least the desire for the longdelayed union with Germany was increasing rapidly. Germany now constituted a major threat to the future existence of the Austrian state, although certainly not its society. Because of the virtual absence of societal

support, an Austrian military mobilization in response to this threat was out of the question. But Dollfuss by seizing power in 1933 could play the international card by enlisting the support of at least one great power in preventing the impending Anschluss. Thus, in March 1934, an alliance between Austria and Italy was signed. Mussolini had shown an early hostility to Hitler that Dollfuss felt he could draw upon in support of his position. Of course, it was unsuccessful, but here we see once again, as in the case of Czechoslovakia, the use of alliances as a source of international security, but this time very late in the day, and in support of the state instead of the nation. In contrast to Czechoslovakia, only by seizing power and effectively denying the wishes of the national majority could Dollfuss hope to perpetuate the existence of the Austrian state in alliance with Italy.

Competing Explanations Clearly, there are other possible sources of the difference in outcomes between the successful and unsuccessful democracies in table 1. Perhaps the most time-honored such explanation is that of economic development. Pioneered by scholars such as Lipset and Dahl,60 it has been found to be widely applicable, yet it does not provide as strong a distinction between successful and unsuccessful democracies as does state security. The values of real product per head given in table 1 are revealing. Austria, with the highest such value is an unsuccessful democracy, whereas Czechoslovakia and Finland, while successful, are substantially below Austria. Ireland demonstrates a value close to but below that of Austria. Despite the absence of comparable data for the Baltics in table 1, from different textual sources, Raun observes that “in 1927–1928, per capita income in Estonia lagged far behind Western Europe and much of Scandinavia, but the level was close to that in Finland and considerably higher than in Poland and Lithuania.”61 Specifically targeting Latvia, Dreifelds tells us that “on the whole, Latvia's achievements were on the level of those of Finland.”62 Thus economic development cannot differentiate between Finland's success and two of the Baltic states that were unsuccessful democracies, while Austria, an unsuccessful democracy, had a higher per capita income than all three of the successful democracies. Even taking into account Przeworski and Limongi's finding of the greater likelihood of democracy's survival in wealthier states,63 it still does not explain the Austrian Page 124 → failure and the Finnish success in comparison with economic peer states. Neither does the worldwide depression help distinguish between the successful and unsuccessful democracies because it affected all European states.64 The civic culture or culture of democracy has been suggested by Putnam,65 among others, as a strong indicator of success in several areas, including democracy and economic development. Yet here too, countries with similar (although certainly not identical) democratic experiences have experienced different outcomes in retaining democracy. Austria, Czechoslovakia, and parts of Poland were all components of the Austro-Hungarian Empire with its tradition of elections to the Reichsrat, albeit under monarchic dominance. Austria itself actually had universal suffrage as of 1907. The Baltic states, Finland, and parts of Poland were components of the czarist empire with its then recent (1906) institution of the Duma, yet they also experienced different outcomes. In the case of Finland it may be argued that prior to the twentieth century and czarist attempts at Russification, Finland experienced considerable autonomy in the form of its Diet, but it was an archaic institution called “antiquated and unrepresentative,” and “more than two thirds of the Finnish people were without the vote at national or local level.”66 Moreover, when mass mobilization did begin in the first decade of the twentieth century, the fastest growing party was the Finnish Social Democrats, which had a distinctly revolutionary Marxist and antidemocratic tone.67 Indeed this party among others of the left provided a base for the Reds in their opposition to the Whites during the Finnish Civil War of 1918.68 During this civil war, the country was equally divided, and indeed the Reds could actually field a larger army than the Whites. The Red forces numbered one hundred thousand to one hundred forty thousand while the Whites mustered at most approximately seventy thousand. The death toll was about thirty thousand, or 1 percent of the population.69 This sort of polarization and civil strife is hardly the basis upon which to establish a stable democracy, but without a direct threat to its sovereignty for much of the interwar period, Finnish democracy could survive.70 Finally, we consider the role of identity conflicts in relation to democracy. As Horowitz suggests, democracy may

suffer because of the existence of multiple identities within the polity.71 Yet here too, we see that such a potential rival explanation cuts across the successful and unsuccessful democracies. Austria was extremely homogeneous yet failed as a democracy, while Czechoslovakia was extraordinarily heterogeneous. Religion was a salient identity for largely Roman Catholic Ireland, but Ireland also had a significant Protestant minority. Heterogeneous Poland with large Ukrainian, Jewish, and German minorities failed as a democracy, but Finland with an 11 percent Swedish minority and Czechoslovakia with a 24 percent dissident German minority succeeded.72 Here too, then, the rival domestic explanation is not as compelling as the international security setting. Page 125 →

Conclusion Among the several possible causes of democracy's success in Europe during the interwar period, the only one that consistently distinguishes between the successes and failures of new democracies is state security. From a realist perspective, this is not surprising, for security, as John Stuart Mill clearly saw, is primary. When democracies are threatened, especially if democracy has not yet been firmly embedded within the polity, it can be jettisoned in favor of a mobilizing leader. Why do these new democracies perceive themselves to be vulnerable? The answer appears to reside in the reality of their political environment. An especially disturbing element for these democracies is the tendency to be rechallenged after an initial crisis. As Grieco found, “controlling for a number of systemic-level factors, and regardless of how we operationalize those systemic factors, democratic defenders appear to be at a higher risk of being re-challenged after an initial crisis than nondemocratic defenders.”73 Although this pattern appeared more strongly in the lower-intensity crises, it is still obviously a disturbing aspect of the environment of a new state. The number of new leaders or revolving door governments may be a particular source of that vulnerability.74 European new democracies during the interwar period were especially prone to that form of instability. Eliminating the potential source of that vulnerability, namely democracy, is a possible, even preferred, solution for states without long democratic experience. Thus, state security, a variable suggested by realist thought, is important in understanding the survival or demise of new democracies. A more complete understanding of the conditions surrounding the democratic peace, especially its perpetuation, depends very much on the international security setting of new democracies. And if a theory so emergent from liberalism and domestic politics can be subject to a realist analysis, then much of our theorizing in international relations should take into account the contributions of more than one paradigm as well as the domestic and international settings. Here is the syncretism necessary not only for understanding the workings of the democratic peace but also for the larger issue of resolving disputes over paradigm applicability. And if a syncretic approach can be successfully applied to a theory as important as the democratic peace, then it may be applicable to the larger issues of the “paradigm wars” as well.

Notes This is a revised and expanded version of a paper prepared for the millennial reflections panel on realism held at the annual convention of the International Studies Association, Los Angeles, Calif., March 14–18, 2000. I am grateful to Joseph Grieco, Patrick James, Jan Kubik, and James Ray for their insightful comments on the earlier version. 1. Bruce Bueno de Mesquita, James D. Morrow, Randolph M. Siverson, and Alastair Smith, “An Institutional Explanation of the Democratic Peace,” American Political Science Review 93 (December

1999): 791–807. 2. Manus I. Midlarsky, “Mature Theories, Second Order Properties, and Other Matters,” in What Do We Know about War? ed. John A. Vasquez (Lanham, Md.: Rowman and Littlefield, 2000), 329–34. 3. Robert Jervis, “Realism in the Study of World Politics,” International Organization 52 (autumn 1998): 972.Page 126 → 4. In an operational sense, a significant domain refers to a nontrivial percentage of the variance explained in a systematic quantitative analysis. But even more important can be the added inclusiveness and expansion of the theory achieved by incorporating elements of more than one paradigm and setting. 5. John Stuart Mill, Utilitarianism (1861; New York: Oxford University Press, 1998), 103. 6. John Gray, “The Light of Other Minds,” Times Literary Supplement, no. 5054, February 11, 2000, p. 12. It appears that Mill's concern for human security is consistent with his overall desire to maximize happiness, as Kant's peace between republics is consistent with his emphasis on rules, here expressed as rules of government. As a major moral philosopher put it, “Kant's view of morality, then, is quite different from those of both Aristotle and Mill. Rather than basing it on the virtues, or on the maximization of happiness, Kant sees morality as constituted by certain rules.” Roger Crisp, editor's introduction to Utilitarianism, by John Stuart Mill (1861; New York: Oxford University Press, 1998), 7; emphasis in original. 7. Although this certainly has been true in the development of international relations theory, analytically it does not have to be the case. A concern for security, a major property of realist thought, can emerge from the domestic environment. Armaments manufactured or procured by an autocratic government to maintain its security against domestic democratic opposition may trigger the security dilemma with neighboring countries that suspect the armaments are for international use. This problem would be especially severe in Third World environments where offensive and defensive weapons are not easily distinguishable from each other because of their relatively unsophisticated levels. To this extent, the two axes of realism-liberalism and domestic-international to be introduced in figures 1 through 3 can be considered to be orthogonal. 8. The place of national or state security at the heart of realism cannot be doubted. For developments of this concept in relation to realism see, for example, John H. Herz, “Idealist Internationalism and the Security Dilemma,” World Politics 2 (January 1950): 157–80; Robert Jervis, “Cooperation under the Security Dilemma,” World Politics 30 (January 1978): 167–213; Jack L. Snyder, “Perceptions of the Security Dilemma in 1914,” in Psychology and Deterrence, ed. Robert Jervis, Richard N. Lebow, and Janice Gross Stein (Baltimore: Johns Hopkins University Press, 1985), 153–79; and Charles L. Glaser, “The Security Dilemma Revisited,” World Politics 50 (October 1997): 171–201. 9. Jeffrey W. Legro and Andrew Moravcsik, “Is Anybody Still a Realist?” International Security 24 (fall 1999): 5–55; John A. Vasquez, “The Realist Paradigm and Degenerative versus Progressive Research Programs: An Appraisal of Neo-traditional Research on Waltz's Balancing Proposition,” American Political Science Review 91 (December 1997): 899–912. 10. Thomas J. Christensen and Jack L. Snyder, “Progressive Research on Degenerate Alliances,” American Political Science Review 91 (December 1997): 919–22; Colin Elman and Miriam F. Elman, “Lakatos and Neorealism: A Reply to Vasquez,” American Political Science Review 91 (December 1997): 923–26; Jervis, “Realism in the Study of World Politics”; John J. Mearsheimer, Great Power Politics (New York: Norton, 2001); Randall L. Schweller, “New Realist Research on Alliances: Refining, Not Refuting, Waltz's Balancing Proposition,” American Political Science Review 91 (December 1997): 927–30.; Stephen M. Walt, “The Progressive Power of Realism,” American Political Science Review 91 (December 1997): 931–35; and Kenneth N. Waltz, “Evaluating Theories,” American Political Science Review 91 (December 1997): 913–17. 11. Michael W. Doyle, “Liberalism and World Politics,” American Political Science Review 80 (December 1986): 1151–69; Bruce M. Russett and Harvey Starr, “From Democratic Peace to Kantian Peace: Democracy and Conflict in the International System,” in Handbook of War Studies II, ed. Manus I. Midlarsky (Ann Arbor: University of Michigan Press, 2000), 93–128. 12. Joseph M. Grieco, “Anarchy and the Limits of Cooperation: A Realist Critique of Neoliberal Institutionalism,” International Organization 42 (summer 1988): Page 127 → 485–507; Ole R. Holsti, “Theories of International Relations and Foreign Policy: Realism and Its Challengers,” in Controversies in International Relations Theory: Realism and the Neoliberal Challenge, ed. Charles W. Kegley Jr. (New York: St. Martin's Press, 1995), 35–65; Andrew Moravcsik, “Taking Preferences Seriously: A Liberal

Theory of International Politics,” International Organization 51 (autumn 1997): 513–53. 13. Stephen M. Walt, The Origins of Alliances (Ithaca, N.Y.: Cornell University Press, 1987). 14. Kenneth N. Waltz, Theory of International Politics (Reading, Mass.: Addison-Wesley, 1979). 15. Robert O. Keohane, After Hegemony: Cooperation and Discord in the World Political Economy (Princeton, N.J.: Princeton University Press, 1984). 16. Manus I. Midlarsky, “Environmental Influences on Democracy: Aridity, Warfare, and a Reversal of the Causal Arrow,” Journal of Conflict Resolution 39 (June 1995): 224–62; Joanne Gowa, “Democratic States and International Disputes,” International Organization 49 (summer 1995): 511–22; William R. Thompson, “Democracy and Peace: Putting the Cart before the Horse?” International Organization 50 (winter 1996): 141–74; Patrick James, Eric Solbert, and Murray Wolfson, “An Identified Systemic Model of the Democracy-Peace Nexus,” Defence and Peace Economics 10, no. 1 (1999): 1–37; Murray Wolfson, Patrick James, and Eric Solberg, “In a World of Cannibals Everyone Votes for War: Democracy and Peace Reconsidered,” in The Political Economy of War and Peace, ed. Murray Wolfson (Boston: Kluwer, 1998), 155–76. For debate on the precedence of democracy or peace, see John R. Oneal and Bruce Russett, “Comment: Why ‘An Identified Systemic Model of the Democracy-Peace Nexus’ Does Not Persuade,” Defence and Peace Economics 11, no. 2 (2000): 197–214; and Patrick James, Eric Solberg, and Murray Wolfson, “Democracy and Peace: Reply to Oneal and Russett,” Defence and Peace Economics 11, no. 2 (2000): 215–29. 17. A. F. K. Organski, World Politics (New York: Knopf, 1958). 18. Here we see a role for liberalism in understanding elements of the power transition; hence it is not positioned at the extreme right of the liberalism-realism continuum. It was economic liberalism of the late eighteenth and nineteenth centuries that was associated with the rise of industrialization, not initially as a base for the power transition but as a way of satisfying the human need for labor-saving devices and ultimately a more comfortable work environment. Only later when the industrialization process was well under way could there be the direct application to increasing force capability and the threats to other powers indicated in the power transition. Interestingly, the major challengers in the power transition have been challengers of the liberal economic order established by the British and French (challenger: Germany) and later by the United States and its allies (challenger: Soviet Union and perhaps now China). One may interpret President Clinton's policies toward China as efforts to incorporate China into that economic order as a means of avoiding another potential challenge. Although Kim found power transition theory to apply to alliances prior to the industrial period, this finding does not reflect on the domestic economic and political changes posited in the theory's original formulation. See Kim Woosang, “Power Transitions and Great Power War from Westphalia to Waterloo,” World Politics 45 (October 1992): 153–72. 19. For example, A. F. K. Organski and Jacek Kugler, The War Ledger (Chicago: University of Chicago Press, 1980). 20. Kim, “Power Transitions and Great Power War.” For a recent review of this literature, see Jacek Kugler and Douglas Lemke, “The Power Transition Research Program: Assessing Theoretical and Empirical Advances,” in Midlarsky, Handbook of War Studies II, 129–63. 21. A mobilizing leader here refers to a leader with the capacity for military mobilization, not social mobilization. Thus, present or former military leaders are probably the best candidates to usurp democratic authority. 22. Midlarsky, “Environmental Influences on Democracy”; Manus I. Midlarsky, “Democracy and Islam: Implications for Civilizational Conflict and the Democratic Page 128 → Peace,” International Studies Quarterly 42 (September 1998): 485–511; Manus I. Midlarsky, The Evolution of Inequality: War, State Survival, and Democracy in Comparative Perspective (Stanford, Calif.: Stanford University Press, 1999). 23. The demise of Polish democracy in 1926 even after the victory over the Bolsheviks in 1920 might seem to belie this conclusion. Although it is a counter-illustration, the Polish case suggests that if there is sufficient threat, as in two great powers simultaneously threatening Poland on her borders—Germany and the USSR—and with no strong allies in the vicinity, even prior success in war by a democracy may not be sufficient to prevent a military coup. For evidence on the positive effect of war on democratization, see Michael Mousseau and Yuhang Shi, “A Test for Reverse Causality in the Democratic Peace Relationship,” Journal of Peace Research 36 (November 1999): 639–63; and Sara M. Mitchell, Scott Gates, and Havard Hegre, “Evolution in Democracy-War Dynamics,” Journal of Conflict Resolution 43 (December 1999):

771–92. 24. Richard Clogg, A Short History of Modern Greece (Cambridge: Cambridge University Press, 1979), 179. 25. Yorgos A. Kourvetaris and Betty A. Dobratz, A Profile of Modern Greece in Search of Identity (Oxford: Clarendon, 1987), 112, emphasis added. 26. Erik Gartzke, “Kant We All Just Get Along? Opportunity, Willingness, and the Origins of the Democratic Peace,” American Journal of Political Science 42 (January 1998): 1–27; and Erik Gartzke, “Preferences and the Democratic Peace,” International Studies Quarterly 44 (June 2000): 191–212. 27. For comprehensive treatments of substitutability in foreign policy, see the special issue of the Journal of Conflict Resolution 44 (February 2000) devoted to that subject. See also Benjamin A. Most and Harvey Starr, “International Relations Theory, Foreign Policy Substitutability, and ‘Nice’ Laws,” World Politics 36 (September 1984): 383–406. 28. The year 1934 was chosen because it is the last year prior to the recognition by the international community of Germany's potentially growing military might. In 1935, for example, the Anglo-German Naval Agreement was signed limiting German naval tonnage to 35 percent of the British, thus signaling efforts to contain and ultimately appease German expansionism. After this time, Germany's military development proceeded apace leading to substantial changes in European and ultimately global international relations. Prior to this time, the European interstate system is without major discontinuities. See Hans J. Morgenthau, Politics among Nations: The Struggle for Power and Peace, 5th ed. (New York: Knopf, 1973), 391. 29. For the domestic sources of Czechoslovak democracy, see Edward Táborský, “The Roots of Czechoslovak Democracy,” in Czechoslovakia Past and Present, vol. 1, Political, International, Social, and Economic Aspects, ed. Miloslav Rechcigl Jr. (The Hague: Mouton, 1968), 117–23. 30. Thomas G. Masaryk, The Making of a State: Memories and Observations, 1914–1918 (New York: Frederick A. Stokes, 1927), 463. 31. Emil Ludwig, Defender of Democracy: Masaryk of Czechoslovakia (New York: McBride, 1936), 149. 32. Toivo U. Raun, Estonia and the Estonians (Stanford, Calif.: Hoover Institution Press, 1987). 33. Edvard Beneš,Masaryk's Path and Legacy (New York: Arno Press and The New York Times, 1971), 27, emphasis in original. 34. As Beneš stated, “On the one side, we were always at work in the interest of peace—ready to co-operate with our adversaries, to come to terms with them in a friendly manner and, where it was necessary, to make such concessions as occasion demanded to the defeated, so that they could be reconciled with the new international order. On the other side, we made all efforts to build up guarantees of collective security, and to strengthen the League of Nations, but at the same time, we prepared for defence in case of a conflict.” Edvard Beneš,Memoirs of Dr. Eduard Beneš: From Munich to New War and New Victory, trans. Godfrey Lias (Boston: Houghton Mifflin, 1954), 3; See Piotr S. Wandycz, “Foreign Policy of Edvard Beneš, 1918–1938,” in A History of the Czechoslovak Page 129 → Republic: 1918–1948, ed. Victor S. Mamatey and Radomir Luža (Princeton, N.J.: Princeton University Press, 1973), 216–38. 35. Ibid., 224. 36. Věra Olivová, The Doomed Democracy: Czechoslovakia in a Disrupted Europe, 1914–38, trans. George Theiner (London: Sidgwick and Jackson, 1972), 145. 37. See Victor S. Mamatey, “The Development of Czechoslovak Democracy, 1920–1938,” in Mamatey and Luža, A History of the Czechoslovak Republic, 9–166. One can ask why Czechoslovak democracy did not break down in 1938–39 when it was under extreme stress due to threats of German expansion. Two answers present themselves. First, the twenty-year period of democracy may have adequately socialized the Czechoslovak public in the democratic ethos. But likely far more important is the fact that Masaryk, the Czech president until 1935, had been the titular head of the Czech Legion in Siberia at the end of World War I that successfully fought its way across Russia against Bolshevik opposition. Further, later under Beneš, Czechoslovakia had two million men under arms. Its defense budget was more than one-third of the total government expenditure. Antonin Basch, “Economic and Financial Policy of Czechoslovakia, 1918–1938,” in Rechcigl Jr., Czechoslovakia Past and Present, vol. 1, 168. In other words, the democratic leadership had successfully mobilized the polity to face the eventuality of war. No replacement by an autocracy therefore was perceived to be necessary.

38. Norman Davies, God's Playground: A History of Poland, vol. 2, 1795 to the Present (New York: Columbia University Press, 1982), 421. 39. Piotr S. Wandycz, Polish Diplomacy, 1914–1945: Aims and Achievements (London: Orbis Books for the School of Slavonic and East European Studies, 1988), 8. 40. R. F. Leslie, Antony Polonsky, Jan M. Ciechanowski, and Z. A. Pelczynski, The History of Poland since 1863 (Cambridge: Cambridge University Press, 1980), 161. 41. Ibid., 159. 42. For example, Davies, God's Playground, 421. 43. Leslie et al., The History of Poland since 1863. 44. Wandycz, Polish Diplomacy, 1914–1945, 20. Also see Anna M. Cienciala and Titus Komarnicki, From Versailles to Locarno: Keys to Polish Foreign Policy, 1919–25 (Lawrence: University Press of Kansas, 1984), 271. 45. Wandycz, Polish Diplomacy, 1914–1945, 18. 46. Quoted in Wandycz, “Foreign Policy of Edvard Beneš, 1918–1938,” 225. 47. Because of the 1921 Peace of Riga recognizing Polish borders, the Soviet Union had no formal claim on Polish territory. However, much of the territory in the East had Ukrainian majorities that had been conquered by Poland in the 1919–20 war with the Soviet Union. That territory was later incorporated into the Soviet Union at the end of World War II. 48. Alfonsas Eidintas, “The Nation Creates Its State,” in Lithuania in European Politics: The Years of the First Republic, 1918–1940, ed. Alfonsas Eidintas, Vytautas Zalys, Alfred E. Senn, and Edvardas Tuskenis (New York: St. Martin's Press, 1997), 53; and V. Stanley Vardys and Judith B. Sedaitis, Lithuania: The Rebel Nation (Boulder, Colo.: Westview, 1997), 36. 49. D. G. Kirby, Finland in the Twentieth Century (Minneapolis: University of Minnesota Press, 1979), 114. 50. Anatole G. Mazour, Finland: Between East and West (Princeton, N.J.: Van Nostrand, 1956), 84. 51. Ibid., 87. 52. Raun, Estonia and the Estonians, 62. 53. Ibid., 123. 54. Juris Dreifelds, Latvia in Transition (Cambridge: Cambridge University Press, 1996). 55. Kirby, Finland in the Twentieth Century, 111. 56. Tom Garvin, 1922: The Birth of Irish Democracy (New York: St. Martin's Press, 1996), 31.Page 130 → 57. Charles Townshend, “The First Peace Process,” Times Literary Supplement, no. 5049, January 7, 2000, p. 22. 58. Eunan O'Halpin, Defending Ireland: The Irish State and Its Enemies since 1922 (New York: Oxford University Press, 1999). 59. Fritz Fellner, “Introduction: The Genesis of the Austrian Republic,” in Modern Austria, ed. Kurt Steiner, Fritz Fellner, and Hubert Feichtlbauer (Palo Alto, Calif.: Society for the Promotion of Science and Scholarship, 1981), 9. Rendering a more contemporary view, “‘We convinced ourselves we'd won freedom by being the first victims,’ says Oscar Bronner, the publisher of Der Standard, a Vienna daily, ‘But we loved the German occupation. In fact, the real occupation here is widely seen as having ended in 1955, when the Allies left!’” Roger Cohen, “A Haider in Their Future,” New York Times Magazine, April 30, 2000, sec. 6, p. 57. 60. Seymour M. Lipset, “Some Social Requisites of Democracy: Economic Development and Political Legitimacy,” American Political Science Review 53 (March 1959): 69–105; and Robert A. Dahl, Polyarchy: Participation and Opposition (New Haven, Conn.: Yale University Press, 1971). 61. Raun, Estonia and the Estonians, 132. 62. Dreifelds, Latvia in Transition, 27. His source for this statement is broad-scale economic data (but not real product) for the Baltics and Finland. Juris Dreifelds, “Belorussia and the Baltic Republics,” in Economics of Soviet Regions, ed. I. S. Korpeckyj and Gertrude E. Schroeder (New York: Praeger, 1981), 364. 63. Adam Przeworski and Fernando Limongi, “Modernization: Theories and Facts,” World Politics 49 (October 1997): 155–83. 64. Other scholars also find the depression alone insufficient in explaining polity revival or collapse. See,

for example, Ekkart Zimmermann and Thomas Saalfeld, “Economic and Political Reactions to the World Economic Crisis of the 1930s in Six European Countries,” International Studies Quarterly 32 (September 1988): 305–34. Czechoslovakia, an industrial state, was especially vulnerable to the great depression yet persisted as a democracy. Industrial exports, a key basis of economic productivity, were particularly hard hit by the slump in international trade. Zora P. Pryor, “Czechoslovak Economic Development in the Interwar Period,” in Mamatey and Luža, A History of the Czechoslovak Republic, 189, 204. 65. Robert D. Putnam, Making Democracy Work: Civic Traditions in Modern Italy (Princeton, N.J.: Princeton University Press, 1993). 66. Kirby, Finland in the Twentieth Century, 23. 67. Ibid., 19. 68. Eino Jutikkala, A History of Finland, trans. Paul Sjöblom (New York: Praeger, 1962). 69. Eric Solsten and Sandra W. Meditz, Finland: A Country Study (Washington, D.C.: Federal Research Division, Library of Congress, 1990), 30. 70. As in the case of Czechoslovakia, one can ask why Finnish democracy survived the later threat of war that would ultimately become the Soviet-Finnish Winter War of 1939–40. Here too, similar arguments are applicable, especially prior success in war as a democracy—the Finnish Civil War of 1918 involving Finnish nationalist Whites against Soviet-supported Reds. 71. Donald L. Horowitz, “Democracy in Divided Societies,” in Nationalism, Ethnic Conflict, and Democracy, ed. Larry Diamond and Marc F. Plattner (Baltimore: Johns Hopkins University Press, 1994), 35–55. 72. John Coakley, “Independence Movements and National Minorities: Some Parallels in the European Experience,” European Journal for Political Research 8 (June 1980): 215–48. 73. Joseph M. Grieco, “Repetitive Military Challenges and Enduring Interstate Rivalries, 1918–1994” (paper presented at the annual convention of the International Studies Association, Los Angeles, Calif., March 2000), 19. 74. Ibid., 19.

Page 131 →

SYSTEMISM AND INTERNATIONAL RELATIONS Toward a Reassessment of Realism Patrick James I don't think that anybody under the sun would deny the statement that if you could have a single theory that would comprehend both international and domestic, both political and economic matters, all in one theory, hey, that would be a lot better than a simple theory of international politics. However, nobody's thought of how to do it. —Kenneth Waltz The vast majority of scholarship in international relations (and the social sciences for that matter) proceeds without conscious reflection on its philosophical bases or premises. —Thomas J. Biersteker This essay introduces the important concept of systemism as an alternative to the extremes of holism and individualism as a point of departure for theorizing about international relations. Systemism means a commitment to understanding a system in terms of a comprehensive set of functional relationships. In the context of international relations, systemism allows for a full range of connections between and among international and domestic “matters” as articulated by Waltz in the first quotation at the outset of this essay.1 Thus the approach stands in opposition to either a “black box” or some form of reductionism, which result, respectively, from holism and individualism.2 As will become apparent from an assessment of realist theory, systemism also provides a coherent philosophical basis for scholarship, which gives it the potential to respond effectively to the problem identified by Biersteker in the second quotation from above.3 Why, it might be asked, is it appropriate to initiate yet another review of international relations by talking about realism? Perhaps further theorizing, whether grounded in systemism or anything else, ought to begin anew elsewhere. Realism, after all, is not without its ongoing problems. Recent reviews sum up realist theory as incomplete; lacking specificity, predictive power, and policy relevance; and out of touch with contemporary developments.4 These and other criticisms commonly are applied to both the Page 132 → system-and unitoriented variants of realism, referring, respectively, to structural realism and “neotraditional” realism.5 While such critiques appear to be accurate at least to some degree, all have been put forward against, in the language of systemism, either holist or reductionist versions of realism. If systemism's basic philosophical point is valid—namely, that a theory should be expressed in terms of a full range of linkages at the unit and system level—then realism is not at a stage where its potential can be evaluated with authority. In other words, until realism is articulated in a manner consistent with systemism, it remains premature to say that it should be discarded as a result of failings associated with one or more variants. Instead, the disappointing performance so far might be a product of individual theories being developed in either reductionist or holist terms as opposed to any fundamental problem with the realist worldview.6 Thus, summing things up to this point, realism is by far the most developed school of thought within international relations,7 so extraordinary “value added” may result from looking at its respective components through the new lens provided by systemism. Both theory and substantive insights could be recombined into a more effective foundation for further work. This analysis of realism and systemism, the two concepts fundamental to this essay, unfolds in several stages. The first section provides a more detailed introduction to systemism. Second, implicit arguments for systemism are revealed within existing theoretical discussions of international relations. The third and fourth sections focus on the problems and prospects of structural realism and neotraditional realism, respectively, in the context of systemism. Fifth, some ideas are put forward for rearticulating and combining realist ideas within a systemist framework. The sixth and final section sums up the analysis.

Systemism, Theorizing, and Concept Formation Systemism is advocated by Bunge as a way of thinking about the world; in principle, it can be applied to any aspect of social life: The alternative to both individualism and holism is systemism, since it accounts for both individual and system and, in particular, for individual agency and social structure. Indeed, systemism postulates that everything is a system or a component of one. And it models every system as a triple (composition, environment, structure), or CES for short, so it encompasses the valid features of its rivals.8 From the standpoint of systemism, the choice between unit and system is, as Brecher might describe it, a “false dichotomy.”9 A useful theory must deal with both systems and units in some way. The question, of course, is how best to proceed. Figure 1 presents Bunge's vision, based on systemism, of functional relations in a social system.10 It traces the full range of effects that might be Page 133 → encountered in any such system, which includes both macro- and micro-variables. (Uppercase and lowercase letters refer to macro- and micro-level variables, respectively.) The logically possible connections are micro-micro (m-m), micro-macro (m-M), macro-macro (M-M), and macromicro (M-m). The essence of systemism is to get away from holism, individualism, and other mind-sets that convey only part of the picture of a working system. Macrovariables may have impact upon each other, as in the case of X affecting Y, with the functional form—perhaps linear in some instances and more complex in others—represented by F. The same is true at the micro-level, as in the case of x affecting y through function f. These linkages represent the limits of theorizing for holism and individualism, in that order. Systemism, by contrast, allows for both micro-micro and macro-macro connections. Furthermore, systemism recognizes that effects also can move across levels. Thus the macro-micro linkage of X to x through function g also appears in the figure, as does the micro-macro connection from y to Y on the basis of function h. Consider the incomplete pictures offered by individualism and holism in comparison with systemism. Individualism focuses on y = f (x), or micro-level processes. This specification is incomplete because it is possible for a causal chain to be traced further back and at a different level altogether. What about (a) the prior effect of X on x, expressed in terms of function g?; and (b) the direct impact of X on Y through function F? Each of these connections is left out of individualism, which cannot incorporate macrolevel processes into its unit-centered vision of the system. Holism, by contrast, begins at the macro level and stays there. It focuses on system properties alone and treats all else as a “black box.” The other processes already noted, which transmit effects from X to Y indirectly, are not incorporated into theorizing. Thus Y = F (X), a relationship based only on macrovariables, becomes the whole story. Systemism demands that a theory tell the full tale of units and systems. A theory that simply holds one type of linkage constant is inherently incomplete, unconvincing, and perhaps most important, unfalsifiable.11 Systemism therefore provides guidance for ontology: the field of observation must include all four logically possible kinds of connections, from macro-macro at one extreme to micro-micro at the other. Page 134 →

Systemism and International Relations Arguments that favor movement toward systemism exist already, in implicit form, and have emerged from varying points of view about international relations. These are not explicit calls for systemism as a unifying approach or philosophical basis for international relations. Instead, some of the ideas put forward about both the shortcomings of existing theory and priorities for the future imply the value of various aspects included in systemism. Salient examples focus on (a) the need to theorize about both units and systems and (b) the range of

linkages that must be explored to achieve a comprehensive understanding of international relations. Several recent studies, with diverse backgrounds in terms of subject matter and methods, reach the conclusion that more must be done to develop frameworks that include both units and systems. While more examples easily could be provided, those noted here are representative of current tendencies. Based on an assessment of great-power conflict and collaboration, Miller concludes that the unit level (i.e., states as actors) will be better at explaining outcomes that result from intended actions, while the system level (e.g., structure) will have the advantage in accounting for unintended outcomes.12 Examples of the latter include inadvertent wars in pre-World War I multipolarity, crisis patterns or tacit rules of the cold war era, and growing problems for crisis management in the current postbipolar era. In relatively encompassing terms, Freyberg-Inan calls for more actor-specific models to be incorporated in “general theories” of international relations,13 while Hall and Paul conclude that “order is both a systemic and a unit-level phenomenon.”14 And perhaps the most conspicuous assertion of this kind comes from Waltz, the creator of structural realism: “More needs to be said about the status and role of units in neorealist [i.e., structural realist] theory.”15 These and other recent discussions of units and systems fall short of citing systemism explicitly, but the preceding assertions are consistent with movement toward its frame of reference. Other studies take issue with holism and seem to approximate the language of systemism with regard to exploration of a wider range of linkages. Studies three decades apart can be used to introduce that point in more detail. Young deserves recognition for issuing a warning against “organismic analogies” in systemic analysis;16 in the language of systemism, this stands as a caution against assuming that any entity can be understood strictly in terms of macro-macro interactions. The point is made explicitly by Rosenau in summing up the contemporary international system: “It is not enough simply to assert that systemic processes at the macro level are sufficient to explain macro outcomes.”17 Thus, as Young and Rosenau point out, theorizing based exclusively on M-M connections cannot tell the whole story of international relations. Understanding of the need to consider hybrid connections also is evident; consider Solingen's implicit endorsement of M-m linkages: “domestic politics are never removed from systemic incentives, and do respond to global Page 135 → restraints.”18 This way of thinking, in which factors from outside (inside) the state are seen as important to explaining what happens inside (outside) the state, is evident in recent treatments of globalization,19 management of security,20 and a wide range of other issues within the realm of international politics. Examples from each side of the macro-micro fence should be sufficient to reveal the negative implications of both individualism and holism for the development of international theory. Neotraditional and structural realism bring out two sides of the same fundamental problem. In each instance, it appears that only one of the four basic kinds of linkage—meaning M-M, m-m, M-m, and m-M—is developed in any systematic way. The others, if they appear at all, are relegated to passing references.

Structural Realism Structural realism as articulated by Waltz illustrates the problem of holism.21 In Theory of International Politics, Waltz links the structure of the international system, defined in terms of the number of great powers under anarchy, with the propensity toward highly destructive war. The main contrast is between international systems with two versus three or more great powers; the former is preferred to the latter for reasons that Waltz sees as permanent and compelling, such as the relative simplicity of conflict management through spheres of influence.22 Thus, for Waltz, the key macrovariables are Yw, the warlikeness of the system, and Xw, the number of great powers defined by the categories of bipolar and multipolar. Yw is expected to be an increasing function of Xw. Waltz explicitly eschews all but this and a few other macro-macro linkages, asserting that structural realism claims to be a theory of international politics rather than foreign policy. Thus it includes neither micro-micro nor hybrid linkages.23 Structural realism's other basic hypotheses also take the form of macro-macro linkages. Waltz asserts that, among

states under anarchy, power balancing will take place and wars will recur.24 These propositions focus on patterns at the level of the system itself and do not suggest a path toward linkages beyond the M-M variety. Thus the complaints about realism alluded to earlier, some of which focused on lack of specificity, predictive power, and relevance to policy, become easy to understand after reviewing its principal system-oriented variant. Progress along the preceding dimensions would require the ability to generate a full range of linkages that include the unit level.25

Neotraditional Realism Neotraditional realism, also referred to as neoclassical realism, represents one reaction to the perceived excesses of Waltz's structural theory.26 Other than this rejection of holism, adherents of neotraditional realism probably have little fully in common, but they do seem committed to opening up the unit level for consideration along several fronts. New and reconstituted directions Page 136 → include traditional realist notions of power balancing (articulated in terms of concerns about relative gains)27 but also other ideas about the focal point of interstate competition. Prominent examples include balance of threat28 and balance of interests.29 Thus neotraditional realism might, on the surface, appear to be on a more dynamic and promising path than structural realism. All, however, is not as it might appear with neotraditional realism. In brilliant and authoritative reviews, Vasquez shows that the diverse set of ideas about balancing creates a basic problem, namely, a lack of falsifiability for the realist paradigm as a whole.30 Vasquez concludes that realism is unscientific because virtually all observed behavior of states, including “bandwagoning” with superior power, seems consistent with one or more variants of the theory.31 Examples of this problem continue to accumulate; consider the following seemingly positive assessment of realist explanations for some major events in the last decade: “After the Cold War, U.S. security and economic strategy have diverged. Security strategy has been more consistent with the predictions of balanceof-threat theory, while economic strategy has followed more closely the expectations of balance-of-power theory.”32 In sum, if power, threat, or interests being balanced all are acceptable as evidence in favor of realism, then, one might ask, “How can it lose?” Systemism can help to put the preceding difficulty into a wider context that facilitates an improved approach. Pursuit of a full set of linkages, beyond the recently near-standard m-m specification, would have the potential to push neotraditional realism in the direction of falsifiability. Consider, for example, Walt's treatment of balance of threat.33 While Walt claims that states will balance against the most serious threats as opposed to the mere aggregation of power, he sees the latter as being relevant to threat assessment. This might be restated, clarified, and linked to other variants of realism through use of language provided by systemism: anarchy produces a concern among states for balancing against threat, understood in terms of the distribution of power and interests, that is, an M-m connection. Ironically, this formulation already exists in the form of expected utility theory.34 Furthermore, unlike Walt's treatment, expected utility is explicit about functional form, which from the standpoint of systemism puts it another step ahead of balance-of-threat theory. While it would be beyond the scope of this essay to delve into expected utility in much greater detail, a brief description of the components from its basic equation related to war should make the preceding point. In words rather than symbols, for ease of exposition: Expected Utility of War (for state i versus potential enemy j) = (utility or value of winning for i) × (i's probability of winning) + (utility or value of losing for i) × (i's probability of losing) The formulas in use today in applied work on expected utility theory go far beyond the preceding simple presentation.35 However, even the rudimentary Page 137 → expression from above is sufficient to show that neotraditional realism could be strengthened by consulting expected utility theory, which works with the same concepts but offers an explicit functional form to bring them together. To be more exact, the value of winning, or utility, is assessed in terms of the interests at issue between states i and j. The probabilities of winning and losing are based on relative power between i and j. The overall expression for expected utility might even be conceived

of, in Walt's terminology, as the degree of threat posed by i to j (or, by reversing the way the expression reads, from j to i). Note that in this form, the concepts that currently find favor among neotraditional realists—interests, power, and threat—come together in a coherent functional form that even can be linked to the system level. In other words, efforts to maximize expected utility would derive from system imperatives created by anarchy, that is, an M-m linkage. While expected utility is not the only potential source for enhanced rigor and system-level connections, neotraditional realism would do well to review its need to improve falsifiability within the preceding context. Neotraditional realism and structural realism end up looking like two sides of the same coin. Each is self-limiting at this stage of development. Systemism points toward a way out of that problem. For international relations the implications of the preceding series of arguments are crucial with respect to building theories: the standard metaphor of puzzle solving encourages an inappropriate approach toward understanding a complex social system. While the idea of treating international relations (and especially the central problem of the causes of war) as a puzzle to be solved finds wide and seemingly expanding support,36 it appears misplaced from a systemist's point of view. The problems that arise from puzzle solving as a point of departure derive from the pathdependence of theorizing and the fundamental difference this makes in what ultimately can be achieved. When solving a jigsaw puzzle, the choice of strategy and tactics will not affect the ultimate result of putting the pieces together; instead, some approaches simply will yield faster results than others. Therefore, unless time is limited or some kind of competition is involved, it does not matter where work begins. Some will start off by selecting pieces of either a particular color or perhaps those along the border of the puzzle and go from there, putting together different sectors in varying orders. It also will not matter whether “dyads” or larger subsets are used in searching for pieces that fit together. The end result, regardless of strategy or tactics, is the same in each instance: a full picture that is identical to all others of its kind. Systemism raises serious doubts about whether starting points and overall strategies and tactics for the study of international relations are analogous to those of puzzle solving. From the standpoint of systemism, the point of entry matters a great deal. A holistic theory, which includes only macro-macro connections, seems unlikely to result in an accurate representation. The “dead end” seemingly reached by structural realism appears to confirm that point. It is possible and even probable that an individualist theory, consisting strictly of micro-micro linkages, also will produce an incomplete and Page 138 → even distorted picture of international relations. This appears to have been the result obtained so far by neotraditional realism, which arose in reaction to structural realism's perceived lack of potential for further growth. By comparison, a point of departure based on systemism necessarily builds in the need to confront all four types of logical connections between and within the micro and macro levels. It encourages, from the very beginning, comparative analysis of potential linkages of all kinds rather than a “false dichotomy”37 that favors one level over another at the outset and discourages or even closes off development of a more comprehensive theory. At this point there is merit in moving on to an initial illustration of how systemism can promote both imagination and greater consistency in application of the realist worldview to international relations.

Systemism and Realism One way to start building up realist theory in the context of systemism is to take propositions of the respective kinds (e.g., M-M, etc.) that have been posited in isolation from each other and search for logically consistent connections between and among them. An illustration of that process will be provided here, through a combination of ideas about dyadic relations and subsystems. Figure 2 shows a network of effects based on realist propositions.38 (Variables in uppercase characters refer to the macro level; those in lowercase to the micro level.) The intended contribution of the figure is to take disparate statements, each of which is consistent by intuition with realism as a worldview, and link them together in a logically consistent way. Striving for such connections at the macro and micro levels generally has been absent so far in the individualism of neotraditional realism and the holism of structural realism.

One of the specific connections that appears in the figure is derived from a recent study and sets in motion what will turn out to be a proliferating set of ideas about how the international system operates. To begin, one of the Mm connections in the figure, SUBSYSTEM COMPETITION⇉dyadic cooperation, is derived from Grieco's analysis of the Maastricht Treaty.39 He draws attention to potentially troubling behavior from a realist point of view, namely, German participation in the European Monetary Union (EMU), which for various reasons would tend to contradict realist predictions about Page 139 → pursuit of relative gains.40 This apparent anomaly is explained as a result of Germany accepting restrictions on its influence in Europe in order for the European Community to balance in monetary affairs against Japan. Convincing as it might be, this hypothesis is completely ad hoc and stands as a further illustration of the point made by Vasquez about the nonfalsifiability of realism.41 When, in particular, would a leading state within a subsystem (e.g., Germany within contemporary Europe) be expected to sacrifice individual gains in the interest of subsystem competitiveness? The M-m linkage in figure 2 calls for this differentiation to be made explicit: namely, when is dyadic cooperation to be expected in response to extrasystemic pressures?

Assuming that conditions for validity of the above-noted connection can be identified, it is possible to move back up to the system level with the m-M connection shown, namely, dyadic cooperation⇉SUBSYSTEM VIABILITY. This connection derives, indirectly, from an idea popularized by Axelrod, namely, that networks of cooperating actors will find advantage when faced with external competition.42 (This linkage would appear to be implicit in Grieco's analysis of German behavior in relation to the EMU.) The preceding linkage, in turn, might suggest the further M-M connections noted in the figure: SUBSYSTEM VIABILITY⇉SUBSYSTEM COMPETITION⇉BALANCE OF POWER⇉ANARCHY. The inducement of subsystem competition makes sense and suggests, perhaps, a balance of power between and among regions. With that in place, reinforcement of anarchy would be expected through interregional competition as opposed to cooperation or even integration. The interesting thing about the preceding sequence is that it becomes so easy to think of these connections once the macro and micro levels are opened up to each other. Consider the next possible step, which is M-m in nature: ANARCHY⇉dyadic competition, a connection that is developed elsewhere in Grieco's realist expositions on political economy and security.43 This last connection, however, would seem to create complications for all that has gone before; the return to the micro level raises questions about contradictory forces. When, in particular, will the macro inducement toward dyadic cooperation be greater, equal, or less than the macro imperative toward dyadic competition? Realism is silent on such questions precisely because it has not previously been cast in terms of systemism. While the connections that appear in the figure are speculative and would need to be developed further in terms of functional form and empirical testing, these linkages are sufficient to suggest (a) the range of hidden assumptions lurking within existing realist hypotheses that focus exclusively on the macro or micro level and (b) the great potential of realism to assemble and test M-M, M-m, m-M, and m-m linkages that previously have been examined either in isolation or not at all.

Conclusions and New Directions This essay has reviewed unit- and system-level variants of realism in the context of systemism, a potentially unifying approach and philosophical Page 140 → foundation for international relations. A reassessment of structural realism and neotraditional realism finds that each is subject to a fundamental flaw—holism on the one hand and individualism on the other. A preliminary attempt to reconsider just one realist hypothesis in the context of systemism shows that existing treatments are highly self-limiting and that realism may have a vast, untapped potential. Further research should continue to seek an expanded range of M-M, M-m, m-M, and m-m connections, logically consistent with each other, in order to derive a more compelling judgment of what realism may yet have to offer the field of international relations.

Notes 1. Kenneth Waltz, quoted in Fred Halliday and Justin Rosenberg, “Interview with Ken Waltz,” Review of International Studies 24 (1998): 379, emphasis in the original. 2. Mario Bunge, Finding Philosophy in Social Science (New Haven, Conn.: Yale University Press, 1996). 3. Thomas J. Biersteker, “Critical Reflections on Post-Positivism in International Relations,” International Studies Quarterly 33 (1989): 265. 4. Charles W. Kegley Jr., “The Neoliberal Challenge to Realist Theories of World Politics: An Introduction, ” in Controversies in International Relations Theory: Realism and the Neoliberal Challenge, ed. Charles W. Kegley Jr. (New York: St. Martin's, 1995), 5–7; Richard W. Mansbach, “Neo-This and Neo-That: Or, ‘Play It Sam’ (Again and Again),” Mershon International Studies Review 40 (1996): 95. 5. John A. Vasquez, “The Realist Paradigm and Degenerative versus Progressive Research Programs: An Appraisal of Neotraditional Research on Waltz's Balancing Proposition,” American Political Science Review 91 (1997): 899–912; John A. Vasquez, The Power of Power Politics, 2d ed. (New Brunswick, N.J.: Rutgers University Press, 1998). 6. This essay will not explore the nuances among concepts that focus on how ideas are put forward in the social sciences. A range of such concepts, from specific to general—including hypothesis, theory, paradigm, ontology, and worldview—is developed in Rosenau and refined in James. The present discussion instead will focus on theories and theorizing as these concepts have been used (often loosely) by exponents and critics of realism in one form or another. See James N. Rosenau, Along the Domestic-Foreign Frontier: Exploring Governance in a Turbulent World (Princeton, N.J.: Princeton University Press, 1997); and Patrick James, International Relations and Scientific Progress: Structural Realism Reconsidered (Columbus: Ohio State University Press, 2002). 7. Kenneth W. Thompson, Masters of International Thought: Major Twentieth-Century Theorists and the World Crisis (Baton Rouge: Louisiana State University Press, 1980); Kenneth W. Thompson, Fathers of International Thought: The Legacy of Political Theory (Baton Rouge: Louisiana State University Press, 1994); Torbjórn L. Knutsen, A History of International Relations Theory, 2d ed. (Manchester, England: Manchester University Press, 1997); George Liska, In Search of Poetry in the Politics of Power: Perspectives on Expanding Realism (New York: Lexington Books, 1998); Jack Donnelly, “Realism: Roots and Renewal,” Review of International Studies 24 (1998): 399–405; Michael Mastanduno and Ethan B. Kapstein, “Realism and State Strategies after the Cold War,” in Unipolar Politics: Realism and State Strategies after the Cold War, ed. Ethan B. Kapstein and Michael Mastanduno (New York: Columbia University Press, 1999). 8. Bunge, Finding Philosophy in Social Science, 264. 9. Michael Brecher, “International Studies in the Twentieth Century and Beyond: Flawed Dichotomies, Synthesis, Cumulation,” International Studies Quarterly 43 (1999): 213–64.Page 141 → 10. Bunge, Finding Philosophy in Social Science, 149. 11. James, International Relations and Scientific Progress. 12. Benjamin Miller, When Opponents Cooperate: Great Power Conflict and Collaboration in World Politics (Ann Arbor: University of Michigan Press, 1995), 3, 166. 13. Annette Freyberg-Inan, “Human Nature in International Relations Theory: An Analysis and Critique of Realist Assumptions about Motivation” (paper presented at the annual meeting of the International Studies Association, Washington, D.C., 1999), 52. 14. John A. Hall and T. V. Paul, “The State and the Future of World Politics,” in International Order and the Future of World Politics, ed. T. V. Paul and John A. Hall (Cambridge: Cambridge University Press, 1999), 406. 15. Kenneth N. Waltz, “Realist Thought and Neorealist Theory,” in Kegley Jr., Controversies in International Relations Theory, 81. 16. Oran R. Young, A Systemic Approach to International Politics, Center of International Studies, Research Monograph no. 33 (Princeton, N.J.: Woodrow Wilson School of Public and International Affairs, Princeton University, 1968), 24. 17. Rosenau, Along the Domestic-Foreign Frontier, 279.

18. Etel Solingen, Regional Orders at Century's Dawn: Global and Domestic Influences on Grand Strategy (Princeton, N.J.: Princeton University Press, 1998), 54. 19. Ian Clark, Globalization and Fragmentation: International Relations in the Twentieth Century (Oxford: Oxford University Press, 1997), 6. 20. Marc V. Simon and Harvey Starr, “Two-Level Security Management and the Prospects for New Democracies: A Simulation Analysis,” International Studies Quarterly 44 (2000): 391–422. 21. Kenneth N. Waltz, Theory of International Politics (Reading, Mass.: Addison-Wesley, 1979), 161–93. This point is ironic because Waltz is most effective in pointing out the problems associated with individualism, the polar opposite of holism, as a type of reductionism. The theory of imperialism is used at length by Waltz in Theory of International Politics as an example of the harmful side effects of unbridled reductionism with respect to theory construction. See Kenneth N. Waltz, “Reflections on Theory of International Politics: A Response to my Critics,” in Neorealism and Its Critics, ed. Robert Keohane (New York: Columbia University Press, 1986). 22. Kenneth N. Waltz, “The Stability of a Bipolar World,” Daedalus 93 (1964): 881–909; Waltz, Theory of International Politics. 23. The use of the terms micro and macro in the present essay is different than that of Waltz and should be made clear for further reference. In Theory of International Politics Waltz claims that structural realism is analogous to a “microeconomic theory” that “shows how the actions and interactions of the units form and affect the market and how the market in turn affects them” (110). By contrast, a “macrotheory is a theory about the national economy built on supply, income, and demand as systemwide aggregates.” The position that will be taken here, however, is that Waltz's structural realism should be regarded as a macrotheory (i.e., holistic) because the causal relations on which it focuses pertain to entities at the level of the international system itself, namely, structure and overall propensity toward destructive conflict. The point of origin for this theory—an application of the microeconomic explanation of behavior among firms in an oligopolistic market—is a different issue. 24. See Waltz, Theory of International Politics; Waltz, “Reflections on Theory of International Politics” Kenneth N. Waltz, “Realist Thought and Neorealist Theory,” in The Evolution of Theory in International Relations: Essays in Honor of William T. R. Fox, ed. Robert L. Rothstein (Columbia: University of South Carolina Press, 1991); Kenneth N. Waltz, “Evaluating Theories,” American Political Science Review 91 (1997): 913–17. 25. Richard Little and Barry Buzan, “Reconceptualizing Anarchy: Structural Realism Meets World History, ” European Journal of International Relations 2 (1996): 405; Randall L. Schweller and David Priess, “A Tale of Two Realisms: Page 142 → Expanding the Institutions Debate,” Mershon International Studies Review 41 (1997): 24; Stefano Guzzini, Realism in International Relations and International Political Economy: The Continuing Story of a Death Foretold (London: Routledge, 1998), 132; K. J. Holsti, “Scholarship in an Era of Anxiety: The Study of International Politics During the Cold War,” Review of International Studies 24 (1998): 21. 26. Jeffrey Legro and Andrew Moravcsik, “Is Anybody Still a Realist?” (Weatherhead Center for International Affairs, Harvard University, 1998). 27. Joseph M. Grieco, “Realist Theory and the Problem of International Cooperation,” Journal of Politics 50 (1988): 600–24. 28. Stephen M. Walt, “The Progressive Power of Realism,” American Political Science Review 91 (1997): 931–35. 29. Randall L. Schweller, “New Realist Research on Alliances: Refining, Not Refuting, Waltz's Balancing Proposition,” American Political Science Review 91 (1997): 927–30. 30. Vasquez, “The Realist Paradigm and Degenerative versus Progressive Research Programs” and Vasquez, The Power of Power Politics. 31. Vasquez, “The Realist Paradigm and Degenerative versus Progressive Research Programs” and Vasquez, The Power of Power Politics. 32. Michael Mastanduno, “A Realist View: Three Images of the Coming International Order,” in Paul and Hall, International Order and the Future of World Politics, 168. 33. Walt, “The Progressive Power of Realism,” 933. 34. Bruce Bueno de Mesquita, The War Trap (New Haven, Conn.: Yale University Press, 1981); Patrick

James, Crisis and War (Montreal: McGill-Queen's University Press 1988); Bruce Bueno de Mesquita and David Lalman, War and Reason (New Haven, Conn.: Yale University Press, 1992). 35. Examples of applied research on expected utility are legion and appear regularly in virtually all of the major journals in international relations. This work is not cited by Walt, “The Progressive Power of Realism.” 36. Dina A. Zinnes, “Three Puzzles in Search of a Researcher: Presidential Address,” International Studies Quarterly 24 (1980): 315–42; Dina A. Zinnes, “Why War? Evidence on the Outbreak of International Conflict,” in Handbook of Political Conflict: Theory and Research, ed. Ted Robert Gurr (New York: Free Press, 1980); John A. Vasquez, The War Puzzle (New York: Cambridge University Press, 1993); Randolph M. Siverson, “Thinking about Puzzles in the Study of International War,” Conflict Management and Peace Science 15 (1996): 113–32. 37. Brecher, “International Studies in the Twentieth Century and Beyond.” 38. This preliminary treatment does not specify a functional form for any of the connections included in the figure. For purposes of simplification, at this point each is assumed to be monotonically increasing. 39. Joseph M. Grieco, “The Maastricht Treaty, Economic and Monetary Union, and the Neo-Realist Research Programme,” Review of International Studies 21 (1995): 21–40. 40. Grieco, “Realist Theory and the Problem of International Cooperation” Joseph M. Grieco, Cooperation among Nations: Europe, America, and Non-Tariff Barriers to Trade (Ithaca, N.Y.: Cornell University Press, 1990). 41. Vasquez, “The Realist Paradigm and Degenerative versus Progressive Research Programs” Vasquez, The Power of Power Politics. 42. Robert Axelrod, The Evolution of Cooperation (New York: Basic Books, 1984). 43. Grieco, “Realist Theory and the Problem of International Cooperation” Grieco, Cooperation among Nations.

Page 143 →

INSTITUTIONALISM

Page 144 → Page 145 →

PROGRESS IN INTERNATIONAL RELATIONS Beyond Paradigms in the Study of Institutions David A. Lake Progress is hard to measure, especially in one's own era. There is certainly no consensus on the general meaning of the term, even less so when applied to scholarly inquiry. Moreover, the implications of new ideas or concepts can take generations to become apparent. It is doubtful that many of the early experimenters with electricity envisioned the role harnessing that force would eventually play in modern life. Nor did Adam Smith, writing against the mercantilist orthodoxy of his day, expect economic theory to develop its axiomatic form or become the new orthodoxy two centuries later. As a subject, progress is as elusive as it is in reality. Nonetheless, it is worthwhile periodically to take stock of where scholarly inquiry stands, where it appears to have succeeded, where it has failed, and which directions appear most promising at a particular juncture in space and time. Well aware of the pitfalls that await, I attempt such a stock taking in this essay. I organize my reflections into three general questions. What is the nature of our enterprise and what should our standards be? Where has there been significant progress in the field in meeting these standards, and why? Why hasn't there been more progress? I address these questions, first, for international relations as a whole and, second, for the study of international institutions.

Standards of Inquiry Our purpose as scholars is to increase understanding of world affairs, to help ourselves and others understand better how the world “works.” Policies for improving the world may or may not follow directly from an improved understanding. Knowledge may suggest changes that enhance human welfare, but it might just as well reveal innate characteristics immune from the interventions of even the most enlightened social engineer. Our goal, then, is not necessarily to improve the world—although, I repeat, this can be an important and valuable by-product of scholarly understanding—but to improve our knowledge of the world we inhabit and, in part, construct. Progress, Page 146 → in turn, can thus be defined as improved understanding at one of two levels: general explanations that fit a variety of phenomena and cases more accurately and particular explanations that apply to a small number of phenomena or cases more fully. There are many ways of understanding, including religious and ideological beliefs, empiricism, and the relationalism of postmodern theories. My preferred approach is that of scientific realism, in general, and the deductive method, in particular. This approach rests upon the explication of basic assumptions, the logical derivation of propositions from those assumptions, the deduction of observable implications, and the testing of implications through carefully designed quasi-experiments. The scientific method is not perfect, to be sure. Scientists have no monopoly on understanding. But the clarity of its procedures and the opportunity to prove arguments wrong make the scientific method superior, in my view, to the major alternatives. Given costly information and the limits of cognitive capacity, no one scholar can produce a complete body of theory and evidence. Active debate and contestation are a necessary part of what is inherently a collective enterprise. Through competition and repeated inquiry, the scientific community probes, refines, and elaborates insights that lie beyond the skills and imagination of any individual. Having clear procedures and standards of evaluation is essential, in turn, to progress. Given sufficient time and resources, scholars employing the scientific method can determine with some degree of objectivity whether assumptions 1, 2, … n produce proposition P, whether P implies hypothesis X⇉Y, and whether the empirical record supports this inference. Over time, inquiry may even approximate the best explanation, as theory gets closer and closer to the real world (but of course never mirroring the real world, as this would then be description). No other method possesses the same potential for “objective” evaluation or, at least, consensus among the scholarly community.1

Areas of Progress There have been significant areas of progress in international relations. At the risk of slighting others, I will focus on two. We have greatly improved our understanding of trade policy. Employing what is sometimes called open economy politics (OEP),2 we now see clearly how domestic groups define their material interests with regard to their positions within the international economy and how those interests are aggregated by political institutions into national policies or, through interstate bargaining, into international outcomes. Peter Gourevitch's now classic study of the first great depression was dismissed as “Marxist claptrap” when it was first submitted for publication, most likely because it assumed that groups pursue their material interests.3 Yet it laid the modern foundation for OEP. These insights were integrated with the pure theory of international trade by Ronald Rogowski, building on the Heckscher-Ohlin model and Stolper-Samuelson theorem (collectively, the HOS model), and Jeff Frieden, drawing on the Page 147 → Ricardo-Viner or specific factors model.4 Although the HOS and specific factors models were treated as alternatives that could be reconciled only over different time scales, Michael Hiscox has recently made the level of factor mobility a variable rather than an assumption, and thus is able to identify when one or the other model ought to apply; in doing so, he significantly improves the fit between theory and evidence.5 We now have a clear, deductive, synthetic, and empirically supported theory that produces non-trivial insights into how economic opportunities structure society, who lobbies for protection, and when they are likely to receive it.6 Cynics can, of course, dismiss this progress by snorting that Schattschneider said it all sixty-five years ago, but the father of pluralism at best described a general and relatively constant process that produced an extreme result at a particular moment in time—the Smoot-Hawley Tariff passed in the opening days of the Great Depression.7 He lacked an explanation for variations in trade policy outcomes. More important, current progress is tempered by significant theoretical anomalies. If we accept that political rhetoric is not entirely instrumental, then a remarkably large number of political leaders and their followers continue to believe that national welfare is enhanced by pursuing principles of competitive advantage rather than comparative advantage. Nor does the wave of liberalization that began in the 1980s seem to be driven by the near simultaneous victory of free traders around the globe.8 National policy still appears to be more than the sum of particularistic interests. Despite these limitations, however, substantial progress is evident. A second area of progress is the so-called democratic peace. This topic first gained prominence as a mix of philosophical inquiry and empirical evidence, 9 but it has since generated a cottage industry of theorists, who have refined and extended Kant's propositions into normative and institutional variants, and empiricists, who have developed measures and challenged findings. The state of the art continues to evolve here, but my reading of the growing literature is that (a) the finding that democracies do not fight one another is well-established; but (b) may be overdetermined, as most democracies during the cold war were on the same side of the bipolar divide;10 (c) there is some support for a normative theory of the democratic peace;11 but (d) the institutionalists have a betterspecified and more robust theory; that (e) generates a surprisingly large number of additional predictions that are also empirically supported.12 Although each of these steps is still debated, we now know substantially more about how differences in regime type affect the chances for war than we did even a decade ago. These two areas of progress have much in common, and a bit of intuition and casual empiricism combine to suggest that those commonalities may have a lot to do with the relative success of these research programs. First, both are disciplined by a strong empirical foundation. They begin with a real-world phenomenon—trade protection and war, respectively—over which there is broad agreement on the historical contours. There are numerous disputes over how best to measure the phenomenon at issue—nominal Page 148 → tariffs versus residuals from a gravity model of trade, war versus crisis—but analysts generally concur on the broad outlines of the targets at which they are shooting. Second, there is an emphasis on deductive rigor. Scholarly debate has forced analysts to clarify the assumptions they make, revealing strengths and limitations, and to state propositions in clear and falsifiable ways, thereby making hypotheses amenable to empirical test. By working within the standards of scientific inquiry, scholarly interactions and cumulative knowledge are facilitated. Both can be seen as exemplars of the power of the scientific method in world politics.

Finally, both literatures have drawn upon and integrated their analyses into broader bodies of theory. In the case of OEP, theorists have grounded their analyses in the pure theory of international trade and theories of collective action. In the democratic peace, effective use has been made of general theories of war, especially those stressing problems of asymmetric information,13 and theories of domestic institutions. Analysts do not reinvent the wheel each time they sit down at their computers. Rather, they extend and tailor theories developed for other purposes to the specifics of the question they seek to answer. In short, they are engaged in what is often disparagingly called normal science.

Impediments to Progress In contrast, many areas of international studies have made relatively little progress. All too often, research has not led to increased understanding but only to more debate and greater confusion. Although not an exhaustive list, I see three general and recurring problems in current research practices. First, we debate assumptions, not implications. In international relations, arguments commonly focus not on assessing the empirical power of theories but on the truth status of the assumptions upon which they are built. The so-called relative gains debate epitomizes this tendency. Joseph Grieco originally charged that by assuming states maximize absolute rather than relative gains, neoliberal institutionalists overpredict the level of cooperation within the international system.14 Without any metric for assessing the level of cooperation or clear deductions of exactly how much cooperation to expect under what circumstances, this was an unanswerable accusation. The debate quickly sank into charges that neoliberals thought cooperation was “too easy” and realists thought it was “too hard,” and the argument became more a test of faith about the world than about the world itself.15 In two important papers, Robert Powell and Duncan Snidal subsequently demonstrated that each approach could be generated as a special case of the other.16 Starting from neoliberal assumptions, Powell showed that if the cost of using force were sufficiently low, states would behave “as if” they were concerned with relative gains. Conversely, building on neorealist assumptions, Snidal showed that as the number of players increased, concerns for relative gains would rapidly dissipate and that in large-N situations states Page 149 → could do no better in maximizing their relative gains than to maximize their absolute gains. Both Powell and Snidal needed to make extra assumptions or impose additional “structure” on the theories to produce these results—and this is controversial—but it reflects more the incompleteness of neorealism and neoliberalism than any particular “agenda” these authors had in this clash. As I read it, the debate was then resolved with Grieco, for the neorealists, and Robert Keohane, for the neoliberals, essentially agreeing that it is an empirical question as to which of the two approaches might apply in any particular situation.17 Revealingly, scholars then immediately dropped this line of inquiry. No research of which I am aware has tried to test the empirical predictions that come out of the Powell and Snidal models. No one has taken up the challenge of identifying empirically the conditions under which one or the other perspective is more appropriate. Once the argument was reframed as an empirical question rather than a debate over first principles, scholars simply lost interest. An important lesson emerges from here. Paradigmatic debates are helpful in highlighting and clarifying assumptions. Without the provocation of Grieco's neorealist critique, an important foundation of neoliberal theory might have remained implicit. Criticism forces all of us to be more clear and precise in our theoretical formulations. But all too often debates dissolve into tests of the truth status of our assumptions rather than tests of their empirical implications. This adds little to our collective understanding.18 Second, we work on separate islands of theory. This is a well-known problem, pointed out by Benjamin Most and Harvey Starr more than fifteen years ago.19 Despite the attention devoted to more synthetic topics like “grand strategy” in recent years, the criticism remains accurate. Economic sanctions and war are treated in separate literatures, for example, even though both are tools used to coerce other states and thereby influence their behavior. The failure to build bridges between these separate islands and to conceptualize adequately the relevant policy

alternatives has hobbled inquiry. Choice theory presumes that actors assess any single policy against the next best options,20 but we typically develop and test our theories only in absolute terms. Thus, we study whether or not countries form alliances, for instance, as if this were a simple dichotomous choice rather than the selection of one option from a range of alternatives. At various times, unilateralism, imperialism, and unification have all been understood as appropriate responses to external threats, as have appeasement and undermining the enemy's own state or empire.21 The real question we ought to be asking, then, is why states ally rather than appease, arm, build empires, form voluntary confederations, or undermine opponents. The failure to specify relevant alternatives, in turn, produces significant selection bias. If we truncate the set of policy alternatives that we are trying to explain—say, studying only alliances rather than a range of alternatives—we systematically (and unwittingly) underestimate the effects of our independent variables. Elements like power disparities that we expect to be associated Page 150 → with the creation of alliances, and which simple extension suggests might have an even greater impact on strategies like appeasement, will appear to be less important and less significant than they really are. Conversely, if we truncate the set of policy alternatives when trying to estimate their effect on something else, we produce highly uncertain results. For instance, if we were to array a variable of, say, deterrence success along the vertical axis and actions of increasing hostility along the horizontal axis, a study of military mobilizations or crises alone would produce a stack of points clustered far from the origin, as “lower”-level or less threatening interactions were intentionally excluded and vertically dispersed, reflecting “natural error.” Nearly any line drawn through this stack of “mobilization” points would be as good as any other. As this is the design commonly employed in studies that seek to discern whether deterrence “works,” we can easily see why reasonable scholars continue to disagree over how to interpret the evidence!22 Selection bias either distorts our results or renders them more uncertain. What we look at—on which island we choose to conduct research—significantly influences what we see. Failing to conceptualize alternatives adequately, even those that we rarely observe in practice, significantly impedes progress. Third, we fail to operationalize our variables, especially our dependent variables. The problems in measuring “cooperation” were alluded to above but provide a particularly clear example of this impediment to progress. Robert Keohane's quite reasonable definition of cooperation as mutual adjustment in policy is now widely accepted.23 Nonetheless, the concept of “adjustment” remains ambiguous, and there is no metric against which change can be measured. We expect there to be “more cooperation” in dilemmas of collaboration rather than coordination,24 but this is a deduction rather than a confirmed empirical finding. How much cooperation occurs? How has the level evolved over time? How does it vary across issue areas? Without answering such basic questions, most theories of cooperation simply cannot be tested. We need to devote more time and resources to developing operational definitions and collecting systematic data. The Correlates of War Project was criticized in its early years for failing to produce important research findings. The payoff to this enterprise, however, is now abundantly clear in the study of the democratic peace. None of the work cited earlier would have been possible without the empirical foundation laid by J. David Singer and his colleagues. Much the same could be said for the Polity data set developed by Ted Robert Gurr and his associates, now generally accepted as an appropriate (if still not perfect) set of measures of “regime type.” Ideally, measures should follow theory, not the other way around. There is something arbitrary, it is easily admitted, about one thousand battle deaths as the definition of war. But we have to start someplace. Just as wars were eventually scaled into the larger class of militarized interstate disputes, so other seemingly arbitrary operationalizations may generate follow-on Page 151 → data sets. Building more systematic measures is a necessary but greatly undervalued activity.

Progress in the Study of International Institutions The study of international institutions, in my view, has enjoyed mixed success. There has been enormous progress,

especially over the last half century.25 Compared with the descriptive summaries of the activities of international organizations in the 1950s and the behavioral analyses of the 1960s, we now have strong theories that explain the formation and maintenance of international institutions, and we have detailed studies of how and where they “matter.” 26 We are now seeing a new generation of innovative midlevel theories designed to account for variations in legalization and institutional design.27 Reflecting this progress, institutionalism is now generally regarded as one of the major, mainstream approaches to the study of international politics. The sources of this success are two. First introduced by John Ruggie, extended by Robert Keohane and Joseph Nye, and the subject of a special issue of International Organization edited by Stephen Krasner, the concept of regime extended the range of what is regarded as an international institution, creating both a larger phenomenon to study and greater variation in the dependent variable used by scholars.28 This progressive “problem shift” in the study of international institutions mitigated—but, as I shall explain below, did not wholly eliminate—the second impediment discussed previously. In addition, international relationists productively drew upon emerging work elsewhere in the social sciences on the sources and roles of institutions, including that by economists Douglass North and Oliver Williamson and sociologist John Meyer.29 By doing so, international relationists were able to identify pitfalls, benefit from clear lines of debate in cognate fields, and move quickly up the learning curve. At the same time, institutionalism has failed to achieve the successes of OEP or the democratic peace, and it suffers in greater or lesser degree from nearly all of the impediments to progress noted previously. After a flowering in the early 1980s, institutionalism has not developed at the same pace, and it has failed to meet the high expectations of many of its proponents. Institutionalism is not, I want to emphasize, a failed research program. Far from it. But there remain grounds for considerable improvement. First, institutionalism lacks a ready operationalization of its key variable, whether regime or institution. Despite a seemingly clear consensus definition, 30 one that most of us can recite in our sleep, there is no consistent agreement on what practically and empirically constitutes a regime. Most agree that formal organizations or agreements constitute regimes, but there is no agreement on how to “count” or “code” organizations. Is the World Trade Organization (WTO) one overarching regime on goods and services or is it a series of smaller and still incomplete regimes on tariffs, property rights, investment, and so on? Even though the concept was explicitly formulated to include informal agreements and persistent practices, the problem Page 152 → of operationalization is magnified for tacit regimes. Is the “nuclear taboo” a regime or not?31 Is there an “atrocities” regime?32 How would we know? The concept of institution—sometimes understood to be broader, the same, or narrower than a regime—suffers from the same ambiguities. More than any other factor, this lack of a clear, operational definition stymies further progress. Second, following from the problem of operationalization, institutionalism lacks an accepted empirical pattern or puzzle that motivates inquiry. By this, I do not mean that institutionalism lacks empirical content. There are many fine studies of how institutions arise, operate, and evolve. Rather, what I mean is that, absent an identified pattern of institutions across issue areas, time, regions, or some other relevant dimension, analysts do not know what it is they are seeking to explain. To date, institutionalists have concentrated on demonstrating that contrary to realism, institutions exist and play an influential role even within an anarchic and perhaps inhospitable international system. Given the intellectual hegemony of realism, this was an important and possibly necessary first step. But to demonstrate that institutions matter does not cumulate naturally into institutional variation that then needs to be explained. Scholars must still impose some analytic dimension on the set of cases and identify some consistent pattern, perhaps by issue area (are institutions more prevalent in trade than monetary affairs, economic than security relations?) or by type of institution (are some institutions “stronger,” more hierarchical, more legalized than others?). In contrast, OEP largely agrees on its dimension of variation, openness to goods and factor movements across national borders, and on certain observable facts, especially variations in openness over time (more countries were more open in 1910 than in 1935, in 2000 than in 1960). The democratic peace is motivated by a single (if still contested) fact, namely that democracies do not fight each other. No similar empirical puzzle exists for institutionalism. We are just now beginning to see work that does pose dimensions of variation, 33 but operationalization remains poor and patterns ambiguous. For further progress, it is essential that scholars agree on the relevant dimensions of institutional variation, develop operational indicators that capture this variation, and

establish a pattern that they can then seek either to explain or to use to explain other phenomena. Third, although institutionalists have drawn productively on other literatures, as already noted, they have not expanded their horizons far enough, and in so doing have fallen into what Keohane and Martin call the “endogeneity trap.”34 Missing from the literature are the institutionalist models so prevalent in American politics and, to a lesser extent, in comparative politics.35 Institutionalism has revolutionized the study of American politics, shifting the emphasis from public opinion and the “imperial presidency,” on the one hand, to Congress and bureaucratic oversight, on the other.36 In highly simplified form, this approach assumes that politics is prone to cycling (whether due to collective preference intransitivities or multiple dimensions) and tends to produce multiple equilibria. Institutions, in turn, structure choice and “induce” one equilibrium from the many possible—in essence, Page 153 → determining which outcome we actually observe.37 In this way, institutions do most of the heavy lifting in explaining policy outcomes. William Riker famously criticized institutionalists for not recognizing that any instability in preferences would immediately translate into cycling over institutions.38 As institutions are endogenous, he pointed out, they cannot “solve” the problem of cycling. Institutionalists responded, quite correctly in my view, that institutions are embedded into larger social structures and reinforced by “vested interests” created by past polity choices. Thus, for many issues and policies, institutions can reasonably be treated as exogenous and theorists can proceed to examine how and why they influence politics without falling prey to the question that so bedevils international relationists—“do institutions matter?”39 International relationists have avoided this line of reasoning, I believe, because we too readily assume that institutions are weak, transient, and therefore part of the “game.” If institutions are always both a product and producer of political choices, it will be difficult—perhaps impossible—to show that they exert an independent effect, which was exactly Riker's point. International relationists fail to take the next step, however, and examine the degree to which institutions are embedded in larger social structures and therefore are more “fixed.” After fifty years, General Agreement on Tariffs and Trade/World Trade Organization rules and norms have had a tremendous effect on the pattern of private investment in economies around the world. As tariffs have fallen and markets opened, uncompetitive industries have been purged, export sectors have grown, and multinational corporations dependent upon the free movement of goods, services, and capital have flourished. Thousands of private investment decisions are made every day premised on the belief that trade will remain open. These decisions, in turn, create and reinforce “vested interests” in the preservation of this international institution. The situation here is not substantially different than that in other, highly constitutionalized areas of politics. The way out of the endogeneity trap is not by searching for the independent effects of institutions but by demonstrating how deeply embedded they are in the social structure of world politics. Finally, international relationists have adopted an overnarrow conception of institutions. Focusing only on “anarchic” institutions composed of sovereign states and based on the principle of sovereign equality, like NATO or the United Nations, they have ignored the broader range of institutions that have existed historically and, as I have tried to show elsewhere,40 continue to exist today. In anticipation of the Versailles peace conference and the difficult problems of constructing new states in Eastern Europe and dividing Germany's overseas empire among the victors, President Woodrow Wilson impaneled a study group known as The Inquiry to prepare background papers on many topics. One volume, a comprehensive study of Types of Restricted Sovereignty and of Colonial Autonomy, identified ten categories of polities ranging from semisovereign states to administered provinces. 41 Covering several hundred years of international history, this volume describes a range of increasingly hierarchical international institutions—a Page 154 → set now virtually forgotten in studies of international relations. Similar forms persist today, although norms of formal sovereignty make it politically incorrect to call such hierarchies by their traditional names—as witnessed by the unwillingness of the Western powers to acknowledge the de facto mandate they exercise over Kosovo. But from the United States' de jure protectorate over Micronesia to its de facto protectorate over Saudi Arabia and Kuwait and Russia's protectorates and informal colonies in the former Soviet Union, hierarchies abound in international politics.42 This variance in institutional form could and should be profitably exploited. The failure to allow for and examine

more hierarchical institutions not only ignores important dimensions of real world politics but—as highlighted previously—creates selection bias in our studies of institutions. As with alliances, truncating the variation in institutions as a dependent variable underestimates the effect of independent or causal variables, implying that those factors that affect international institutions are more significant than past studies have revealed. Transaction costs, informational asymmetries, the need for credibility, and other variables that institutionalists have used to explain institutions may be more important than they themselves recognize.43 Limiting variation in institutions as an independent variable, in turn, makes our estimate of their effects more uncertain. Implicit in most studies is a comparison of anarchic institutions, which may be common internationally, and hierarchic institutions, which are often presumed to characterize domestic political systems.44 The real question for international relationists is not whether international institutions matter in an absolute sense but how much and in what ways they matter relative to the supposedly more established and effective institutions of domestic politics. As with the case of deterrence failures, if one were to plot points for institutions of increasing hierarchy on the x-axis, thereby capturing this comparison, and “effectiveness” on the y-axis, the research design most commonly employed in international relations would produce a stack of points near the “anarchic” origin, with some vertical dispersion reflecting natural variation in effectiveness; with no further observations at other values of the x-axis, almost any regression line drawn through this stack of points would be just as valid as any other, thus preventing us from determining whether international institutions matter a little, a lot, or not at all. Given this research design, it is not surprising that reasonable scholars have been able to read the record of institutional effectiveness very differently.45 In short, by truncating variation in institutions, international relationists forfeit much-needed leverage over their subject. By extending the continuum of institutions to include hierarchies we gain a better understanding of institutions both as products of political choices and as producers of political outcomes.

The Way Forward The study of international institutions needs to become the study of governance. Governance is a generic problem of all polities and can be defined Page 155 → as the design, construction, and maintenance of mechanisms to reach collective decisions and to enforce agreed-upon bargains.46 The WTO, International Monetary Fund, and Organization for Economic Cooperation and Development are all “anarchic” institutions that govern relations between states, just as protectorates, confederations (like the European Union), and even the state itself are hierarchic institutions that govern relations between polities. The conceptual shift from institution—a human product that we often seek to explain—to governance, the function that institutions are designed to fulfill, naturally enlarges the purview of international relations and shifts attention to the question of why certain institutions and not others are chosen under what circumstances. Focusing on governance is no guarantee against artificially truncating the range of variations in institutions, but it makes unintentional selection bias less likely. Perhaps more important, a focus on governance endogenizes the state—that primordial institution of modern politics. The nature of the state, its forms and functions, its size and shape, all become properly part of the study of international institutions. Moreover, the question of why some governance functions are assigned to or appropriated by states while others are either settled cooperatively by mutual agreements between states or allocated to private actors—like nongovernmental organizations or multinational corporations—or supranational organizations becomes a central concern. By making the shift to governance, the study of institutions will be considerably enriched.

Conclusion International relations has a long history of paradigmatic contests. Indeed, much of the history of our field can be defined in terms of its “great debates.”47 As I suggested, these contests have performed a useful role. But they also produce or at least exacerbate the problems just identified. Paradigm warriors are particularly prone to argue about

the truth status of assumptions, to defend islands at the core of their approach rather than build bridges to those occupied by others, and to neglect research design in hopes of scoring debating points. In the continuing debate with neorealists on whether institutions “matter,” institutionalists have continued in this not-so-grand tradition. We are stuck in a convention trap, where however much we may individually abhor some of the standards and practices in our field we are nonetheless compelled to perpetuate them for our own success and that of our graduate students.48 Who hasn't sniffed at the work of a colleague as being merely “normal science”? Who hasn't reviewed an article and concluded that it is quite solid but not sufficiently innovative to warrant publication in a top journal? Who hasn't counseled a promising graduate student to be bold and provocative rather than solid and empirical? And yes, despite the complaints developed in this paper, I teach the introductory graduate course in international relations as a series of paradigmatic clashes, spending a week or more on each of the various “isms” that populate our field. Page 156 → A necessary step in changing conventions is to expose the insidious side effects of current practices. Evidence of a possible realignment of priorities is already under way in the nature of the “training” we now expect advanced graduate students to receive. But more needs to be done. We must insist that scholars ask interesting empirical questions. We must ensure that normal science has a prominent place in our journals. We must encourage scholars who seek only to debunk received wisdom or to add incrementally to our understanding of world affairs. We must reward scholars who contribute the years necessary to define concepts systematically, code cases, and build data sets. And we should continuously ask paradigm warriors the important question “What have you helped us understand lately?”

Notes 1. A colleague once asked a postmodern theorist of international relations how he would evaluate two works on the same topic. If he had to decide which was better, how would he choose? After a brief pause, the theorist replied that he would prefer the one that revealed the most hidden power. My colleague was too polite to ask the obvious follow-up: how would he know this result? 2. The label is from Robert H. Bates, Open-Economy Politics: The Political Economy of the World Coffee Trade (Princeton, N.J.: Princeton University Press, 1997). 3. Peter A. Gourevitch, “International Trade, Domestic Coalitions, and Liberty: Comparative Responses to the Crisis of 1876–1896,” Journal of Interdisciplinary History 8, no. 2 (1977): 281–313. Personal communication, summarizing the reviews received when the paper was submitted to a leading journal of international political economy. 4. Ronald Rogowski, Commerce and Coalitions: How Trade Affects Domestic Political Alignments (Princeton, N.J.: Princeton University Press, 1989); Jeffry A. Frieden, Debt, Development, and Democracy: Modern Political Economy and Latin America, 1965–1985 (Princeton, N.J.: Princeton University Press, 1991). Work parallel to that of Gourevitch, Rogowski, and Frieden was occurring in economics under the heading of “endogenous tariff theory.” See Stephen P. Magee, William A. Brock, and Leslie Young, Black Hole Tariffs and Endogenous Policy Theory: Political Economy in General Equilibrium (New York: Cambridge University Press, 1989). 5. Michael Hiscox, “Class versus Industry Cleavages: Inter-Industry Factor Mobility and the Politics of Trade,” International Organization 55, no. 1 (2001): 1–46. 6. OEP has also proven superior to the next best alternative. Although I think the theory of hegemonic stability was dismissed prematurely (see David A. Lake, “Leadership, Hegemony, and the International Economy: Naked Emperor or Tattered Monarch with Potential?” International Studies Quarterly 37, no. 4 [1993]: 459–89), I admit that OEP clearly provides a better explanation for the cross-sectional pattern of

protection we observe than this systemic alternative. Even on comparative grounds, OEP constitutes a better theory. 7. E. E. Schattschneider, Politics, Pressures, and the Tariff (New York: Prentice Hall, 1935). 8. Beth Simmons and Zachery Elkins, “Globalization and Policy Diffusion: Explaining Three Decades of Liberalization” (paper presented to the annual meeting of the American Political Science Association, Washington, D.C., September 2000). 9. Michael W. Doyle, “Liberalism and World Politics,” American Political Science Review 80, no. 4 (1986): 1161–69. 10. Henry S. Farber and Joanne Gowa, “Politics and Peace,” International Security 20, no. 2 (1995): 147–84.Page 157 → 11. Zeev Maoz and Bruce Russett, “Normative and Structural Causes of Peace, 1946–1986,” American Political Science Review 87, no. 3 (1993): 624–38. 12. See Bruce Bueno de Mesquita et al., “An Institutional Explanation of the Democratic Peace,” American Political Science Review 94, no. 4 (1999): 791; and Kenneth A. Schultz, “Do Democratic Institutions Constrain or Inform? Contrasting Two Institutional Perspectives on Democracy and War,” International Organization 53, no. 2 (1999): 233–66. 13. James D. Fearon, “Rational Approaches for War,” International Organization 49, no. 3 (1995): 379–414. 14. Joseph M. Grieco, “Anarchy and the Limits of Cooperation: A Realist Critique of the Newest Liberal Institutionalism,” International Organization 42, no. 3 (1988): 485–507. 15. John Mearsheimer, “The False Promise of International Institutions,” International Security 19, no. 3 (1994–95): esp. 19–20. 16. Robert Powell, “Absolute and Relative Gains in International Relations Theory,” American Political Science Review 85, no. 4 (1991): 1303–20; Duncan Snidal, “Relative Gains and the Pattern of International Cooperation,” American Political Science Review 85, no. 3 (1991): 701–26. 17. Joseph M. Grieco, “Understanding the Problem of International Cooperation: The Limits of Neoliberal Institutionalism and the Future of Realist Theory,” in Neorealism and Neoliberalism: The Contemporary Debate, ed. David A. Baldwin (New York: Columbia University Press, 1993); Robert O. Keohane, “Institutional Theory and the Realist Challenge after the Cold War,” in ibid. 18. There is, of course, a postmodern critique that our discourse itself shapes the world in which we live, but this raises a deeper set of issues than I want to dwell on in this essay. 19. Benjamin A. Most and Harvey Starr, “International Relations Theory, Foreign Policy Substitutability, and ‘Nice’ Laws,” World Politics 36, no. 3 (1984): 383–406. 20. David A. Baldwin, Economic Statecraft (Princeton, N.J.: Princeton University Press, 1985). 21. See David A. Lake, Entangling Relations: American Foreign Policy in its Century (Princeton, N.J.: Princeton University Press, 1999); Robert P. Hager Jr. and David A. Lake, “Balancing Empires: Competitive Decolonization in International Politics,” Security Studies 9, no. 3 (2000): 108–48. 22. Christopher H. Achen and Duncan Snidal, “Rational Deterrence Theory and Comparative Case Studies, ” World Politics 41, no 2 (1989): 143–69. 23. Robert O. Keohane, After Hegemony: Cooperation and Discord in the World Political Economy (Princeton, N.J.: Princeton University Press, 1984), 51–54. 24. Arthur A. Stein, “Coordination and Collaboration: Regimes in an Anarchic World,” in International Regimes, ed. Stephen D. Krasner (Ithaca, N.Y.: Cornell University Press, 1983). 25. Lisa L. Martin and Beth Simmons, “Theories and Empirical Studies of International Institutions,” International Organization 52, no. 4 (1998): 729–57. 26. For the former see Keohane, After Hegemony. For the latter see Lisa L. Martin, Coercive Cooperation: Explaining Multilateral Economic Sanctions (Princeton, N.J.: Princeton University Press, 1992); Ronald Mitchell, Intentional Oil Pollution at Sea: Environmental Pollution and Treaty Compliance (Cambridge: MIT Press, 1994). 27. For the former see Judith Goldstein et al., Legalization and World Politics, special issue of International Organization 54, no. 3 (2000). For the latter see Barbara Koremenos, Charles Lipson, and Duncan Snidal, The Rational Design of International Institutions, special issue of International Organization 55, no. 4 (2001).

28. John G. Ruggie, “International Responses to Technology: Concepts and Trends,” International Organization 29, no. 3 (1975): 557–83; Robert O. Keohane and Joseph P. Nye Jr., Power and Interdependence: World Politics in Transition (Boston: Little, Brown, 1977); Krasner, International Regimes. 29. Douglass C. North, Structure and Change in Economic History (New York: Norton, 1981); Oliver E. Williamson, The Economic Institutions of Capitalism Page 158 → (New York: Free Press, 1985); John W. Meyer, “The World Polity and the Authority of the Nation-State,” in Studies of the Modern World-System, ed. A. Bergesen (New York: Academic Press, 1980). 30. Krasner, International Regimes, 2. 31. Nina Tannenwald, “The Nuclear Taboo: The United States and the Normative Basis of Nuclear NonUse,” International Organization 53, no. 3 (1999): 433–68. 32. Christopher Rudolph, “Constructing an Atrocities Regime,” International Organization 55, no. 3 (2001): 655–91. 33. Goldstein et al., Legalization and World Politics; Koremenos, Lipson, and Snidal, The Rational Design of International Institutions. 34. Robert O. Keohane and Lisa L. Martin, “Institutional Theory as a Research Program,” in Progress in International Relations Theory: A Critical Assessment and Application of Imre Lakatos's Methodology of Scientific Research Programs, ed. Colin and Miriam Elman (forthcoming). 35. An important exception is Lisa L. Martin, Democratic Commitments: Legislatures and International Cooperation (Princeton, N.J.: Princeton University Press, 2000). 36. See Gary W. Cox and Mathew D. McCubbins, Legislative Leviathan: Party Government in the House (Berkeley: University of California Press, 1993) and D. Roderick Kiewiet and Mathew D. McCubbins, The Logic of Delegation: Congressional Parties and the Appropriations Process (Chicago: University of Chicago Press, 1991) for examples. 37. Kenneth Shepsle, “Institutional Arrangements and Equilibrium in Multidimensional Voting Models,” American Journal of Political Science 23, no. 1 (1979): 27–59. 38. William H. Riker, “Implications from the Disequilibrium of Majority Rule for the Study of Institutions,” American Political Science Review 74, no. 2 (1980): 432–46. 39. On endogenizing institutions in international relations, see Peter A. Gourevitch, “The Governance Problem in International Relations,” in Strategic Choice and International Relations, ed. David A. Lake and Robert Powell (Princeton, N.J.: Princeton University Press, 1999). 40. Lake, Entangling Relations. 41. W. W. Willoughby and C. G. Fenwick, The Inquiry Handbooks, vol. 16 (originally published as Types of Restricted Sovereignty and of Colonial Autonomy, Washington, D.C.: GPO, 1919; reprint, Wilmington, Del.: Scholarly Resources, 1974). 42. For the former see Lake, Entangling Relations, 147–48 and ch. 6. For the latter see Kathleen Hancock, “Specific Assets and White Knights: Explaining Ukraine's Success in Escaping Russia's Grasp” (paper presented at the annual meeting of the American Association of Slavic Studies, Denver, Colo., November 2000). 43. Most importantly, this selection bias has led analysts to largely ignore what I have elsewhere called “governance costs.” See Lake, Entangling Relations. 44. Kenneth Waltz, Theory of International Politics (Reading, Mass.: Addison-Wesley, 1979), 81–93; alternatively Helen Milner, “The Assumption of Anarchy in International Relations Theory: A Critique,” Review of International Studies 17, no. 1 (1991): 67–85. 45. Although the “problem” of institutional effectiveness has not previously been formulated in precisely this manner, recognizing it helps explain the focus of many institutionalists on the European Union. As the most hierarchical of the contemporary and voluntarily negotiated institutions, it provides crucial analytic leverage on the question of institutional effectiveness. Yet it is only one additional data point, and it may well be an outlier on which we would be well advised not to place much credence. 46. Miles Kahler and David A. Lake, “Globalization and Governance” (paper presented to the annual meeting of the American Political Science Association, Washington, D.C., September 2000).Page 159 → 47. Peter A. Katzenstein, Robert O. Keohane, and Stephen D. Krasner, “International Organization and the Study of World Politics,” International Organization 52, no. 4 (1998): 645–85.

48. For examples of such convention traps, see Gerry Mackie's chilling discussion of footbinding and infibulation. Gerry Mackie, “Ending Footbinding and Infibulation: A Convention Account,” American Sociological Review 61, no. 6 (1996): 999–1017.

Page 160 →

INSTITUTIONAL THEORY IN INTERNATIONAL RELATIONS Robert O. Keohane Joseph S. Nye has discussed our joint work on “transnational relations and interdependence” in his paper for this volume. We accordingly agreed that I would focus on my own work on institutional theory and cooperation, exemplified most clearly in my After Hegemony: Cooperation and Discord in the World Political Economy,1 and subsequent work by myself and others. These lines of work are closely related. Our joint analysis of asymmetrical interdependence and power, and “complex interdependence,” describes the context of contemporary world politics within which my institutional theory is set. Furthermore, some of the key elements of that institutional theory appear in our book Power and Interdependence: World Politics in Transition, especially in chapter 3.2 Indeed, our jointly developed “international organization model of regime change” provided the basis for my subsequent analysis of institutions. In this model, “networks, norms and institutions are independent factors for explaining regime change.”3 Like Joseph Nye, I have long been dissatisfied with realism but unwilling simply to jettison its insights, particularly the powerful understanding of international structure and its constraining effects developed by Kenneth Waltz in Theory of International Politics.4 The institutional theory that I have played a role in creating shares with realism that it takes actors and preferences over outcomes as given. It is therefore, as both students of comparative foreign policy and constructivists have emphasized, incomplete. Indeed, it is somewhat ironic that after the battles between institutional theory and realism of the 1980s and 1990s, they now seem to have much in common—particularly this limitation of beginning with already-constituted actors, having exogenously determined preferences over outcomes. It is important to note that although preferences over outcomes are exogenously determined, strategies are not. One of the key claims of neorealism is that strategies depend on the distribution of power. Institutional theory accepts this point but adds that strategies are also affected by the institutional configuration, which affects transaction costs of collective action, and by information conditions. I find it useful to think of long end-means Page 161 → chains. At the base are fundamental preferences—“life, liberty, and the pursuit of happiness,” or maybe happiness itself. Strategies that are means to more fundamental goals are themselves ends with respect to strategies meant to achieve them. National power, for instance, is not an end in itself, except for megalomaniacs, but a means to other goals. If we define what we mean by national power, we can speak coherently of it as a goal or interest; but we need to be aware that it is not, properly speaking, a preference over an outcome. In this short paper, I will sketch what I view as some of the accomplishments of institutional theory and some of the challenges that it faces. Many of my points reflect joint work—notably with Joe Nye and Lisa Martin.5 My criteria for evaluation of theoretical insights are essentially an adaptation from the insights of Imre Lakatos,6 although I recognize that his work does not provide precise guidelines. Really good new explanatory theories predict “new facts” either previously unrecognized or at least not seen as relevant to the issues on which the theory focuses. Examples include the orbit of Mercury for Einstein; iridium in the earth's crust for the dinosaur extinction theory of Alvarez and Alvarez; and underproduction of collective goods for Mancur Olson's theory of collective action. Descriptive interpretations (sometimes called “theories”) identify new phenomena and put them into a coherent context of what Clifford Geertz referred to as “thick description.”7 When the phenomena are important and the interpretation illuminating, such thick description is an important contribution, even when explanation is lacking. Institutional theory accepts three basic realist assumptions: (1) states are the primary actors in world politics; (2) they can be analyzed as if they were rational; and (3) they are not altruistic but, rather, are broadly “selfinterested.” However, neorealism assumes that information, although important, is a constant: information is scarce and of low quality, forcing states to lean toward worst-case scenarios in selecting strategies. In stark contrast, institutional theory treats information as a variable that can be affected by human actions and that is

embedded in institutions. Furthermore, states have interests in creating and using institutions to improve the quality of information—both what they receive and the credibility of what they send. States therefore construct institutions in an attempt not only to reduce transaction costs but to improve the information environment, thus enabling them to overcome some of the dilemmas and defensive stances identified by realism. Let me say here that I never claimed that institutions were “efficient.” “Egoism” and “empathy” appear in the index to After Hegemony, but “efficiency” does not. Institutional theory argues that the members of an international regime seek gains for themselves, which they can often only achieve through agreements with others. But it emphatically does not claim that institutions will be efficient, in either of two senses: 1. Even for members of an institution, informational asymmetries and transaction costs will keep participants away from the Pareto frontier.Page 162 → 2. Even if the members were to build institutions that were efficient for themselves, there is no reason to believe that those institutions would be socially efficient. Indeed, if other potential members were excluded, there is every reason to believe that they might suffer as a result of the institutions. This theory made some correct predictions, notably that international organizations should engage heavily in monitoring state behavior. Enforcement should remain decentralized, as realism expects, while monitoring, through more or less elaborate procedures, should be extensive. These predictions seemed to hold up not only in areas of international political economy, for which they were first devised, but also for environmental issues and even with respect to security institutions.8 The application of institutional theory to real situations has helped to illuminate how international regimes operate: there has been real progress in the subfield. We have done rather well at extending institutional theory to other issue areas (trade, environment, human rights). The field has been pretty good at recognizing new trends—“complex interdependence,” increases in nongovernmental organizations and their networks, transgovernmental relations, the information revolution, globalization—which Joe Nye discusses in his paper. We have developed new categories and taxonomies to make sense of the otherwise bewildering changes that we observe. We think about Nye's “soft power”9 as well as “hard power,” and we are alert to how institutions can change information and strategies; the field is certainly not trapped in the old realists' limited frame. Indeed, it seems to me that students of world politics have been pretty good at keeping what is helpful about realism and discarding some of the encrusted doctrine. These accomplishments have been significant and on the whole independent of the “great debates” involving realism, various forms of liberalism, and constructivism. These macroperspectives are not mutually incompatible in any case: the issue is how to use them to promote good analysis of world politics, not to score points or to trumpet the triumph of one paradigm over others. Yet there are good reasons to be self-critical. The most general point is that we have been better at extending our descriptions to new issue areas, and incorporating new trends into them, than we have been at improving the explanatory power of our theories. That is, our record is better at description, and even descriptive inference, than at explanation. The irony here is that institutional theory's much-criticized limitations should have been a source of scientific strength. Having eschewed the study of the origin of state preferences, institutional theory has sought to use game theory and other techniques to understand strategy and bargaining within international institutions, particularly among states. Admittedly, we have too often been distracted into defending these self-imposed limitations against other macroapproaches—in the “interparadigm debates.” Yet even when we focus on the stripped-down essentials, deliberately ignoring the sources of “state preferences,” our progress has been slow. Let me list a few salient problems: Page 163 → 1. Distributive issues have been recognized as important, but institutional theory has been slow to integrate them into its analysis. Recent progress has been made here,10 but it has taken the form of identifying the

1. issues more clearly and has not yet yielded a conditional theory specifying the effects of distributive conflict on cooperation. 2. We have been slow to develop systematic databases of institutions, which could be used for sophisticated quantitative work. Oran Young and his colleagues, and recently Ronald Mitchell, have started projects designed to construct such databases. It is difficult work, since so many of the relevant questions require extensive knowledge and sound judgment to answer; this is not a matter simply of extracting data from statistical volumes or tapes. 3. Methodologically, we face the usual problem of strategic interaction: rarely are there unique equilibrium solutions. Since international institutions are characterized by multiple actors and audiences, the range of possible outcomes is enormous. Interactions and feedback effects are pervasive, so the simple logic of explanatory and dependent variables only takes us a small part of the way. Even if we had better data, therefore, it would be quite difficult to go beyond identifying causal mechanisms and describing the frequency of different patterns to a genuinely conditional analysis of strategies. 4. In particular, there is a serious problem of endogeneity embedded in institutional theory itself. Institutional theory acknowledges that institutional configurations reflect underlying power structures—Nye and I made this point in chapter 3 of Power and Interdependence. Yet we institutionalists also believe that institutions have an independent impact on strategies and outcomes. Theoretically, one response to this “endogeneity critique” is to show that theories of agency and delegation are consistent with both horns of the dilemma: that structural conditions matter, but so do the actions of agents.11 Methodologically, however, the issue is more difficult: it is very tricky to disentangle the effects of underlying structure from the effects of institutions. We cannot control for structure while varying institutions, given that institutions are endogenous to structure. This, it seems to me, is a crucial problem that needs to be solved—I expect, through a combination of qualitative and quantitative approaches—before we will be able scientifically to establish the effects of international institutions on world politics. I suspect that we will make progress on these issues during the next decades only if we are tolerant both of separation and synthesis. Separation of specific tasks from broader theoretical questions is essential for scientific progress: disciplines from physics to paleontology, from biology to economics, have shown that decomposing problems into subparts (even artificially) is essential to scientific progress. It is quite conceivable that solutions to issues of strategic interaction or endogeneity in institutional theory will come from basic scientific discoveries in other fields that seem at first esoteric and irrelevant. Game theory, agent-based simulations, mathematical explorations of path-dependence, and high-powered quantitative methodology may all come to our rescue, someday, in ways that we cannot now imagine. Original innovations in these areas should be encouraged even by those of us who barely understand them—at least as long as their proponents Page 164 → (if not their theoretical innovators) seek to show how they could be relevant to real-world institutional issues. It would, however, be mistaken to aim so exclusively for science that we fail at our task of interpretation and description. Change in world politics is coming rapidly, from many directions. Journalists, activists, and officials often lack even basic social scientific skills of qualitative inference. Many of them know very little history and are not very reflective by nature. Even if they are, they do not have time for extensive reflection. Describing contemporary world politics in a coherent way is too important to be left to the journalists and the advocates. Our students expect it of us; and the public needs it, whether it knows it or not. Scientists making basic discoveries in game theory or methodology do not supply this essential interpretative discussion. The aspiring scientists among us—those most ambitious for the development of cumulative bodies of intersubjectively testable knowledge—should be made humble by recognizing that during the last several decades, they have had less impact on how decision makers see events, or how they have acted, than their colleagues who employ looser forms of interpretation. Neither intellectual quality nor relevance lies only on one side of the ledger. Intellectual progress requires both arrogance and humility. One has to be somewhat arrogant to believe one can make an original contribution in a field so complex, riddled with uncertainty, and littered with the remains of failed theories and research programs. But to get very far, we need also to be humble—to recognize our own limitations, the strengths of other perspectives, and the need for synthesis as well as separation. We are better at

being arrogant than being humble, so it is more important to work at the latter!

Notes 1. Robert O. Keohane, After Hegemony: Cooperation and Discord in the World Political Economy (Princeton, N.J.: Princeton University Press, 1984). 2. Robert O. Keohane and Joseph S. Nye Jr., Power and Interdependence: World Politics in Transition, 3d ed. (New York: Addison Wesley Longman, 2000). 3. Ibid., 54. 4. Kenneth N. Waltz, Theory of International Politics (Reading, Mass.: Addison-Wesley, 1979). 5. Robert O. Keohane and Lisa L. Martin, “Institutional Theory as a Research Program” (paper presented at the annual meeting of the American Political Science Association, Atlanta, August 1999). 6. Imre Lakatos, “Falsification and the Methodology of Scientific Research Programs,” in Criticism and the Growth of Knowledge, ed. Imre Lakatos and Alan Musgrave (New York: Cambridge University Press, 1970). 7. Clifford Geertz, An Interpretation of Cultures (New York: Basic Books, 1973). 8. Helga Haftendorn, Robert O. Keohane, and Celeste A. Wallander, Imperfect Unions: Security Institutions over Time and Space (Oxford: Oxford University Press, 1999). 9. Joseph S. Nye Jr., Bound to Lead: The Changing Nature of American Power (New York: Basic Books, 1990). 10. James D. Fearon, “Bargaining, Enforcement, and International Cooperation,” International Organization 52, no. 2 (1998): 269–306. 11. Keohane and Martin, “Institutional Theory as a Research Program.”

Page 165 →

TRANSNATIONAL RELATIONS, INTERDEPENDENCE, AND GLOBALIZATION Joseph S. Nye Jr. When I was a graduate student, I took a course taught by Hans J. Morgenthau. I was impressed by the power of the realist model and remain so. When I was in a policy position and responsible for reconstructing American alliance relations in East Asia, I relied heavily on realism.1 In my textbook Understanding International Conflicts, I insist that students begin by understanding the realist model before they turn to liberal and constructivist theories.2 But even in the early 1960s, I was equally impressed by the notion that realism captured only part of what was important in world politics. Ideas, institutions, and economic interdependence were left out, and I wanted to know more about them. As I pursued fieldwork on my thesis on Pan-Africanism and East African integration, I found the theoretical work of Ernst Haas enormously helpful.3 While the regional integration theory that was developed by Haas, Deutsch, Lindberg, Puchala, Scheingold, Schmitter, and others over the course of a decade outstripped the realities of regional integration,4 it was a healthy antidote to the single-mindedness of the realist model, and it laid the basis for future work on transnational relations, interdependence, and many of the themes now stressed by constructivist writers. Haas has been unduly harsh in his evaluation of his legacy.5 In the early 1960s, International Organization, the premier journal in the field of international institutions, tended to have a restricted and traditional view of the field. Toward the end of the decade, the board of editors decided that it needed to add younger members with a new perspective, among whom Robert O. Keohane and I were included. At our first board meeting, we joined forces to plead for more attention to transnational relations and interdependence, and in 1971 we produced a special issue of the journal devoted to those topics.6 Subsequently, Keohane became editor and I became chair of the board, and the journal began to develop the broader approach to international institutions (which we sometimes called “small i, small o”) that characterizes it today. Keohane and I also began collaboration on a number of articles and our book Power and Interdependence, published in Page 166 → 1977. Much of the basic work was done while we were both at the Center for International Affairs at Harvard, where we benefited enormously from colleagues such as Stanley Hoffmann, Samuel Huntington, Stephen Krasner, and other critics. Above all, we profited from the stimulus provided by Raymond Vernon, whose book on the multinational corporation, Sovereignty at Bay, was a pioneering work.7 We began working on the ideas in our book at a time when economic interdependence was gaining prominence. Events such as the collapse of the Bretton Woods fixed exchange rate system and the Arab oil embargo made daily headlines. Even a classic realist like Henry Kissinger, then secretary of state, began to include interdependence in his speeches. Hans Morgenthau called the transfer of wealth to OPEC historically unprecedented. Some enthusiasts proclaimed the end of the nation-state; others predicted a new international economic order of many OPECs dominated by the rising South of the globe. The hyperbole did not help the cause of analysis of interdependence, and although we pointed that out at the time, many readers categorized our work as a rejection of realism. As headlines and foundation grants shifted back to traditional security concerns in the 1980s, many realists breathed a sigh of relief. In 1989, in the preface to the second edition of our book, Keohane and I observed how theories of international relations are susceptible to the influence of current events. We noted the revival of realism during the 1980s “little cold war,” and how different the political climate was from the decade in which the book was written. Nonetheless, we argued that the perspectives on interdependence were still relevant. At the beginning of a new century and millennium, many aspects of world politics resemble the politics of the 1970s more than those of the 1980s, but it would be as much a mistake to discard the insights of realism today as it was to ignore insights of liberalism in the 1980s. The consistent theme of our argument has been to combine the two great theoretical traditions. In the 1980s, my work focused on ethical and policy problems that grew out of my experience managing nuclear

nonproliferation policy in the Carter administration. Later, I turned my attention to the fallacies of the popular declinist theories and tried to provide a framework for understanding the future of America's hard and soft power. Keohane focused his work on developing a neoliberal case for institutions based on the rational actor framework that neorealists themselves use, and that is the subject of his reflection paper in this symposium.8 The 1980s debate between neorealists and neoliberals greatly increased the simplicity and elegance of theory, but in my view it did so at the cost of discarding much of the rich complexity of the classic theories. Although their approach lacks theoretical elegance, I welcome the corrective provided by constructivists. I tend to agree with Miles Kahler that “by the end of the 1980s, the theoretical contest that might have been was reduced to relatively narrow disagreements within one statecentric rationalist model of international relations.”9 In that context, and with globalization becoming a new buzzword, Keohane and I began working on a third edition of Power and Interdependence. Page 167 → Of course, the world today bears only a superficial resemblance to that of the 1970s. An information revolution is strongly affecting the relations of states, economies, and civil societies in the world. Rapid technological advances in computers, communications, and the Internet have led to dramatic decreases in the cost of information. It is generally agreed that we are still in the early stages of the information revolution. Nonetheless, the effects of the information revolution are already significant. Nongovernmental actors can organize transnationally around various issues at very low transaction costs, blurring the distinction between domestic and international politics. Many of these new organizations are more like networks than bureaucratic hierarchies.10 If one envisions social space comprising markets, governments, and civil societies, all three sectors are affected, as well as the relations between them. Markets are becoming more global and corporate structures more like networks. In civil societies, new forms of transnational community and identity are developing that cut across traditional geographical jurisdictions. At the same time, the form and functions of government are changing in relation to the economy and society, resulting in less centralized states. While the Westphalian system of sovereign states is still the dominant pattern in international relations, one can begin to discern a pattern of crosscutting communities and governance that bears some resemblance to the situation before 1648. Of course the information revolution is not the only powerful trend undermining the Westphalian system. Globalization and marketization also have powerful effects. Both are related to the information revolution, but they also have independent roots. Globalism is networks of interdependence at multicontinental distances.11 It is not new, and its economic dimensions grew greatly in the nineteenth century before being disrupted by World War I and the Great Depression. The recent wave of globalization goes back to American strategy after World War II and the desire to create an open international economy to forestall another depression and to contain communism. The institutional framework and political pressures for opening markets were a product of American policy, but they were reinforced by developments in the technology of transportation and communications that made it increasingly costly for states to turn away from global market forces. With the fall of the Berlin Wall symbolizing the demise of the bipolar political barriers, the pace of globalization has accelerated. Marketization is related to globalization but also has independent roots in shifting domestic political coalitions in leading capitalist states, the inflation that followed the oil crises of the 1970s, the failure of the planned economies, and the development of new financial instruments. Power has diffused from governments to markets in such critical functions as maintaining the value of the currency, choosing the form of the economy, taxation, providing infrastructure, countercyclical policy, and protection from crime.12 In the United States, the growth of government's share of gross national product has tapered off at roughly a third. Even in Western Europe, where government accounts for nearly half of GNP in many countries, market Page 168 → forces have become more powerful and many public corporations have been privatized. In Eastern Europe and less developed countries—sometimes categorized as emerging markets—the change has been even more dramatic. Like globalization, the development of markets was also strongly affected by the information revolution. At a macro level, the failure of the Soviet planned economy to respond adequately to the information revolution led to the collapse of Soviet power and the end of bipolarity. At the micro level, information is central to effective markets, and the growth and diffusion of information has lowered transaction costs and made all types of

contracting much easier. Thus the information revolution is not the sole cause of the current changes in international relations, but it has accentuated other causes as well as produced its own independent effects. Some of these changes were already discernible when we wrote Power and Interdependence. As we said then, “Jet aircraft have brought Asia and America within a day's journey from each other. Synchronous orbit satellites have brought the costs of intercontinental phone calls into the same range as intercity calls. Picture-phone transmission and intercontinental computer communications promise to continue to shrink many of the costly barriers imposed by distance.” And we described the views of those we called “modernists,” who “see telecommunications and jet travel as creating a ‘global village’” and believe that “the territorial state, which has been dominant in world politics for the four centuries since feudal times ended, is being eclipsed by nonterritorial actors such as multinational corporations, transnational social movements, and international organizations.” We were careful, however, not to associate ourselves solely with the modernists. Instead, we sought to integrate the new dimensions of interdependence with the traditional realist model based on the distribution of power among states. We showed how asymmetrical interdependence, particularly in trade, provided a form of power that states could use in very traditional ways. However, we also invented an ideal type of politics that we called “complex interdependence” that would result from reversing the assumptions of the traditional realist model. Contrary to the expectations of some theorists, the information revolution has not greatly decentralized or equalized power among states.13 If anything, thus far it has had the opposite effect. But what about reducing the role of governments and the power of all states? Here the changes are more interesting. In the second chapter of Power and Interdependence, we asked what the world would look like if one reversed the three critical assumptions of the realist model that had previously dominated most portrayals of international politics. Doing so led us to posit an ideal type called complex interdependence with three conditions: (1) a minor role for military force, (2) absence of hierarchy among issues, and (3) multiple channels of contact among societies. To what extent has the world moved in a direction that approximates those conditions? There was great variation among regions and across issues that we studied in the 1970s. Force and hierarchy of issues were of minor significance in the relations between the United States and Canada, or between France and Page 169 → Germany, but of crucial significance in the U.S.-Soviet relationship or among the states of the Middle East. The biggest change since the 1970s has been the end of the cold war and of the Soviet Union. With the demise of the bipolar world, the structure of world power has become like a three-dimensional chess game, with a military board marked by unipolarity, an economic board that is multipolar, and a transnational board that has a dispersed structure of power. It would be a mistake to argue that the world has changed from geopolitics to geoeconomics as did some commentators in the early 1990s. The end of the cold war removed one set of military-related tensions but left some in place, as in the Middle East, and allowed others to emerge in the domestic violence of failed states in the Balkans and Africa. Even in East Asia, the scene of rapid economic growth, pre-cold war rivalries maintain the relevance of military force. It is interesting (and consistent with what we wrote about the indirect role of force in reducing insecurity) that contrary to early post-cold war predictions, NATO remains popular in Western and Central Europe, and the American military presence is welcome in most of East Asia. All markets occur within a political framework. To ignore the role of military security in an era of economic growth is like forgetting the importance of oxygen to our breathing. Even in some areas of high economic wealth, realist assumptions about the importance of military force still have considerable relevance. If the approximation of complex interdependence is less than some have predicted in regard to the realist assumptions about the role of military force and hierarchy of issues, it is certainly much greater in the third assumption relating to multiple channels of contact among societies. Here is the real change, and it brings us back to our work on transnational and transgovernmental politics. We see an order of magnitude shift as a result of the Internet. When we wrote in the 1970s, transnational contacts were growing, but the people involved were relatively small numbers of elites involved in multinational corporations, scientific groups, and academic institutions.14 Now the Internet is creating low costs of transnational communications for many millions of

people. The number of nongovernmental organizations (NGOs) increased from six thousand to twenty-six thousand in the 1990s, and some fifteen hundred NGOs used the Internet to coordinate their Seattle protest in 1999.15 Mixed coalitions of NGOs, some governments, and celebrity figures successfully pressed for a treaty banning landmines. Not only is there a great increase in the number of transnational and transgovernmental contacts compared with what we described two decades ago, but there has been a change in type. Earlier transnational flows were heavily controlled by large bureaucratic organizations, like multinational corporations or the Catholic Church, that could profit from economies of scale. Such organizations remain important, but the lower costs have now opened the field to loosely structured network organizations and even to individuals. These NGOs and networks are particularly effective in penetrating states without regard to borders and using domestic constituencies for agenda setting.16 Interactivity at low cost allows for the development of Page 170 → new virtual communities. While still in their infancy, transnational virtual communities are likely to grow and more complex identities and loyalties to develop. Whatever the future effects of interactivity and virtual communities, one political effect of increased information flows through multiple channels is already clear: states have lost much of their control over information about their own societies. Geographical communities still matter most, but governments will lose some of the “nontariff barriers” to information flow that protected officials from outside scrutiny. Rather than constituting an irrelevant ideal type, one of the assumptions of complex interdependence is rapidly becoming realistic enough to affect power resources that even hard-nosed realists must take into account (though many won't). This unlocks a rich agenda for research on globalization and governance. Contrary to some prophetic views, the nation-state is not about to be replaced as the primary instrument of domestic and global governance. There is an extensive literature on the effects of globalism on domestic governance that reaches more nuanced conclusions.17 Instead, the nation-state is being supplemented by other actors—private and third sector—in a more complex geography. The nation-state is the most important actor on the stage of global politics, but it is not the only important actor. If one divides the three sectors of social and political space in terms of three levels of activity, one obtains a nine-cell matrix, pictured in table 1. More governance activities will occur outside the box represented by national capitals of nation-states. James Rosenau's essay in this collection makes a similar point. Not only is the geography of governance more complex, but so are its modalities at all three levels. As Lawrence Lessig argues, governance can be accomplished by law, norms, markets, and architecture.18 Taking a local example, one can slow traffic through a neighborhood by enforcing speed limits, posting “children at play” signs, charging for access, or building speed bumps in the roads. Lessig describes an Internet world in which governance is shifting from law made by governments to architecture created by companies. “Effective regulation then shifts from lawmakers to code writers.”19 At the same time, private firms press governments for favorable legal regimes both domestically and internationally, as do actors from the third sector. The result is not the obsolescence of the nation-state but its transformation and the creation of politics in new contested spaces. Many writers in discussing the governance of globalism use what Hedley Bull referred to as the “domestic analogy.”20 It is commonplace for people to think of global governance as global government, because the domestic analogy is so familiar. Just as the nationalization of the American economy in the nineteenth century led to the nationalization of American government in the Progressive era, so globalization of the world economy should lead to world government. But the structure of federalism already existed in the United States, and it rested on a common language and political culture. (And even that did not prevent a bloody civil war in the middle of the century.) Page 171 → Another example is the United Nations World Development Report, which portrays global governance in terms of strengthening UN institutions. It calls, for example, for a bicameral General Assembly, an investment trust that

will redistribute the proceeds of taxes on global transactions, and a global central bank.21 But it is state structures, and the loyalty of people to particular states, that enable states to create connections among themselves, handle issues of interdependence, and resist amalgamation, even if it might seem justified on purely functional grounds. Hence, world government seems highly unlikely, at least in the absence of an overwhelming global threat that could only be dealt with in a unified way. In the absence of such a threat, it seems highly unlikely that peoples in some two hundred states will be willing to act on the domestic analogy for well into this century. Although world government seems unfeasible, one cannot be complacent about the effects of globalization without some coherent means of governance. Karl Polanyi made a powerful argument that the inability of polities to cope with the disruptive effects of nineteenth-century globalization helped cause the great disturbances of the twentieth century—communism and fascism.22 Without regulation—or what was traditionally known as “protection”—personal insecurity for many individuals can become intolerable. As Polanyi put it, “to allow the market mechanism to be sole director of the fate of human beings and their natural environment … would result in the demolition of society.”23 If world government is not feasible and laissez-faire is a recipe for a back-lash, scholars of international institutions need to search for an intermediate solution: a set of practices for governance that improve coordination and create safety valves for political and social pressures, consistent with the maintenance of nation-states as the fundamental form of political organization. Such arrangements will involve a heterogeneous array of agents—from the private sector and the third sector as well as from governments. And the governmental agents will not necessarily be operating on orders from the “top levels” of governments. The efficacy of these agents will depend on the networks in which they are embedded and their positions in those networks. And no hierarchy is likely to be acceptable or effective in governing networks. Keohane and I refer very generally to the governance structures we envisage as “networked democracies.” Networked—because globalism is best Page 172 → characterized as networked rather than as a set of hierarchies. Democracies—because governance at the global level will be legitimate only if it does not supersede democratic national governance and if its intrusions into the autonomy of states and communities are clearly justified in terms of cooperative results. To speak of “networked democracies” is, of course, not to solve the problems of global governance but merely to point toward a generic response to them. In particular, such a phrase begs the question of accountability, which is crucial to democratic legitimacy. Democracy is government by the people. In simplified form, this has meant the majority of the people (albeit with protections for individuals and minorities in liberal democracies). Historically, democracy has meant government by the majority of the people who regard themselves as a political community. The key question for global governance is who are “we the people” when there is no sense of political identity and community, and the political world is organized largely around a system of unequal states. In thinking about legitimacy, it may be helpful to separate the inputs and outputs of democratic government. On the input side, elections determine ruling majorities. But what are the boundaries of the relevant electoral constituencies in which votes are held? If the moral claim for democracy rests on the worth and equality of individuals, then a basic rule is one person, one vote. One state, one vote is not democratic because a Maldive Islander would have one thousand times the voting power of a citizen of China. A cosmopolitan view, however, that treats the globe as one constituency implies the existence of a political community in which citizens of 198 states would be willing to be continually outvoted by a billion Chinese and a billion Indians. As Pippa Norris has shown, there is no evidence that national identities are changing in a manner that would make that feasible for a long time to come.24 In the absence of such a community, the extension of domestic voting practices to the world scale would make little normative sense, even if it were feasible. Most meaningful voting, and associated democratic political activities, occurs within the boundaries of nation-states that have democratic constitutions and processes. Minorities are willing to acquiesce to a majority in which they may not participate directly because they feel they participate in some larger community. This is clearly absent at the global level and creates severe normative as well as practical problems for the input side of global democracy.25

At the same time, voting is not the only feature of the input side of democratic government. Many democratic theorists would argue that people should have a voice on issues that have important effects on their lives, and that voice is raised frequently in the long intervals between elections. The mechanisms stretch from polls to protests. The boundaries for this type of input are less clearly defined than in electoral constituencies. In a well-functioning domestic democracy, popular politics and the organization of interest groups lead directly to legislation and to the implementation of such legislation. These connections are lacking at the international level. Intergovernmental organizations (IGOs) that make binding rules often lack the democratic legitimacy that comes from having transparent Page 173 → procedures, institutional arrangements that facilitate accountability, and activities by politicians seeking reelection by appealing to publics. At the same time, the private and NGO sectors that agitate political issues internationally do not have any greater claim to democratic legitimacy. Despite their claims to represent civil society, they tend to be self-selected and often unrepresentative elites. The disjunction between international arrangements facilitating such public involvement and multilateral cooperation on binding decisions leads to disputes over legitimacy and dangers of stalemate in intergovernmental institutions. This suggests the need for a more appropriate measure for judging democratic legitimacy than the so-called democratic deficit based on the domestic analogy. The development of appropriate normative theory for judging global institutions will be an important part of the agenda for scholars in this field.26 It is unlikely that the literature based on the European Union with its close links to the domestic analogy is appropriate for global institutions for reasons already given. Nor will new theories based on the potential for direct voting in cyberdemocracy be sufficient. One can imagine technology enabling the world to engage in frequent plebiscites that collect the votes of vast numbers of people interested in an issue. But it is more difficult to envisage the effective processes of deliberation in the absence of community that would make such voting meaningful in a normative sense. With time, such obstacles may be overcome and practices gradually develop, but that is not imminent. In the interim it will be important to develop more modest normative principles and practices to enhance transparency and accountability not only of IGOs but of corporations and NGOs that constitute global governance today. For example, increased transparency is important to accountability, but transparency need not be instantaneous or complete—witness the delayed release of Federal Reserve Board hearings or the details of Supreme Court deliberations. Similarly, accountability has many dimensions, only one of which is reporting up the chain of delegation to elected leaders. Markets aggregate the preferences (albeit unequally) of large numbers of people, and both governments and transnational corporations are accountable to them. Professional associations create and maintain transnational norms to which IGOs, NGOs, and government officials can be judged accountable. The practice of “naming and shaming” of transnational corporations with valuable brand names by NGOs and the press also provides a sort of accountability. Similarly, the naming and shaming of governments engaged in corrupt practices helps create a type of accountability. While trisectoral cooperation and mixed coalitions are to be welcomed, competition among sectors and among mixed coalitions is useful for transparency and accountability. I cannot pretend to predict the information revolution's and globalization's future directions or their effects on international institutions. They may take unpredictable turns: some aspects may be reversed by historical surprises, as has happened in the past. World war is unlikely but not impossible, and it is worth rereading Polanyi on the failures of politics in the Page 174 → last century to cope with the effects of economic globalization.27 My argument is that the tools and concepts that Keohane and I developed to understand transnational relations and interdependence three decades ago still help us to understand something about the new world politics. And contrary to the view that nothing is cumulative in research in this area, some scholars are using those tools. Maybe we were right before our time, or maybe like a stopped clock that is right twice a day, we are just back at a lucky phase of a cycle. In any case, we have enjoyed the ride.

Notes 1. Joseph S. Nye Jr., “East Asian Security: The Case for Deep Engagement,” Foreign Affairs 74 (JulyAugust 1995). 2. Joseph S. Nye Jr., Understanding International Conflicts: An Introduction to Theory and History, 3d ed.(New York: Longman, 2000), ch. 1. 3. Ernst B. Haas, Beyond the Nation-State: Functionalism and International Organization (Stanford, Calif.: Stanford University Press, 1964). 4. See Leon N. Lindberg and Stuart A. Scheingold, eds., Regional Integration: Theory and Research (Cambridge: Harvard University Press, 1971); as well as Joseph S. Nye Jr., Peace in Parts: Integration and Conflict in Regional Organization (Boston: Little, Brown, 1971). 5. Ernst B. Haas, “Turbulent Fields and the Theory of Regional Integration,” International Organization 30 (spring 1976): 173–212. 6. Transnational Relations and World Politics, special issue of International Organization 25 (summer 1971). 7. Raymond Vernon, Sovereignty at Bay: The Multinational Spread of U.S. Enterprises (New York: Basic Books, 1971). 8. Robert O. Keohane, After Hegemony: Cooperation and Discord in the World Political Economy (Princeton, N.J.: Princeton University Press, 1984). 9. Miles Kahler, “Inventing International Relations: International Relations Theory after 1945,” in New Thinking in International Relations Theory, ed. Michael W. Doyle and G. John Ikenberry (Boulder, Colo.: Westview, 1997), 38. 10. Robert O. Keohane and Joseph S. Nye Jr., “Power and Interdependence in the Information Age,” Foreign Affairs 77 (fall 1998): 81–94. 11. Robert O. Keohane and Joseph S. Nye Jr., “Globalization: What's New? What's Not? (And So What?),” Foreign Policy, no. 118 (spring 2000): 104–19; David Held et al., Global Transformations: Politics, Economics, and Culture (Stanford, Calif.: Stanford University Press, 1999); James N. Rosenau, Turbulence in World Politics: A Theory of Change and Continuity (Princeton, N.J.: Princeton University Press, 1990). 12. Susan Strange, States and Markets, 2d ed. (New York: Pinter, 1994), 14, 73; Linda Weiss, The Myth of the Powerless State (Ithaca, N.Y.: Cornell University Press, 1998); Richard Rosecrance, The Rise of the Virtual State: Wealth and Power in the Coming Century (New York: Basic Books, 1999). 13. See for example the breathless account in Frances Cairncross, The Death of Distance: How the Communications Revolution Will Change Our Lives (Boston: Harvard Business School Press, 1997). 14. See Kjell Skjelsbaek, “The Growth of International Nongovernmental Organization in the Twentieth Century,” in Transnational Relations and World Politics, ed. Robert O. Keohane and Joseph S. Nye Jr. (Cambridge: Harvard University Press, 1972). 15. “A Semi-Integrated World,” The Economist, December 11, 1999, p. 42. 16. Margaret Keck and Kathryn Sikkink, Activists beyond Borders: Advocacy Networks in International Politics (Ithaca, N.Y.: Cornell University Press, 1998).Page 175 → 17. Robert O. Keohane and Joseph S. Nye Jr., introduction to Governance in a Globalizing World, ed. Joseph S. Nye Jr. and John Donahue (Washington, D.C.: Brookings Institution Press, 2000). 18. Lawrence Lessig, Code and Other Laws of Cyberspace (New York: Basic Books, 1999), 88. 19. Ibid., 207. 20. Hedley Bull, The Anarchical Society: A Study of Order in World Politics (New York: Columbia University Press, 1977), 46. 21. United Nations Development Program, Human Development Report, 1999 (New York: Oxford University Press, 1999). 22. Karl Polanyi, The Great Transformation (New York: Rinehart, 1944). See also John G. Ruggie, “International Regimes, Transactions, and Change: Embedded Liberalism in the Postwar Economic Order,” in International Regimes, ed. Stephen Krasner (Ithaca, N.Y.: Cornell University Press, 1983); as well as Dani Rodrik, Has Globalization Gone Too Far? (Washington, D.C.: Institute for International Economics, 1997).

23. Polanyi, The Great Transformation, 73. 24. Pippa Norris, “Global Governance and Cosmopolitan Citizens,” in Nye and Donahue, Governance in a Globalizing World. 25. Robert A. Dahl, “Can International Organizations Be Democratic? A Skeptic's View,” in Democracy's Edges, ed. Ian Shapiro and Casiano Hacker (New York: Cambridge University Press, 1999). 26. Karl Kaiser raised this issue in our volume on transnational relations. See his essay “Transnational Relations as a Threat to the Democratic Process,” in Keohane and Nye, Transnational Relations and World Politics. 27. Polanyi, The Great Transformation; Rodrik, Has Globalization Gone Too Far?

Page 176 →

ARE INSTITUTIONS INTERVENING VARIABLES OR BASIC CAUSAL FORCES? Causal Clusters versus Causal Chains in International Society Oran R. Young How should we think about the roles that institutions play as determinants of the content of collective outcomes in international society? Many main-stream analysts, whether they embrace the precepts of neorealism or lean toward the tenets of neoliberalism, assume (often implicitly) that we can understand these outcomes by envisioning causal chains in which institutions emerge as proximate forces coming into play only in the later stages of specific chains. They conclude, on the basis of this perspective, that institutions should be treated as intervening variables in contrast to basic causal forces.1 But this is by no means the only way to think about the role of institutions in international society, and some alternative approaches assign a more fundamental role to institutions. Taking a fresh look at this issue, I argue that the question articulated in the title of this essay obscures some of the more important features of causation in international society and, for that matter, in most other social settings. To fill the gap left by this conclusion, I provide an initial sketch of an approach to causation in international society that highlights causal clusters or sets of interactive driving forces in contrast to causal chains. Because my own substantive work deals with issues relating to natural resources and the environment, I turn to this realm repeatedly for illustrations of the theoretical arguments under consideration. Specifically, I refer regularly to issues pertaining to large atmospheric, marine, and terrestrial ecosystems in which growing anthropogenic impacts have given rise to human-dominated ecosystems.2 But the question about the roles that institutions play as determinants of collective outcomes is generic. It arises with respect to the full range of issues involving security, economics, and human rights as well as natural resources and the environment.

Mainstream Perspectives Most students of international relations adopt an instrumental perspective in which the development of institutions or, to be more specific, the creation Page 177 → of issue-specific regimes constitutes a—more or less successful—response human actors can and often do make to the emergence of collective-action problems. Resource regimes, on this account, are mechanisms for overcoming problems like the tendency of appropriators to overexploit living resources as a consequence of the dynamic widely known as the tragedy of the commons.3 Trade and monetary regimes are designed to avoid or overcome problems like trade wars arising from competing impositions of tariffs and other trade barriers in interactive situations. Arms control regimes are mechanisms created to avoid arms races in which participants invest large quantities of scarce resources without improving their security. Farsighted actors may anticipate the occurrence of such problems and take steps to create effective regimes before fish stocks become depleted, trade wars get under way, or arms races escalate. But that does not alter the logic underlying this view of the role of institutions in international society. The basic causal forces are factors such as power, interests, and knowledge, and the role of institutions is to intervene between these forces and collective outcomes in such a way as to prevent the occurrence of joint losses or to secure the achievement of joint gains.4 Starting from this instrumental perspective, individual analysts adopt varying attitudes toward the role of international regimes based on their assessment of the capacity of institutions to influence or channel the impact of the basic causal forces. In principle, this can give rise to a full spectrum of views, ranging from the perspective of those who believe that institutions have little capacity to influence the content of collective outcomes to the views of those who see institutions as effective mechanisms for solving, or at least alleviating, a wide range of collective-action problems.5 In practice, however, it is possible to identify a few clusters of positions that contain most of the mainstream perspectives on this issue.

One major cluster includes neorealists such as Susan Strange and John Mearsheimer who regard institutions as epiphenomena that have little or no independent impact on the content of collective outcomes. As Strange famously put it, regimes in international society are surface phenomena that reflect underlying bargains among states and other actors that wield the power to determine collective outcomes; they will experience more or less fundamental changes when the allocation of power among key actors shifts or these actors renegotiate the bargains on which regimes rest.6 As a result, according to Strange, “the study of regimes is, for the most part, a fad, one of those shifts of fashion not too difficult to explain as a temporary reaction to events in the real world but in itself making little in the way of a long-term contribution to knowledge.”7 Encapsulating this view, Mearsheimer has emphasized what he calls the “false promise of international institutions.”8 This relatively extreme view is captured graphically in panel a of figure 1 in which the causal arrow goes directly from basic forces like power to collective outcomes, and institutions are products of the operation of the basic forces that have little or no independent effect on the character of these outcomes. Yet not all those who adhere to the precepts of neorealism adopt such an extreme view regarding the role of institutions in international society. Page 178 → Those who focus on states as unitary and self-interested actors and who think predominantly in utilitarian terms, in particular, are apt to highlight the importance of collectiveaction problems9 and to see regimes as mechanisms that states can and sometimes do create to avoid or alleviate these problems.10 These analysts typically speak of the demand for regimes, direct attention to institutional bargaining, and make a concerted effort to understand the conditions under which regime members can be expected to comply with institutional rules and other obligations they assume under the provisions of specific regimes. Because international society lacks a public authority of the sort we would ordinarily think of as a government, compliance emerges as a critical issue for this cluster of analysts. Although they do not dismiss the relevance of decentralized compliance mechanisms altogether, these analysts expect the problem of compliance to become increasingly severe as the depth of cooperative arrangements embedded in regimes increases.11 This perspective is captured in panel b of figure 1a-b in which the width of the causal arrows indicates the relative importance of institutions and basic causal forces as determinants of the content of collective outcomes. For their part, neoliberals assign greater causal significance to institutions, even though they, too, adopt an instrumental perspective on the roles that institutions play. Members of this cluster expect the impact of issuespecific regimes to go beyond what can be deduced from conventional utilitarian Page 179 → premises. Partly, this is a matter of the role of the logic of appropriateness in contrast to the logic of consequences.12 Regime members frequently comply with rules or the results of institutionalized decision-making procedures because they regard these arrangements as legitimate and because they see compliance as a matter of proper conduct. In part, neoliberals treat regimes as sticky arrangements in which members participate through the force of habit or the operation of standard operating procedures and which, therefore, often remain in place even in the wake of power shifts or changes in the bargaining positions of key members. This does not mean that institutions are fixed and unchanging or that individual actors have no capacity to break out of the confines of specific regimes at the international level. But as panel c of figure 1 suggests, this way of thinking assigns substantially greater weight to the causal arrows running through institutions than either of the neorealist perspectives. Some students of international relations go a step further, drawing a distinction between the provisions of individual regimes and the deeper, more encompassing constitutive features of international society. On this account, international society itself is an institution encompassing constitutive Page 180 → rules governing membership and interactions among members, quite apart from the provisions of issue-specific regimes. It follows that collective outcomes are doubly affected by the force of institutions. The constitutive rules of international society shape issue-specific regimes that in turn influence the content of specific collective outcomes. Panel d of figure 1 portrays this causal picture by adding international society as a basic causal force, while maintaining the place of issue-specific regimes as intervening variables. Here, the instrumental role of institutions remains prominent. But now questions begin to arise about the persuasiveness of assigning institutions exclusively to the category of intervening variables in contrast to that of basic causal forces. Accordingly, this is a good point at which to turn to alternative perspectives on the roles that institutions play as determinants of collective outcomes in international society.

Alternative Perspectives The appeal of treating institutions as instrumental arrangements is not hard to understand. Unlike many other forces, institutions are subject to conscious design or manipulation on the part of human actors endeavoring to solve collective-action problems or, for that matter, seeking to promote any number of individualistic goals. Even if the operation of institutions accounts for only a modest proportion of the variance in collective outcomes, therefore, there is much to be said for a strategy of paying particular attention to matters of institutional design. Better to concentrate, in other words, on forces that human actors have some capacity to control than on causal forces that are important but generally beyond the reach of human control. But this is not the only way to think about the roles that institutions play in international society. In this section, I explore the view that institutions often emerge as basic causal forces in their own right, shaping the roles of a range of other factors treated as basic causal forces in most mainstream accounts. Consider first the well-known formula I = PAT devised to explain the impact of human actions on ecosystems ranging from small-scale local systems to global systems.13 According to this formula, we can explain the impacts (I) of humans on ecosystems as products of population (P), affluence (A), and technology (T). The perspective embedded in this formula is appealing; most accounts of the growing impact of human action on specific ecosystems point to some combination of population pressures, economic growth, and the introduction of powerful new technologies. But where do institutions fit into this causal model? At first glance, it appears that the model treats institutions as irrelevant or inconsequential when it comes to explaining the impacts of human actions on ecosystems. But a little more thought leads to the conclusion that this may not be the case at all. In fact, institutions often emerge as basic causal forces that play important roles in determining the values or shaping the operation of P, A, and T. The size, rates of growth, and distribution of human populations are all influenced, to a greater or lesser degree, by the character of prevailing institutions. Page 181 → One of the more striking determinants of human fertility across both space and time, for example, is the status of women. In systems where prevailing institutions accord women a full complement of rights, fertility rates are strikingly lower than in situations where the rights of women are sharply limited. The incentives of people to have children of their own are closely tied to the (non)existence of institutional arrangements under which states provide a range of services addressing health, education, and welfare needs. Movements of people from one place to another and especially across national borders, whether as guest workers, immigrants, or refugees, are subject to more or less elaborate rules articulated by national governments in most parts of the world. It is true, of course, that numerous individuals succeed in violating or evading these rules. But governments regularly invest substantial resources in efforts to enforce the rules, and debates over rules governing immigration are frequent and intense in the political discourses of most members of international society. Similar observations are in order regarding the institutional roots of affluence. Affluence in the I = PAT formula is essentially a measure of the role of economic growth as a driver of human impacts on ecosystems. It is not necessary to look far for convincing evidence of the roles that institutions play as determinants of economic performance. The new economic historians, for instance, have argued persuasively that institutional innovations involving property rights, contractual arrangements, and credit systems played an essential role in the rise of the Western world in the period from A.D. 900 to 1700.14 But there is no need to reach so far back in time for evidence of the roles that institutions play as determinants of economic performance. There is little doubt, for example, that many of the most severe problems the Russian Federation has encountered in making the transition from socialism to capitalism are attributable to the absence or underdeveloped character of a range of institutional arrangements that are necessary to the operation of market mechanisms but that are often taken for granted in societies that have well-developed market systems.15 At the international level, moreover, most analysts would agree that the introduction of a set of interlocking regimes (e.g., the General Agreement on Tariffs and Trade) explains a significant proportion of the variance in international economic relations between the first half of the twentieth century, which was marked by trade wars, and the second half of the century, which became an era of remarkable growth in world trade. The case of technology is equally if not more dramatic. There is little doubt, as prominent observers like Barry

Commoner have argued,16 that the introduction of new and more powerful technologies has played an important role as a driver of the impact of human actions on ecosystems. The growth of farming and fishing power, advanced transportation systems, and urban population centers attest to that. But note that both the development and the diffusion of technology are influenced—often dramatically—by the rules of the game relating to such matters as patents and other mechanisms for protecting intellectual property, product liability, and responsibility for the consequences of externalities. Not only would the incentives of developers Page 182 → be sharply limited in the absence of an effective patent system, but also corporations facing strict rules regarding product liability are likely to be markedly more cautious about bringing new products on stream than those who do not face such rules. In recent years, courts have begun to award punitive damages to those whose health has been severely impaired by the use of specific products (e.g., cigarettes). Imagine what would happen if chemical companies were held liable for the damages to large atmospheric, marine, and terrestrial ecosystems arising from the development, production, and consumption of ozone-depleting substances or a wide range of persistent organic pollutants (POPs). And as the negotiation of international agreements regarding these matters in recent years makes clear, there are good reasons to believe that the rules of the game regarding the development and diffusion of such technologies are now undergoing important changes. Lest the discussion of I = PAT seem esoteric to readers steeped in the traditional concerns of international relations, let me turn now to a consideration of the institutional roots of power, interests, and knowledge, factors that figure in one fashion or another as basic causal forces in most work in this field. There is no reason to deny that various combinations of power, interests, and knowledge loom large as causal forces in international society. Yet each of these forces is affected in turn by underlying institutional arrangements. Even power in the material sense is allocated and protected by the constitutive rules of international society dealing with matters like sovereignty and jurisdiction. Of course, great powers can and sometimes do ignore these rules, intervening in the domestic affairs of weaker states in ways that affect the distribution of power in the material sense. For the most part, however, the constitutive rules regarding such matters are simply taken for granted; great powers respect the jurisdiction of lesser powers not because they lack the capacity to intervene but rather because the rules of the game mandate such respect. Moreover, changes in the rules of the game can have far-reaching consequences for the distribution of material power in international society. The creation of exclusive economic zones during the 1970s and 1980s, for instance, allowed coastal states to assume control over about 11 percent of the world ocean and approximately 95 percent of the world's marine living resources.17 Nor is it any accident that states fight hard over rules governing access to resources like deep seabed minerals and the electromagnetic spectrum. Beyond these material concerns, it is worth noting as well that the rules of the game frequently affect the bargaining power of those who participate in institutionalized activities at the international level.18 Permanent membership coupled with the veto power in the UN Security Council is a dramatic case in point, particularly when we consider that permanent membership no longer correlates well with the distribution of power in the material sense in international society. But other interesting cases abound, including the rules governing weighted voting in the International Monetary Fund and the World Bank, the composition of the executive body of the Montreal Protocol Multilateral Fund, and the complex decision-making procedure employed by the Global Environment Facility. Page 183 → What is true of the institutional roots of power is all the more true of interests and knowledge. Those who have adopted a constructivist perspective and written about the agent-structure problem take the view that the identities of the actors in international society are determined, at least in part, by the character of the institutional settings in which they operate.19 On this account, the interests of individual actors, which are derived from their identities, are shaped in large part by the nature of international society treated as a set of constitutive institutional arrangements. But it is not necessary to adopt this strong constructivist perspective to see that issue-specific regimes can have substantial impacts on the ways in which actors think about and operationalize their interests. Interests are tied to the roles that actors play, and roles are affected by the rules of the game within which actors pursue their interests. The creation of exclusive economic zones, for instance, has given coastal states a proprietary interest in the use and management of large sections of the world ocean that actors previously regarded

as open-to-entry resources to be exploited on a competitive and unrestricted basis. Similarly, the creation of the ozone regime has given actors a clear-cut interest in the development of permissible alternatives to chlorofluorocarbons (CFCs) and a number of related ozone-depleting substances. As this example suggests, the impacts of regimes on actor interests may affect a variety of nonstate actors as well as the member states themselves. Thus, creation of the ozone regime has significantly affected the interests of major producers like DuPont as well as those of the governments of member states. But the basic message is clear. Actor interests are not exogenously determined and immutable. Changes in the rules of the game can and often do give actors incentives not only to frame new interests but also to redefine preexisting interests. Somewhat similar remarks are in order regarding the relationships between institutions and knowledge. Without doubt, the causal arrows run in both directions in this connection. Issue-specific regimes frequently reflect and embed both discourses that serve to structure thinking about the issues at stake in a general way and relatively well-defined models that provide the rationale underlying the rules and decision-making procedures of specific regimes.20 Consider the neoclassical economic models underlying the regime governing international trade and the logistics models that provide the foundation for many regimes dealing with living resources as cases in point.21 But issue-specific regimes also serve to legitimize and empower knowledge claims as well as to guide efforts to broaden and deepen knowledge in the relevant issue areas.22 Thus, the trade regime is not only based on neoclassical economic models—its existence and capacity to bestow legitimacy has also served to protect and strengthen those models from those who think in terms of dependency theory or models emphasizing the social and cultural destruction arising from the set of processes known as globalization. What makes the confrontation between the global regime dealing with biological diversity and many regional regimes addressing the conservation of living resources particularly significant is that the biodiversity regime calls for ecosystems management in which the welfare of all species is valued, while the conservation regimes typically focus on the goal of Page 184 → achieving maximum sustainable yields from commercially valuable and individually targeted species. It is not difficult to see both that the policy implications of these arrangements will often diverge and that the two sets of arrangements are likely to encourage and reward the production of different models and knowledge claims. Over time, successful regimes can have profound impacts on the ways in which we think about relevant issues and on the policies we adopt to solve or alleviate the problems that lead to the creation of issue-specific regimes.

Initial Implications What should we make of the argument set forth in the preceding section? Does it provide a compelling reason for proponents of mainstream perspectives to conclude that institutions are basic causal forces and revise their thinking accordingly? My sense is that most neorealists and, for that matter, neoliberals are likely to resist pressure to alter their views regarding the roles that institutions play as determinants of collective outcomes and for reasons that are perfectly understandable. But I also hope to show that the resultant interplay of ideas sets the stage for the introduction of an alternative view of causation that will improve our understanding of collective outcomes in international society and many other social settings as well. The obvious rejoinder for mainstream analysts to the argument of the preceding section is to reemphasize the idea of the causal chain and to insist that basic causal forces, such as power, underlie each institutional link in the chain. If I say that changes in the rules of the game (e.g., the creation of exclusive economic zones) have led to a reallocation of power in the material sense, for example, those who see power as a basic causal force can respond by saying ‘yes,’ but the changes in the rules themselves are best explained as the outcome of an exercise of power on the part of the dominant actors in the relevant setting. Similarly, if I assert that the decision rules used by bodies like the UN Security Council serve to strengthen the bargaining power of certain actors, those who emphasize the explanatory significance of power can say ‘yes,’ but the provisions of the UN Charter dealing with the decision-making procedures of the Security Council were dictated by the outlooks and preferences of the victors in World War II. In short, behind every institutional determinant of power, on this account, lies an antecedent exercise of power on the part of key players in the system. Similar responses are available to those who prefer to think in terms of interests and knowledge as driving forces

in international society. If the creation of exclusive economic zones has redefined the role of coastal states and provided those states with new interests, for example, those who espouse interest-based explanations of collective outcomes in international society can argue that the motive force behind the creation of exclusive economic zones was the interests of coastal states that worked hard in the institutional bargaining surrounding the adaptation of the law of the sea to ensure that their preferences would be accommodated in the provisions set forth in part V of the 1982 UN Convention on the Law of the Sea.23 For their Page 185 → part, analysts who emphasize the role of knowledge—or ideas more generally—are apt to see institutions as governance systems that grow out of and reflect developments in our understanding of the problems at stake. The creation of the ozone regime, on this account, owes a great deal to a shift in the discourse surrounding this issue that made efforts to control the production and consumption of ozone-depleting substances a legitimate activity, even before the specific biogeochemical mechanisms involved in ozone depletion were fully understood.24 The lack of consensus regarding a number of major aspects of climate change, by contrast, is often mentioned as a factor that impedes efforts to devise an effective regime dealing with emissions of greenhouse gases. How should those who see institutions as basic causal forces respond to these arguments? Should they concede to the mainstream perspectives regarding the distinction between basic causal forces, like power, interests, and knowledge, and intervening variables, like institutions? I can see no reason why they should. Naturally, all causal chains become more complex and difficult to comprehend the further we seek to follow them away from the phenomena to be explained. But those who emphasize the importance of institutions can respond to the arguments of those who favor explanations focusing on power, interests, and ideas in the same way that their counter-parts do, proceeding further along the causal chain to suggest that each of these factors is shaped by the impact of institutional arrangements located at some antecedent stage in the chain. They need only point to the fact that the bargaining power of those who worked hard to bring exclusive economic zones into existence was molded in significant measure by the institutional substructure of international society. Similarly, they can take the view that the growth of knowledge that ultimately produced the paradigm shift required to understand and take seriously the problem of stratospheric ozone depletion was shaped by the institutional setting within which large-scale atmospheric investigations are conducted. My objective in describing this rejoinder is not to demonstrate that the perspective of those who see institutions as basic causal forces is correct. Rather, the basic message I derive from this discussion is that neither side has a compelling case for the distinction it seeks to establish between basic causal forces and intervening variables.

What Is to Be Done? The obvious insight arising from this discussion is that efforts to answer the title question in terms of the idea of causal chains trigger a regress that has no natural stopping point. Those who see power as the basic driver will invariably find a way to explain institutional arrangements as surface phenomena that reflect underlying relations of power. But those who prefer to think of institutions as driving forces will inevitably counter with arguments that show how relations of power reflect underlying institutional arrangements. And the same goes for parallel arguments relating to forces such as interests and knowledge. What should we conclude from a recognition Page 186 → of this problem? My answer is that the title question assumes a model of causation in international society that has serious drawbacks, a conclusion that suggests it will be more productive to reexamine our thinking regarding the nature of causation in this realm than to devote more time to sorting out the merits of claims regarding basic causal forces and intervening variables. How should we proceed in light of this conclusion? The approach I propose and seek to support in the following paragraphs centers on the development of multivariate analyses in which collective outcomes in international society are treated as products of clusters of variables that operate more or less simultaneously rather than forming causal chains. To explore the implications of this proposal, I would begin by focusing on a set of empirically tractable dependent variables, such as trends in global carbon dioxide emissions, the volume of international trade, or expenditures on arms and armed forces. This would provide a clear-cut target for analysis along with a reasonable prospect for efforts to produce quantitative measures of changes in dependent variables thought to be associated with any of a number of independent variables.

Initially, this perspective suggests that we should seek to move forward by identifying a suite of independent variables and devising procedures for determining the relative weights of individual members of the suite or, in other words, the proportion of the variance in the dependent variable of interest that can be accounted for in a convincing manner by reference to specific independent variables. Panel a of figure 2 captures this perspective in graphic terms. In my own field, this approach has the attractive feature of allowing us to ask questions about the relative importance of anthropogenic forces (e.g., fishing effort) and nonanthropogenic drivers (e.g., changes in water temperature) in accounting for dramatic changes in large marine and terrestrial ecosystems (e.g., major shifts in species composition). But the method obviously allows us to ask about the relative weights of a range of social variables, such as power, interests, knowledge, technology, and institutions, as well. Anyone familiar with the basic ideas underlying statistical inference will realize immediately that this approach resembles a form of statistical analysis known as regression. This technique, widely used in both the social sciences and the natural sciences, provides a method for sorting out the effects of a number of variables that operate simultaneously as determinants of measurable outcomes in a variety of systems. In some respects, this is a satisfying conclusion in the context of the present discussion of approaches to causation in international society. But it takes only a little reflection to see that there are major problems with this line of thinking, even in cases where the dependent variable is empirically tractable. One problem arises from the well-known fact that universes of cases, at least in international society, are notoriously small and heterogeneous. In dealing with resource depletions, trade wars, and arms races, for example, we are seldom in a position to make use of the procedures of statistical inference in any quantitatively rigorous and analytically satisfactory manner. This problem is obviously Page 187 → serious. But it is not necessarily fatal. In some cases, there are opportunities to expand the universe of cases under consideration,25 and many commentators have suggested proceeding in a manner that corresponds as closely as possible to the logic of statistical inference, even in the analysis of qualitative data or case studies.26 Yet this line of thinking leads to a second and in some respects more fundamental problem with associating the model of causation I have proposed with the idea of regression analysis. As panel b in figure 2 suggests, we regularly find ourselves dealing with situations in which there are more or less complex and influential interactions between and among the members of the cluster of driving forces. In fact, the discussion in the preceding sections has already anticipated this observation. Not only do factors like power and institutions interact with one another, but also there are extensive interactions between and among forces like power, interests, and knowledge. To take some simple examples, it is easy to see that bargaining power is often a function of the location of an actor in a particular distribution or configuration of interests, and there is much to be said for the view that the possession or control of knowledge can be translated into power under a variety of conditions. Under the circumstances, it seems clear that a simple conception based on the logic of regression analysis will not solve Page 188 → the problem of devising an appropriate model of causation to serve as an analytic foundation for the efforts of those seeking to understand collective outcomes in international society. In my judgment, two other procedures for thinking about causation can help us to move beyond this impasse. One centers on what philosophers of science often describe as genetic explanations or tendency-finding analyses.27 In contrast to measures of association, which dominate statistical or variation-finding analyses, tendency-finding procedures seek to determine what combination of forces or factors, operating together, have brought about or produced a given outcome (e.g., rising levels of anthropogenic carbon dioxide emissions). The goal here is to build up explanations that can account for specific occurrences in a satisfactory manner rather than to develop and test generalizations of the sort we associate with scientific laws (e.g., statements of necessary and sufficient conditions) or, for that matter, probabilistic measures of association of the sort flowing from statistical inference. On this account, the general effects of individual causal factors may be much the same across a wide range of circumstances. Uncertainty regarding the nature of environmental problems, for instance, can be expected to have similar consequences with respect to efforts to solve many specific problems. But the particular combinations of forces that come together to produce collective outcomes in individual cases may be uncommon or, in extreme situations, unique. The challenge associated with this way of thinking is to isolate clusters of independent variables and to show how they work together to produce particular outcomes of interest to the analyst. Another procedure that is worthy of serious consideration in this context is dynamic systems modeling. As those

who have developed general circulation models to explore the dynamics of Earth's climate system have noted, this procedure makes it possible to simulate and, therefore, to explore the dynamics of systems in which a number of drivers operate simultaneously and usually interactively to determine the content of outcomes at the macro level. It is important to recognize that simulations of this type do not lead to predictions about the character of collective outcomes in specific real-world systems.28 But they do make it possible to explore different combinations of drivers and different values for key drivers in a systematic fashion. What makes this procedure attractive in the context of the current discussion is the prospect that we can use dynamic systems modeling to think systematically about interactions between institutions and other causal forces, including power, interests, and knowledge, as determinants of the content of collective outcomes in international society. The objective here is not to assign weights to individual variables in the interests of making statements about their relative importance as determinants of collective outcomes. Rather, this mode of analysis directs attention to the identification of causal mechanisms in which several variables interact with one another in a fashion that determines the content of outcomes that are systemic in character (e.g., overall levels of carbon dioxide emissions or trends in the total volume of trade at the international level). Page 189 → It is perfectly possible that the use of these procedures on the part of students of international relations will lead to the accumulation of a growing collection of explanation sketches or, to use a more elegant phrase, analytic narratives in which there are few patterns.29 But this need not be the case. It is entirely possible that we will discover a number of distinct patterns involving clusters or combinations of interactive driving forces that occur on a more or less regular basis. My own recent effort to shed light on the effectiveness of international environmental regimes certainly points in this direction.30 Again and again, I am finding that clusters of factors pertaining to the character of the problem, the nature of the actors, and the state of the social setting combined with various attributes of the regimes themselves can provide us with a convincing account of outcomes arising in specific issue areas (e.g., the production and consumption of CFCs, emissions of carbon dioxide and other greenhouse gases, losses of biological diversity).31 Needless to say, it will take time and sustained programs of research to justify any strong conclusions about the existence and prevalence of these patterns. But I can see no reason to be pessimistic at this stage about the prospects for success in efforts to identify recurrent patterns of this sort.

Conclusion My response to the question posed in the title of this essay, then, is that this is the wrong question to ask. This question makes sense only in the context of a model of causation in which the focus is on causal chains and the challenge is to determine the points in these chains at which different variables or types of variables come into play. If the argument I have presented in this essay is correct, this way of thinking about causation is not likely to get us far, at least as an approach to understanding the content of collective outcomes in international society. As an alternative, I suggest a model in which clusters or combinations of interactive variables operate together to determine the content of collective outcomes. The challenge here is not to assign weights to individual variables in an effort to determine which causal forces are the more important ones. On the contrary, we should be thinking about the various ways in which interactive clusters of variables lead to causal mechanisms that leave us with the feeling that we have a satisfactory and satisfying understanding of the reasons why collective outcomes in international society take specific forms. Finally, let me emphasize that nothing in this account precludes the adoption of an instrumental approach to the design and operation of international regimes. Although institutional factors usually constitute only a single element in clusters of driving forces, it is nonetheless true that they are the factors that are most conducive to conscious control or manipulation on the part of human actors. No doubt, power relations, interests, and knowledge claims all change in more or less significant ways over time. But it is hard to think of effective methods to control or channel these changes in the interests of solving common problems or promoting the common Page 190 → good. It is perfectly possible, by contrast, to launch concerted efforts to design and create new institutions or to redesign existing institutions. In the realm of environmental affairs, the last several decades

have witnessed a steady stream of efforts to form new regimes and to reform existing regimes. In this connection, it is essential to bear in mind not only that institutional design is a political process featuring bargaining rather than social engineering but also and especially that institutions always affect the content of collective outcomes as elements in clusters of driving forces. As a result, a regime that performs well in dealing with one problem may prove far less successful or even fail completely as a response to other problems. This makes it easy to understand why one size does not fit all problems even in a single issue area and why it is essential to design institutions that fit well with a range of other forces at work in connection with specific problems. It follows that a cookbook approach to regime formation will produce poor results and that we should be extremely cautious in adopting arrangements that prove successful in one setting (e.g., ozone depletion) in the construction of regimes to deal with problems arising in other settings (e.g., climate change).32 Yet nothing in these remarks suggests that there is anything in-appropriate about the current interest in the formation of issue-specific regimes as a means of addressing a variety of important problems in international society.

Notes 1. Stephen D. Krasner, “Structural Causes and Regime Consequences: Regimes as Intervening Variables,” in International Regimes, ed. Stephen D. Krasner (Ithaca, N.Y.: Cornell University Press, 1983). 2. Peter Vitousek, Harold Mooney, Jane Lubchenko, and Jerry Melillo, “Human Domination of the Earth's Ecosystems,” Science 277 (1997): 494–99. 3. Garrett Hardin, “The Tragedy of the Commons,” Science 162 (1968): 1343–48. 4. Krasner, “Structural Causes and Regime Consequences.” 5. David A. Baldwin, ed., Neorealism and Neoliberalism: The Contemporary Debate (New York: Columbia University Press, 1993). 6. Susan Strange, “Cave! Hic drangones: A Critique of Regime Analysis,” in Krasner, International Regimes, 337–54. 7. Ibid., 337. 8. John Mearsheimer, “The False Promise of International Institutions,” International Security 19 (1994–95): 5–49. 9. Russell Hardin, Collective Action (Baltimore: Johns Hopkins University Press, 1982). 10. Robert O. Keohane, After Hegemony: Cooperation and Discord in the World Political Economy (Princeton, N.J.: Princeton University Press, 1984). 11. George W. Downs, David M. Rocke, and Peter N. Barsoom, “Is the Good News about Compliance Good News about Cooperation?” International Organization 50 (1996): 379–406. 12. James G. March and Johan P. Olsen, “The Institutional Dynamics of International Political Orders,” International Organization 52 (1998): 943–69. 13. Paul R. Ehrlich and John P. Holdren, “Impact of Population Growth,” Science 171 (1971): 1212–17. 14. Douglass C. North and Robert Paul Thomas, The Rise of the Western World: A New Economic History (Cambridge: Cambridge University Press, 1973). 15. Mancur Olson, Power and Prosperity: Outgrowing Communist and Capitalist Dictatorships (New York: Basic Books, 2000).Page 191 → 16. Barry Commoner, The Closing Circle (New York: Bantam Books, 1972). 17. Alf Hakon Hoel, “Performance of Exclusive Economic Zones” (scoping report prepared for the Project on the Institutional Dimensions of Global Environmental Change, Dartmouth College, Hanover, N.H., 2000). 18. Oran R. Young, International Governance: Protecting the Environment in a Stateless Society (Ithaca, N.Y.: Cornell University Press, 1994). 19. Alexander Wendt, Social Theory of International Politics (Cambridge: Cambridge University Press, 1999). 20. Karen T. Litfin, Ozone Discourses: Science and Politics in Global Environmental Cooperation (New

York: Columbia University Press, 1994). 21. John Gerard Ruggie, “International Regimes, Transactions, and Change: Embedded Liberalism in the Postwar Economic Order,” in Krasner, International Regimes ; J. A. Gulland, The Management of Marine Fisheries (Seattle: University of Washington Press, 1974). 22. Virginia Walsh, Background Knowledge in World Politics (forthcoming). 23. James K. Sebenius, Negotiating the Law of the Sea (Cambridge: Harvard University Press, 1984). 24. Litfin, Ozone Discourses. 25. Ronald B. Mitchell, “Quantitative Analysis in International Environmental Politics: Toward a Theory of Relative Effectiveness,” in Regime Consequences: Methodological Challenges and Regime Strategies, ed. Arild Underal and Oran R. Young (forthcoming). 26. Gary King, Robert O. Keohane, and Sidney Verba, Designing Social Inquiry: Scientific Inference in Qualitative Research (Princeton, N.J.: Princeton University Press, 1994). 27. Ernest Nagel, The Structure of Science: Problems in the Logic of Scientific Explanations (New York: Harcourt, Brace and World, 1961); David Dessler, “The Architecture of Causal Analysis” (paper presented at the Harvard Center for International Affairs, Harvard University, 1992). 28. Bert Bolin, “Scientific Assessment of Climate Change,” in International Politics of Climate Change: Key Issues and Critical Actors, ed. Gunnar Fermann (Oslo, Norway: Scandinavian University Press, 1997), 83–109. 29. Robert H. Bates, Avner Greif, Margaret Levi, Jean-Laurent Rosenthal, and Barry R. Weingast, Analytic Narratives (Princeton, N.J.: Princeton University Press, 1998). 30. Oran R. Young, ed., The Effectiveness of International Environmental Regimes: Causal Connections and Behavioral Mechanisms (Cambridge: MIT Press, 1999). 31. Oran R. Young, Governance in World Affairs (Ithaca, N.Y.: Cornell University Press, 1999). 32. Lawrence E. Susskind, Environmental Diplomacy: Negotiating More Effective Global Agreements (Oxford: Oxford University Press, 1994).

Page 192 → Page 193 →

ALTERNATIVE AND CRITICAL PERSPECTIVES

Page 194 → Page 195 →

ALTERNATIVE AND CRITICAL PERSPECTIVES Steve Smith This chapter is a rewritten version of the paper originally presented on one of the millennial reflections panels, but the basic argument and structure of the paper remain the same. I have kept the argument and structure largely because I was one of the few who wrote my paper in exactly the format requested by the organizers; that is, I wrote a paper that addressed each of the six questions the panel organizers asked. My fellow panelists interpreted the invitation in different, and equally legitimate, ways, but none of them answered the precise exam paper they had been sent! At the panel, one of them, Yosef Lapid, responded in detail to the text of my (precirculated) paper, and he clearly had important disagreements with the way I had characterized the current state of the discipline. The other participants raised important points about the problems of characterizing, compressing, labeling, and representing “alternative and critical perspectives.” In rewriting my paper for publication, I have decided to retain my main arguments and claims, mainly because it was clear that some of the other participants would want to disagree with or dissent from my paper when they came to write up their own. Thus, what follows is an extended version of the paper as presented, and this will hopefully allow the other participants to continue their disagreements with me in print. It would have been scarcely ethical for me to change my arguments, when I knew that some of them wished to take those arguments as their starting points! Tempting, but unethical! Like all participants on the panel I was asked to focus on six tasks, all designed to get me to “confront [my] own limitations, stimulate debate about their most significant accomplishments and shortcomings, and discuss research paths for the years ahead.” I interpreted this as an invitation to assess where alternative and critical perspectives stand at the start of the new millennium. Aside from the obvious fact that these approaches involve a variety of different (and sometimes contradictory) positions since there are few assumptions that they share, I was also troubled by the expectation of offering a definitive statement of how things stood. Thus, what follows is a very partial and personal account of the development of these approaches. Page 196 → My other initial health warning is simply that I was not really sure what “alternative and critical perspectives” included. But from the lineups of the other panels I deduced that the panel should focus on critical theory and postmodernist accounts of international relations (IR). But within that general label, there is a real problem over gender and feminist work. Although one panel dealt solely with feminist and gender scholarship, it struck me as interesting, to say the least, that those scholars were not involved in the panel on which I had been invited to participate. My presumption was that the organizers saw this as a way of giving feminist/gender work its own platform, rightly signifying the importance of the work in this area of the discipline. Yet, paradoxically, this decision seemed to backfire, as it was widely seen as a way of marginalizing feminist/gender work into its own ghetto. But probably more important than this was that the division of work into feminist/gender on the one hand and alternative and critical on the other created the unfortunate implication that feminist/gender work was not alternative or critical. This is not a statement about the motives of the organizers, but it evidently did impart a dynamic to the alternative/critical panel and certainly made me feel decidedly uncomfortable about the absence of women on the panel. I said this in my presentation but nonetheless got criticized for taking part in an all-male panel (dealing with, of all things, alternative and critical work) given the enormously important role that female scholars have played in the development of these parts of the discipline. With this proviso about the implied content of the terms alternative and critical, my paper will concentrate on what is usually defined as nonmainstream theoretical work within IR. The very move of delineating what is and is not alternative and critical work within the discipline of IR has enormous political and professional consequences, especially if those who define what counts as alternative and critical perspectives are mainstream scholars. Let me be clear, I am not so much concerned with the precise boundaries that are drawn between mainstream and alternative scholarship, nor about the boundary between alternative and feminist work; rather I want to stress the enormous implications of being able to define X (whatever X is) as “an alternative” or as “critical.” Think of the

implications of such a move: the very notion of an alternative carries with it a set of connotations, as does the label critical (though note that the term critical is the chosen description of a school of thinking that self-consciously wants to challenge conventional ways of interpreting and explaining the social world). It may not have been the organizers' intention, but the label does imply that these approaches are not “normal,” and it even implies a certain stigma of being more than a little fashionable and wacky, providing alternatives or being merely critical. I now turn to the six set tasks, in the order that they were presented to me. I. Engage in self-critical, state-of-the-art reflection on accomplishments and failures, especially since the creation of the ISA more than forty years ago. My first comment is that it is far too easy to overemphasize the achievements of what can broadly be called alternative approaches to IR. I think Page 197 → that alternative perspectives have not really made significant inroads into the mainstream of the discipline, for one simple reason: they do not share the epistemological and methodological assumptions of that mainstream. I think this is the key issue. Many discussions of alternative perspectives focus on their different ontological assumptions, and thus have argued that their alternative status derives from the fact that they look at nonmainstream issues. Such was evidently the case with peace research and environmental approaches; more recently a similar response has been made to feminist work within IR. But I think this is not the main judgment that is at work. Rather, I think that the alternative perspectives are seen as such because they do not work within the model of social science that dominates the U.S. IR academic community (and thus the world community of IR scholars). Hence the incredible power of the critiques of a number of writers who claim that alternative approaches are not part of the social science enterprise. Let me note three examples: the first is that of Keohane, Krasner, and Katzenstein in their 1998 International Organization paper,1 when they argue that postmodern work in IR is not really social science and therefore cannot be assessed by the usual academic standards. They argue that the journal has published little postmodern IR work “since IO [International Organization] has been committed to an enterprise that postmodernism denies: the use of evidence to adjudicate between truth claims. In contrast to conventional and critical constructivism, postmodernism falls clearly outside of the social science enterprise, and in IR research it risks becoming self-referential and disengaged from the world, protests to the contrary notwithstanding.”2 A second example is the now famous claim made by Robert Keohane in his 1988 address as president of the International Studies Association (ISA). Noting that reflectivist approaches had important criticisms to make about rationalism, he nonetheless noted that their greatest weakness was “the lack of a clear reflective research program. … Until the reflective scholars or others sympathetic to their arguments have delineated such a research program and shown in particular studies that it can illuminate important issues in world politics, they will remain on the margins of the field, largely invisible to the preponderance of empirical researchers, most of whom explicitly or implicitly accept one or another version of rationalistic premises.”3 Reflectivist scholars needed to develop “testable theories” without which “it will be impossible to evaluate their research program.”4 The most significant point about this challenge was that it was, not surprisingly, made on the epistemological terrain of rationalism: it is frankly impossible to see just how reflectivist accounts could conceivably provide answers that Keohane would accept given the gap between their epistemological starting points. The final example of this delegitimization of reflectivist approaches comes in Stephen Walt's review of the state of international relations theory.5 Although Walt notes that the central debate in IR theory remains that between realism and liberalism, he notes that there is an alternative, but this is not reflectivism or any of its components; rather it is constructivism. His only discussion of any reflectivist work is the following reference to the Page 198 → “deconstruction” approach: “because these scholars focused initially on criticizing the mainstream paradigms but did not offer positive alternatives to them, they remained a self-consciously dissident minority for most of the 1980s.”6 Even constructivism comes out of this survey badly. He lists its “unit of analysis” as “individuals” and its “main instruments” as “ideas and discourse.” Its “post-cold war prediction” is “agnostic because it cannot predict the content of ideas,” and its “main limitation” is that it is “better at describing the past than anticipating the future.” Not only are alternative and critical approaches ignored in his overview, but constructivism receives the following disciplining of its role: “The ‘compleat diplomat’ of the future should remain cognizant of realism's emphasis on the inescapable role of power, keep liberalism's awareness of domestic forces in mind, and occasionally reflect on constructivism's vision of change.”7

These claims have enormous professional consequences for anyone working in the alternative perspectives. They immediately delegitimize such scholars' work as not serious research, and one can only guess at the implications of this for their careers. In short, I do not think that the alternative perspectives have achieved much in the way of challenging the mainstream of the discipline for the simple reason that they are dismissed as outside the standards of social science. Such is the political power of assumptions about epistemology. This does not mean that there have been no successes, only that those successes have never changed the nature of the mainstream. Taken together, the works of the writers such as those who appeared on the panel with me have made it impossible for the discipline to pretend that there are no criticisms of the dominant assumptions. The work of Cox, Walker, Ashley, and Lapid, along with that of Shapiro, Der Derian, Campbell, Linklater, Sylvester, Enloe, Tickner, and many more, has offered a sustained critique of the assumptions of the mainstream literature and has thus opened up major avenues for research. But, to repeat, this does not mean that this research has been accepted as relevant or legitimate by the mainstream. It certainly means that much of the work in the leading graduate schools is informed by the work of these writers, but this has yet to have a significant impact on the everyday politics, psychology, and sociology of the profession. Why is this? In my view it is not that the work carried out in alternative perspectives is less academic or less thorough; rather it is that the mainstream has tended to reconsolidate itself so as to continue the U.S. social science project. The classic case has been the fate of neorealism since Waltz's 1979 book Theory of International Politics was published. Despite the critiques of it, and despite the rather inconvenient fact that the structure of the international system changed in a way unforeseen by the theory, neorealism has remained a central part of the mainstream and now constitutes one of two main poles of that mainstream, the other being neoliberalism. A similar and very significant development is occurring with social constructivism. What I find amazingly important is the way that (a particular version of) social constructivism is becoming part of the developing mainstream. Note how Wendt in his new book8 places himself firmly on the side of IR Page 199 → as social science: “Epistemologically, I have sided with positivists. … our best hope is social science.”9 He is, he states, “a strong believer in science. … I am a positivist.”10 This position is, presumably, one opposed to “Parisian social theory.”11 Similarly, note the co-opting noises made by Keohane, Krasner, and Katzenstein, and by Stephen Walt, in their reviews noted previously. By these moves the mainstream can say that it deals with the ontological questions raised by the alternative perspectives, but that it does so within the canons of U.S.-style social science. I have written about the dominance of the discipline by the U.S. IR academic community for the last twenty years, 12 but in an important sense for a long time I missed the point. I tended to see the reason for this as the importance of the United States in world politics and the effect of this on how the discipline has evolved. Whilst I would not want to ignore that point, I now see the impact of the U.S. model of social science as absolutely critical in explaining the lack of impact of alternative perspectives. They simply do not seem to fit within what the discipline defines as the right way to carry out academic study; together they question the four main tenets of U.S. social science—that there is a distinction between facts and values, that empiricism is the appropriate epistemology, that naturalism is possible, and that the task of the researcher is to uncover regularities in the social world. What is of central importance in all this is that these assumptions are not incompatible with studying identity, perceptions, culture, or normative concerns: it is only that they are to be studied from a social science perspective. It is in this way that the concerns of the alternative perspectives can be dealt with by the mainstream, and in a way that fits comfortably within the pressing and powerful norms of social science U.S.-style. In this light the main failure of the alternative perspectives is that whilst they have done much to undermine the epistemological assumptions of the mainstream, they have not succeeded in establishing an alternative epistemology. This is not surprising for two main reasons: first they disagree amongst themselves over questions of epistemology (look at the differences between, say, Robert Cox and Rick Ashley or Rob Walker to take an example from the other panel participants). Second, the grip on the discipline of positivism U.S.-style is so great as to close down any space for alternatives. This is the nub of the issue, since I genuinely do not think that any alternative perspective can replace the mainstream if it does not fit within positivism, and here of course is the problem: alternative perspectives cannot fit because they are not based on positivism. This has considerable influence on the development of the discipline in the rest of the world, where IR communities have to deal with

not only the ontological and policy commitments of U.S. IR but also the dominance of a particular view of social science. This situation has been brilliantly analyzed by Ole Waever in a recent paper.13 As Waever notes: “The internal intellectual structure of American IR explains both the recurring great debates and why American IR generates global leadership. It has a hierarchy centered on theoretical journals, and scholars must compete for access to these. This they have not had to do in Europe, where power historically rested either in sub-fields or in local universities, Page 200 → not in a disciplinary elite. American IR alone generates an apex that therefore comes to serve as the global core of the discipline.”14 Waever argues persuasively that in the future there will be a parting of the ways between the work going on in the U.S. IR community and that of the other main IR communities, especially those in Europe. The U.S. IR community will tend toward “national professionalization,” with the mainstream concentrating on rational choice theory. These methods will not “travel well” to IR communities such as Germany, France, and the United Kingdom, and those communities will both develop their own national identities and, at the same time, combine more to form a European IR community. These trends will result in “a slow shift from a pattern with only one professional and coherent national market—the United States, and the rest of the world more or less peripheral or disconnected—toward a relative American abdication and larger academic communities forming around their own independent cores in Europe.”15 These, then, are some of the main issues at the start of the new millennium concerning the politics of international politics, and the social relations of international relations. 2. Assess where we stand on key debates and why we have failed to resolve them. My answer to this question follows directly from what I have already argued. The key debates are defined as those within the mainstream of the discipline, and any debate with alternative perspectives is axiomatically defined as impossible because the alternatives do not adhere to the same forms of scholarship as are found in the mainstream. In my view the mainstream has been very successful both in setting out the Kuhnian normal science puzzles and in stating the ways in which these kinds of questions should be answered. I view the neo-neo synthesis to be an extraordinarily powerful boundary-making and boundary-enforcing exercise for defining and assessing the key debates. Alternative perspectives do not fit within these boundaries. Similarly, note the way in which the Wendtian arm of social constructivism is now setting out the terrain of the next key debate with the neoliberalist arm of the neo-neo synthesis as one of the ways in which ideas construct and influence policy choices. I can easily see this becoming the key debate, and all within the U.S. model of social science.16 In short, I would argue that the key debates have been very largely solved within the mainstream. I think that the neo-neo debate does involve some genuinely important and complex questions, mainly those to do with whether anarchy can be mitigated by institutions and whether state actors pursue relative or absolute gains. The social constructivism/neoliberal institutionalism debate will likewise focus on the important issue of the role of ideas in international relations; but there is much more to international relations than this, and I can see no easy way in which the alternative perspectives can enter into any debate with the mainstream if the standards for scholarship are so positivistic and so fixed. Within the alternative camp, there is some debate between rival perspectives, but frankly there is little agreement on epistemological matters. Thus critical theory requires exactly the kind of foundational epistemology that postmodernism denies is possible, just as standpoint feminism stands Page 201 → in contrast to postmodern feminism. Alternative may be a good catchall term, but any term that includes work inspired by writers with as distinctive positions as Gramsci, Habermas, Butler, Spivak, Virilio, and Foucault hides more than it reveals. Within each of the alternative perspectives there is lively debate around a series of research questions, but these cannot be expected to add up to one alternative perspective to challenge the mainstream. Just as it is naive to ask for the feminist perspective on an issue, so is it misguided to ask for the alternative perspective on an issue. But do note how politically powerful it is to ask for that answer and then to consign the “well, it all depends on which alternative you are talking about” response to outside the walls of the academy, to the land of nonserious research and a (regretful) overlap between the explanatory and the normative. In summary, the question assumes two things: first, that the key debates are debates open to all; second, that they

can be resolved. The first assumption is false because I do not think that the mainstream is willing to assess alternative perspectives on any terms other than its own, which, by definition, they do not accept. The second assumption is false because it is itself an assumption based on a positivist view of inquiry. Maybe the key debates cannot be resolved because there are no answers, if by answers one means the kind of social scientific answer that positivism assumes. Maybe, the key debates revolve around not questions of explanation but questions of understanding, or even possibly questions of ethics or morality. Resolving requires giving an answer, but what if there is no one answer? What if the answers are ultimately dependent on which “facts” one selects and what one sees as the main issues in the debate? Think for a second about the concept of globalization: how might one resolve a debate about its impact on world politics if one cannot agree on who is the relevant unit for assessing this impact, or if one sees it as having both positive and negative effects? Resolving the debates may simply not be possible; indeed the belief that it should be is itself just one example of how deeply ingrained are the assumptions of positivistic social science in the U.S. academic community. 3. Evaluate what intrasubfield standards we should use to evaluate the significance of theoretical insights. This is an important question because a common version of the problems of the alternative perspectives is that they have no standards for demarcating good academic work from bad. This is especially the charge leveled at postmodern work, with the usual version of the charge being that since for postmodernism anything goes, there can be no standards for deciding between good and bad scholarship. This was clearly a major issue on the panel, as it indeed is with any attempt to bring together mainstream and alternative approaches and then discuss them in terms of standards of assessment. The alternative and critical approaches tend to resist any attempt to make the kinds of assessments about “progress” or “success” that characterize mainstream assessments of the field. The reason is obvious: the language of “assessment,” “progress,” and “success” smacks of a modernist notion of knowledge, whereby we are all involved in a process of learning more and more, with a hint of an increasing ability to use this knowledge in the world of policy and politics. The organizers Page 202 → have tried to be open in their guidelines for participants to write up these chapters, stressing that they do not expect conformity to one model of social science: instead we have been asked to “at least define what you consider to be a fair measure of success /failure in [sic] regard to your own subfield, and then assess the extent to which core objectives (whatever they may be) have or have not been met, and why.” The problem here for alternative and critical approaches is the familiar one of not being willing to buy into this model of assessment. This is clearly a fundamental difference between the mainstream and the alternative and critical literatures, and I see no easy way to bridge it. My own view, for what it is worth, is that the acid test for the success of alternative and critical approaches is the extent to which they have led to empirically grounded work that explores the range and variety of world politics. My “core objective” has been to create space for thinking and for politics, and to refuse attempts to determine what counts as scholarship on the basis of a culturally and historically specific notion of social science. My concern is not so much to push a particular view of what counts as knowledge but to examine the (often hidden) assumptions of those approaches that claim to have solid epistemological foundations, so as to show that these foundations are far less secure than those who propound them like to think. With regard to the criteria used to assess alternative and critical work, I have not found that there is an absence of standards for distinguishing between good and poor scholarship. This is simply not the case in my experience. I would like to make my point by referring to my own direct experience of this matter. I edit the Cambridge University Press series Studies in International Relations. It is a big series, with more than seventy-five books published or under contract, and in excess of 150,000 copies sold since it published its first volume in 1988. In that time the series has had more than 750 submissions, with a success rate of about one in ten. In the 1990s we have attracted a lot of work from what would be considered the alternative perspectives. Now, in all that time, I have never felt that there was anything particularly problematic about intrasubfield standards in alternative perspectives. Referees have seemed to have no problems in making tough judgments as to the scholarly standing of manuscripts, and since the usual practice has been to have two or three referees for each submission, it is important to note that there has been no noticeable difference between the pattern of referees' reactions in the alternative perspectives and those found in the mainstream. I think this is a very important point: of course there are gatekeepers and inflexible scholars working in the alternative perspectives, but there are certainly no more of them than one would find in, say, social constructivism or neorealism. It does not appear any more difficult to

determine intrasubfield standards in the alternative perspectives than in any other area. Of course, there can and should be debate over what these are, and it is obviously the case that those working within the alternative perspectives do not appeal to the relatively straightforward (if demanding) standards of positivist scholarship. But there does appear to be roughly the same degree of agreement among, say, feminist scholars over what constitutes Page 203 → scholarship in their field as one finds in the mainstream. These standards tend to revolve around the categories of coherence, tight argumentation, knowledge of the relevant literature, the use of empirical evidence, persuasiveness of the major analytical moves, innovative scholarship, and the standard of writing and expression. To put it bluntly, I have sent out quite a few manuscripts for review and have had back very tough and ultimately negative reports from individuals working precisely in that part of the alternative and critical perspectives. It is not the case that referees support proposals because they work in the same area. However comforting it might be for those who deride this work as not being in accordance with the canons of social scientific inquiry to say that adherents to a given alternative perspective promote the work of others in that area, that is simply not the case in my experience. The problem is not one of intrasubfield standards; the problem is one of intersubfield standards, and here the situation is that the standards of the mainstream skew the situation. They become the standards for assessing work and are never seen as themselves open to critique or question. They are presented as if they were beyond reproach, beyond debate, and beyond the realm of politics. They are the standards of the academy, and that is that. In my view, of course, they are the place where the discipline needs to begin its self-examination at the start of this millennium. 4. Explore ways to achieve fruitful synthesis of approaches, both in terms of core research questions and appropriate methodologies. As can be deduced from the foregoing comments, I have problems with this question. The obvious reason is that any answer I give to it will end up disciplining, by definition, the content of the discipline of IR, and I do not want to do that. Rather than cut down the area of study, rather than prescribe what should be the research questions and the appropriate methodologies, I want instead to insist on a pluralism about research questions and methods. I am interested in supporting the best scholarship within any of the approaches. Another personal experience might be relevant: when Cambridge University Press (recently voted the number-one publisher in political science and IR by the American Political Science Association) undertook its latest fiveyearly review of the series I have edited since 1986, the press asked a lot of academics what they thought of the list. The most interesting feature of the U.S. responses (as opposed to the responses from other parts of the world) was the perception that the series lacked an identity, a position on what to study and how to study it. This was seen as a weakness: the series needed to have an identity and a common approach. That was never the idea behind the series: as far as I was concerned the whole point of it was to publish the best scholarship across the entire gamut of approaches to IR. This was the first statement in the initial proposal to Cambridge University Press, and I think it remains as relevant today as it was then. I think that this says a lot about the difference between what can broadly be termed U.S. and European/Australasian IR. Accordingly, I think the way forward is to insist only on research being carried out to the highest standards possible as defined in the research area. I want the series to publish the best postmodern feminist work, the best Page 204 → neorealist work, the best quantitative work, and so on. I think this is the right way to answer the question of what are the core research questions and the appropriate methodologies. Any other route, say by defining the core questions according to the policy interests of the United States, or in accordance with the views of the dominant approaches in the discipline, results in a closing down of space for scholarship. As I mentioned, my interest is in opening it up and trying as far as possible to assess scholarly standards by reference to the best standards in each perspective. Let me end this answer with two further personal experiences: in 1993 Martin Hollis and I submitted an article on the agent-structure debate to a leading U.S. journal, arguing that there were always at least two stories to tell about the social world (explaining and understanding). We noted that there could be no common standard for evaluating the rival stories, since the epistemologies and methodologies behind each were so very different. Specifically we noted that quantitative data could not be used to solve the debate. Well, the article was refereed and given a revise and resubmit, with the main requirement being the use of quantitative empirical evidence to demarcate between the two sets of claims. We sent it to the leading U.K. journal, which, after getting three referees' reports, accepted

it as it stood. Note, I am not claiming that the U.S. journal should have accepted it, only that its standards were based on an overarching view of scholarship that precluded almost by definition work that was not explanatory social science. The second example is that in 1995 Ken Booth and I published the edited book International Relations Theory Today.17 It has sold well (including several thousand copies in its U.S. imprint) and has been well reviewed. But our U.K. publisher initially discussed the book with one of the two or three leading U.S. presses, and that press had it refereed. The interesting thing is that the press liked the book but asked us to rewrite it so that it had a clear theme and argued for a specific view of international theory. Our point in editing the book, of course, was to bring together a variety of different approaches. Now, it may be that the book wasn't good enough for the press and that its reason was a nice way of rejecting our work, or it may be the case that the themes were not well integrated, but my honest view is that the press simply didn't like the idea of a set of essays arguing for different interpretations of IR. I know of the dangers of arguing from one's own experiences, but I do think that these are interesting reflections of the dominance of a particular view of how research should be undertaken and what it should be studying—as is the above task when it speaks of synthesis as the way forward. Epistemological assumptions lie deep, and they affect far more than how we study: in practice they end up determining what we can study, how we assess our studies, and how they should be related together (as compatible parts of one social world that exhibits regularities and is “knowable” from the outside). 5. Address the broader question of progress in international studies. Frankly, I do not see much progress in IR. This is because I remain convinced that the discipline has remained wedded to a particular view of the world throughout the period of the ISA's history. Broadly speaking, I accept John Page 205 → Vasquez's view of the “power of power politics,” the “colouring it Morgenthau” thesis.18 I think that the ontological assumptions of contemporary mainstream IR theory are little changed from the assumptions that dominated the discipline when the ISA was founded. I see the history of IR since the war as one of a continuing debate between realist and liberal worldviews, and it is particularly interesting to note how the two of them have come together in the neo-neo synthesis. The U.S. study of IR has always had its liberal and its realist wings (as Brian Schmidt has convincingly shown in his recent book),19 and I see the differences between these two positions as relatively minor when compared with the wider political spectrum, or the world of social theory. Think of the similarities between the current phase of mainstream IR theory and that of the late 1950s and early 1960s, or that of the early 1970s: in all cases the “debate” in the discipline was between broadly liberal (integration studies, the nonstate actor literature respectively) and a dominant realist account. The mainstream has, in other words, remained relatively unchanged over the last half century. Of course, this is not the dominant self-image of the U.S. IR community, which presents the story as one of progress in terms of the refinement of specific models and theories of IR. The liberal peace theory and neoliberal institutionalism are two such “progressive” research paradigms. But in my view, they are not so much progressive as political views of IR: they define a set of parameters as given and then examine behavior within those parameters, never looking at what aspects of world politics are excluded from the analysis. These debates take place within a powerful framework that defines the political in a very specific way, and which excludes a whole set of “politics” (defining these as “not political” or as “private” or as “domestic politics”). Thus economic power is separate from political power, and even the dominant international political economy work in the United States can end up presenting economics and politics as two separate things that need to be combined. Thus, the liberal peace theory defines both peace and democracy in a very U.S. way, and neoliberal institutionalism deals with the role of ideas in constituting actors' identities and interests in an equally specific (and normative) way. In this light, IR presents itself as a discipline that explains (and I use the word deliberately) international politics. Its progress comes, it is said, from its growing sophistication and explanatory power. My view is that IR does far more than this; specifically I see IR theory as partly constituting the practice of international relations, and therefore I see its progress as more instrumental than emancipatory. Ultimately, IR is unavoidably normative, both in the narrow sense that by choosing to study some subjects rather than others it normalizes and naturalizes the world, and in the more general sense that it is overwhelmingly focused on explaining that world. I maintain that explanation is not the name of the game in the social world, and to argue that it is leads to a view of the social

world that is ultimately limiting and normative in its choice of what to study and how to study it. This is not to deny that there have been significant developments in IR theory since the ISA was founded, only that these developments have all Page 206 → taken place within a very restricted view of the political, the international, and the social. 6. Select an agenda of topics and research questions that should guide the subfield during the coming decades. I am not going to pronounce on the topics and questions for all the perspectives presumably included under the label alternative, although my general view is that the issues that will be studied by the profession at large will continue to be informally dominated by the policy agenda of the United States, and these will be studies in an essentially positivistic way (although one significantly influenced by the Wendtian form of social constructivism). Referring to the work in the alternative perspectives, in my view it is for those who work in each of those areas to study those aspects of IR that they want to study. I decidedly do not want to prescribe what should be studied or how it should be studied. I have my own intellectual biases and views on these questions, but that is all they are—judgments and opinions, fallible, hopefully humble, and always provisional. Just as it would be entirely wrong for me as editor of a series (or as journal referee) to reject a book proposal (or article) because of the topic or the approach or to try and stop a scholar being appointed to my own or any other department because of his or her approach, so it should decidedly not be down to me or to anyone else to lay out the agenda for others to study; more important, even, I do not think that a commitment to studying certain issues in a given way should become a view that only that issue or that way is legitimate. My views of what might be studied, then, argue a case; they do not seek to delegitimize other views. Having said which, what I do want to do under this heading is to make three points. First, I hope that theoretical work, of whatever persuasion, should be informed by, and in turn inform, empirical material. There is nothing as practical as a good theory, and I find the most valuable work to be that which is empirically grounded. There will be some work dealing with metatheoretical issues, but I do think that this has its place. I write as someone who has spent much of the last decade dealing with metatheoretical issues, but during that time I have also published a lot of empirically grounded material. Indeed, my publication list to date splits almost equally between metatheory and empirical work. Interestingly, I got heavily involved in metatheory only because my empirical work on arms control and foreign policy seemed to run into some rather serious theoretical problems. Second, I hope that the discipline will foster intellectual pluralism and open rather than close down analytical space. I would like to see work evaluated by the standards of its own perspective rather than by pretending that there is one undisputed metatheoretical standard of scholarship for all approaches, especially when that approach takes such a dogmatic, and blinkered, position on what constitutes a social science. Finally, I want to return to the notion of alternative perspectives. The simple and beguiling move of classifying perspectives as alternative is in my view an act that disciplines the discipline. However well-meaning and however much it is intended to give a voice to those perspectives that might Page 207 → otherwise be ignored, my worry is that the simple act of defining perspectives as alternative reinforces that definition and their standing in the profession. Presumably they are alternatives to the dominant or mainstream: they are not “normal” they are not “standard” they are not the most important perspectives. A genuine attempt to be pluralistic may end up reinforcing the very marginality of the perspectives. My worry is that the very act of defining perspectives as alternative and critical performs an important gatekeeping task. My hope is that these comments stimulate discussion and debate on the current state of the discipline. Despite the thrust of some of the foregoing comments, I am in fact quite heartened by the developing pluralism in areas of the discipline. Whether one focuses on journals or on the panels at conferences, one sees a variety of perspectives in the discipline, and I feel that this is a sign of a vigorous and healthy discipline. However, I remain concerned that this pluralism does not extend into the heartland of the discipline in the United States. I wait to see significant numbers of leading alternative and critical scholars appointed to the most prestigious departments in the United States. I remain to be convinced that one historically and culturally specific model of social science is not the touchstone for judgments about academic quality, be that about the fate of Ph.D.'s or about the fate of job applicants.

In conclusion, I am worried about two main features of the discipline at the turn of the millennium. The first is the sheer dominance of the U.S. IR community. A good example was provided by the simple mathematics of the millennium panels. Although I know that the organizers were trying to ensure that a variety of voices were heard, those were overwhelmingly North American, male, and white voices. There were ten panels on the program, with two (“alternatives” and “gender”) dealing with the nonmainstream. Interestingly, there was no place for social constructivism. But in terms of who spoke for the discipline, I noted that according to the list of confirmed participants, of the forty-nine speakers only six were from outside North America, and four of those six appeared on the feminist theory panel (the other two were myself and Mick Cox, who appeared on the “editors” panel). In terms of male/female, only twelve of the forty-nine speakers were women, and five of those twelve were on the feminist theory panel. Indeed, had it not been for the feminist theory panel, there would have been only two nonNorth Americans and seven women out of the forty-four speakers! I am much less certain when it comes to identity in terms of color, but I could only readily identify one nonwhite participant from the list. I think these figures tell us a lot about the kind of IR that dominates the discipline at the start of the millennium. My second concern is even more fundamental and influential, and it is the main theme running through this paper and indeed my work over the last decade. I need to say little more about it at this juncture, except to reiterate it: it concerns the continuing dominance in IR of a very specific definition of social science. It is this dominance that makes alternative perspectives alternative, that links the discipline of IR to Page 208 → the practice of international relations, and that, by naturalizing the social world, has constituted the hidden politics of the profession of IR since, as well as before, the founding of the ISA.

Notes 1. Peter Katzenstein, Robert Keohane, and Stephen Krasner, “International Organization and the Study of World Politics,” International Organization 52 (1998): 645–85. 2. Ibid., 678. 3. Robert O. Keohane, International Institutions and State Power: Essays in International Relations Theory (Boulder, Colo.: Westview, 1989), 173. 4. Ibid., 173–74. 5. Stephen Walt, “International Relations: One World, Many Theories,” Foreign Policy, no. 110 (1998): 29–46. 6. Ibid., 32. 7. Ibid., 44. 8. Alexander Wendt, Social Theory of International Politics (Cambridge: Cambridge University Press, 1999). 9. Ibid., 90. 10. Ibid., 39. 11. Ronald Jepperson, Alexander Wendt, and Peter Katzenstein, “Norms, Identity, and Culture in National Security,” in The Culture of National Security: Norms and Identity in World Politics, ed. Peter Katzenstein (New York: Columbia University Press, 1996), 34. 12. See Steve Smith, ed., International Relations: British and American Perspectives (Oxford: Blackwell, 1985); Steve Smith, “The Development of International Relations as a Social Science,” Millennium: Journal of International Studies 16 (1987): 189–206; Steve Smith, “Hegemonic Power, Hegemonic Discipline? The Superpower Status of the American Study of International Relations,” in Global Voices: Dialogues in International Relations, ed. James Rosenau (Boulder, Colo.: Westview, 1993), 55–82; Steve Smith, “Foreign Policy Theory and the New Europe,” in European Foreign Policy: The EC and Changing Perspectives in Europe, ed. Walter Carlsnaes and Steve Smith (London: Sage, 1994), 1–20; Steve Smith, “The Self-Images of a Discipline: A Genealogy of International Relations Theory,” in International Relations Theory Today, ed. Ken Booth and Steve Smith (Cambridge, England: Polity Press, 1995), 1–37; Steve Smith, “Positivism and Beyond,” in International Theory: Positivism and Beyond, ed. Steve Smith,

Ken Booth, and Marysia Zalewski (Cambridge: Cambridge University Press, 1996), 11–44. 13. Ole Waever, “The Sociology of a Not So International Discipline: American and European Developments in International Relations,” International Organization 52 (1998): 687–727. 14. Ibid., 726. 15. Ibid., 726. 16. See Steve Smith, “Wendt's World,” Review of International Studies 26 (2000): 151–63. 17. Booth and Smith, International Relations Theory Today. 18. John Vasquez, The Power of Power Politics: A Critique (London: Frances Pinter, 1983); see new edition published as The Power of Power Politics: From Classical Realism to Neotraditionalism (Cambridge: Cambridge University Press, 1998). 19. Brian Schmidt, The Political Discourse of Anarchy: A Disciplinary History of International Relations (Albany: State University of New York Press, 1998).

Page 209 →

UNIVERSALITY IN INTERNATIONAL STUDIES A Historicist Approach Robert W. Cox One of the problems in participating on a panel entitled Alternative and Critical Perspectives is that the very nature of “alternative” is diversity and nonconformity. There can be no single alternative perspective nor any alternative school. Nor can I think of alternative as a “subfield.” Alternative is a residual category for all who are not considered to be “mainstream.” It has a clearer exclusionary meaning for those who consider themselves to be mainstream than it does for those who are so excluded. As a participant in an alternative panel, I can only write about the intellectual problems that confront me in my own work. I cannot discuss alternative perspectives in general. “Critical” is a different matter. It describes an approach toward understanding and action. So far as I am concerned, there is no “critical school,” no orthodoxy of criticism. It is, however, possible to distinguish a critical approach to a topic from other approaches that express different purposes—description, explanation, restoration, reconstruction, promotion, for example—and to ask what the critical approach consists of, how the critical theorist approaches the subject. In this brief paper I focus on one theme that I found important in my own work: the issue of universality versus relativism (which includes the issue of positivism versus historicism). Among the tasks assigned to the panel members by the conveners, what I have to say probably relates most of all to item four—synthesis of approaches and appropriate methodologies—and item six—agenda of topics and research questions.1 I approach the theme mentioned from the standpoint of critical theory, and I ask where it leads me toward a research program. Critical theorists have been accused of not having a substantive research program. In my perspective (and I speak only, of course, for myself) the question of a research program follows from the nature of the critical approach itself. I would outline this approach through three questions: 1. What is critical theory and how has it come into international studies? 2. What is the core philosophical issue inherent to critical theory?Page 210 → 3. What ontology can the critical approach propose as a framework for inquiry and action?

Critical and Problem-Solving Theory Problem-solving theory takes the existing world as its framework and considers how things work within that framework and how disturbances or dysfunctions may be remedied. It is essentially synchronic in its way of conceiving reality. The existing world is a given (datum), a set of supposedly stable relationships that may be disturbed. The analyst is concerned with dysfunctions, with homeostasis, and with the restoration of equilibrium. Critical theory addresses itself to the frameworks for action, how these frameworks came into being, what are the conflicts or contradictions that call them into question, and what directions of change in frameworks are possible and desirable.2 Economics and political science have been primarily problem solving in their approaches—dealing with the issues that concern them respectively within the assumed or “given” frameworks of markets and of existing sets of political institutions. Critical theory takes a diachronic perspective, one of history informed by sociological theory. It deals with facts, that is, “mades,” rather than data or “givens,” and, as Giambattista Vico argued, facts are to be understood through the minds of their (individual or collective) makers, that is, as products of history.3 The analyst, the historically oriented social scientist, goes through a double process of historical thinking. The first phase is self-consciousness of one's own historical time and place, which determines the questions that claim attention. The second phase is the effort to understand the historical dynamics that brought about the conditions in which these questions arise.

Critical theory in recent decades has entered international studies through the revival of international political economy (IPE), although, of course, some earlier writers such as E. H. Carr were critical scholars. Initially, IPE was a branch of political science addressing the politics of economic policy.4 It evolved, however, into the study of structural change in the world economy—that is, the focus came to be on changes in the organization of the economy and the changes in social and political structures and in cultural values associated with economic change. While primacy was given to the economic as what is to be explained, other factors in change came to have comparable weight, depending upon the topic and the theorist. IPE remained a convenient name but would be a misnomer if it were to obscure the attention required in an examination of structural change to demography, migration, gender, race relations, cultural differences, and humanity's place in the biosphere. IPE opened the way to a comprehensive critical inquiry into structural change and also to the infusion of a normative concern, whether for maintenance of the status quo or for emancipatory action, as the purpose of inquiry. Page 211 →

The Core Philosophical Issue: Universality The issue of universality arises in two ways: about norms and about knowledge. Are there universal norms, for instance about human rights? Can there be a universally knowable definitive, final, complete form of human knowledge about anything? These questions take us to assumptions about human nature and the human mind that are the basis for considering both forms of universality. We can start our inquiry with the broad consensus of the European Enlightenment that human nature was universal and unchanging in time and space (e.g., in both Hobbes and Rousseau) and that political truths could be deduced from this construction. The Enlightenment, by and large, had a low opinion of history as an inferior, imprecise kind of knowledge. The natural sciences were the model in which scientists were observers of physical phenomena whose regularities could be demonstrated to be universal laws. The assumption of a universal human nature opened the way to explore the possibility that human activity in history and society could be treated like the phenomena of physical nature by social scientists who would assume the position of external observers and catalogue human activities in search of universal laws. Already in the eighteenth century a dissident strand of European philosophy challenged the hegemony of the Enlightenment. In this dissident perspective the study of history and society was fundamentally different from physical science. Kant, of course, made the distinction between the phenomena that were the object of scientific observation and the “thing-in-itself” that could not be known. Giambattista Vico, before Kant, reversed the Enlightenment view about the supremacy of natural science over what we would now call social science by affirming that a thing could only be truly known by its maker. For Vico, the physical universe could only be known truly in the mind of God, its maker, whereas human history was intelligible to the minds of human beings who are the makers of history. Nineteenth- and twentieth-century theorists like Wilhelm Dilthey, Benedetto Croce, R. G. Collingwood, and Martin Heidegger pursued the analysis of historical thought as a distinct tradition. This led initially to the sense of a dichotomy between the world as nature and the world as history. The same distinction between positivism and historicism emerged within Marxism in the thought of Antonio Gramsci. Following Vico, human nature could no longer be regarded as universal. Human nature was a creation of history, made and remade by human beings, and therefore changeable over time (albeit very slowly) and diverse in space. The key point here is that, for Vico, there was enough similarity among human minds for one mind to be able empathetically to reproduce the mental processes of another so as to be able to understand minds differently shaped by history. Today, a more extreme position would deny that possibility, affirming that human natures have no common foundation, that they are distinct products of historical development, which in Page 212 → effect denies any possibility of historical or social knowledge that is not particularistic.5

In the late twentieth century, the classic distinction between an “objective” positivist natural science and a subjective (or better, intersubjective) historical and social science has been narrowed from the natural science side. The second law of thermodynamics and entropy, the relativity of space and time, Heisenberg's principle of uncertainty, the introduction of reflexivity recognizing that the scientist is part of the experiment and not just an external observer, complexity in cybernetic systems, and chaos theory (recognizing that there is an order in chaos) have all meant that chance gains over predictable determinism.6 Contemplation of the infinitely big and the infinitely small, the cosmos and the atomic particle, has made the conventional laws of physics less absolute, more relative. The gap between the natural and social sciences has been reduced. Notions of universality in the natural sciences have been weakened so that the recognition of reflexivity in historical and social knowledge no longer need seem so inferior as it did to the Enlightenment philosophers. The subjects of knowledge, actual human beings, the concrete instances of the species Homo sapiens, are products of a long evolution. Homo sapiens has existed for somewhere between fifty thousand and one hundred thousand years, having evolved out of hominids during a passage of some four million years. Some people have speculated upon a biological evolution beyond Homo sapiens, involving a significant increase in brain cells, trading in the “end of history” (of Homo sapiens) for a “posthuman” world, which according to some such speculations, could lead to a self-organizing network of beings not needing formal institutions. The future is open and unpredictable. The speculation does, however, underline that our humanity is itself a product of the very longue durée of biological history that, having had a beginning, will by implication come to its end through either self-destruction or a long biological evolution.7 Our capacity for reasoning is, for practical purposes, stuck with the condition of Homo sapiens. And this reason we know is not a means of access to knowledge about Kant's noumenal world. It is a practical tool for making the best of the particular predicaments in which various specimens of Homo sapiens find themselves. Reason is colored by the circumstances that give rise to our use of it. Knowledge is reflexive—it knows itself in relation to its specific historical experience. Accordingly, several kinds of knowledge coexist as the intersubjective truths of different groups of people. This brings us back to Vico's challenge: can a person participating intersubjectively in one kind of truth, in one perception of reality, painstakingly reproduce the intersubjective meanings of another group and its truth. Carl Jung has suggested a possible answer in defining three levels of consciousness and unconsciousness: First is personal consciousness, which is the common awareness we have of our perceptions and reasoning. This consciousness can be reflexive, that is, self-aware of the conditions in which it is being applied.Page 213 → The second level is the personal unconscious, which includes the occulted influences that affect our conscious awareness and may bias our perceptions and evaluations. The deepest level is the collective unconscious consisting of archetypes, that is, myths or forms that have a variety of different contents among different peoples and different eras but, as archetypes, are common to all humanity beyond time and space. This third and deepest level is the closest thing to universality in Jung's thinking. It could be the basis, the common foundation, upon which rests the possibility of penetrating the minds of others constituted differently by history. Jung's method is indeed reminiscent of Vico's in tracing myths and symbols common to the different “nations.”8 Even this third level, the closest to universality, would only be universal for the era of Homo sapiens, the longest durée. And the implication that common archetypes have been given different content in the stories and myths of different peoples suggests that what is universal as a starting point has become diverse in time and space while at the same time retaining its common base in the human psyche. It should therefore be possible to retrace the development of diverse mentalities because of this common foundation. What, then, is to be done?

In recognizing that values are products of history for particular peoples and particular times, does this thereby undermine those values? That is what relativism sometimes is held to mean. The dilemma was confronted by Isaiah Berlin when he wrote (quoting Schumpeter, I believe, but without attribution): “To realize the relative validity of one's convictions and yet stand for them unflinchingly is what distinguishes a civilized man from a barbarian.” And Berlin went on: “To demand more than this is perhaps a deep and incurable metaphysical need; but to allow it to guide one's practice is a symptom of an equally deep, and far more dangerous, moral and political immaturity.”9 The difficulty is that there is more than one way of becoming civilized, and each way has its barbarians to cope with. As between different civilized values, active pretensions to universality are ultimately reducible to power. Moreover, the context in which absolute values are advanced is always complex. Absolutist claims obscure other motives of an economic, geopolitical, or institutional kind. There are no absolutely pure motives as there are no just wars. In such a context of divergence or conflict in values, we are led to Max Weber's distinction between an ethic of ultimate ends and an ethic of responsibility.10 The ethic of ultimate ends makes absolutist claims and brooks no compromises. The ethic of responsibility looks to the likely consequences of action eschewing absolutist claims.

Ontology and a Research Program An ontology for critical research (the specification of the key issues and foci of study) depends upon purpose—whether to maintain existing values and Page 214 → structures or to transform them, whether to conceive the future world as homogeneous or pluralistic. In a purpose geared to restoration and maintenance, it may be appropriate to focus on specific institutions or practices that have become dysfunctional, assuming stability in the whole. Where the purpose is for change and transformation and where homogeneity is abandoned in recognition of pluralism, the ontology has to be broadened to include all those factors that will have a bearing upon the fixing of goals for societies and the processes of transformation. In the latter perspective, states and markets remain key institutions, but they function within a broader framework. Popular forces have become suspicious of both states and markets. In the established democracies large segments of the public have become alienated or nonparticipant in formal political processes; and under authoritarian regimes public participation is repressed. In both cases, efforts at self-organization articulate dissent in a new kind of politics. The “battle in Seattle” of December 1999 when civil society groups from North America and many other parts of the world confronted what they saw as the alliance of state and corporate powers in the World Trade Organization illustrated both the breadth and the diversity of dissent at both national and transnational levels. Similar confrontations have followed with other high-level meetings of managers of world and regional economies. A focus on civil society gives an indication of the issues and the forces for change. The concept of civil society itself requires reformulation. It no longer has the connotation it had for Hegel or Adam Smith as the realm of private interests distinct from the public interest realm of the state. That eighteenth-century notion of civil society was equated with the bourgeoisie. Today's civil society distinguishes itself from both the state and economic power and seeks to bring both under social control. Encompassing the totality of activities in state, markets, and civil society is the common sense of an era and of a community, or the set of intersubjective meanings that give people a sense of reality and enable them to communicate meaningfully with one another. I would equate this with civilization, and since there are coexisting forms of common sense, there are coexisting civilizations. Civilizations exist in people's minds, but they are formed in the practices and institutions through which people confront their material conditions of existence. The understanding of civilizations is to be approached through the interaction between common sense ideas and materially conditioned practices.11 Understanding civilizations enables one to understand how people whose common sense is differently formed by history perceive and interpret events differently. Civilizations are not, however, a methodological a priori. They are shaped by the development of civil society in the longue durée.

There is a reciprocal relationship or dialectic between civil society and civilization in which states and markets may be mediating factors. Civil society is a terrain of conflict among social forces. Exclusionary as well as inclusionary forces compete; right-wing populism and fascism are probably stronger now than at any time since the 1930s.12 Page 215 → Covert forces—organized crime, intelligence services, money-laundering banks, the drug and arms trades, people smuggling, mercenary military forces, even hermetic religious cults—influence surreptitiously the outcomes of overt political and economic processes. One hypothesis is that the weaker civil society is, the greater space there is for the influence of the covert world. A more developed civil society enhances transparency and accountability. The totality of human activity takes place within the biosphere within which humanity is interdependent with other forms of life and life-sustaining substances. The future of humanity depends upon evolving ways of living that maintain the biosphere, and this imposes a categorical constraint upon ecologically destructive practices and is an injunction to collective change in lifestyles. This in turn raises the ethical issue of who is to make the changes and of equity in sharing the burden and benefits of change. Finally, there is the matter of social cohesion. This is ignored by methodological individualism but has been one of the oldest questions of political philosophy. Ibn Khaldun called the quality required to make a state ‘asabiya, a group consciousness generated in the warrior band. Machiavelli called the quality virtú. He thought his own contemporaries were too corrupt to possess it and looked to a prince to embody it. Gramsci sought for it in the political party (the modern prince).13 Today we may look for it in the revival of civil society. If you accept that a primary purpose for international studies is to prepare the ground for a pluralist world of diversity in social organization and ethical norms, then a research program toward this end would give emphasis to the study of civilizations, civil society and its complex context including the covert world, the maintenance of the biosphere, and the problem of social cohesion for the maintenance of collective entities and their institutions. These constitute the framework for understanding and coping with the more immediate issues of poverty, racism, discrimination, and violence. In this approach, “universality” has a different meaning from that bequeathed by the Enlightenment. In the Enlightenment meaning universal meant true for all time and space—the perspective of a homogeneous reality. In the approach just sketched, universality means comprehending and respecting diversity in an ever-changing world.

Notes 1. I have discussed some of these points in “Social Cleavages and the Way Ahead: Towards a New Ontology of World Order,” in Critical Theory and World Politics, ed. Richard Wyn Jones (Boulder, Colo.: Lynne Rienner, 2001). 2. I have discussed problem-solving and critical theory in “Social Forces, States, and World Orders: Beyond International Relations Theory,” in Robert Cox with Timothy J. Sinclair, Approaches to World Order (Cambridge: Cambridge University Press, 1996), 87–91. 3. Giambattista Vico, The New Science of Giambattista Vico, trans. from the 3d ed. (1744) by Thomas Goddard Bergin and Max Harold Fisch (Ithaca, N.Y.: Cornell University Press, 1970). 4. For example, in the textbook by Joan Spero, The Politics of International Economic Relations, 1st ed. (New York: St. Martin's, 1977). Earlier, Susan Strange Page 216 → in “International Economics and International Relations: A Case of Mutual Neglect” (International Affairs 46, no. 2 [1970]: 304–15) had appealed for a rapprochement of the two disciplines. Strange later articulated in States and Markets (London: Pinter, 1988) a structural view of political economy that contributed to a critical approach. 5. Richard Rorty, as I read him, affirms that there is no core human nature, no residue once the creations of

history have been peeled off such as Vico thought, and that only the experience of pain is common to human beings (and to animals). In Contingency, Irony, and Solidarity (Cambridge: Cambridge University Press, 1989), Rorty sees no common nature in the past but the possibility of creating one in the future. 6. Among the earliest exponents of this approach to the natural sciences was Jacques Monod, Le hasard et la nécessité (Paris: Seuil, 1970). His thinking was extended with implications for the social sciences by Edgar Morin, Science avec conscience (Paris: Fayard, 1982). An American overview is in M. Mitchell Waldrop, Complexity: The Emerging Science at the Edge of Order and Chaos (New York: Touchstone, 1993). 7. See, for example, Edgar Morin, Le paradigme perdu: La nature humaine (Paris: Seuil, 1973). 8. Jung wrote: “When I first took this direction I did not know where it would lead. I did not know what lay in the depths of the psyche—that region which I have since called the ‘collective unconscious,’ and whose contents I designate as ‘archetypes.’” Modern Man in Search of a Soul (New York: Harcourt Brace, 1933), 241. Compare this with Vico: “To discover the way in which this first human thinking arose in the gentile world, we encountered exasperating difficulties which have cost us the research of a good twenty years. We had to descend from these human and refined natures of ours to those quite wild and savage natures, which we cannot at all imagine and can comprehend only with great effort.” New Science, para. 338. Jung's brief definition is: “We mean by collective unconscious, a certain psychic disposition shaped by the forces of heredity; from it consciousness has developed. In the physical structure of the body we find traces of earlier stages of evolution, and we may expect the human psyche also to conform in its makeup to the law of phylogeny.” Jung, Modern Man in Search of a Soul, 165. 9. Isaiah Berlin, Two Concepts of Liberty (Oxford: Clarendon, 1958), 57. Berlin has also written, referring to Vico and Herder: “In the house of human history there are many mansions: this view may be unChristian; yet it appears to have been held by both these pious eighteenth-century thinkers.” “This doctrine is called pluralism. There are many objective ends, some incompatible with others, pursued by different societies at various times, or by different groups in the same society, by entire classes or churches or races, or by particular individuals within them, any one of which may find itself subject to conflicting claims of incompatible yet equally ultimate and objective ends. Incompatible these ends may be but their variety cannot be unlimited, for the nature of men however various and subject to change, must possess some generic character if it is to be called human at all. This holds, a fortiori, of differences between entire cultures.” Isaiah Berlin, The Crooked Timber of Humanity (London: Fontana, 1991), 79–80. 10. Max Weber, “Politics as a Vocation,” in From Max Weber, ed. H. H. Gerth and C. Wright Mills (London: Routledge and Kegan Paul, 1948), 118–28. 11. I have tried to develop this idea of civilizations in “Civilizations and the 21st Century: Some Theoretical Considerations,” International Relations of the Asia-Pacific 1 (2001): 105–30. 12. I have expanded on this point in “Civil Society at the Turn of the Millennium: Prospects for an Alternative World Order,” Review of International Studies 25, no. 1 (1999): 3–28. 13. In some recent work the quality has been called “social capital,” reflecting a market mentality. Robert D. Putnam, “Bowling Alone: America's Declining Social Capital,” Journal of Democracy 6, no. 1 (January 1995): 65–78.

Page 217 →

THE CONTINUING STORY OF ANOTHER DEATH FORETOLD Radical Theory and the New International Relations Michael Cox

Introduction If international relations is to be judged by its capacity not just to understand the world as it is but by its proven ability to anticipate future global trends—a not unreasonable demand—then by any measure it has to be judged to have been one of the more dismal of the social sciences. The list of its failures is too long to bear repetition here and is possibly best forgotten. Others, however, might not let us forget quite so easily. As a senior editor of a leading foreign policy magazine once reminded us, consider what we, the so-called experts, either managed to get wrong or failed to get right.1 Who, he asked, called the following? The East Asian crisis? Or the collapse of the Soviet Union? The end of the cold war in Eastern Europe? Yugoslavia's demise or Japan's swift transformation from global economic powerhouse to financial wreck? The answer, he concluded, was no one.2 Nor, he might have added, did many people get very close to guessing the shape of the post-cold war order either. In fact, so complicated did the world look after 1989 that many writers took refuge in old intellectual certainties—with either dire predictions coming from one section of the international relations community about the future looking very much like the past or Kantian optimism emanating from another about a new international system in formation where the pacifying force of democracy, the integrating power of the market, and the stabilizing role of multilateral institutions would combine together to create a world of peace and relative prosperity.3 But such simplicities could not begin to capture the structural complexity of the new international system. Indeed, so complex did it seem that the term most frequently deployed by writers to describe it—other than the muchdisputed notion of globalization and the less-than-original idea of unipolarity—was Page 218 → the prefix post. Thus, according to whom you happened to be reading at the time, we were now entering into, or had in fact entered, a post-cold war era whose chief feature was that the cold war happened to be over, a postsocialist order whose main ideological characteristic was that there was no longer any socialism, or a postmodern world where apparently there was no order at all. For the first time in history it appeared that we were trying to make sense of the world around us not in terms of what it was, but rather in terms of what it happened to have superseded. Intellectually, at least, it was all rather depressing.4 In spite of the fact that few of the major academic trends can claim very much credit for either predicting the end of the cold war or what followed, one, it would seem, has suffered the greatest damage of all as a result of what happened in 1989: that loosely defined, deeply divided, and heterogeneous current in international relations known rather vaguely as “radical.” Since the “fall,” realism, it appears, has recovered its nerve.5 Liberalism has experienced something of a resurgence. And constructivism has done especially well.6 But radicalism, according to some accounts at least, has virtually disappeared off the intellectual map. This point was made somewhat clearer than the truth by the influential Stephen Walt in a recent survey of the discipline after the cold war. The death of realism, he felt, had “been largely exaggerated” liberalism meanwhile was experiencing a new lease of life following the “defeat of communism” while the end of the cold war had played “an important role” in actually “legitimating constructivist theories because realism and liberalism both failed to anticipate the event and had some trouble explaining it.” Radicalism, however, had a bleak future. Indeed, in its largely Marxist form, it appeared to have no future at all—and for good reason according to Walt. First, it had failed to provide a serious theory of intercapitalist relations in the modern world. It had then advanced a false theory of dependency that had proven quite unable to explain relations between the North and the South. And finally, it could not really account for what had happened under “actually existing socialism” in particular, it had no means of explaining the often aggressive and frequently fissiparous nature of communism as a system. For all these reasons (and no doubt others could think of more) radical analysis had lost any appeal it might have once had and held no particular allure for students of international relations in the post-cold war era.7

The argument that radical theory in its old historical materialist form has had its day is certainly a popular one. My purpose here, basically, is to suggest that just as with realism, it might still be a little premature to declare it dead altogether. In part this is based on a simple empirical observation: there is still lots of it out there, as any brief skim through the programs of either the British International Studies Association or the International Studies Association annual conferences would quickly reveal. If radical analysis has disappeared from the agenda as Walt, and no doubt many others seem to believe, then nobody appears to have told a sizable number of the profession who continue to do applied research on such subjects as Gramscian theories of hegemony,8 the geography of the world economy,9 U.S. foreign Page 219 → policy,10 capitalism and law,11 and global governance. If radical theory has suffered a setback since 1989, it is not one that has proved fatal.12 This brings me to make a second, related point, which in many ways runs directly counter to the common-sense view that the death of communism dealt a body blow to radical theory:13 whereas the cold war with its focus on the superpower competition and the strategic arms race was never that susceptible to radical materialist theorizing, unfettered capitalism (which is what we've now got) is. As the doyen of liberal American historians has recently pointed out, the irresistible dynamic of our modern form of “unbridled capitalism” makes very fertile ground indeed for radical analysis14—a view also endorsed by one of Mrs. Thatcher's onetime academic supporters.15 Nor should this come as much of a surprise. Marxism after all was a theory of global market relations, not of state conflict, and under modern conditions where the market is geographically dominant, the tools developed by Marx and others like Marx are just as likely to be useful (if not more so) in understanding the modern world as those employed by liberalism, realism, or constructivism. Stanley Hoffmann made much the same point in a useful intervention in the debate in 1995. We have, he suggested, to make a distinction between what existed in countries like the USSR and the analytical categories of historical materialism. The former may have collapsed, but that does not mean the latter is therefore invalidated. Marxism might have given birth to a utopian politics. This does not mean however that it lacks serious analytical power.16 In the end of course nothing is given, and the credibility of radical analysis is going to depend more on its scholarship than on its materialist categories or its skepticism toward capitalism. Nothing after all is inevitable, and if radicals fail as badly in their attempts to explain the shape of the post-Soviet world as they did in analyzing the contradictions of the former Soviet Union itself, then they have—and will deserve—no future. It might therefore be useful to see how well the different varieties of radicalism have thus far come to terms with a world they, and nearly everybody else for that matter, never anticipated. In what follows I shall therefore try to summarize a large, very uneven, and deeply schismatic literature. In my brief review I will include many writers who some might not even include within the fold, and others who themselves would not be comfortable being labeled as Marxist. Robert Cox for example is often associated with Marxism, though he himself could hardly be described as orthodox.17 The same might be said of André Gunder Frank, whose most recent work seems to challenge the whole edifice of Marxist thought on the history of capitalism.18 Noam Chomsky moreover is no Marxist. On the other hand, his writings on the new international order are highly critical of the status quo. Furthermore, unlike many of the analysts mentioned in this paper, his works have a fairly wide readership, especially outside the narrow confines of the international relations' profession.19 I have divided the discussion into four sections. In the first, I look at the way radicals have tried to come to terms with the death of actually existing socialism and the associated international result in the shape of the “end of Page 220 → the cold war.” In the second, I examine their efforts to decode the meaning of the term globalization. Part three then sees whether radical analysts have developed a theory of crisis. Finally, part four examines the way or ways radicals have tried to come to terms with American power. In the concluding section I explore what is far and away the greatest problem facing radical analysis today (possibly its greatest problem throughout the twentieth century): identifying the source of political regeneration in a world where there is possibly as much, if not more, suffering than at any point over the last fifty years, but little sense that much can be done to alleviate it.

The End of the Cold War The collapse of Soviet power and with it the end of the cold war was as big a surprise to most radicals as it was to more mainstream analysts, perhaps more so because many on the Western Left had a certain regard for “actually existing socialism”: quite a few out of a misplaced sense of political loyalty, some because they just didn't like

capitalism, and others because they felt that the Soviet Union (whatever its faults internally) played an internationally progressive role by counterbalancing the power of the United States while underwriting numerous anti-imperialist regimes—especially in the less developed countries of what was then, but is no more, referred to as the Third World.20 Lacking a proper political economy of communism (for reasons which would take too long to discuss here) radical analysts have in the main tried to deal with the politically problematic question of the fall of official socialism not by confronting the problem head on but rather by finessing the issue. They have done so in a number of different ways. The first way, quite simply, has been to deny that the regimes in Eastern Europe or the former Soviet Union were genuinely socialist. Thus what “fell” between 1989 and 1991 was not the real article, but some odd hybrid that had little or nothing to do with what American Marxist Bertell Ollmann once tried to describe as “Marx's vision” of the new society. This inclination to deny the socialist authenticity of the former USSR can in fact be found in the writings of many radicals, though it is perhaps most strongly articulated in the work of Hillel Ticktin, editor of the journal Critique and the only Western Marxist to have developed a detailed political economy of the USSR before its disintegration. According to Ticktin (who was one of the few radicals to have ever lived in the USSR for any extended period of time) the Soviet system was not just repressive but economically far less efficient than what existed in the West. For this reason he did not believe that what he termed this “economy of waste” could endure over the long term. This view—first articulated as early as 1973—in turn became the basis of a very specific politics, which meant that Ticktin at least was somewhat less surprised than other Marxists by the Soviet Union's subsequent collapse. Indeed, in his opinion, until the USSR passed from the stage of history, there was little chance of a genuine “new” Left ever emerging in the West.21 Page 221 → If Ticktin detected deep and life-threatening economic flaws in the Soviet system, this was not the position of most socialists. Indeed, one of the more obvious ways in which other radicals have tried to come to terms with the fall of the Soviet Union has been to imply that the system did not have to go under at all: and the only reason it did was because of ill-fated attempts to reform the country in the 1980s. Thus one of the better-informed socialist economists has argued that although the Soviet Union had its fair share of problems, it did not face a terminal crisis. What brought it down, in the end, was not its flaws but Gorbachev's contradictory policies.22 The American radical Anders Stephanson also insists that there was nothing inevitable about the demise of communism in the USSR. In his assessment it was largely the result of what he terms “contingency.”23 Halliday too has suggested that the Soviet system did not (in his words) “collapse,” “fail,” or “break down.”24 Rather, the Soviet leadership after 1985 decided—albeit for good objective reasons—to rule in a different way: and it did so not because of massive internal difficulties but because the Soviet elite finally realized that the USSR could neither catch up with nor compete with the West. Once this became manifest, the ruling group effectively lost its historical nerve.25 The argument that the old Soviet system might not have been suffering incurable economic cancer has also led certain radical writers to the not illogical conclusion that if the system was not doomed because of its internal problems, it was in the end external factors that caused it to implode. This is certainly implicit in the influential work of Halliday, who, significantly, says little about the USSR's domestic weaknesses but a good deal about the impact Western economic performance had upon Soviet elite perceptions. Others have stressed a more direct connection. Thus in the view of the Dutch Marxist Kees van der Pijl, though the final transformation of the Soviet system was the result of several internal factors, one should not underestimate the role played by the United States and its declared objective of quite literally spending “the Soviet Union into bankruptcy.”26 Robert Cox appears to have come to much the same conclusion in his writings. Like van der Pijl, Cox accepts there is no simple explanation of what happened in the former USSR after 1985. Nonetheless he still concludes that “the arms race provoked by the Reaganite phase of the Cold War was too much for an unreformed Soviet economy to sustain.”27 Michael Ellman has also laid great stress on the importance of U.S. strategy: taken together Reagan's rearmament program, Star Wars, and U.S. support for anticommunist guerillas throughout the world were, in his opinion, “key external factors” in bringing about Soviet economic collapse in 1991.28

But if the Left has faced very real problems in coming to terms with the fall of the Soviet Union, it has tried to compensate for this in two very different ways. One has been to understate the impact the communist collapse has actually had upon the “essential” nature of the international system. Hence, in the view of Noam Chomsky, there is nothing “new” at all about the new world order: the rich remain rich, the poor South remains the poor South, and the United States still remains in charge.29 Robert Cox concurs. Page 222 → The new international system, he thinks, looks very much like the old one. Indeed, in the most basic of ways, the cold war, he argues, has “not ended” at all, and its more basic structures continue “to live in the West” in the shape of high military spending, in the operation of intelligence services, and in the unequal distribution of power among the various states.30 This also appears to be the position of the doyen of world systems theory, Immanuel Wallerstein. Unfortunately, in his metastructural (and highly abstract) output over many years, one had no real sense that the cold war ever had much meaning at all. Thus its conclusion was unlikely to have a great deal of impact upon a world system that had existed since the sixteenth century, that had been in some fairly unspecified “crisis” since the late 1960s, and that would presumably remain in crisis until it came to an equally unspecified end twenty or thirty years down the historical line.31 Finally, if writers such as Wallerstein have tended to minimize the impact of the end of the cold war, others have argued (perhaps rather more convincingly) that its passing has in fact created new political spaces that did not exist before. This is more or less the position adopted by Bogdan Denitch. Starting from the not unreasonable assumption that the division of Europe rested upon an illegitimate form of Soviet domination over the East, and a legitimate form of American hegemony in the West, Denitch concludes that in the new united continent there are now great opportunities. Unlike many politically active Marxists, Denitch is no utopian. Thus in his view the new openings are unlikely to free the workers from the grip of capitalism. Yet 1989 does make possible the deeper integration of Western Europe, and upon this basis Europe—in his opinion—will be able to develop a new social democratic third way between a highly dynamic but politically unacceptable American-style liberal capitalism and a moribund Soviet-style communism. This hardly amounts to the same thing as world revolution. Nonetheless, in a postcommunist world, the possibility of building a new progressive Europe is one that should animate intelligent radicals more than pointless calls to man barricades that nobody wants to build and few want to stand behind.32

New World Economic Order: Globalization Though analysts like Denitch have tried to find some crumbs of comfort from the events of 1989, overall the collapse of planning in Eastern Europe, followed as it was by the adoption of radical market strategies in countries such as Poland and the Czech Republic, had an enormously debilitating political impact on Western radicals. Yet, as I have implied earlier, the pessimism did not last long, and those who managed to survive the baptism by fire began to take intellectual heart somewhere around the mid-1990s. There were two reasons for this. The first, quite simply, was that the much-heralded transition in the former communist countries turned out to be far more problematic than most market triumphalists had originally anticipated. In the case of Russia of course the socalled transition to something better and higher soon turned into a minor tragedy for the Russian people.33 Page 223 → The other, equally important, reason was the birth of a new world order in which economics assumed center stage. In fact, it almost looks now as if it required a healthy dose of capitalism to reignite radical analysis and provide the intellectual Left with a clear focus. But perhaps we should not be so surprised by this. In an age of geoeconomics where even staid bankers were now prepared to use words like “capitalism,” where the former editor of the London Times talked menacingly of the “coming depression,”34 and an American president paid tribute to a book that speculated in almost Leninist terms about the coming economic struggle for dominance between the great powers in the twenty-first century35—it was almost inevitable that radicals would begin to take heart! Having been ground under politically since 1989, it looked to some of them at least (and at last) that historical materialism had finally come of age. Rather than trying to provide a detailed reconstruction here of a single radical or “neo-Marxist” analysis of latetwentieth-century capitalism (one that has yet to be written), it might be more useful to briefly point to some of the issues now being debated in the growing left-wing literature. Not surprisingly, one issue that has been discussed more than most is globalization. While there is no agreed-upon radical view on the subject, four quite reasonable questions have been asked of the concept since it literally exploded onto the academic agenda in the early 1990s.36

The first has perhaps been the most challenging: namely, what exactly is so new about the idea? As many on the Left have argued, the apparently novel thesis that national economies have become mere regions of the global economy and that productive forces have expanded far beyond the boundaries of the nation-state is not novel at all.37 Indeed, one of the first writers to advance the argument was no less than the abused and much-ignored Karl Marx, who in the Communist Manifesto made it abundantly clear that the central feature of the capitalist epoch was the “universal interdependence of nations.” Moreover, this simple but critical idea ran like a red thread through Marxist thinking thereafter. It was, for instance, repeated by Lenin in his 1916 pamphlet Imperialism in an effort to provide a materialist explanation of World War I. Trotsky also deployed very much the same argument in his critique of Stalin's claim that it was possible to build socialism in one country. And later theorists of dependency took it as read that until the less developed countries could break away from the spidery economic web of the world market, they had no chance of overcoming the limits of backwardness.38 Globalization might have become a fashionable concept in the 1990s amongst those desperately looking for a “relevant” topic now that traditional security questions no longer seemed to be interesting. But like most intellectual fads and fashions it was only a recycled version of a very old idea. In fact, according to Burnham, even the idea itself was not a very good one.39 The theme of continuity is also developed in the work of Hirst and Thompson. However, rather than attacking mainstream academics for failing to recognize the radical antecedents of the concept, they question whether globalization is even an accurate description of the world economy Page 224 → in the late twentieth century. In a much-cited study whose underlying purpose is as much political as it is economic, they conclude that the image of globalization has for too long mesmerized analysts.40 In their opinion the theory can be criticized on at least two grounds. The first is in terms of its descriptive power. In their view the “present highly internationalized economy is not unprecedented” at all; indeed, “in some respects the current international economy is less open and integrated than the regime which prevailed between 1870 and 1914.” Moreover, “genuinely transnational companies appear to be relatively rare,” while the world economy itself, far from being genuinely “global,” is still very much “dominated by the Triad of Europe, Japan and North America.” They also question its political implications and suggest that far from being powerless as the theory implies, the state can still make a difference. As they argue, in this less than completely globalized economy, there are still opportunities for the development of governance mechanisms at the level of the international economy that neither undermine national governments nor hinder the creation of national strategies for international control. In other words the world of “markets” remains susceptible to conscious intervention. To this extent, the world economy is not out of control: politics, politicians, and the people—in other words the state under conditions of democracy—can make a difference.41 A third line of radical attack has not been to question the reality of globalization so much as to point to its appalling human consequences. In a world where the market is “unbound”—they argue—where there is in effect no alternative to the market, capitalism has assumed an increasingly aggressive posture: and this has led to the most extreme forms of inequality and economic polarization.42 Nor is this accidental; nor can it be overcome without challenging the foundations of the system itself. Furthermore, if global economic integration has generated what one critic has rather mildly termed “underconsumptionist tendencies,” it has tended to do so not only within capitalist countries but also across them.43 This is why the gap between the have and have-not nations has tended to increase rather than decrease in an era of global capitalism. According to one study, in an unregulated capitalist world economy, the outcome has been that those countries and regions that already possessed abundant resources and power have remained powerful and prosperous, while those that did not have become poorer and even more dependent than before.44 Finally, many radical critics have wondered whether some of the advocates of globalization have tended to underestimate the anarchic and competitive character of the world capitalist system. Though accepting the more general thesis about global interdependence, many on the Left do not accept the liberal corollary that we have moved beyond the age of conflict. And though war in the more traditional sense is highly unlikely, this does not rule out intense competition at the other levels. Realist by inclination, radicals have noted several areas in the world today where antagonism rather than cooperation is the norm. America's intense rivalry with France over trade, Germany's attempt to exercise economic hegemony over Europe at the expense of the United Kingdom, and the United States' more recent drive Page 225 → to open up the markets of Asia-Pacific all point to a slightly less

benign view of economic reality than that suggested by the “globalists.”45

Capitalist Contradictions Perhaps the most serious difference, however, between radicals and their more orthodox peers concerns the longterm stability of world capitalism. Though few but the most orthodox economists would subscribe to a simple theory of global economic equilibrium, there is an underlying assumption amongst most non-Marxists that even though the international economic system might go through periodic booms and busts, these movements are either functional to the system overall (Schumpeter, you will recall, once talked of “creative destruction”) or can easily be resolved.46 Naturally enough radicals do not share this sense of optimism. Nor in one sense can they, given their opposition to the status quo. The problem for the Left of course is that they have too often cried wolf before to be taken seriously now. However, with the onset of the Asian economic crisis it at last seems as if their prediction of economic doom has finally turned out to be true; and inevitably they have drawn some comfort from the fact that the so-called Asia-Pacific miracle has turned into a nightmare—one that has even prompted The Economist to ask whether the world as a whole is on the cusp of a new slump.47 But long before the collapse of capitalist optimism in Asia, radicals had already begun to articulate a theory of crisis. Basically, this consisted of a number of distinct arguments. The first part was in essence an updated version of Hobson mediated via Keynes and restated in different forms by radical economists like Sweezy and Magdoff: and what this amounted to in effect was a belief that there was a fundamental contradiction between the world economy's capacity to produce and the people's ability to consume. In other words, there existed what Marx had frequently referred to in his work as a systemic tendency to overproduction—one that he thought could not be overcome as long as capitalism continued to exist. Though long consigned to the proverbial dustbin of history, this particular theory has enjoyed something of a revival over the past few years, and not just amongst radicals but also more mainstream economists concerned that the great boom of the 1990s could well easily be followed by the great crash of the early twenty-first century. Indeed, the thesis itself has been given an enormous boost with the publication of the evocatively subtitled study by William Greider, The Manic Logic of Global Capitalism. Though Greider's underlying argument is not designed to support revolutionary conclusions, that his study evokes the ghost of Marx (and has become an instant best-seller) would suggest that what one radical reviewer has called this “powerful and disturbing book” has touched a very raw nerve amongst more orthodox analysts.48 Greider's pessimism, however, is not just based upon a general argument about the overproductive character of modern capitalism. It also flows from a more detailed analysis and awareness of the increasingly integrated and Page 226 → highly open character of the international economy, which allows billions to be moved, or lost, in a matter of hours; and where events in one country or set of countries are very rapidly felt around the world in a domino process that once set in train becomes very difficult to stop. A good example of this is provided by what happened in Hong Kong in late 1997. Here a 25 percent fall on the Hong Kong Stock Exchange quickly led to a major decline in share prices around the world. In the same way, the meltdown in Indonesia has had a profoundly depressing effect upon the rest of Asia, while the financial crisis in Japan has sent seismic shocks right across the Pacific to the United States. And so it will go on according to radicals until governments either decide to reflate the world economy—which they are scared to do politically—or there is the crash predicted by Greider. The argument that the system might well be spinning out of control is further supported in radical analysis by the argument that we are now living in an era when finance rather than industry—and finance capital rather than productive capital—has assumed the dominant role in the world economy.49 While there may not be anything especially “radical” about that particular empirical observation, there is about some of the conclusions radical theorists tend to draw from it. First, in their view, the preponderance of finance effectively means that capitalism today has little or no interest in supporting industrial policies that sustain full employment. This therefore means that the system overall is now less able to fulfil at least one basic human right: the right to work. More generally, the overwhelming power of finance capital introduces enormous instability into the system as those with money either seek speculative gain with little concern about the political consequences of their actions or move their

money at very high speed if and when conditions change. What makes the situation all the more volatile of course is that there are no national or international means for controlling these various movements and flows. Consequently, a very dangerous and apparently unbridgeable gap has opened up between those institutions that are supposed to manage the world economic system in the general interest and the specific interests of the banks, the large pension funds, and the insurance companies. Finally, the tendency to crisis in the post-cold war epoch has been reinforced it has been argued by the end of the cold war itself. Though not all radicals adhere to the argument that the cold war was good for capitalism (there was a powerful current of thought that suggested the opposite), there are those who maintain that even though the superpower conflict was costly these costs were more than offset by the benefits. By the same token, while there has been obvious economic benefits accruing from the termination of the cold war, its ending has created major problems for the West as a whole. First, governments in key capitalist countries like the United States and the United Kingdom no longer have military spending as a way of pump priming their economies.50 Second, there are the unforeseen but really quite huge costs involved for Germany caused by reunification—costs that have transformed Germany from being a boom economy into one of the great underperforming economies of Europe. Third, though the end of the cold war has Page 227 → been followed by what most radicals see as a temporary boom in the United States, its passing has fundamentally weakened America's capacity to act abroad. And without American leadership, the international order in general and the world capitalist system in particular are bound to suffer. Working on the good realist assumption that American power was an essential element in the postwar reconstruction of international capitalism, a number of radicals believe that now that the cold war is over and America can no longer exercise its hegemony so effectively, the world system is likely to become a good deal less stable.51 Difficult times lie ahead for the last remaining “superpower without a mission.”52

Hegemonic Still? The United States This brings us quite logically to the question of the United States: the source of most radical distaste during the cold war and the cause of much intellectual anguish since now that it has seen off the only power in the world capable of limiting its reach. Though less vilified in the 1990s than it was previously, the United States nonetheless continues to fascinate radicals in ways that no other nation does. The reasons for this are clear. No other country is as powerful, dynamic, or as “exceptional” as the United States of America. Indeed, according to one “European” analyst, even American radicals and Marxists are more interesting and “have been more intellectually productive and innovative” than their comrades across the Atlantic since the late 1960s!53 Three issues have been of greatest interest to radical critics: one concerns the use of American power; another, the nature of American foreign policy in the post-cold war era; and the third, America's position within the larger international system. Let us deal briefly with each. The question about American power has been an especially problematic one for the Left in the post-cold war period. Naturally enough, most (but by no means all) radicals opposed American intervention against Iraq in 1991. However, since then, many have found themselves in the somewhat paradoxical position of attacking the United States not for being too interventionist but for not being interventionist enough. The issue that led to this rather odd state of affairs was of course the war in ex-Yugoslavia. Here the Left found itself caught between a rock and a hard place. On the one hand, being radicals they were deeply suspicious of any American involvement on the continent of Europe. On the other hand, it was palpably clear that if the United States did not get involved, the genocide in Bosnia would continue. There was no easy squaring of this particular circle. Some therefore decided to stick to their ideological guns and opposed any American role.54 Others, however, bit the political bullet and urged Washington on. Indeed, one of the greatest ironies of this particular tragedy was that many radicals who had earlier criticized the United States for having intervened in the Gulf because of oil now pilloried the United States for not intervening in the Balkans because there was no oil. Moreover, having been the strongest critics of the American military before Bosnia, some on the Left at least now Page 228 → became the most bellicose advocates of tough military action against the Serbs. If radicals seemed to have had serious problems in dealing with U.S. power in the post-cold war world, they

appear to have had none at all in attacking America's self-proclaimed goal of promoting democracy. To be fair, they did not oppose the United States because they were against democracy as such, but rather because they thought that its championing of the policy was either a sham or, more obviously, a device designed to obscure America's economic objectives in the larger capitalist system. Chomsky in fact has even argued that the United States has actually deterred democracy, 55 while Furedi in his broadside has attacked all Western talk of making the world a better place as little more than a cover for neocolonialism.56 Others have been slightly less harsh, or at least more subtle. This is certainly true of the important work undertaken by Gills and Robinson in their analysis of the Third World. In an attempt to move the argument forward, both have proposed the thesis that certain forms of “low intensity democracy” have had an important role to play in both containing popular protest while legitimizing painful economic reforms being advocated by Washington.57 Robinson indeed has put forward a whole historical argument concerning the complex interplay between social change and elite rule in Latin America. Deploying the much-used and admired Gramsci, he argues that by the 1980s it had become clear to the dominant group that the old repressive methods were no longer workable in an age of globalization; and supported by the United States, they therefore replaced coercive means of social control with consensual ones. Though the policy carried certain risks, in the end it achieved precisely what it had been designed to: namely, to secure political stability in a period of social upheaval caused by Latin America's more complete integration into the world capitalist system. The third and final “great debate” has turned around the hoary old problem as to whether or not the United States is, ever has been, or will be in decline. Regarded by many in the international relations profession today as a nonissue (though this was not the view between the Vietnam War and the end of the cold war), it continues to inform a large part of the modern radical discussion about the United States. Some like Bernstein and Adler have little doubt that the United States is in decline—and has been so since the late 1960s.58 Others like Stephen Gill assert that the “declinist” thesis is quite false.59 Clearly there is no consensus on the issue, and it would be misleading to suggest that there was one. Yet whereas radicals before the collapse of the USSR were more inclined to believe that the United States was on the way down, since 1991 they have tended (along with nearly everybody else) to assume that there is still a good head of steam left in the engine of the American capitalist machine. U.S. success in the cold war, its easy victory over Iraq, the financial crisis in Japan, Europe's inability to resolve the situation in ex-Yugoslavia, and the economic boom in the United States after 1992 have in fact convinced many that American power is still something to be reckoned with. Indeed, according to one study written by a Latin American with impeccable left-wing credentials, Marxists have to face up to the Page 229 → unpalatable fact that the United States is not only not in decline but can actually look forward to the future with enormous self-confidence. The prophets of (relative) doom like Paul Kennedy may know their history according to Valladao. However, they are in his view a century or two adrift; and if historical analogies must be drawn then it should be with Rome in triumph after its victory over Carthage, not with Britain in the postwar period, or Spain in the sixteenth century. The twenty-first century will be American.60

Changing the World The issue of American power leads finally to the question of political renewal on the Left. The two are obviously connected. After all, if America is in decline as some on the Left believe, the political possibilities for radicals would seem to be bright. If, on the other hand, we can look forward to continued American hegemony, then capitalism by implication must be secure: and if capitalism is secure, then the possibility of a radical breakthrough is highly unlikely. This in turn raises the even larger problem of what radicals or Marxists are supposed to do in a world where on the one hand the “socialist alternative” seems to have failed and where on the other the market in spite of its manifest contradictions looks like “it's the only game in town.” The intellectual Left has responded to this dilemma in a number of ways. Two deserve special mention here. The first has been to accept that for the time being there may in fact be no alternative to the market, and the only thing one can do therefore is develop strategies that seek to build areas of opposition and resistance within the larger interstices of “civil society”—either at the national or global level.61 With this in mind, no doubt, many on

the Left have endorsed campaigns to extend the realm of democracy or increase the degree of information available to the public at large. In Britain a good deal of radical energy has also been expended on supporting constitutional change, while in the United States radical activists have been involved in an as yet unsuccessful effort to develop a comprehensive health system. The Left has also engaged in numerous other campaigns covering a range of questions, from women's rights and trade union recognition right through to the increasingly popular issue of the environment—on which there is now a vast academic literature. Indeed, one of the more interesting developments over the past few years has been the marked rise in a radical discourse on major environmental questions. One might even be tempted to suggest that the struggle to save “mother earth” from what some now see as impending environmental catastrophe has taken over from the equally influential movement ten years previously to prevent the collapse of the world into nuclear war. The second way in which the Left has responded to the current situation has been to explore ways and means by which the dynamics of globalization can either be slowed down or even arrested entirely—a perspective explored with typical intellectual sensitivity by Robert Cox. According to Cox, there Page 230 → is no reason for despair insofar as the dynamics of globalization are bound to throw up various forms of resistance around the world. This will come from many layers affected by the internationalization of production, including groups outside of the production process proper. Resistance to global capitalism, however, is also bound to involve workers themselves who have been placed under unremitting pressure by the logic of a global capitalism constantly seeking to weaken the position of organized labor.62 According to radical critics, moreover, the proletarian genie is not just a figment of some rabid left-wing imagination. In country after country—from South Korea to France, from Germany to the United States itself—workers have begun to take action to resist attempts to make them mere robots in a world without frontiers, where capital owes no loyalty—except of course to its shareholders.63 They may not yet have united, but at last the workers are beginning to act. However, as Cox would be the first to admit, resistance to globalization is not exactly the same thing as a positive or coherent strategy, and until there is such a strategy, there will always remain what he calls “a vacuum to be filled—a challenge to critical thinking on the left.” And at the heart of this challenge “is the question of the motive force for change.”64 For if the working class is not the universal class described by Marx, and the vision of a new society has been besmirched by the experience of the USSR, then there is little possibility of major political transformation. Furthermore, even if the world is in crisis, or in what Hobsbawm prefers to call a “state of social breakdown,” without a vision of a different society, nothing can fundamentally change.65 This presents radicals with a major problem. For however sound their analysis, if the world remains the same (or even gets worse) they will simply be left standing where they have been for a very long time: on the sidelines of history. In this sense their greatest challenge perhaps is not so much intellectual as political; and until they can provide a coherent answer to the question of what it is they are for rather than what it is they are against, they will remain what they have been, in effect, for more years than they would care to admit—well-informed rebels without a political cause.

Notes The title is of course a derivative and draws very directly from that used by Stafano Guzzini in his excellent discussion of another premature fatality. See Stafano Guzzini, Realism in International Relations and International Political Economy: The Continuing Story of a Death Foretold (London: Routledge, 1998). 1. Moises Naim, “Editor's Note,” Foreign Policy 110 (spring 1998): 9–11. 2. The admittedly controversial issue of prediction, and predicting the end of the cold war in particular, is dealt with at length in Michael Cox, ed., Rethinking the Soviet Collapse: Sovietology, the Death of Communism, and the New Russia (London: Pinter, 1998). 3. Some of these issues are dealt with in Michael Cox, G. John Ikenberry, and Takashi Inoguchi, eds., American Democracy Promotion: Strategies, Impacts (Oxford: Oxford University Press, 2000).

4. For one attempt to explore the various features of the international system after the “fall” see Michael Cox, Ken Booth, and Tim Dunne, eds., The Interregnum: Controversies in World Politics, 1989–1999 (Cambridge: Cambridge University Press, 1999). 5. On the state of realism see the useful collection of essays in Ethan B. Kapstein and Michael Mastanduno, eds., Unipolar Politics: Realism and State Strategies after the Cold War (New York: Columbia University Press, 1999).Page 231 → 6. Though who is and who is not a “constructivist” these days remains unclear. See the lengthy discussion occasioned by the publication of Alexander Wendt's book Social Theory of International Politics (Cambridge: Cambridge University Press, 1999) in the Review of International Studies 26 (January 2000): 123–80. 7. Stephen Walt, “International Relations: One World, Many Theories,” Foreign Policy 100 (spring 1998): 29–46. 8. See Mark Rupert, Producing Hegemony: The Politics of Mass Production and American Global Power (Cambridge: Cambridge University Press, 1995). 9. For a tour through the new radical geography, see Paul Knox and John Agnew, The Geography of the World Economy, 3d ed. (London: Arnold, 1998). 10. See Nicholas Guyatt, Another American Century? The United States and the World after 2000 (London: Zed Books, 2000). 11. Claire Cutler, “Critical Reflections on the Westphalian Assumption of International Law: A Crisis of Legitimacy,” Review of International Studies 27 (April 2001): 133–50. 12. For a very useful and comprehensive guide to radical theory in an age of globalization see Steve Hobden and Richard Wyn Jones, “Marxist Theories of International Relations,” in The Globalization of World Politics, 2d ed., ed. John Baylis and Steve Smith (Oxford: Oxford University Press, 2001). 13. For a more nuanced position, however, see Peter J. Katzenstein, Robert O. Keohane, and Stephen D. Krasner, “International Organization and the Study of World Politics,” in their Exploration and Contestation in the Study of World Politics (Cambridge: MIT Press, 1999), esp. 24–26. 14. See Arthur Schlesinger, “Has Democracy a Future?” Foreign Affairs 76 (September-October 1997): 12–22. 15. See John Gray, False Dawn: The Delusion of Global Capitalism (London: Granta Books, 1998). 16. See the interview with Stanley Hoffmann, “Democracy and Society,” World Policy Journal 12 (spring 1995): 35. 17. See Robert Cox, “Social Forces, States, and World Orders: Beyond International Relations Theory,” Millennium 10, no. 2 (1981): 126–55; and “Gramsci, Hegemony, and International Relations: An Essay in Method,” Millennium 12, no. 2 (1983): 162–75. 18. See André Gunder Frank and Barry Gills, eds., The World System: Five Hundred Years or Five Thousand? (London: Routledge, 1993). 19. See, for example, Noam Chomsky, World Orders: Old and New (London: Pluto, 1994). 20. Fred Halliday, The Making of the Second Cold War (London: Verso, 1993). 21. See also Hillel Ticktin, The Origins of the Crisis in the USSR: Essays on the Political Economy of a Disintegrating System (New York: Sharpe, 1992). 22. See Michael Ellman and Vladimir Kontorovich, eds., The Disintegration of the Soviet Economic System (London: Routledge, 1992). 23. See Anders Stephanson, “Rethinking Cold War History,” Review of International Studies 24 (January 1998): 119–24. 24. See Fred Halliday, Rethinking International Relations (London: Macmillan, 1994), esp. 191–215. 25. See Fred Halliday's essay “The End of the Cold War and International Relations,” in International Relations Theory Today, ed. Ken Booth and Steve Smith (University Park: Pennsylvania State University Press, 1995), pp. 38–61. 26. See Kees van der Pijl, “Soviet Socialism and Passive Revolution,” in Gramsci, Historical Materialism, and International Relations, ed. Stephen Gill (Cambridge: Cambridge University Press, 1993), 237–58. 27. See Robert Cox with Timothy J. Sinclair, Approaches to World Order (Cambridge: Cambridge University Press, 1996), 217. 28. Michael Ellman, “Multiple Causes of the Collapse,” RFE/RL Research Report 2, no. 23 (1993): 56.

29. See Noam Chomsky, Deterring Democracy (New York: Vintage, 1992). 30. Cox with Sinclair, Approaches to World Order, 34.Page 232 → 31. See Immanuel Wallerstein, “The Inter-State Structure of the Modern World-System,” in International Theory: Positivism and Beyond, ed. Steve Smith, Ken Booth, and Marysia Zalewski (Cambridge: Cambridge University Press, 1996). See also Immanuel Wallerstein, ed., The Capitalist World-Economy (Cambridge: Cambridge University Press, 1979). 32. Bogdan Denitch, The End of the Cold War: European Unity, Socialism, and the Shift in Global Power (London: Verso, 1990), 3–14. 33. See the essays by Stephen Cohen, Stephen White, and Bob Arnot in Cox, Rethinking the Soviet Collapse. 34. See James Dale Davidson and Lord William Rees-Mogg, The Great Reckoning: Protect Yourself in the Coming Depression (New York: Touchstone Books, 1993). 35. The book in question was Lester Thurow, Head to Head: The Coming Economic Battle among Japan, Europe, and America (New York: Morrow, 1992). 36. According to Anthony Giddens, “globalization is almost worth not naming now: it is less a phenomenon, it is simply the way we live…. You can forget the word globalization: it is what we are.” Cited in John Lloyd, “Interview: Anthony Giddens,” New Statesman, January 10, 1997. 37. For a materialist though not necessarily orthodox Marxist account of the long history of capitalist economic interdependence, see Fernand Braudel, Capitalism and Material Life, 1400–1800 (New York: Harper and Row, 1973); Civilization and Capitalism, 15th–18th Century, vol. 2, The Wheels of Commerce (New York: Harper and Row, 1982); and Civilization and Capitalism, 15th–18th Century, vol. 3, The Perspective of the World (London: Collins/Fontana, 1984). On Braudel, see Randall Germain, “The Worlds of Finance: A Braudelian Perspective on IPE,” European Journal of International Relations 2, no. 2 (1996): 201–30. 38. The two radical classics on dependency and development are Paul Baran, The Political Economy of Growth (New York: Monthly Review Press, 1957), who later influenced André Gunder Frank, Capitalism and Underdevelopment in Latin America (New York: Monthly Review Press, 1969). 39. See Peter Burnham, “Globalisation: States, Markets, and Class Relations,” Historical Materialism 1, no. 1 (1997): 1–16. 40. See Paul Hirst and Graham Thompson, Globalization in Question (Cambridge: Polity, 1996). In her radical critique of the view that states are now powerless to make policy because of globalization, Linda Weiss even talks about “the enhanced importance of state power in the new international environment.” See her “Globalization and the Myth of the Powerless State,” New Left Review 225 (September-October 1997): 3–27. The Canadian Marxist Leo Panitch is another radical who disputes the notion that globalization has rendered national politics meaningless. See his “Globalization and the State,” in Socialist Register, 1994, ed. Ralph Miliband and Leo Panitch (London: Merlin, 1994), 60–93. 41. According to the blurb on the back of one noted study on the new “borderless world,” “nation states are dinosaurs waiting to die…. [they] have lost their ability to control exchange rates and protect their currencies…. they no longer generate real economic activity…. the fate of nation states is increasingly determined by choices made elsewhere.” See Kenichi Ohmae, The End of the Nation State: The Rise of Regional Economics (New York: Harper Collins, 1996). 42. See Elmar Altvater and Birgit Mahnkopf, “The World Market Unbound,” Review of International Political Economy 4 (autumn 1997): 448–71. For useful discussions of the relationship between globalization and inequality, see Andrew Hurrell and Ngaire Woods, “Globalisation and Inequality,” Millennium 24, no. 3 (1995): 447–70; and Julian Saurin, “Globalisation, Poverty, and the Promises of Modernity,” Millennium 25, no. 3 (1996): 657–80. 43. Ronen Palan, Underconsumptionism and Widening Income Inequalities: The Dynamics of Globalization, Newcastle Discussion Papers in Politics, no. 4 (Newcastle, England: University of Newcastle, 1993), 45.Page 233 → 44. Andrew Sayer and Richard Walker, The New Social Economy: Reworking the Division of Labour (Oxford: Oxford University Press, 1992). 45. See James Petras and Morris Morley, “U.S.-French Relations in the New World Order” (July 1997, unpublished), 39.

46. In an attack on what he identified as a “Marxist vision” of a world economic crash, the American economist Lester Thurow argued that elected governments could and would in the end act to prevent the destruction of “both the economic system and democracy itself.” See his “The Revolution Is upon Us,” Atlantic Monthly 279 (March 1997): 97–100. 47. See “Will the World Slump?” The Economist, November 15, 1997, 17–18. 48. William Greider, One World, Ready or Not: The Manic Logic of Global Capitalism (New York: Penguin, 1997). 49. One well-established Marxist has argued that the era of “financial expansion represents the ‘autumn’ of a prevailing capitalist order as it slowly gives way to a new order.” See Giovanni Arrighi, The Long Twentieth Century: Money, Power, and the Origins of Our Time (London: Verso, 1994). See also the article by the German Marxist Elmar Altvater, “Financial Crises on the Threshold of the 21st Century,” in Socialist Register, 1997, ed. Leo Panitch and Colin Leys (London: Merlin, 1997), 48–74. 50. See Ann Markusen and Joel Yudken, Dismantling the Cold War Economy (New York: Basic Books, 1992). 51. Thomas J. McCormick, America's Half-Century: United States Foreign Policy in the Cold War and After (Baltimore: Johns Hopkins University Press, 1995). 52. See Michael Cox, U.S. Foreign Policy after the Cold War: Superpower without a Mission? (London: Royal Institute of International Affairs, 1995). 53. Goran Therborn, “Dialectics of Modernity: On Critical Theory and the Legacy of Twentieth-Century Marxism,” New Left Review 215 (January-February 1996): 59–81. 54. For the radical case against U.S. intervention in Bosnia see James Petras and Steve Vieux, “Bosnia and the Revival of U.S. Hegemony,” New Left Review 218 (July-August 1996): 3–25. 55. Chomsky's discussion of democracy and America's role in supporting it in the postwar period is somewhat more nuanced than the actual title of his book—Deterring Democracy—would suggest. See, in particular, “The Decline of the Democratic Ideal,” ibid., 331–50. 56. Frank Furedi, The New Ideology of Imperialism (London: Pluto, 1994). 57. See Barry Gills, Joel Rocamora, and Richard Wilson, eds., Low Intensity Democracy: Political Power in the New World Order (London: Pluto, 1993); and William I. Robinson, Promoting Polyarchy: Globalization, U.S. Intervention, and Hegemony (Cambridge: Cambridge University Press, 1996). 58. Michael A. Bernstein and David E. Adler, eds., Understanding American Economic Decline (New York: Cambridge University Press, 1995). 59. See Stephen Gill, “American Hegemony: Its Limits and Prospects in the Reagan Era,” Millennium 15, no. 3 (1986): 311–39; and American Hegemony and the Trilateral Commission (Cambridge: Cambridge University Press, 1990). For a critique of hegemonic stability theory from a Marxist perspective see Peter Burnham, The Political Economy of Postwar Reconstruction (London: Macmillan, 1990). 60. Alfredo Valladao, The Twenty-first Century Will Be American (London: Verso, 1996). 61. See David Held, Democracy and the Global Order: From the Modern State to Cosmopolitan Governance (Cambridge: Cambridge University Press, 1997). 62. Robert W. Cox, “The Global Political Economy and Social Choice,” in Cox with Sinclair, Approaches to World Order, 191–208. 63. See “Global Economy, Local Mayhem,” The Economist, January 18, 1997, 15–16. 64. Cox, “The Global Political Economy and Social Choice,” 192. 65. Eric Hobsbawm, The Age of Extremes: A History of the World, 1914–1991 (London: Pantheon, 1994), 459.

Page 234 →

HOW WE LEARNED TO ESCAPE PHYSICS ENVY AND TO LOVE PLURALISM AND COMPLEXITY Ernst B. Haas and Peter M. Haas

Introduction Is cumulative knowledge and progress possible in international relations (IR)? While this essay focuses on international institutions, it may be taken as a mere illustration of much wider issues in the field of IR and social science more generally. Students of IR remain divided on the implications of international institutions for the understanding of contemporary international relations. This is largely due, we believe, to the incommensurate epistemological and ontological positions within the discipline of IR that characterize most studies and interpretations of international institutions. In this essay we try to frame a pragmatic-constructivist approach for the study of international institutions, and of IR more generally. “Pragmatic” derives from the philosophy of science tradition of “pragmatism.”1 Because our objective is to sidestep the philosophical chasms now dividing the discipline we also seek to avoid some of the dichotomies that now bedevil us. One such opposes ideographic to nomothetic studies—unnecessarily. Another claims that there is a deep fissure between explanation and understanding—an overstated caricature of systematic research. Still another opposes positivism to reflectivism. Our method of sidestepping these procrustean constraints on inquiry features a consensus theory of truth: we argue that it is possible for followers of any and all of these approaches to knowledge to agree with one another in solving a problem if and when they can also agree that they accept a given solution to be “true,” if only temporarily and for a restricted purpose. We also argue that the means for ascertaining such a “truth”—truth tests—can also become consensual by means of sustained dialogue among theorists and practitioners. Ontologically Page 235 → and epistemologically the truth ascertained by these operations is neither as absolute as positivists and scientific realists demand nor as biased, subject to someone's domination, or hegemonic as relativists proclaim.2

Ontology “Institutions” remains an essentially contested concept. This is irksome since “any discussion of causation must presuppose an ontological framework of entities among which causal relations are to hold, and also an accompanying logical and semantical framework in which these entities can be talked about.”3 Scholars disagree about the meaning and consequences of international institutions. Most neoliberal institutionalists study the consequences of variations in formal arrangements of organizational properties in light of an, often unacknowledged, rational choice approach to an understanding of actors' preferences and motivations. On the other hand, those informed by a more sociological bent tend to focus more intensively on informal institutions. They regard formal institutions as reflecting common goals (and values) at the founding; that is, institutions are artifices that need not be simply efficiency-enhancing, or transaction cost-reducing, contrivances. Moreover, sociologically informed institutionalists consider the effect of transforming the understandings, interests, and preferences of actors in ways that were not anticipated by them at the institution's founding.4 The dividing line between the kinds of institutionalists is between those who believe in structural determinism as opposed to those who favor actor-initiated change. Our early modern mentors in international relations theory did not differ over values: both sides preferred peace to war, life to death, wealth to poverty. They differed over the question of freedom of the will, or the agencystructure problem, as we call it today. The proponents of anarchy held (and hold) that structure triumphs, that humankind chooses under frighteningly tight constraints. The defenders of cooperation argued (and argue) that opportunities exist for the exercise of innovative choices, albeit constrained by factors outside human control. This

basic difference in the assumptions, propositions, and research associated with each side has not been bridged to this day. There are few signs of either side winning a decisive victory. It is difficult to formulate a singular theoretical way to study institutions. Our conceptions go back to the very origin of the modern state system and to our theories about it. With a little legerdemain it could even be argued that the fissure was already present in ancient Greek thought, pitting the realism of Thucydides against the idealism of the Stoics. That fissure divides theories of anarchy from notions of contract; doctrines of competition and war from dogma about cooperation and possible peace; Machiavelli, Hobbes, and Rousseau from Grotius, Locke, and Kant. The arguments have become more nuanced, the data much richer, and the methods now include statistics, gaming, and modeling, but the fissure remains as deep as ever. Not only do the ontologies of students of institutions differ, but the epistemological battles among constructivists, neorealists, neoliberal institutionalists, radical Page 236 → political-economy perspectives, and postmodernists continue to rage unabated.

Positivism versus the Tyranny of Competing Epistemologies Scholars in the natural sciences do not face the problem of reconciling competing protean principles. Their units of analysis lack free will; at least, none has been empirically demonstrated in atoms, molecules, and cells. Therefore, the positivist epistemology that has proved so successful in physics is unlikely to prove equally powerful in the social sciences because the entities studied “talk back” to the scholar. It is relatively simple to agree on what constitutes intellectual progress when there is consensus about methods, about how we establish truth. The schools of international relations scholars to be reviewed here disagree over the espousal of positivism as the sole legitimator of truth; but they agree on rejecting the relativism of postmodernists and some feminist theorists. Hence what may be considered as intellectual progress remains in the realm of contestation. Neorealists and neoliberals do profess to be positivists. Their tolerance for other approaches is always conditioned by their insistence that the arguments of other schools must meet positivist truth tests. Peace theorists in the Kantian tradition disagree, though those who come from social psychology do not. Many world systems theorists eschew positivism, but they may subscribe to scientific realism or pragmatism. Some constructivists also claim to be scientific realists; others follow the pragmatist tradition. Scientific realism, to our minds, collapses to positivism by default or exhaustion. No constructivist accepts positivism as the sole and authoritative guide to sound knowledge.5 We consider ourselves evolutionary “epistemologists” who also embrace rigor; rigor is possible without falling prey to the temptation of positivism. While positivism's achievements in the natural sciences are very impressive, its utility does not carry over unproblematically to the social sciences where social scientists cannot be sure that the “reality” being studied is not a product of the concepts chosen to study it. Even physicists have not uncritically accepted positivist claims. Early physicists Max Planck and Ernst Mach debated the extent to which physics is grounded in actual reality or in intellectual constructs.6 Such debates require social scientists to carefully assess a positivism based on a glorification of physics as an approach for understanding the social sciences. Positivism, and to a lesser extent its cousin scientific realism, relies on a correspondence theory of truth and knowledge. The reality to be comprehended is “real,” whether we perceive it correctly or not. Not so when it comes to the study of international institutions, where discussion mostly depends on a social construction of reality over which various schools of scholarship obviously and painfully disagree. That being so, many truth tests valid in the natural sciences cannot be made to work in the study of international institutions. While modeling and quantification may well add value to less formal modes of discourse, their Page 237 → weight in validating an explanatory argument is weaker in the social than in the natural sciences, and even feebler when it comes to prediction. That, of course, does not mean that positivist procedures are not entirely appropriate in the social sciences when and where their epistemological requirements do not clash with the social construction of reality, which stresses human intention and perception rather than reliance on fixed conditions expressible as crisp variables.

More often though, and certainly for matters that are important and interesting, social science inquiry is outside the positivist domain. A generative mechanism for behavior eludes this approach to knowledge for fundamental reasons. It is difficult to formulate universal claims over time and across cultures because of the mutable nature of institutions and the potential role of free will (that is, of actors' ability to change their minds and pursue new goals).7 Researchers may interfere with the subject of study or find themselves subject to the same set of potentially biasing social influences that shape their very object of study. Mathematics may not be the natural language of human interaction, or at least not a sophisticated mathematics that is incapable of autopoetic and recursive calculations. No wonder Imre Lakatos warned against applying his theory of scientific progress to the social sciences. He did not envision a discipline marked by competition among a number of partly incommensurable intellectual traditions, all enjoying high legitimacy in the eyes of their adherents. He thought of a single established theory undergoing degenerative or progressive evolution, a theory whose adherents shared basic ontological and epistemological commitments. International relations hardly fits this view. Neorealists accord importance to international institutions—defined as organizations and law—only to the extent that they serve to secure the hegemon's interests. World systems theorists regard institutions as stages in, and beyond, the historical evolution of capitalism or as tools of capitalist exploitation. Neoliberals think of institutions—as behavior patterns or as organizations—as devices for coping with uncertainty and limiting transaction costs. Peace theorists and classical liberals equate institutions with arrangements for improving international cooperation, especially if they have their roots in social psychology or in law. Constructivists, though they agree that actors construct their own reality and that ideas are important causes of behavior, exhibit at least three different stances with respect to international institutions. Constructivists who stress state identity subordinate interest in institutions as such to the roles assumed by state actors; constructivists who privilege norms as shapers of behavior see the world much as peace theorists do when it comes to international cooperation: they see institutions as agents of change. Finally, constructivists who embrace an evolutionary epistemology regard institutions partly as arenas for designing change and also as arrangements that bring about change as they alter to the perceptions of their members. Thus, evolutionary constructivists can contribute to a dialogue between observer-scholars and actor-politicians. Page 238 → These diverse positions are troublesome for intellectual progress for two reasons: the different meanings of core terms inhibit discourse in general, and these differences infuse and confuse the discussion of causality in particular. Realists and neoliberals subscribe to a naturalistic conception of reality, in line with their positivist proclivities. Pragmatist-constructivists disagree with their scientific realist cousins about the extent of social construction.

Can the Study of Institutions Be Rescued? The difficulty of attaining a consensual theory of international institutions is overwhelming. Differing notions about institutions are usually pitched at different levels of analysis. Bridging levels in a single theory or approach is possible, but experience has shown that one scholar's claim of successful bridging is often not accepted as persuasive by the entire community of scholars. No Lakatosian cumulation occurs because different approaches are embedded in different scholarly traditions whose competing ontological claims continue to burden our intellects. World systems theory is embedded in historical materialism, neorealism and neoliberalism in microeconomics and utilitarianism. Peace theory has its roots in the social psychology of human imperfection; increasingly it is also inspired by general systems thinking of a very abstract kind, by notions of self-organizing systems of ever greater complexity. Constructivists, by and large, see themselves as the heirs of the sociology of collective behavior. Yet they too remain divided over whether to favor structural predominance, “structuration,” or the evolutionary logic of change that combines intentional causality with natural selection. As scholars we are often asked to specify the “research frontier” in our field. Those who call for that specification

seem to think that once that elusive frontier is crossed, “progress” has occurred or will be made. In the natural sciences that idea is not problematic. Discoveries about nature shape our ideas of “reality out there.” More recent discoveries may demonstrate an imperfect understanding or subsume earlier notions of reality. Yet there is a consensus that incremental increases of understanding of how nature works constitute scientific progress. In the social sciences nothing is “discovered.” Our understanding of social reality is subject to the reflexivity of our minds. Scholars analyze and reanalyze, interpret and reinterpret, conceptualize and reconceptualize the same core questions time after time. Social scientists do not make discoveries in the sense of the natural sciences. Many contemporary discussions of international institutions would be quite accessible to Machiavelli, Grotius, Hobbes, Locke, Rousseau, and certainly to Kant, Bentham, and J. S. Mill. But these foundational scholars would certainly be astonished to learn about new issues on the political agenda that have their origins in global trends and conditions specific to our times. They would be surprised by the advent of a truly global society. The proliferation of new actors would Page 239 → astound them. So would the concerns triggered by humankind's intensified interaction with the material environment. But would all these novelties amount to new constitutive principles of global order unsuspected by our forerunners? The appreciation of the importance of the value of healthy ecosystems to human survival, and the willingness to sacrifice gains in wealth and power to achieve ecological improvement, may well be the only important change in actors' understanding of their role in the world that cannot be explained through traditional IR theoretical accounts of actors' motivations, understandings, and patterns of behavior. Even in the natural sciences consensus on basic theory is never permanent, always subject to challenge and to reconceptualization. The grand theories of today—quantum mechanics, the evolutionary synthesis, and the big bang origin of the universe—may not survive unscathed. Why then burden the social sciences to ever emulate the natural sciences in elaborating grand theories of this type, and expect the entire scholarly community to accept them as true? But the inability of international relations scholars to grasp the positivist grail of discovery is not a sufficient reason for seeking rigor and relevance. We should continually reexamine our theories about institutions and revise them. We ought to continue to debate the significance of international institutions. We as scholars of international relations ought to know more about what these institutions can, or cannot, contribute to peace, wealth creation and distribution, human rights, and sustainable development. We ought to persist in inquiring how institutions fit into larger events and trends, such as growing (or declining) global income inequalities, stronger (or weaker) nationstates. We ought to do this even if basic differences among approaches cannot be synthesized or transcended, if consensual grand theories continue to elude us. We ought to be able to do this without worshipping physics as an exemplar and without embracing positivism as the only valid method.

Pragmatism and Pluralism When Nobel laureate Robert Solow was asked whether economics had made any progress he replied, “My message is that economics makes progress the way any other discipline makes progress; by thinking carefully and respecting evidence.”8 Since we all think carefully and certainly respect evidence, yet need a meeting of the International Studies Association to discuss whether progress has been made, this advice seems to beg the very question. IR specialists have sought to make progress for more than fifty years; yet they continue to disagree over whether they have produced any. Is there something wrong with the very notion of progress as applied to international studies? Yes and no. If intellectual work is judged by the Lakatosian measure, progress is not occurring because the deep chasm between believers in free Page 240 → will and adherents of structural determinism is not being bridged. Moreover, this deep chasm entails continuing differences over types of causality. But students of international institutions are not compelled to judge themselves by these extremely demanding standards. A theoretical contribution may be considered progressive if it changes the way most scholars think

about the problem or puzzle that they consider as calling for an explanation and if it offers the possibility of bridging existing theories at a level below that of grand theory. Under those conditions rival notions of causality can be sidestepped rather than reconciled. Progress occurs, for instance, when two theories, one positivist and tied to a causality of statistical covariation and the other constructivist and based on intentional causality, nevertheless illuminate the same puzzle, or explain away the same problem, or provide synergistic explanatory accounts.9 In that case fundamentals are not reconciled, but interesting phenomena are confirmed intersubjectively. IR trends may be fruitfully explored, and the causal understanding of the interplay between forces typically analyzed by discrete schools is advanced. Pragmatists, including ourselves, while remaining within the family of scholars who believe in the possibility of establishing the truth of propositions about the world, nevertheless differ from positivists because we eschew any crisp notion of causality. Pragmatism rests on the following foundations, and thus is particularly attractive to constructivists.10 The world is not immediately accessible to actors. Actors behave willfully, but not all share common perspectives or understandings. Willful actors, through intersubjective activities, construct the world. Institutions are thus causes and effects of this process. Different propositions need to be developed to deal with each step. Understanding can only be made intelligible via a consensus theory of truth, since our concepts and understandings about the world do not necessarily correspond to the world as such. Consensus works by shared epistemological commitments by a community of scholars embedded in a receptive and supportive milieu. Knowledge cumulation is a social activity conducted by communities of scholars rather than by individuals working in isolation. Conclusions reached through a consensual process are more likely than individual efforts based on a correspondence theory of truth, and a positivist approach to understanding, to yield policy-relevant solutions that work. Because the determinations are more likely to be true in a limited sense, having been derived through a collective debate, and because the conclusions themselves are politically attractive, they are seen in democratic societies to be impartial and uncompromised by political bias. Finally, pragmatists retain the Enlightenment faith in reason that political emancipation may be achieved through publicizing warranted social findings through the double hermeneutic or reflexive social science. Unlike positivism, such results are collective and warranted by group appraisal. Truth—an emancipatory political project—is most likely to arise in a democratic culture of tolerance for public discourse and intellectual curiosity and with some independence from funding sources. Unlike radical interpretivism, truth claims are warranted Page 241 → by the process by which they are generated rather than their resonance with actors' norms. Pragmatists hold that truth is provisional. They prefer to think in terms of complex relations among variables, not simply “A is explained by B” because the two are logically associated with each other. Unfortunately, some positivists mistake statistical association for causation, and also for a direct correspondence with the “natural world.” They also make their peace with the short half-life of most theories, and thus few aim at finding general covering laws. Some constructivists, however, shade into a commitment to scientific realist epistemology because they wish to demonstrate the reality of phenomena not amenable to direct observation, an ability most other constructivists do not claim to possess. Progress can be had if we lower our ambitions. There will be no consensual grand theory for us. Neither synthesis nor cumulation seems within our grasp. What to do? We might focus our explorations on substantive areas of general concern. Instead of asking, “How do we explain the origin of institutions?” we should ask “Which institutions are most likely to bring about a peaceful (or egalitarian, or wealthy) world?” Or we might profitably ask, “Why have international institutions failed to deliver significant economic development?” or, “Under what circumstances may institutions with particular designs affect behavior in a desirable way?” It makes sense to speak of a problem-focused synthesis rather than continuing to pitch progress at the macrotheoretical level because different approaches may well illuminate different aspects of the puzzle and still allow a comprehensive response to the question.11 If the strictures of positivism or realism can be satisfied in the conduct of this research, so much the better. If not, all ought to be able to make their peace with the looser notion of causality urged by pragmatists.

It bears repeating that the evolutionary epistemologist stance, a variant of pragmatist-constructivism, is by no means consistently hostile to all aspects of positivism, though it is less positivistic than are constructivists who focus on state identity. Positivism supposedly undergirds “rational” ways of studying international institutions; constructivism is alleged to rely on Verstehen, on “reflection,” on “interpretive sociology.” It is simply incorrect to argue, as some do, that all branches of constructivism ought to be sharply juxtaposed to the epistemologies of “hard science.” Max Weber, the inventor of interpretive sociology, was more judicious in his formulation. He distinguished between “observational” and “explanatory” understanding. The former involves directly watching and then interpreting the significance of some item of actor conduct, such as joy or anger. The latter calls for the determination of the actor's reason for acting in the observed manner, the study of motivation. Weber is careful to specify that the mode of explanatory understanding is “rational understanding of motivation, which consists in placing the act in an intelligible and more inclusive context of meaning.”12 Actors are assumed to be acting in accordance with what they see as the best way to attain their objectives; motives are “rational” Page 242 → in that sense, even if not in the perspective of microeconomics or game theory. Therefore they can be studied with systematic methods and subjected to rigorous hypotheses that may even be falsifiable. Interpretation does not have to be the equivalent of mushiness. Hence patterns of behavior ascertained by such methods can be fully consistent with the assumptions of evolutionary epistemology. Nevertheless, the progressive cumulation of knowledge of the Lakatosian positivist universe is not likely in the more contingent style of inquiry of this branch of constructivism. Moreover, policy relevance remains possible if concerns are kept modest. Even though IR scholars cannot claim to have made important discoveries at more modest levels of discourse, something like progress occurs. All schools of thought in international relations (except for the realists) now agree that the shape of things has fundamentally changed because of these developments in the real world of the last fifty years: sovereignty has become a contested concept, global civil society is emerging, domestic civil society has blown up the unitary state, equality among states is becoming more than a legal conceit, the redistribution of wealth is accepted as given by most actors, sustainable development is a goal for all. Theory has always followed diplomatic and military practice. The arrival of these notions makes it evident that theory and practice continue to reinfect each other in the face of macrotheoretical commitments that ignore more modest insights.

Methodological Injunctions If knowledge is socially constructed, how can one, within a social setting, analyze social processes? We suggest that unpackaging types of social construction can accomplish this. Fundamentally, there are three different types of knowledge, as defined by philosophers of science. “Brute facts” are those that are subject to intersubjective observation and verification by analysts.13 This is the traditional area of natural science where, ultimately, nature can bite back. Second, there are “hybrid facts,” where analysts apply frames in excess of their principal understanding of the phenomena. These domains are much more common in the study of international institutions, including environmental and economic regimes. Last is the domain of “social facts,” which exist ultimately by definition and have no meaning independent of the social world in which they exist, such as international law, human rights, and even military deterrence. Being self-conscious about each of these types of “fact” is essential for making a methodologically sound constructivist argument. Jumping from one to the other, such as mistaking a hybrid fact for a social fact and back again, discredits the reliability of constructivism. Cumulative progress in creating consensual knowledge about institutions rests on a number of individual steps: making individual claims in single works done in the constructivist style; reconciling individual claims within constructivism (intraparadigm progress); and conducting interparadigm midlevel discussions that try to resolve different interpretations of similar phenomena and conceptual applications that may lead, ultimately, Page 243 → to some degree of provisional closure and dispute resolution between paradigms. First is the development of good work by individuals, an exercise that is closest to positivism in the need for rigor and self-conscious assumptions, or of the propositions that guide the research. Individual claims should rest on the logical consistency behind the analysis, possible falsification of the underlying claims, and some degree of empirical testing. By way of methods, constructivist work that heeds seriously the epistemological injunctions

about theorizing about ineffable actor beliefs is probably best performed with qualitative methods, rather than the statistical and behavioral ones typically associated with positivism, because of the difficulty of developing reliable indicators of such variables. Thus, focused comparative studies, process tracing, and even counterfactuals are able to focus attention on the vital variables that constructivists believe to drive institutional politics.14 Reconciling individual contributions within a common theoretical tradition should be possible through the same techniques of careful testing of propositions and findings in compatible domains. Cumulation and progress may occur when lessons are drawn from such individual contributions. How do they fit with views from other approaches, are they complementary in terms of assumptions, and do they qualify or challenge key beliefs? Ideally progress should occur between paradigms via a Lakatosian process of hypothesis testing and Popperian falsification, with an eye toward the value of contributions to developing new insights rather than marginal improvements on a given tradition. However, we have difficulty recalling the last time we saw scholars admit to error or publicly change their minds on the basis of these procedures. In practice cumulation probably occurs over a span of generations through a sociological process of cohorts of trained graduate students replacing a previous generation of scholars. In the shorter term debates are often characterized by rhetorical and polemical interparadigm debates that convince no one. Still, there are pragmatic institutional design lessons about how to promote consensus through cumulative exercises that inhibit egregious bias. Conscientious debates among well-intentioned scholars should be encouraged in peer-reviewed journals, including multiauthored works and edited volumes organized around common themes.

Conclusion Our notion of progress implies the legitimacy of a plurality of approaches. It suggests tolerance for competing ontologies and epistemologies. However, we are not being tolerant for the sake of being tolerant, of being nice to our intellectual antagonists. Tolerance is an inevitable by-product of the double hermeneutic in which we, as pragmatist-constructivists, are stuck. Intellectual pluralism is inescapable because of the commitment to pragmatism. We concede that our social construction of reality cannot be proved superior to anyone else's. In analyzing the social construction of actors we remain embedded in our own, which will differ from those of other scholars. The Page 244 → achievement of modest progress depends on continuing the conversation among all of us. Nobody finally wins or loses in this perspective. The grand theoretical stakes are low because any cumulating that may be achieved is temporally and spatially limited and unlikely to challenge any deep commitments. But a few puzzles get solved along the way.

Notes This paper was originally prepared for the millennial reflections panel on institutions held at the annual meeting of the International Studies Association, March 16, 2000. We are immensely indebted for critical comments to Anne Clunan, Peter Katzenstein, Craig Murphy, and M. J. Peterson. 1. A good collection of the seminal articles is presented in Louis Menand, ed., Pragmatism (New York: Vintage Books, 1997). 2. The philosophy of science against which these categories are constructed is elaborated in Michael Martin and Lee C. McIntyre, eds., Readings in the Philosophy of Social Science (Cambridge: MIT Press, 1994); Richard J. Bernstein, Beyond Objectivism and Relativism (Philadelphia: University of Pennsylvania Press, 1983); Ernest Sosa and Michael Tooley, eds., Causation (Oxford: Oxford University Press, 1993); David Braybrooke, Philosophy of Social Science (Englewood Cliffs, N.J.: Prentice Hall, 1987); John Searle, The Construction of Social Reality (New York: Free Press, 1995); Anthony Giddens, The Constitution of Society

(Berkeley: University of California Press, 1984); Martin Hollis and Steve Smith, Explaining and Understanding International Relations (Oxford: Oxford University Press, 1990). 3. Jaegwon Kim, Philosophy of Mind (Boulder, Colo.: Westview, 1996), 5. 4. Peter M. Haas and Ernst B. Haas, “Learning to Learn,” Global Governance 1 (September-December 1995): 255–84; Ernst B. Haas, When Knowledge Is Power (Berkeley: University of California Press, 1990). 5. We use these philosophy of science categories as explained in Larry Laudan, Science and Relativism (Chicago: University of Chicago Press, 1990). See also his Beyond Positivism and Relativism (Boulder, Colo.: Westview, 1996) for the notion of pragmatism as we use it. Evolutionary epistemology is defined and explicated by Donald T. Campbell in “Evolutionary Epistemology,” in The Philosophy of Karl Popper, P. A. Schilpp (La Salle, Ill.: Open Court Publishing, 1974), 413–63; and Kai Hallweg and C. A. Hooker, eds., Issues in Evolutionary Epistemology (Albany: State University of New York Press, 1989). For more on the application of evolutionary epistemology and consensus theories of truth see Donald T. Campbell, Methodology and Epistemology for Social Sciences (Chicago: University of Chicago Press, 1988); and Steven Toulmin, Human Understanding (Princeton, N.J.: Princeton University Press, 1972), app. 1. 6. Stephen Toulmin, ed., Physical Reality (New York: Harper Torchbooks, 1970). 7. R. Harre and P. F. Secord, The Explanation of Social Behavior (Oxford, England: Basil Blackwell, 1972), 9–12ff. 8. Bulletin of the American Academy of Arts and Sciences, December 1982, 30. 9. Syncretic individual volumes include Glenn H. Snyder and Paul Diesing, Conflict among Nations (Princeton, N.J.: Princeton University Press, 1977); Janice Gross Stein and Raymond Tanter, Rational Decision-Making (Columbus: Ohio State University Press, 1980); Graham Allison, The Essence of Decision (Boston: Little, Brown, 1971); Peter M. Haas, Saving the Mediterranean (New York: Columbia University Press, 1990); Daniel Deudney and G. John Ikenberry, “Soviet Reform and the End of the Cold War,” Review of International Studies 17 (July 1991); Daniel Deudney and G. John Ikenberry, “Who Won the Cold War?” Foreign Policy 87 (summer 1992); Richard Ned Lebow and Thomas Risse-Kappen, eds., International Relations Theory and the End of the Cold War (New York: Columbia University Press, 1995). For complementary works that include different approaches and yet address a similar question, see the following. On the Cuban missile crisis see James Bright, Cuba on the Brink (New York: Parthenon, 1993); and James Bright and David Welch, On the Brink (New York: Page 245 → Hill and Wang, 1989). For environmental studies from multiple perspectives see Richard Benedick, Ozone Diplomacy (Cambridge: Harvard University Press, 1998); Karen Litfin, Ozone Discourses (New York: Columbia University Press, 1994); Peter M. Haas, “Banning Chlorofluorocarbons,” in Knowledge, Power, and International Policy Coordination, ed. Peter M. Haas (Columbia: University of South Carolina Press, 1997); and Edward Parson, “Protecting the Ozone Layer,” in Institutions for the Earth, ed. Peter M. Haas, Marc A. Levy, and Robert O. Keohane (Cambridge: MIT Press, 1993). 10. Prominent pragmatists who are clustered together for the purpose of this summary include Charles Peirce, William James, John Dewey, Jurgen Habermas, Paul Ricoeur, Charles Taylor, Hilary Putnam, and Richard Rorty. See Menand, Pragmatism, for a useful summary. 11. We are comforted by this dictum of Max Weber's: “Meaningfulness naturally does not coincide with laws as such, and the more general the law the less the coincidence. For the specific meaning which a phenomenon has for us is naturally not to be found in those relationships which it shares with other phenomena…. In the cultural sciences the knowledge of the universal or the general is never valuable in itself.” Quoted by David Dessler in “Constructivism within Positivist Social Science,” Review of International Studies 25 (1999): 136. Emphasis in original. Taken from Max Weber, “‘Objectivity’ in Social Science and Social Policy,” in Methodology of the Social Sciences, ed. E. Shils and H. Finch (Glencoe, Ill.: Free Press, 1949), 76. 12. Max Weber, “The Interpretive Understanding of Social Action,” in Readings in the Philosophy of the Social Sciences, ed. May Brodbeck (New York: Macmillan, 1968), 25–26. For an extended argument that this mode of inquiry is consistent with positivism, albeit not with its most demanding form that insists on strict correspondence between nature and the concepts used in discovery, see Gil Friedman and Harvey Starr, Agency, Structure, and International Politics (New York: Routledge, 1997), ch. 4. 13. G. E. M. Anscombe, “On Brute Facts,” Analysis, no. 184 (March 1958): 69–73. 14. Alexander L. George and Timothy J. McKeown, “Case Studies and Theories of Organizational

Decision-Making,” Advances in Information Process in Organizations 2 (1985): 21–58; David Collier, “The Comparative Method,” in Political Science: The State of the Discipline, ed. Ada W. Finifter (Washington, D.C.: American Political Science Association, 1993); Andrew Bennett and Alexander George, eds., Case Studies and Theory Development (Cambridge: MIT Press, forthcoming).

Page 246 →

EN ROUTE TO KNOWLEDGE Is There a “Third Path” (in the Third Debate)? Yosef Lapid A third path leads beyond modern and postmodern methodological debates in the social sciences, history, and the humanities. It turns out that choices between the routes of science and interpretation, history and theory, objectivism and relativism are more illusionary than real.1 Notwithstanding the general understanding that the end of the millennium may well represent an arbitrary or insignificant temporal marker, few scholarly disciplines (international relations [IR] included) have been able to resist the temptation of using this festive occasion as a pretext for another round of extensive and vigorous stock taking. Michael Burawoy, a Berkeley sociologist, perceptively notes that at stake in such evaluative exercises is not so much “what we don't know, but how to interpret what we do know”2 in our respective disciplinary domains. Our presentation guidelines suggest that such was also the intent behind the ten “millennial panels” that occasioned this theoretical collection. Participants were invited to address six carefully formulated and closely related tasks. Of these, however, only the last one (dealing with future themes and research agendas) addressed specifically the issue of what we “don't know” (or what we need to know) in international relations. The other five engaged in different ways the far more difficult task of interpreting or assessing what we (think) “we do know” about our notoriously elusive subject matter. In the following, I focus my comments on the first and most general task, namely a “self-critical state-of-the-art reflection on accomplishments and failures.” In accord with my panel assignment, “alternative and critical perspectives” will be the referent object of my comments. However, insofar as the meaning of being “critical” (or alternative) is anything but self-evident, some preliminary clarifications are in order. In the aftermath of more than a decade of “third debating,” it should be evident that there is no unitary “alternative” position, nor a simple or single way of demarcating the “critical” from the “noncritical.” In international relations as in other scholarly fields critical voices are informed by vastly different (neo-Marxist, poststructuralist, feminist, postpositivist, and so on) ontological and epistemological Page 247 → premises. Reproducing multiple lines of division evident more generally in recent social theory, IR's alternative and critical perspectives thus constitute a fragmented and slippery domain that firmly resists undifferentiated assessments. Since a suitably nuanced mapping of this vigorously contested domain is precluded by the limited scope of this essay, I will engage the issue of IR criticality through the related concepts of pluralism, dialogue, and reflexivity. To be sure, such a demarcation of criticality will be deemed excessively narrow and anemic, especially by those who insist on infusing critique with a massive dosage of political engagement. As it happens, I tend to agree that a high level of social concern on the part of the scientist is indeed a defining attribute of any scholarly enterprise that claims a “critical” mantle. However, consistent with the main argument developed in this paper, with respect to the question of how “emancipatory” can a critical inquiry become without compromising its scholarly value, I wish to highlight, in particular, the wisdom of exploring more fully the promise of possible, but currently still submerged, “third way/middle ground” positions.3 At least two justifications can be invoked at this point in support of the following attempt to assess accomplishments and failures of critical and alternative IR approaches with a triple optic of pluralism, dialogue, and reflexivity. First, pluralism, dialogue, and reflexivity are widely recognized across the humanities and the social sciences to be both closely associated with, and supportive of, critical and alternative agendas. Second, and perhaps more important, pluralism, dialogue, and reflexivity have been widely recognized as potential moments of special productivity in IR's disciplinary transition known as the “third debate.” Initiated by a loose coalition of postpositivist/postmodernist/poststructuralist/constructivist/feminist/normative and other “critical” and/or “dissident” challengers, the third debate registered extensive demands for radical change. These included calls for

a reflexive reexamination of ontological, epistemological, and axiological premises and a serious reassessment of disciplinary options and practices in a rapidly changing intellectual and political landscape. Some envisioned an epistemologically more pluralistic and creative discipline, willing to experiment with new ways of doing science. Others sought an intellectually less parochial, less insular, and more interdisciplinary knowledge-producing enterprise. Still others hoped to institute “interpretation” or “emancipation” (as opposed to “explanation”) as the defining telos of a more vibrant science of international relations. Most, however, looked forward to a more open, inclusive, and dialogically engaged scholarly community capable of communicating effectively (if not consensually) across separate categories of all sorts (levels of analysis, paradigms, disciplines, genders, continents, and so on). In this sense, pluralism, dialogue, and reflexivity are properly understood as common denominators of sorts in the change-oriented disciplinary transition known as the third debate. To the detriment of the entire third-debate project, most of these initial demands were put on the table by some of the most vocal and polemical proponents of radical change. As a result, it was initially rather difficult to Page 248 → discern that this “opening-up” momentum was tacitly sustained also by a more general sentiment (spreading deep into “mainstream” positions) that things have become excessively stale in the study of international relations. My main argument in this paper is based on the premise (and hope) that this polarized situation may now be finally changing. In recent years, we have witnessed multiple (but only partially convergent) attempts to nudge the center of gravity in the third debate away from polarized (either/or) confrontations in the direction of more moderate via media (both/and) type postures. This median space is still consolidating and will likely face mounting opposition from both mainstream and critical theoretical quarters. Should it prevail, however, disciplinary ramifications for pluralism, dialogue, and reflexivity can be far reaching and largely positive.4 With this general argument in mind, and with the benefit of more than a decade's hindsight, I now revisit the merits of contending disciplinary responses triggered by the third-debate transition. I do so in two stages. First, I offer a brief impressionistic reading of the current status of the three disciplinary responses (i.e., despair, celebration, and reconstruction) mentioned originally in my third-debate article published in 1989.5 Building on this reading, I then engage in a more detailed discussion of how the IR discipline may better situate itself to become a main beneficiary of the current interest in via media solutions to seemingly insoluble knowledge problems. Throughout this discussion, the main questions remain the same: What is the relationship between the third debate and theoretical progress in international relations? Was the third debate able to, at least partly, honor its promise of a vastly improved scholarly enterprise? Are we still in the third debate or have we moved, as Ole Waever and others argue, to a “fourth” or perhaps “fifth” debate?6 And, finally, are there any reasons to be guardedly optimistic regarding the “next stage” in IR's continuing disciplinary transition?

A Decade Later: Despair, Celebration, or Reconstruction? In view of the fact that knockout-type eliminations are as rare in international relations as in other social sciences, it is hardly surprising that all three responses (despair, celebration, and reconstruction) that were embryonically present back in 1989 are still visible today. This is, however, different from affirming a static or stagnant picture. For subtle changes in both the constitution and relative standing of these three responses are clearly evident. Starting with the despairing response, my sense is that it has largely maintained its relative vigor and standing in the IR domain. One detects, however, a noteworthy change in the constitution of this response. Back in 1989, the despairing camp drew most of its support from manifestly alarmed mainstream quarters. At the time, mainstream opponents seemed intent on stemming the tide of what they perceived to be cleverly camouflaged “knowledge viruses,” ingeniously designed by sinister third-debate programmers to infect, erase, and/or otherwise “deconstruct” the meager fund of “positive” knowledge that had been slowly Page 249 → accumulating in the IR field. Switching metaphors, in the late 1980s and early 1990s, mainstream opponents operated with a crude “roach motel” image of the third debate: IR theories checked in, but they did not check out. Needless to add, such a negative posture left little room for optimism regarding the third debate. As evidenced, however, by the recent spectacular rise of constructivism in international relations, mainstream circles are nowadays far less suspicious or paranoid about third debate-minted ideas. This more relaxed attitude may have been there already in 1989, but it was fully eclipsed by the fiercely polarizing logic of the opening

salvos in the third debate. For our present purposes, suffice it to say that the despairing camp has not been seriously weakened by recent mainstream defections because of parallel new defections of deeply disillusioned scholars from the reconstructive and/or celebratory camps. Many applicable examples come to mind. Barry Buzan and Ole Waever—whose work on identity and security, for instance, is exemplary in the theoretical “reconstruction” mode—have come up recently with separate and rather strong indictments of the IR theoretical enterprise.7 Similarly, Ann Tickner—whose initial response to the third debate was rather celebratory—has grown, over time, increasingly pessimistic over the prospects of ever reaching a meaningful conversational encounter between IR feminists and the IR mainstream.8 And, finally, Steve Smith's contribution to this millennial panel offers yet another compelling example of a despairing retreat into pessimism on account of a putatively decisive victory by the U.S./IR academic mainstream in its battle with critical and alternative approaches over the authorization of knowledge in international relations. As in 1989, I remain fairly skeptical of such despairing disciplinary stories. I continue to believe that it is simply wrong and counterproductive to invoke the imagery of a scholarly expedition aboard a Titanic, tragically headed toward some epistemological, ontological, or axiological “iceberg.” Ten years into the third debate, IR's journey to the land of better knowledge continues. Having taken, or at least scouted, the interpretive, the linguistic, the constructivist, the rhetorical, the normative, the critical, the spatial, and many other real or imaginary “turns,” the IR scholarly expedition has not stumbled into any unscientific black hole. To the contrary, as noted recently by Samuel Makinda, “A distinctive feature of recent IR debates is that most theorists tend to address more critically such themes as science and interpretation; objectivity and subjectivity; norms and practices; and materialism and idealism.”9 To that extent, the journey has become more pluralistic, more lively, and also more promising. Which takes me directly to the celebratory response. Back in 1989, I paid almost exclusive attention to this optimistic reaction because, at the time, it seemed to be ascendant in the IR field. The situation is quite different today. In recent years the celebratory camp has been constantly losing altitude, as serious pessimism started to set in regarding the actual limits of the revitalizing theoretical energy released by the third debate. To better Page 250 → understand the reason why this growing pessimism has managed to gain such momentum, it is helpful to distinguish between two items on the celebratory agenda that happened to be destined to very different fates in the actual unfolding of the third debate. The first agenda item can be provisionally characterized as an “opening-up/pluralism” project. It sought a more open, diverse, and interdisciplinary theoretical enterprise. It tried to drive home the message that having a productive discipline in no way requires that diversity and interdisciplinarity be suppressed. This agenda item registered in my view significant and, most likely, irreversible achievements. To be sure, additional work remains to be done, for instance, on the “IR-as-an-American-discipline” front. There is also considerable substance to the lingering suspicion that much of this “opening” may have been more cosmetic and cynical than real.10 But things cannot be so terribly wrong if even as disgruntled a participant as David Campbell can now concede that “where once we were all caught in the headlights of the large North American car of international relations theory, now the continental sportster of critical theories has long since left behind the border guards and toll collectors of the mainstream—who can be observed in the rearview mirror waving their arms wildly still demanding papers and the price of admission—as the occupants go on their way in search of another political problem to explore.”11 Arguably, what holds true for “the continental sportster of critical theories” is equally true for many other approaches whatever their destination, preferred mode of transportation, and location vis-à-vis “the large North American car of international relations theory.” Further testimony that such is the case can be adduced, for instance, from James Der Derian (the actual originator of the catchy “large North American car” metaphor) who in a recent publication simply refuses to offer any apologies for his “eclectic group of travel companions” and flatly asserts instead that “whichever theorist helps me best understand the subject of my inquiry gets to the head of the class.”12 There was, however, a second, and far less successful, item on the third-debate agenda. This second agenda item expressed the hope for a more dialogically oriented and communicatively engaged scholarly enterprise. To the extent that scholarly disciplines are well described as conversational communities with a tradition of argumentation, it is clear that the third debate has energized “argumentation” at the expense of “conversation.”

Under the third debate, IR theoretical discourses have proliferated, and some have become considerably more vibrant but also correspondingly less “engaged.” The growing disarray of concepts, theories, and approaches has, in fact, become so overwhelming as to prompt descriptions of the IR field as a “paradigmatic dysfunctional family.”13 Occasional attempts to engage in sustained dialogues have been regularly undermined by oppositional rhetoric that quickly turned them into bitterly contested battlegrounds. Lost in such escalations was the realization that a discipline's (communicative) ability to impart or share information is not necessarily less important than its (epistemological) ability to produce and certify “true” knowledge, and the two are, of course, intimately related. Page 251 → The problem is now well understood in our discipline, and the urgent desire to find new workable solutions seems nearly universal. This became evident, for instance, in a Visions of International Relations series held recently under the auspices of the Walker Institute at the University of South Carolina. In his introduction to a forthcoming volume based on this project, Donald Puchala forcefully captures this sentiment. “Pluralism,” he says, with positive affect attached, turns up in text after text, and a definite point of convergence among the contributors is their common concern about the fragmentation of scholarship in International Relations—the “insularity of scholarship,” the “incommensurability of discourses,” the “mutual inattentiveness” of specialists, the growing intolerance of intellectual diversity, and even the degeneration of civility in the academy are all bemoaned. The call of almost all in this volume is for “bridge building” across the paradigmatic divides, “dialogue” and “interaction” among contending schools and different specializations, avoiding extremes, seeking the via media while generally attending to the task of trying to better understand the real world despite the distractions of academia.14 The same sentiment was voiced in the annual addresses of two past ISA presidents.15 Finally, a similar concern was echoed also by Thomas Bierksteker in his opening editorial article for a newly (re)launched IR journal, where he suggests that we must “give primacy to approaches that can accommodate incompatible or incommensurable differences, rather than to try to choose a single ‘winner’ out of competing claims.”16 It is against this general background that the question posed in the title of this essay can be best understood. If our main disciplinary problem today is the eruption of “flabby”—as opposed to “engaged”—pluralism,17 is taking the third way in the third debate a possible solution? Is there, in other words, an intellectually viable middle ground that carries the promise of dialogue and deliberation (but not necessarily consensus or agreement) across deeply divisive ontological, epistemological, and axiological lines? In the next section, I cautiously answer this question in the affirmative. Caution is needed because there is nothing automatically or necessarily redeeming in the median space. To the contrary, the idea lends itself to serious misunderstanding and misappropriation. Indeed, there are already some reasons to suspect that the recent popularization of the via media idea through IR constructivism (Emanuel Adler and Alex Wendt, in particular) may turn out to be detrimental to a broader and more rewarding disciplinary engagement with the median space.

The Third Way: From “Flabby” to “Engaged” Pluralism in International Relations An interesting set of developments, both internal and external to the IR discipline, has combined to elevate the appeal of third way-type reasoning. Page 252 → Whatever its intellectual or political merits, third way thinking seems now increasingly in tune with the emerging spirit of our times.18 Such is particularly the case in the academic world, where one detects a growing fascination with, and broadening support for, via media positions. In a recently published book Francois Dosse vividly captures the rising centrality of this intellectual median space, which, in his view, opens up exciting possibilities for new alliances between the exact sciences, the social sciences, and the humanities.19 Interestingly enough, a parallel reorientation toward the middle ground is now tentatively evident also in international relations.20 This is, I submit, a promising but far from risk-free development. If properly managed, it may eventually blossom into a productive disciplinary discussion regarding the “next stage” of the third debate.

To facilitate such an outcome, in the following, I sketch a more analytical understanding of the median space (i.e., the third way, via media, middle ground) idea and spell out some of its ramifications for reconfiguring pluralism, dialogue, and reflexivity in the IR domain. The main objective throughout will be to explore the possibility of transcending sterile paradigmatic fragmentation with a more carefully fine-tuned third way heuristic. Erik Doxtader's insightful elucidation of the “communicative qualities” of the “middle” in public life will start us in this general direction.21 Doxtader argues “that a manifold exploration of the middle of public life may help to unravel the puzzle of how public transgression and opposition facilitates dialogue and mutual agreement.”22 His rich and philosophically informed discussion presents the middle as a “space of deliberation that contains both opposition and agreement” as a “movement between transgressive and intersubjective modes of communication” and as a “norm that moderates but does not mediate the dynamic elements of public deliberation.”23 Following Hegel, Doxtader portrays the middle as a productive space that is both “found” and “created” it is also a reflective space “which is simultaneously interior and exterior to the self.”24 To reiterate, however, Doxtader's main interest in the middle is as a “heuristic” that explains “how the negativity of opposition can create the ground of mutual dialogue.”25 At the end of his article, Doxtader extends an open invitation to all scholars to “find other middles and investigate their communicative qualities.”26 I suggest that the IR scholarly community would be well advised to accept this invitation and proceed with a serious and reflective investigation of the “middle” of the third debate. Unfortunately, with few exceptions, this is not the direction in which the rising IR interest in the median space is currently evolving.27 International relations' via media may be “under construction,” but there is little evidence of a plausible and imaginative architectural design to guide a productive implementation of such a critical disciplinary project.28 Missing, in particular, is any sustained reflection on whether and how this third way can facilitate new types of pluralism, dialogue, and reflexivity in international relations. And yet, in lieu of such reflection, there is ironically a serious risk that the via media project may itself become an additional bone of contention in an already overly fragmented and polarized discipline. Page 253 → Spirited, if often confused, responses elicited by high-powered constructivist claims to exclusive ownership of the median space in the current disciplinary debate vividly attest to the gravity of the aforementioned risk.29 While clearly justifiable, the decision of some English School proponents to present a historical counterclaim to the very same space may be equally misguided.30 Indeed, facile presentations of such claims and counterclaims can only invite further third way maneuvers (from international law, for instance) “that would make even Prime Minister Tony Blair proud.”31 Whether such maneuvers can also make the IR discipline proud is, of course, another matter. Now, to the extent that with respect to the middle ground, the point is “not to argue in favour of any specific approach,”32 what else can IR scholars do with it in the context of the third debate? My answer returns to the themes of pluralism, dialogue, and reflexivity. To further explicate the promise of rethinking these three themes in the context of a via media approach, I refer briefly to the work of Eviatar Zerubavel (a Rutgers sociologist), who profiles three intellectual “mind-sets” (the rigid, the fuzzy, and the flexible) that correspond to three modes of “sculpting” academic identities.33 The rigid mindset feels most comfortable in a context of clear-cut distinctions and sharp demarcations. It promotes a world made up of discrete, insular entities. It discourages contacts and bridge building between such entities, and it leads a sustained campaign against the vague, the in-between, and the nested. In the academic context, the result is a highly compartmentalized, that is, fortresslike, pluralism. The fuzzy mind-set is the diametrical opposite (the Other) of the rigid mind-set. It is suspicious of all borders and resentful of all efforts to curb curiosity and focus intellectual attention. When applied to an academic context, the fuzzy mindset results in nondisciplinary, or even contradisciplinary, sculpting practices, as it refuses to partake in the constitution and sedimentation of separate academic identities. The result is “flabby” (i.e., “anything goes”) pluralism. Zerubavel prefers, however, the flexible mindset, which offers an attractive via media between absolute rigid-mindedness and virtual fuzzy-mindedness. In his own words, the flexible mind-set “fosters an intellectual environment that would allow for both order and creativity, structure and open-mindedness, focus and change.”34 Against this background, it is reassuring to note that the emergence of a flexible mind-set/middle ground orientation as a guiding vision for would-be disciplinary “sculptors” may be sufficiently important to justify

speculations on a new stage in IR's evolving disciplinary debate.35 Relevant in this context is the fact that academic identities sculpted under the influence of the flexible mind-set can mature into a type of pluralism approvingly described by Richard Bernstein as “engaged pluralism.” Bernstein believes that this type of pluralism “has been characteristic of what is best in our own American philosophic tradition…. One accepts the multiplicity of perspectives and interpretations. One rejects the quest for certainty, the craving for absolutes, and the idea of a totality in which all differences are finally reconciled. But such a pluralism demands an openness to what is different and other, a willingness to risk one's prejudgements, seeking for common ground without any guarantees that it will be found.”36 To that extent, via Page 254 → media-type pluralism can teach us how “to think and act in the ‘in-between’ interstices of forced reconciliation and radical dispersion.”37 A better understanding of the manner in which via media approaches may facilitate new and more productive understandings of pluralism in the social sciences (IR included) also brings to the fore the realization that some lingering scholarly disagreements may be rooted in “differing ideas about the nature of dialogue, or about how social scientists participate in dialogue.”38 Indeed, a better understanding of the communicative qualities of different types of dialogical encounters can be a prime benefit of a more energetic engagement with the middle of the third debate along the lines advocated in this essay. As aptly pointed out by Nicholas Burbules, “Dialogue represents, to one view or another, a way of reconciling differences; a means of promoting empathy and understanding for others; a mode of collaborative inquiry; a method of critically comparing and testing alternative hypotheses; a form of constructivist teaching and learning; a forum for deliberation and negotiation about public policy differences; a therapeutic engagement of self-and Other-exploration; and a basis for shaping uncoerced social and political consensus.”39 Furthermore, dialogues can yield a wide range of possible outcomes including agreement, consensus, understanding, tolerance, and even communicative failure or “incommensurability.” For as pointed out by Burbules, “There are instances in which the very encounter with a radically different, unreconciled, and unreconcilable point of view, value, voice, or belief can serve important educational purposes; to cause us to question the horizons of our own assumptions, to explore within ourselves (and not only within the other) the cause of why dialogue ‘fails’ and to consider the possibility of a radically different way of approaching the world. Dialogue in the mode of resolving or dissolving differences provides no tools for coping with such encounters or deriving meaning from them; it regards them as failures or breakdowns, and not as limitations within the model of dialogue itself.”40 It is at this point that the themes of dialogue and pluralism tie into the broader theme of reflexivity. As noted by many observers reflexivity can be, and has been, practiced dogmatically in the IR domain even by “critical,” “dissident,” or “alternative” scholars.41 An additional promise of the median space is to facilitate more genuinely reflexive practices in the IR domain. That such a promise is entailed in a third way reorientation is suggested by Zerubavel's notion of a “flexible mind-set,” by Doxtader's depiction of the middle as a “reflexive space,” and by Bernstein's ideal of “engaged pluralism.” To date, the IR scholarly community seems only dimly aware of the many intellectual and communicative opportunities that can be seized by a serious engagement with the middle of the third debate. It is, however, to the credit of this “discipline-defining” debate that the possibility that we may actually seize upon these opportunities cannot be entirely precluded.

Concluding Remarks: Bringing Goldilocks Back into IR Theory? Evaluative appraisals of the kind undertaken in this essay are frequently referred to nowadays as disciplinary “stories” or “fairy tales.” If so, this was Page 255 → neither a “Sleeping Beauty and the prince” nor a “Titanic and the iceberg” story. Rather, it was a “bringing Goldilocks back into IR theory” story. Scientists refer to the idea that a planet needs to be “just right” (i.e., neither “too hard” nor “too soft”) to support life as the “Goldilocks principle.” I submit that a Goldilocks principle of sorts is at work also in the constitution and operation of vibrant scholarly communities. My argument in this paper suggests that the “next stage” in the third debate may depend to a great extent on our understanding of where the IR discipline is situated today vis-à-vis this principle. To be sure, there is nothing original or new in such third way contentions. Inside and outside the academic world, they have been known to please some people and to infuriate others. I can readily anticipate similar reactions to this presentation. Indeed, as with any other trail, good things and bad things may, and do, happen on the third path

depending on human choices and actions. As noted by Tim Dunne (with respect to the English School), theories located in the middle ground can be prone to misidentification and “open to predatory strategies of assimilation or rejection.”42 Similarly, critical and postmodernist scholars have good reasons to be nervous about rampant bridge building because, as Cynthia Weber justly notes, “often the bridge becomes the alien other's back.”43 Besides, difficult analytical choices will always await IR scholars of different theoretical persuasions willing to venture into the middle ground. For, as aptly put by John Ruggie, “even coalitions of the willing may find the going tough as they discover analytical limits beyond which their respective approaches cannot be pushed.”44 And those who naively, or strategically, decide to nonetheless push against such “limits” may end up producing meaningless or monstrous togetherness in which “several incompatible things are … being bolted together.”45 Be that as it may, I still believe that the facilitation of new and intellectually better-equipped “coalitions of the willing” should be prioritized at this stage of the disciplinary debate. Despite its numerous imperfections, the third path seems highly promising in this respect. Triumphant, unreflective, and/or hegemonic “middlegroundism”—under which all IR scholars would mindlessly line up and embrace each other on the third path—is undesirable and, of course, also impossible. As a pluralist, I would be first to concede that the third way is itself but one trail in the forest (albeit an important and, until recently, also a rarely traveled one). All considered, however, a reflective cost-benefit analysis still leaves us, I think, deep inside positive terrain. “Would it not be refreshing,” asks Donald Puchala, “if such continuing conversation, and not periodic great debates, becomes the intellectual mode of International Relations?”46 The third way in the third debate may help turn this vision into a reachable goal.

Notes 1. John R. Hall, Cultures of Inquiry (Cambridge: Cambridge University Press, 1999), 1. 2. Michael Burawoy, “A Sociology for the Second Great Transformation?” Annual Review of Sociology 26 (2000): 693.Page 256 → 3. Noretta Koertge, “Science, Values, and the Value of Science,” Philosophy of Science 67, no. 3 (2000): S45–S57; for an early articulation of a similar position in IR, see Yosef Lapid, “Quo Vadis International Relations? Further Reflections on the ‘Next Stage’ of International Theory,” Millennium: Journal of International Studies 18, no. 1 (1989): 84–85. 4. Yosef Lapid, “Sculpting the Academic Identity: Disciplinary Reflections at the Dawn of a New Millennium,” in Visions of International Relations: Assessing an Academic Field, ed. Donald Puchala (forthcoming). 5. Yosef Lapid, “The Third Debate,” International Studies Quarterly 33 (1989): 235–54. 6. Ole Waever, “Figures of International Thought: Introducing Persons instead of Paradigms,” in The Future of International Relations, ed. Iver B. Neumann and Ole Waever (New York: Routledge, 1997), 1–37. 7. Barry Buzan and Richard Little, “Why International Relations Has Failed as an Intellectual Project and What to Do about It?” (paper presented at the annual convention of the International Studies Association, Los Angeles, Calif., March 14–18, 2000); Ole Waever, “The Sociology of a Not So International Discipline, ” International Organization 52, no. 4 (1998): 687–727. 8. Ann Tickner, “You Just Don't Understand: Troubled Engagements between Feminists and IR Theorists,” International Studies Quarterly 41, no. 4 (1997): 611–32. 9. Samuel Makinda, “Reading and Writing International Relations,” Australian Journal of International Affairs 54, no. 3 (2000): 389. 10. Cynthia Weber, “IR: The Resurrection or New Frontiers of Incorporation,” European Journal of International Relations 5, no. 4 (1999): 435–50. 11. David Campbell, Writing Security, 2d ed. (Minneapolis: University of Minnesota Press, 1999), 215. 12. James Der Derian, “Virtuous War/Virtual Theory,” International Affairs 76, no. 4 (2000): 778.

13. David Bederman, “Constructivism, Positivism, and Empiricism in International Law,” Georgetown Law Journal 89, no. 2 (2001): 469. 14. Donald Puchala, “Visions of International Relations: Continuing the Dialogue among Scholars,” in Puchala, Visions of International Relations, 7. 15. Margaret Hermann, “One Field, Many Perspectives: Building the Foundation for Dialogue,” International Studies Quarterly 42 (1998): 605–24; Michael Brecher, “International Studies in the Twentieth Century and Beyond: Flawed Dichotomies, Synthesis, Cumulation,” International Studies Quarterly 43 (1999): 213–64. 16. Thomas Biersteker, “Eroding Boundaries, Contested Terrain,” International Studies Review 1, no. 1 (1999): 8. 17. Richard Bernstein, The New Constellation (Cambridge: MIT Press, 1992). 18. For a different opinion see Stephen Walt, “Fads, Fevers, and Firestorms,” Foreign Policy, no. 121 (November-December 2000): 37. 19. Francois Dosse, Empire of Meaning: The Humanization of the Social Sciences (Minneapolis: University of Minnesota Press, 1999), xix. 20. See Alexander Wendt, “On the Via Media: A Response to the Critics,” Review of International Studies 26 (2000): 165–80; Gillian Wylie, “International Relations' Via Media: Still under Construction,” International Studies Review 2, no. 3 (2001): 123–26; Georg Sorensen, “IR Theory after the Cold War,” Review of International Studies 24 (1998): 83–100; Marc Doucet, “Standing Nowhere: Navigating the Third Route on the Question of Foundation in International Theory,” Millennium: Journal of International Studies 28, no. 2 (1999): 289–310; Emanuel Adler, “Seizing the Middle Ground: Constructivism in World Politics,” European Journal of International Relations 3, no. 3 (1997): 319–63; Nicholas Onuf, “Constructivism: A User's Manual,” in International Relations in a Constructed World, ed. Vendulka Kubalkova, Nicholas Onuf, and Paul Kowert (London: Sharpe, 1998). 21. Erik Doxtader, “Characters in the Middle of Public Life: Consensus, Dissent, and Ethos,” Philosophy and Rhetoric 33, no. 4 (2000): 336–69. 22. Ibid., 338.Page 257 → 23. Ibid., 339. 24. Ibid., 346–47. 25. Ibid., 351. 26. Ibid., 362. 27. See in particular Sorensen, “IR Theory after the Cold War,” 85–92. 28. Wylie, “International Relations' Via Media,” 123–26. Exemplifying this problem, Wylie locates the main weakness of Wendt's otherwise impressive new book in “Wendt's determination to forge a via media through the ‘Third Debate,’” 124. 29. Wendt, “On the Via Media: A Response to the Critics,” 165–80. 30. Richard Little, “The English School's Contribution to the Study of International Relations,” European Journal of International Relations 6, no. 3 (2000): 395–422. 31. Bederman, “Constructivism, Positivism, and Empiricism in International Law,” 469. 32. Sorensen, “IR Theory after the Cold War,” 91. 33. Eviatar Zerubavel, “The Rigid, the Fuzzy, and the Flexible: Notes on the Mental Sculpting of Academic Identity,” Social Research 64, no. 4 (1995): 1093–98. 34. Ibid., 1097. 35. Waever, “Figures of International Thought,” in Neumann and Waever, The Future of International Relations, 25. 36. Richard Bernstein, “Metaphysics, Critique and Utopia,” Review of Metaphysics 42 (1988): 257. 37. Richard Bernstein, The New Constellation (Cambridge: MIT Press, 1992), 9. 38. James Bohman, “Pluralism, Indeterminacy, and the Social Sciences: Reply to Ingram and Meehan,” Human Studies 20, no. 4 (1997): 452. 39. Nicholas Burbules, “The Limits of Dialogue as a Critical Pedagogy,” in Revolutionary Pedagogies: Cultural Politics, Instituting Education, and the Discourse of Theory, ed. Peter Trifonas (New York: Routledge, 2000), 252. 40. Ibid., 260.

41. Mark Neufeld, “Reflexivity and International Relations Theory,” Millennium 22, no. 1 (1993): 53–76. 42. Tim Dunne, “Constructivism and the Altered Landscape of IR Theory” (paper presented at the annual meeting of the International Studies Association, Washington, D.C., February 1999), 10. 43. Cynthia Weber, “IR: The Resurrection or New Frontiers of Incorporation,” 446. 44. John Ruggie, “What Makes the World Hang Together? Neo-Utilitarianism and the Social Constructivist Challenge,” International Organization 52, no. 4 (1998): 883. 45. Friedrich Kratochwil, “Constructing a New Orthodoxy? Wendt's ‘Social Theory of International Politics’ and the Constructivist Challenge,” Millennium 29, no. 1 (2000): 73–101. 46. Donald Puchala, “Marking a Weberian Moment: Our Discipline Looks Ahead,” International Studies Perspectives 1, no. 2 (2000): 142.

Page 258 →

ALTERNATIVE, CRITICAL, POLITICAL R. B. J. Walker This paper offers a necessarily highly condensed argument about what it might mean to make judgments about the state of international relations or international studies as a scholarly discipline from some kind of alternative or critical perspective. I begin by sketching some tentative grounds on which I would presume to make judgments about judgments in this context. I then identify what I take to be four among many possible areas of convergence among the heterogeneous literatures that are currently identified, and disciplined, as alternative and critical. I then conclude by emphasizing my overall concern with practices of authorization, both in a specific discipline and in the practices that discipline seeks to examine, and try to end on an upbeat. I apologize in advance for the density and unqualified character of some of the formulations. I hope it will be clear that the whole business of offering judgments about judgments is a complex matter, one fraught with contradictions that are constitutive of what it is we think we are referring to when speaking of international relations/studies. I also hope it will be clear that the issues I try to emphasize here are important less in relation to the relatively trivial debates that occur in a specific academic site than to questions of authority that arise in relation to substantive claims provoked by analyses of the rapidly changing character of contemporary political life. My primary argument in brief, though I make no pretense to work through its underlying logic in any detail, is that the literatures that can be loosely even if controversially classified as alternative and critical are interesting primarily as attempts to politicize practices that have been profoundly depoliticized. This, I might add, is hardly a novel preoccupation. My secondary argument is that, read in this way, these literatures lead to engagements with the constitutive practices of a specifically modern politics, especially with the practices of sovereign authorization—again a familiar and longstanding concern. Linking these arguments is a claim that even though these literatures might be framed in relation to some kind of “outside,” whether as phenomena that are somehow (temporally) post- or as phenomena that are somehow (spatially) beyond (while also being constitutive Page 259 → of) the state, they express a renewed skepticism about a modern politics of autonomous subjectivities “inside.” What impresses me about much of the theoretical literature that has been portrayed in recent popular culture as radically different is less the claim that something has been left behind (modernity, structuralism, and so on) than the rather single-minded and often extremely rigorous engagement with what it could possibly mean to be radically different. In a formulation I have used too many times already, though one that I think remains especially important for thinking about the explicitly spatial preoccupations of international relations theory, one way of understanding many of the problems of contemporary politics is to think of the difficulties of imagining an outside to a politics that is already constituted as a structure of insides and outsides.1 This formulation can also be framed as the difficulty of imagining an alternative to a politics that already expresses a structure of the norm and the alternative (or the exception, or the state of nature, or the international anarchy, or the barbarian, or the oriental, or the romantic, or the utopian, to name only some of the best-known variations on this theme). It can also be framed, in a way that I currently find more interesting, as the difficulty of imagining forms of critique other than those depicted in the standard modern (Kantian) opposition between dogmatism and critique.2 Some might argue that there is no need to pose such problems. Many eminent people would agree with them. I wish them well. I also wish they would stop justifying themselves by making highly contentious claims about the real, the rational, the unchanging necessities of modern political life, or the natural goodness of modern liberalism: claims that often amount to much the same thing. I take it as fairly obvious that contemporary political life is changing more rapidly than our capacity to comprehend such change. This is especially so given that modern politics rests upon a specific understanding of the relationship between stasis (as norm) and change (as exception), or, more portentously/pretentiously, between space and time. The extraordinary explosion of diverse meanings awkwardly subsumed by the term globalization, or the barely articulated forms of dissent to currently hegemonic

governmentalities expressed since the Seattle trade talks, offers some indication of the limits of the modern political imagination in this respect. What we see in so many disciplinary and disciplining responses to the attempt to articulate alternative and critical approaches can be understood very largely as a repetition of familiar dogmatic assertions of the natural necessity of the categories of modern liberalism. These are categories that I, along with many of the canonical liberal texts, would insist are a site of very difficult problems rather than of principles that must be upheld so as to save us from anarchy and damnation, or worse, relativism. Many of these canonical texts have long insisted that liberalism must turn dogmatic at the limit. This, one might say, is the whole point of all those “political realisms” that now seem so embarrassing (and why they nevertheless remain much more interesting than their more stridently utilitarian successors). It is also why so many theorists of international relations constantly bowdlerize Page 260 → Hobbes and Kant (or refuse to look Max Weber or Carl Schmitt in the face, or look no further than John Rawls and the local economics department to construct their account of what liberalism involves) in order to sustain a systematic refusal to engage with the conditions under which the categories of modern liberalism could be naturalized. In this context, it is hardly surprising that attempts to repose questions about politics are so often met with reenactments of Hobbes's arbitrary assertion of the sovereign authority that guarantees legitimate knowledge. Still, there is considerable irony (and some crucial political practices to be analyzed) in the deployment of metaphors of a state of nature (accusations of relativism, immorality, absent foundations, and absent knowledge) in a discipline that has so often (and so unwisely) invoked this metaphor to describe its most essential problematic. More positively, what I find interesting about contemporary international relations/studies is that it is becoming a key site at which the necessity to reengage with questions about authority and the authorization of authority is becoming most obvious. I would readily admit that some people may not find this interesting at all. Such questions can be a little disruptive to scholars who are more or less content to accept prevailing ontological and ideological assumptions and just want to get on with the specific research that interests them. Nevertheless, the most important contestations in any discipline of knowledge concern the questions that are authorized to be interesting or even askable/answerable. This is why, I will suggest, it is so much easier to invoke the name of Hobbes (to take only the most obvious example) as a guide to the alternative and the critical than as an expression of conventions and necessities that must be obeyed. It is also important to insist at the outset that there is no field, subfield, or scholarly community that can be identified easily or uncontentiously as alternative or critical in relation to a practice called international studies /relations. In the first place, to identify an alternative is to assume a norm against which exceptions are to be identified. Here one might point, to take only three obvious lines of analysis, to (1) a hegemonic system of institutional practices (to be examined in terms of, say, the sociology of a profession, its recruitment practices, and so on); (2) a hegemonic discourse identifying the intellectual traditions through which scholarship should be constituted and judged (to be examined in terms of any of the now vast family of analyses that were once so easily referred to as ideology-critique); (3) a hegemonic account of the core problem or problems that enable prevailing intellectual traditions and institutional practices (to be examined especially in relation to claims about the canonical inheritance and central debates). For my present purposes, I assume that these three possibilities are identified in increasing order of importance. I recognize that what count as alternative or as critical approaches have been driven in very large part by the quite extraordinarily controlled, elitist, and nationalist character of the way in which “international relations” has been shaped as a subfield in the specific American discipline of “political science,” as well as by the more ideologically Page 261 → driven and shoddy formulation of many of the intellectual traditions that have become conventionalized in this context. Nevertheless, the appropriate tools for analyzing institutional and discursive hegemonies are fairly readily to hand even in the traditional crafts of the history of ideas and the sociology of knowledge. Consequently, I take it that it is in relation to the formulation and legitimation of core questions and debates that what have been framed as alternative approaches have been most sustained and serious. Insofar as one can judge the achievements and limits of these approaches, it is primarily in terms of their effect on what counts as an interesting and important scholarly agenda, and thus on the authorization of what counts as a legitimate

repertoire of scholarly questions. This is a theme that concerns not only this specific discipline. It is also central to the more general problems faced by attempts to develop forms of political pluralism that are sensitive to claims to “difference.” It is in this context especially that I would identify many of the key disagreements among those who think of themselves as somehow alternative in the specific context of international relations/studies. The key and in many ways quite obvious difficulty here is not the need to increase sensitivities to differences and pluralisms. It is the need to come to terms with the historically specific framing of the relationship between identity and difference, or universal and plural, that has set up modern political practice as a series of impossible but necessary antinomies, not least between the necessary and the possible, the norm and the exception, the natural and the critical. In the second place, to identify something as critical is to encounter a complex field of both popular and technical usages of this hopelessly overburdened term. In some of its popular senses, and in some of the literatures about international relations/studies, critique might encompass distaste, whining, and doctrinal disagreement. In most of its technical and scholarly senses it refers to some (of the many) versions of the Kantian distinction between the critical and the dogmatic. Complexities then arise not least because there are so many ways of insisting on the need to insist on a proper (critical) awareness of the conditions under which claims to knowledge are made (an insistence that applies to all forms of scholarly knowledge, whether empirical or hermeneutic as well as the supposedly alternative) and because of the ease with which claims to critical knowledge turn into forms of dogmatism (again whether empirical, hermeneutic, alternative, or critical). The Kantian distinction between the critical and the dogmatic remains an unavoidable point of orientation (even if one that is very difficult to sustain in recognizably Kantian terms). I would also claim that what is most interesting in those intellectual traditions that have been framed as radically different, especially as Foucauldian, Deleuzean, Derridian, and so on, remains tied to the possibility of sustaining and extending this distinction, though it does so by recognizing that the grounds for critique must be immanent rather than transcendent. Moreover, I hope that notice of this assumption is sufficient to ward off the abusive identification of the alternative and the critical with rabble-rousing accounts of some postmodernity, Page 262 → and to insist that indictments of some kind of relativism, nihilism, or Thrasymachian antiethics from some presumed position of rationality or modernity are merely expressions of problems that need to be addressed with great seriousness rather than a tactic about which anyone can be proud. To be clear here: my comments rest on the assumption that the distinction between the critical and the dogmatic remains crucial, even if difficult. It applies across all fields of knowledge. From this ground, most of the other distinctions and categories through which the possibility of critique has been raised in international relations /studies—the pretense to distinctions between first, second, third, and nth debates, between realists and liberals, between modernists and postmodernists, and between either rationalists or empiricists and either reflectivists or constructivists are perhaps the most perplexing, but there are many others—are almost entirely irrelevant. Most of these disciplinary artifacts constitute a philosophical swamp, a sign that a little knowledge is indeed a dangerous thing, especially in the hands of those who would be disciplinary king. They are a massive hindrance to constructive discussion. Quite apart from all the wasted energy required to negotiate this swamp, however, the pursuit of critical knowledge has always been a very tricky affair, and it has certainly not become any less tricky over the past few decades. Insofar as one can judge the achievements and limits of approaches that aspire to be critical in the specific context of a professional discipline, it is perhaps primarily in terms of their capacity to successfully insist on the very necessity of critique in the face of entrenched institutionalized conventions. But the challenge of critique in international studies is especially interesting because this possibility is necessarily raised in and through the various contemporary challenges to the subject of modernity—the (free, equal, autonomous, and so on) individual subject constituted as citizen in and by the (sovereign, national) state—that has for so long been both the ground and aspiration of the possibility of critique. The institutionalized practices of a discipline may be annoying, not to say career destroying, but as a site for engaging with the crucial questions of contemporary politics, “the international” poses enormous intellectual challenges for which even our most sophisticated traditions are clearly

wearing very thin. (This, I might add, is why forms of critical possibility associated with Habermas, cosmopolitanism, Critical Theory, and related attempts to privilege a Subject of History, seem to me to be expressions of rather than positive responses to the problem at hand.) To put this in a slightly different language, if international studies/relations is somehow concerned with “the state, ” then it also has to be concerned with the question of authority—that is, not only with the effects of the state as the most effective site of modern authority but with the practices of authorization that have enabled the state. Hobbes may be the patron saint of international relations, but only to the extent that, like Machiavelli, Kant, and all the other poor abused caricatures of the prevailing canon, he is understood to have been concerned precisely with the practices of authorization. Page 263 → Hobbes is especially important as a paradigmatic account of the practices of discrimination and authorization necessary for reading texts and making claims to scholarly knowledge quite as much as for the practices of states, legal regimes, and declarations of war. Hobbes's questions, I would say, like the questions of a Machiavelli, a Rousseau, a Kant, and many others, are much closer to the questions posed by various alternative and critical theorists than to most of what is claimed in their name in the standard disciplinary texts. This is not surprising in that they were asking about the very possibility of a modern politics. Their answers, of course, or the procedures through which they tried to work out their answers, are a different matter. Again, one might want to claim that their questions are uninteresting, or merely historical, or merely theoretical, or culturally parochial, or even that their answers are perfectly adequate. Such claims might be debated. But they can be debated usefully only while recognizing that it is difficult to find much of a theory of international relations in any of these canonical sources—accounts of the practices of authorization certainly, accounts of a (always highly problematic) relationship between a political community and an absence of political community also, but clear accounts of what international relations is really like, very rarely. That enterprise starts much later and has a very interesting history, a history that makes it into the standard texts in the most extraordinarily simple-minded form we all know and loathe. Every introductory student learns it. The categories are hammered into memory, then it's social science all the way. Good-bye questions about authorization, about the very possibility of a modern politics, let alone a politics that is not monopolized by the modern state. Hello the desperate need for critique. The key achievement of supposedly alternative and critical literatures over the past two decades has been to open up at least some possibility of asking questions about the location and character of the political. Or rather, this is the context in which I would try to make judgments about the contributions of alternative and critical literatures. Moreover, I would want to insist that areas of scholarship that are not formally identified as alternative and critical also have much to say. I would say this especially of some feminist political theory, some postcolonial theory, some political economy, some literary theory, some cultural theory, some legal theory, some ecological theory, some empirical theory, some rationalist theory, some institutional theory, and so on. In all cases, the emphasis is on the some: not all claims to critique are critical; not all claims to be alternative are alternative; not all claims made under the signs of empiricism, rationalism, and the rest are conventional or dogmatic. Difficult questions of judgment here are all too often erased by caricatured identities and a mania for easy classifications. Moreover, it is not as though alternative and critical theories of international relations/studies can be considered as closed communities, no matter how much people identified with these categories might hang around together in the annual meetings of the professional associations. I doubt very much that publications about either alternative /critical international relations/studies or publications about the professional discipline Page 264 → of international relations/studies in general account for more than a very small proportion of what any of us read in order to make sense of what is going on, or to keep up with the concepts and techniques necessary for either research or teaching. Indeed, I would consider it to be a sign of professional incompetence if this were the case. What I think ultimately links many of those who are located as alternative and critical in relation to international studies/relations, then, is a deep skepticism about the received account of where and what the political must be. This is an account that is notoriously expressed as a simple fact of life by those who claim to speak of a political

realism in international relations and naturalized even by many if not most attempts to challenge the presumptions of political realism. This skepticism, this concern with the contemporary fate of modern practices of authorization, provokes different scholars to move in many theoretical directions. The differences among these directions is perhaps much more interesting than what unites them as somehow alternative or critical, but for purposes of this discussion I will focus on some commonalities. There seems, to me at least, to be a shared concern with four core themes. First, the tendency to interrogate the received assumptions about what the world is taken to be—that is, to privilege the critical analysis of categories that express (and help constitute) assumptions about the ontology of politics. Second, the tendency to engage with sovereignty as an enormously complicated site of political practice, one that most accounts of modern politics pass over in silence. Third, the tendency to get very annoyed when told that difficult questions about ontology and sovereignty ought to be adjudicated on the ground of epistemology and method. And fourth, the insistence that there are interesting parallels (not to say ironies) between the constitution of international relations as a discipline concerned with the study of crisis and the constitution of accounts of what it means to have alternative or critical perspectives on that discipline. Of course it would come as no surprise to me to learn that others who have come to represent the alternative and the critical find that this way of framing a specific site of scholarly engagement completely misses the point of what they are up to. Moreover, each of these four themes can be made to seem highly arcane, trivial, philosophical, and so on; the rhetorical procedures are familiar. Nevertheless, these four themes seem to get at political practices that are of considerable practical importance. My own experience has been that it's much easier to talk about them with policymakers and activists than with theorists of international relations. Whether one takes this claim to practical relevance seriously or not, however, I would want to insist that it has to be couched not in terms of the familiar and profoundly ideological distinction between theory and practice, but of questions about the conditions under which claims about the arcane, the trivial, the theoretical, and the practical get to be authorized as authoritative. First, then, perhaps the most crucial constitutive move of modern Anglo-American social science has been to insist that all theoretical disputes be resolved on the ground of epistemology, a ground that is then usually narrowed Page 265 → either to the possibilities of empiricism, the techniques of method, or both. Not everyone, not even most (or perhaps even many) Anglo-American social scientists have been fooled by this move, but it has been especially influential in the political science subfield of international relations. There are two obvious countermoves to this insistence on the priority of epistemology and method. One is to note that the very possibility of modern epistemology has depended on a specifically modern ontology, classically a dualism of spatially differentiated subjects and objects. This is an observation that might at least serve to emphasize the highly contested character of contemporary accounts of legitimate knowledge rather than to legitimize claims about a store of scientific techniques. The problem of knowledge, one might say, is precisely a problem, and a very interesting and important one at that. The other countermove is to note that all the noise about epistemology and method has largely served to mask a sustained silence about the ontological categories and assumptions that are thereby reified. It is in both contexts that we can understand much of the insistence in various critical and alternative literatures on and importance of examining, and largely rejecting, assumptions that are taken for granted in many claims about the priority of epistemological adjudication. Sometimes this rejection is made on grounds of philosophical incoherence and sometimes on grounds that these assumptions are politically interested/suspect; the possibilities are obviously extensive. Thus, if there were an opportunity to survey these literatures in detail, I would especially stress the implications of the following: Critiques of various practices of reification (in the senses identified by both Georg Lukacs3—as commodification—and Alfred North Whitehead4—as the fallacy of misplaced concreteness), especially in relation to claims about state, nation, history, structure, cultural identity, gender, power, rationality, and the necessary virtue of the standard liberal antinomies (as assumed not least in the familiar “levels of analysis”

schema). Hence the “post-” in poststructuralism, the methods of genealogy, the injunctions to privilege verbs and actions rather than nouns and things, and so on. Refusals of Cartesian and Lockean accounts of the relation between language and world (and thus a concern with the politics of representation and questions of authorization rather than with, say, the “influence” of “ideas” or “culture”). Explorations of multiple subjectivities in relation to claims about cultures, genders, political communities, citizenships, rights, ecologies, and so on. Explorations of spatiotemporal configurations other than those identified as national histories contained in horizontal territories (and thus of “speed,” flows, lines of flight, rhizomatics, networks, horizons, global /locals, and so on). To privilege questions about ontology is, of course, to question the most basic constitutive categories of the discipline. On the one hand it is hard to Page 266 → see how this is anything to which anyone could object, given the degree to which these categories have been challenged by just about everyone else. On the other, it is easy to see how these questions are unlikely to find favor in disciplines that depend on these categories for their very legitimation. Second, to focus on questions of ontology is especially to focus attention on the principle that is claimed to define the phenomena that is international studies/relations studies and yet that is simply taken for granted by those studies. This is the principle (and practice, and institution, and problem) of sovereignty, a principle that, not least, offers our paradigmatic account of how questions about ontology and questions about epistemology should be related. International relations theory typically takes sovereignty to be an expression of (the already achieved condition of) a plurality of communities, communities that are always potentially in conflict because of the absence of any common community. It is only rarely taken to be a practice through which the relation between plurality and commonality is produced and reproduced. (This, I have argued elsewhere is the crucial cause and effect of the discourses through which “international relations” has been distinguished from “political theory.”)5 Sovereignty then ceases to be a site of contradictions and paradoxes—of, to invoke the terms of a variety of analytical traditions, the paradox of founding, of the groundless authorization of a ground of authority, of the ontotheological production of ontology, and of a decision to include/exclude that defines what can then be included/excluded. It is no longer understood to be the site of a politics that constitutes a politics (Hobbes, standing on the shoulders of Plato and Aristotle) or a site at which exceptions are made to the norm (Schmitt, the distiller of epigrams from logics worked out by Hobbes, Kant, and Weber), or a site at which the sign of rationality asserts itself on a ground of charade, rhetoric, irrationality, and the yawning gap between word and thing (the standard puzzle of nominalism bereft of transcendental guarantees). Sovereignty becomes merely an achieved condition, though we might have some interesting and important debates about exactly when it was achieved, if ever. Indeed, one might, and I would, want to insist that sovereignty can never be an achieved condition, in principle; careful readings of Hobbes are especially helpful in this context. There is, though, a paradox in the way the discipline of international relations has so easily assumed sovereignty as a foundation (been mesmerized by the top turtle, perhaps) and then constituted itself as the study of a world of multiple sovereignties. The rise of various alternative and critical modes of analysis can be explained in large part as a consequence of conditions under which it is no longer sensible to allow this paradox to work as if it is the most natural thing in the world. Hobbes, of course, hoped that his account of the necessity of sovereign necessity would seem to be the most natural thing in the world, indeed insisted that it must be so if we moderns are to forestall the necessary consequence of liberal individualism: relativism. Kant is not so very different in this respect. If Hobbes and Kant are still to be used as canonical figures Page 267 → in thinking about international relations/studies, then at least they ought to be read seriously as sites in which the problem of sovereignty is explored rather than as the dimwitted utilitarians and protodemocrats that pop up in so many texts in this field. It may be the case that much of the alternative and the critical is articulated with the help of writers who seem excessively post- or excessively

Parisian, or excessively literary, or excessively philosophical. Still, it can hardly be said that the ways in which they engage precisely with the sovereign practice of authorization has only marginal relevance to international relations/studies. In any case, excessive reification is arguably a far, far worse problem, one that is far more common, and, of course, one that is subject to a multitude of rhetorical tactics intended to make it seem utterly benign. Third, questions about how we know what we claim to know are dependent on what we assume we are trying to know. If one starts from an Augustinian distinction between necessity and free will, for example, one will end up doing the familiar dance of structure and action. One might prefer the tango—or at least try to come to terms with how modern accounts of sovereign authority constitute a modern account of freedom as necessity, whether under the law of the state or the law of universal morality. We are free to act, as we have been told time and time again, as long as we do what we have to do. This is a contradiction that plays rather a large role in the constitution of most of our modern political categories, not least those that are used in claims about rational action. It is not one that is likely to be resolved by epistemology, though it is one that we might begin to see as very poorly framed in the first place. Epistemologies have their place, but they are the product of ontological contestation, not referees; they are authorized grounds for contestation in which referees can be deployed, not authorizations of those authorized grounds. Epistemology remains crucially important. Many of the characteristic concerns of alternative and critical theories are not directed against the importance of epistemological matters, indeed quite the contrary, but against the prevalence of caricatured accounts of what it means to make knowledge claims about the world. These caricatures, moreover, have now become almost inseparable from a utilitarian or instrumental account of rational action that serves as an obvious ideology of more or less identifiable political interests. Furthermore, much of what actually goes on under the rubric of epistemology turns out to be a more or less elaborate production of classifications, of inclusions and exclusions, about which a great deal might and should be said from a critical perspective. In any case, demands that critical and alternative scholarship submit to the arbitrary and long-contested accounts of what a proper social science should look like are unpersuasive, to say the least. However, demands that all claims to knowledge be susceptible to critical analyses of their conditions of possibility remain crucial. Critique trumps social science, not the other way around. Or rather, there is no social or any other kind of science without critique. Contestations arise in this context because there are so many claimants to the status of critique—and hence many opportunities for epistemological Page 268 → debate—but also so many attempts to claim this status by fiat, and hence so many tactics for the practice of discipline building. Fourth, there are grounds to be skeptical about many scholarly claims produced by the discipline of international relations. These grounds are not sufficient to undermine its importance as a characteristic discourse of the modern state, whether as, for example, an expression of its constitutive outside, as the negation of the modern claim to political identity and community that serves to affirm the claims to identity and community that is being negated, as a legitimation of specifically modern practices of nationalism, as a sequence of claims about where and what modern politics must be, or as a set of profoundly normative codes framed as the most realistic of natural necessities. At least, these are the characteristics of a contemporary site of political practice that have interested me. The practices of authorization in this discipline reproduce practices of authorization in the phenomena it seeks to study. Forms of critique and claims about alternatives in the discipline express wider doubts about modern authorizations of authority. This is why I think there is something quite interesting about the ways in which a discipline that has been framed more than any other as being about crisis—about war and necessary violence at the point at which authority disappears—should be said to be somehow in crisis. It is perhaps telling that we are now so willing to speak about policy, about governance, about governmentalities, about regimes, about global civil societies, about global social movements, about nongovernmental organizations and international nongovernmental organizations, and perhaps now especially about ethics, yet we remain so reluctant to speak about politics. It is also perhaps telling that so many of the practices of the discipline still

largely depend on the framing of the key debates on precisely the dyadic possibilities that both enabled and reproduced modern accounts of sovereign necessity/freedom. This is apparent not least in the recent constitution of various strands of social constructivism, a term that makes little sense to me unless taken as some vintage opposition to a claim to nature, to essence, or a methodological injunction to reify; in this context, surely the key issue is not whether constructivism (see, for start, Hobbes, Locke, Rousseau, Smith, Marx, Mill, Weber, and the rest) but what kind, and especially what ontologies are being described as an act of construction and why we should refer to this act as social rather than political. Modern politics, of course, hinges on a capacity to draw the line—between humans and citizens, legal and illegal, norm and exception, public and private, political and nonpolitical, political and economic, political and ethical, political and cultural, and not least, the domestic and the international. It is thus of considerable interest to see how this discipline, which has been framed especially as a concern with life on the edge, with the state of emergency, with the inscription and transgression of boundaries, continues to enact lines of sovereign jurisdiction and legitimate authority. My own view is that the extensive metaphorical/discursive field through which lines drawn in (Euclidean) space have come to be read as lines between the legitimate and the illegitimate, or the included and the excluded, Page 269 → offers an increasingly (but always) difficult ground on which to understand contemporary accounts of both subjectivity and authority. It is not a ground on which I would try to fix an analysis of contemporary political life. The modern art of discrimination, in both the positive and negative sense of this term, is in trouble. At this historical juncture, it may make very little sense to speak of critical and alternative perspectives in international studies/relations except as the familiar Other that serves to legitimize some prevailing norm. The tight corsets of the cold war years have clearly come unstuck, but hardly as a result of a handful of scholars who had read some political, social, and literary theory. Intellectual horizons may have been reshaped and broadened, some possibilities explored, many new connections made across disciplines and institutions, but these depended on broader historical changes that made the existing order unsustainable in any case. In some ways, the relatively open and cosmopolitan character of contemporary international studies is quite remarkable, especially given the state of the discipline thirty years ago. This openness may still have fairly obvious limits, which become even more obvious as one moves toward the key institutions of intellectual hegemony. Still, it does seem to me that questions about authorization, about the political practices through which authorities are constituted and reconstituted, can now be asked in more and more interesting ways. Such questioning is, of course, the very opposite of the claim to political realism as we have come to know it in this context, although it is far from unknown in some of the key sources of this claim, among which I would especially single out Max Weber. It is also precisely what would interest a Machiavelli, a Hobbes, or a Kant, as well as the Foucauldians, Derridians, and Deleuzeans, even if the contemporary plague of Benthamites might wish us to believe—and I insist on the word believe—otherwise. But let me try to be clear here also: to insist on the necessity of critique, something that seems to me to be rather conventional (which is why I read Foucault, Derrida, Deleuze, and the rest as people engaged with rather conventional questions and problems, even if in provocative and I think quite productive ways), is also to insist that the very meaning of critique, and of what it might mean to engage with alternatives, is highly contentious. The standard Hobbesian, Kantian, and Schmittean authorizations of authority are increasingly difficult to sustain, as are all the other less stringent liberal attempts to draw the line that ultimately rests upon these more dramatic theorists of the limit condition. The meaning of critique is contentious because modern forms of authorization are in widespread dispute. The authorizations of this specific discipline are interesting because the realm of international relations/studies is one in which difficulties of authorization are so obviously acute. Hence, presumably, all those synonyms that try to elude references to politics. Hence also the difficulty of coming to terms with contemporary authorizations enacted in the World Trade Organization, the European Union, the United Nations, the conventions on this and that, and so on. And hence, consequently, the good old-fashioned wisdom of looking for politics in places where it is not supposed to be. Page 270 →

In the end, then, it is this fourth set of convergences that seems to me to be the most significant. The great paradox of modern international relations/studies is that its disciplinary conventions have for so long discouraged any concern with the practices of authorization, and yet the problem of authorization (the Hobbesian problem of sovereignty rather than the supposedly bare Bodinian or Westphalian fact of sovereignty if one prefers, and thus all those myriad practices through which sovereign authorizations are authorized and reauthorized) is perhaps nowhere as interesting as in the phenomena these conventions seek to examine. In the end, the key feature of international relations theory, as of its enabling twin, modern political theory, has been an almost complete refusal to take the problem of sovereignty seriously (and I might add that modern political theorists have generally been even more irresponsible than modern theorists of international relations in this respect). I at least take it as a sign of some progress that this refusal, while still popular, is increasingly treated as a sign of sloppy scholarship, as indeed I would insist it is, and on the most conventional of grounds. Contrary to various attempts to link critique, alternatives, emancipations, and so on to the discovery of firm foundations in ethics, economics-in-the-last-instance, or some such extrapolitical realm, it is important to stress that these grounds are anything but certain. These are, after all, the grounds on which modern politics, and thereby modern international relations, was initially constituted against theocracy and empire, even if this was on terms always susceptible to a reversion to theological models of authorization in the horizontal spaces of modern states rather than the vertical space of empire. But then, perhaps we really are now being asked to conform to the necessities of empire. Perhaps questions about authorization have become redundant. Perhaps a further deification of the market and its rational actors is all we can look forward to. If not, there is certainly a lot of work to be done.

Notes 1. This is the puzzle I seek to examine in my International Relations as Political Theory (Cambridge: Cambridge University Press, 1993). 2. For an attempt to explore what this might involve, see my After the Globe/Before the World (Cambridge: Cambridge University Press, forthcoming). 3. Georg Lukacs, History and Class Consciousness (1923; London: Merlin, 1971). 4. Alfred North Whitehead, Science and the Modern World (New York: Macmillan, 1925). 5. R. B. J. Walker, Inside/Outside: International Relations as Political Theory (Cambridge: Cambridge University Press, 1993).

Page 271 →

THE GLOBALIZATION OF GLOBALIZATION James N. Rosenau As one whose professional life exceeds that of the International Studies Association, I have mixed feelings about the invitation to reflect critically on the accomplishments, failures, debates, standards, approaches, and future agendas of international relations (IR) and its subfields. On the one hand, I have no doubt that our shared past is marked by enormous growth and progress. The conceptual and methodological equipment with which IR is probed today is far more elaborate, incisive, and diverse than was the case at the outset—back in the 1950s when a few isolated, non-Ivy League scholars first came together around common interests to form a professional association, replete with a journal (called Background before being changed to the International Studies Quarterly) and with a membership so small that it convened annually on campuses because it was unable to reach the minimum registration required for reduced rates at hotels. On the other hand, for all the progress that has marked IR's evolution, I am plagued with doubts about the field's capacity for adapting to the transformations at work in the world today. And my doubts extend to questioning the wisdom of engaging in introspective assessments of inter-and intraparadigmatic debates—perhaps out of fatigue generated by a history of vigorous involvement in such debates and perhaps because I have progressively moved away from the IR mainstream, but more likely out of a conviction that change has left us so far in arrears that we need to focus our energies on assessing substantive dynamics rather than evaluating our colleagues, their theories, and their standards. Several years ago I gave up wondering, worrying, and writing about the fit between my thinking and that of realists and liberals. To articulate such concerns struck me as needless ritual that diverted valuable time and space from focusing on what was transpiring outside academe in what students call “the real world.” The problem was me and my understanding, which required an agenda framed by my own reasoning and not one devised Page 272 → by others. Put differently, exercises like the present one can be a substitute for thinking about the nature of the world and how it should be studied, and I think that too often we resort to this substitute when the course of events seems too complex and obscure to comprehend. So the ensuing paragraphs stray beyond the assigned topics. They are rooted in a sense of déjà vu all over again, to quote an insightful philosopher. In 1968 I wrote that all the signs were pointing in the same direction, that a spate of articles, textbooks, conferences, and curricula signified the emergence of a subfield (comparative foreign policy),1 and now an even greater variety and number of indicators are heralding the arrival of another new preoccupation, one that is so pervasive that it has not only spread quickly through our ivory towers, but it has also galvanized action in the world's streets, markets, foreign offices, boardrooms, legislative chambers, city halls, cyberspaces, and every other site where people converge. And its pervasiveness in academe is such that it may emerge as our field and not just as a subfield. On Tuesday, January 19, 2000, while preparing a syllabus for a new course on globalization, three insightful newspaper stories, two soon-to-be-published manuscripts, a large book of readings, invitations to two conferences (one in Washington and one in Israel), and the lead article in the December issue of Foreign Affairs came my way—all of them concentrating on one or another aspect of globalization. That was just one day's take. Previously the flow of such indicators was no less continuous or voluminous, with word of new courses, conferences, dissertations, books, and papers on the subject amounting to a flow of tidal proportions. In effect, the preoccupation with the dynamics of globalization, both good and bad, has undergone globalization. But there is a difference between the 1968 article and the recent déjà vu experience. The former involved a flow located entirely in the IR academic community, whereas the second is predominantly inundating other communities both outside and within the academy. The outside flow is evident in the huge extent to which globalization encompasses a cluster of issues at the top of the agenda of communities at every level of aggregation. The inside preoccupation with such issues is sustained by scholars in anthropology, sociology, economics, geography, and social psychology, with political science and IR trailing way behind. More accurately, within the American IR community, globalization is not a primary focus of most inquiries. The articles, books, and conference proposals I received January 19 were the work either of colleagues abroad or of journalists,

defense officials, and economists. It is almost as if our political science and IR communities are shying away from the forces of globalization because their complexity does not fit readily within the extant frameworks. These forces are so pervaded by nonlinear feedback mechanisms, by confounding nuances, that confronting them runs the risk of finding that one's hard-won intellectual perspectives are inadequate to the explanatory tasks and may have to be abandoned, or at least greatly modified. If that seems like sour grapes and a violation of my commitment not to worry or write about the approach of others, let me articulate my out-of-the-mainstream perspectives not in terms of others having gone astray but Page 273 → in terms of four major dimensions of globalization that strike me as urgently needing to be addressed. There are, of course, a number of other dimensions that should command focused attention, but these four can be seen as prerequisites to focusing on the others.

Change If one assumes, as I do, that the world, its societies and its people, is undergoing transformations so profound as not to be fully appreciated, then a major conceptual challenge needs to be faced: how do we know change when we see it? How do we differentiate between evolutionary and breakpoint change? At what levels of aggregation are deep and enduring changes most likely? Do changes at, say, micro levels necessitate comparable changes at macro levels, and vice versa? Are some forms of change illusory, amounting to no more than brief disruptions of underlying patterns? With few exceptions, such questions have not been the focus of conceptual inquiries by IR scholars.2 Most of us tend to take for granted that salient changes in the actors and structures of world affairs will be manifest as they unfold. Sure, when a regime collapses or an alliance breaks up, when terrorists demolish skyscrapers, when markets decline precipitously, or when situations deteriorate abruptly, we have little difficulty discerning the end of one historical sequence and the onset of another. Ascribing change to such developments is easy, but assessing the durability of the changes, or discerning the early indicators of regime, alliance, terroristinduced, situational, or market collapses, is where our conceptual equipment is rudimentary, if not altogether lacking. This is why all too often we are surprised by the turn of events. There are, of course, no magic formulas for understanding and anticipating different forms of change. Still, there are ways of maximizing our ability to assess when transformations may ensue. One is to assume that systems are always on the edge of collapse, an assumption that compels us to be sensitive to, even in awe of, the capacity of systems to get from one moment, week, year, or decade in time to the next. To proceed from the opposite assumption—that systems are likely to persist—is to limit our readiness to recognize the formation and early stages of change dynamics. Another way of coping with the challenge of change is to allow our variables to vary—that is, to mentally imagine a wide range of possible shifts in the values of all the variables relevant to our concerns. Most of us, for example, did not allow for the possibility that the cold war and the Soviet Union would come to abrupt ends. In retrospect, such failures border on the inexcusable. Or at least if we had been more sensitive to the susceptibility of systemic collapse and thus been alert to variations in the two structures expressive of such tendencies, it may have been less surprising.

Fragmegration Of the numerous change dynamics presently shaping world affairs, two clusters stand out as paramount—those that foster globalization, centralization, Page 274 → and integration on the one hand and those that promote localization, decentralization, and fragmentation on the other. While these polarities move the course of events in opposite directions, they are continuously, simultaneously, and often causally interactive, giving rise to the prime tensions with which individuals and their collectivities must contend. To capture the inextricable and pervasive character of these interactions, I use the concept of “fragmegration,” which is admittedly awkward and grating but at the same time reminds us that the processes of fragmentation and integration are interactive and, often, causally linked.3 Indeed, it can readily be argued that the emergent epoch is one of fragmegration and not simply one of globalization. The so-called battle of Seattle in late November 1999 offers a quintessential illustration of fragmegrative dynamics: as the representatives of states gathered for the integrative purpose of negotiating new trade agreements, so did various nongovernmental organizations (NGOs) and individuals take to the streets to

demonstrate their opposition and to highlight the fragmenting nature of such agreements. And the terrorist actions against the United States might well be described as the “mother” of all fragmegrative situations: they pulled the country together even as they destroyed one of its major landmarks. While anthropologists, sociologists, journalists, and business executives have recognized the importance of fragmegrative processes4 and even coined their own labels for them,5 these crucial dynamics have not been accorded centrality by the IR community. To the extent that globalizing or localizing forces are examined, they tend to be explored and traced separately, with only passing attention being paid to the ways in which each set of forces affects the other. One reason for this conceptual gap probably concerns the simultaneity of fragmegrative dynamics. The processes whereby the several polarities are linked to each other are comprised largely of feedback mechanisms, of nonlinear sequences that present enormous methodological dilemmas (see last section of essay). Most of us are accustomed to linear analysis, to discerning how dependent variables vary in response to the operation of various independent variables. The idea that each dependent variable becomes instantaneously an independent variable tends to be too mind-boggling to acknowledge, much less serve as the basis for inquiry. So fragmegrative studies languish for want of an effective methodology. Equally important, the conceptual gap is bypassed because of the number and variety of sources that contribute to and sustain the processes of fragmegration. One of these sources consists of what I have elsewhere labeled “the skill revolution” wherein people everywhere are increasingly able to construct scenarios that trace the course of distant events back into their homes and pocketbooks.6 A second source involves the large degree to which collectivities around the world are undergoing authority crises, by which is meant the paralysis and stalemates that prevent them from framing and moving toward their goals. A third focuses on the bifurcation of global structures whereby the long-standing statecentric world now has a rival in an emergent multicentric world of diverse actors such as ethnic minorities, NGOs, professional societies, transnational corporations, and the many Page 275 → other types of private collectivities that now crowd the global stage. A fourth is what I call the “organizational explosion” that has witnessed a huge proliferation of associations and networks at every level of community. A fifth I call the “mobility upheaval,” by which is meant the vast and ever-growing movement of people around the world, a movement that includes everyone from the tourist to the terrorist and from the jet-setter to the immigrant. A sixth consists of the many microelectronic and transportational technologies that have collapsed time and space. A seventh involves the complex processes through which territoriality, states, and sovereignty have weakened to the point where it can be reasonably asserted that landscapes have been supplemented—and in some cases replaced—by mediascapes, financescapes, technoscapes, ethnoscapes, and ideoscapes.7 An eighth concerns the large degree to which national economies have been globalized. How these eight major sources (and doubtless others could be identified) interactively generate and sustain the dynamics of fragmegration is an enormous analytic challenge, but even prior to assessing their interaction is the need to probe each of them for their fragmegrative content and consequences.8

Micro-Macro Links Still another critical aspect of fragmegrative dynamics involves the links between individuals at the micro level and collectivities at the macro level. Like change and fragmegration, these links have not been the focus of extensive conceptualization and investigation. While few analysts would deny that the flow between the two levels is central to how collectivities come into being and sustain themselves through time, how their macromacro relationships are configured, and how people are shaped by the collectivities to which they belong, the interaction across these levels has been largely taken for granted or, in one well-known case, assessed to be beyond systematic comprehension.9 We simply do not have any viable theory that anticipates how individuals will vary in response to varying macroinputs or how the structures and policies of macrocollectivities might be undermined, redirected, sustained, or otherwise affected by new patterns at the micro level. Again, this is a preoccupation in some of the social sciences,10 but IR scholars have essentially ignored the puzzles posed by the links among the levels of aggregation. Indeed, a major paradigm in the field, realism, proceeds from the premise that the only relevant action is that of states at the macro level, that individuals at the micro level can be assumed to follow the lead of their states.

The reasons for this seeming obliviousness to micro-macro links are not difficult to identify. Tracing such links is extremely difficult theoretically and thus even more challenging empirically. Those of us who do not subscribe to realist formulations intuitively know that the links are endlessly operative, that what collectivities and individuals do on the global stage is in part reactions to each other, but faced with the task of tracing their interactions, we tend to find it easier to take them for granted than to wrestle with the puzzles they pose. I do not claim to have made any progress in Page 276 → solving the puzzles, but I do contend that comprehension of world affairs increasingly requires us to address the challenge and frame models that offer a chance to fit some of the pieces of the puzzles together.11 As the skill revolution, authority crises, structural bifurcation, the organization explosion, the mobility upheaval, microelectronic technologies, the weakening of territoriality, states, and sovereignty, and the globalization of national economies accelerate and extend their impact on fragmegrative dynamics, so does it become all the more urgent that we collectively work to confront the causal interactions thereby established. Table 1 is an effort to highlight the vast domain across which such theorizing must roam. The entries in its cells are no more than impressionistic hypotheses about how the sources of fragmegration may play out at the micro, macro, macro-macro, and micro-macro levels, but hopefully they are suggestive of some of the paths inquiry into such relationships should follow.

Methodology The conceptual challenges posed by the dynamics of change, fragmegration, and micro-macro interactions share a methodological dilemma. Each is pervaded with feedback processes, and thus they pose the difficult question of how to explore them systematically. In other words, each is rooted in events that unfold virtually simultaneously, making it fruitless to cast analyses in terms of the interaction of independent and dependent variables. More accurately, analyses cast in a conventional linear framework seem bound to fall short in terms of capturing the interactive, high-speed nature of the phenomena of interest. How, then, to proceed? What methodologies might be available for probing nonlinear sequences of interaction? Are we bound to rely exclusively on case studies that, hopefully, somehow reveal the underlying tendencies that drive the transformative impacts of fragmegration and micro-macro links? Can nonlinear methodologies drawn from mathematics and statistics be adapted to the needs of the IR student? If we are not ourselves adept at such methods, how do we make an effective case for students employing them? Never having been very sophisticated as a methodologist, I do not have very precise answers to these questions. I know they are important, and I presume there are colleagues who can answer them with some authority. Still, ignorance is no excuse. If the mysteries of a fragmegrative world are to be fathomed, we cannot shy away from the methodological questions on the grounds of inexperience. Herewith, then, a partial, somewhat informed response to the methodological challenge: while case studies can surely be of value, there are also nonlinear procedures that have become more feasible as a result of advances in computer technologies. As I understand it, there is now the prospect of computer chips that will be ten billion (yes, ten billion) times faster than those in use today.12 This heightens the potential of using computer simulations based on complexity theory, of building nonlinear feedback mechanisms into models that simulate the dynamics of change, fragmegration, and micromacro interactions. To recur to points Page 277 → made earlier, conceptualizations of these three sets of dynamics are in short supply and obviously need to be refined before computer models can be applied. If it is assumed that such refinements will eventually be developed, however, then computer simulations may prove to be a useful methodology in unraveling the mysteries of a fragmegrative world. As for the problem of motivating and equipping students to take advantage of the technological advances that lie ahead, let me report on a teaching aid that I have found highly effective semester after semester for the last several years. It is a book entitled Complexity: The Emerging Science at the Edge of Order and Chaos, by M. Mitchell Waldrop.13 In fifty-one years of teaching I have never given an assignment that has had such pervasive consequences. Each semester the first week's assignment is to read the entire book and write a five-page evaluation of it. The reactions are consistently impressive. Even though the book refers only occasionally and peripherally to the relevance of complexity theory for world affairs, and even though much of the book is about matters far removed from the concerns of social science—it is the story of the Santa Fe Institute in New Mexico—the students find the outlines of complexity theory eye-opening and virtually every week of the semester

one or another student mentions the book and voices an idea he or she picked up from it. No book I have ever assigned lingers so persistently in the memory banks of students. In the last few years several students have rethought their study plans in the direction of tooling up in computer science. In sum, I find myself persuaded that the future of our field, its capacity to confront the huge conceptual and methodological challenges that have eluded our generation, lies in the training of those who will enter the field in the future. We need to acknowledge our own limitations and alert those we train to the necessity of their breaking with past paradigmatic assumptions and finding new ways of understanding and probing the enormous challenges posed by the dynamics of change, fragmegration, and micro-macro interactions. If we can orient our students along these lines, the globalization of globalization will have moved on to a higher and more secure analytic plane.

Notes This paper was prepared for presentation at a millennial reflections panel at the annual meeting of the International Studies Association, Los Angeles, Calif., March 16, 2000. 1. James N. Rosenau, “Comparative Foreign Policy: Fad, Fantasy, or Field?” International Studies Quarterly 12 (September 1968): 296–329. 2. Among the exceptions are Barry Buzan and R. J. Barry Jones, eds., Change and the Study of International Relations: The Evaded Dimension (London: Pinter, 1981); K. R. Dark, The Waves of Time: Long-Term Change and International Relations (London: Pinter, 1998); Joshua Goldstein, Long Cycles: Prosperity and War in the Modern Age (New Haven, Conn.: Yale University Press, 1988); K. J. Holsti, Change in the International System: Essays on the Theory and Practice of International Relations (Brookfield, Vt.: Elgar, 1991); and George Modelski, Long Cycles in World Politics (New York: Macmillan, 1987). 3. The fragmegration concept was first developed in James N. Rosenau, “‘Fragmegrative’ Challenges to National Security,” in Understanding U.S. Strategy: A Reader, ed. Terry Heyns (Washington, D.C.: National Defense University, 1983), 65–82. For a subsequent and more elaborate formulation, see James N. Rosenau, Along the Domestic-Foreign Frontier: Exploring Governance in a Turbulent World (Cambridge: Cambridge University Press, 1997), chap. 6. Page 278 → Page 279 → Page 280 → 4. See, for example, Stuart Hall, “Old and New Identities, Old and New Ethnicities,” in Culture, Globalization, and the World-System: Contemporary Conditions for the Representation of Identity, ed. Anthony D. King (Minneapolis: University of Minnesota Press, 1997); Ulf Hannerz, Transnational Connections: Culture, People, Places (London: Routledge, 1996); Arjun Appadurai, Modernity at Large: Cultural Dimensions of Globalization (Minneapolis: University of Minnesota Press, 1996); and Kevin R. Cox, ed., Spaces of Globalization: Reasserting the Power of the Local (New York: Guilford Press, 1997). For fragmegrative analyses by business executives and journalists, see Dee Hock, Birth of the Chaordic Age (San Francisco: Berrett-Koehler, 1999); and Thomas L. Friedman, The Lexus and the Olive Tree (New York: Farrar, Straus and Giroux, 1999). 5. Among the single-word labels designed to suggest the contradictory tensions that pull systems toward both coherence and collapse are “chaord,” a label that juxtaposes the dynamics of chaos and order; “glocalization,” which points to the simultaneity of globalizing and localizing dynamics; and “regcal,” a term designed to focus attention on the links between regional and local phenomena. The chaord designation is proposed in Hock, Birth of the Chaordic Age; the glocalization concept is elaborately developed in Roland Robertson, “Glocalization: Time-Space and Homogeneity-Heterogeneity,” in Global Modernities, ed. Mike Featherstone, Scott Lash, and Roland Robertson (Thousand Oaks, Calif.: Sage, 1995), 25–44; and the regcal formulation can be found in Susan H. C. Tai and Y. H. Wong, “Advertising Decision Making in Asia: ‘Glocal’ versus ‘Regcal’ Approach,” Journal of Managerial Issues 10 (fall 1998):

318–39. I prefer the term “fragmegration” because it does not imply a territorial scale and broadens the focus to include tensions at work in organizations as well as those that pervade communities. 6. For a full discussion of the skill revolution as a microdynamic, see James N. Rosenau, Turbulence in World Politics: A Theory of Change and Continuity (Princeton, N.J.: Princeton University Press, 1990), passim. For a test of the skill revolution as an empirical hypothesis, see James N. Rosenau and W. Michael Fagen, “Increasingly Skillful Citizens: A New Dynamism in World Politics?” International Studies Quarterly 41 (December 1997): 655–86. 7. These “scapes” are discussed in Arjun Appadurai, Modernity at Large: Cultural Dimensions of Globalization (Minneapolis: University of Minnesota Press, 1996), 33–36. 8. For a brief discussion of the eight sources of fragmegration, see James N. Rosenau, “Emergent Spaces, New Places, and Old Faces: Proliferating Identities in a Globalizing World,” in Worlds on the Move: Globalization, Migration, and Cultural Scarcity, ed. Jonathan Friedman and Shalini Randeria (London: I. B. Tauris, forthcoming). 9. J. David Singer, “The Levels-of-Analysis Problem in International Relations,” World Politics 14 (October 1961): 77–92. 10. See, for example, J. C. Alexander et al., eds., The Micro-Macro Link (Berkeley: University of California Press, 1987); and K. D. Knorr-Cetina and A. V. Cicourel, eds., Advances in Social Theory and Methodology: Toward an Integration of Micro- and Macro-Sociologies (Boston: Routledge and Kegan Paul, 1981). 11. Recently I have undertaken to address the challenge in James N. Rosenau, Distant Proximities: Dynamics beyond Globalization (Princeton, N.J.: Princeton University Press, forthcoming). 12. John Markoff, “Tiniest Circuits Hold Prospect of Explosive Computer Speeds,” New York Times, July 16, 1999, A1. 13. Torchbooks (Simon and Schuster), New York 1992.

Page 281 →

FEMINIST THEORY AND GENDER PERSPECTIVES

Page 282 → Page 283 →

THE FISH AND THE TURTLE Multiple Worlds as Method L. H. M. Ling Once upon a time, there lived a colorful and proud little fish. He had lots of friends and loved the water surrounding him. He swam swiftly, quickly, and gracefully—and was very agile at getting his food. One day he met a turtle, an old friend whom he had not seen in a long time. The fish greeted the turtle. “Hello, Sister Turtle, how are you? I have not seen you in a long time. Where have you been?” The turtle replied, “I am fine, thank you, Fish. I was away on earth for an errand.” “Oh, really? What is earth? Is there something beyond this lovely water?” “Very much so.” “What does it look like?” The turtle paused. It was difficult to find the right words to describe something that the fish had neither experienced nor seen. But the impatient fish interrupted the turtle's thoughts. “Is earth like water?” “Uh, no….” “Can you swim in it?” “No.” “Do you feel the same pressure as you go deeper?” “No, it's not like that at all….” “Does it dance with sparkling lights when the sun shines on it?” “No, not really….” Impatient, the fish got mad. “I have asked you many questions about earth, and all you can answer is no. As far as I am concerned, that earth of yours does not exist.” And with a disdainful flash of his colorful tail, he swam proudly away. The turtle sighed, “How can one know something new when one's questions are based on the prejudices of the old ?”1 This Vietnamese folktale illustrates for me the central problematique facing our discipline of international relations (IR) today: multiple worlds and how Page 284 → to know them. In the case of IR, the fish is more like a whale. The latter not only rules the seas; it can experience another life above the sea—when it surfaces to spout, for example—but chooses not to recognize it as so. For too long, the whale of IR—realism2—has subjected our discipline to an illusory universalism propagated through notions of power and the state. Though many states now populate our “international system” and each exercises power within its limits, these facts alone do not verify the claim that, therefore, every state is the same and acts on power in the same way. Doing so would require two, related corollaries: (1) state making and power construction are identical processes across time and space because (2) the Westphalian interstate system has succeeded in eradicating all previous “imagined communities” of clan, tribe, nation, language, religion, and/or civilization. Precisely because this is not so (September 11 is the latest example), realists vow to renew the fight against “instability,” “pariahs,” and “the clash of civilizations.” Feminists and other critical theorists have led the charge against such realist one-worldism.3 First, they have

identified this perspective as precisely that. Neither an objective law of nature nor an intrinsic element of the human condition, realist one-worldism is but one representation of the world. Its longevity relies on a deceptively simple self-justification: that is, because realists believe that the world is nasty, brutish, lonely, poor, and short, they behave accordingly, thereby ensuring that the world is indeed so, which in turn affirms their belief that the world is nasty, brutish, lonely, poor, and short. Meanwhile, realists proclaim loudly their goal of averting a global holocaust by entrenching us further into this one-world logic—of which the latest manifestation is rational choice theory. More than other critical theorists, feminists have exposed the inherently gendered nature of such world making. Not only is realist one-worldism a social construction, but it reflects the experiences, practices, norms, and institutions of a particularly small segment of the world's population: propertied, white males. What traditional IR has passed off as “universal” or “human,” feminists have pointed out, should be described more accurately as “particular” and “androcentric.” For this reason, IR as a field of study and practice implicitly discriminates against anyone who does not or cannot abide by realism's version of hegemonic masculinity—even those bodies that may be white, male, propertied, and even heterosexual but who may subscribe to an alternative logic of understanding or constructing the world. In this sense, feminists have highlighted at least two worlds operating in IR: one of men, masculinity, and patriarchy and a second, hidden one of women, femininity, and matriarchy. Cutting-edge theorists like Cynthia Weber have deconstructed this binary further to include a “queer,” third world of blurred pairings between men /women, masculinity/femininity, patriarchy/matriarchy.4 In her analysis of U.S. foreign policy toward five Caribbean neighbors (Cuba, Dominican Republic, Grenada, Panama, Haiti), Weber has found a consistent resort to hypermasculine tactics (invasion, bombing, kidnapping, embargo) on the part of the U.S. government to cover for a sense of national emasculation that burns from unrequited desire. Similarly, Page 285 → in my own work,5 I have shown how the West's historical conviction that Asia must be a woman who should be conquered (with Puccini's Madame Butterfly as cultural exemplar) actually reveals a quaking, lovesick, inverted Western Butterfly who wants nothing more than to be loved by the Oriental Other (with David Henry Hwang's M. Butterfly as counterexemplar). Herein lies the challenge for feminist and critical theorizing in IR. While feminist theorizing has helped us realize the world's gendered split and multiple personalities, we need further explorations into other worlds and their ways of world making. Postcolonial feminists, for instance, have surfaced a deep-sea “whale” within liberal (white, Western) feminism.6 It proclaims “global sisterhood” for all but denies equal standing for nonliberal (nonwhite, non-Western) feminisms and feminists.7 Liberal feminism presumes that gender supersedes race, class, and culture, thereby allocating one categorical patriarchy to rule over all. Doing so effectively upholds a de facto hegemonic identity that is white, heterosexual, and North American/western European.8 In silencing unequal power relations among women, liberal feminism reframes “difference” into the familiar model of “pluralism.” That is, while difference acknowledges deeply embedded, multiple “lifeworlds” in constituting who and what we are, pluralism cares not for the content of such lifeworlds but merely recognizes their form. Indeed, according to Anthony Arblaster, “inclusion” in liberal theory historically means tolerance of, not acceptance of or, more important, engagement with difference.9 This pluralist logic invariably objectifies women, despite feminist avowals to the contrary. Celebratory of a linear, progressive logic of capitalist development, it returns Third World women (including those in the First World) to a center/periphery dichotomy. White women always sit comfortably in the center while Others either “crouch” (to use Frantz Fanon's famous characterization of the colonized) covetously on the periphery or straddle precipitously on the “borderlands” of multiple worlds.10 Such hegemonic self-involvement loses out on potential alliances that could be forged among feminists who, despite their differences, still share a common purpose in social emancipation. Equally important, such hegemonic self-involvement prohibits learning about who and what we are. Accordingly, the question remains: how to know multiple worlds, let alone communicate across them? Various strategies have been suggested: “strong objectivity,”11 cultural “conversations,”12 “empathetic cooperation,”13 “mobile subjectivities,”14 cultural “traveling.”15 These efforts are reinforced by a greater recognition of “contact

zones”16 between cultures—specifically those of colonizer and colonized—that allow their common progeny to “autoethnographize” their own subjectivities. Even so, the concept of subjectivity remains unitary. Despite all this “traveling,” “conversing,” “cooperating,” “mobilizing,” or “autoethnographizing,” an implicit assumption still holds: that is, subjectivity as an analytical unit may be multiplied—even “deconstructed”—to reflect greater complexity in real-life experience, but it remains intact as a core or essential source of difference. Harding's notion of “strong objectivity” serves as an Page 286 → apt example. She calls for an additive approach to multicultural knowledge: that is, by hearing from more voices, especially those that have been silenced or marginalized, we may “know” more. But the very condition of communicating across multiple worlds induces an intimate mixing of subjectivities, where the Self's boundaries begin to blur with those of the Other. Distinct labels like “southern,” “minority,” or “Other” do not simply add to more familiar ones like “northern,” “majority,” or “Self.” Instead, the interaction between these two identities triggers the eventual rise of a third, hybrid subjectivity. It compels a recognition that the southern, minority, Other exists in part within the northern, majority, Self, and vice versa. Put in terms of the folktale that opened this essay, we need the fish and the turtle, first and foremost, to continue their mutual inquiry.17 But the turtle needs to speak up. It cannot just react to the impatient fish/whale's questions that, essentially, reconstruct the unfamiliar in terms of the familiar, thereby never getting access to the unfamiliar.18 The turtle needs to show the fish/whale how to “see” the other, unfamiliar world through its relationship to the familiar one. (Alas, epistemological richness is both a privilege and a burden!) The turtle can best accomplish this task by discarding the temptation to objectify either Self or Other as distinct identities. Instead, we must recognize history as it is and always has been:19 a constant process of hybrid world making to ensure survival.20 For her part, the turtle needs to search within her own subjectivity for that commonality that binds all turtles, fishes, and whales. The fish/whale, in turn, should listen, patiently and openly, while developing a method or hermeneutic of asking questions.21 These must pertain to the turtle's knowledge, rather than reaffirm what the fish/whale already knows. For such knowledge is not just an interesting aside that adds spice to life, nor does it simply cumulate information for predictive purposes, nor is it the “right” or “fair” thing to do. Rather, the turtle's wisdom helps the fish/whale survive better in a universe of multiple worlds, each fraught with hazardous rocks, coasts, currents, and beaches. For this reason alone, the fish/whale has a vested interest in forgoing age-old hierarchies of anarchy/order, center/periphery, Self/Other, or masculinity/femininity. The whale of IR must learn, finally, that these offer the mere illusion, rather than actuality, of power and control.

Notes Many thanks to Anna M. Agathangelou, Celine M. Larkin, Thanh-Dam Truong, and Augusta del Zotto for their comments on this paper. The paper was prepared for a millennial reflections panel on feminist theory and gender studies held at the annual meeting of the International Studies Association, Los Angeles, Calif., March 15–20, 2000. 1. Many thanks to Thanh-Dam Truong for recounting this folktale to me. 2. By realism, I refer to the classical tradition of Hans Morgenthau and Henry Kissinger but also include its latest variant: neoliberal institutionalism. 3. Ann Tickner has characterized this condition of nonunderstanding between conventional and critical /feminist IR theorists as one of talking across whole worlds. Robert Keohane has taken this gap to mean the difference between problem-solving theory (which accepts the world as is) and critical theory (which seeks to change the world). As a counter to Tickner, he cited Hans Morgenthau as one realist who problem solved in order to change the world. But Tickner's point, reinforced here, was that Morgenthau and other likeminded realists still believed in one world. Hence, whether he sought to change the world or not was irrelevant. Ultimately, Morgenthau subscribed to an authoritarian Page 287 → assumption about the world.

J. A. Tickner, “You Just Don't Understand: Troubled Engagements between Feminists and IR Theorists,” International Studies Quarterly 41 (1997): 611–32; R. O. Keohane, “Beyond Dichotomy: Conversations between International Relations and Feminist Theory,” International Studies Quarterly 42 (1998): 193–98. 4. C. Weber, Faking It: U.S. Hegemony in a “Post-Phallic” Era (Minneapolis: University of Minnesota Press, 2000). 5. L. H. M. Ling, Postcolonial International Relations: Conquest and Desire between Asia and the West (London: Palgrave, 2001). 6. This critique could be applied to Marxian feminism as well. But liberal feminism takes the brunt of this charge given its spread through neoliberal globalization in the world today. 7. Well-known feminists like Susan Moller Okin and Martha Nussbaum, for example, have argued that only Western liberalism can “save” Third World women from their Third World traditions. Neither recognizes that there are elements of Third World traditions that may be as emancipatory as their own liberal one, nor that Western liberalism may be as incarcerating as some Third World traditions. S. M. Okin, “Is Multiculturalism Bad for Women?” in Is Multiculturalism Bad for Women? Susan Moller Okin with Respondents, ed. J. Cohen, M. Howard, and M. C. Nussbaum (Princeton, N.J.: Princeton University Press, 1999), 7–26; M. Nussbaum, introduction and “Emotions and Women's Capabilities,” in Women, Culture, and Development, ed. M. Nussbaum and J. Glover (Oxford, England: Clarendon, 1995), 1–15, 360–95. 8. This hegemonic subjectivity excludes even those descendants of white, heterosexual, North American /western Europeans in former colonial locations like Australia, New Zealand, and Latin America. Similarly, it exiles those from transitional societies in Russia and eastern Europe who may be categorized as racially “white” but considered “Third World” politically, economically, and culturally. 9. A. Arblaster, The Rise and Decline of Western Liberalism (London: Basil Blackwell, 1984). 10. While we should not ignore the pain, confusion, and sheer tediousness of those caught “in between” or belonging to “neither-nor” worlds (see C. Moraga and G. Anzaldua, eds., This Bridge Called My Back [New York: Kitchen Table, Women of Color Press, 1983]), neither should we project a glorious “centeredness” for those who are not considered “ethnics,” “minorities,” or “Third World.” Such a presumption is not only self-defeating (“we are never centered”) but also inaccurate. R. W. Connell has shown in his work on masculinities that even those who are supposedly the most centered in today's world—heterosexual, white men—face similar dilemmas of dealing with a societally imposed identity that may conflict with or undermine their personal subjectivity. R. W. Connell, Masculinities (Berkeley: University of California Press, 1996). 11. S. Harding, Is Science Multicultural? (Bloomington: Indiana University Press, 1998). 12. D. L. Blaney and I. Inayatullah, “Prelude to a Conversation of Cultures in International Society? Todorov and Nandy on the Possibility of Dialogue,” Alternatives 19 (1994): 23–51. 13. C. Sylvester, Feminist Theory and International Relations in a Postmodern Era (Cambridge: Cambridge University Press, 1994). 14. K. Ferguson, The Man Question (Berkeley: University of California Press, 1993). 15. Moraga and Anzaldua, This Bridge Called My Back. 16. M. L. Pratt, Imperial Eyes (London: Routledge, 1992). 17. In this sense, the dialogue between Keohane (“Beyond Dichotomy: Conversations between International Relations and Feminist Theory”) and Tickner is a good beginning. Tickner, “You Just Don't Understand,” 611–32. 18. Keohane's continued attempt to reframe feminist IR into conventional positivist categories of hypothesis testing and question formation, despite avowals to the contrary, is but one example of this. Keohane, “Beyond Dichotomy.” 19. As Stuart Hall noted: “In this ‘post-colonial’ moment, these transverse, transnational, transcultural movements, which were always inscribed in the history Page 288 → of ‘colonisation,’ but carefully overwritten by more binary forms of narrativisation, have, of course, emerged in new forms to disrupt the settled relations of domination and resistance inscribed in other ways of living and telling these stories.” S. Hall, “When was ‘the Post-Colonial’? Thinking at the Limit,” in The Post-Colonial Question: Common Skies, Divided Horizons, ed. Iain Chambers and Lidia Curti (London: Routledge, 1996), 251. 20. For example, Robert Wright has posited that evolution favors systems that figure out ways of coexistence, rather than the old Darwinian model of survival of the fittest. R. Wright, The Moral Animal:

Evolutionary Psychology and Everyday Life (New York: Vintage Books, 1995). 21. Trinh T. Minh-ha has described this as an attitude of learning through indirection: “One can only approach things indirectly. Because, in doing so, one not only goes toward the subject of one's focus without killing it, but one also allows oneself to get acquainted with the envelope, that is, all the elements which surround, situate or simply relate to it.” Trinh T. Minh-ha (in conversation with Annamaria Morelli), “The Undone Interval,” in Chambers and Curti, The Post-Colonial Question, 4. Gayatri Chakravorty Spivak has offered another suggestion. It would apply to all “Selves” coping with their “Others”: “Rather than imagining that women automatically have something identifiable in common, why not say, humbly and practically, my first obligation in understanding solidarity is to learn her mother-tongue. You will see immediately what the differences are. You will also feel the solidarity every day as you make the attempt to learn the language in which the other woman learnt to recognize reality at her mother's knee. This is preparation for the intimacy of cultural translation.” Quoted in C. Bulbeck, Re-Orienting Western Feminisms: Women's Diversity in a Postcolonial World (Cambridge: Cambridge University Press, 1998), vi.

Page 289 →

ON THE CUT(TING) EDGE V. Spike Peterson I would first like to thank the editors for inviting me to contribute to the Millennial Reflections Project. In preparing my remarks I have had the benefit of reading the essays presented at the “Feminist Theory and Gender Studies” and “Alternative and Critical Perspectives” panels. Indeed, my remarks often build on points elaborated by my colleagues and are especially indebted to Marysia Zalewski's essay. The latter inspired the themes and the title of my own contribution. Rather than repeat arguments already articulated, my essay focuses on identifying the strengths of feminist international relations (IR)—what renders it “cutting edge”—and considering how these strengths are related to the successes of feminisms and feminist IR. I will also consider how these very strengths of feminisms are related to the “failures” of feminist IR—what renders it “cut”—in relation to mainstream IR. In other words, I wish to explore more fully Zalewski's key insight: that feminisms' strengths explain both why and how feminist IR is cutting edge and why and how it also gets cut by the mainstream. Finally, I will consider these strengths in relation to the current state of IR and cutting-edge horizons.

Feminisms and Feminist IR Before proceeding, I feel obliged to clarify in my own words how feminisms constitute a wide spectrum of analytical and political positions. A variety of overlapping feminist projects constitute a continuum that spans positivist and “alternative” (postpositivist, poststructuralist) orientations. Typically, feminist interventions in any discipline begin by exposing the omission of actual women and their activities, while also documenting how “woman” is represented as deviant from or deficient in respect to male-as-norm criteria. For example, the model of human nature (as atomistic, self-interested, acquisitive, and competitive) that underpins the dominant discourse in IR is in fact based upon interpretations by and of a particular subset of humans (elite males) in a particular context (modern Europe). Hence, these are suspicious as universalizing claims about all humans—or even all males—at all times. Page 290 → Probably the most familiar feminist work involves the subsequent project that “corrects” androcentric bias by asking “where are women and gender?” in the context of conventional topics.1 “Adding women” exposes the androcentric assumptions of conventional accounts, inserts actual (embodied) women in our picture of “reality,” and reveals women as agents and activists, as well as victims of violence and the poorest of the poor. But it also, and significantly, forces a rethinking of foundational categories by revealing the extent to which excluding women /femininity is a fundamental structuring principle of conventional thought. Indeed, women cannot simply be “added” to concepts and categories that are constitutively masculine: the public sphere, the military, rationality, political identity, objectivity. Either women as feminine cannot be added (that is, women must become like men) or the constructions themselves are transformed (that is, adding women as feminine alters masculine premises and conceptual boundaries). For many feminists, these studies reveal gender as not simply a trait of embodied individuals but a historically institutionalized, structural feature of social life: it is a socially imposed and internalized “lens” through which individuals perceive themselves and the world, and the pervasiveness of gendered meanings shapes concepts, practices, identities, and institutions in patterned ways. In short, gender is not simply an empirical category referring to substantive men and women but an analytical category, pervasively shaping how we conceptualize, think, and “know.” How we think, of course, both shapes and is shaped by who we (think we) are, how we relate and act, and what we do. Understanding gender as an analytical category enables feminists to criticize not only the exclusion and/or denigration of females (as a sex category) but also the masculinist constructs that underpin philosophy (reason, abstraction), political theory (atomistic individualism, sovereignty), economic models (waged labor, rational choice), and science (objectivity, dichotomies). It thus links “women” and gender in the more

complex project of reconstructing theory. As we move along the continuum, feminisms focus less on sex as an empirical variable and more on exploring the interdependence of masculine-feminine, the centrality of gendered identities, and the significance of gender in how we conceptualize and theorize. Conventional categories and dichotomies are not taken for granted but problematized, and there is more attention to symbols, discourse, foundational claims, epistemologies, and theoretical debates. The point here is to reiterate that feminisms are not only about “adding women”—as an empirical category—to existing frameworks and the IR mainstream. These are crucial projects, but understanding gender only as an empirical category misses too much of what makes feminisms intellectually exciting and politically transformative. Perhaps we could even say that it is gender as an analytical category that makes feminisms truly cutting edge.

Feminist IR as Analytically and Politically Cutting Edge Zalewski identifies feminisms as cutting edge because they so thoroughly destabilize “what we think we already know and how we know it.”2 In this Page 291 → sense, feminisms force us to reexamine presumed givens and stretch us to address the resulting “shake up” of “conventional ideas about subjectivity, epistemology, ontology, agency and political action in IR.” What I want to explore in more detail is how feminisms and more specifically, feminist IR, are cutting edge both analytically and politically.3 I will subsequently consider how this positions feminist IR positively and negatively in relation to mainstream IR. Analytically There are numerous interweaving threads in the claim that feminisms are cutting edge and also destabilizing.4 First, given that gendered identities, bodies, concepts, and practices permeate social relations and empirical conditions, the study of gender requires and produces transdisciplinary orientations. We can observe that institutionally (if not intellectually or ideologically), conventional disciplines tend toward conceptual and methodological reification. This suits certain purposes (e.g., cumulative knowledge claims, or Cox's “problemsolving” theory) and inhibits others (e.g., diachronic reflection, or Cox's “critical theory”). An extensive literature argues the advantages and disadvantages of each. Not surprisingly, cutting-edge work is favored by (but not exclusive to) transdisciplinary orientations insofar as these by definition stretch or transgress familiar and often constraining boundaries. Similarly, transdisciplinary scholars are more likely to be exposed to, and therefore be aware of and engaged with, a plurality of methods and interdisciplinary debates, and these conditions favor (without ensuring!) an epistemological sophistication that is less required or cultivated by mono-disciplinary orientations. As one consequence, from transdisciplinary starting points feminists generate more wide-ranging and encompassing analyses of social relations. For example, in their study of world politics feminists draw upon and link areas of inquiry (e.g., developmental psychology, family relations, domestic violence, welfare economics, heterosexist religion, ecological issues) neglected in mainstream accounts. Providing a different example, Christine Sylvester notes how Cynthia Enloe “spots the internationally destined woman as body beautiful—Carmen Miranda—and links it to international trade in tropical fruits and discusses it in the context of America's Good Neighbor Policy of yesteryear and also says something about land reform in Latin American development strategies.”5 As these examples suggest, the ubiquity of gender prompts feminists to transgress not only disciplinary boundaries but the boundaries separating public and private and “levels of analysis”—boundaries that are paradigmatic in political science and IR. For example, feminists examine how individual internalization of gender stereotypes (e.g., masculinity as competitive, unemotional, and “in control”) interacts with sociocultural inputs (e.g., patriarchal religious beliefs, pornography, violent entertainment) and with gendered divisions of labor (e.g., politics and the military as “men's work”) to fuel international relations that are competitive and conflictual (e.g., dominated by competitive men seeking autonomy and control through Page 292 → direct and indirect violence). This linking of the personal (and psychological) to the cultural, political, and international is especially well developed in feminist studies of nationalism, which exemplify feminisms' cutting-edge status in relation to analyses of identity.6

As Jindy Pettman notes in her essay, “feminists have particular interests and investments in the politics of identification, with gender and sexuality crucial components in themselves, as well as constituting and marking other boundaries and power relations.”7 Whereas feminists have an enduring interest in the politics of identity, scholars in political science and IR have conventionally neglected these areas of inquiry. Until recently, the latter disciplines tended to take for granted both a Eurocentric model of the modern subject (identified as unitary, autonomous, interest maximizing, and rational) and a (Eurocentric) spatial model of public sphere agency and territorial states. In these accounts, the dichotomy of public-private locates political action and identity in one but not the other sphere; the dichotomy of internal-external distinguishes the identity of citizens and order within from “Others” and anarchy without; and the dichotomy of culture-nature (civilized-primitive, advanced-backward, developed-undeveloped) “naturalizes” First and Third World identities coded by uneven geopolitical power. Empirical and epistemological transformations challenge the adequacy of these conventional accounts. In terms of the former, statecentric political identity increasingly shares the stage with nonterritorial claimants. Subnational and transnational social movements transgress territorial boundaries in favor of identities “grounded” on ecological, antinuclear, ethnic, feminist, religious, and other non-state-based commitments. Uneven flows of information, culture, people, technology, and capital involve human migrations and diasporic identities that complicate territorial frames of reference. The globalization of production and finance undercuts national economic planning, eroding state sovereignty and the political identities it presupposes, even as supranational forces alter state power and subnational conflicts expose the illusion of homogeneity promoted in nationalist narratives. In short, identities conventionally “grounded” in state territoriality are losing ground to a politics of new or even “non” space(s). In terms of epistemological transformations, critiques of positivist empiricism have altered our understanding of agents, identity, and subjectivity. In contrast to the modernist conception of a unitary rational actor, contemporary social theory illuminates the multiplicity of subject locations (implying multiple identifications) and their dynamic interaction “within” the self and in relation to the self's environment. In short, identities are socially constructed as ongoing processes: they are embedded in and interact with historically specific social contexts composed of intersubjective meaning systems, practices, institutional structures, and material conditions. Hence, the study of identities must be historical, contextual, and dynamic: asking not only how identities are located in time and space but also how they are (re)produced, resisted, and reconfigured. These remarks suggest that any cutting-edge scholarship must engage the theory and practice of identity, identification, and subjectivity, and feminists are at the forefront of such work. Page 293 → It is not simply that attention to identity raises issues that are typically neglected in mainstream accounts (where positivist-empiricist orientations marginalize them as psychological, private sphere, emotional, and too subjective). It is also that feminists bring identity and subjectivity into relation with more familiar features of mainstream study. My own work casts this relational orientation as an alternative analytical framing, or “triad analytics.” The latter posits identities, social practices/institutions, and meaning systems as three co-constituting dimensions of social reality. More specifically, it juxtaposes “who we are” (identity, subjectivity, sexuality, allegiance, etc.), “what we do” (actions, practices, social structures, etc.), and “how we think” (concepts, discourse, ideology, etc.). This triadic framing moves away from dichotomized constructions in favor of understanding the material (empirical, structural) and symbolic (analytical, discursive) relationally, that is, as interdependent (mutually constituting), and it insists that “what we do” and “how we think” are also inextricable from “who we are.” It thus affords an interactive, multilevel, multivariable, and transdisciplinary orientation and can facilitate conversations among diverse theoretical perspectives. On the one hand, the triad makes reference to, or can be linked with, features of social life that are familiar in modernist, positivist, and critical accounts (e.g., agents, ideology, institutions). On the other hand, the triad takes subject formation seriously, privileges relational rather than dichotomized analytics, and “opens new spaces” for theory/practice; these are “moves” more familiar in critical and “alternative” accounts. To return to an earlier example, consider how gender is a structural feature of nationalist projects in codetermining senses: the identities mobilized in nationalist activities are gender differentiated (resolute leaders, supportive caretakers); the practices of nationalism are gender differentiated (men at the war front, women at the home front),

and the symbols and discourse of nationalism are deeply gendered (masculine protectors of the motherland, feminine signifiers of the nation and cultural identity). We see these relationships horrifically at work in the example of rape as a weapon of war. “Men” are expected to defend their motherland and protect “their women.” Rape is not only a metaphor of national humiliation, suggesting the failure of men to defend their territory and the loss of manhood in defeat, but also an embodied practice whereby the Other threatens group reproduction by impregnating the group's women with “alien” seed. Ethnonationalist conflicts in the 1990s have demonstrated all too well how rape is a weapon of war, how performing gender identities involves violence, and how the female body is not only figuratively but literally a battleground. In short, the gendering of nationalisms is not a coincidence but integral to how we identify with, think about, and go about making war. The point here is that feminisms are cutting edge analytically in part because they are transdisciplinary, translevel, and multidimensional, in part because they pioneered studies of identity, and in part because they address complexity through innovative and relational analytics. The nationalism example begins to suggest the linkages between feminisms as cutting edge not just analytically but politically. Page 294 → Politically Feminists acknowledge that their work is informed by normative/political commitments. The specifics of that commitment vary tremendously, but the very acknowledgment of a commitment to “improving the conditions of women” links and strengthens feminists even as it also works against feminist projects by fueling resistance from those who deny the politics of all knowledge claims or denounce gender equality. By definition and conviction, then, feminists are engaged in critical studies, analyzing and resisting status quo masculinism and gender hierarchy. But self-reflection and critical politics are integral to feminist theory/practice in several additional senses. First, like all marginalized and subordinated groups, feminists must be consciously “political” if they are to survive, much less prosper, in a typically indifferent and frequently hostile environment. For feminist academics, this involves career -and life -defining trade-offs in the pursuit of research, publishing, professional, personal, teaching, mentoring, and activist priorities. An important consideration is how the need to build and sustain an institutional/professional presence for feminist IR diverts precious time from more “serious” (read: research and publishing) activities. Given uninstructed and resistant audiences, feminists must also spend precious time defending their research orientation and repeating basic argumentation, which depletes time available for forging ahead with an expansive research agenda. The lived politics of this struggle are part of what my colleagues mean by debating whether to “ask feminist questions in IR” or “IR questions in feminism.”8 Second, and very significantly, the diversity among women has forced feminists to reflect critically (and uncomfortably) on the meaning of feminism, definitions of “woman,” the politics of representation, and the dangers of universalizing claims. “Sisterhood” aspirations have always been in tension with differences of ethnicity/race, class, age, physical ability, sexuality, and nationality, especially so in the global context that engages feminist IR. As Lily Ling reminds us, postcolonial feminists “have surfaced a deep-sea ‘whale’ within liberal (white, Western) feminism [that] proclaims ‘global sisterhood’ for all but denies equal standing for nonliberal (nonwhite, non-Western) feminisms and feminists.”9 However one assesses the success or failure of feminisms to address the challenges of difference, I believe feminists have taken those challenges more seriously, and moved more responsibly to address them, than most oppositional groups. This is due in part to taking their commitment to social justice seriously (an uneven record) and in part, I believe, to the unique situation of feminists in the academy. Struggling with these complicated and arduous challenges, feminists both drew upon and expanded their transdisciplinary orientations and, especially, their analyses of identity and identification. These were somewhat unique resources, and the resulting scholarship is surely one of feminisms' claims to cutting-edge status. At the same time, addressing these questions involved analytical development in additional respects, for example, in regard to ontological claims, epistemological Page 295 → debates, and theoretical advances. No less significant, addressing these questions involved developments

in political practice, for example, in regard to movement priorities, organizational politics, and long-term, “big picture” strategies. What I want to emphasize is the inextricable link between analytics and politics, which is especially evident in feminist theory/practice in regard to structural hierarchies. On the one hand, feminisms have transdisciplinary and complex analytical resources for investigating and theorizing about identity, difference, and historically specific hierarchies of oppression (heterosexism, racism, classism, etc.). On the other hand, feminist claims to political relevance and critique have “forced” them to address embodied differences of power: compared with other academic units, feminist scholars are expected to “walk” their (egalitarian) “talk.” In struggling to do so, feminists draw on and expand their analytical resources, and generate cutting-edge analyses of hierarchical power. The particular contribution I have in mind demonstrates these points and illustrates the transformative implications of gender as an analytical category. I refer to the feminist analysis of gender hierarchy as fundamental to domination in its many guises. Viewed empirically, females suffer disproportionally under systems of domination insofar as females constitute one-half of most subordinated groups and are systematically rendered more vulnerable to sexual and other violence, inadequate health care, political subordination, and economic impoverishment. Viewed analytically, the naturalization of women's oppression—taking gender hierarchy as “given” rather than historically, politically constructed—serves as the model for depoliticizing exploitation more generally, whether of groups or of nature. That is, feminists argue that the subordination of women and femininity to men and that which is privileged as masculine is not “natural” or inevitable—as it is typically represented—but socially constructed and historically contingent. Dominant (masculinist) ways of thinking, however, naturalize the marginalization, objectification, and corollary exploitation of all who are denigrated by association with the feminine: not only “women” but also nature, racialized minorities, effeminate men, and colonized Others. The ostensible “naturalness” of sex difference and masculine dominance is thus generalized to other forms of domination, which has the effect of legitimating them as equally “natural” hierarchies. Eliminating the justification of oppression as natural does not eliminate oppression, nor preclude other justifications of it. But the ideology that treats hierarchies as natural serves powerfully to legitimate and reproduce domination: through the internalization of oppression, the silencing of protest, and the depoliticization of exploitative rule and global hierarchies.10 In this crucial sense, feminist critiques are not simply about male-female relations but about all social hierarchies that are naturalized (depoliticized, legitimated) by denigration of the feminine. The transformative insight that this affords is that “adding women” (as an empirical category) may be an effective strategy for improving the lives of (some) “women,” but it falls far short of taking gender (as an analytical category) seriously. The latter entails Page 296 → deconstructing the naturalization of structural hierarchies in their multiple (and intersecting) forms. Hence, it involves improving the lives of all—not just women—whose oppression is naturalized by denigration of the feminine. The point here is the uniqueness of feminisms in transforming an initial critique of patriarchy into cutting-edge theory/practice that not only takes difference seriously and analyzes the intersection of structural hierarchies but also informs and enacts emancipatory political practice. Perhaps we could even say that this constitutes “progress”?

Feminist IR as the Cut Edge Zalewski identifies feminisms as “cut” by the IR mainstream because destabilizing “what we think we already know and how we know it” is not attractive to all knowers (or to any knower at all times). Considered through conventional IR lenses, the transgression of disciplinary boundaries, of public-private divides, or of levels of analysis is typically not even considered worthwhile, much less applauded. For those steeped in positivist empiricism, identities and subjectivities are detours at best, derailments at worst, in the pursuit of “objective” social science. Moreover, embracing complexity and framing analyses relationally are unlikely—even counterproductive—paths if one is seeking parsimonious explanation. In short, what renders feminisms analytically cutting edge effectively cuts them out of the picture framed by dominant starting points in IR.

Similarly, acknowledging the politics of making knowledge claims is not a commonplace in IR. There are several overlapping issues here. First, positivist commitments preclude such acknowledgment and prevent discussions of power in epistemological debates. Quite simply, it is deemed epistemologically “incorrect” to associate knowledge with power. Second, positivist disavowal of power cultivates a silence in regard to power at work in the discipline and its disciplining practices. We rarely hear the mainstream engage in critique of gatekeeping practices; old boy networks; U.S. dominance of the discipline; or race, gender, class, and nationality exclusions. Quite simply, privilege seems to be taken for granted and not interrogated, as if differential power were not of interest in the study of world politics (!) or had no effect on the quality of our scholarship and our collective lives. Third, insofar as allegiance to specific political commitments/ policies is acknowledged, liberalism (in the broadest sense) would seem to be the norm among the mainstream. Suffice it to say that this is a limited breeding ground for activists seeking structural transformation, and it is little surprise that the mainstream literature is so muted in regard to self-reflection or radical critique. In short, what renders feminisms politically cutting edge is cut down as suspicious and inappropriate by a mainstream that claims its knowledge is apolitical, its privilege is either irrelevant or unproblematic, and its desire for change is limited. That IR has little time for the cutting edge of feminisms is not to say that it cuts out all feminisms or all contributions of feminist IR. By reference to Page 297 → the continuum identified earlier, IR feminists have been especially productive in regard to research and publications that “add women” to existing frames of reference. Simply making this research available makes a difference in the discipline: for giving feminism and gender questions visibility; for easier access to empirical data and feminist analyses; for encouraging would-be feminist graduate students and scholars; and for adding feminist materials to mainstream courses and expanding resources for specifically feminist courses. Given the distance we have traveled in little more than a decade, and the variety of obstacles we have confronted, I consider these no small successes. We earned them the hard way and are right to be proud of them. But even work that simply adds women and thus conforms to the discipline's preferred empirical methods has a limited impact on the mainstream insofar as the latter remains focused on traditional topics where women are “apparently” absent: politics, foreign policy, military and security activities. In the absence of more complex understandings of power, the intersection of public and private, the force of identity politics, and so on, adding women does little to affect mainstream analyses. Stated differently, as long as IR understands gender only as an empirical category (e.g., how do women in the military affect the conduct of war?), feminisms appear largely irrelevant to the discipline's primary questions and inquiry. Understanding gender as an analytical category opens up entirely new areas and reframes old questions (e.g., how does masculinism affect—even entail—warring identities, attitudes, and practices?). But positivist commitments preclude seeing, much less appreciating, this expanded and admittedly less stable terrain. In the absence of a willingness to transgress conventional boundaries and deconstruct foundational dichotomies, gender as an analytical category is “out of sight” for the mainstream of IR. More specifically, the intellectually exciting and politically transformative work in feminist IR theory is effectively invisible because it is outside of what the mainstream can see through positivist and politically conservative lenses.11

Conclusion: Feminisms, Feminist IR, IR, and the Future With many of my feminist and alternative theory colleagues, I am persuaded that positivist commitments12 dominating the discipline are a major impediment to more productive conversations between feminists (many of whom subscribe to alternative perspectives), alternative theorists (some of whom are feminists), and the IR mainstream (who differ among themselves, but share key epistemological commitments). In particular, these commitments prevent understanding of gender as an analytical category and foreclose the openings and transformations this insight provides. That I consider this situation a loss for all concerned, and devastating for the quality and relevance of our scholarship, is no surprise. But what is to be done?13 Although feminist work has had a limited impact, and abandoning IR terrain in favor of “asking IR questions in feminism” has a certain logic (and emotional appeal!), there are good reasons for continuing to “ask feminist Page 298 → questions in IR.” It is important to remember that limited impact is not the same as no impact, and I

believe that IR needs feminist and other critical interventions. First, continuing to work “within IR” is important for the differences it already makes and will continue to make. Not least among these is the availability of feminist IR (and other critical studies) as an area of inquiry for students and an academic option for graduates. In the absence of more transdisciplinary offerings and job placements, building critical sub-fields directly benefits participants and indirectly promotes changes in the mainstream. Second, IR as a discipline needs to enhance its analytical contributions to our understanding of the world(s) we live in, and feminisms have important resources to share in this regard. Given the importance of transnational and global dynamics in our lives today, one might expect that IR would be the prominent discipline offering sophisticated analyses of these dynamics and possible responses to them. That this is not the case is due in part to epistemological starting points and analytical frameworks that preclude socially relevant questions being asked and limit the political relevance of explanations delivered. IR's analytical toolkit is demonstrably inadequate. In spite of the strengths of conventional approaches, their glaring failure to predict the collapse of communism suggests important limitations, even if the scale of local-national-regional-global problems and the accelerating pace of change did not otherwise demand revising and expanding our analytical capacities. The world “needs” more adequate and more relevant analytics in regard to globalization, and IR should be one important source of these. Perhaps we could even say that IR must take the cutting edge more seriously if it is to be relevant as a substantive area of inquiry and as a contributor to contemporary social theory. Third, and related, IR as practice has important political effects, yet the discipline (in contrast to many feminisms)14 engages in little self-reflection about its role in reproducing power relations. Consider the lack of attention to structural hierarchies and exclusionary practices within the discipline, as reflected, for example, in the organization and composition of the millennial reflections panels. Let me say first that the inclusion of feminist and alternative panels was no doubt well intended, and panel participants acknowledged this. But they also pointed out that “separate” panels effectively reproduces a division between mainstream panels engaging in “real” research and “Other” panels doing whatever it is that they do. Even more telling is the composition of the panels: of forty-eight participants “only five are from outside North America, and only ten are women.”15 Racial diversity is not quantified here, but it seems reasonable to note that diversity has not traditionally been a prominent feature of International Studies Association meetings, its mainstream panels, or its leadership rosters. This is not to deny that many individuals bemoan these conditions and that small efforts are made to “improve the numbers”—but we can hardly claim that racial diversity is a priority for the association or the discipline more generally. Of particular concern (and effectively reproducing these conditions), is how racism as a feature of international relations—that is, as worthy of Page 299 → study—appears almost nowhere in mainstream scholarship. A growing number of critical, postcolonial, and feminist scholars, myself included, argue that this “omission” is especially indefensible in today's world. When we “leave race aside,” we not only produce deficient scholarship but we also are complicit in reproducing structural oppression. And we know that structural oppression—whatever other factors must also be recognized—is key to conflict both within and between nation-states. In short, however great the strengths of conventional approaches, their glaring failure to take racism (and other structural oppressions) seriously has impaired our analyses of power and reproduced conflictual relations. Perhaps we can even say that the objective of a cutting-edge IR would be to understand and transform structural hierarchies and their violence-making effects. Perhaps most of the discipline shares this objective. But I am persuaded we cannot achieve it as long as we cut out critical, alternative, postcolonial, and feminist IR.

Notes 1. Betty Reardon Sexism and the War System [New York: Columbia University Teachers College, 1985]), Jean Bethke Elshtain (Women and War [New York: Basic Books, 1987]), and Cynthia Enloe (Bananas, Beaches, and Bases: Making Feminist Sense of International Politics [Berkeley: University of California

Press, 1990]) pioneered this line of inquiry in IR, and feminist IR scholars followed by exploring gender in /and security (J. Ann Tickner, Gender in International Relations: Feminist Perspectives on Achieving Global Security [New York: Columbia University Press, 1992]), states (V. Spike Peterson, ed., Gendered States: (Re)Visions of International Relations Theory [Boulder, Colo.: Lynne Rienner, 1992]), foreign policy (Nancy E. McGlen and Meredith Reid Sarkees, eds., Women in Foreign Policy: The Insiders [New York: Routledge, 1993]; Cynthia Weber, Faking It: U.S. Hegemony in a “Post-Phallic” Era [Minneapolis: University of Minnesota Press, 1999]), world politics (V. Spike Peterson and Anne Sisson Runyan, Global Gender Issues, 2d ed. [Boulder, Colo.: Westview, 1993/1999]; Peter Beckman and Francine D'Amico, eds., Women, Gender, and World Politics [Westport, Conn.: Bergin and Garvey, 1994]; Francine D'Amico and Peter Beckman, eds., Women in World Politics: An Introduction [Westport, Conn.: Bergin and Garvey, 1995]; Jan Jindy Pettman, Worlding Women: A Feminist International Politics [London: Routledge, 1996]; Jill Steans, Gender and International Relations [New Brunswick, N.J.: Rutgers University Press, 1998]), international organizations (Deborah Stienstra, Women's Movements and International Organization [New York: St. Martin's Press, 1994]; Sandra Whitworth, Feminism and International Relations: Towards a Political Economy of Gender in Interstate and Non-Governmental Institutions [London: Macmillan, 1994]), development (Marianne Marchand and Jane Par-part, Feminism, Postmodernism, Development [London: Routledge, 1994]), revolution (Mary Ann Tetreault, ed., Women and Revolution in Africa, Asia, and the New World [Columbia: University of South Carolina Press, 1994]), peace studies (Simona Sharoni, Gender and the Israeli-Palestinian Conflict [Syracuse: Syracuse University Press, 1995]), international political economy (Eileen Boris and Elisabeth Prügl, eds., Homeworkers in Global Perspective [New York: Routledge, 1996]; Christine B. Chin, In Service and Servitude: Foreign Female Domestic Workers and the Malaysian “Modernity” Project [New York: Columbia University Press, 1998]; Marianne Marchand and Anne Sisson Runyan, eds., Gender and Global Restructuring: Sightings, Sites, and Resistance [London: Routledge, 2000]), and IR theory (Gillian Youngs, International Relations in a Global Age [Cambridge, England: Polity, 1999]).Page 300 → 2. Marysia Zalewski, “Feminism and/in International Relations: An Exhausted Conversation?” in this collection. 3. I hasten to add that I understand analytics (theory, knowledge) and politics (practice, power) as inextricable; I separate them here for heuristic and organizational purposes only. 4. Again, not all feminisms or all IR feminists conform to the generalizations presented here. 5. Christine Sylvester, “‘Progress’ as Feminist International Relations,” in this collection. 6. For a discussion of the importance of identities in feminist theory/practice, see V. Spike Peterson, “Sexing Political Identity/Nationalism as Heterosexism,” International Feminist Journal of Politics 1, no. 1 (1999): 21–52; on gendered identities in IR, see Marysia Zalewski and Cynthia Enloe, “Questions about Identity in International Relations,” in International Relations Theory Today, ed. Ken Booth and Steve Smith (Oxford, England: Polity, 1995); Marysia Zalewski and Jane Parpart, eds., The “Man ”Question in International Relations (Boulder, Colo.: Westview, 1998). 7. Jan Jindy Pettman, “Critical Paradigms in International Studies: Bringing It All Back Home?” in this collection. 8. Zalewski, “Feminism and/in International Relations: An Exhausted Conversation?” and Sylvester, “‘Progress’ as Feminist International Relations,” in this collection. 9. Lily Ling, “The Fish and the Turtle: Multiple Worlds as Method,” in this collection. 10. Consider how the reigning ideology of neoliberalism both naturalizes “obscene global inequalities” and demonizes any attempt to criticize these practices and effects as “unnatural.” Zillah Eisenstein, Global Obscenities: Patriarchy, Capitalism, and the Lure of Cyberfantasy (New York: New York University Press, 1998). 11. Although divided by the panel organizers, alternative scholars share with feminists this critique of the mainstream's resistance to recognizing a plurality of methods and epistemologies, which effectively delegitimizes alternative and feminist theorizing. 12. See, for example, J. Ann Tickner (“Feminist Theory and Gender Studies: Reflections for the Millennium” in this collection) on positivist methodologies as inadequate for analyzing gender, hierarchical social relations, or the mutual constitution of theory and practice; and Steve Smith (“Alternative and Critical Perspectives” in this collection) on the dominance of U.S.-style positivism as delegitimizing alternative

perspectives as “outside the standards of social science.” 13. Ultimately, of course, what to do is a personal decision, and my own perspective is simply that. My comments here follow the discussion I have elaborated. 14. The point here is less that feminists have “the answers” or offer a definitive model than that their efforts, while continually warranting critical reflection and improvement, produce more diversity and inclusion than manifested in conventional ISA panels and activities. 15. Steve Smith, “Alternative and Critical Perspectives,” in this collection.

Page 301 →

CRITICAL PARADIGMS IN INTERNATIONAL STUDIES Bringing It All Back Home? Jan Jindy Pettman It is a daunting task to respond to the challenge to assess the state of the art of international studies (IS), or more particularly, feminist IS: to reflect on where we've come from and where we are now (and indeed who “we” are); and to ask, What's at stake here? For me, writing as a feminist in IS, there is always a double edge, asking both what happens to the international when we take feminism seriously in IS and what happens to women, sex, gender, and feminism when we take the international seriously. I argue that IS needs feminism, to give better and more inclusive accounts of the world; and that feminism needs an international perspective, especially in these days of intensifying globalization. This double move is typical of feminist IS, working both in and beyond—or even against—the discipline. I begin with several claims regarding feminism and IS, before elaborating on the difference I think feminism has made to our understanding of “the international.” This elaboration provides an opportunity to assess “progress”1 and to address the questions that guide this collection. I will conclude with two illustrations of feminist approaches, regarding war and the Asian crisis.

Making Claims 1. Feminism came late to IS. The first thirty years of the International Studies Association (ISA) were largely feminist free. Feminism began its formalized presence in the ISA with the establishment of the Feminist Theory and Gender Studies section (FTGS) in 1990, much later than equivalent sections and caucuses in other disciplinary fields. Does this late coming say something about the ways IS is constituted as a field, and its particular masculinisms? Or does it say something about feminism, as less useful for or amenable to IS? 2. Feminism appears to have had comparatively little effect on IS judging by most course offerings, articles or chapters, or references in many IS texts. Page 302 → But while the center holds, feminism does have a remarkably robust, diverse, and productive presence in and around IS—as a glance at the FTGS-sponsored panels in recent ISA conventions and the contents pages of the recently launched International Feminist Journal of Politics (IFjP)2 demonstrates. 3. Disciplinary recognition of and concern about women and gender lag behind the increasing (though still marginal) visibility of women and feminist-informed critique in the “real world” in terms of international agenda items, media reporting, international nongovernmental organization and UN forums, and some state rhetoric (though not necessarily state practice). Even the World Bank now declares an interest in women, and in gendersensitive data. This lag is worth reflecting on. One way to apply “standards” or judgments to assess theoretical insights of both mainstream IS and its various subfields is to ask how good at or close to reflecting or making sense of the “real world” each is. However, this is tricky—for the world is experienced rather differently by different people, whose identity, location, political interests, and power (or lack thereof) translate into different investments in and understandings of world politics. Nowadays, Euro-North American media frequently notice women.3 There are for example exceptional women in world politics—Madeleine Albright, one of the most powerful people in the world under the U.S. Clinton administration; or Aung San Suu Kyi, whose influence rests on democratic support and peaceful challenge to militarized violence in Burma. Often, though, women appear as victims or as symbols of the costs of world politics—of war, civil strife, environmental disaster. Women figure as problems for the international community or organization, through representations as the majority of refugees, or amongst the poorest of the poor. Women are now a category with special needs or a separate section in UN and agency reports. (Note that this category is

often actually one of “womenandchildren.” This effectively invisibilizes the effects both of/on youth and young children and fails to “gender” the account by discriminating between boy and girl children's experiences of war, violence, work, sex tourism, and so on.) Such representations do make women visible, but usually in ways that perpetuate the men/doer-women/done dichotomy, presenting men as the agents and women as the victims of world politics, whatever individual men or women are doing. At times, women do appear in more positive guises, as survivors or witnesses, as for example in the recent international war tribunal held in Tokyo against Japanese military sexual slavery in World War II. Some women do figure as actors, advocates, or claimants in international politics. Women function as a (difficult) identity or constituency. Women have long organized across state boundaries, especially in the cause of peace or women's rights. International conferences have proved to be rich sites for such organizing. Women's caucuses and networks flourished through the 1990s, around the 1992 Rio Environment and Development Conference, the 1993 Vienna Human Rights Conference, the 1994 Cairo Population and Development Conference, and the 1995 Women's International Conference in Beijing. As well, Page 303 → many women network and politic transnationally, in alliances built around shared experiences or advocacy, against sex tourism or military base sex, for example. They have succeeded in introducing “new” issues to the international agenda. Those now visible in the media and international forums include war rape, international trafficking of women and children, international service industries, labor migration, child soldiers, restructuring, and international health issues. This coverage is often a long way short of analyses that reveal how gender functions to constitute other identities and power relations, but it does provide clues and signs that still often get missed in IS. Recognizing that women are in international politics and attending to women as international actors opens up new spaces for feminist work; though how “feminism” figures here is another puzzle. There is no simple or single feminist position on/in IS (and it continues to surprise me how often I'm asked to give the lecture or write the chapter on feminism and …). Different kinds of feminism have been mapped for/in IS too,4 including schema naming feminist divides along liberal/empiricist, standpoint, socialist/materialist, postmodern, and more recently postcolonial lines. Some have commented on the difficult relations or lack of conversation between feminists in IS, too.5 Critical approaches with an explicit emphasis on theoretical issues predominate in feminist IS, though these do not necessarily exclude more empirical work. But while feminist IS as a collective is more likely associated with other critical approaches in IS, its relations with nonfeminist critical IS scholars, too, can be difficult or occasional.

Feminism Comes to IS Feminism in IS is one of many manifestations of second-wave feminism, driven by a fascinating and at times awkward combination of women's advocacy and academic critique. (The relationship between the category women, women's movements, and feminisms in the academy remains problematic in the international context, too.) Coming to the discipline comparatively late has consequences for feminism in IS, for there were already many feminisms, including nuanced and sophisticated offerings in relevant areas. Notable amongst these relate to development, peace and war, nationalism, and citizenship. As well, there are years of fierce feminist controversy around identity and difference between women to draw on. These differences often reflect the social impact of international processes and relations, including colonization, migration, race, and ethnicity. These relations have also attracted very little attention amongst most IS scholars and remain marginal.6 Until the last decade, and still in many places and texts, women, gender, and feminist scholarship were effectively excluded from IS. The assumption appeared to be either that IS was gender neutral, affecting women and men in similar ways, or that IS was indeed men's business, as its practitioners in the discipline and the management of world politics seemed to confirm. Some saw these positions as stemming in part from IS's failure to notice Page 304 → people, preferring a focus on states, markets, and other apparently impersonal forces. First-generation feminists in IS sought to deconstruct IS, to refute these assumptions and challenge routine, taken-for-granted exclusions. These works made space for “women,” and in the process posed questions about the discipline more broadly, including about who owns IS and what its territory, subjects, concepts, and questions should be. Issues of identity, including but not only gender, were often central.

Unsurprisingly, feminists have particular interests and investments in the politics of identification, with gender and sexuality crucial components in themselves, as well as constituting and marking other boundaries and power relations.7 As well, serious political or intellectual interest in “women” needs notice that women are many things, and what these things are can never be taken for granted. This relates to but is a different problem from that of tracking gender through all social relations including international politics and IS … a point I will return to.

The Difference Feminism Makes? An evaluation of feminism's impact, or lack of impact, on IS might make for discouraging reading. Rather, I wish to consider the contribution of feminist IS, the ways feminists have contributed, on its own terms8 (the question of the relations between the discipline and “its” feminists is taken up by the other feminist contributors to this collection). Feminists working in IS 1. radically expand the kinds of topics, issues, and content in IS, as a glance through the ISA program reveals. For example, the ISA 2001 program includes feminist panels on postcolonialism, critical methodologies, global-local feminisms, Internet activism, ethnic and national conflicts, women's narratives of violence and war, sex in the military, queer theory, migration, and hunger. While feminist issues are still mostly concentrated in FTGS panels and issues of the IFjP, they are now visible and do attract an audience beyond feminists in IS. Together, they unsettle the discipline. They suggest new ways of thinking about core issues, including power, security, and the state, and introduce new ones. 2. ask different kinds of questions: looking elsewhere, or differently; revealing habitual absences and structured exclusions. A beginning question, “where are the women?”9 has proven enormously productive, uncovering women in some unexpected places, or women absent because they are women, or women identified and acting as men. Asking “who are the soldiers?” reveals gender (only men?), sexuality (no gays?), (boy?) children, masculinities, ethnicity, and race. Asking why those working in export-processing zones are overwhelmingly young women and are often racialized as, for example, “Asian” or “immigrant,” delivers clues about the international sexual division of labor and the gendered nature of global capital, state transformations, and restructuring strategies. Page 305 → 3. reveal the masculinist assumptions, investments, and strategies of mainstream IS. The soldier, citizen, and state are still often presumed to be male; dominant group men's experiences and interests inform much of the discipline's business and culture. Critique can be labeled as negative, but it is actually productive, disrupting the territory, pushing the boundaries, and clearing a space for feminist reconstruction and revisioning. 4. in the process, widen definitions of the political in the international, too. Women's participation is more likely in liberation movements and social movements than on multinational corporation boards or state councils—clues to the gender of state power and signs to other sites of transnational politics. “Women's issues” have become international political issues, too: the recent Millennium special issue celebrating ten years since the first Women and International Relations conference held at the London School of Economics in 1988 included four articles related to international sex work, for example. 5. disrupt or transgress disciplinary boundaries. Feminists in IS often explore other ways of being in the world, bringing ways of thinking and seeing back from other disciplines: critical social theory, cultural studies and postcolonial studies, and especially feminism beyond IS.10 Increasingly, they also find other kinds of texts and sites for interrogating the construction and representations of gendered identities, including films and other forms of cultural production. These travels lead more conservative IS colleagues to question whether what feminists in IS do is really IS at all, and whether they are capable of displaying the desired levels of disciplinary loyalty to earn recognition. 6. internationalize the account—raising questions about the identity of the discipline, too. How do issues to do

with representation play out in the discipline? IS appears predominantly Anglo-North American and especially U.S. dominated in terms of jobs and journals for example.11 (This may seem an accurate reflection of global power.) First-wave feminist IS insisted on gender difference but initially remained largely within the same geopolitical territory. (I deliberately leave aside questions about the difference it makes being a woman doing IS—and of course not all women in IS are feminist.) While often located in these same IS heartlands, more graduate students and recent Ph.D.'s do come from outside the centers, and may be working on “somewhere else” (though of course the relation between origin and identity and intellectual and political interests is highly problematic). Exciting new work that might be characterized as second generation has helped push the boundaries further, in terms of locale, subjects, and methodology/theory.12 7. gender the account. While “gender” still often functions in IS as a euphemism for women, the constitutive role of gender in world politics is now the subject of keen interrogation. There is an extensive feminist literature on the construction of particular kinds of masculinity and femininity in different international sites, and on gender in IS, too.13 This literature develops more mobile, nuanced, and political understandings of gender relations—meanings and associations that do not necessarily attach to particular sexed bodies. Gender coding mobilizes particular kinds of roles and Page 306 → obligations, directs us to understand who are the good guys or what it is necessary to do, and rules out possibilities of thinking or acting differently.14 8. sex the account. While gender figured from the beginning of feminist intervention in IS, identifying the roles that sex and sexuality, and heterosexism, play in international politics is a more recent concern. This includes the frequent sexualization of the Other and deployment of sexual fears and fantasies in domination relations, and in representations of difference that mark boundaries and binaries in world politics.15 Gendering and sexing the account together provide 9. the possibility of an embodied IS. There is now a vibrant accumulation of feminist writings that identify particular kinds of bodies in different international engagements and the privilege or penalty that can flow from being identified as male or female, inside or outside, exotic or “normal,” in these exchanges. The body you are in can determine which side of the killing line you are on, how exploitable or sexualized you are, and the kinds of attention that your state and other states pay you. These writings also come closer to measuring the actual body impact of international politics and political economy. This is a reminder of the strangeness of a discipline that so relentlessly disavows women, sex, and gender and yet manifests and utilizes gendered discourse and gender meanings and associations to describe, and do, its intellectual business. It often “misses the body”16 and the bloody bodily impact of different kinds of international politics, which do occasionally become visible in concern about the domestic political effects of the body bag count, for example. Sex and bodies might sound new for IS, though they were there all along, in the utilization of body metaphors, for example. But what does emerge from these illustrations is how good feminism is to think with in terms of older and more familiar IS concerns. I was enticed into this field three decades ago by two puzzles—the problem of war and the problem of culture. The first I found centrally located, but with deafening silences and tantalizing glimpses. The second, I found hardly ever mentioned but always practiced (like gender, really). I will conclude this contribution with a brief consideration of the difference feminist questions can make to a study of war; and then with an example from my current research, venturing beyond the Euro-North American heartlands, seeking gender and sex in the politics around the Asian miracle and Asian crisis. I hope to use these to revisit the question of “standards,” or how we might assess the usefulness and possibilities of particular perspectives and ways of knowing or seeking to know.

Regarding War Along with nationalism, war is the subject of much of my writing. Definitive answers elude me. Feminist IS reveals that our (and others') work is always in process and can never be fully completed, that productivity and performance might be better judged by the questions we ask and especially useful Page 307 → in pointing out the kinds of questions that have not been or could not be asked.

I made Waltz's Man, the State, and War the center of my international relations course in the late 1960s. When I returned to teaching IS in 1991, feminism gave me more, and more useful, ways of interrogating the kinds of relations that make organized large-scale violence possible. I moved from “man” to men, to asking whether certain kinds of masculinity are more violent than others, to whether state militaries need certain kinds of masculinity. And certain kinds of sexuality, too—debates about gay men in state militaries generate as much heat as do those about women serving in combat positions. As the 1990s progressed, feminist and some new masculinities writings provided more clues about kinds of masculinity constructed in war and other kinds of international relations, too. A significant IS collection was entitled The “Man” Question in International Relations,17 refocusing attention on men, this time qua men—including in IS/A. Issues having to do with whether technology and current forms of warfare require different kinds of military masculinity are receiving attention. However, investigation of questions regarding cultural difference and different state and subculture notions of what it takes to be a good (military) man are only beginning.18 It is a political, ethical, and intellectual question of immense importance to ask how men (mainly) who would not resort to violence in their personal or social lives can be prevailed upon to use violence against those outside their own state or nation; and how people move from holding multiple and often shifting identities to a primary identity for which they will fight, kill, or die.19 Feminist interrogations of identity and difference and the politics of boundaries provide insights into nationalist passions and gendered civic obligations, including that of citizen soldiers. They help to denaturalize those popularly held explanations of war that center on “human nature” or ancient hatreds or the inevitability of violence in an anarchic world. They trace the history and politics of particular conflicts and wars, and so bring back the possibility of other kinds of response or intervention. In the process, they reveal the gender dynamics that infuse identities, political subjectivities, and power relations in interstate relations, too.

The Asian Other? New feminist IS writings trace the extent to which IS is manifest in a gaze from the center/s, overlooking other places and ways of being in the world.20 “Asia” has long functioned as Other for the worlds and ways of thinking that IS exemplifies. Asia was long constructed as feminine and ripe for conquest or domination by the West in colonial and in cold war times. IS's refusal to take coloniality and postcolonialism seriously reinforces and reproduces the circulation of racialized international identities like the West, Asia, and the Third World, which in turn influence analysis and policy toward Others. State and regional identities are effectively homogenized, essentialized; disguising both differences within these big identities (including Page 308 → gender differences) and the constant and considerable interchange and mutual constitution that the colonial experience has bequeathed to the contemporary world. These international identities are themselves constituted through gender as well as race/cultural difference. Changing international power relations signaled through the rise of East Asian states and the Asian miracle were seen as a threat to Western state power and manhood. East Asian triumphalism and the self-confident declaration of Asian values (from Singaporean and Malaysian leaders especially) claimed that a newly dynamic, successful Asia had recovered from its colonial emasculation. Instead, in a kind of gender reversal, “the West” was seen to be faltering, losing the manly qualities of hard work, self-control, competition, and strength, becoming decadent, immoral, obsessed with consumption and pleasure—in short, feminized. The 1997 Asian crisis ushered in new configurations in gender representations of state power, national identity, and international relations. As usual, gender was mobilized along with and to stand in for “culture.” Both Asian miracle and crisis were culturalized: what formerly explained success, in terms of miracle (for example, Confucian capitalism) now explained failure and crisis (for example, crony capitalism, corruption). These forms of essentializing disguise the gendered, and class, costs of both miracle and crisis. They do so through complicity with IS's notorious statecentrism, discouraging us from looking within and across states, and from asking questions such as Who pays during miracle, and crisis, and who gains? How do particular constructions of family and gender relations help subsidize states and markets through the construction of women's labor as cheap labor? And legitimize privileged elites by mobilizing gendered notions of citizenship and

hierarchy that undercut the rights and claims of labor, human rights, and women activists? How do reconstructions of national gender roles feed into interstate identity relations? My current research on Australia's relations with “Asia” asks what is obscured by the failure of academic and other commentators to take women, gender relations, and feminist writings seriously in analyzing these relations.21 This failure disguises the gendered impact of miracle and crisis, and the ways women's experiences become the basis for transnational organizing and campaigns that generate a feminist politics of region. But more than that: miracle, crisis, and military intervention (for example in East Timor in 1999) become sites for the gendered reconstruction of state power and interstate relations. Gender and sexuality mark hierarchies and power relations among states, too, and signal contests and changes in these relations. These illustrations demonstrate the usefulness of using feminist perspectives in explorations of key topics already close to IS: militarization, intensifying globalization and state transformations, national interest, and interstate relations. These are hardly add-ons or special interests. They go to the heart of the discipline, and of contemporary world politics and political economy. Feminism brings to IS more inclusive, indeed better, accounts of the world. Page 309 → But it's not only how “the world” works that feminism in IS thinks about. We have been forced to think seriously about how we work: about the politics of representation and the disciplinary politics behind who is heard or seen as an authority. Feminism is a political project, which removes the possibility of occupying an innocent position in knowledge making. It helps us to address key questions about the discipline: Who is it for? What is at stake in doing IS? Who is listening? Who (if anyone) are we accountable to? How do issues of power and identification play out in IS, too? Interrogating the politics of our own work, of the discipline, and of knowledge making more broadly removes the possibility of expecting a resolution of our differences, within feminism or with IS. Our debates and views will always be contingent, contested, situated, and in some ways quite particular. We are therefore called to reflect critically upon our own practice. Who do we recognize as authorities? Whose experiences do we take seriously and admit as evidence? Who do we talk to, read, and cite? Who do we imagine as our audience? Would the people we write about or those whose lives are directly affected by the processes or events we analyze recognize our accounts? Are those whose experiences we seek to understand in any way present in our work? The struggles around difference and voice in feminism push us to reflect more on our own practice in research, writing, teaching, supervising, and relating to colleagues and to the discipline more broadly. (Which is not to say that only feminists—or all feminists—do this. There are other, critical collaborators, those who work from different political and intellectual terrain to make complementary demands, including some represented in this collection.) Nor have these issues been resolved; but they are now part of our conversations and examples of emerging good practice.22

Looking for (a) Home? It is a cause for celebration that feminism is now well into critical mass in IS. Clearly, however, feminists are (not yet?) at home in IS. Our relations with the father/master discipline remain difficult, and will continue to be so, for feminist IS is disruptive, and hopefully transformative—and therefore rightly perceived as a threat by those who currently exercise disciplinary power and execute border patrols. And for a discipline that has an ongoing fascination with the workings of power, IS remains remarkably incurious (or resistant?) about its own power relations and conditions. For many feminists, the FTGS section and women's caucus offer a home or at least a safe visiting and recovery place while within the IS orbit. There is a tension between a wish to remain at home, to present one's work to colleagues who, however robust the exchanges, form an affinity grouping. Yet staying close to home might reinforce the borders and foreclose possible conversations that some feminists in IS still wish to pursue. And, as usual, home is not always safe—not all feminists feel at home in FTGS. Issues of diversity and difference within feminism are present within feminist IS, too. Page 310 →

Feminism in IS is becoming more international and more representative. Indeed, it is broad and rich enough to raise questions about its own centers and margins. There are important questions here about who “we” are, including in terms of background, generation, and current location, both geographic and institutional (each shorthand for very complicated feminist identity politics). This range, diversity, and at times discomfort provide the context within which we ask what it means to read, write, converse, and organize as feminists, in and around IS. These questions are especially pertinent in terms of building a critical feminist community around the ISA, FTGS, and IFjP, which was conceived and initially developed within the FTGS community. IFjP went to launch in mid-1999. It is located at the intersection of international studies, politics, and women's studies. Its first two volumes illustrate something of the range of contemporary international feminist engagements. It includes articles on girls in war zones, nationalism as heterosexism, travel, feminist nongovernmental organizations and UN conferences, the “temp” industry, and citizenship in these global times. It highlights feminist politics and analysis in amongst others Zimbabwe, Kuwait, Venezuela, Israel, Russia, and France. The “Conversations” section includes fascinating reflections on teaching, commentaries on films like GI Jane and Saving Private Ryan, and interviews with international activists. The book review section showcases the current state of feminist international studies, including through review essays and reflections on what ought to be taught in IS. The journal appears to be flourishing; though whether it is having any effect on the wider IS community has yet to be tested. Concluding, then: feminist IS is a new field, and one in which postgraduate work figures largely. Its recent development follows no fixed plan and is not easy to map. Like the real world, it is messy, jumpy, sometimes cumulative, sometimes contradictory or doubling back. Its double view, both feminist and international, is understood and pursued through very different disciplinary, politicogeographic, and cultural lenses. It is not possible, then, to propose a future research agenda. For my own interests, though, I am especially interested in the move to internationalize the agenda, to generate more feminist accounts from outside the usual centers of power, and to do so in ways that combine big questions around globalization and international power relations with careful tracing of the gendered, cross-cultural, and lived experiences of the everyday. These can never be constrained entirely within the discipline; they will continue to be enriched by travels beyond its boundaries.

Notes 1. Though see Christine Sylvester on this problematic, and contested, notion. 2. For a sample copy, send an e-mail message to [email protected] with the word info in the body of the message. 3. Jindy Pettman, “Gendering Australia's Foreign Policy,” in Issues in International Relations for the Twentieth Century: An Australian Perspective, ed. W. Tow and M. Hanson (Sydney, Australia: Oxford University Press, 2001).Page 311 → 4. For example, Sandra Whitworth, Feminism and International Relations (New York: St Martin's Press, 1994); L. H. M. Ling, “Feminist International Relations: From Critique to Reconstruction,” Journal of International Communications 3, no. 1 (1996): 26–41. 5. J. Ann Tickner in this volume. 6. Exceptions include Philip Darby, ed., At the Edge of International Relations: Postcolonialism, Gender, and Dependency (London: Pinter, 1997). 7. V. S. Peterson, “Reframing the Politics of Identity: Democracy, Globalisation, and Gender,” Political Expressions 1, no. 1 (1995): 1–16. 8. For my take on feminist international studies to the mid-1990s, see Worlding Women: A Feminist International Politics (New York: Routledge, 1996). For recent writings see V. S. Peterson, “Feminisms and International Relations,” Gender and History 10, no. 3 (1998): 581–89; Terrell Carver, “Gendering International Relations,” Millennium 27, no. 2 (1998): 343–51; recent ISA programs and issues of IFjP. 9. Cynthia Enloe's groundbreaking text Bananas, Bases, and Beaches: Making Feminist Sense of

International Politics (London: Pandora, 1990). 10. J. Ann Tickner in this volume. 11. Robert Crawford and Darryl Jarvis, eds., International Relations—Still an American Social Science? (New York: State University of New York Press, 2001). While considering feminist IS, we might add Still a Social Science? See J. Ann Tickner in this volume. 12. J. Ann Tickner's contribution refers to the work of Katherine Moon and Christine Chin in this context; see also Lily Ling's work. 13. Marysia Zalewski and Jane Parpart, eds., The “Man” Question in International Relations (Boulder, Colo.: Westview, 1998); Carver, “Gendering International Relations.” 14. For example, Carol Cohn, “Wars, Wimps, and Women: Talking Gender and Thinking War,” in Gendering War Talk, ed. M. Cooke and A. Wollacott (Princeton, N.J.: Princeton University Press, 1993). 15. V. S. Peterson, “Sexing Political Identities/Nationalism as Heterosexism,” International Feminist Journal of Politics 1, no. 1 (1999): 34–65; Cohn, “Wars, Wimps, and Women” Cynthia Weber, Faking It: U.S. Hegemony in a Post-Phallic Era (Minneapolis: University of Minnesota Press, 1999). 16. Jan Jindy Pettman, “Writing the Body: Transnational Sex,” in Political Economy, Power, and the Body: Global Perspectives, ed. G. Youngs (London: Macmillan, 2000). 17. Zalewski and Parpart, The “Man” Question in International Relations. 18. But see Cynthia Enloe's writings, including Maneuvers: The International Politics of Militarizing Women's Lives (Berkeley: University of California Press, 2000). 19. Jan Jindy Pettman, “Theorizing Gendered Violence” (paper presented at the annual conference of the International Studies Association, Washington, D.C., February 16–20, 1999); Jan Jindy Pettman, “Sukupuoli ja vakivalta rauhantutkimuksessa” (Theorizing gendered violence), Kosmopolis (special issue on gender and political violence) 30, no. 1 (2000). 20. Lily Ling and Christine Sylvester in this volume. 21. “A Feminist Perspective on ‘Australia in Asia,’” in Race, Colour, and Identity in Australia and New Zealand, ed. J. Docker and G. Fischer (Sydney, Australia: University of New South Wales Press, 2000); “Identity and Security in Australia-Southeast Asia Relations,” in Security, Community, and Emancipation, ed. K. Booth (Boulder, Colo.: Lynne Rienner, 2002). 22. For example, the “Forum on Teaching: Globalising Women's Studies” in International Feminist Journal of Politics 1, no. 2 (1999): 256–98.

Page 312 →

“PROGRESS” AS FEMINIST INTERNATIONAL RELATIONS Christine Sylvester One of the questions put to us by the organizers of the millennial panels at the 2000 International Studies Association (ISA) conference concerns progress: what might we, from our particular positions within a field, say about the progress of international relations (IR) in general? As it happens, I have been ruminating about progress lately, quite independently of this particular query from the ISA. I have done so mostly in the context of productions of women and progress in Zimbabwe1 but also with respect to feminist IR and whether it has brought us—and who is that?—progress. “Progress” is an odd thing. It twists my head around. Feminists require “it” for the advancement of “women.” Much of IR tries to reach “it” cumulatively, using the right methods harnessed to the right research programs—now neorealism, now neoliberal institutionalism, now constructivism, now rationalism, now … I delight in feminist IR as a sign of an avant-garde movement making waves and making a certain type of progress in the IR field possible. Simultaneously, knowing that some avant-gardes have been described as working to extricate themselves from the distractions of women,2 the double edges of avant-garde progress also come into view. Indeed, progress is a cre(a)p and a leap, a jump, and a crawl. It strives for and against, thrusts, moves, makes a crumble, flows. It is a fact, a sorrow, a make-believe, a virtuality, a compromise, a cynicism. Progress advances. It stands us still. It writhes. It invites nostalgia. It is at once a common notion, easily grasped by the modern mind, and something difficult to understand, make happen, or repudiate. Progressive deterioration even bespeaks progress. Tatters talk. Yet there is movement in life and in theory—around and back, in new circles. Certainly there is process, rhythm, and change, the unfolding of potential and intent, simultaneous destruction and renewal. For some, there is steady or at least erratic improvement in their lives, which we often think of as progress: the advance toward more promise for tomorrow than exists today or was possible yesterday. For others, progress is someone else's news today, yesterday, and tomorrow. Page 313 → Consider a few shades of “progress.” A “miracle” precipitously lifts standards of living in some parts of the world, prompting predictions of an upcoming Asian century. In 1997 the lift deflates somewhat, and we ask ourselves how long miracle progress sustains. At the other end of things, only a miracle could bring recognizable progress to certain territories of intractable conflict. In Angola, surviving is a form of advance. A circuitous route to the status quo ante can also mark progress. A Haitian pays the last bit of money for passage to Miami—to the city of progress—but the rickety vessel is intercepted and turned back. The traveler manages to elude arrest and to survive the seas, and progress occurs in the simple fact that s/he has not died. In another case, people adjust to someone else's sense of progress when development feminists turn up at the village toting a water pump (the local women would prefer a school building). And then, of course, there is progress on a Hollywood/Washington scale: Toni Morrison has it that Bill Clinton was the first black president of the United States (and after that an election produces No-body as president of the United States). Now we're really getting somewhere? Life is ever in progress. Whether it progresses in positive or negative or in-between ways, when, where, and how is an enormous issue. The term eludes definitional clarity. The jury, it seems, is still out. We face a new millennium without a name for “it,” or with a name that is everywhere known and there/not there. Is it not rather risky, therefore, to talk about progress … in IR?

Elshtain and Enloe Progress IR Feminist international relations poses an unfinished journey (some progress in progress, you might say) around and about the field of IR.3 The voyagers set off in earnest in the 1980s, when feminism was in some quandary about Marxist, radical, liberal, and socialist women questions and pitched to new queries around the existence, meaning, discourse, and variety of “women.”4 In that decade, IR came under the influence of Hedley Bull's The

Anarchical Society: A Study of Order in World Politics and Kenneth Waltz's Theory of International Politics. Both tomes put emphasis on anarchies, norms, and modes of conflict and cooperation over agents of them rather than people.5 Toward the end of the 1980s, feminism and IR made a linkup around Cynthia Enloe's Bananas, Beaches, and Bases: Making Feminist Sense of International Relations and Jean Bethke Elshtain's Women and War.6 Then, with rapid progress, came the lively Feminist Theory and Gender Studies section of the ISA in 1990 and an ever-diversifying set of topics defining the feminist IR genre. For the most part, conventional IR did not seize the moment to tie itself to feminism in a field-defining or field-debating way, which is not to say that there was no progress observed. Elshtain and Enloe certainly did their best to advance “our field.” They established themselves as innovators inclined toward framing questions of international relations around feminist concerns rather than around IR agendas. What a difference a starting point would make! Feminists generally Page 314 → begin with gender, bodies, sexuality, difference, voice, patriarchy, subjectivity, representation, and/or travel theory.7 Enloe wonders where women are in a world that keeps maneuvering patriarchy to the fore.8 Elshtain finds narratives of soldierly and motherly bodies intriguing points of entry into the puzzle of war and its sexed stories; she is also on about citizenship and civic virtue and the ways America needs to reflect on itself better in order to better engage the world.9 Meanwhile, conventional IR still typically begins with states, war, trade, regimes, anarchy, and/or rationality. It maps relations of force, power, decision, threat, and agreement. Much of it endeavors these days to sort out the rationalities, construct the codeterminations, find the elusive international society, demarcate globalization, or map the ethics. Working late into the night on such topics, there are inevitable spills of cold coffee on a world atlas that is already stained with burger droppings. There goes Bhutan. The Maldives and Bangladesh smudge, their cities becoming illegible. No worries for IR the Abstract. No power is at stake in those places anyway. Such spills do not go unnoticed by feminists interested in international relations; indeed, some among them can purposefully cause IR to go missing in the mop up. Elshtain and Enloe write of wars and bases, which (unlike bananas and beaches) are items of interest in IR. Yet they nearly ignore IR as they journey, which means that they rarely work from its frameworks or systematically critique its myopias. Both pick a few IR themes—for Enloe it is power and for Elshtain, realism—needle the field about the flaws in its preoccupations, and then advance their own ideas with dispatch. The results startlingly recompose the world as a question mark in the study of “women.” Here is a form of scholarly progress that takes back the people-alienating night of IR. As they venture into that abyss, feminists tarry at atypical sites of IR research. Elshtain enters Mrs. Von Clausewitz's parlor in the evening light, where confessions of an unknown wife publishing her husband's work enliven the atmosphere. Enloe takes us to the night laundry of a Jamaican tourist hotel or runs the track telling stories about the gendered history of sneakers. In such IR-estranged places feminists locate events, histories, forms of agency, and people far from great-power concerns that nonetheless have the audacity to contribute to (and therefore have place within) international relations. Hans Morgenthau famously declared autonomy from nonrealist standards of sociality. Enloe relies on those “other standards” to spotlight women at the center of certain realms of international relations. She spies Carmen Miranda, for example—the showgirl with all those exotic fruits on her head—as an international actor linked to corporate marketing strategies, to America's Good Neighbor Policy of yesteryear, and to issues of land reform in Latin America. Starting from one glamorous but hardly heroic woman of the world from south of the North, Enloe unravels a series of relations international that IR misses. She and others refuse grandiose sweeps of the world, ignore Waltz's caution about the importance of remaining above the state level of analysis on a systems footing, and typically do not think about strategic games that no particular body plays.10 Page 315 → Enloe and Elshtain are unique—they are the innovators who first brought IR questions to feminism.11 Not all feminists thereafter roaming international relations have taken up their feminist-first agenda. Some prefer starting with a recognizable IR topic and bringing to it new gender-highlighting questions,12 which lately telescope feminist angles on international political economy, foreign policy, war, peace, globalization, the state, nationalism,

trade, and so forth.13 Nonetheless, most feminists engaged with international relations share interest in ethnographic and anthropological methods of analysis, with human geography replacing geopolitics and bottomup research transplanting abstractions. Quantitative analysis, formal modeling, gaming—there are few followers of these approaches in the feminist corner of “our field.” Feminists explore the lower-than-low politics that comprise, write, and enact relations international alongside the heroics. Such research maneuvers provincialize IR, to borrow an out-of-context thought from Dipesh Chakrabarty,14 by making the usual field concerns a stage on which other actors perform compelling and salient plays. A brief change of (mise-en-) scene illustrates the types of performances that feminists observe, enact, and bring into being.

Interlude We are standing at a painting, oil on board, framed. Unlike Helen Brocklehurst's painting of IR,15 the one that eyes me is very small. It is also abstract and seemingly an anarchy of blacks on black, like Robert Ryman does whites. In the corner is a drip of yellow. Imagine that Deborah Moggach writes a novel around this painting and calls it Power Fever.16 Imagine that her story starts something like this: “I am seated across the table from a man who has spinach caught between his teeth, but who does not know the meaning of embarrassment. He fills the entire dining room with his smoke and fire and words. He neither notices that I am noticing nor assesses himself before standing and solemnly progressing to the heavy front door. He is going out to discuss the small black painting (yellow edging) with important colleagues. None of them will mention the spinach, I know, so I rush forward to spare him from the embarrassment I realize he is not capable of mustering. He smiles at me benignly and bows out. The spinach I have tried to dislodge has been jostled but holds stubbornly, crookedly to its place. The man arrives at the club and takes his seat among others like himself. No one notices his spinach because all have a bit of it caught in their own teeth.” Intervention by this field-guarding feminist is foiled by patronizing phantoms of sophistication. The stage set changes to Langelinie Harbour. Abronze woman's bare back is barely visible among the rocks. “She” is bent away from all the industrial and military installations across the water, turned against ferries that ply waters to Malmo and that also bring visitors to take a look at her. An electric train travels in the background, humming. She does not hum back. She does not look at us looking, does not appear to hear the din or to record the significance of herself. Yet she pulls us toward her as though she knows us and our sounds. “She” is tiny and young and dignified and odd and not modern Page 316 → and a study in postmodern nostalgia. Her eyes downcast, her body pawed at, leaned against, photo-opped, she evades the myth of her docility; for it is clear that she commands the harbor and its spectators, installations, ferries, and international visitors. She is what people come to see as Copenhagen, no matter that local art critics and some feminists decry her old-fashioned style, her docile image of woman, her crass popularity. Gentle she is and unwavering, even with her head chopped off now and then. She is there winter after summer in unheroic pose and nonmonumental leadership. She, the creature, has achieved a divorce from creators Edvard Eriksen and Hans Christian Andersen. She has been written/has written herself as determined vulnerability rather than Jeanne D'Arc bravery or ladylike weakness. They say she wants a soul. That's why she is made to sit there with her smooth body in the land of man. Won't some nice young man marry her and thereby give her a soul? That would be progress in a story that has it that soul is a man's to give and that it makes a girl a true human. Yet it is she who is there long after the man passes by, or drools over her, or dresses like her, or whatever it is he does there. Soul passes round, and she is doing the passing. Who has it, who wants it, and who gives it to whom? And who says? Who says! That is one bunch of devilish queries that a keeper of untidy questions of international relations in feminism can ask but that the feminist questioner of IR might not wish to ask or be ignored for asking/answering by an audience that suffers power fever.

More Progress Ahead As we gaze at mermaid we notice that she has no clothes for the impending Danish winter. She is not fussed, though, because her world persists whether others choose to notice her and help her out or not. In fact, we who gaze at her are the ones written by her authorial capacity to command our attention. Outlined in an unlikely place,

national symbolism inclined against the panorama of ships and tanks, buskers, and lookers, mermaid is a living sculptural story of complex gender shaping the world. Surely she is an international relation of ours in that veritable carnival where she dwells—out in the spectacle of beasts, sorcerers, and average people. Mermaid and entourage bring that carnival to feminist-minded readers who journey to ask questions and to read spectacular interpretations.17 Where else can we go to relations international that have been neglected in IR? Myself, I like to take to the field for close-up sightings. By “field” I mean not only the usual destinations of IR case studies—Russia, France, the International Monetary Fund (IMF), the State Department—where handmaids may lurk uncited in power.18 I also mean the fields we currently fence off and call the Humanities. I recommend forays into novels, choreography, drama, poetry, music, and visual arts. These places of fiction, expression, and emotion assist those of us trained in IR to escape narrow vision in order to see other international actors, interactions, products, exchanges, and dramatizations. If we work inside the usual IR frameworks all the time, we may find it difficult to imagine worlds beyond or within the one IR has Page 317 → imagined for us. Call it avant-garde or call it finding lost “data,” the advantages of reassigning the arts to social science fieldwork—as vehicles for training in-sight—could be immense. Entering different worlds than those of IR, we might see prisoners, children, mothers, the global goods, petty tyrannies, and murderous attitudes that threaten the relations of many in the world. We will surely stumble onto women and men and transsexuals and texts and terrorism, to say nothing of hyphenators, resisters, laborers, oozers, seepers, and lackers of Phallic access. We will note that the stuff of international relations gets done by women who shoot bullets or women who sew seams for Banana Republic. In rare cases, she is Al(l)bright or Merle Laderman Ukeles, the latter known to wash the floors of a museum of fine arts in a performance of the international-domestic “arts” of women.19 If we keep journeying, we might sight signs put up by the Gorilla Girl movement asking: Do Women Have to be Naked to Enter the Metropolitan Museum of Art? We can then more easily picture ourselves entering and performing a different script at the ISA than the one we all know so well.20 We can imagine the signs we could install there asking the passerby: Must Feminists Do IR in Order to Do International Relations? Fields of vision might widen even more if we enter archives of the Museum of Modern Art to research Georgia O'Keeffe's pineapple paintings for the Dole Corporation, talk to sex workers in Bangkok about their German clients, hold children of Kosovar refugees in Tasmania, kick up our tourist skirts at the Moulin Rouge, stand among women villagers in Java as they worry about forests burning nearby, or ask women microentrepreneurs in Pusan, Korea, why they advertised IMF sales in 1998.21 It is progress of some sort to find that “women” and the relations of gender that shape international relations come in statistical form, in verse, in camp, in camps, in comics, in mass graves, in frames, in shops, in the World Bank. They hang on the wall, paint the wall, bloody the wall, and line the walls of diplomatic parties.22 As more of us get better at sighting the habitations of international relations, our “findings” will yield heteroskedasticities that “our field” may come to consider its artful central tendency.

The Theory End But. While participating in and applauding such sightings as progressive, I am also thinking about how progress can be gauged at the end of a day in the brothel, on the streets of the export zones, after reading the last paragraph of a feminist IR article, while wondering one more time why Lee Krasner's paintings were so thoroughly backseated to Jackson Pollock's internationally enamored work. My particular worry conjures up old ghosts: Where is, where should be, theory at the end of our travails? Should IR questions in feminism lead us to new feminist theories of international relations? Has it done this? Should research on feminist questions in IR contribute new IR theories: has it done this? Do feminists need to find recurring patterns, as, for example, Enloe's extended work on militarism seems to do vis-à-vis masculinity; or do we work best highlighting anomalies? Page 318 → More questions now. Is the very idea of building feminist IR theory archaic or is it “not yet,” as distinct from “now”—to use terms I like when I hear Chakrabarty using them? Is it that a certain progress is being written by what we do and say more than it is the case that we are consciously writing a field theoretically? Is critical theory

(small c) where the IR questions in feminism and feminist questions in IR should be? Might it be that whatever we theorize eludes definitional clarity once we mix up the disciplines—literary and art theory, postcolonial studies, cultural studies, psychoanalysis, women's studies—to become (feminist) IR? Life is ever in progress, and thus so is the theory side of life. Yet as feminist IR/IR feminist articles accumulate, feminist international relations theories go missing. Feminist IR, impressive as it is, has not capitalized sufficiently on its progressive tendencies by theorizing the connective tissues that bind the genre. It paints more and more pictures that possibly chase away (from) theory—as a thing purposively, progressfully, perhaps, to avoid.23 IR, by contrast, has invested heavily in theory, but of a sort that produces too few pictures of salience to much of the world. Certainly it has failed to cotton onto the sites of gender and women over the years, for reasons, in part, of its chosen locations for seeing the international. Needed is an exercise of theorizing around a millennial question: of what does theory consist at the fulcrum of international relations and feminism? Theory need not be lawlike. It need not be problem solving. What should it, can it, be in this context? If we do not want to do theory or believe we are theorizing in what we are doing—either way, the assumptions must become clearer and subject to more debate. Feminists cannot afford to toss theory to the rationalist and constructivist wings of IR, or to international society traditions that fail to get all the occupants of society counted. We can no longer duck the theory question by hedging it. In taking up a theory challenge we will want to avoid some common errors of avant-gardism, like seeing ourselves as the saviors of international relations, or as progressively more progressive than foremothers like Elshtain and Enloe—thus forgetting where we have been. Feminists will want to say something mermaidenly memorable and mermaidenly elegant in the midst of the theory talk and steer clear of self-celebratory posturing as a stand-in for progress. There are many cautions, many advances, and many lessons about feminist IR/IR feminism that the humble may weep over and the overconfident regret not heeding. Such is the “nature” of progress. We might keep that bit of spinach and that Langelinie mermaid in our sights as the signifying power of differently positioned theory. And then perhaps we can state the unstated. Or will theory power give us self-induced fever—to ask nothing of whether Angolan villagers and Haitian boaters will survive while more progressive moments of a thing called IR unfold?

Notes 1. See, for example, Christine Sylvester, “‘Progress’ in Zimbabwe: Is ‘It’ a ‘Woman’?” International Feminist Journal of Politics 1, no. 1 (1999): 89–90; and Christine Page 319 → Sylvester, Producing Women and Progress in Zimbabwe: Narratives of Identity and Work from the 1980s (Portsmouth, New Hampshire: Heinemann, 2000). 2. John Docker, Postmodernism and Popular Culture: A Cultural History (Cambridge: Cambridge University Press, 1994). 3. Christine Sylvester, Feminist International Relations: An Unfinished Journey (Cambridge: Cambridge University Press, 2002). 4. Denise Riley, “Am I That Name?” Feminism and the Category of “Women” in History (Minneapolis: University of Minnesota Press, 1988); and Judith Butler, Gender Trouble: Feminism and the Subversion of Identity (New York: Routledge, 1990). 5. Hedley Bull, The Anarchical Society: A Study of Order in World Politics (New York: Columbia University Press, 1977); and Kenneth Waltz, Theory of International Politics (Reading, Mass.: AddisonWesley, 1979). 6. Cynthia Enloe, Bananas, Beaches, and Bases: Making Feminist Sense of International Relations (London: Pandora, 1989); and Jean Bethke Elshtain, Women and War (New York: Basic Books, 1987). 7. See Chris Weedon, Feminism, Theory, and the Politics of Difference (Oxford, England: Blackwell, 1999); and Sandra Harding, Is Science Multicultural? Postcolonialisms, Feminisms, and Epistemologies

(Bloomington: Indiana University Press, 1998). 8. Among Cynthia Enloe's feminist IR books are The Morning After: Sexual Politics at the End of the Cold War (Berkeley: University of California Press, 1993) and Maneuvers: The International Politics of Militarizing Women's Lives (Berkeley: University of California Press, 2000). 9. See Jean Bethke Elshtain, Real Politics: At the Center of Everyday Life (Baltimore: Johns Hopkins University Press, 1997); and Jean Bethke Elshtain, New Wine and Old Bottles: International Politics and Ethical Discourse (Notre Dame, Ind.: University of Notre Dame Press, 1998). 10. Hans Morgenthau, Politics among Nations: The Struggle for Power and Peace (New York: Knopf, 1966); Waltz, Theory of International Politics; Robert Keohane, After Hegemony: Cooperation and Discord in the World Political Economy (Princeton, N.J.: Princeton University Press, 1984); Robert Axelrod, “The Emergence of Cooperation among Egoists,” American Political Science Review 25 (1981): 306–18. 11. This phraseology is obviously beholden to Sandra Harding's The Science Question in Feminism (Ithaca, N.Y.: Cornell University Press, 1986). 12. J. Ann Tickner's Gender in International Relations: Feminist Perspectives on Achieving Global Security (New York: Columbia University Press, 1992) is a classic in this vein. 13. For example: Marieke de Goede, “Mastering ‘Lady Credit’: Discourses of Financial Credit in Historical Perspective,” International Feminist Journal of Politics 2, no. 1 (2000): 58–81; Nancy McGlen and Meredith Sarkees, Women in Foreign Policy: The Insiders (New York: Routledge, 1993); Carolyn Nordstrom, Girls and Warzones: Troubling Questions (Uppsala, Sweden: Life and Peace Institute, 1997); Eleanor Kofman and Gillian Youngs, eds., Globalization: Theory and Practice (London: Pinter, 1996); Gender and International Relations in a Nordic Context, special issue of Cooperation and Conflict 36, no. 2 (2001). 14. Dipesh Chakrabarty, Provincializing Europe (Princeton, N.J.: Princeton University Press, 2000); also Dipesh Chakrabarty, “Postcoloniality and the Artifice of History: Who Speaks for ‘Indian’ Pasts?” Representations 37 (winter 1992): 1–26. 15. Helen Bocklehurst, “Painting International Relations,” International Feminist Journal of Politics 1, no. 2 (1999): 314–23. 16. Deborah Moggach, Tulip Fever: A Novel (London: Heinemann, 1999). 17. This is one version of the thesis advanced by Minrose Gwin in “Space-Travel: The Connective Politics of Feminist Reading,” Signs: Journal of Women in Culture and Society 21, no. 4 (1996): 870–905. 18. Christine Sylvester, “Handmaids' Tales of Washington Power: The Abject and the Real Kennedy White House,” Body and Society 4, no. 3 (1998): 39–66.Page 320 → 19. For a discussion of Ukeles, see Lucy Lippard, The Pink Glass Swan: Selected Feminist Essays on Art (New York: New Press, 1995). 20. In fact, for the Pan European Conference on International Relations held in Canterbury in September 2001, Costas Constantinou organized sessions to perform IR. 21. See the IR-relevant feminist work (albeit not strictly IR feminist work) on these and related subjects by Thanh-Dam Truong, Sex, Money, and Morality: Prostitution and Tourism in South-East Asia (London: Zed Press, 1990); Gloria Anzaldua, ed., Making Face, Making Soul: Haciendo Caras (San Francisco: Aunt Lute, 1990); Chandra Mohanty, Ann Russo, and Lourdes Torres, eds., Third World Women and the Politics of Feminism (Bloomington: Indiana University Press, 1991); Caren Kaplan, Norma Alarcon, and Minoo Moallem, eds., Between Woman and Nation: Nationalism, Transnational Feminisms, and the State (Durham, N.C.: Duke University Press, 1999). 22. For some “strange” recent sitings, see for example Lisa Bloom, ed., With Other Eyes: Looking at Race and Gender in Visual Culture (Minneapolis: University of Minnesota Press, 1999); and Poetic International Relations, special issue of Alternatives: Humane Governance and Social Transformation 25, no. 3 (2000). 23. See discussions in Judith Butler and Joan Scott, eds., Feminists Theorize the Political (New York: Routledge, 1992); and Anne Phillips, ed., Feminism and Politics (Oxford: Oxford University Press, 1998).

Page 321 →

FEMINIST THEORY AND GENDER STUDIES Reflections for the Millennium J. Ann Tickner A forty-year retrospect is a difficult challenge for feminist theory and gender studies. For thirty of those years there was almost total silence on gender issues in international relations (IR); women were barely visible as scholars and foreign policy practitioners, as well as in the subject matter of the discipline. The invisibility in the discipline was not because gender was, and for some still is, irrelevant but because international relations, in both its theory and practice, was so thoroughly gendered that, in Cynthia Enloe's words, no one noticed that women were missing.1 So clearly, feminist approaches to IR have come a long way since the late 1980s when, after several conferences at which feminist and IR scholars began to talk together about these issues, feminist IR research began to appear.2 Probably the most significant accomplishments have been the enormous productivity of a relatively small group of scholars and their success in creating a supportive space within the International Studies Association (ISA) within which this type of work is being pursued.3 Besides an impressive list of publications and a growing number of panels at the ISA, we have recently witnessed the inauguration of the International Feminist Journal of Politics. There has been a genuine attempt to recognize a diversity of voices and approaches on feminist ISA panels. And a volume reflecting the state of the discipline at the millennium that acknowledges this work is clearly an enormous step for a discipline that has had trouble understanding why gender should be relevant to its scholarly concerns. My first thought on being asked to write these reflections was how difficult a critical assessment of such a young field would be. While there are differences amongst IR feminists, engaging in self-criticism seems premature as feminist approaches still struggle to be heard in a discipline that has not been particularly open to gender approaches. I have also been struggling with a more fundamental problem as I attempt to formulate responses to the questions posed to the participants in this millennium project—namely that these questions are not ones that many feminists would ask when engaging in self-evaluation. Words such as “state of the art” and “cumulation,” Page 322 → as well as the selection of certain scholars to speak on behalf of the rest, are in tension with an approach that celebrates diversity in terms of its subject matters, normative focus, and methodologies. Asking contributors to compare feminist approaches with realism, for example, is like asking for a comparison of a whole array of theoretical and methodological approaches that cross the disciplines, from the humanities to the social sciences, with one IR approach that, for the most part, has been located within the discipline of political science. Nevertheless, I welcome this opportunity to elaborate on some “progressive” paths in IR feminism and to identify some of the important research questions that have guided feminist scholars in the past and will probably continue to do so in the future. I will also engage in some self-criticism; to this end I will conclude by raising some issues relating to methodological differences amongst IR feminists. First, some comments on the questions themselves. We have been asked to assess the degree of “integration” and “cumulation” in our field, as well as to decide which core research questions might be the most promising for the future. Such an assessment is framed in terms suited to a social scientific discipline that strives to emulate the universalistic theory-based knowledge cumulation model of the natural sciences. Since many of those who identify themselves as IR feminists do not use social scientific methodologies, preferring instead methodologies more compatible with hermeneutic, interpretive epistemological traditions, they would be unlikely to use these standards of evaluation. It also assumes a particular definition of the meaning of theory. Marysia Zalewski has identified three types of theory: theory as a tool for understanding the world; theory as critique, or understanding how the world got to be as it is so that it can be changed; and theory as practice in which people engage as they go about their everyday life.4 Conventional IR theorists, by which I mean scholars who conform to positivist methodologies broadly defined, usually employ theory in the first sense, as a tool to develop explanations for the political and economic behavior of states in the international system. For those who

define theory in this sense, its separation from political practice and, as far as possible, from the values of the researcher are thought to be important goals. The reflections and evaluation we have been asked to undertake are quite appropriate ones for those who use theory as a tool to develop better explanations. Feminist approaches, however, are more likely to define theory in Zalewski's second or third sense, which means that they would be likely to use different standards of evaluation. For many feminist theorists, knowledge construction is explicitly linked to emancipatory political practice, primarily that of transforming the unequal power relationships between women and men. Sandra Whitworth has claimed that contemporary feminism has its roots in social movements; feminism is a politics of protest directed at transforming the unequal power relationships between women and men.5 Therefore, a key goal for IR feminist theory used in this sense is to understand how the existing social order, one many feminists believe is marked by discrimination and oppression, Page 323 → came into being and how this knowledge can be used to work toward its transformation. For many IR feminists, therefore, knowledge is practical and explicitly normative. Christine Chin has claimed that these emancipatory concerns suggest the need for restructuring the ways in which we conceive and execute research problems; she argues for undoing received disciplinary and epistemological boundaries that segregate the pursuit of knowledge.6 Claiming that the kind of practical knowledge they seek emerges from political practice, many feminists do not believe in, nor see knowledge as requiring or benefiting from, the separation between theory and practice. Theory as practice, Zalewski's third definition of theory, means that we need to take into account many more human activities than would be thought necessary by those who use theory as a tool. The goal of this type of practical knowledge is not the improvement of theory but of practice; thus, it requires different standards of evaluation. Explicitly rejecting the separation between observers and observed, considered desirable by practitioners of the first definition of theory, practical knowledge is intended to yield greater understanding of people's everyday lives in order to improve them. Given these very different views of both the meaning and goals of theory, it is unlikely that feminists would use “scientific” criteria, such as integration and cumulation, to evaluate them. Hilary Charlesworth has described feminist methodology as an “archaeological dig” where different methods are appropriate at different levels of excavation.7 She claims that no single method is sufficient for analyzing complex social phenomena. Since feminists are using tools that are rarely included in a standard IR methodological training, their methodologies tend to be eclectic. There is a sense that research should be grounded, whether it is in people's everyday activities or in the close reading of texts that can offer interpretations about how people construct their world and, therefore, act upon it. While it is important to do research that is linked to concrete historical cases, it is also important to rethink the theoretical assumptions that led to consideration of these cases in the first place. Defining gender as a social construction of inequality and preferring critically interpretive, hermeneutic epistemologies and a variety of related research strategies, such as ethnography and discourse analysis, for discovering these structures of inequality makes it difficult to assess feminist theoretical approaches primarily in terms of explanatory power. As mentioned before, most feminists celebrate difference and resist synthesis, positions that are more acceptable in the humanities than in the social sciences. Comparing feminist IR with other IR approaches that share a single set of assumptions and methodological tools is also difficult. IR feminists have drawn on a variety of feminist theoretical approaches variously described as liberal, radical, psychoanalytic, socialist, postmodern, and postcolonial. Besides seeking better understanding of women's subordination, most of these approaches see themselves as politically engaged in the practical tasks of improving women's lives. While liberal feminists have generally relied on empiricist methodologies, other approaches have questioned these positivist methodologies. Arguing from standpoint or postmodern Page 324 → epistemological positions, they claim that “scientific” theories that admit to the possibility of neutrality of facts and a universalist objectivity hide an epistemological tradition that is gendered and built on knowledge mostly about men. These different feminist approaches range across the disciplines, from the natural to the social sciences and the humanities. Forging these transdisciplinary bonds and respecting and using different methodologies is a position to which many feminist scholars aspire.

The editors' assessment of international studies at the millennium speaks of clusters of separable identifiable research questions such as globalization, international institutions, security policy, and gender and international relations as one of the few, but inadequate, signs of cumulation in the discipline. Feminists would probably question the labeling of gender and international relations in these terms; rather than identifying gender and international relations as a separable (and potentially cumulative) research program, a better measure of success would be when those studying each of these realms of activity recognize gender as intrinsic to their subject matter. Rather than defining itself as a separable body of research, feminist IR has ranged across all these research domains (and others) asking how gender as a category of analysis can improve our understanding of them. Feminists have suggested that gender considerations are vital to how we understand security, the construction of international institutions, economic globalization, human rights, democratization, and many other issues in international relations. To this end, feminists point to gender in international relations rather than gender and international relations. In spite of my reservations about the adequacy of these questions for evaluating feminist research, I have found them useful in reflecting on some of the accomplishments and limitations of feminist IR. First, there has been progress although defined in rather a different way. To assess this progress, I prefer to draw on the work of Peggy McIntosh, who outlines five phases of curriculum change necessary for introducing gender into the scholarly disciplines.8 The first phase is what McIntosh describes as a womanless world; this type of analysis describes only the activities of those holding high positions of power, usually in dominant states; much IR literature is still at this stage. Phase two notes the absence of women and adds the “famous few,” such as Margaret Thatcher and Madeleine Albright, to the curriculum; several IR textbooks have begun to do this. While the addition of the “famous few” provides role models for women, it does nothing to change the discipline in ways that acknowledge that anything different can be learned from women's experiences. In phase three, the absence of women is seen as a problem as we begin to understand the politics implicit in a curriculum constructed without the inclusion of women's experiences; in this phase, women are typically seen as victims. Moving to phase four involves seeing women as valid human beings whose various life experiences have shaped the world in which we live, even though their contributions involve tasks that are often unacknowledged. The final phase of McIntosh's curriculum development brings us to the point where the subject matter of the discipline genuinely includes the experiences of all individuals regardless of race, culture, Page 325 → class, and gender. Such a discipline, which includes us all, would require a radical redrawing of the boundaries of its subject matter. While phase five has not yet been realized and while much of international relations scholarship remains at stages one and two, feminist IR has generally progressed to stages three and four. Scholarly progress in feminist IR has involved moving from studies, informed by a liberal empiricist approach, that try to fit women into existing paradigms and frameworks to work that acknowledges the centrality of gender as a category of analysis for explaining, among other things, why women have so often been invisible in international politics. There has also been much more attention paid to race, class, and geographical and historical location, issues that have not received much attention in the discipline as conventionally defined. Challenging a discipline that was, for so long, silent about women, feminist IR has evolved, as has feminist theory more generally, from early studies of women and development and women and the military to more recent empirical studies that situate women within gendered structures of power. For example, Kathy Moon has challenged us to recognize the plight of prostitutes serving the U.S. military in Korea and to understand how their lives are embedded within the broader framework of U.S.-Korean security relations.9 In her study of prostitution around U.S. military bases in South Korea in the 1970s, Moon shows how these people-to-people relations were actually matters of security at the international level. Cleanup of prostitution camps by the South Korean government was part of its attempt to prevent withdrawal of American troops that had begun under the Nixon Doctrine of 1969. Crossing levels of analysis, Moon demonstrates how weakness of the Korean state in terms of influencing the U.S. government resulted in authoritarian sexist control at the domestic level. In other words, national security translated into social insecurity for these women. Moon describes her methodology as ethnographic; her interviews are not intended to offer statistical evidence but “to give voice to people who most Koreans and Americans have never considered as having anything important to say.”10 Moon's stories help us locate women in places not normally considered relevant to IR and to link their experiences to

wider gendered processes and hierarchical structures of power from the local to the international levels. Likewise, Christine Chin has demonstrated how overseas domestic workers in Malaysia are supporting the Malaysian state and the political economy of the East Asian region to the detriment of their own security.11 The Malaysian state has needed Malaysian women to fill labor shortages in the industrial sector; thus, foreign workers are being used in the home for family-related tasks, and modernization is being achieved at the expense of the working poor, many of whom are immigrants left unprotected and often working under extremely harsh conditions. Describing her research as ethnographic, Chin rejects the survey method because it oversimplifies the complexities of life, which, she claims, cannot be distilled in a series of hypotheses to be tested. She describes her work as multimethod ethnographic research and offers field note quotations as evidence that allows her subjects to use their own words and the right to speak about any issue they Page 326 → pleased. Chin describes her analysis of her interviews as a study of narrativity or how we come to construct our identities by locating ourselves within our life stories. Narrative is a method sometimes used by feminists to further their goal of constructing practical knowledge that comes out of people's everyday life experiences. Such work grows out of multiple sites and multiple perspectives. In other words, feminist scholars are challenging us to look at IR from the margins, to ask questions not normally thought to be relevant to the discipline, and to use different methodologies to answer them. Frequently, investigations begin at the level of the individual, or level one, and move up to levels two and three, the state and the international system. IR feminists believe that such investigations can also help us to understand that IR's levels of analysis are not discrete concepts but mutually constituted. Using theory in Zalewski's third sense, Jane Mansbridge has claimed that if good theory evolves from practice and from seeing with a feminist eye, then it will never be possible to predict the most important future issues in feminist theory.12 Nevertheless, we can identify some broad research questions that have motivated IR feminists over the past ten years and that will probably continue to do so in the future. As mentioned before, they are questions that cross issue areas and perspectives. One question with which feminists often begin is “where are the women?” Finding women involves looking in places not usually investigated in international relations. Acknowledging that we need to look in unconventional places not normally considered within the boundaries of IR, Cynthia Enloe has asked whether women's roles, as secretaries, clerical workers, domestic servants, and diplomats' wives, are relevant to the business of international politics.13 Moon, Chin, Enloe, and many others have demonstrated that women in their various roles as domestic servants, prostitutes, and homeworkers are relevant and vital to the business of international politics. Locating women must include placing them within gendered structures. Typically, feminist research questions have to do with investigating how the international system and the global economy contribute to the subordination of women and other subjugated groups. IR feminists ask what kind of evidence might further the claim that the practices of international politics are gendered. Through what mechanisms are the types of power necessary to keep unequal gender structures in place perpetuated? To fully understand the role of gender and other social hierarchies, questions about men and masculinity are also beginning to be asked. For example, does it make any difference to states' behavior that their foreign and security policies are so often legitimated through appeals to various types of hegemonic masculinity? It is important to emphasize that these feminist questions are often investigated with emancipatory goals in mind and with the understanding that the choice of questions we ask is never neutral. Feminists will continue to push the boundaries of the discipline, question our understanding of “common sense,” and hypothesize a world that could be otherwise. Questioning the way we have come to understand the world as well as the forms of power necessary to sustain dominant forms of interpretation demands quite different Page 327 → methodologies from those generally used by conventional IR. Recovering the experiences of subjugated people demands methods more typical of anthropology and sociology than political science. Given these goals, typically these and other feminist questions are investigated using critical rather than social scientific methodologies; many IR feminists, myself included, have claimed that positivist methodologies are inadequate for building a body of knowledge that takes gender identity and hierarchical social relations as its central framework and that acknowledges the mutual constitution of theory and practice. However, while many IR

feminists have rejected social scientific methodologies, the introduction of feminist approaches into the discipline has also stimulated a growing body of research on women in international politics that draws on these more conventional methodologies. Indeed, there is a long tradition of this type of research in American politics, dealing with issues such as voting, public opinion, and the advancement of women in different arenas of public and private life. An important (and uncomfortable) question for many IR feminists is whether our tolerance (and celebration) of difference can extend to this type of work. In international relations, feminists writing from a social scientific perspective have investigated the role of gender in foreign policy, particularly the evidence for a fairly substantial gender gap, in the United States at least, in attitudes about war and peace. Women in the United States have consistently shown less support for forceful means of pursuing foreign policy goals than men, and this gender gap continues to grow.14 This leads to questions that are also being asked as to whether the presence of more women in elite foreign policy-making positions would make any difference in the conduct of foreign policy. For example, Mary Caprioli has quantitatively tested the relationship between state militarism and domestic gender equality; her study shows that domestic gender equality has a pacifying effect on state behavior on the international level.15 Another study that uses public opinion data on attitudes toward the Arab-Israeli conflict found a strong positive correlation between attitudes toward support for equality of women and support for diplomacy and compromise.16 The literature on the democratic peace has made these questions particularly compelling for scholars working on these issues. Lisa Brandes, a scholar who does empirical research on the gender gap in popular attitudes, has suggested that I, amongst others, have dismissed these types of research questions largely on methodological grounds.17 She argues that rather than dismissing a research program out of hand, it is incumbent upon IR feminist scholars to do the hard investigative work to find out what the results will be, before making claims about their likely value. Yet, since women do not come close to parity in any state in terms of occupying positions of power in foreign policymaking, I remain skeptical of efforts to explain the foreign policy orientations of liberal states in these terms. Given the pitfalls of McIntosh's stage two analysis discussed earlier, I am also suspicious of attempting to understand gender hierarchies by adding women to existing paradigms and research traditions. Nevertheless, I do think that IR Page 328 → feminists in postpositivist traditions should consider how to engage more seriously with this type of work. While bridging the methodological divide between scientific traditions of explaining and hermeneutic traditions of practical understanding is difficult and, according to some, impossible, feminists need to talk more, albeit acknowledging their differences, across these theoretical divides.18 In a relatively short span of ten years, feminists have challenged the gendered foundations of IR; they have introduced new issues and brought new questions into a discipline in which women were largely absent. In order to answer their questions, many IR feminists have ranged far from the discipline as conventionally defined and drawn on methodologies unfamiliar to many in IR. They have challenged disciplinary boundaries and methods that, they believe, impose limitations on the kinds of questions that can be asked and the ways in which they can be answered. However, these transdisciplinary excursions and methodological innovations have consequences. Most graduate education in IR, at least in the United States, still occurs within political science departments where training in the kind of critical methodologies used by many feminists may not exist. Moreover, power differences between conventional and critical approaches, which often play out by drawing disciplinary boundaries around subject matter and methods, will continue to render judgment of feminist approaches as less than adequate. Nevertheless, feminists must continue to challenge and critique a discipline that has been quite ethnocentric both in terms of its subject matter and its methodology. At the same time, those who have gone outside the discipline of political science to find tools for understanding global politics must converse with those feminists more conventionally trained. Since a basic assumption of postpositivist feminist IR has been the inadequacy of conventional methodologies for understanding the issues with which it is concerned, can (or should) we be truly tolerant when fundamental questions about epistemologies are at stake? These are difficult questions both for IR feminists and for a discipline much of which judges itself in terms of social scientific standards. At the end of a century that has witnessed the founding and institutionalization of this IR discipline, as well as a great deal of conflict and injustice, much remains to be done. Could we not all agree on the additional, even more pressing need to measure the progress of feminist scholarship in terms of its contributions to the betterment of the lives of

women and men?

Notes 1. Cynthia Enloe, Bananas, Beaches, and Bases: Making Feminist Sense of International Politics (Berkeley: University of California Press, 1989). 2. Much of the early work in feminist IR was generated by a series of conferences in the United States and the United Kingdom. Conferences were held at the London School of Economics (LSE) in 1988, the University of Southern California in 1989, and Wellesley College in 1990. Grant and Newland's Gender and International Relations and Peterson's Gendered States were the products of the LSE and Wellesley conferences respectively. Rebecca Grant and Kathleen Newland, eds., Gender and International Relations (Bloomington: Indiana University Press, Page 329 → 1991); V. Spike Peterson, ed., Gendered States: Feminist (Re)Visions of International Relations Theory (Boulder, Colo.: Lynne Rienner, 1992). 3. Some of the early texts include Enloe, Bananas, Beaches, and Bases; Grant and Newland, Gender and International Relations; Peterson, Gendered States; Christine Sylvester, Feminist Theory and International Relations in a Postmodern Era (Cambridge: Cambridge University Press, 1994); J. Ann Tickner, Gender in International Relations: Feminist Perspectives on Achieving Global Security (New York: Columbia University Press, 1992); Sandra Whitworth, Feminism and International Relations (New York: St. Martin's Press, 1994). 4. Marysia Zalewski, “All These Theories Yet the Bodies Keep Piling Up,” in Steve Smith, Ken Booth, and Marysia Zalewski, eds., International Theory: Positivism and Beyond (Cambridge: Cambridge University Press, 1996). 5. Sandra Whitworth, “Gender in the Inter-Paradigm Debate,” Millennium: Journal of International Studies 18, no. 2 (1989): 265–72. 6. Christine Chin, In Service and Servitude: Foreign Female Domestic Workers and the Malaysian “Modernization” Project (New York: Columbia University Press, 1998). 7. Hilary Charlesworth, “Feminists Critiques of International Law and Their Critics,” Third World Legal Studies (1994–95): 1–16. 8. See Peggy McIntosh, “Interactive Phases of Curriculum Re-Vision: A Feminist Perspective” (Wellesley College Center for Research on Women, Wellesley, Mass., 1983). McIntosh is using history as her example, but her analysis could equally well be applied to the discipline of international relations. 9. Katharine Moon, Sex among Allies: Military Prostitution in U.S.–Korean Relations (New York: Columbia University Press, 1997). 10. Ibid., 15. 11. Chin, In Service and Servitude. 12. Jane Mansbridge, “Frontier Research and Crucial Questions for Research in Feminist Theory” (paper presented at the Frontiers of Women and Politics Research Workshop, APSA Women and Politics Seminar, Harvard University, 1998). 13. Enloe, Bananas, Beaches, and Bases, 8. 14. Nancy Gallagher, “The Gender Gap in Popular Attitudes toward the Use of Force,” in Women and the Use of Military Force, ed. Ruth Howes and Michael Stevenson (Boulder, Colo.: Lynne Rienner, 1993), 29. 15. Mary Caprioli, “Gendered Conflict,” Journal of Peace Research 37, no. 1 (2000): 51–68. 16. Mark Tessler and Ina Warriner, “Gender, Feminism, and Attitudes toward International Conflict: Exploring Relationships with Survey Data from the Middle East,” World Politics 49, no. 2 (1997): 250–81. 17. Lisa Brandes, “Research Frontiers in Gender and the International Order” (paper presented at the Frontiers of Women and Politics Research Workshop, APSA Women and Politics Seminar, Harvard University, 1998). 18. This issue is discussed in some detail in Martin Hollis and Steve Smith, Explaining and Understanding International Relations (Oxford: Oxford University Press, 1990). These authors conclude that bridging this

divide is not possible.

Page 330 →

FEMINISM AND/IN INTERNATIONAL RELATIONS An Exhausted Conversation? or

FEMINISTS DOING INTERNATIONAL RELATIONS The Cut(ting) Edge of Contemporary Critical Theory and Practice? Marysia Zalewski

Questions about Questions In common with the other presenters on the millennial reflections feminist theory panel, I found the task at hand something of a challenge. At one level this should not be surprising; writing any academic paper usually is—and should be—a challenge. But there was something about the kinds of questions we were being asked to consider that clearly troubled our group. J. Ann Tickner, for example, suggested that the questions asked were not ones that “feminists would ask when engaging in self-evaluation.”1 Jan Jindy Pettman claimed that “it is a daunting task to respond to the challenges laid down for this panel”2 and further explained that one of the things feminists do is ask different kinds of questions about different things. L. H. M. Ling, through her tale about the fish and the turtle, implied that “old questions” would not tell us anything about “the new.”3 And Christine Sylvester cautioned us to “mind the easy answers” when asking questions about progress within feminism and IR—or indeed IR in feminism.4 Why were the questions we were asked to consider so problematic for us? Surely it makes eminently reasonable intellectual sense to engage in “critical self-reflection” or to “assess where we stand on key debates and address why we have failed to resolve them” or to suggest “what intrasubfield Page 331 → standards we might use to evaluate the significance of feminist theoretical insights”? I think there are at least two main reasons that we found these kinds of questions so problematic. One is that what makes “reasonable intellectual sense” is structurally defined by the traditional or neotraditional center of the discipline, and therefore, by definition, a critical approach such as feminism will have an alternative sense of what counts as reasonable in the realm of the intellectual. A second reason is that we were being asked to make authoritative statements about the impact of the work on feminism and gender in international relations. This is problematic as it relies on the assumption that feminist work in international relations can be fitted into a coherent framework upon which such judgments can then be made. This is not the case as feminism is a vast and sometimes contradictory body of work that makes it therefore difficult (and perhaps inadvisable) to make sweeping and authoritative generalizations about. But the fact that we had “questions about the questions” did not inevitably make the questions we were asked useless or irrelevant! Indeed, the original questions instigated a plethora of other questions, which, in my view, is a positive intellectual practice. For example, when asked to “look back” at the development or trajectory of feminist work within the discipline of international relations over the last ten or more years: Where do we start? Where will we look? How would we know what “difference” feminisms have made? Or, where do we go in the discipline to find the answer to these questions—whether they are about accomplishments, shortcomings, or where we stand on key debates? How do we know what a key debate is anyway? What are the correct answers to all these questions? Of course, there is no such thing as the “correct answer.” It depends on which perspective one starts from and crucially on how one is situated in relation to the core of the discipline. If, for example, one is working in the area of neorealism, the key debates are well defined and indeed well protected. Scholars may disagree over the most appropriate approach to dealing with interstate relationships and have vigorous debates about this. But the value and centrality of the debate itself is well protected by the very self-definition of the discipline. This is not the case

for those who work on feminism and gender. By definition, these scholars not only debate amongst themselves—but crucially we are challenging the self-definition of the discipline. Following from this last point a serious issue needs to be raised that gives some further indication as to one of the reasons we found the questions challenging. This is that the questions asked of the millennial reflections panelists had the potential to consolidate and reinforce some areas of debate in the discipline (the mainstream and neomainstream ones) but significantly also had the potential to render vulnerable those areas that, by their very nature, challenge the status quo.5 For example, being self-critical within and amongst the margins is a very different enterprise with very different issues at stake than being self-critical within the established center. No wonder we found these questions problematic. Nevertheless, the challenge was taken up by the participants of the feminist theory and gender panel, but in Page 332 → ways that I believe did not facilitate a falling into the very gendered space of vulnerability just alluded to. In my own case I was particularly vexed with the idea of having to tell one “story” about feminism and international relations. But I decided that my propensity to see more than one story at a time was something I should work with. As Irigaray suggests, “don't worry about the ‘right’ word. There isn't any.”6 And there isn't any right word or “correct” answer to all the questions asked of us. There isn't any “right” or “truthful” story about feminism and international relations because it depends who speaks the words and who has the authority to decide which voices and which stories are the authentic and legitimate ones. This led me to choose two stories of the many stories that could be told about “feminism and IR.”

Story One: An Exhausted Conversation? As I try and “time travel” back to the late 1980s and “reflect” on my early, formal introductions to the “feminism and IR” debate, two memories demand my attention. The first is of a graduate student speaking at the 1988 London School of Economics and Political Science (LSE) conference on “Women and IR.” She was speaking on the topic of abortion, suggesting that no time limit be imposed on when a woman could have an abortion. There was a stunned silence in the audience. It was clear the student was expecting such a reaction. “Why,” she asked, “why wouldn't we trust women enough to make humane and reasonable decisions in this context?” Is the assumption that removing the legal constraint here would result in a deluge of abortions of near-to-full-term babies? Why would we fear this? Why do we think the law is so necessary to contain women's behavior? I can't remember who that graduate student was7—but I do remember being both shocked and impressed by what she said. The second memory is of a paper given by Sarah Brown at the conference.8 I remember how her paper seemed to sit uneasily alongside many of the other papers at the conference, which were eventually published in a special issue of Millennium.9 Perhaps more pertinently, I recall that Sarah Brown's paper was not included in the followup volume edited by Grant and Newland.10 Maybe one reason Brown's paper “sat uneasily” with many of the others is because it was her view that “a feminist critical theory of International Relations is fundamentally a political act of commitment to understanding the world from the perspective of the socially subjugated.”11 This was quite a radical view at the time (and probably still is) especially as it appeared to be departing from the traditional agenda of international relations in a big way, further, perhaps, than other feminists were prepared to go. Do these two memories have anything to tell us about the contemporary relationship between feminism and IR? As I asked myself this question, I was reminded of a more recent piece on feminism and international relations by Molly Cochran in which she suggests that “the key to the possibility of Page 333 → real dialogue between feminists and the discipline rests on whether or not IR is prepared to accommodate epistemological doubt.”12 What would Sarah Brown think of Molly Cochran's suggestion? Perhaps Brown would insist that questions of epistemology necessarily implicate questions of ontology. Maybe she would suggest that if “real dialogue” were possible between feminists and the discipline, IR would have to accommodate not only epistemological doubt but

also ontological doubt. But even further, she would probably suggest that what is necessary is more than an accommodation and more than simply doubt. Only a radical restructuring of many of IR's epistemological, ontological, and political beliefs will do, alongside the commitment to make real and effective changes that would accompany such a restructuring. But many feminist international relations scholars and international relations scholars then and now have not been willing to embrace the idea or reality of such a radical restructuring. Speaking on behalf of the discipline in 1993, the question I asked of Brown's work was “why would IR scholars want to change the focus of the discipline from war, peace and interstate behaviour and turn their attention solely to issues of gender—thereby reconstructing the starting point of International Relations enquiry?”13 To start addressing this question we would surely have to know where to start. And Where Do We Start? The usual place to start is IR. Consider this encyclopedic entry on feminism and IR: International Relations is a paradigmatically male dominated subject. The academic discipline is overwhelmingly male and the conventional practices of international relations are typically associated with men. War is still regarded as the core topic of international relations. The discipline itself originated in 1919, instigated by the carnage of World War One, inspiring the laudable aim to understand wars in an attempt to prevent them. This concern with war leads international relations scholars to be pre-occupied with power, security, international structures, and relations between states especially in the realms of foreign policy and military strategy. A concurrent underlying theme of international relations is a desire to achieve order and stability on a global scale.14 As the story unfolds, it might seem that IR conceptualized in this way has little in common with feminism, in all its manifestations. Indeed, feminism appears significantly mismatched with conventional international relations. Feminism's questions about women and domestic/personal politics and interest in disrupting the traditional (patriarchal) order seem inherently incompatible with international relations. As Sandra Whitworth remarks, it is perhaps not surprising that there is such resistance to feminist intrusions into international relations as feminism and IR appear to be “politically at odds with one another.”15 Page 334 → Despite the (political) odds, feminist intrusions have continued. The encyclopedic entry continues. There is the classic “add women and stir” approach, influenced by liberal feminist theory and politics. In the “real world” of international politics, this is a logical move, especially in the context of western liberal democracies apparently committed to equal rights. A brief look around the world's senior politicians and diplomats, combat soldiers or military commanders reveals the scarcity of women. However, the ontological move of “adding women” revealed something initially paradoxical; that women are already in international relations. This did not simply mean some women are in traditionally powerful international roles; but rather that traditional (and therefore invisible) women's activities ensure the international system works smoothly and efficiently. But as we know, there is more to feminism than this. A second approach, evolving from the liberal inspired one, is more concerned with epistemology. This has at least two strands; one linked with radical feminism and standpoint feminism; the second influenced by postmodernism and cultural studies. The radical feminist focus on the lives and experiences of women and the standpoint feminist interest in the gendered construction of knowledge encourages some feminists to investigate how we understand traditional topics in international relations. Security, for example, is conventionally understood as militarily defending the state from attack. Feminists studying women's lives ask if this idea about security is a “male” one, shot through

with the typically masculinist belief that power concerns control rather than empowerment.16 If women defined security, the important issues might be different because women are more likely to be attacked by men they know rather than strangers from other states.

But the feminism and IR story clearly doesn't end there. Beginning the story of IR with its own self-defined origins and purpose perhaps makes it more difficult for many within the discipline to understand that “feminist work goes beyond a desire to be included onto the traditional agenda.”17 Feminist insistence on the relevance and importance of women's lives and the significant power of gendered imaginations on the structuring, representation, and acting out of matters international logically goes way beyond a desire to be included onto the traditional agenda and instead moves us into the realm of radical restructuring. If this is the case—what does this mean for feminist interventions in IR? Then What Happened? As we know, much has been written in the area of gender, women, feminism, and international relations and continues to be so. But it is abundantly Page 335 → clear that the starting point of international relations inquiry has not been fundamentally reconstructed. In response to one of the questions asked by the panel organizers, does this mean feminists have failed? For most feminists it's not so much a question of failure—or if it is, one might argue that the failure belongs to the mainstream in IR who resist or indeed refuse to listen.18 This mainstream and neomainstream resistance and refusal has led to a number of feminist calls to work out a better and more successful way to have the conversation with the mainstream in the pursuit of a more constructive dialogue between the “two communities,” such as that suggested by Cochran. But still the discipline looks unlikely to be restructured in the ways implied by much feminist work, which begs the question “Is the conversation between feminists and the discipline now an exhausted one?” For those who leave the discipline, like Sarah Brown, it clearly is an exhausted conversation. Relying on the ideals and practices of liberal humanism in an effort to persuade the mainstream and neomainstream to widen their epistemological and ontological boundaries to take on board feminist insights radically underestimates the relationship between epistemology, ontology, and power.19 There would appear to be little hope of achieving the separation of epistemology and ontology from power that is necessary to gain the radical restructuring demanded by scholars like Brown. But should we be disappointed about this? Not necessarily.

Story Two: Feminism in International Relations—The Cut(ting) Edge of Contemporary Critical Theory and Practice? Caricature and forgetting is the great[er] danger haunting women's political actions. —Lynn Segal, Why Feminism? “Looking back” over the past in the context of the relationship between feminism and IR has led me to ask myself two questions in particular. What did feminists in IR expect would happen, and what does feminist theory do to IR? As with any revolutionary political movement in the “early” days, there is often a sense of euphoric optimism. Areas that need changing are identified and targeted, which goes along with a willingness to make the changes and a sense of vision that change will be achieved. In the context of feminism and IR—what did all of that mean? At a basic level it meant a more equitable inclusion of women both “out there” in the world of international politics as well as “in here” in the institution. But crucially it also meant a legitimate place for feminist scholarship in the discipline. Let me concentrate on the latter issue of legitimacy by moving on to my second question asked

above: what does feminist theory do? What Does Feminist Theory Do? As Patricia Clough puts it, “Feminists have made the academy itself a site of feminist politics.”20 The challenge that the graduate student threw out to Page 336 → the audience at the Women and IR conference at the LSE in 1988 surely gives us a clue as to one of the things “feminism does” and one of the ways in which feminism makes the academy a site of feminist politics. I mean by this that her comment was shocking, and additionally it didn't seem to have anything to do with IR—the latter being a common reaction to feminist questions in IR. In the “early days” it seemed that people in the discipline were excited about the contributions feminists could make to IR. There was an overall keen interest in hearing what feminists had to say. This did mean that for a short time, feminists could say anything—even the comment made by the unnamed graduate student! Because of this interest alongside a certain “freedom” of speech, it did seem that in the late 1980s there was a sense that feminist scholarship was well on its way to achieving a legitimate place within the discipline. But legitimacy and subversion are not very compatible! And if feminism is (in large part) about blurring disciplinary boundaries, exposing the hard work that goes into maintaining institutional and hierarchical parameters and privileges, and subverting more traditional academic practices, this makes it an unlikely candidate to be embraced by the discipline as legitimate. For as we know, disciplines have boundaries; institutions have parameters that they jealously guard; the academy has a “purpose” and will resist attempts to subvert it. This suggests that the ways available to the discipline of international relations to respond to feminism are actually somewhat limited. What this means in practice is that since feminists “intruded” in the discipline we have witnessed an incremental resistance to and refusal of this work—or rather the implications of this work. This does not mean that many of the more traditional international relations scholars are not genuinely interested in the work that feminists produce. My point is not one about individual motivation—but about how scholars are positioned in relation to the discipline and what this entails for how they will be able to respond to challenges feminist work poses. Let us consider further what these challenges are in the context of what feminism “does to IR.” Think of the millennial reflections panels held at the International Studies Association convention in Los Angeles in March 2000 at which an earlier version of this paper was presented. Does the organization of these panels tell us anything about “feminist effects” and “disciplinary resistances”? The “feminists” were all gathered together on one panel, which certainly appears to divide “feminists” and the “rest of IR” sharply into “two communities.”21 But are we two separate communities—“the feminists” and the rest of IR? “You? I? That's still saying too much. Dividing too sharply between us: all.”22 Of course we have all been involved in affirming the idea of “two communities”—but one effect arising from this is the idea that dialogue between the two communities is necessary. This also logically leads to another effect in the feminism and IR context—which is that the dialogue between the two communities isn't working very well partly because the communities Page 337 → appear to have very different agendas but also significantly because the two communities are unequally positioned within the hierarchy of the discipline. My recent move to Queen's University, Belfast, in Northern Ireland has forced me to think more deeply about the notion of two communities and the necessity of dialogue. Having heard many of the “leading actors” speak, it seems that without exception, “dialogue is seen to be key.” My intention is to (hopefully) not fall into the trap of drawing inappropriate parallels—rather it is to question what the idea of two communities does, especially in the context of apparently “failed dialogue” and the persistent attempts to improve dialogue. An initial suggestion about what follows from the construction of two communities—particularly as evidenced by the construction of the millennial reflections panels—is that it gives the impression that feminists are not involved in any of the other debates currently being rehearsed within IR.23 Is it really the case that no feminist scholars are involved in the neorealist debate or the critical theory/postmodernism debate? This certainly appears to be the case

as there were no “out” feminist scholars on any of the other reflections panels. And on the “other side” of that particular coin, there were hardly any women on the other panels either and none on those particular two panels! Of course these are just panels at a conference—where else might we look for evidence of what “feminism has done to IR” and the implications of thinking of feminist scholars and IR scholars as two separate communities? What about the arena of publishing? A cursory glance at some of the main IR journals will probably show that feminist work is still poorly represented; in the main (and best-selling) textbooks there may be a token page or two—or even chapter or two—but it seems that “IR” is still … IR, taking us right back to the idea that there are these “two communities.” What if we abandoned the idea of having to create and sustain a functioning dialogue with the discipline of IR? This might mean that instead of looking at what “feminism has done to IR”—we might want to look at how feminists “do” IR. Let's take the example of teaching. What is happening there? It can be difficult, in the first instance, getting students to think about the effects and impact of gender. One has to find something—some example—that will grab their attention. One thing that always really grabs my students' attention is the image of a mainstream political theory or IR theory module on which the vast majority of the authors are female. I usually tell them I have taught on many undergraduate core political theory modules that have included a number of authors such as Plato, Aristotle, Schopenhauer, J. S. Mill, Kant, Marx, Weber, Rousseau, Rawls, and so on. And I would come in and give the Mary Wollstonecraft lecture somewhere near the end of the course. Now of course the students see that there is possibly an issue of male dominance and the invisibility of women at stake here. But the move that really gets them to start understanding what gender does is one in which I suggest the whole mainstream module might instead be based Page 338 → on female authors—maybe with one token lecture on a male author. Looking at their incredulous faces, the questions I then ask them are—wouldn't it feel so biased? Wouldn't it seem ridiculous at some level? Doesn't this tell us something about the “naturalness” of the visibility of “things masculine;” and the naturalness of the invisibility of “things feminine”? (One might try the same thought experiment about the reflections panels at the ISA. An all-female “realist” or “alternative” panel perhaps?) This “moment” of destabilization takes us right back to the unnamed graduate student's challenge at the LSE in 1988. And I think it demonstrates one of the most exciting and important things that feminist work does—particularly at the level of teaching. It can destabilize what we think we already know and how we know it. In the context of the contemporary academy this destabilization may not appear to be the most appropriate or best thing to do! Of course students have assignments to do, essays to complete, exams to sit, and degrees to get, and we need to help them succeed in that. But I maintain that a vastly important part of a university education is to destabilize established knowledges about how things work and to emphasize the relationship between how we think and what we do and what happens to us outside in the real world. Because of this I would argue that feminist work is at the cut(ting) edge of contemporary critical theory and practice within international relations. What I mean by cut(ting) is that feminisms are “cut” (or attempts to cut are made) from what is counted as the “real,” the “important,” the “significant,” the “relevant” in international relations. But one of the reasons so many attempts are made to cut feminisms out or pare feminisms down is precisely because this work is so cutting edge. And I say this work is cutting edge because “[feminist] studies alters the very subject of knowledge by calling into question what is considered knowledge in any discipline.”24 By asking the question “who is woman, what is woman, how is woman?,” which is where much feminist work starts—we are set off on a veritable avalanche of deconstructive and destabilizing processes and practices that have the potential to radically shake up conventional ideas about subjectivity, epistemology, ontology, agency, and political action in international relations. Opening the doors for other deconstructive theories and practices such as critical theory, postmodernism, poststructuralism, and queer theory, feminist work is one of the most upsetting things to have happened to international relations. Like the notion of “play” in postmodern approaches, the notion of “upset” is not used lightly—although I am

aware it can be seen as too negative a strategy. But of course the point is—there is no choice to be had. Feminist work is inherently upsetting in a number of ways because it calls into question what constitutes the discipline. Where Do We Go from Here? . . . To New Conversations? When initially working on this paper, my first conclusion about the current state of feminism and IR led me to the first story told in this chapter—one Page 339 → in which the relationship between feminism and international relations could be characterized as an “exhausted conversation.” Perhaps I reached this conclusion as I was too concerned with thinking about the notion of “two communities” and the presumed necessity and/or impossibility of dialogue between the two in the context of exploring and expanding feminist questions. Having spent much time thinking about and working on these issues and having numerous conversations with friends and colleagues working in this area, I have come to the conclusion that if we think about feminists and IR as two separate communities we are necessarily drawn toward the conclusion that the conversation between these two communities is not going to go very far because of their very different agendas and positionings. But if we shift how we think about the positionings of the so-called communities we can tell a different story. For example, if we begin to see feminist scholars as IR scholars, then we can look at the conversation within this community and come up with a story, which tells us that this conversation is actually doing rather well. The two stories I have told in this paper may both be “true” (or not), but I think much more momentum and energy is to be found by starting with the feminist IR community (in all its guises) than by starting with what we traditionally think of as IR. Additionally it is interesting to think about where some of our most fruitful and innovative conversations occur. In my and many of my colleagues' experience, the students who choose to do feminist work or who are genuinely curious about it generally love this work—and courses are highly popular. This is surely a wonderful sign of success. Additionally, I think we have started a trend in “backpack” books—by which I mean the books that some of our number writes are actually to be seen being read by people going on holiday! I'm thinking in particular here of Cynthia Enloe's Bananas, Beaches, and Bases25 (and I am sure there are many more). This must surely imply that this work is reaching a wider audience than just university students and thus must be entering into any number of everyday conversations. How many more conventional or neoconventional IR books can we say this about? And staying with the idea of “conversations,” the new journal that all of us who presented papers on the millennial reflections panel are involved with, the International Feminist Journal of Politics, has a section called “Conversations.” I wonder if Sarah Brown would write a piece for this section as this is a space where we hope all manner of radical and innovative and unusual voices might be heard.

Returning to the Questions and Questions about Questions I want to return to the original questions asked of us, specifically the request by the organizers of the reflections panels that we include in our revised conference papers a consideration of what we thought would be a fair measure of success or failure in our own field and then to assess the extent to which “core objectives” have or have not been met and why. Page 340 → In this chapter I have suggested that we could think about two of the initial objectives of feminist work in IR as to increase the numbers of women in the realms of international relations and international politics and to find a legitimate place for feminist scholarship in the discipline. Have these objectives been met? How can we find out? And what would the answers to these questions tell us about the successes or failures of feminist work in IR? There probably are more women now in the public world of international politics and the academic world of international relations. But it is still the case that most of these women are clustered at the lower and middle levels of the hierarchies. Does this mean feminists have failed in their objectives? Or does this mean that the institutions have failed feminists?

With regard to the legitimacy of feminist scholarship—it clearly depends where one goes to find the criteria of legitimacy. As I have intimated in this chapter, if we judge feminist work by the standards of conventional and neoconventional IR, much feminist work would probably fail the test of legitimacy. However, if we use feminist criteria of legitimacy—a different story can be told. Of course feminists do not all agree with each other, and some feminists will find some feminist work inappropriate. But that the debate itself is legitimate and centrally important is unlikely to be in doubt. And this makes a lot of difference. At the end of the day—or more accurately, at the beginning of a new millennium—as I see it, the task for scholars like myself is not to identify “intrasubfield consensus,” yet I do not see this as a cause for lament. Indeed, if after a decade and more of feminist international relations scholarship there was a synergy of interests, concerns, and approaches between feminists and the mainstream and neomainstream of the discipline—this would indeed be a cause for lament; we would have lost the battle. This, I know, sounds paradoxical. Wasn't feminism supposed to change IR into something “feminist friendly,” “gender sensitive,” even “emancipatory”? Possibly. Of course the questions we were asked were problematic. It is the traditional practice of an institution like IR to try and secure itself against radical change. Indeed, IR's task in terms of “reflection” is to see itself reflected back. However, I share Jenny Edkins's view that the role of the critical IR theorist—feminist or otherwise (though I happen to think feminists do it better …)—is not to secure the discipline from theoretical danger but indeed quite the opposite.26 This can be quite a risky practice and may not be one agreed upon by all the presenters on the panel from which these chapters are a result. This is not surprising, as there are many differences between feminists! But it is a practice that feminists continue to engage in because despite the differences between us, my guess is that we all still believe “it matters.” That is, whatever we think feminism is, there is a sense that injustices continue, hierarchies persist, and people are consistently harmed. Therefore “it matters” to continue to be involved in the practice of insecuring the discipline and to not only provide “answers” but to question what it is we are to provide “answers” to. Page 341 →

Notes I would like to thank the following for conversations about this essay: Helen Brocklehurst, Cynthia Enloe, Cynthia Weber, Antje Wiener, and the organizers and participants at the seminar organized by the Gender and Women's Studies Seminar series at the University of Bradford in March 2000, especially Ruth Jacobson and Anne Scott. I would also like to thank Michael Brecher and Frank Harvey for organizing this series of panels and the subsequent publication of the papers and to commend them for their patience. 1. J. A. Tickner, “Feminist Theory and Gender Studies: Reflections for the Millennium,” in this volume. 2. J. J. Pettman, “Critical Paradigms in International Studies: Bringing It All Back Home?,” in this volume. 3. L. H. M. Ling, “The Fish and the Turtle: Multiple Worlds as Method,” in this volume. 4. C. Sylvester, “‘Progress’ as Feminist International Relations,” in this volume. 5. This potential is a function of the structure of the discipline and not the intentions of the organizers. 6. L. Irigaray, This Sex Which Is Not One (Ithaca, N.Y.: Cornell University Press, 1985), 213. 7. I apologize for not being able to name her—perhaps this tells a story in itself. 8. S. Brown, “Feminism, International Theory, and International Relations of Gender Equality,” Millennium 17, no. 3 (1988): 461–76. I believe that Sarah Brown was a graduate student at Sussex University at the time of the LSE conference. She has since left the profession. 9. R. Grant and David Long, eds., Women and International Relations, special issue of Millennium 17, no. 3 (1988). 10. R. Grant and K. Newland, Gender and International Relations (Milton Keynes, England: Open University Press, 1991). 11. Brown, “Feminism, International Theory, and International Relations of Gender Equality,” 472. 12. M. Cochran, “Talking with Feminists about What We Can Know in IR,” Cambridge Review of

International Affairs 12, no. 2 (1999): 46–56, 51. 13. M. Zalewski, “Feminist Theory and International Relations,” in From Cold War to Collapse: Theory and World Politics in the 1980s, ed. M. Bowker and R. Brown (Cambridge: Cambridge University Press, 1993), 128. I would not ask this kind of question now. 14. M. Zalewski, entry on “Feminism and International Relations,” in The Encyclopaedia of Feminist Theory, ed. Lorraine Code (New York: Routledge, 2000). 15. S. Whitworth, Feminism and International Relations (London: Macmillan, 1994), ix. 16. J. A. Tickner, Gender and International Relations (New York: Columbia University Press, 1992). 17. C. Sylvester, Feminist Theory and International Relations in a Postmodern Era (Cambridge: Cambridge University Press, 1994), 134. 18. M. Zalewski, “Where Is Woman in International Relations? To Return as a Woman and Be Heard,” Millennium 27, no. 4 (1999): 847–67. 19. M. Zalewski, “Playing Like a Girl: A Response to Molly Cochran,” Cambridge Review of International Affairs 12, no. 2 (1999): 57–60, 59. 20. P. T. Clough, Feminist Thought (Oxford, England: Basil Blackwell, 1994), 2. 21. My point is not to imply that the organizers of the panels intended to marginalize feminist work. Indeed, the inclusion of this panel indicated precisely the opposite. My aim is rather to peruse the effects of the categorizing practices we are all involved with. 22. Irigaray, This Sex Which Is Not One, 218. 23. See note 21. 24. J. Gallop, Reading Lacan (Ithaca, N.Y.: Cornell University Press, 1985), 18. 25. C. Enloe, Bananas, Beaches, and Bases: Making Feminist Sense of International Politics (London: Pandora, 1989). 26. J. Edkins, Poststructuralism and International Relations: Bringing the Political Back In (Boulder, Colo.: Lynne Rienner, 1999), 142.

Page 342 → Page 343 →

METHODOLOGY

Page 344 → Page 345 →

FORMAL METHODS IN INTERNATIONAL RELATIONS Michael Nicholson

Introduction Approaches to the analysis of international relations using formal methods are now well established. Models of international relations problems appear frequently in many of the leading journals. While there are still those who hold that formal methods only manage to render obscure what had hitherto been clear, the achievements of such methods, commonplace in all the other social sciences, make such positions hard to justify. The purpose of this paper is to discuss the principles behind formalization. I also wish to justify its extensive use, mainly by conceptual argument and by referring very briefly to some of the achievements. These achievements are genuine and require attention from those outside the formal tradition as well as those within it. Ultimately the aim of formal modeling is the same as that of any other tradition in international relations—to understand better the international system and point in the direction of making the world a pleasanter place in which to live. Has formal modeling contributed to this, and, if so, what are the directions in which it might continue to move? Scholars have tried to model almost every area of empirical international relations with some attempts to look at the normative also. Thus issues of cooperation, conflict, deterrence, alliances, arms races decision structures, hostility patterns, and so on have passed under the formal modelers' eyes with varying success. They search for the nature of the logic of various situations such as conflict situations; they try to investigate the consequences of various sorts of assumptions being true; they also try to structure theories to make them amenable to empirical test by statistical means, something that is difficult unless some minimum degree of formality is allowed. Such analyses are necessary if many of the problems central to international relations are not to be vague and indeterminate. I am not claiming that all aspects of international relations have to be formalized in order to make progress, or even that they can be. However, I do argue that some very important problems in the discipline require formal approaches to make progress. In this we are no different from the rest of the social sciences. The same broad conceptual problems face them all. Page 346 → The underlying basis of my argument is as follows. Natural languages, by which I mean the languages we speak, such as French, English, Chinese, and so on, are good at description but quickly become awkward and clumsy when faced with significant chains of deductive argument. By contrast, formal languages, and I mean particularly mathematics as it is normally understood, are very good at deductive arguments though rather weaker in fine description. Theories involve deductive structures, at least within the conventional empiricist tradition of the philosophy of science. Hence, if they are to be of any complexity they are going to require complex deductive arguments. In their turn these require mathematical methods. Further, we aim to explain events in international relations. An explanation shows how the events to be explained follow from certain initial conditions and a theory. Again we require deductive arguments that quickly become complex. Thus, for those issues of which complex deductive arguments are a part, we require formal methods. I assert there are many such issues. We can restate this argument by looking briefly at the concepts of syntax and semantics. We use syntax to mean the internal structure of a language. Thus, “the king of France is a hobgoblin” is a perfectly syntactically correct sentence in the English language irrespective of its external truth claims. Using the word language in a slightly extended though perfectly conventional sense to include mathematics, we can also assert that if “>” is a transitive relationship and A > B > C, then A > C irrespective of the interpretation of the letters or the symbol “>.” We use semantics to mean the interpretation of the language to the world outside itself. Thus, a syntactically correct sentence can be true or false, or even relate to imaginary entities (such as hobgoblins), while being syntactically correct. The semantics deals with the external meaning of the sentences whereas the syntax deals with the internal structure of the language. Thus, the sign “>” is used, for example, to mean “is numerically greater than” and also,

in the context of choice theory, to mean “is preferred to” as in the sentence “A is preferred to B.” The statement is syntactically correct in both interpretations. Natural languages are very good at semantics. We can say all sorts of interesting and subtle things about the world with the uses of natural language. Indeed, the whole richness of the world is captured in language.1 However, natural languages are inflexible as far as syntax is concerned. We can rearrange things a little, but we cannot carry out more than basic logical deductions within them. Anything much beyond the traditional syllogism leads to cumbersome structures that are hard to understand and follow when expressed in natural languages. Mathematical languages have the opposite virtues and vices. Their descriptive capacity is limited. “My love is like a red, red rose” would only translate as an unwieldy collection of symbols with little aesthetic appeal and whose meaning would be obscure. The semantic relationship of a mathematical language to the real structure it is describing needs to be precise but often rather stripped down. However, as we commented earlier, they are excellent for carrying out logical deductions. The implications of a set of mathematical statements can be worked out and often result in amazing and counterintuitive conclusions. They can also Page 347 → assert with precision whether propositions are true or false within the given calculus, proving various propositions. We do not need to simply say that something “looks true,” but we can show that it is true. “True,” of course, relates to syntactic truth within the mathematical system. However, we can assert that if the premises are semantically true then the conclusions, however far they lie from the original premises, are likewise semantically true. The truth has been carried through various steps on which we can rely with total confidence. It is this versatility with complex logical operations that gives mathematical languages their great power. With mathematical systems we can deal with high orders of syntactical complexity that can, of course, reflect complex semantic systems.

Some Applications of Formal Methods in International Relations Many of the applications of formal methods in international relations theory have been in the area of decisions and choice.2 Rational choice theory and the closely related game-theoretic approaches are dominant here. It is hard to see how these could have developed without the use of mathematical methods. It is also hard to deny that they have been of great significance in the development of international relations theory. This of course emphasizes the common link between international relations and the rest of the social sciences. Rational choice theory and game theory developed from microeconomics although they have been applied to virtually every form of human behavior both as an explanatory and normative theory. Many of the most significant developments, such as Axelrod's work on cooperation,3 have not been cultivated within a narrow disciplinary framework but apply to cooperation as a general human phenomenon. Such developments have important implications for international behavior and can become incorporated within international relations theory. However, this theory is all part of a more general social theory. The initial applications of the theory of games to international relations came in the early 1960s with the work of Boulding, Rapoport, and Schelling.4 This marked a big step, and the nature of deterrence and other conflict phenomena were greatly illuminated by this sort of analysis. Though now taken for granted, the insights developed by the formulation of such basic and simple games as chicken and prisoners' dilemma were profound. Nevertheless, after the big burst in the early 1960s, the development of the application of the theory of games to international relations seemed slow, though perhaps more so at the time than it seems in retrospect. Nigel Howard's development of metagames provided some development, but the applications to international relations were relatively few.5 His later work on hypergames and drama theory has been more influential though arguably more neglected than it deserves.6 The developments of game theory in international relations were nevertheless moving on with Brams's work, culminating, at least so far, in the theory of moves (TOM) reinforcing the tendency to look at games in the extensive form rather than the normal form as had earlier been more conventional. Kilgour, Zagare, and others have worked very much in Page 348 → the same sort of tradition. It is a tradition that comes directly out of the theory of games, though moving a long way from the original formulation (as one would hope and expect given the original work was first published in 1944). This work has been angled toward international relations, where such aspects of human behavior as deterrence have been prominent. However, it has more general application as all the authors are very aware, particularly in the case of TOM.7

Given that conflict and cooperation are central issues in international relations, it is not surprising that game theory and its offshoots have been prominent. However, they do not wholly define the field, and various general decision theory models such as the expected utility model have been powerfully employed, notably in the work of Bueno de Mesquita.8 Expected utility models bring in the issues of risk and uncertainty in an explicit way. Given that the world is inherently very uncertain this seems like a necessary thing to do. Issues of uncertainty are quite prominent in recent discussions such as those of Claudio Cioffi-Revilla.9 Rational choice and game theory have been the dominant forms of analysis based on decisions, but it is a mistake to regard them as making up all of decision theory. Other decision theories such as prospect theory also have application in international relations, though more in potentiality than actuality so far. They too are stated in mathematical terms and would be hard to analyze otherwise. One of the issues that the whole rational choice tradition has illuminated is the ambiguity of the concept of rationality itself when placed in a competitive context. The game of prisoners' dilemma itself exposes this ambiguity. Even when we have clearheaded actors, with coherent and transitive preferences, we cannot always say just what a rational action consists of. This issue, though a negative one, is very important as far as discussions of rational behavior are concerned. This has provoked a lot of discussion on the edge of formal theory itself about the whole nature of rationality in human conduct. Ainslie, Elster, Sen, and many others have participated in this debate.10 The centrality of decision theory in formal approaches to international relations (or any other approaches for that matter) is not surprising. We believe many phenomena in international relations can be explained as the consequences of decisions. However, not everything can, and there are some approaches that do not center on decisions, or, inasmuch as they imply decisions somewhere, they are taken as input where the output is the central issue. Classic amongst these are the Richardson models of arms races and hostility. There have been extensions of this at various times including the various hostility models of Dina Zinnes and her collaborators.11 Schrodt has also developed Richardson-style models to discuss arms balances and other similar problems.12 This class of model explores the consequences of simple input-output rules where the ensuing patterns are often counterintuitive. Despite the historical importance of Richardson's work in the development of our discipline, these sorts of model have tended to be secondary to the models based on various forms of decision theory. Page 349 →

The Nature of Formal Modeling Broadly, formal modeling involves the use of mathematical, diagrammatic, and symbolic methods in the study of problems in international relations. The term is often used synonymously with “mathematical modeling,” though perhaps here we should be cautious. Clearly the use of mathematical methods is an aspect of formal modeling. However, we might broaden the definition. Formal theory can be regarded as any theory that is expressed in any terms other than natural language. This would then include a statement of the one-shot prisoners' dilemma13 as part of formal methods. This is a clear case of where something was obscure in natural language, but when stated more formally the concept became clear. It is a very simple problem. Once stated, the deductive steps are minimal and obvious. Nevertheless this simple statement of the problem has been enormously insightful in all branches of the social sciences. We might also include in formal methods any diagrammatic representation such as a flow diagram where we seek to illuminate a problem by representing it in a nonverbal manner. Often such diagrams are just a restatement of a verbal argument in different terms to make the argument clearer. No new information is added, and there need be no deductive steps involved in the argument that reveal unsuspected implications of the argument. If we use diagrams to illustrate a verbal argument, this is a wholly pragmatic issue and is judged solely by whether some or all the readers of the argument find it helpful. However, deduction starts quite early. Once deductive arguments get complicated, they quickly need formal, mathematical methods to carry them through. This is no longer a matter of taste or convenience but of necessity. Richardson models are a clear case in point, though this is not an isolated instance.

Clearly we have to be careful with our vocabulary, so some elaboration on the basic terms will be useful. The words are not always used consistently, though the underlying concepts are. I follow the usage of R. B. Braithwaite in what follows.14 A theory is a set of generalizations about the world. The generalizations can be true or false, something that is determined by observation. A theory starts off with a set of postulates that are regarded as true for the purpose of the exercise. Other propositions are deduced from these postulates, and the whole set of propositions is the theory. The deduced propositions are called lower-level propositions within the theory. It is the theory as a whole that is testable against observation. Not all individual propositions need be testable (or refutable in Popperian terms).15 The theory stands or falls as a whole as a description of the world.16 A model17 is a simplified picture of reality. In structure it is the same as a theory. It starts from postulates from which other deductions are made. However, because of its simplified nature, it does not always follow that it is testable as the reality it describes is an artificial one. It is also in part the prototype of a theory. A model is progressively complicated to bring it closer to reality until it is sufficiently close to stand in as a picture of the reality. Page 350 → The boundaries between theory and model are sometimes rather vague (which also makes the issue of refutability more ambiguous in practice than it is in theory). However, models are often used to explore the underlying structure of some process even when it is quite a way from a description of any international or other social process. The initial statement of the one-shot prisoners' dilemma is usually told in terms of an absurdly implausible story. Even iterated models of the prisoners' dilemma are a long way from any direct application to the real world except, from time to time, that of the psychological laboratory. However, they do appear to have some structural correspondence with “real” processes, and for that reason we think we can get some greater understanding of these processes such as the significance of time and the discount factor. It might lead us to theory, but it is a long way before we get there. As Jon Elster remarked, “Rational choice theory tells us what to look for, not what we will find.” However, the real world even then intrudes on our models, and we can fit various findings into them. Usually we assume simple exponential discount functions. However, there is increasing evidence that people use hyperbolic discount functions despite the fact that this is in some sense “irrational” and can produce switching preferences over time. However, if we are hoping to get theories of how the world behaves rather than as we think it “ought” to behave then these must be incorporated into our models as precursors of what will go into our theories. Mathematics are used, and need to be used, when there are significant deductive steps in the argument. A mathematical structure, sometimes called a calculus, starts from axioms, which are assumed to be true, from which theorems are deduced. Thus, they parallel the structures of models and theories. The abstract terms of the calculus are paralleled and interpreted in terms of the theory. Thus, if there is a correspondence between the axioms of the calculus and the postulates of the theory, the theorems of the calculus correspond to the lower-level propositions within the theory. A calculus can often be interpreted in many different contexts. Thus, differential equations can be used to describe arms races,18 the growth of plants,19 and the movements of bodies in motion (as in classical mechanics). We can have nonmathematical models and nonmathematical theories, but in general theories of any complexity are mathematically expressed. The final purpose of any theory is to test it against the real world. There is obviously a close relationship between the use of formal methods and the use of statistical methods. Statistics deals with classes of events and generalizations. So, for the most part, do formal methods. Thus, many of the statistical tests used are tests of relationships such as linear relationships. This immediately relates statistics to formal theory, and the two are intertwined. To test relationships statistically one must also manipulate them into easily testable form, either in the sense of restating the relationships (testing for log-linearity instead of direct linearity for example) or in working out implications of a proposition to find those that are readily testable that will imply the truth also of the basic proposition under review. Any theory Page 351 → that is to be tested by statistical means is going to be at least minimally formal. Statistical theory itself is, of course, mathematically based, though this is a separate issue. It should not be concluded from this that all formal methods necessarily require strong forms of measurement. Mathematics deals in qualitative shapes as well as measurement though such qualitative methods are not

widespread in international relations. On the other hand a lot of work does depend only on ordinal measures where greater than and less than and equal to are the only categories. Quite a lot of decision theory, including basic games like prisoners' dilemma, can be posed in ordinal terms.

What Is Added by the Use of Formal Methods? Formal languages, beyond the simplest, are technical languages. This raises the problem of communication and understanding across different academic cultures. We have to learn mathematics. Often the problem with understanding mathematical arguments is not a lack of knowledge of the mathematics involved so much as the lack of the habit of following mathematical arguments. One “reads” mathematics in a rather different way from the way one reads natural languages. This can be an impediment to the novice though it need not be an impediment to anyone interested in the results and the initial premises from which they have been derived. Given that it takes time and effort to learn mathematical techniques, which is something of an ordeal for many people, why should we bother applying formal methods to the study of international relations? An argument one often hears—literally, for though I have heard it often in conversation it is rarely committed to paper—is that all formal theories do is restate the obvious in complicated language or move off into mathematical fantasies that have nothing to do with the real world. The discipline of international relations can move ahead quite satisfactorily without the intervention of mathematics. In its extreme form, this argument is hard to justify, though in a more muted form something may be said for it. It is probably the case that some efforts at formal theory have been justified more by mathematical enthusiasm than by a substantive problem. Intrigued by the mathematics, scholars have pursued problems further and further away from any plausible application. Even this might be justified in that, prior to any investigation, it is often hard to know where it will lead. Most formal theorists are, I presume, like the author in confining the bulk of their output to the wastepaper bin. In this paper I argue that formal theory both has and will contribute much more than that. However, it is important for formal theorists to make some effort to communicate their results in nonmathematical language to those who are involved in other fields. This is often possible providing that the actual workings of the processes are taken on trust. Thus, using game-theoretic techniques, Wagner20 has shown that a balance-of-power system does not require the five actors that some conventional theorists have asserted.21 The working out of this result is complex and requires some knowledge Page 352 → of game theory. However, the outsider needs to know the initial postulates of the theory—that is, what Wagner has assumed—and the final result. It does not require much, if any, formal knowledge to understand these; indeed, these are the general concepts that go into a balanceof-power theory anyway. If these postulates are accepted, the rest follows. If these postulates are regarded as incomplete, the nonformal theories can suggest a new approach. This is not a question of not doing formal theory so much as making sure that there is a maximum of accessibility. One major reason for at least attempting to apply mathematical methods to international relations is their extraordinary success in other disciplines. Physics, of course, is the paradigm case. From its earliest days, physics could develop only with the extensive use of mathematics. Only the most extreme of postmodernists would want to deny its great success. The use of mathematics in the natural sciences is generally uncontroversial. Likewise mathematics has been widely used with great success in the social sciences. Economics, in particular, has made big advances by the use of mathematical techniques. Even those who regard economics as going overboard in the overdevelopment of technique as opposed to substance nevertheless acknowledge the necessity of mathematical methods in developing even basic theory.22 While economics stands as the physics of the social sciences, the application of mathematics has been widespread to other areas. While the extent of the application of mathematics to psychology, sociology, and so on is controversial, the application of some mathematics is widely if not universally accepted amongst practitioners of the various disciplines. Thus, it would seem foolish to neglect the possibility that similar methods in international relations and cognate subjects would also be fruitful. They may yield magnificent returns as some of the earlier practitioners such as Richardson hoped and even expected. That at least they would yield some profitable insights would almost be expected. To neglect the use of methods that have been so outstandingly successful in other areas would seem simply irresponsible. Further, there seems no special reason why the subject matter of international relations should mean that it is exempt from analysis by

mathematical means. Scholars of international relations study a form of social behavior. If any form of social behavior is susceptible to mathematical analysis, then the presumption is that international relations would be also.

The Covering Law The use of deductive methods became central with the covering law of explanation. An explanation of an event is an interpretation of an event in the context of a theory; in slightly different words, the event is seen as an instance of the theory. This form of explanation has been most clearly expounded by Hempel in his covering law model of explanation.23 Hempel argues that all true explanations can be stated in this form, and an analysis of theories in international relations would seem to confirm that we, too, follow this pattern. One of the more explicit recent uses of this in international relations is in Geller and Singer.24 Page 353 → In the covering law model, we have a theory consisting of the statements L(1), L(2) … … …. L(n). These statements are generalizations about sets of events (which can be in natural language or a formal language) that are deductively linked so that those that are not postulates are logically implied by groups of others. We also have a set of initial conditions: C(1), C(2) … … …. C(m). These are statements about singular events. We have another set of singular events, E(1), E(2) … … …. E(k), that are the events we want to explain. Thus the search for an explanation involves an inquiry into why, if the C's occur, we should expect the E's to occur. In a strict form of explanation (such as is commonplace in physics), the E statements and the L statements together logically imply the E statements. This is illustrated in the following diagram.25 The point is that a logical deductive argument is carried out. While this does not necessarily imply the use of formal methods, once the argument gets at all complex they become necessary. Clearly this form of explanation in its strict form, though a commonplace in physics, tends to get blurred in the social sciences. Our facts in our theories are not as precisely defined as in physics, and our explanations are more commonly “explanation sketches” (in Hempel's terms) rather than strict explanations. However, this does not alter the need for the deductive argument and the mathematics that follows from that. The covering law model is a standard approach to explanation in the empiricist or positivist approaches to the social sciences. Mathematics is not excluded if we adopt approaches that do not center on generalization, though their scope and range is more restricted. I have argued elsewhere that even approaches that on the surface deny the centrality of generalization in fact need some concept of generalization if they are to be any use at least in any policy context. This opens the door to a mathematical approach.

Some Counterarguments It could still be argued that while the above arguments are satisfactory as far as other social sciences are concerned, international relations is different in some way. There is no scope for significant deductive arguments in this context. This argument is hard to accept. Apart from the successful cases where formal deductive methods have been used, it is hard to see what is distinctive about this form of human behavior that makes it exempt from the principles that apply to others. If significant deductive arguments are useful in other parts of the social sciences, it would seem odd if international relations were somehow different. It is an argument that has been used generally Page 354 → about social scientific methods in international relations but with as little basis. It might appear that an analysis in terms of natural language is to be preferred to a formal analysis in that it is more widely understood. Formal theorists can understand natural language in a way in which the reverse does not happen. For this reason, only in situations where a formal analysis is manifestly superior should it be used. This argument has two weaknesses. First, it is often not clear at the beginning what the payoffs of writing out a problem formally will be. It may turn out to be an exercise in futility. However, it may turn out to reveal a richness that was obscure at the beginning. In some ways, formal theorists resemble gold prospectors who get

inured to the disappointments of mining dross by the promises of the occasional successful insight. In many ways problems are always worth trying out. Second, formal theorists are not obliged to make every step of every argument clear to everyone irrespective of their mathematical knowledge. Indeed, there can be debates amongst formal theorists, sometimes on purely technical matters, that would be unclear to outsiders. There is nothing particularly wrong in this. What is important is that when significant conclusions are reached that might be expected to be of general interest to the profession as a whole, efforts are made to communicate to the broader professional public in generally comprehensible language. The more technical our arguments, the more crucial should be our expository efforts. This is no more than we would expect of other technical areas. International relations specialists are interested in what comes out of the Soviet archives about, say, the Cuban missile crisis. However, we leave the technical arguments about the reliability or authenticity of some documents to the specialists whose debates might be incomprehensible to outsiders. At best, the outsiders would lack means of evaluating the evidence. Thus, we wait for the end results of the argument and expect them to be explained to us. This is similarly the case with work done in the formal theory tradition.26 This does not mean that all problems can be usefully analyzed by mathematical means. If description is dominant, then natural language is the most appropriate and will do the job better. Indeed, mathematical languages will add nothing except, perhaps, a little confusion. An account of the actions of the United Nations over the last five years will almost certainly be better done in natural language. However, an analysis of the power within UN committees might well have a useful formal aspect to it, drawing on the many powerful propositions from the formal theory of committees.

The Origins and Development of Formal Modeling in International Relations Formal modeling became noticeable in international relations in the early 1960s. The works that took it from being a rather specialist niche, which excited a handful of scholars—usually not professionally involved in inter Page 355 → national relations—were those of Boulding, Rapoport, and Schelling. The approach became visible and hard to ignore except, possibly, by those who believed they had found the truth in traditional methods. Schelling concentrated on extending game theory to make it more readily interpretable in terms of international relations and practical conflict issues, though not always in a way that appealed to the standard game theorists. Both Boulding and Rapoport, though in rather different ways, used game theory and Richardson's interaction models to analyze general conflict problems, especially relating to international relations. Boulding further ransacked the economists' collection of models, particularly duopoly and oligopoly models, to interpret them in more general contexts. Rapoport, a mathematician but also a general polymath, concluded his book with a dialogue between a supporter of the capitalist system and a supporter of a communist system to illustrate the concept of what he called “debates.” The dialogue, and the whole spirit behind it, has strong hints of what later became called constructivism.27 Though this is not directly related to the concept of formal modeling, it is illustrative of the spirit of intellectual innovation that was characteristic of this whole style of thought. Though this period saw a vastly increased interest in formal methods, the origins were much earlier. Richardson was the grandparent of formal methods as well as of statistical methods in international relations. His work goes back to the early 1920s when he published a monograph written while an ambulance driver in World War I. He developed this during World War II and the immediate postwar period (he died in 1953). His central mathematical work was his celebrated arms race model, now regarded as a central piece but initially ignored. Rapoport published an extended article in the first issue of the Journal of Conflict Resolution on “Lewis Fry Richardson's mathematical theory of war.”28 This is still the best exposition of Richardson's work and its breadth. The recognition of Richardson as an extraordinary thinker and pioneer in the development of international relations and peace research was confirmed with the importance given to him in the works by Boulding and Rapoport mentioned earlier. The development of formal methods did not come in isolation, either intellectual or political. There were two factors that made the developments unsurprising in the late 1950s and early 1960s. First, the intellectual context of the time was one where the powers of mathematical analysis were being recognized more generally in the social sciences. Second, the political context was the cold war, which included fears of nuclear annihilation. A common

feeling amongst those who were deeply concerned about the nuclear problem was that traditional international relations scholarship needed at best augmenting. It seemed to offer little prospect of peace. The dark pessimism of the realists such as Morgenthau who appeared to hold that what we had for an international system was the best we could hope for seemed unduly dreary. If the traditional methods could only help us bemoan our sad lot at a higher level of sophistication then perhaps other methods should be used that might offer some better prospect. At least one should try. This, of Page 356 → course, applied to the whole of the social scientific approach to the peace research tradition and was not confined to formal methods as such. However, it was a significant stimulus to the formal approaches. However, the whole development came in an intellectual as well as a political context. Economics had taken a powerful mathematical and statistical turn and appeared to promise results. Attempts to use mathematical analysis in all forms of social science from sociology to psychology were developing apace. Political science, often regarded as the nearest disciplinary neighbor to international relations, was using formal methods enthusiastically and could claim to have done so in limited forms since the eighteenth century when Condorcet formulated his famous paradox of voting. Votes rather naturally lend themselves easily to both theoretical and practical work, but the mathematization of political science has gone well beyond the analysis of voting. The move toward mathematization was associated with the view that the divisions into separate disciplines of the social sciences were arbitrary. The divisions into economics, psychology, and so on were convenient, possibly, but indicated no fundamental ontological differences (to use a word that was rarely used at the time and would not have been used in contemporary justifications of this argument). The basic issue was the study of human behavior where there was the presumption that this would be done in similar ways whatever the manifestation. Appropriate techniques were a matter of convenience, not of principle. Even these techniques, such as many mathematical techniques, had broad application over both natural and social sciences. In this context, the extension of formal methods into international relations and conflict studies was not only natural but almost inevitable. The role of the political background of some of the early workers in the formal tradition was significant. Richardson was a Quaker pacifist, of the generation who were deeply influenced by World War I, which he had seen firsthand. He passionately believed in rebuilding a world in which peace was the norm. His work was a model of objectivity. Nevertheless, his underlying position was that we must alter the world but first we need to understand it as it is. Indeed, it would be impossible from reading his work to know what his political position was. Boulding and Rapoport were both within this same tradition, though it was World War II and the subsequent nuclear world that was the center of their political experience and concerns. They were studying the world with a long-term view of bringing about fundamental change. They certainly believed in change. Schelling occupies a more ambiguous position. He certainly provided a more fundamental analysis of conflict than the traditional theorists could, and nor was he a crude cold warrior. However, his view was less radical. He was more warmly welcomed than the others by the liberal end of the conventional establishment. There was a lot more work done using formal methods in what can be more clearly seen as a problem-solving tradition. All sorts of work in operations research went on at organizations such as the RAND Corporation into such issues as missile siting, convoy size and deployment, and all the other more down-to-earth aspects of military procedures that rest on the edge of practical Page 357 → strategy and international relations. Moving toward the more political and abstract realm of international relations theory and such issues as how to have agreements in contexts where there is cheating come close to a sort of Fabian approach to fundamental change.

Conclusion: What of the Future? Formal approaches to international relations theory have added a lot to our understanding of the operation of the international system and bode well for doing rather more in the future. We know things, for example, about alliance behavior and the structures of conflict, quite apart from the structuring of problems to make them susceptible to statistical testing, that would not have been possible otherwise. This is all sufficient justification in itself. What of the future of formal methods? We face the problem that Popper raised about the prediction of ideas. If one can predict an idea one has already had to have it so one is caught in a trap. All we can do is suggest some

plausible lines of progress that may be shown in the future to have been good or bad guesses. However, it is hard to suppose that deductive modeling is going to be less needed. As theory advances, it is more likely to become more deductively complex rather than less. Guetzkow's famous metaphor of “islands of theory” will expand and link up needing more not less formality to construct them adequately. Likewise polymetric techniques are likely to advance requiring more ingenious rearrangements of theoretical relationships to maximize the effective confrontation of theory with fact. Even simplification requires more formality: the derivation of simple patterns from complex ones requires insightful deduction. The big successes of formal theory have mainly been with one form or another of decision theory, mainly rational choice theory. Indeed, sometimes formal theory is mistakenly referred to as if it were a near synonym of rational choice theory. In the classic debate, formal methods have tended toward agency rather than structure. This itself has been elaborated in the formal tradition by Starr and various coauthors. I would speculate that structural theories will become more and more developed from a mathematical point of view. Other speculations are that simulation methods will be used more frequently and that these will generate a whole set of quasi-experimental worlds. These simulations can be of the structure of the system or of an artificial intelligence form that might (or might not) integrate more closely with choice theory. We might also move to consider further still the presuppositions behind not just formal theory but in particular those of rational choice. Often, as some of the critics of rational choice have maintained, the apparent obviousness of the maximizing hypotheses makes one blind to the possibility that, as hypotheses about actual behavior, they may be wrong. Theories about choice are empirical theories and depend on their being accounts of how people actually choose. We can have doubts about some of the central Page 358 → principles of rational choice. For example, people do not calculate their degrees of belief as the theory of probability suggests they do; they do not necessarily have exponential discount functions even if such functions do lead to well-behaved patterns; they do not always use backward induction as they should as many experiments of iterated prisoners' dilemma with a known ending point show. Choice theories will have to embody these issues. People do not always behave as they “ought” to do from the point of view of the modeler. The conceptual problems also need to be dealt with with care. There is already a well-trodden field that elaborates the presuppositions of such concepts as “rationality,” but there is further to go. The efforts to bring emotions into the analysis of choice represent welcome moves away from the aridity of earlier rational choice theory. Indeed, it would be better to get rid of the word rational altogether as it has prescriptive overtones that are inappropriate for a supposedly empirical analysis. As the applications of formal methods in international relations are likely to increase rather than decrease, we should also pay more attention to making our results comprehensible to people who have neither the training nor the interest to engage in the technicalities of the problems. This does not mean that we have to make every technical exercise clear to everyone—far from it—or that everyone should be constantly explaining things to the nonmathematically inclined. It simply means that some of us should tackle the problem of exposition seriously. Formal approaches are an important and necessary means of tackling problems, but they are not the only way, nor are all problems suitable for formal analysis. Thus, we do not have to suppose we shall colonize the whole of international relations knowledge. We must, at all cost, avoid becoming a mathematical priesthood purporting to dispense knowledge from behind a secure barrier of incomprehensibility.

Notes 1. There is also the view, stemming from the later Wittgenstein, that language does not just reflect the world but actually constitutes the world. I would not deny that language is, at times, constitutive but assert that it also has a reflective role as in the correspondence view of language. 2. This paper is not intended as a comprehensive, or even a partial, survey of the applications of formal

methods in international relations. I give various examples along the way to illustrate my argument. All these illustrations I use are, in my view, significant developments, but the omission of some work (and I omit most work) does not imply that I regard it as any less important. 3. Robert Axelrod, The Evolution of Cooperation (New York: Basic Books, 1984). 4. Kenneth Boulding, Conflict and Defense (New York: Harper and Row, 1962); Anatol Rapoport, Fights, Games, and Debates (Ann Arbor: University of Michigan Press, 1965); Thomas Schelling, Strategy of Conflict (Cambridge: Harvard University Press, 1960). 5. Nigel Howard, Paradoxes of Rationality (Cambridge: MIT Press, 1971). 6. Nigel Howard, “Drama Theory and Its Relationship to the Theory of Games,” Group Decision and Negotiation 3, no. 2 (1994): 187–206. 7. Steven J. Brams, Theory of Moves (Cambridge: Cambridge University Press, 1994).Page 359 → 8. See Bruce Bueno de Mesquita, “The Contribution of Expected-Utility Theory to the Study of International Conflict,” in Handbook of War Studies, ed. Manus I. Midlarsky (Boston: Unwin Hyman, 1989): 143–69. 9. Claudio Cioffi-Revilla, Politics and Uncertainty: Theory, Models, and Applications (Cambridge: Cambridge University Press, 1998). 10. George Ainslie, Picoeconomics: The Strategic Interaction of Successive Motivational States within the Person (Cambridge: Cambridge University Press, 1992); Jon Elster, Ulysses and the Sirens (Cambridge: Cambridge University Press, 1979); A. K. Sen, “Rational Fools: A Critique of the Behavioral Foundations of Economic Theory,” Philosophy and Public Affairs 16, no. 4 (1977): 317–44. 11. Dina Zinnes, Conflict Processes and the Breakdown of International Systems, Merriam Seminar Series on Research Frontiers (Denver: University of Denver Press, 1983); Dina Zinnes and Robert Mulcaster, “Hostile Activity and the Prediction of War,” Journal of Conflict Resolution 28, no. 2 (1984). 12. Philip Schrodt, “Richardson's N-Nation Model and the Balance of Power,” American Journal of Political Science 22, no. 2 (1978). 13. The position of the apostrophe in “prisoners'” is deliberate. There are two of them, and the word is plural. The dilemma only arises because there are more than one. It is, so to speak, jointly owned. 14. R. B. Braithwaite, Scientific Explanation (Cambridge: Cambridge University Press, 1955). 15. Karl Popper, The Logic of Scientific Discovery (New York: Basic Books, 1959). 16. This point is widely misunderstood, particularly by critics of an empiricist or positivist approach. Not all propositions within a theory need be directly observable. If this were required it would impoverish the natural as well as the social sciences. See Braithwaite, Scientific Explanation, and Michael Nicholson, Causes and Consequences in International Relations: A Conceptual Study (New York: Pinter, 1996). 17. This use of theory, at least in the sense of a scientific theory, is fairly generally accepted, as is the use of axiom. The words model and postulate are not always used consistently. Axiom is used as I (following Braithwaite) have used it though sometimes it is used also to include what I am calling a postulate. 18. Lewis F. Richardson, Statistics of Deadly Quarrels (Pittsburgh: Boxwood Press, 1960). 19. D'Arcy Thompson, Growth and Form (Cambridge: Cambridge University Press, 1917). 20. Harrison Wagner, “The Theory of Games and the Balance of Power,” World Politics 38, no. 4 (1986): 546–76. 21. Joseph Frankel, International Relations in a Changing World (Oxford: Oxford University Press, 1988). 22. The eminent economist Joan Robinson once remarked, “I knew no mathematics and therefore had to think instead.” This was meant as a jibe at the mathematical economists. Her work, by the standards of economics at the time and still more modern economics, is undoubtedly not mathematical. However, her skillful use of diagrammatic methods would put her firmly in the ranks of the formal theorists if they were applied to international relations even today. 23. Carl G. Hempel, Aspects of Scientific Explanation: And Other Essays in the Philosophy of Science (New York: Free Press, 1965). 24. Daniel S. Geller and J. David Singer, Nations at War: Scientific Study of International Conflict (Cambridge: Cambridge University Press, 1998). 25. This diagram is very minimally formal. There is no deduction involved nor does it contain anything that is not explained in natural language in the preceding paragraph. However, I think it makes the argument clearer, which it probably does for some readers and not for others. Its justification is purely heuristic.

26. I think this issue is important. I argue a similar point in “What's the Use of International Relations,” Review of International Studies 26, no. 2 (2000).Page 360 → 27. Constructivism emphasizes the interpretation and understanding of things and events and argues that these interpretations can alter, though not necessarily easily. Alexander Wendt argues this in The Social Theory of International Relations (Cambridge: Cambridge University Press, 1999). Of particular interest in light of Rapoport's approach is K. M. Fierke's Changing Games, Changing Strategies (Manchester, England: Manchester University Press, 1998). 28. Anatol Rapoport, “Lewis Fry Richardson's Mathematical Theory of War,” Journal of Conflict Resolution 1, no. 1 (1957): 249–99.

Page 361 →

CUMULATION, SYNTHESIS, AND RESEARCH DESIGN FOR THE POST—FOURTH WAVE Harvey Starr

Theory, Logic, and Method in the Post—Fourth Wave It is both fitting and useful for the International Studies Association to ask its members to reflect on where we are in the study of international/transnational/global politics and relations, and where the study might be going. It is similarly useful to think about areas of advance as well as the puzzles and problems that remain, and how we can do better. It is also appropriate to do so at the start not only of a new century but a new millennium. Thus, in a memo to the participants in the millennial reflection panels held at the 2000 International Studies Association annual meeting, Michael Brecher and Frank Harvey noted: The goal of the ten millennium Reflection panels, then, is to challenge proponents of specific paradigms, theories, approaches and substantive issue areas to confront their own limitations, stimulate debate about their most significant accomplishments and shortcomings, and discuss research paths for the years ahead. On the eve of a new millennium, scholars of International Studies have an opportunity and an obligation to engage in genuine self-criticism as a guide to identifying important theoretical, methodological and epistemological credentials and prospects of International Studies. While this point in time is a propitious one for reflection, the members of the quantitative methods panel have not had to wait for the millennium to engage in a variety of self-reflective and stock-taking activities. Their careerlong search for cumulation, better theory, and the best ways to test and evaluate that theory is quite evident. For example, more than a quarter of a century ago, I addressed stock taking and cumulation in the discussion of the application of quantitative methods to the study of international phenomena. 1 More than a decade ago, Ben Most and I addressed a set of impediments to cumulation, which included a self-critical appraisal of the failure of quantitative scholars to match theory and logic to research design and Page 362 → method. We identified these three elements—theory, logic, and method—as the components of the “research triad.” The research triad served as the core of our presentation, as we argued that “scholars need to recognize the existence of a research triad … and that each leg of this triad is critical for advancing our knowledge of international phenomena.”2 Scholars were likened to “jugglers” as “each element of the triad needs to be held in the air at the same time in a complex set of interrelationships, indicating that for the juggler to be successful, all of the balls (elements of the triad) must be kept going simultaneously.”3 Part of the charge to members of the millennial reflection panels was to identify and promote “intrasubfield consensus.” As such, I think that I can do no better than to continue the call for attention to the research triad and the promotion of creative and rigorous research design—no matter what subfield is under investigation, what specific theoretical framework is used to guide the research, or what specific methodologies are used to evaluate and compare theoretical frameworks. While Most and Starr pointed to a set of ongoing feedback loops among the elements of the triad, as well as the more specific elements of research design,4 the driving element always was theory. Thus, theory must be at the forefront of our endeavors, whatever the subfield or whatever the methodology employed. Clearly, I do not think this is unique to me or members of the quantitative methods panel; see, for example, the 1985 symposium in International Studies Quarterly, where international relations scholars from very different methodological and substantive camps all agreed on the centrality of rigorous theorizing.5 The think pieces we were enjoined to prepare for the millennial reflection panel in order to facilitate discussion and debate were asked to address one or more of six tasks. A careful reading of these tasks indicates that they are specifically related to questions of cumulation and synthesis. Specifically, task number two was “to assess where we stand on key debates and why we have failed to resolve them,” while task five was “to address the broader question of progress in international studies.” Task four was “to explore ways to achieve fruitful synthesis of

approaches.” In addition to drawing our attention once again to the research triad, these tasks lead me to maintain my position that researchers should use some form of an agent-structure theoretical framework, as I continue to see that approach as the most useful way to synthesize across levels of analysis, disciplines, and different forms of causal relationships. After directing some comments toward tasks two and five, I will argue for some form of agent-structure framework (partially addressing task six regarding an “agenda of topics”) and note that this approach presents us with three major methodological challenges. A final element of introduction is in order. As the latter decades of the twentieth century have been filled with various “post”-ists and “post”-isms, perhaps I can describe this essay as thinking about the study of international phenomena in terms of a “post-fourth wave.” In the 1960s Karl Deutsch characterized the study of international relations as going through four “waves,” starting first with international law, then to a second focusing on Page 363 → diplomatic history and international organization. Deutsch noted “a third wave of advance … underway since the 1950s. It has consisted in the reception of many relevant results and methods from the younger of the social and behavioral sciences such as psychology, anthropology, and sociology.”6 While the third wave continued, a fourth began in the 1960s: “the rise of analytic and quantitative research concepts, models and methods, a movement toward the comparative study of quantitative data and the better use of some of the potentialities of electronic computation.”7 The third wave led to questions with massive data requirements—data that previously had not been considered necessary or relevant. Obviously, if such data had not even been thought of, it had not been collected. Therefore, much of the activity in the third and fourth waves found itself concerned with data collection and analysis; much of the work of the fourth wave continued to be inductive and descriptive. It had also begun to be aware of, and take advantage of, the astounding capabilities of computers. I think we are in a post-fourth wave period that has been fueled by significant advances in computational and datahandling technology (both hardware and software) and by a massive influx of mathematical and formal models. I'm not sure that the stage we are now in clearly qualifies as a “fifth wave”—after all Lewis F. Richardson was building mathematical models in the decades before World War II (covering such phenomena as arms races or diffusion), and computers were being used in data analysis with regularity by the late 1950s and early 1960s. However, it seems clear to me that the range of the “analytic” techniques now being used, the frequency of their use, and their sophistication involve a qualitative (pardon the expression!) movement beyond the fourth wave that Deutsch described. The number and caliber of mathematical and formal models being employed and their technical complexity, investigating every level of analysis and choice situation (most based on microfoundations of economic choice), constitute a significant difference from Deutsch's fourth wave. A reading of the papers presented by Zinnes, Bueno de Mesquita, and Brams on the quantitative methods panel (in this volume) strongly confirms this view. The formal nature of the post-fourth wave is important, because it can be inclusive of, and synthesize, both the research triad and agent-structure approaches. In large part, the fourth wave continues with the tasks of empirically testing and evaluating these formal models. This is an important component of the fourth wave, and I think the extent of this empirical testing of formal models in international politics is more extensive than the picture presented by Zinnes in her paper. I do agree with Zinnes that “a mathematical model may be beautiful and provide wonderful insights, but it is not a theory about international politics until it has been subjected to an evaluation. Story and data must meet.” This is the heart of the research triad—the linkage between theory and research. Both the fourth wave and the post-fourth wave involve the testing of formal models, both for evaluation of existing models and the generation of better models (theory). Again, I agree with Zinnes that this is not easy. In Page 364 → many cases the task of testing formal constructs constitutes a formidable challenge to our ability to generate appropriate research design. However, in terms of the research triad, formal models both require and reflect a high degree of theoretical grounding. Formal models translate theory into mathematical and logical constructs. As such, they ensure the internal logical consistency of the formalized translation of theory.8 As noted, the challenge then confronting the researcher is somehow to capture these formalized relationships in empirical representations and evaluations. The relationships among theory, formalized models, research design, and research

results epitomize the research process described by Most and Starr,9 as well as the scholar who must “juggle” all three components of the research triad.

Cumulation/Progress and Theory Tasks two and five essentially address the issue of cumulation. The cumulation of knowledge in a specific discipline or area also served to focus the discussion in Most and Starr.10 Most and Starr argue that theory must be central to research and propel the research process.11 The centrality of theory in the research enterprise is also key to Zinnes's notion of “integrative cumulation”12 or Bueno de Mesquita's focus on “progress”13 in a Lakatosian sense. The obvious position any student of research design must support is that the methods selected by any researcher must be appropriate to the questions being investigated. Those questions, in turn, must be theory driven. Thus, as noted earlier, any researcher must begin with well-grounded and well-specified theory. The importance of theory is obvious. Russett and Starr write, “Theory organizes and simplifies reality; thus helping to separate the important from the trivial by pointing out what we really wish to look at and what is unimportant enough to ignore. This is why theory is so important—it affects not only which answers we come up with, but also what questions we ask in the first place!”14 Most and Starr stress the dynamic feedback loops that exist among theory, research design, and findings, as each informs and modifies the other.15 Theory thus not only affects the design and the research product but is itself continually modified and updated by the research process and the results of that process, in a dynamic combination of induction and deduction. As the research process unfolds and theory is modified, the researcher may be led to investigate additional phenomena originally omitted from the study. This view means that theory is central to the actual enterprise of empirical research. Given the current state of international relations “theory” at some U.S. institutions as well as abroad, it is important that basic issues concerning methodology (and epistemology) that link theory, research design, and the actual enterprise of research are discussed and returned to the forefront of the scholarly enterprise. This must be argued explicitly because, for some academics, the idea of “theory” has been, and too often continues to be, used to mean only broader philosophical worldviews or ideologies. As such, theory has been divorced from research by those who use the term in this way. Social scientists need to be reminded that theory's central contribution Page 365 → rests in the way in which it informs, shapes, and is in turn shaped by empirical research, that theory is a tool in the study of the world of politics. The theory-research loop is crucial here. This loop is irrelevant when theory is conceived of as ideology—where the answers are known, and, thus, research is irrelevant. Only when theory is seen as central to research, and not used as ideology, can it be the basis for integrative cumulation. This theory-research feedback loop was a prominent component of Most and Starr's Inquiry, Logic, and International Politics.16 The work in that book, and with other colleagues since then, was concerned with a set of key issues in the logic of inquiry that directly addressed this relationship. These issues also helped to comprise a truly critical logic that challenged researchers to examine the theoretical bases of their research. This involved clarifying the logical consistency of the relationship between theory and research design. It also involved the researcher considering the logical structure of the components of his or her research designs. In sum, the set of issues presented by Most and Starr that comprised the critical logic included the form of the relationship under study; matching the logic of theory/models to the logic of research design; development of process models; broader versions of the agent-structure problem; identifying the appropriate units of analysis; case selection; and such vital concepts as “substitutability” and “nice laws.” Perhaps most important for both cumulation and synthesis, the critical logic—along with its subcomponents and its attendant issues—was not limited to any specific area of international politics or political science.17 Nor is it limited to any specific methodological tradition. This critical logic is indeed able to deal with most critiques that are directed at systematic empirical work from outside the scientific method-based tradition of inquiry.18 That is, most of the critiques raised by those critical of “positivism” (always an ambiguously used concept) can be subsumed under issues raised in the critical logic of Most and Starr. Using the critical logic of Most and Starr, the scholar need not “stand outside” the scientific method tradition to raise and deal with such issues as “choice posing as truth” or “reality as a social construction” or the broad issue of relativism.19 To repeat, the key point

here is that the critical logic and its components apply across methodologies and to all aspects of research design—as long as one's true enterprise is to investigate some aspect of the empirical world.

The Synthetic Use of Agent-Structure Approaches: Three Challenges The opportunity-and-willingness construct20 is based on the work of Harold and Margaret Sprout. This construct is simply one version of a more general agent-structure approach and is captured in the “ecological triad” of the Sprouts. The ecological triad consists of the entity, the environment, and the entity-environment relationship. The first challenge I want to raise reflects a major methodological problem scholars currently face—and will constitute a major future challenge; it is how to get at the entity-environment Page 366 → relationship. This is, again, simply one version of the more general “agent-structure problem.”21 The original article by Starr (“‘Opportunity’ and ‘Willingness’ as Ordering Concepts in the Study of War”) was only vaguely suggestive of the ways in which opportunity could relate to willingness and how willingness might relate to opportunity. Yet the nature of these (inter)relationships is essential to theory building and testing. As Elster notes in discussing “opportunities” and “preferences”: “Scholars disagree on the relative importance of preferences and opportunities in explaining behavior.”22 This was a key question throughout Most and Starr's Inquiry, Logic, and International Politics, one that was approached in a number of ways. While a set of logical analyses demonstrated the joint necessity of opportunity and willingness in explanation, other parts of the book, such as the simulation analyses of system structure and willingness in chapter 6, began to lead toward a focus on willingness that is most strongly articulated in Friedman and Starr's ultimate development of an empirical framework based on methodological individualism.23 The relationship between opportunity and willingness is further developed in Cioffi-Revilla and Starr.24 Here, through the use of Boolean logic, the authors demonstrate how opportunity and willingness as elements in first-order causality are related to substitutability, which is central to second-order causality. Looking at opportunity and willingness and then adding the important auxiliary concept of substitutability presented by Most and Starr, Cioffi-Revilla and Starr distinguish between a first-order causality of world politics (willingness and opportunity) at the analytical level and a deeper second-order causality of substitutability at the operational level. The first-order (necessary) elements of opportunity and willingness are linked by the Boolean and, while the range of possible modes of (sufficient) second-order substitutability are connected by the Boolean or. Cioffi-Revilla and Starr then formalize and analyze the political uncertainty of international behavior, along with willingness, opportunity, and their substitutability, at both the analytical and operational levels. In so doing, they mathematically derive a number of interesting insights into the agent-structure problem, especially in regard to the relationship between agency and structure. This article clearly demonstrates the many possible ways in which the opportunity-and-willingness framework can be developed and the important nonintuitive implications that can be derived through formal analyses. Friedman and Starr investigate the opportunity-and-willingness framework as it relates to the broader debate in the literature concerning the agent-structure relationship.25 The agent-structure problem refers to questions concerning the interrelationship of agency and structure and to the integration of this dialectical relationship into the explanation of social phenomena. The agent-structure problem raises a variety of issues in the philosophy of science and the logic of inquiry. Friedman and Starr argue that the current treatment of the agent-structure problem in the international relations literature is problematic. Using the opportunity-and-willingness framework and related concepts such as substitutability, they critique the Page 367 → confusion among ontological, epistemological, and methodological issues and analyses. For example, they argue that the postulation of causal interrelationships among agency and structure requires that these concepts be defined autonomously and encompass explicitly delimited parameters of variability. The structurationist view that each must be endogenized is incomplete, ignoring the basic nonrecursive nature of agency and structure (opportunity and willingness). The use of opportunity and willingness also promotes recognition that structuration-theoretic approaches to international politics inadequately attend to the variability of agency identity and interest, and to the social structural foundations of this variability. From these and other metatheoretic considerations, Friedman and Starr derive a methodological individualist model of the dialectical interrelationship between multidimensional and

dynamic sociopolitical contexts and individual international political elite choice processes. This methodological individualist perspective is clearly picked up by models of expected utility, prospect theory, and other formal models of choice, where factors of structure/opportunity are built into the payoff structure, resource constraints, the form of the game, and so on. This is one reason why I have found these approaches so useful and why developments in methodology must continue in the areas of choice (agency) under a variety of conditions of constraint (structure). A key problem in getting at the opportunity-and-willingness relationship is often due to the lack of variance within the opportunity structure. That is, it is very difficult to find enough cases where the different structural or environmental conditions necessary to evaluate a model exist. Recall also that Most and Starr in their discussion of necessity, sufficiency, and research design note that it is not enough to study the presence of some independent variable x or some dependent variable y, but not-x and not-y as well.26 It is for this reason that in both Most and Starr and in the ongoing Two-Level Security Management Project,27 I have argued that the use of simulations can be most helpful, and have moved in that direction. Given the lack of cases and variance,28 simulations permit the control and experimentation necessary to look for causal relationships. Using simulation we are able to analyze a model for its internal consistency and to examine the plausibility of its assumptions by generating implications for hypothetical systems of states (or any other set of “environed entities”). Simulations, such as that created for the Two-Level Security Management Project, can demonstrate that the model's predictions conform to the implications of other accepted theories and to the empirical findings of previous studies. And, as with other formal models, they can reveal nonintuitive relationships.29 A second challenge, briefly noted earlier, concerns the form of the relationships investigated with an agentstructure model. That is, given opportunity and willingness, substitutability, and nice laws, how do we construct research designs to deal with the logical differences between models of necessity and sufficiency? In particular, how do we conduct the empirical analysis of necessity? For example, analysts such as Frank Harvey30 have usefully Page 368 → and appropriately raised the issue that theories or models of deterrence are models of necessity not sufficiency but that critics of those models test them with designs only appropriate for sufficiency. It is interesting to highlight the parallels between critics of rational models of deterrence and the critics of Bueno de Mesquita's expected utility theory, as discussed by Most and Starr in Inquiry, Logic, and International Politics. In both cases analysts are proposing models of how people are supposed to behave if certain processes are at work. Neither process is itself observable; indeed, deterrence is about actions that do not occur! Both “theories” set out conditions that must be present in order for certain outcomes—the results of these processes—to occur: that is, necessary conditions. Neither theory is presented in an “if-then” structure that represents sufficiency. However, while the conditions may obtain, the proposed results may not appear. This startles the critics of such approaches! Yet this is exactly what is to be expected with models of necessity. And it is what makes them difficult to test. One of our ongoing challenges, then, is to design methods to study nonobservable phenomena or outcomes that do not occur. As Most and Starr point out in regard to expected utility, the number of cases that occur in a cell where the conditions of expected utility obtain but no conflict initiation occurs is irrelevant.31 And, again, we're trying to analyze behavior that does not occur. Note carefully that this corresponds directly to arguments by Lebow and Stein,32 and similar critics, that deterrence theory fails because while all the conditions for deterrence are present, there is “no deterrence.” In Bueno de Mesquita's expected utility analysis, no conflict initiation reflects the nonoccurrence of behavior. With deterrence, critics do not realize they are in the same boat, because a lack of deterrence is reflected in some party actually behaving—initiating an attack or making a challenge of some kind. Using the research triad of logic, theory, and research design, it is clear what is happening. The critics of such models of necessity think that the sort of behavior they can see permits them to treat the relationship as a sufficient relationship. It does not. Note that by using deterrence we can illustrate the methodological challenges posed by the study of necessity: a. if deterrence is successful there is no behavior to see; b. if deterrence fails behavior does occur and can be observed; c. thus, critics study deterrence failure because it is observable;

d. in so doing they think they are solving the not-y problem, which must be solved to establish sufficiency; e. they conclude that the factors of rational deterrence theory do not predict to successful deterrence, that either there is no relationship or there is no sufficient relationship. The former is not true; the latter is essentially irrelevant. Logic has helped us understand these deterrence debates. Some promising research designs, following Harvey and others, include the use of Boolean analyses and comparative studies in order to study necessity. Page 369 → The third challenge concerns how we are to investigate the phenomenon of substitutability (and the related idea of nice laws). As noted in Cioffi-Revilla and Starr, substitutability is closely related to opportunity and willingness, providing multiple paths to either opportunity or willingness—“alternative modes of redundancy.”33 More generally, substitutability indicates that any single cause may have multiple effects and that any single effect may have multiple causes. Different problems may lead to similar responses or that a given problem may be dealt with in multiple ways. Thus, for any given set of conditions, the same policy choice is not equally likely across time or space for elites.34 Substitutability (as well as nice laws) thus lies at the heart of the use of agent-structure frameworks, where the choices of agents are analyzed in regard to the contextual structures within which they exist. It is not simply that substitutable modes (or redundancy) are related to greater “complexity” in the world; these modes actually make political behavior in a complex world more possible. The combination of opportunity and willingness with substitutability as a way to deal with the agent-structure problem reveals the complexity and uncertainty in international phenomena (a complexity that has not been adequately addressed by either realism or neorealism or most variants of transnational liberalism). Yet as CioffiRevilla and Starr demonstrate, “the fundamental uncertainty of politics is scientifically tractable.”35 With concepts such as nice laws, substitutability, context, and adaptation, the opportunity-and-willingness framework can deal with both the analytical and operational levels of theory and give us a better grasp as to what behaviors are more or less likely to occur under what conditions. The challenge of how to study substitutability is discussed at length in Starr as part of a special issue of the Journal of Conflict Resolution on the use of substitutability in the study of international politics.36 The recognition of substitutability complicates research design as it forces the analyst out of single-variable or single-indicator ways of thinking about political phenomena. Only by understanding the broader concept or phenomenon under study could one fully grasp exactly what one should be investigating in terms of the full range of operational observations. If some range of behaviors were all substitutable means for achieving some foreign policy outcome, then the whole range would need to be included in any particular study. Focusing on only one would mean a failure to provide full coverage of the possible outcomes and lead to incomplete results that failed to cumulate (or even make sense when compared). The results would fail to capture the theory or model being tested (as only one part was being tested). And, returning to the logic component of the research triad, focusing on only one possible outcome (the “successful” outcome that is observed) leads to the logical problems that exclude the study of sufficient relationships. The authors of the special issue of the Journal of Conflict Resolution provide important clues as to how we can study substitutability. Some note that the status quo, or “no action,” must be part of the model. They also argue that the substitutability problem “translates” to a time-series situation, Page 370 → concerned with sequences of responses that require data dealing with temporal variance.37 With such time-series needs in mind, the best techniques to deal with empirical estimation are sought. Both Regan and Bennett and Nordstrom argue that the use of multinomial logit analysis is the most appropriate way to deal with the data in substitutability situations: analyses where the dependent variable is presented in multiple, unordered categories. The possibility of multiple outcomes that capture the same overriding concept—such as security or viability—clearly presents the analyst with substantial problems of research design. Choices that have not been taken need also to be considered in our

analyses, as difficult as that may be. The fact that there are multiple possible outcomes—directed at different internal and external audiences playing two-level games—may also complicate analyses that are based on a straightforward trade-off between two possible policies or actions.

Conclusion I have briefly—too briefly—tried to indicate some of the methodological challenges that face students of world /global/international politics/relations as we enter the twenty-first century. Cumulation and “progress” in the study of global phenomena will depend on the quality and rigor of our theories and our methods. Synthesis will follow broad agent-structure approaches that cut across more standard levels of analysis and disciplinary boundaries. The challenges facing researchers arise from finding the appropriate methods by which to study the agent-structure problem. These include how to cut into the continuous feedback loops between agent and structure—between endogenous and exogenous factors; how to design studies of necessity; how to find the correct methodological tools for studying phenomena that do not occur, for studying not-x and not-y and not just selecting on either the independent or dependent variables; and how to craft the proper designs and methods to study substitutability. I also think we must encourage the use of computer simulations to overcome problems of variance in opportunity /structure. All of the aforementioned will be needed to test fully the empirical implications of the formal models that are being developed to investigate literally every aspect of world/global/international politics/relations. Formal models, whether developed from economics, game theory, expected utility theory, or other sources will most certainly constitute the main analytical approach of the post-fourth wave, as a natural means to synthesize agency and structure, and choice (willingness) given constraints (opportunity). The prospects for continued cumulation and progress in the expansion of our knowledge are both daunting and exciting. Page 371 →

Notes 1. Harvey Starr, “The Quantitative International Relations Scholar as Surfer: Riding the Fourth Wave,” Journal of Conflict Resolution 18, no. 2 (1974): 336–68. 2. Benjamin A. Most and Harvey Starr, Inquiry, Logic, and International Politics (Columbia: University of South Carolina Press, 1989), 2. 3. Ibid., 10. 4. Gil Friedman and Harvey Starr, Agency, Structure, and International Politics (London: Routledge, 1997). 5. Bueno de Mesquita began his reply to the other articles by saying, “I am greatly encouraged by the extent of agreement regarding the advantages of axiomatic, deductive theorizing over ad hoc hypothesis testing reflected in the exchange between myself, Robert Jervis and Stephen Krasner. Convergence on this critical epistemological issue is far more important than the remaining differences over quantification, case study analysis, and the like.” Bruce Bueno de Mesquita, “Reply to Stephen Krasner and Robert Jervis,” International Studies Quarterly 29, no. 2 (1985): 151. 6. Karl Deutsch quoted in Starr, “The Quantitative International Relations Scholar as Surfer,” 345. 7. Deutsch quoted in ibid., 345. 8. In discussing the strengths and contributions of game theory in his quantitative methods panel paper (in this volume), Bueno de Mesquita includes game theory's reliance on explicit logic, along with its transparency in assumptions, reasoning, and propositions. These features ensure that the scholar must be involved in rigorous theorizing. 9. Most and Starr, Inquiry, Logic, and International Politics, ch. 1. 10. Ibid.

11. See also Harvey Starr, “Visions of Global Politics as an Intellectual Enterprise: Three Questions without Answers,” in Visions of International Relations, ed. Donald J. Puchala (Columbia: University of South Carolina Press, 2000). 12. Dina A. Zinnes, “The Problem of Cumulation,” in In Search of Global Patterns, ed. James N. Rosenau (New York: Free Press, 1976). 13. Bueno de Mesquita, “Reply to Stephen Krasner and Robert Jervis” or Bruce Bueno de Mesquita, “The Contribution of Expected-Utility Theory to the Study of International Conflict,” in Handbook of War Studies, ed. Manus I. Midlarsky (Boston: Unwin Hyman, 1989). See also “Symposium: Methodological Foundation of the Study of International Conflict,” in International Studies Quarterly 29, no. 2 (1985): 121–54. 14. Bruce Russett and Harvey Starr, World Politics: The Menu for Choice, 5th ed. (New York: Freeman, 1996), 30. 15. Most and Starr, Inquiry, Logic, and International Politics. 16. Ibid. 17. See, for example, the title of Claudio Cioffi-Revilla and Harvey Starr, “Opportunity, Willingness, and Political Uncertainty: Theoretical Foundations of Politics,” Journal of Theoretical Politics 7, no. 4 (1995): 447–76. 18. See especially Friedman and Starr, Agency, Structure, and International Politics. 19. These are among the “five major insights which constitute the promise of postmodernism,” which are presented in John Vasquez, “The Post-Positivist Debate: Reconstructing Scientific Enquiry and International Relations Theory after Enlightenment's Fall,” in International Relations Theory Today, ed. Ken Booth and Steve Smith (University Park: Pennsylvania State University Press, 1995), 218. While it would take another paper to demonstrate fully, I hold that the critical logic (and/or positions it can represent or lead to) is capable of dealing with each of these insights. It is not that postmodernism does not have useful things to say, only that such comments can be generated by critical investigation from within the scientific tradition. For a much fuller development of this argument, see Friedman and Starr, Agency, Structure, and International Page 372 → Politics. I have not been persuaded by the postpositivist or postmodernist position that the world eludes our ability to study it in any intersubjective way. Individuals do so every day; scholars across the social sciences and natural sciences have done so successfully for decades (if not centuries). See Pauline Rosenau, Post-Modernism and the Social Sciences (Princeton: Princeton University Press, 1991); Vasquez, “The Post-Positivist Debate” Oyvind Osterud, “Antinomies of Postmodernism in International Studies,” Journal of Peace Research 33, no. 4 (1996): 385–90; and Oyvind Osterud, “Focus on Postmodernism: A Rejoinder,” Journal of Peace Research 34, no. 3 (1997): 337–8, for examples of the counterarguments that I find compelling. 20. See Harvey Starr, “‘Opportunity’ and ‘Willingness’ as Ordering Concepts in the Study of War,” International Interactions 4, no. 4 (1978): 363–87; and Most and Starr, Inquiry, Logic, and International Politics, ch. 2. 21. Friedman and Starr (Agency, Structure, and International Politics, 3) elaborate: Agency and structure are the defining components for the understanding of human interaction within a society and of the explanation of social phenomena. The agent-structure problem refers to the general set of questions concerning the interrelationship of these two components, and to the ways in which explanations of social phenomena integrate them. International political systems, like all social systems, are comprised of agents and structures. What is more, agency and structure are interrelated. This basic tenet of social theory is shared by the three most widely acclaimed modern social theorists—Durkheim, Weber, and Marx. See also Alexander Wendt, “The Agent-Structure Problem in International Relations Theory,” International Organization 41, no. 3 (1987): 335–70. 22. Jon Elster, Nuts and Bolts for the Social Sciences (Cambridge: Cambridge University Press, 1989), 15. 23. Friedman and Starr, Agency, Structure, and International Politics.

24. Cioffi-Revilla and Starr, “Opportunity, Willingness, and Political Uncertainty.” 25. Friedman and Starr, Agency, Structure, and International Politics. 26. Most and Starr, Inquiry, Logic, and International Politics. 27. For example, Marc V. Simon and Harvey Starr, “Extraction, Allocation, and the Rise and Decline of States,” Journal of Conflict Resolution 40, no. 2 (1996): 272–97; and Marc V. Simon and Harvey Starr, “A Two-Level Analysis of War and Revolution: A Dynamic Simulation of Response to Threat,” in DecisionMaking on War and Peace: The Cognitive-Rational Debate, ed. Nehemia Geva and Alex Mintz (Boulder, Colo.: Lynne Rienner, 1997). 28. For example, in looking for systems with different polarity, as in Most and Starr, Inquiry, Logic, and International Politics, ch. 6; or the different conditions regarding internal and external conflict and the rise and decline of states, as in Simon and Starr, “Extraction, Allocation, and the Rise and Decline of States”; or the number of democracies in the system, the number of endangered democracies, and countries with different configurations of neighbors, as in Marc V. Simon and Harvey Starr, “Two-Level Security Management and the Prospects for New Democracies: A Simulation Analysis,” International Studies Quarterly 44, no. 3 (2000): 391–422. 29. Additionally, as Simon and Starr note: “In a sense, we use simulations as what Eckstein (1975) called a ‘plausibility probe’ to determine what kind of empirical work is warranted—to lead to further empirical analysis.” Simon and Starr, “Extraction, Allocation, and the Rise and Decline of States,” 275. 30. Frank Harvey, “Rigor Mortis or Rigor, More Tests: Necessity, Sufficiency, and Deterrence Logic,” International Studies Quarterly 42, no. 4 (1998): 675–707; Frank Harvey, “Practicing Coercion: Revisiting Successes and Failures Using Boolean Logic and Comparative Methods,” Journal of Conflict Resolution 43, no. 4 (1999): 840–71. 31. Most and Starr, Inquiry, Logic, and International Politics, 56–7.Page 373 → 32. Richard N. Lebow and Janice Gross Stein, “Rational Deterrence Theory: I Think Therefore I Deter,” World Politics 41, no. 1 (1989): 208–24. 33. Cioffi-Revilla and Starr, “Opportunity, Willingness, and Political Uncertainty,” 456–57. 34. See, for example, Patrick M. Regan, “Substituting Policies During U.S. Interventions in Internal Conflicts,” Journal of Conflict Resolution 44, no. 1 (2000): 90–106. Nice laws are similarly related to context and the complexity of choice within complex notions of context. This term refers to “‘sometimes true,’ domain specific laws” (Most and Starr, Inquiry, Logic, and International Politics, 98). Most and Starr remind us that while we should aim for generality, the “right type of law” is one that is clearly specified, that the relationships among variables that it proposes will work only under specified conditions. Most and Starr question whether social scientists will ever generate important “universal” laws. They note, however, that “it may be useful to recognize that there could very well be laws that are in some sense ‘good,’ ‘domain specific,’ or ‘nice’ even though the relationships they imply are not necessarily very general empirically.… it may be more productive to think of laws each of which is always true under certain conditions (or within certain domains) but which is only ‘sometimes true’ empirically because those conditions do not always hold in the empirical world.” Most and Starr, Inquiry, Logic, and International Politics, 117. 35. Cioffi-Revilla and Starr, “Opportunity, Willingness, and Political Uncertainty,” 468. 36. Harvey Starr, “Substitutability in Foreign Policy: Theoretically Central, Empirically Elusive,” Journal of Conflict Resolution 44, no. 1 (2000): 128–38; See Glenn Palmer, ed., Substitutability in Foreign Policy: Applications and Advances, special issue of Journal of Conflict Resolution 44, no. 1 (2000). 37. See Regan, “Substituting Policies During U.S. Interventions in Internal Conflicts” and D. Scott Bennett and Timothy Nordstrom, “Foreign Policy Substitutability and Internal Economic Problems in Enduring Rivalries,” Journal of Conflict Resolution 44, no. 1 (2000): 33–61.

Page 374 →

ACCOMPLISHMENTS AND LIMITATIONS OF A GAMETHEORETIC APPROACH TO INTERNATIONAL RELATIONS Bruce Bueno de Mesquita Game-theoretic reasoning emerged as an important form of analysis in international relations, especially regarding security studies, with the publication in 1960 of Thomas Schelling's The Strategy of Conflict. The remainder of the 1960s saw a proliferation of carefully reasoned, policy-relevant studies grounded in the techniques of formal logic and rational choice modeling. A small sampling includes Daniel Ellsberg's application of expected utility reasoning to deterrence strategy, Martin McGuire's investigation of secrecy and arms races, Anatol Rapoport and Albert Chammah's examination of the prisoner's dilemma in the context of cooperation and conflict, Bruce Russett's study of the calculus of deterrence, and many others.1 Over the course of the next thirty or forty years, game theory became analytically more sophisticated, contributed greatly to the acquisition of cumulative knowledge in international relations, and gradually emerged as a tool for generating hypotheses worthy of empirical study and as a contributor to policy debate. Game-theoretic analyses today increasingly are closely linked to empirical tests and to questions of great policy significance. This essay summarizes one person's views of the strengths and contributions of game theory as a tool for studying international relations and assesses the limitations inherent in the current state of the science. I do not review individual studies here in any depth, focusing instead on broad patterns and generalizations. The interested reader should see several excellent reviews and evaluations of this literature, as well as my own lengthier discussions of it.2

Contributions of Game Theory to Studies of Social Phenomena Game theory, and rational choice models more broadly conceived, make at least five contributions to the study of human interaction, including the study of international relations. These include (1) the provision of a motivational foundation for action; (2) an explanation of strategic behavior; (3) an ability to integrate other theories and approaches; (4) cumulative knowledge; Page 375 → and (5) insights into methodological issues influencing analyses of decision making. With the possible exception of its attentiveness to strategic decisions, game theory cannot claim a unique position in advancing any of these components of the study of international relations. No theory or methodology can make such a claim. But formal models, particularly game-theoretic models and spatial models, offer specific and important advantages in contributing to progress on all five of the fronts I have mentioned. What is more they have succeeded in doing so despite the fact that only a tiny fraction of international relations scholars employ these models, perhaps at present not more than two dozen out of, for instance, the more than three thousand members of the International Studies Association. In the pages that follow, I explore this claim, highlight important achievements, and then turn to a discussion of some of game theory's limitations.

Motivation for Action Rational choice approaches provide a motivational foundation for action so that the link between correlation and causation is readily discerned. Rational choice models of international relations stipulate that decision makers are interested in maximizing their welfare. Theorists differ in their perspective regarding the content behind this motivation, with some looking at the maximization of the national interest,3 while others investigate individual decision-maker interests,4 thereby distinguishing the welfare of leaders from the well-being of their states. By emphasizing motivations, such models ensure that political strategy enters into the explanation of international affairs, with emphasis on how domestic and foreign institutions, and international structures, constrain, but do not determine, behavior. The assumption that decision makers are rational is not a subject of inquiry in models that accept this perspective but rather an initial condition that, in conjunction with other assumptions, leads to implications about choices and

actions. The rationality assumption says no more than that decision makers do what they believe is in their best interest given the information they possess at the time a choice is made. The “other assumptions” can relate to whether the “decision maker” is a state treated as a unitary actor, a group with shared interests, or individuals with perhaps competing or conflicting interests. Additionally, a rational choice theory might stipulate that all actors maximize the same thing, whether it be national power, security, or welfare, or the individual's term in office, policy goals, religious fervor, and so on. Alternatively, different actors might be assumed to maximize different interests or different goals, such as relative gains or absolute gains, personal political gain or national security, and so on. The assumption of rationality does not restrict the range or complexity of objectives decision makers want to achieve in theory or in reality. Despite its flexibility, the rationality assumption helps formulate the microfoundations for causal arguments by tying action to objectives. Other approaches, such as psychological theories of cognitive dissonance, groupthink, Page 376 → and so on, can also help provide microfoundations so that they form part of a relatively small set of approaches that facilitate a focus on what motivates action. Ultimately, the explanations offered by alternative theories can only be judged on the basis of their logical integrity and their empirical bite. Although some might argue that internal consistency is not critical,5 without it we can make no sense of what a theory does or does not claim. Likewise, without empirical testing we can make no headway in choosing among internally consistent theories that make different predictions about behavior and outcomes.

Strategic Interaction Game theory is a body of reasoning designed explicitly to attend to the logic of strategic interaction. Strategic interaction, in which decision makers select a course of action taking into account expectations about how others will respond, is central to all of international affairs. One consequence of the emphasis on strategic interaction is that propositions derived from game theory models are contingent hypotheses. This is true because any decision maker's actions are contingent on what he or she believes will prompt the most favorable response. Therefore, what one chooses to do depends on how others are expected to respond. Non-game-theoretic points of view often assume away or greatly simplify the most interesting features of strategic interaction in international affairs. In the course of doing so, they frequently make noncontingent, linear predictions. For example, it is commonplace to find realists arguing that greater military advantage over an adversary is always better than less advantage.6 Yet there are numerous instances of weaker states emerging victorious in military confrontations with stronger states. The observable military balance is contingent on a willingness to use available resources. Asymmetries in motivation, for example, can offset the apparent military superiority of one or the other combatant.7 Of course, it is not necessary to use game theory to form contingent hypotheses, but because all game theory models rely on the logic of strategic interaction, the use of game theory is sufficient to make contingent predictions. The very idea of equilibrium in game theory depends on contingent expectations. Decision makers consider alternative courses of action. They weigh the expected consequences of choosing one path or another, perhaps across many choices. In each case, they consider what they believe other decision makers will do, and then they, like all decision makers, choose the action that they think most advances their goals. Their actions are restricted to the set that enhances their welfare over all other actions, not at every step of the way, but in terms of the expected payoff at the end of the game. Psychological theories, like game theory models, provide motivations for action, thereby fulfilling the first of the five conditions I discuss. They do not, however, provide clear and explicit rules for decision making in situations involving strategic interaction. That is, psychological theories lack an equivalent to game theory's notion of equilibrium. Even as highly developed Page 377 → and formalized a psychological theory as prospect theory contains no foundation from which to discuss how anticipated responses by others shape decision making. For example, consider the well-known experimental observation that individuals told by a physician that they have a 10 percent chance of recovery (and, implicitly, a 90 percent chance of dying) given a proposed medical procedure are unlikely to choose the procedure, whereas those told they have a 90 percent chance of recovery (and,

implicitly, a 10 percent chance of dying) tend to elect the procedure. Prospect theorists, lacking a theory of strategic interaction, point to this result as evidence that standard expected utility axioms are violated. Yet a simple signaling game readily predicts the observed behavior. Imagine a physician bound by professional ethics to report accurately to her patients the average recovery rate from a medical procedure. The physician may believe that her patient has a better or worse than average chance of recovery but is restricted to report the average recovery rate. By choosing to frame the rate as a recovery rate or a failure rate, the physician conveys additional information that she believes will help the patient make a more informed choice. The patient, comprehending the signal, takes the additional information into account and chooses accordingly. Because prospect theory and other psychological theories lack a well-formed approach to strategic interaction, they are unable at this time to achieve the subtlety of game-theoretic accounts when choices involve strategy or the transmission of private information. Instead, they interpret their experimental results as a challenge to all models grounded in expected utility's axioms, whereas many—though perhaps not all—of the experimental findings make perfectly good sense in a setting that includes strategic interaction. For problems involving strategic choice—and this surely includes most interesting problems in international relations—game theory is significantly advantaged over other approaches.

Integrating Approaches Models based on game theory facilitate incorporation of the concerns and assumptions of structural, behavioral, constructivist, and psychological theories and so can help to integrate the important knowledge derived from these other approaches. Game theory readily synthesizes knowledge from diverse perspectives while retaining the logical rigor imposed by its axiomatic foundations. It does so because these other approaches do not contradict game theory assumptions. Structure is a central element in games of sequential decision making in which choices are constrained by the situation decision makers find themselves in. Indeed, a significant body of research exists revolving around the theme of structure-induced equilibria. This literature examines how structures (such as bipolarity or multipolarity, balance of power or preponderance of power, etc.) and political institutions (such as democracy, autocracy, or monarchy, federalism or unitary governance, etc.) shape choices, influencing the strategies decision makers reject or accept. Page 378 → Through a strategic analysis of preferences and beliefs game theory provides a means to examine attitudes, perceptions, uncertainties, and learning on the part of decision makers. At the same time, game theory provides a systematic means of analyzing and predicting behavior across large classes of events whether they involve sincere behavior, credible commitments, bluffing, or other forms of strategic decision making. These all seem to be crucial elements in the study of international affairs. Other approaches may take these features into account, but game theory is the only method with which I am familiar that explicitly requires attentiveness to all of these concerns. Having said that, we should keep in mind that there is not a single game theory of international relations. Rather, game theory is a mathematical foundation from which to construct different, often competing theories. As such, game theory is an axiomatically based tool for studying decision making. In international relations, game theory provides a method for explaining and analyzing strategic behavior. The quality of any given theory, of course, depends on the insights of the researcher. Game theory as a method is beneficial because it ensures that crucial factors, including the ones I have enumerated, as well as explicit, transparent arguments, are taken into account. With normal language it is easy for clever rhetoric to mask logical inconsistencies. With the greater precision of mathematics, formal models reduce the risk that seemingly consistent arguments substitute for actually consistent arguments. Since inconsistent arguments are incoherent, either “predicting” everything or nothing, the explicit logic of game-theoretic analysis is a boon to weeding out ill-conceived theories while still leaving room for the proverbial thousand flowers to bloom.

A few examples help illustrate how rational choice models, especially game-theoretic ones, clarify logical relationships between seemingly competing theories or help separate internally consistent implications from less rigorously surmised associations within extant theories. Power transition theory purportedly predicts that the likelihood of major, system-transforming wars increases when rivals are approximately equal in power, one is growing faster than the other, and the rivals disagree over the rules and norms governing international intercourse.8 Kim and Morrow, Powell, and Bueno de Mesquita and Lalman offer formal representations of power transition theory.9 Each finds support for some of that theory's core hypotheses, but each also raises doubts about other, central portions of the power transition argument. None of the formal models, for instance, finds that equal power between rivals is either necessary or sufficient for war. The formal models suggest new propositions as well as challenging some original ones, and point to more nuanced tests (on, for example, differences in risk-taking propensities between challengers and targets that are associated with war). Power transition predictions about allies are shown to be a special case of a more general argument about allies and war, an argument that subsumes the seemingly contradictory predictions stated by power transition theorists and balance-of-power theorists.10 Page 379 → Though not unique in its ability to integrate ideas or compare arguments, game theory's formalism conveys an advantage. To assess the logical implications of alternative assumptions or alternative statements of a theory, it is necessary first to stipulate precisely what the theory says or what it includes and excludes. When assumptions and their logical connections are explicit, it is straightforward to alter one or another component of a theory to ascertain how the theory's predictions are modified by the change. In doing so, we learn whether the theory's predictions follow from its assumptions; follow but are not robust against minor changes in the assumptions; follow and are robust to changes in assumptions; or do not follow. This kind of systematic, detailed evaluation of a theory is hard to do and, therefore, is rarely done outside the formal literature in international relations. Within that literature, however, it is common.11

Cumulative Knowledge No approach or methodology corners the market on generating cumulative knowledge about international relations. Game-theoretic models, because of their formalism, however, offer some important advantages in the efficient discovery of cumulative knowledge. They do so because explicit assumptions make it easy to determine what the logical consequences are of relaxing one or another condition, substituting alternative assumptions, or restricting the theory further by adding new assumptions as discussed earlier. In this way, anomalous cases can be carefully explored and modified theories constructed that may account for known facts and the seeming anomalies, while also uncovering new facts. Fearon's audience costs argument has just this quality, as does Powell's research into guns versus butter or Morrow's assessment of endogenous regime formation.12 The past few years have witnessed the development of a body of research dedicated to testing competing rational actor models, thereby helping to distinguish the empirical content of alternative formulations.13 Reiter and Stam, for instance, contrast competing explanations of the democratic peace.14 Some argue that democracies win wars more often than do their adversaries because democratic norms lead to greater support for war efforts among the citizenry. Others argue that democracies win more often because leaders risk losing their jobs if their state is defeated in war and so select the wars to enter carefully, negotiating and offering concessions in those instances where they do not believe they will prevail militarily. Because at least some of these arguments have been formulated in explicit, game-theoretic terms, the contingencies under which democracies are expected to fight and win are spelled out with sufficient detail to construct tests of the alternative explanations. Gelpi and Grieco15 also contributed significantly to this debate by identifying conditions under which audience effects, as predicted by Fearon, are expected to increase the willingness of democracies to fight and when institutional constraints, as predicted by Bueno de Mesquita and Lalman, diminish Page 380 → the willingness of democracies to fight. These at least partially competing hypotheses (depending on the contingent conditions) both derive from formal, game theory models of international interaction. Smith has similarly expanded the general understanding of international interaction, specifically regarding the reliability of alliances, by using game theory

to focus attention on the interdependency between alliance formation, alliance reliability, and conflict initiation.16 Schultz and Morrow have expanded our understanding of the nexus between domestic affairs and international conflict in their models in which decision makers maximize personal welfare rather than national welfare.17 These studies raise significant questions about the common notion that partisanship should end at the water's edge, suggesting political and national benefits to be derived from partisan competition in foreign policy. The tie between the formal literature and empirical tests has led to more sophisticated testing and to progress in understanding the contingencies under which war is more or less likely,18 under which alliances are more or less reliable,19 under which threats of sanctions are likely to succeed or fail,20 under which deterrence is effective or ineffective,21 and so forth. The empirical track record associated with game-theoretic approaches to international relations is growing in breadth, depth, and the significance of findings.

Methodological Insights Game theory contributes broadly to international relations research, even among non-game theorists, by drawing greater attention to endogenous choices and their implications for path dependence; selection effects in theory and in data and how they can distort inferences from historical analysis; the importance of independence between arguments and the evidence used to evaluate their merits if we are to distinguish between description, explanation, and prediction; and prediction as a means of evaluating the potential of scientific inquiry to help improve future international affairs. These four items—endogeneity, selection effects, independence between argument and evidence, and prediction—represent areas where rational choice models are especially helpful in clarifying empirical problems that frequently arise in other modes of analysis. By endogeneity I mean choices or conditions determined by the logic of the situation rather than given or assumed from outside. For instance, in neorealism the structure of the international system is taken as given, as exogenous, but neorealists state that their objective is to explain changes (as well as constancy) in system structure. Their stated objective indicates that they believe the system's structure is endogenous, yet they do not lay out a theory of choice that helps explain the relationship between individual actor choices and consequences for system structure. Quite the contrary, they deny the existence of such a link. Yet when individual leaders choose to form alliances, build domestic military capabilities, or conquer others' territory, they make choices that can alter the polarity and balance of power in the international system. They also contemplate choices that can precipitate Page 381 → countermoves by their rivals, leading to still further changes in the system's structure. Thus, their choices are constrained by their expectations about the consequences of their actions, one set of which has to do with the international structure. In this more game-theoretic view, the structure is endogenized rather than taken as given. Selection effects are closely tied to endogeneity and are central to understanding the nature of contingent hypotheses. For example, as mentioned earlier, Smith shows that one nation is more likely to attack another if the would-be attacker believes the target's allies are unreliable.22 If the target's allies are believed to be reliable then the likelihood of attack is diminished, and when attacks take place, allies are more likely to prove unreliable. Thus, tests of the reliability of allies cannot look only at cases following an attack as the attacker is likely to have selected targets based on the reliability of its target's allies. Failure to attend to this selection effect can lead to grossly misleading empirical inferences. Independence between the statement of hypotheses and the evidence for or against the hypotheses is essential if we are to develop a better understanding of how the world of international politics works. There is no inherent advantage to induction or deduction in arriving at “correct” answers to the puzzle of how the world works. There are, perhaps, efficiencies to be gained in beginning deductively and then progressing to induction. Hypotheses derived from first principles provide ideas to test that are not derived from observed patterns in known data. Because the deduced propositions from an axiomatic model follow logically from the assumed initial conditions, the propositions have the advantage of being closely based on causal logic. In this way the researcher gains the efficiency of evaluating the hypotheses by looking at data chosen to approximate the initial conditions of the theory, rather than being chosen based on known values for the dependent variable. If the propositions or hypotheses are borne out, then confidence is built in the theory. Inductive research, in contrast, typically derives hypotheses from the observed patterns in the data. Then, until a second, independent data set is generated, the first

data set has not provided a test of any hypotheses, but rather a description of the hypotheses. Unfortunately, it is the rare social science study that collects a second, independent data set and then tests to see if the precise explanation extracted from the first data set (or case study) holds in the second. If the structure of the hypotheses changes from data set to data set, then the evidence refutes the original hypotheses, new hypotheses describe current data (which may be refuted by the original data set), and yet another independent data set may be required to evaluate the new claims. One great advantage of out-of-sample testing is that it provides a useful way to evaluate whether a hypothesis derived from one body of data is robust in the face of new bodies of data. Out-of-sample testing moves knowledge forward by providing predictive tests. As formal models do not derive their hypotheses from the data, they facilitate prediction because they offer specific, contingent descriptions of how variables relate to each other rather than merely describing some known data. Then predictive tests are as readily conducted as ex post tests. Page 382 → Of course, with genuinely predictive tests there is no danger that the ex ante, stated hypothesis is designed to fit the known outcome because there is not yet a known outcome. Induction does not preclude real prediction as evinced by empirical models of economic performance and presidential election outcomes. But few empirical theories are applied to prediction, whereas formal models are as readily tested on future events as on past events. Elsewhere I have summarized some of the literature concerning formal models and prediction in international relations.23

Summary of Benefits The emphasis on strategic interaction inherent in game theory coupled with its explicit logic, transparency in assumptions, and reasoning and propositions have led to substantial progress in knowledge. Game-theoretic models and empirical tests of them have advanced our understanding of alliance formation, reliability, and termination. They contributed significantly to the debate on the democratic peace, providing theoretical bases on which to account for the known empirical regularities, as well as suggesting novel empirical insights. Rational choice models are at the core of deterrence theory and played a central role in the development of nuclear and conventional strategy throughout the cold war. Recent research on economic sanctions as a tool of diplomacy shows the limitations of sanctioning as a strategy. Much of that research is grounded in game theory.24 A burgeoning literature on the paths to cooperation in international affairs grows directly out of investigations of repeated games.25 Knowledge of the factors leading to conflict initiation, escalation, and termination have been significantly improved by game-theoretic analyses. I could go on but will stop here to turn attention to the limitations of this approach. Clearly, much remains to be done, with the promise of benefits from this approach—an approach still in its infancy and lacking a critical mass of scholars—being greater than the substantial accomplishments to date.

Limitations Since its initial formal statement in the 1940s,26 game theory has made tremendous technical progress. The earliest formulations assumed that players were bound to honor promises they made. Cooperative games of this sort precluded the possibility of investigating such inherently interesting problems as the prisoner's dilemma. After all, a key feature of that game is that players can promise to cooperate and fail to do so because nothing other than their self-interest governs their behavior. By the 1950s, noncooperative game theory emerged as a means to consider situations in which players could renege on promises. The 1950s also saw the emergence of the critical solution concept for noncooperative games, the Nash equilibrium. A Nash equilibrium is a complete plan of action for each player (i.e., a strategy) such that no player believes he can be better off by switching to a different strategy. Page 383 → By the 1960s, Harsanyi showed how to model games in which players were, in colloquial parlance, uncertain of what game they were playing.27 That is, he showed how to examine situations in which, for example, players were uncertain about prior moves in the game (that is, the history of play) or, loosely, about the payoffs their fellow players attach to particular outcomes. With his insights it became possible to address much higher levels of

uncertainty than had previously been possible in game theory. With the introduction of subgame perfection as a refinement of Nash's equilibria, sequential games and their game trees came into their own, largely supplanting “normal form” games. Now analysts could evaluate how the sequence of action influences behavior and, in the process, could eliminate on logical grounds many Nash equilibria. Only a subset of Nash equilibria survive the subgame perfection requirement that players are restricted to making best replies, thereby eliminating incredible threats. Progress continued apace throughout the 1980s and up to the present. With the development of sequential equilibria, Bayesian perfect equilibria, and other refinements, it became possible to deal increasingly effectively with still more complex forms of uncertainty, larger numbers of players, learning, and one-stage, multistage, and infinitely repeated games. By any yardstick, game theory has matured markedly over the past six decades, but much remains to be done. Game theory endures significant limitations, some of which are being overcome as the state of knowledge improves, others of which are inherent to the approach and are unlikely to be removed, though they can be improved upon by borrowing from other perspectives. I divide my discussion along these lines. First I discuss limitations that may be resolved in time. Then I turn to limitations that reflect the complementarity between rational actor models and other perspectives.

Information Game theory models are demanding in terms of information assumptions. I hasten to add, however, that the demands regarding information are not what many critics seem to think they are. For instance, any game must contain some element of common knowledge. Critics often mistakenly believe that this means that every bit of information in a game-theoretic setting must be known by everyone. This is odd since game theory has always dealt with problems of uncertainty in the guise of incomplete and/or imperfect information. Many of the structurally simplest games, such as the prisoner's dilemma or chicken, involve decision making under uncertainty. Huge strides, as already noted, have been made in adding sophistication and rigor to the treatment of uncertainty and of learning in game theory models. However, some element of any game must be common knowledge. That is, each player must know the element, each player must know that everyone else knows it, and each player must know that everyone else knows that each player knows it, and so on. This requirement means that games that involve learning must start with the premise that everyone knows everyone else's prior (initial) beliefs (though not their prior knowledge, which can be Page 384 → private). Players can disagree about prior expectations, thereby holding different beliefs, but games assume that everyone knows what the prior beliefs held by others are, whether they agree with them or not. Several prominent game theorists are working on alternative approaches to remove this constraint, so this is an area where we can expect significant progress in the future.

Multiple Equilibria Many games have multiple Nash equilibria. For games with multiple equilibria (and even many simple games like chicken or battle of the sexes contain multiple equilibria), game theory can predict a probability distribution over actions but cannot say definitively when one action or another will be taken provided each such action is part of an equilibrium strategy. However, there have been and continue to be important steps toward reducing the problem of multiple equilibria. Solution concepts like subgame perfection, universal divinity, Bayesian perfect equilibria, and so forth “refine away” many Nash equilibria, thereby making predictions more precise. Each refinement is itself a Nash equilibrium so that all of the foundational rigor established through the application of Nash equilibria is preserved. Consider the following simple example. Table 1 depicts a normal-form payoff matrix for battle of the sexes. Player A prefers to go to the opera with B (A's utility if both choose opera is 3) over going to the movies with B (in which case A's utility is 1), but A prefers going to the movies with B over going to the opera or the movies alone (in each case A's utility for going alone is 0). B's utility for going to the movies with A is 3 and for going to the movies alone is 0, while B's utility for going to the opera with A is 1 and for going to the opera alone is 0. That is, A and B prefer to coordinate, doing something together rather than doing something alone, but they disagree over what they would most like to do together. This simple game as stated has three equilibria: both go to the opera; both go to the movies; or A goes to

the opera with probability .75 and A goes to the movies with probability .25, while B goes to the opera with probability .25 and to the opera with probability .75. The game matrix in table 1 is consistent with three games written out as trees, also called the extensive form (see figure 1). In the top-left game, player A moves first. Of the three Nash equilibria for the game in table 1, Page 385 → only one survives as a subgame perfect Nash equilibrium. The equilibrium strategies are depicted with dotted lines. B's plan of action is to choose the movies if A chooses the movies and to choose the opera if A chooses the opera. A's plan of action is to choose the opera, so that the outcome is that A and B go to the opera together. The top-right game shows B moving first, again with dotted lines depicting the equilibrium actions. B's threat to go to the movies is not credible in the top-left game, just as A's threat to go to the opera is not credible in the top-right game. The lighter dotted line in the bottom game indicates that when B chooses, B does not know what choice A made. I have not depicted the equilibria for this game as they are the same as in table 1. Yet the top two games each contain only one equilibrium although each, if written in matrix or normal form, would look exactly like table 1. This illustrates several principles. Refinements of the Nash equilibrium improve the precision of game-theoretic predictions by eliminating some strategies from the equilibrium set. Whether these refinements improve predictive accuracy, however, remains unknown. To my knowledge there are no systematic empirical tests of the match between outcomes and predictions under alternative refinements. Page 386 → When games contain mixed strategy equilibria, such as the probabilistic choices of movies and opera just discussed, game theory predicts the probability distribution of actions. In such situations, the prediction can be evaluated only against a large number of cases as the theory makes no prediction about a specific case, but rather a prediction about the distribution of choices across cases. Then a single case can neither be said to support nor to contradict the theory's prediction. Of course, as the previous example also illustrates, not all games contain multiple equilibria. Game theorists have not stopped with refinements of the Nash equilibrium in efforts to refine and improve prediction. Other efforts, such as quantal response equilibria, introduce non-Nash refinements that nevertheless are grounded in a rigorous, axiomatic foundation.28 These refinements are designed partially to address the reality that people make errors in their choices and partially to address estimation problems in models with perfect information. There is considerable additional research on improving the criteria for selecting among equilibria, but this is not the place for an extended discussion of these technical matters. Despite the improvements in equilibrium selection concepts, much remains to be done. Here is an area with clear complementarity between approaches. Constructivism focuses on socially created views of reality. As such, it may develop into a theory of how people choose among alternative strategies. It is a possible way to resolve the issue of common conjectures. Nash assumed that people share common conjectures about how others would play a given game and that this would lead to equilibrium selection. Focal points likewise were seen as ways that people would choose one equilibrium strategy over another. But no game theorist has a theory about the formation of common conjectures or about the formation of preferences. Constructivism and the empirical literature on socialization are two possible sources for identifying behavioral bases for selecting among equilibria and for predicting and explaining the formation of preferences over outcomes. Constructivism lacks a theory of action, like equilibrium theory in games, but game theory lacks a view on how preferences, focal points, and common conjectures are formed or communicated. Together, constructivism, or theories of socialization, and game theory might provide a powerful body of insights with which to resolve equilibrium selection.29 It is a pity, then, that more constructivists have not been attracted to the idea of using their insights to test empirically whether their approach can help discriminate preferences over equilibria. Perhaps this will prove to be a promising and fruitful area of future research.

Skill and Bounded Rationality Another well-known limitation of game theory is that all players are assumed to play with equal skill. That does

not, however, mean that everyone plays the same way. The concept of player types is prominent in game-theoretic investigations of situations with limited information, especially when one does not have a clear idea of what outcome others prefer or by Page 387 → how much. Signaling games, for instance, have become prominent as a means to address the problems that arise regarding player types and the credibility of their actions. Players can learn more about the true preferences and intentions of other players through close observation of their actions. By distinguishing between cheap talk signals and costly signals, game theory has refined the analyst's ability to distinguish conditions under which a bluff is credible from those in which it is not. One prominent approach to addressing the reality that skills are not equally distributed is the vast literature on bounded rationality. Little of the bounded rationality literature confronts problems of strategic interaction, focusing instead on simpler decision-theoretic problems. Most prominent are the efforts, mentioned earlier, by Kahneman and Tversky to develop prospect theory.30 Unfortunately, there is little real-world, nonlaboratory empirical evidence beyond anecdotes to substantiate the claims for prospect theory, and few of its adherents in international relations are carefully attentive to its demanding mathematical structure. Nevertheless, prospect theory holds out the possibility of evolving into an axiomatic foundation for explaining rational, but non-utilitymaximizing, behavior. One recent effort to compare prospect theory with an expected utility model highlights some important, generally overlooked limitations of prospect theory and raises questions about its practicality or power as a predictive tool.31 In this study, a prospect theorist chose the elements of the Northern Ireland Good Friday Agreements as a vehicle for comparing the two approaches. Data were collected on the Good Friday Agreements in May 1999, and explicit predictions were made about the implementation, delay, watering down, or outright breakdown of talks surrounding each element and the whole package. In August 2000, the researchers identified what actually had happened and compared the results to the predictions. Several important lessons were learned from the exercise. First, the expected utility approach made detailed predictions about the actions of each decision maker over time, as well as predicting the consequences of those actions for decisions taken on each element and the whole of the Good Friday Agreements. Prospect theory could only be precise about predicting who among the decision makers was expected to be risk averse and who was expected to be risk acceptant. It allowed less precise predictions about outcomes. Still, explicit conditions were stipulated that would falsify the less precise prospect theory predictions. The expected utility model made eleven predictions, each of which turned out to be correct. The prospect theory predictions were correct between four-elevenths and seven-elevenths of the time, depending on how some are interpreted. The conditions that were stipulated in advance as falsifying the specific prospect theory claims in fact occurred in this case so that the theory was not upheld. Of course this is only one test, so too much should not be made of it. Nevertheless, it is the first systematic, ex ante test of the effectiveness of the prospect theory viewpoint in comparison with a standard, expected utility-grounded game theory model. Theories of bounded rationality address an important limitation of game theory models, but they do not yet address those limitations effectively. In Page 388 → particular, bounded rationality models proliferate the range of actions predicted to be possible. That is, in models of bounded or limited rationality, the sequence in which options present themselves determines actions while no theory of what governs the sequence of choices is provided. A fundamental factor—the order in which options are presented to players—is taken as exogenous rather than as strategically shaped by rivals. Theories of bounded rationality, then, are capable only of weak, vague (sometimes non-falsifiable) predictions rather than the more precise (and falsifiable) predictions possible with game theory. Even when game-theoretic predictions are somewhat vague, as in cases of multiple equilibria, bounded rationality models are vaguer, with even more equilibria.32 Theories of bounded rationality, rather than simplifying or clarifying expectations, lead to a proliferation of equilibria beyond those supported in rational choice models. A promising approach for solving many of the information restrictions of game theory is to blend game theory with evolutionary models. Evolutionary games are beginning to enjoy considerable attention.33 Evolutionary models allow players to have different levels of skill or insight in strategic situations and then select players on the basis of their success. In the short term, many inefficient strategies survive, but this literature generally shows that over the longer run, strategies and players fail to survive on the basis of their inefficiency so that eventually

evolutionary processes seem to favor players who choose Nash equilibrium strategies. This is encouraging both for rational choice theorists and for non-rational choice theorists who utilize the power of evolutionary models. Indeed, I suspect that in the long run, game theory as we now know it will give way to just such dynamic, evolutionary models. For now, however, game theory provides a simpler mathematical structure than do evolutionary models within which to investigate strategic interaction in international affairs. As the tools for dealing with computational complexity improve, we may perhaps look forward to evolutionary models or dynamic games replacing the comparative static analysis prevalent in current game theory.

Conclusions As a tool for studying international affairs game theory is, relatively speaking, in its infancy. Despite the recency of the approach and the paucity of researchers trained in its use, game theory and other rational choice approaches already have had a significant impact on how we think about and test theories of international affairs. Rational actor models yield subtle, contingent propositions, many of which have already been tested and found to be consistent with the historical record. But game theory also poses challenges to the analyst. These challenges concern ways to improve or reduce dependence on demanding conditions. Still, even when the demands of assumptions are stringent, rational actor models offer a tool that facilitates testing the effects of relaxing or substituting alternative assumptions while maintaining analytic rigor and transparency. Other approaches lack this degree of clarity and logical precision. In the end, we are unlikely to uncover Page 389 → whatever laws govern human interaction until we impose rigorous standards on the logical consistency and empirical assessment of our theories. Rational choice models are one important step in that direction.

Notes 1. Daniel Ellsberg, “The Crude Analysis of Strategic Choice,” American Economic Review 51 (1961): 472–89; Martin McGuire, Secrecy and the Arms Race (Cambridge: Harvard University Press, 1965); Anatol Rapoport and Albert M. Chammah, Prisoner's Dilemma: A Study in Conflict and Cooperation (Ann Arbor: University of Michigan Press, 1965); Bruce M. Russett, “The Calculus of Deterrence,” Journal of Conflict Resolution 7 (June 1963): 97–109. 2. David Lalman, Joe Oppenheimer, and Piotr Swistak, “Formal Rational Choice Theory: A Cumulative Science of Politics,” in Political Science: The State of the Discipline, 2d ed., ed. Ada Finifter (Washington, D.C.: American Political Science Association, 1993); Bruce Bueno de Mesquita, Principles of International Politics (Washington, D.C.: Congressional Quarterly Press, 2000); Jacek Kugler and Yi Feng, eds., special issue of International Interactions 23, no. 3 (1997); Michael Brown, Owen Cote Jr., Sean Lynn-Jones, and Steven Miller, eds., Rational Choice and Security Studies (Cambridge: MIT Press, 2000). 3. Steven J. Brams, Superpower Games (New Haven, Conn.: Yale University Press, 1985); Emerson Niou, Peter Ordeshook, and Gregory Rose, The Balance of Power (New York: Cambridge University Press, 1989); Robert Powell, Nuclear Deterrence Theory: The Search for Credibility (New York: Cambridge University Press, 1990); Robert Powell, In the Shadow of Power: States and Strategy in International Politics (Princeton, N.J.: Princeton University Press, 1999); Woosang Kim and James D. Morrow, “When Do Power Shifts Lead to War?” American Journal of Political Science 36 (1992): 896–922; Bruce Bueno de Mesquita and David Lalman, War and Reason (New Haven, Conn.: Yale University Press, 1992); Alastair Smith, “Alliance Formation and War,” International Studies Quarterly 39 (1995): 405–25. 4. James D. Fearon, “Domestic Political Audiences and the Escalation of International Disputes,” American Political Science Review 88 (1994): 577–92; James D. Fearon, “Signaling Foreign Policy Interests: Tying Hands versus Sinking Costs,” Journal of Conflict Resolution 41 (February 1997): 68–90; Bruce Bueno de Mesquita and Randolph Siverson, “War and the Survival of Political Leaders: A Comparative Study of Regime Types and Political Accountability,” American Political Science Review 89 (1995): 841–55; Kenneth A. Schultz, “Domestic Opposition and Signaling in International Crises,” American Political

Science Review 92 (1998): 829–44; Bruce Bueno de Mesquita, James D. Morrow, Randolph Siverson, and Alastair Smith, “An Institutional Explanation of the Democratic Peace,” American Political Science Review 93 (1999): 791–807. 5. Stephen Walt, “Rigor or Rigor Mortis? Rational Choice and Security Studies” International Security 23, no. 4 (1999): 5–48. 6. Henry Kissinger, White House Years (Boston: Little, Brown, 1979); Paul K. Huth, Standing Your Ground: Territorial Disputes and International Conflict (Ann Arbor: University of Michigan Press, 1996). 7. Alexander L. George, David K. Hall, and William E. Simons, The Limits of Coercive Diplomacy (Boston: Little, Brown, 1971); Bueno de Mesquita, Principles of International Politics. 8. A. F. K. Organski, World Politics (New York: Knopf, 1958); A. F. K. Organski and Jacek Kugler, The War Ledger (Chicago: University of Chicago Press, 1980); Jacek Kugler and Douglas Lemke, eds., Parity and War (Ann Arbor: University of Michigan Press, 1996); Ronald L. Tammen, Jacek Kugler, Douglas Lemke, Allan C. Stam III, et al., Power Transitions: Strategies for the 21st Century (New York: Chatham House/Seven Bridges Press, 2000).Page 390 → 9. Kim and Morrow, “When Do Power Shifts Lead to War?” Robert Powell, “Uncertainty, Shifting Power, and Appeasement,” American Political Science Review 90 (1996): 749–64; Bueno de Mesquita and Lalman, War and Reason. 10. Bueno de Mesquita, Principles of International Politics, 413–14. 11. Christopher K. Butler, Superpower Dispute Initiation: Status-Quo Evaluations and Strategic Timing (Ph.D. dissertation, Michigan State University, 2000); Byeonggil Ahn, “When Hawks are Doves and Doves are Hawks: Domestic Conditions and Foreign Conflicts,” Journal of International and Area Studies 6, no. 1 (1998): 21–44; Powell, “Uncertainty, Shifting Power, and Appeasement” Kim and Morrow, “When Do Power Shifts Lead to War?” 12. Fearon, “Domestic Political Audiences and the Escalation of International Disputes” Powell, In the Shadow of Power; James D. Morrow, “Modeling the Forms of Cooperation: Distribution versus Information,” International Organization 48 (1994): 387–423. 13. For a small sample of such investigations, see for instance Christopher Gelpi and Joseph Grieco, “Democracy, Crisis Bargaining, and Audience Costs: Analyzing the Survival of Political Elites” (paper presented at the annual meetings of the American Political Science Association, Boston, 1998); D. Scott Bennett and Allan C. Stam III, “Comparative Theory Testing: Expected Utility versus All Comers” (paper presented at the annual meeting of the International Studies Association, Minneapolis, 1998); D. Scott Bennett and Allan C. Stam III, “A Universal Test of an Expected Utility Theory of War,” International Studies Quarterly 44 (2000): 451–80; Dani Reiter and Allan C. Stam III, “Democracy, War Initiation, and Victory,” American Political Science Review 92 (1996): 377–89; Erik Gartzke, “War Is in the Error Term,” International Organization 53 (1999): 567–88. 14. Reiter and Stam, “Democracy, War Initiation, and Victory.” 15. Gelpi and Grieco, “Democracy, Crisis Bargaining, and Audience Costs.” 16. Smith, “Alliance Formation and War.” 17. Schultz, “Domestic Opposition and Signaling in International Crises” James D. Morrow, “Electoral and Congressional Incentives and Arms Control,” Journal of Conflict Resolution 35 (June 1991): 243–63. 18. Bennett and Stam, “A Universal Test of an Expected Utility Theory of War.” 19. Smith, “Alliance Formation and War.” 20. Lisa L. Martin, Coercive Cooperation: Explaining Multilateral Economic Sanctions (Princeton, N.J.: Princeton University Press, 1992); Alastair Smith, “The Success and Use of Sanctions,” International Interactions 21, no. 3 (1996): 229–45. 21. Jacek Kugler and Frank C. Zagare, “The Long-Term Stability of Deterrence,” International Interactions 15, no. 3 (1990): 255–78; Frank C. Zagare, The Dynamics of Deterrence (Chicago: University of Chicago Press, 1987); Frank C. Zagare and D. Marc Kilgour, “Asymmetric Deterrence,” International Studies Quarterly 37 (1993): 1–27; Samuel Wu, “To Attack or Not to Attack: A Theory and Empirical Assessment of Extended Immediate Deterrence,” Journal of Conflict Resolution 34 (September 1990): 531–52. 22. Smith, “Alliance Formation and War.” 23. Bueno de Mesquita, Principles of International Politics, ch. 17. 24. Martin, Coercive Cooperation; Smith, “The Success and Use of Sanctions.”

25. Rapoport and Chammah, Prisoner's Dilemma; Michael Taylor, Anarchy and Cooperation (New York: Wiley, 1976); Robert Axelrod, The Evolution of Cooperation (New York: Basic Books, 1984). 26. John von Neumann and Oskar Morgenstern, Theory of Games and Economic Behavior (Princeton, N.J.: Princeton University Press, 1944). 27. John C. Harsanyi, “Games with Incomplete Information Played by ‘Bayesian’ Players,” Management Science, ser. a, 14 (1967–68): 159–82, 320–34, 486–502. 28. Richard D. McKelvey and Thomas R. Palfrey, “Quantal Response Equilibria for Extensive Form Games,” Experimental Economics 1, no. 1 (1998): 9–41. 29. I benefited from discussions with James D. Morrow in developing ideas about the linkage between common conjectures and constructivism.Page 391 → 30. See for instance Daniel Kahneman and Amos Tversky, “PT: An Analysis of Decision under Risk,” Econometrica 47 (1979): 263–291; Daniel Kahneman and Amos Tversky, “Choices, Values, and Frames,” American Psychologist 39 (1984): 341–50; Rose McDermott, Risk-Taking in International Politics: Prospect Theory in American Foreign Policy (Ann Arbor: University of Michigan Press, 1998). 31. Bueno de Mesquita, Principles of International Politics; Bruce Bueno de Mesquita, Rose McDermott, and Emily Cope, “The Expected Prospects for Peace in Northern Ireland,” International Interactions 27, no. 2 (2001): 129–67. 32. Kenneth Binmore, Essays on the Foundations of Game Theory (Oxford, England: Blackwell, 1990); Thomas J. Sargent, Bounded Rationality in Macroeconomics (Oxford, England: Clarendon, 1993). 33. Axelrod, The Evolution of Cooperation; David P. Barash, Sociobiology and Behavior, 2d ed. (New York: Elsevier, 1982); Andrew Farkas, State Learning and International Change (Ann Arbor: University of Michigan Press, 1999).

Page 392 →

GAME THEORY IN PRACTICE Problems and Prospects in Applying It to International Relations Steven J. Brams

Introduction In the last twenty or so years, there has been a surge of interest in modeling both national security and international political economy issues using the tools of game theory.1 I will not try to cite this now extensive literature here but instead will (1) outline four major theoretical problems that have bedeviled various attempts at game-theoretic modeling of international relations (IR) and (2) propose an alternative approach, called the “theory of moves,” that is grounded in game theory and that I and others have found attractive in modeling dynamic play. I argue that it captures the thinking of decision makers and realistically delineates their strategic choices in a parsimonious but historically informed way. Thereby I will indicate how some of the failures of game theory over the past half century, especially in modeling the dynamics of strategic choice in IR, can be ameliorated by an alternative, though controversial, theory.2 I illustrate how the theory can be applied to the well-known game of prisoners' dilemma. I then use the theory to analyze Anwar Sadat's peace initiative to Israel in 1977, which startled the world but eventually led to a peace treaty between Egypt and Israel in 1979. Sadat's bold attempt to break the impasse between the two countries and end a thirty-year conflict seems, in retrospect, to be the product of an eminently rational statesman, though it probably cost Sadat his life when he was assassinated in 1981.

Modeling Problems I start by outlining four major pitfalls that plague game-theoretic modeling of IR phenomena and processes. 1. Misspecifying the Rules A game is “the totality of rules that describe it.”3 To be sure, what constitute rules is not a simple matter. Generally speaking, rules both prescribe and Page 393 → proscribe behavior. Any game-theoretic models, therefore, should propose rules of play that reflect how players think and act in the strategic situation being modeled. Based on this reality check, standard game theory often misses the mark. Players do not usually choose strategies simultaneously, or independently of each other, as assumed in the normal or strategic form of a game that can be represented by a payoff matrix; or adhere to a specified sequence of choices, as assumed in the extensive form of a game that can be represented by a game tree. Rather, play usually starts at some initial state, or status quo point. Players look ahead to ascertain whether they can do better by moving or staying in this state. Their choices will vary, depending on the initial state and the rules that indicate what kinds of moves are possible and the means that best satisfy their goals. Generally accepted rationality postulates, applied to the normal or extensive game forms, are frequently violated. For example, sometimes players choose dominated strategies in prisoners' dilemma, or they do not use backward induction, starting from the endpoints of a game tree, when prisoners' dilemma is repeated and the number of rounds is known. These violations, I suggest, are remediable, but they require different rules of play to render them rational and intelligible. 2. Confusing Goals with Rational Choice

Most people have goals, whether implicit or explicit, and they choose the best means to satisfy them, subject to environmental constraints like incomplete information. Emphatically, these constraints do not impair their rationality. Thus, if a person's goals are short term, then he or she will be rational with respect to them; if long term, this person's behavior may be quite different. Some theorists have suggested that short-term goals mirror a thin kind of rationality and long-term goals a thicker or deeper kind. But, in fact, the observable differences in the behavior of the thin and thick types usually arises from their different goals, not their rationality or lack thereof. Put another way, thin types are not being irrational but instead choosing (rationally) not to look deeply or far ahead, presumably because this is too costly. Rationality is a concept appropriately applied to the efficiency or efficacy of the means, or instruments used, to attain desired ends. What are the costs and benefits of different means, and are people making efficacious choices to achieve their ends? Rationality does not concern the ends themselves, which are neither rational nor irrational. To be sure, it is an important developmental question how people come to harbor the goals that they do, but that question is not pertinent to any instrumental notion of rationality that game theory postulates. Page 394 → 3. Arbitrarily Reducing the Multiplicity of Equilibria Many games have multiple Nash equilibria; the 2 × 2 game of chicken, for example, has two equilibria in pure strategies and one in mixed strategies. Worse, in repeated play, almost any outcome of any game can be supported as a Nash equilibrium under certain conditions. With this embarrassment of riches, a game-theoretic model may not be helpful in narrowing down a plethora of stable outcomes to a few precise predictions. Some theorists argue that norms like reciprocity, when they take hold in a population, can induce cooperative outcomes in collective-action situations, thereby helping actors avoid inefficient equilibria. Indeed, there is a good deal of empirical evidence to support not only the reciprocity norm but also the buildup of trust, and the nurturing of reputations of honesty and probity, to stabilize cooperation in games. Unfortunately, these explanations for cooperation are often ad hoc: they do not follow from tenets of rationality in the games being studied. Rather, reciprocity, reputation, and trust are imported as behavioral assumptions, in part because these characteristics of game play are actually observed. Institutions may also play a role in fostering cooperation, but they are not as robust in IR as in domestic politics. An important theoretical direction of game theory has been to endogenize some of its more restrictive assumptions by coarsening information structures, relaxing common knowledge of rationality, offering refinements of Nash equilibria, and the like. In effect, the definition of a game is altered by changing what players know, how they apprehend others' calculations, and the nature of the outcomes they seek. For example, there are now several models that demonstrate how cooperation can emerge from repeated play of prisoners' dilemma when information is incomplete. But repeated play is very much an artifact when modeling player choices over time. Do players really play a single so-called stage game again and again, with no variation in its strategies, outcomes, or payoffs, accumulating payoffs at each stage? In fact, strategies are often in flux, and possible outcomes are constantly changing, undermining the assumption of repeated play of a single stage game. Also problematic in the modeling of IR games, such as protracted wars, is that payoffs may not be realized in stages but only after considerable time has elapsed. While evolutionary game theory and ecological models offer more sophisticated dynamics, they are complex, even if they narrow down the equilibria to relatively few. In addition, few IR processes seem plausibly modeled by a dynamic involving the random matching of players—with possible mutations—as assumed in biologicalevolutionary models. 4. Forsaking Backward Induction

On the one hand, backward induction is “the oldest idea in game theory” that has “maintained its centrality to this day.”4 On the other hand, real Page 395 → players do not painstakingly work backward from the endpoints of a game tree, with many levels and a multitude of branches, to try to determine what they should do at the start of play. Rather, players develop heuristics and use rules of thumb that simplify this process, not because they are short-term thinkers and have abandoned their intelligence but because they have found good approximations to backward induction. But whence these rules, if not from robust and time-tested simplifications that approximate the calculations of backward induction?5 Thus, we should not dismiss backward induction too quickly. We use simple contingent strategies, like tit for tat, because they work well most of the time, even if they are not optimal in every situation. As I shall argue next, players make more exacting calculations in specific games, looking two, three, or four moves ahead, especially when the outcome is important to them.

TOM: An Alternative Approach What I call the theory of moves (TOM) embeds extensive-form games within the normal form, deriving advantages of both forms: the nonmyopic thinking of the extensive form disciplined by the economy of the normal form.6 In my view, the rules of TOM, which allow players to move and countermove within a payoff matrix, capture the changing strategic nature of situations that evolve over time. TOM's ability, as well, to incorporate differences in power—moving, order, or threat—reflects an asymmetry of player capabilities in many IR games. Because prisoners' dilemma is a central focus of IR, let me illustrate the nonmyopic perspective that TOM provides by showing what constitutes rational play in the simple 2 × 2 version of prisoners' dilemma: If the play starts at the noncooperative outcome, players are stuck, no matter how far ahead they look, because as soon as one player departs, the other player, enjoying its best outcome, will not move on. Outcome: the players stay at the noncooperative outcome, which makes it a “nonmyopic equilibrium.” If play starts at the cooperative outcome, then neither player will defect, because if it does, the other player will also defect, and they both will end up worse off. Thinking ahead, therefore, neither player will defect. Outcome: the players stay at the cooperative outcome, which makes it a nonmyopic equilibrium. If play starts at one of the win-lose outcomes (best for one player, worst for the other), the player doing best will know that if it is not magnanimous, and consequently does not move to the cooperative outcome, its opponent will move to the noncooperative outcome, inflicting on the best-off player its next-worst outcome. Therefore, it is in the best-off player's interest, as well as its opponent's, that the former act magnanimously, anticipating that if it does not, the noncooperative outcome (next to worst for both), rather than the cooperative outcome Page 396 → (next to best for both), will be chosen. Outcome: the best-off player will move to the cooperative outcome, where play will stop. These rational moves, grounded in backward induction, are not beyond the pale of most players. They are frequently made by those who look beyond the immediate consequences of their own choices.7 Such farsighted players can escape the dilemma in prisoners' dilemma—as well as poor outcomes in other games—provided play begins at a state other than the noncooperative one in prisoners' dilemma.8 Hence, TOM does not predict unconditional cooperation in this game but, instead, makes it a function of the starting point of play.9 TOM's history-dependent approach has been applied to the fifty-six other 2 × 2 strict ordinal conflict games—in which players can rank outcomes, but there is no outcome that is mutually best—and its predictions empirically tested in several IR studies.10 Not only have its predictions been generally supported, but it has also explained well the dynamics of these conflicts. I turn next to a real-life application of TOM that makes a prediction different from classical game theory. It illustrates how a deliberate move by one player (surpriser) to a Pareto-inferior state (i.e., one worse for both players) may induce the other player (surprisee) to implement an outcome better for both. This case suggests how surprise can engender cooperation rather than conflict in foreign policy, though it may entail risk.

Sadat's Peace Initiative The disengagement of forces after the October 1973 Yom Kippur War, engineered by U.S. secretary of state Henry Kissinger, left Israel in control of the Golan Heights of Syria and the Sinai Peninsula of Egypt, the same territory it had captured in the Six Day War in June 1967. With no settlement in sight four years after the Yom Kippur War, there seemed every prospect of still another Arab-Israeli war when Anwar Sadat, president of Egypt, reported “thinking along the following lines”: Why shouldn't I go to the Israelis directly? Why shouldn't I stand before the Knesset and address the Israelis themselves as well as the whole world, putting forward the Arab cause and stating its dimensions? I conjured up what the reaction might be to such a move, which no one would expect. It would be said that it would be an uncalculated gamble.11 Although Sadat offers, as justification, such philosophical musings as “This is my fate” and “The hour is coming, have no doubt,”12 it seems that there was considerable calculation behind his gamble, despite his impulsive nature.13 For one thing, he wanted “to show the Israelis they were dealing with a new style of Arab leadership,” not “monsters who wanted only to drive Israel into the sea.”14 For another, he was dismayed by the Page 397 → prospect of a reconvening of the Geneva Peace Conference between Israel and its Arab neighbors, which was supported by the superpowers but which Sadat believed was fraught with procedural obstacles and would end in stalemate.15 After his historic visit to Jerusalem on November 19–21, 1977, Sadat could accurately claim that “the Israeli people themselves became a pressure group in favor of peace.”16 As for other Arab leaders, Sadat bitterly remarked that after the Yom Kippur War, “when the entire Arab world made a lot of money out of oil and added to their wealth, Egypt by contrast was drained of its resources”17 indeed, she faced a severe economic crisis in 1976–77.18 Sadat acknowledged that “we knew we could not fight the United States,”19 which would likely side again with Israel if there were another war. More than any other factor, Telhami claims that it was the economic and military power of the United States that moved Egypt from the Soviet toward the American camp and a willingness, under U.S. auspices, to negotiate with Israel.20 Israeli leaders also felt that the United States must be a principal participant in the peace process.21 In addition, Israel's military thinking had changed considerably as a result of the 1973 war. Secure borders were still considered vital, but it was recognized that dynamic and mobile warfare could to a large extent compensate for natural borders.… As a result of less steadfast American backing, Israel's bargaining position also changed … weakened by the Carter administration's more sympathetic policy toward the Arab states and Egypt in particular.22 Thus, when Sadat made his peace initiative to the Egyptian parliament on November 9, 1977, Prime Minister Menachem Begin, though caught unprepared, took only four days to extend an official invitation for Sadat to come to Jerusalem “to conduct talks for a permanent peace between Israel and Egypt.”23 While the “psychological barrier” was broken,24 it was more than psychology that produced what I call the Sadat initiative game (table 1). Egypt's (Sadat's) choices were to initiate (I) or not initiate (Ī) a lasting peace by visiting Jerusalem, and Israel's (Begin's) choices were to cooperate (C) or not cooperate () with this initiative. A brief description of the resulting four possible outcomes and the preferences of the players for each follows (I proceed clockwise from the upper-left outcome in table 1).25 1. ĪC—(4,1). Israel undermines her bargaining position by being cooperative—presumably, by offering concessions—before Egypt takes the first step and commits herself to a permanent solution to their conflict. (This would be especially damaging to the reputation of Begin, who was the newly elected prime minister from the hawkish Likud party.) What is worst (1) for Israel is, in this case, best for Egypt (4), because any resulting

agreement between the two countries would favor Egypt insofar as it requires fewer concessions from her. Page 398 → 2. —(2,3). Without a peace initiative from Egypt or cooperation from Israel, the status quo stays in place. Continuing occupation of the Sinai is next worst (2) for Egypt, whereas it is next best (3) for Israel since her security is ensured, though at a high price. 3. I—(1,2). Israel's frustration of Egypt's peace initiative is Egypt's worst (1) outcome because Sadat's big gamble is to no avail. But this outcome is quite unsatisfactory (2) for Israel as well, because she will lose American support and almost surely have to fight another war with Egypt, which will have little recourse to try to regain its lost territory. 4. IC—(3,4). Israel's favorable response to Egypt's peace initiative not only breaks the psychological barrier but also offers a good prospect of permanent peace. With appropriate security guarantees, this is the best (4) outcome for Israel and the next best (3) for Egypt, which will face the wrath of its Arab neighbors and the loss of monetary support from other Arab countries for negotiating a separate peace. In the 2 × 2 game in table 1,26 Egypt has a dominant strategy of not initiating peace: it is better whatever strategy Israel chooses. Anticipating this choice, Israel's best response is not to cooperate, resulting in payoffs of (2,3) at the status quo of . This is the unique Nash equilibrium, from which neither player would have an incentive to depart unilaterally, making status quo the outcome that classical game theory would predict when both players are assumed to choose their strategies independently. To be sure, Sadat, by taking the initiative and choosing I, changed the 2 × 2 game into a 2 × 4 game—in which he moves first and Begin responds—which makes IC, associated with (3,4), a dominant-strategy Nash equilibrium Page 399 → in the resulting 2 × 4 game. Although Sadat moved first, it still leaves open the question that both Arabs and Israelis have struggled to answer: Why did Sadat take the initiative that he did?27 The answer, I submit, is that it is rational, according to TOM, for Egypt to move to (1,2) and Israel to countermove to (3,4), resulting in a better outcome for both players than they receive at the status quo of (2,3). This is because if Egypt moves from the status quo of to I, changing the state from (2,3) to (1,2), Israel will suffer too unless it responds by moving from I to IC, which gives Israel its best state of (3,4). Notice that the latter state, which is a negotiated settlement, is better for both players than the status quo of (2,3). It is, therefore, in both players' interest that Egypt seize the initiative and move, temporarily, to (1,2), thereby affording Israel the opportunity to countermove to (3,4). On the other hand, if Israel had taken the initiative and moved first to ĪC from , Egypt—going from (2,3) to (4,1)—would have had no incentive subsequently to move on to a negotiated settlement at (3,4). Thus, TOM also explains why it was Egypt, not Israel, that had to seize the initiative that would break the deadlock at (2,3). After the peace accords negotiated at Camp David in September 1978 were translated into a peace treaty that was signed in April 1979, both sides clearly benefited from Sadat's dramatic visit to Jerusalem. In retrospect, there seems to be nothing irrational about Sadat's first move, because it did precisely what Sadat intended it to do in breaking the “psychological barrier.”

Discussion I do not pretend that all surprise can be captured in the switch by a player from a dominant to a dominated strategy, as illustrated in game 27. However, both the Sadat case of diplomatic surprise and the case of military surprise modeled in Brams's “The Rationality of Surprise”—namely, the Japanese attack on Pearl Harbor in December 1941—suggest that such switches may well catch a target completely off guard, generating great shock. It turns out that a move by one player (say, Row) from its dominant-strategy Nash equilibrium, which in turn

induces a countermove by Column to a different state, is rational in exactly six 2 × 2 games. In three of these games, this is to Column's benefit as well as Row's (both Egypt and Israel benefited from Sadat's visit to Jerusalem), but in the other three games it is only to Row's benefit (see “The Rationality of Surprise,” which defines a generic surprise game). In classical game theory, the incentive to move from a unique dominant-strategy Nash equilibrium in a 2 × 2 game also has a rational-choice justification, but it is awkward. It requires comparing the Nash equilibrium in the 2 × 2 game with the Nash equilibrium that Row can induce, if it moves first, in a 2 × 4 game. If the equilibrium in the 2 × 4 game is better for Row, then Row would have an incentive to move first. TOM, by contrast, does not presume different games. Rather, play starts at states, with the order of moves endogenous. Players move and counter-move Page 400 → according to rules of play and rationality rules that enable them to look ahead and anticipate ending up at a nonmyopic equilibrium from each state. Thus, (3,3) is a nonmyopic equilibrium from any state except (2,2) in prisoners' dilemma, whereas (3,4) is a nonmyopic equilibrium from all states in game 27. The surprise of the Sadat peace initiative stems from thinking that the player with the dominant strategy (Egypt) would never have an incentive to depart from its strategy. The reason is that the outcome induced by the other player's best response, which is (3,4) in game 27, has no equilibrium status, not to mention state (1,2) that the players must pass through to get to (3,4) from (2,3). According to TOM, however, IC is the unique nonmyopic equilibrium in game 27, wherever play starts. Hence, the players will be motivated to move there and realize payoffs of (3,4). The migration of Sadat and Begin from the status quo to a negotiated settlement is, therefore, not so surprising. The fact that there was so much surprise leads me to postulate that, even if the game played was one of complete information (i.e., each player knows the other player's strategy choices and preferences for states, as given in table 1), there was incomplete information about either the rules of play or the rationality rules. Indeed, there was probably incomplete information about both player preferences and the rules, but it is the latter kind of incomplete information that has been much less studied, either empirically or theoretically. Because different rules may generate radically different equilibrium results, it is important to try to ascertain those that are most plausible in a situation. The rules of TOM, in my opinion, are highly plausible in many situations. Players often do think more than one step ahead and are able to determine the series of moves and countermoves that they and an opponent are likely to make. Without diminishing the standard explanations for surprise, especially in war, including lack of warning and noise, 28 it seems to me that a neglected alternative explanation is anticipation failure. This occurs when the surprised (myopic) player fails to anticipate the nonmyopic thinking of the other player.29 In the case of Sadat's peace initiative, the initially myopic player (Israel) acted very much according to TOM after it got over its initial shock. Had Sadat's dramatic gesture been anticipated, Begin might have tried to steal some of his thunder. But, as I showed earlier, Israel could not have induced (3,4) in game 27. Moreover, because Israel stood to benefit from the train of events that Sadat's move set off, it would have been foolish of Begin to try to undermine Sadat's efforts.

Conclusions TOM incorporates into the framework of game theory an initial state in a payoff matrix; the moves and countermoves required to reach a nonmyopic equilibrium; and threat, moving, and order power that reflect asymmetries Page 401 → in the capabilities of the players (which I have not discussed here). It also allows for incomplete information, which in some games may lead to misperceptions and even flawed play.30

Because TOM postulates that players rank outcomes but not attach cardinal utilities to them, it seems eminently applicable to strategic situations in which it is hard to estimate players' precise valuations of the outcomes. In addition, it overcomes some problems of classical game theory by providing realistic rules for dynamic play, restricting nonmyopic equilibria to those that can be reached from where play commences, and using backward induction that enables players to make farsighted calculations. What TOM offers that classical game theory does not is a simple and dynamic rationale for the players' making their moves and countermoves, which was observed in the empirical case analyzed herein. In specific games, it provides detailed prescriptions for optimal play, depending on where play starts and the powers of the players, that could aid foreign policy decision makers, especially in crises.31 But TOM is no panacea. It has been mostly developed for two-person games, whereas strategic situations in IR, including alliances and collective-action situations more generally, are often best modeled as n -person games.32 While TOM can be extended to more complex games, more general theoretical structures need to be built, based on plausible rules of play. In my view, these need to take into account (1) players' strategic choices of coalition partners and (2) the courses of actions that flow from these choices.33 Power differences among those coalitions that form, and the history of play that delimits equilibria—by ruling out later choices that are precluded by earlier ones—should provide foundations for a more general dynamic theory.

Notes 1. This paper is adapted from my paper “Game Theory: Pitfalls and Opportunities in Applying It to International Relations” (International Studies Perspectives 1, no. 3 [2000]: 221–32), in which the Iran hostage crisis of 1979–81 is used as the illustrative case study instead of, as here, Sadat's peace initiative in 1977; the latter is drawn from Steven J. Brams, “The Rationality of Surprise: Unstable Nash Equilibria and the Theory of Moves,” in Decisionmaking on War and Peace: The Cognitive-Rational Debate, ed. Nehemia Geva and Alex Mintz (Boulder, Colo.: Lynne Rienner, 1997), 103–29 (updated version, with a response and rebuttal, in Mathematica Japonica, 49, no. 2 [1999]: 293–316). In both this article and “Game Theory,” technical details are omitted that can be found in Steven J. Brams, Theory of Moves (Cambridge: Cambridge University Press, 1994); Brams, “The Rationality of Surprise” and other works on the theory of moves that are cited later. I gratefully acknowledge the support of the C. V. Starr Center for Applied Economics at New York University. 2. See the following recent exchange on the theory: Randall W. Stone, “The Use and Abuse of Game Theory in International Relations: The Theory of Moves,” Journal of Conflict Resolution 45, no. 2 (2001): 216–44; Steven J. Brams, “Response to Randall Stone: Heresy or Scientific Progress?” Journal of Conflict Resolution 45, no. 2 (2001): 245–54. 3. John von Neumann and Oskar Morgenstern, Theory of Games and Economic Behavior, 3d ed. (Princeton, N.J.: Princeton University Press, 1953), 49.Page 402 → 4. Robert J. Aumann, “Backward Induction and Common Knowledge of Rationality,” Games and Economic Behavior 8, no. 1 (1995): 6. 5. Backward induction itself is not always robust; it may be sensitive to the parity of the number of rounds of play and the boundedness of the game (i.e., whether it will end after some finite number of rounds). Steven J. Brams and D. Marc Kilgour, “Backward Induction Is Not Robust: The Parity Problem and the Uncertainty Problem,” Theory and Decision 45, no. 3 (1998): 263–89. 6. Steven J. Brams, “Theory of Moves,” American Scientist 81, no. 6 (1993), 562–70; Brams, Theory of Moves. 7. As a case in point, consider the following statement by Theodore C. Sorensen about the deliberations of the Executive Committee (Excom) during the October 1962 Cuban missile crisis: “We discussed what the Soviet reaction would be to any possible move by the United States, what our reaction with them would have to be to that Soviet reaction, and so on, trying to follow each of those roads to their ultimate

conclusion.” Quoted in Ole R. Holsti, Richard A. Brody, and Robert C. North, “Measuring Affect and Action in International Reaction Models: Empirical Materials from the 1962 Cuban Missile Crisis,” Journal of Peace Research 1, no. 1 (1964): 188. 8. However, Willson's nonmyopic rules, which allow for backtracking, provide an avenue of escape by rendering even a sequence of moves that carries the players from the noncooperative to the cooperative state in prisoners' dilemma rational. Alternatively, I show how credible threats can induce cooperation in prisoners' dilemma and related games. See Stephen J. Willson, “Long-Term Behavior in the Theory of Moves,” Theory and Decision 45, no. 3 (1998): 201–40; Stephen J. Willson, “Axioms for the Outcomes of Negotiation in Matrix Games,” Mathematical Social Sciences 39, no. 3 (2000): 323–48; Brams, Theory of Moves, ch. 5. 9. Under the TOM rules of play that allow players to move and countermove within the payoff matrix, prisoners' dilemma is, of course, no longer the classical version of this game. Because there is almost always some status quo in most IR games, however, I believe it more realistic to start play at states rather than have players commence by making simultaneous or independent strategy choices. Contrast the cooperation that emerges from this framework, based on backward induction, with that that emerges from a framework that postulates ad hoc behavioral assumptions, like reciprocity. 10. Brams's Theory of Moves includes analyses of the Cuban missile crisis in 1962 and the U.S. bombing of Vietnam in 1972. Other applications of TOM to IR can be found in Brams, “The Rationality of Surprise” Steven J. Brams, “To Mobilize or Not to Mobilize: Catch-22s in International Relations,” International Studies Quarterly 43, no. 4 (1999): 621–40; Zeev Maoz and Ben D. Mor, “Enduring Rivalries: The Early Years,” International Political Science Review 17, no. 2 (1996): 141–60; Tansa George Massoud, “Theory of Moves and the Persian Gulf War,” in The Political Economy of War and Peace, ed. Murray Wolfson (Amsterdam: Kluwer Academic, 1998), 247–66; Ben D. Mor, Decision and Interaction in Crisis: A Model of International Crisis Behavior (Westport, Conn.: Praeger, 1993); Ben D. Mor, “The Middle East Peace Process and Regional Security,” Journal of Strategic Studies 20, no. 1 (1997): 172–202; Ben D. Mor and Zeev Maoz, “Learning and the Evolution of Enduring International Rivalries: A Strategic Approach,” Conflict Management and Peace Science 17, no. 1 (1999): 1–48; Marc V. Simon, “When Sanctions Can Work: Economic Sanctions and the Theory of Moves,” International Interactions 21, no. 3 (1996): 203–28; Lester A. Zeager, “Negotiations for Refugee Repatriation or Local Settlement: A Game-Theoretic Analysis, ” International Studies Quarterly 42, no. 2 (1998): 367–84; Lester A. Zeager, “The Role of Threat Power in Refugee Resettlement: The Indochinese Crisis of 1979” (preprint, Department of Economics, East Carolina University); Lester A. Zeager and Jonathan B. Bascom, “Strategic Behavior in Refugee Repatriation: A Game-Theoretic Analysis,” Journal of Conflict Resolution 40, no. 3 (1996): 460–85. 11. Anwar Sadat, Those I Have Known (New York: Continuum, 1984), 104–5.Page 403 → 12. Ibid., 105. 13. Michael I. Handel, The Diplomacy of Surprise: Hitler, Nixon, Sadat (Cambridge: Center for International Affairs, Harvard University, 1981), 324–25; Eric Silver, Begin: The Haunted Prophet (New York: Random House, 1984), 177. 14. Sadat, Those I Have Known, 105. 15. Frank Gervasi, The Life and Times of Menachem Begin: Rebel to Statesman (New York: Putnam, 1979), 48–54; Handel, The Diplomacy of Surprise, 318–20. 16. Sadat, Those I Have Known, 106. 17. Ibid., 106. 18. Handel, The Diplomacy of Surprise, 322–23. 19. Sadat, Those I Have Known, 107. 20. Shibley Telhami, Power and Leadership in International Bargaining: The Path to the Camp David Accords (New York: Columbia University Press, 1990), 13. 21. Janice Gross Stein, “Prenegotiation in the Arab-Israeli Conflict: The Paradoxes of Success and Failure,” in Getting to the Table: The Processes of International Prenegotiation, ed. Janice Gross Stein (Baltimore: Johns Hopkins University Press, 1989), 179. 22. Handel, The Diplomacy of Surprise, 265–66. 23. Moshe Dayan, Breakthrough: A Personal Account of the Egypt-Israel Peace Negotiations (New York: Knopf, 1981), 75.

24. Anwar Sadat, In Search of Identity: An Autobiography (New York: Harper and Row, 1978), 300–302. 25. After developing this game, I reread Maoz and Felsenthal and discovered, to my amazement, that they had used exactly the same game to model Sadat's peace initiative. However, their strategies focus on the concrete terms of the eventual peace treaty: for Egypt, to recognize or not recognize Israel; for Israel, to withdraw or not withdraw from the Sinai. They interpret Sadat's visit to Jerusalem as being a self-binding commitment to peace that was made especially credible by its public nature. TOM shows why Sadat moved first—and why Begin responded as he did—but it says nothing about why Sadat felt it necessary to create such a public spectacle: not being trusted by the Israelis, he wanted to convince them with a well-nigh irrevocable commitment to peace. Zeev Maoz and Dan S. Felsenthal, “Self-Binding Commitments, the Inducement of Trust, Social Choice, and the Theory of International Cooperation,” International Studies Quarterly 31, no. 2 (1987): 177–200. 26. Game 27 in Brams, Theory of Moves. 27. This question is asked in, for example, Ishmail Fahmy, Negotiating for Peace in the Middle East (Baltimore: Johns Hopkins University Press, 1983); Dayan, Breakthrough. 28. See Brams, “The Rationality of Surprise,” for citations. 29. It is worth noting that a game-theoretic model incorporates the choices of both players, whereas the surprise literature distinguishes two schools of thought: the “victim's school,” whose “central goal is to uncover the factors that constrain a victim's attempt to avert surprise” and the “surpriser's school,” which “seeks to explain the means an actor is likely to use to deceive an adversary.” Alex Roberto Hybel, The Logic of Surprise in International Conflict (Lexington, Mass.: Lexington Books, 1986), 3. A game may be seen as a way of combining these different perspectives in a payoff matrix, which links the strategies and preferences of both the victim and the surpriser and is a powerful device that Schelling highlighted a generation ago (Thomas C. Schelling, “What Is Game Theory?” in Contemporary Political Analysis, ed. James C. Charlesworth [New York: Free Press, 1967], 224–32). Even earlier, Schelling suggested a dynamic approach to the study of surprise attack that has much in common with the dynamics incorporated in TOM. See Thomas C. Schelling, The Strategy of Conflict (Cambridge: Harvard University Press, 1960). 30. Examples are given in Brams, Theory of Moves. 31. For its recent application to the Northern Ireland conflict, see Steven J. Brams and Jeffrey M. Togman, “Agreement through Threats: The Northern Ireland Page 404 → Case,” in Being Useful: Policy Relevance and International Relations Theory, ed. Miroslav Nincic and Joseph Lepgold (Ann Arbor: University of Michigan Press, 2000), 325–42. 32. Steven J. Brams and D. Marc Kilgour, “Fallback Bargaining,” Group Decision and Negotiation (forthcoming 2000), and Willson, “Long-Term Behavior in the Theory of Moves,” among others, have offered extensions to n-person games. 33. Steven J. Brams, Michael A. Jones, and D. Marc Kilgour, “Single-Peakedness and Disconnected Coalitions,” Journal of Theoretical Politics (2002).

Page 405 →

REFLECTIONS ON QUANTITATIVE INTERNATIONAL POLITICS Dina A. Zinnes Like other members of the millennial reflections panels, I received a set of questions to guide my comments. To a large extent, however, I found the questions hard to answer within the context of a panel on quantitative methods. I am not, for example, sure that there are “unresolved debates” in regard to quantitative methods; I don't know what “theoretical insights” we have had regarding methodology; nor am I clear as to the meaning of “a fruitful synthesis of approaches” within this context. So while I will reflect on the uses and abuses of methodology in the general field of international studies/ relations, it will be more of a personal statement about where we have been, where we seem to be going, and what we need to consider in the decades ahead. First, where have we been? To answer this question I had a graduate student look through the past decade of the main journals in political science to determine the number of international relations (IR) articles in each journal and the percentage of those that used either statistics or mathematical modeling or both. I was curious as to whether there had been an increase (or decrease) in the use of quantitative methods and particularly whether there were differences in trends between the use of statistics or the use of mathematical modeling. I was also interested in two subquestions: (1) was there a strict dichotomy between the use of statistics and modeling or were there some attempts to use both? and (2) was mathematical modeling primarily game theoretic? The survey covered American Political Science Review (APSR), American Journal of Political Science (AJPS), Journal of Conflict Resolution (JCR), Conflict Management and Peace Science (CMPS), Journal of Peace Research (JPR), and of course International Studies Quarterly (ISQ) for the years from 1990 through 1999. Table 1 provides the results of this investigation, reported in two five-year intervals. With the exception of JPR, especially from 1990 to 1994, we discover that, based on these journals, most of the published work in the field is quantitative—that is, it uses either statistics, mathematical modeling, or both. We also see that quantitative research in Page 406 → Page 407 → IR has not decreased but increased or held steady (ISQ). However, quantitative research in our field is heavily dominated by statistical analyses as seen by comparing the numbers in columns 3 and 4. Indeed, with the exception of APSR, the use of statistical analyses has increased from the 1990–94 period to the 1995–99 period, while the use of mathematical models, with the exception of JPR, has held constant or decreased slightly. Moreover, a comparison of columns 4 and 5 with column 6 quickly demonstrates that the overwhelming form of mathematical modeling is game theory. Finally, column 5 shows that very few quantitative articles contain both mathematical models and statistical analyses. So this is where we have been. Where are we going? If the “trends” over the two subperiods are any indication, it appears that while we are moving in the quantitative direction, we are doing so largely through the use of statistical analyses. Moreover, given that most mathematical models are game theoretic, the questions of international politics subjected to mathematical modeling are seen principally through the eyes of rational actors. Finally, the separation of the two forms of quantitative analysis—very little research contains both mathematical models and statistical analyses—suggests that most mathematical models go untested, and this does not seem to be changing. I find these results disturbing on several counts. The typical statistical study begins with an argument, formulated verbally, from which one or more hypotheses are generated. Data are collected and statistics are used to assess the likelihood that the hypothesis is false, or conversely that the data are adequate to conclude that the hypothesis is supported by the data. But what is the relationship between the verbal argument and the hypothesis? Without the use of a formal language—that is, a mathematical model—it is next to impossible to demonstrate that the hypothesis actually flows from the argument. The statistical analysis of the hypothesis thus becomes, by default, the “theory” that is being evaluated. The argument that supposedly produced the hypothesis is essentially

irrelevant. But the argument that was meant to produce the hypothesis is usually where the interesting ideas are found, the place where the researcher is explaining why certain things happen. We have reduced what may have been a rich interpretation of the international world to an often uninteresting if-then statement. Consequently if IR research is becoming increasingly quantitative through the ever greater use of statistics, we are replacing possible explanation with far less interesting statements about what is associated with what. This is not to decry the use of statistics. Statistics are extremely important in helping us make decisions about the noisy world of data. What is unfortunate then is not the increasing use of statistics but that we have not correspondingly increased our use of formal languages to help us tie the hypotheses being tested with the arguments that supposedly produce them. The disturbing fact is that the use of mathematical models has not increased over the last decade. The increasing use of statistical analyses has fooled us into thinking that we have become more “theoretical,” whereas in actual fact it might reasonably be argued that we have become less theoretical. Page 408 → Thus what is needed in the years to come is a far greater emphasis on the use of mathematical models to carefully represent our arguments, theories, about the international world so that we can demonstrate beyond a reasonable doubt what conclusions do indeed flow from an argument. Only then can we say that the test of a hypothesis gives us confidence in our argument; only then will we become more theoretical as a field. How do we do this? The first step is to move statistics courses out of the limelight of graduate programs. These should not be among the first courses that graduate students meet in their training as they produce a mind-set that makes potential researchers see everything through “if … then …” glasses. The first courses of a graduate program should force the student to think in terms of puzzles, things that don't make sense, don't fit with the way we think about the world, and teach the student how to “tell stories” that fit these anomalies into what we know about the way the world works. This is the beginning of theorizing. The next step is to teach the student how to translate these stories into a mathematical language so that it is possible to empirically evaluate various aspects of our stories. Only after we have developed the theory (= story) and made it precise using a formal language should we turn to statistics to help us determine whether conclusions from our theory are indeed supported by data. Statistics should come toward the end of a graduate program; these courses should be the caboose rather than the engine of a research program. But how do we teach the translation of stories into formal languages? The fact that most research involving mathematical models is game theoretic in nature indicates that we need to broaden our horizons. Game theory is an important and exciting way to translate stories into a formal language. But it is not the only formal language available, and indeed it may not be the best formal language for the story that is being told. Mathematical languages, like natural languages, come in different shapes and sizes—some are better than others for expressing certain ideas. I understand that the language of the Eskimo is rich in its ability to talk about snow but less adequate to express other ideas. Similarly, game theory is an excellent modeling language for capturing conflictual decision-theoretic situations, but not all stories involve rational actors. Our graduate programs must therefore include introductions to more than a single modeling language and provide the student with some feeling for how to make appropriate choices between languages. But having learned that the nature of theorizing lies in telling stories about puzzles and even having learned how to choose among modeling languages and translate a story into an appropriate mathematical structure is not sufficient. We now need to go the last mile. A mathematical model may be beautiful and provide wonderful insights, but it is not a theory about international politics until it has been subjected to an evaluation. Story and data must meet. And this raises some very interesting and curious issues. How does one “test” a game-theoretic model? Does the estimation of the parameters of a dynamic model constitute a test of that model? Because statistics and mathematical modeling have gone their separate ways, questions Page 409 → like these have seldom been addressed. But if we are unable to assess our models, if indeed we do not know how to falsify our theories, then they are nothing more than fun fairy tales.

So in conclusion, we have made considerable progress in making the field of international politics answer to the real world by becoming more aware of the need to empirically test our ideas. But we have a long way to go in making our ideas, our theories, explicit and an even longer way before we understand how to test our models.

Page 410 →

REFLECTIONS ON MILLENNIA, OLD AND NEW The Evolution and Role of Quantitative Approaches to the Study of International Politics James Lee Ray

Introduction As we enter the new millennium, the time seems right for a broad retrospective, introspective, and prospective analytical review of all categories of human endeavor, even including the academic study of international politics. This paper focuses on the evolution of the subfield of international politics over the last forty years, with a particular emphasis on the development of quantitative approaches. It focuses on its shortcomings but also points out what to this author at least seem worthy of being considered some accomplishments. It concludes with a few brief prescriptions for the future. It is presumptuous of me (and, to be fair to myself, of any one person) to take on such a task. I have two excuses. One is that the task was assigned to me. The other is that for good or ill I currently occupy something of a “gatekeeper” role in the field, that is, editor of a journal. That being the case, my views on these matters may be of some interest to practitioners in the field (assuming they are desirous of getting their works published, no matter how idiosyncratic and otherwise lacking in intrinsic merit they may be).

Where Have We Been? The International Studies Association was created more than forty years ago. At that time, the academic field of international politics was, according to a general conception, much wrapped up in the debate fostered mostly by the appearance of Hans Morgenthau's Politics among Nations regarding the utility of “realism” versus “idealism.”1 This preoccupation of the 1950s was about to be replaced, in the 1960s, by the “behavioralist” revolution. These may seem relatively simple, straightforward statements. But one of the points of this essay is that it is quite difficult to characterize, organize, and categorize in any comprehensive way all of the research efforts even in a relatively small field such as the academic study of international politics. If we analyze two editions of International Politics and Foreign Policy, edited Page 411 → by James Rosenau,2 we can find evidence supporting the view of the 1950s as dominated by the debate between realists and idealists, while in the 1960s attention focused instead on the behavioral revolution. For example, the 1961 version of this volume contained not a single selection authored by J. David Singer. In the later 1969 edition, there are five selections authored or coauthored by Singer, including one focusing on “The Behavioral Science Approach to International Relations.”3 But a closer look at the 1961 version of International Politics and Foreign Policy reveals that in fact the behavioral revolution has roots and origins that predate significantly the onset of the 1960s. That volume (based on works originally published in the 1950s, for the most part) has a substantial number of works based on and/or advocating quantitative, statistical, mathematical, and “scientific” approaches to the study of international politics. Richard C. Snyder advocates game theory as a tool of analysis.4 There is an early exemplar from the Project on International Conflict and Integration at Stanford under Robert North, relying on content analysis to provide a reproducible, statistical analysis of the onset of World War I.5 One also finds a contribution by Karl Deutsch advocating quantitative studies of international politics,6 and other articles urging practitioners in the field to investigate the utility of computer simulations7 and experiments,8 as well as mathematics.9 And of course, Rapoport devoted his contribution to a discussion of the work by Lewis Fry Richardson on arms races, and “Richardson published a mathematical treatment of the causes of war … as long ago as 1919.”10 In short, the idea that the behavioral revolution in international politics burst upon the academic scene in the 1960s is rather misleading; that movement can, in fact, be seen as well under way by the arrival of the 1960s.11 A review of academic work in the field in the 1950s and 1960s also provides a basis for what is probably a healthy sense of “the more things change, the more things etc., etc.” “Realists” and “idealists” fought passionate battles in

the 1950s. In the 1990s, Steve Smith observes that “what is termed idealist thinking is still powerful in contemporary theoretical debates, and realism still probably dominates the teaching of the discipline.”12 International Security just last year published a lively exchange regarding the utility of mathematical, deductive “rational choice” versus more intuitive and inductive approaches to the study of international politics.13 Rapoport observes that “the contribution of … [those] inclined toward mathematical theorizing is in showing that it does not matter where the cycle of observation-hypothesis-theory-verification begins. It may be at times useful to begin with theory in abstracto and at other times with the compilation of data. The cycle must be traversed several times anyway before a coherent picture can arise.”14 Though it is a healthy antidote against faddism, perhaps, to see such continuity in debates over the last four decades, such continuity can also lead to skepticism. Perhaps the field is just spinning its wheels, continually engaged in the same old debates over and over again? I believe such skepticism should be resisted and that there are sound bases for doing so. One can, for example, get some sense of progress by Page 412 → reading the editor's introduction to Oliver Benson's contribution to the earlier International Politics and Foreign Policy on the use of computers: “Professor Benson explores the possibility of using high speed computers in research in the international field. Other disciplines have found that these newly developed machines facilitate the handling of problems which once seemed unresearchable. What at one time might have required thousands of man-hours can now be accomplished in but a few computer-minutes. It would seem that students of international politics and foreign policy cannot afford to ignore the potentialities of these aids to research.”15 If one reviews the two Rosenau readers, one can see (with some benefit of hindsight) a striking lack of attention to economic factors or variables, either as explanatory factors or as issues to be addressed in a theoretical fashion. This began to change, perhaps, with neo-Marxist (or neo-Marxist-Leninist) analyses of the Vietnam War.16 Attention to economic factors was reinforced in the 1970s by the onset of détente between the United States and the Soviet Union (and thus a decrease in the extent to which the traditional issues of security and war seemed to deserve priority) and serious international economic problems associated importantly with OPEC's oil price increases in 1973 and 1979. The result was the emergence of what has become a thriving subfield of international political economy,17 as well as increased sensitivity to nonstate actors such as multinational corporations.18 This subfield experienced some de-emphasis perhaps, with the rebirth of the cold war in the 1980s (and the resulting reemphasis on matters of national security and international conflict). However, what might be termed the second end of the cold war at the end of the 1980s has brought about a rekindling of interest in nongovernmental actors19 and a solidification of the place in the study of international politics of international political economy.20 Surely this entry and integration of economic matters into the sub-field of international politics represents real, tangible progress for it over the last forty years. According to Vasquez, “an increasing portion of the field … has until recently moved further and further away from quantitative analysis…. [But] the tide has turned the other way with findings on the democratic peace increasing the interest in scientific research even among senior scholars who had long been hostile to such modes of analysis.”21 Tracing research on the relationship between regime type and international conflict does in fact offer important insights, perhaps, into the evolution of quantitative approaches to the study of international conflict over the last three or four decades. James Rosenau's article on “pre-theories” posited that one of the more important fundamental distinctions between states for attempts to understand foreign policies was type of political system.22 Several empirical analyses from that era23 provided evidence that democratic states are somewhat less conflict prone than autocratic states, but the evidence was not all that striking, and the finding had no apparent or lasting impact on the field. Dean Babst argued on the basis of a reasonably thorough review of data on interstate wars that no two democratic states had ever fought a war against each other.24 Rummel devoted five whole volumes25 to developing a theoretical Page 413 → structure supporting the idea that “the fundamental variable relating to the occurrence of international … war … is political rights, civil liberties, and economic freedom within nations. The less freedom within two nations, the more likely violence is between them; and between democracies, war should not occur at all.”26 These volumes had little apparent or immediate impact. Nor did Rummel's articles following up on this theme27 draw much attention, except from critics who disputed his

claims.28 Not until the 1990s did research on the “democratic peace” draw all the attention that led Vasquez to conclude as the new millennium approaches that it constitutes an important indication as well as a source of inspiration for a “rebirth” of sorts of quantitative approaches to the study of international politics.29 It has been argued by this author and others that this sequence of developments regarding the unfolding of the democratic peace research program can be explained primarily by international political developments in the “real” world.30 According to this argument, early work on peace between democratic states attracted little attention or support because of the cold war atmosphere surrounding it. Perhaps, however, there is a simpler, even more parsimonious explanation (which will, however, surely strike some as stunningly naive). It is possible that the recent fate of the democratic peace research program in terms of the amount of attention and support it has received can be explained primarily by reference to the quality of the evidence, and even the coherence of the theoretical structure on which it has been based. Rummel provided a massive theoretical structure on which to base his observation that no two democratic states have fought wars against each other.31 But the most specific systematic empirical evidence he provided in support of that assertion was basically limited to a review of data on interstate wars generated by the Correlates of War Project.32 An attempt by Small and Singer to analyze the relationship between democracy and peace in a somewhat more thorough fashion provides an intriguing baseline against which to compare more recent research efforts.33 Small and Singer had no annual data on regime type for the time period on which they focused. Therefore, they could not compare rates of warfare for democratic pairs of states with those of pairs of states not jointly democratic. (What they did instead was to compare wars in which democratic states participated with wars in which they did not.) They noted on the basis of the same kind of visual inspection of available data on wars relied on by Rummel (as well as Babst)34 that democratic states had rarely if ever warred against each other. But they concluded that this was because wars are most often fought between neighbors, and “bourgeois democracies do not border upon one another very frequently over much of the period since 1816.”35 That conclusion, too, had to be based on a kind of summary review of information on the geographic location of states vis-à-vis each other, because computerized data on contiguity were not readily available then and were certainly not utilized by Small and Singer. Starting with Maoz and Abdolali, and into the 1990s, such analyses as those by Maoz and Russett, Bremer, and Oneal and Russett have benefited from several types of progress achieved by quantitatively Page 414 → oriented analysts of international politics since the 1970s.36 Annual observations of regime types have become available.37 Computerized data on contiguity have allowed analysts such as Bremer and Maoz and Russett to establish beyond much doubt that the relationship between democracy and peace persists even when contiguity is controlled for.38 These more recent analysts have also been able to control for a rather lengthy list of possibly confounding variables in a way that was simply not available to Small and Singer in the 1970s. They have done so using databases consisting of hundreds of thousands of cases that would have been difficult at best to deal with in the 1970s. And recent research has been based on analytical techniques39 and refinements40 unknown to international relations (IR) specialists in the 1970s. I also believe that the democratic peace research program shows signs of having an impact in the area of theory that may ultimately be at least as important as that created by the empirical observation regarding the absence of wars between democratic states. These theoretical developments focus on substituting for basic realist and neorealist assumptions regarding states seeking power or security the alternative assumption that state leaders place the highest priority on staying in power.41 This basic assumption provides theoretical purchase enabling the integration of the “two-level games”42 faced by all state leaders and a means for integrating domestic and international factors within explanations of foreign policies and international politics in a manner increasingly attractive to an expanding sector of global politics specialists.43 It may not be hopelessly naive, then, to conclude that the increased attention paid to and the impact of research on the relationship between regime type and interstate conflict in the 1990s is not simply a function of changing political tastes brought on by current events. Perhaps it is instead more importantly produced by the improvements in the quality of the evidence, as well as the theoretical bases on which this rapidly growing stream of research relies.

Where Are We (and Where Should We Be) Going?

As insensitive as this essay may appear up to this point, it is in fact not true that I am unaware that “the idealist /realist/behaviouralist progression is in fact a narrow, and particularly political, reading of international theory. Where, for example, is class, or ethnicity, or gender in this self-image? Where are the concerns of developing countries to be found in this canon? It is, in fact, a Western/white/male/conservative view of international theory.”44 So far excluded from discussion in this review have been feminist contributions to IR theory or any mention in general of the “post-positivist” contributions making up the “third wave.”45 In principle, I am certainly favorably disposed to the idea that the increase in diversity among specialists in IR in the last ten or twenty years is a good thing. It would be even better if communication among these diverse specialists were more effective. While cognizant of the possibility that my perceptions on this point are a product of my own shortcomings, what I see occurring in recent years Page 415 → among much of this diverse population of IR specialists are miscommunication, misperceptions, and misunderstandings. As partisan as I may be, however, I would not claim that all of these misperceptions are the fault of specialists making up part of the third wave who fail to comprehend what “scientific-,” quantitative-, or rational choiceoriented specialists are up to. I can see how quantitatively oriented specialists in particular help create misperceptions and misunderstandings about what they aim to accomplish, for example. In fact, I myself am confused about the ideal model or theory or product toward which large-N analysts are striving. The typical multivariate model consists of one relationship of special interest and then several obligatory control variables included in what most often appears to me to be a sort of haphazard, arbitrary fashion. Is the ideal-typical model or theory capable of being summarized in some sort of extra-large multivariate equation? A set of simultaneous equations? A computer simulation? What? (I have my own ideas, but unfortunately space limitations preclude their full discussion here.) And somehow, second-wave specialists have convinced third-wave specialists, it seems to me, that we consider ourselves capable of being entirely objective, above prejudice, that what it means to be scientific, in our view, is to be sufficiently detached to allow the data to speak for themselves. Perhaps some second-wave specialists even believe that. Perhaps in the new millennium specialists making up the second wave can be more clear about the ultimate aims of their research endeavors and dispel this notion that to be scientific means somehow to be “above” emotions, commitment, or ideological convictions. And perhaps third-wave specialists could communicate more clearly, too, in the new millennium. I for one am aware that many third-wave specialists (as well as many second-wave specialists, actually) are suspicious of the tendency for IR theory to focus on interaction among states. It is less clear to me what alternative type of political structure postpositivist or postmodernist (or even feminist) specialists would prefer to see take the place of states as the type of political unit preeminent in the current global political system. UN secretary-general Kofi Annan has observed recently that “during the 1990s there has been a remarkable and little-noticed reduction in global warfare. More old wars have ended than new ones begun.”46 He refers almost entirely to internal, civil, ethnic conflicts. He notes also that between 1992 and 1998, the number of democratically elected governments in the world increased by about a third, and that “the evidence is in line with the well-established, if little publicized finding that democracies have far lower levels of internal violence than non-democracies.… In an era when more than 90% of wars take place within, not between states, the import of this finding for conflict prevention should be obvious.”47 In short, Kofi Annan sees the emergence of stronger and more democratic states as a trend and a solution to the scourge of political violence. Granted, many new states have emerged in recent decades, and many of them are woefully weak and ineffective. Does this indicate that the time has come to move beyond the state? Toward what? Page 416 → Ken Booth is one of those analysts who is obviously quite critical of the current state of “academic international

relations theory” and of its emphasis on interactions among states.48 He is quite clear about what else he is against. “We do,” he declares with admirable candor, “have an enemy, and an enemy of global proportions at that.” He concludes that “the enemy is us. Western consumerist democracy … is the problem.… In a capitalistconstituted world the consumer is sovereign.”49 From this it is clear to me that Booth is opposed to democracy, at least in its current “Western” form, as well as capitalism. But what is he for? What would he prefer to substitute for “Western consumerist democracy”? Eastern aesthetic autocracy? To be fair, I should note that Booth does call for a “community of communities” within a much looser state framework, and for decision-making power to devolve downward to regional and local groups, as well as upward to continental and global functional organizations.50 To me, this image raises more questions than it answers. Would these communities be democratic? Would they be capitalist or market oriented? (Would consumers, in other words, be somehow less “sovereign”?) Why should we believe that they would be less consumerist or better able to deal with pressing problems than national governments? (Have local or state governments in the United States, for example, been consistently more “progressive” or effective or admirable than the national or federal government?) What will prevent the most powerful state governments from dominating “global organizations” in the future the way they do now? What if some less-“enlightened” states refuse to devolve power to local and regional groups, perhaps expressly for the purpose of taking advantage of those states who do? And finally, considering all the problems that seem to result in important part from weak state governments in many parts of the world, is a decline in the influence, strength, and competence of states (and their governments) something to be welcomed or encouraged? One hope I have for the new millennium is that it will be easier to discern what opponents of “statecentric” theories or models are proposing and why they think these proposals should be adopted or advocated by the field of global politics as a whole.

Notes 1. Hans J. Morgenthau, Politics among Nations (New York: Knopf, 1948). As Hoffmann said: “The theory which has occupied the center of the scene in this country during the last ten years is Professor Morgenthau's ‘realist’ theory of power politics.” See Stanley Hoffmann, “International Relations: The Long Road to Theory,” World Politics 11, no. 3 (1959): 346–77, reproduced in International Politics and Foreign Policy, ed. James N. Rosenau (New York: Free Press, 1961), 421–37. Morgenthau became something of a bête noire for advocates of the behavioral revolution in the 1960s, who tended to see him as the leading advocate of traditional, antiquantitative, and antiscientific work in international politics. Hoffmann's point in the article just cited is to advocate historical and philosophical approaches to international politics. One might assume, then, that he would be sympathetic to Morgenthau's realist approach. However, Hoffmann categorizes Morgenthau with advocates of “scientific approaches” Page 417 → such as Morton Kaplan and Karl Deutsch and is energetically critical of all of them. 2. James N. Rosenau, International Politics and Foreign Policy (New York: Free Press, 1961, 1969). 3. J. David Singer, “The Behavioral Science Approach to International Relations: Payoffs and Prospects,” in James N. Rosenau, International Politics and Foreign Policy, 2d ed. (New York: Free Press, 1969), 380–91. It may also be interesting to note in light of the image of Singer that evolved in later years as an enthusiastic advocate of scientific methods, with an emphasis on inductive, statistical (as opposed to mathematical and deductive) approaches, that one of Singer's contributions to the second edition of the Rosenau volume was entitled “Inter-Nation Influence: A Formal Model.” It is, perhaps, equally intriguing, and consonant with the theme here regarding the difficulties confronted by efforts to characterize theories, approaches, and perhaps most of all, theorists, to take note of the conciliatory theme with which Singer concludes his article advocating the “behavioral science approach”: “It is … important that … [scientific research] not proceed at the expense of, and unrelated to, the excellent work that the traditional scholars

have done for decades.… We can be expected to move steadily in the integrative direction, with the ‘traditionalist’ and the ‘behavioralist’ both reaching out in the other's direction, borrowing, adapting, and recombining in such a way as to gain the great benefits of the newer approach while retaining all that is valuable in the older.” Singer, “Inter-Nation Influence,” 69. 4. Richard C. Snyder, “Game Theory and the Analysis of Political Behavior,” in Rosenau, International Politics and Foreign Policy (1961), 381–90. 5. Dina Zinnes, Robert C. North, and Howard E. Koch Jr., “Capability, Threat, and the Outbreak of War,” in Rosenau, International Politics and Foreign Policy (1961), 469–82. 6. Karl W. Deutsch, “Toward an Inventory of Basic Trends and Patterns in Comparative and International Politics,” in Rosenau, International Politics and Foreign Policy (1961), 450–60. 7. Oliver Benson, “A Simple Diplomatic Game,” in Rosenau, International Politics and Foreign Policy (1961), 504–11. 8. Herbert Goldhamer and Hans Speier, “Some Observations on Political Gaming,” in Rosenau, International Politics and Foreign Policy (1961), 498–503. 9. Anatol Rapoport, “The Mathematics of Arms Races,” in Rosenau, International Politics and Foreign Policy (1961), 492–97. 10. Ibid., 492. 11. Nevertheless, the 1960s can accurately be viewed as a decade of great ferment and change in the field. The Rosenau 1969 edition retained only five out of the fifty-five contributions in the 1961 version. 12. Steve Smith, “The Self-Images of a Discipline: A Genealogy of International Relations Theory,” in International Relations Theory Today, ed. Ken Booth and Steve Smith (University Park: Pennsylvania State University Press, 1995), 1–37. 13. Stephen M. Walt, “Rigor or Rigor Mortis? Rational Choice and Security Studies,” International Security 23, no. 4 (1999): 5–48; Bruce Bueno de Mesquita and James D. Morrow, “Sorting through the Wealth of Notions,” International Security 24, no. 2 (1999): 56–73; Lisa L. Martin, “The Contributions of Rational Choice: A Defense of Pluralism,” International Security 24, no. 2 (1999): 74–83; Emerson M. S. Niou and Peter C. Ordeshook, “Return of the Luddites,” International Security 24, no. 2 (1999): 84–96; Robert Powell, “The Modelling Enterprise and Security Studies,” International Security 24, no. 2 (1999): 97–106; Frank Zagare, “All Mortis, No Rigor,” International Security 24, no. 2 (1999): 107–14. See also Jonathan Cohn, “When Did Political Science Forget about Politics? Irrational Exuberance,” New Republic, October 25, 1999, . 14. Rapoport, “The Mathematics of Arms Races,” 497. Rapoport concludes that “one suspects that questions of scientific strategy have no general answers.” Ibid. Compare Rapoport's statement with that of Ordeshook of almost forty years later: “Regardless of the mathematical rigor of models, we need to drop Page 418 → the view of science as an enterprise directed by academics armed with theorems and lemmas or by experimentalists scurrying about in white smocks. Science proceeds less coherently, through induction and deduction.…” Cited by Walt, “Rigor or Rigor Mortis?”48. 15. Benson in Rosenau, “A Simple Diplomatic Game,” 504. 16. Harry Magdoff, The Age of Imperialism (New York: Monthly Review Press, 1969); Gabriel Kolko, The Roots of American Foreign Policy (Boston: Beacon Press, 1969). 17. Robert O. Keohane and Joseph S. Nye Jr.'s Power and Interdependence (Princeton, N.J.: Princeton University Press, 1977) is one important precursor. 18. Raymond Vernon, Sovereignty at Bay (New York: Basic Books, 1971). Charles F. Kindleberger's The World in Depression, 1929–1939 (Berkeley: University of California Press, 1973) also deserves mention in this context. 19. For example, Jessica Mathews, “Power Shift,” Foreign Affairs 76 (January-February 1997): 50–66. 20. As indicated quite neatly, perhaps, by the appearance of the second edition of Keohane and Nye's Power and Interdependence in 1989, as well as the fifth edition of Joan Edelman Spero and Jeffrey A. Hart's The Politics of International Economic Relations (New York: St. Martin's Press, 1997). 21. John A. Vasquez, The Power of Power Politics: From Classical Realism to Neotraditionalism (Cambridge: Cambridge University Press, 1998), 2. 22. James N. Rosenau, “Pre-Theories and Theories of Foreign Policy,” in Approaches to Comparative and International Politics, ed. R. Barry Farrell (Evanston, Ill.: Northwestern University Press 1966), 27–92.

23. Maurice East and Philip M. Gregg, “Factors Influencing Cooperation and Conflict in the International System,” International Studies Quarterly 11, no. 3 (1967): 244–69; Stephen A. Salmore and Charles F. Hermann, “The Effect of Size, Development, and Accountability on Foreign Policy,” Peace Science Society Papers 14 (1969): 16–30; Maurice A. East and Charles F. Hermann, “Do Nation-Types Account for Foreign Policy Behavior?” in Comparing Foreign Policies, ed. James N. Rosenau (New York: Wiley, 1974), 209–69; Dina A. Zinnes and Jonathan Wilkenfeld, “An Analysis of Foreign Conflict Behavior of Nations,” in Comparative Foreign Policy, ed. Wolfram F. Hanrieder (New York: McKay, 1971), 167–213. 24. Dean Babst, “A Force for Peace,” Industrial Research 14 (April 1972): 55–58. An earlier version of this paper was published as Dean Babst, “Elective Governments—A Force for Peace,” Wisconsin Sociologist 3, no. 1 (1964): 9–14. 25. R. J. Rummel, Understanding Conflict and War, vols. 1–5 (New York: Sage, 1975–1981). 26. R. J. Rummel, “Roots of Faith II,” in Journeys through World Politics, ed. Joseph Kruzel and James N. Rosenau (Lexington, Mass.: Lexington, 1989), 323. 27. R. J. Rummel, “Libertarianism and International Violence,” Journal of Conflict Resolution 27, no.1 (1983): 27–71; R. J. Rummel, “A Test of Libertarian Propositions on Violence,” Journal of Conflict Resolution 29, no. 3 (1985): 419–55. 28. For example, Steve Chan, “Mirror, Mirror on the Wall … Are the Freer Countries More Pacific?” Journal of Conflict Resolution 28, no. 4 (1984): 617–48; Erich Weede, “Democracy and War Involvement,” Journal of Conflict Resolution 28, no. 4 (1984): 649–64. 29. Vasquez, The Power of Power Politics. 30. James Lee Ray, “Promise or Peril? Neorealism, Neoliberalism, and the Future of International Politics,” in Controversies in International Relations Theory, ed. Charles W. Kegley Jr. (New York: St. Martin's Press, 1995), 335–55; and, for example, Nils Petter Gleditsch, ed., Focus on Democracy and Peace, special issue of Journal of Peace Research 29, no. 4 (1992): 369–434. 31. Rummel, Understanding Conflict and War. 32. J. David Singer and Melvin Small, The Wages of War, 1816–1965: A Statistical Handbook (New York: Wiley, 1972). 33. Melvin Small and J. David Singer, “The War-Proneness of Democratic Regimes, Page 419 → 1816–1965,” Jerusalem Journal of International Relations 1 (summer 1976): 50–69. 34. Babst, “Elective Governments” Babst, “A Force for Peace.” 35. Small and Singer, “The War-Proneness of Democratic Regimes, 1816–1965,” 67. 36. Zeev Maoz and Nasrin Abdolali, “Regime Types and International Conflict, 1816–1976,” Journal of Conflict Resolution 33, no. 1 (1989): 3–35; Zeev Maoz and Bruce Russett, “Alliance, Contiguity, Wealth, and Political Stability: Is the Lack of Conflict among Democracies a Statistical Artifact?” International Interactions 17, no. 3 (1992): 245–67; Zeev Maoz and Bruce Russett, “Normative and Structural Causes of Democratic Peace, 1946–1986,” American Political Science Review 87, no. 3 (1993): 624–38; Stuart Bremer, “Dangerous Dyads: Conditions Affecting the Likelihood of Interstate War, 1816–1965,” Journal of Conflict Resolution 36, no. 2 (1992): 309–41; John R. Oneal and Bruce Russett, “The Classical Liberals Were Right: Democracy, Interdependence, and Conflict, 1950–1985,” International Studies Quarterly 41, no. 2 (1997): 267–94. 37. Ted Robert Gurr, Polity Data Handbook (Ann Arbor, Mich.: Inter-University Consortium for Political and Social Research, 1978); Ted Robert Gurr, Keith Jaggers, and Will H. Moore, Polity II Handbook (Boulder: Department of Political Science, University of Colorado, 1989); Keith Jaggers and Ted Robert Gurr, “Tracking Democracy's Third Wave with Polity III Data,” Journal of Peace Research 32, no. 2 (1995): 469–82. 38. Bremer, “Dangerous Dyads” Maoz and Russett, “Normative and Structural Causes of Democratic Peace, 1946–1986.” In addition, Gleditsch provides evidence that democratic states have not in fact been separated geographically substantially more than nondemocratic pairs of states. Nils Petter Gleditsch, “Geography, Democracy, and Peace,” International Interactions 20, no. 4 (1995): 297–323. For an early example of an interest in the impact of “distance” on relationships between them, see Quincy Wright, A Study of War (Chicago: University of Chicago Press, 1942, 1965): 1466–72. 39. Gary King, Unifying Political Methodology (New York: Cambridge University Press, 1989). 40. Nathaniel Beck, Jonathan N. Katz, and Richard Tucker, “Taking Time Seriously: Time-Series-Cross-

Section Analysis with a Binary Dependent Variable,” American Journal of Political Science 42, no. 4 (1998): 1260–88. 41. Bruce Bueno de Mesquita, James D. Morrow, Randolph M. Siverson, and Alastair Smith, “An Institutional Explanation of the Democratic Peace,” American Political Science Review 93, no. 4 (1999): 791–807. 42. Robert D. Putnam, “Diplomacy and Domestic Politics: The Logic of Two-Level Games,” International Organization 42, no. 3 (1988): 427–60. 43. James Lee Ray, “A Lakatosian View of the Democratic Peace Research Programme: Does It Falsify Realism (or Neorealism)” (paper delivered at the annual convention of the International Studies Association, Washington, D.C., February 16–20, 1999), and in “Progress in International Relations Theory, ” ed. Colin Elman and Miriam Fendius Elman (manuscript, 1999). 44. Smith, “The Self-Images of a Discipline,” 17. 45. Yosef Lapid, “The Third Debate: On the Prospects of International Theory in the Post-Positivist Era,” International Studies Quarterly 33, no. 3 (1989): 235–54. 46. Kofi Annan, “Preventing Conflict in the Next Century,” in The World in 2000, special issue of the Economist, ed. Dudley Fishburn (London: Economist Newspaper Limited, 1999), 51. 47. Ibid., 51. 48. Ken Booth, “Dare Not to Know: International Relations Theory versus the Future,” in Booth and Smith, International Relations Theory Today, 344. 49. Ibid., 344. 50. Ibid., 344.

Page 420 →

QUANTITATIVE INTERNATIONAL POLITICS AND ITS CRITICS Then and Now Russell J. Leng One way to provide an accounting of the state of quantitative international politics at the turn of the millennium is by evaluating its record against the skepticism of its early critics. Traditional international relations scholars, in fact, took a rather jaundiced view of the scientific study of world politics when the subfield was in its infancy at the beginning of the second half of the twentieth century. One of the most skeptical critiques came from the British classicist Hedley Bull who argued that the scientific approach was “likely to contribute very little to the theory of international relations, and in so far as it is intended to encroach upon and ultimately displace the classical approach, it is positively harmful.”1 Bull based his dire predictions on the presumed intractability of the subject matter: an unmanageable number of variables and the impossibility of controlled experiments or even direct observation of key phenomena. Those adopting a scientific approach would be forced to work at the “marginalia of the subject that are susceptible of measurement and direct observation.” The prospect of achieving any cumulative findings, Bull predicted, was “very bleak indeed.”2 There was more than a hint of defensiveness in Bull's critique. One of his complaints was that scientists do not appreciate the quality of contributions from classical scholars. A fetish with scientific method, Bull feared, would lead to a neglect of philosophy and history in university teaching. We would be producing students and scholars well trained in statistical methods, game theory, or model building but ignorant of diplomatic history, lacking a feel for politics, and with scant understanding of the moral issues with which classical scholars have been wrestling over the centuries. A couple of years passed before J. David Singer, who became the point man for advocacy of the scientific study of world politics, published a response.3 Singer was willing to recognize the past contributions of the classicists, but he asserted that they “have gone almost as far as they can go in adding to social science knowledge in any appreciable way.”4 Singer held out an olive branch, but it was in the form of an offer to accept converts. Page 421 → The relationship between the two approaches resembled that of a new faith confronting an established religion, say Christianity claiming to supersede Judaism, or Islam claiming to supersede Christianity. The new faith is willing to accept some of the revelations of the old, while claiming to move beyond them, but the old religion cannot accept the revelations of the new without being rendered obsolete. The arguments of the participants may be based on theological or epistemological differences, but the passions with which they are expressed are fueled by political stakes. In the first two decades of what Karl Deutsch called the “fourth wave” of world politics research,5 the only cross-fertilization between the two fields was in the form of mudslinging. I return to the issue of the relationship between the classical and scientific traditions in the concluding section. But the first task is to consider the extent to which quantitative international politics research has been able to meet the challenges posed by its foremost early critic.

The Record Bull had a laundry list of complaints about scientific research, but they can be reduced to three general categories: (1) triviality; (2) a failure to achieve cumulative knowledge; and (3) a harmful substitution of methodological training for an understanding of the content of the subject matter and appreciation of knowledge gained through centuries of classical scholarship. After considering each of these challenges, I will return to the issue of the contentiousness between the classicists and scientists. Research Questions

Bull asserted that scientific research was doomed to remain on the margins of the central questions of world politics because he assumed that the research would be limited to phenomena that were easily measured. The big questions would be beyond the bounds of quantitative research. Leaving aside statistical issues related to different levels of measurement, the problem to which Bull refers is that of obtaining valid and reliable descriptions of the phenomena of interest. That problem, however, is not limited to the scientific study of international politics. It has confronted every serious student of human affairs, beginning with the ancients. Consider, for example, the disclaimer offered by Thucydides at the beginning of his history of the Peloponnesian Wars as he described the problems he faced as he attempted to reconstruct speeches for which there was no written record: Some I heard myself, others I got from various quarters; it was in all cases difficult to carry them word for word in one's memory, so my habit has been to make the speakers say what was in my opinion demanded of them by the various occasions, of course, adhering as closely as possible to the general sense of what they really said.6 Page 422 → Thucydides faced a similar problem in his attempt to construct a narrative of events occurring during the war, and his comments illustrate his concern with reliability as well as validity: “I did not even trust my own impressions, but [my account] rests partly on what I saw myself, partly on what others saw for me, the accuracy of the report being always tried by the most severe and detailed tests.”7 The most easily described political phenomena are those that are tangible and relatively static, such as the size of a country's armed forces, the number of diplomatic missions within a state, or its gross national product. The variables most difficult to operationalize are those that are less tangible and dynamic, for example, the rate at which a dispute escalates or the influence strategies employed by the contending states. Reporting is more complete in the modern age, but the validity and reliability problems faced by students of international conflict are essentially the same as those faced by Thucydides. The difference lies in the explicitness of the procedures for generating the descriptions. We have no way of knowing where Thucydides obtained his information, how many sources he consulted, or how much of the content of any of the great battles or speeches he described consists of reconstructions based on his perception of the context in which the event occurred. We have no way of knowing what constituted his “most severe and detailed tests” of the information. Issues of descriptive validity and reliability continue to be a problem in classical historical research. In fact, they provide the fuel for repeated cycles of revisionist history. A major advantage of the scientific approach is its methodological explicitness, its replicability. Thucydides' history has attained its unparalleled greatness because it deals with the heart of interstate politics—great debates, decisions, and actions of states and their leaders. Bull's “triviality” challenge presumed that the scientific approach would never get that close to the stuff of international politics. It must be admitted that the evolution of scientific studies of world politics has been driven to some extent by data availability issues. Too often research agendas have been determined by the availability of data, or the relative ease with which new data could be generated, rather than by the most interesting questions. The problem, however, has been less that of subject matter intractability than a lack of adequate funding, imagination, and energy, as has been demonstrated by the success of those who have been willing to undertake ambitious data generation enterprises. Notable of these efforts is Singer's own Correlates of War Project,8 which was beginning to get off the ground as Bull wrote his critique. Within a decade, Singer was publishing studies dealing with questions that Bull claimed to be beyond the reach of scientific investigations, such as the relationship between the structure of the interstate system and war. Other major projects that have come to fruition more recently are Brecher and Wilkenfeld's work on the attributes of crises, Gurr's on intercommunal conflict, and Bercovitch's on mediation to name a few.9 My own research on interstate crisis behavior has generated a data set designed to obtain the fullest possible descriptions of the behavior of states in evolving crises.10 Page 423 →

There are, to be sure, some areas that remain beyond the bounds of a solely empirical approach. Normative and legal questions cannot be answered completely by scientific investigations. But even when dealing with normative issues, scientific findings can clarify our choices. They can tell us what means are necessary to achieve certain ends, and they can inform us of the consequences of employing those means. My research on learning during enduring rivalries encountered questions, particularly those related to the perceptions of national leaders, that required qualitative judgment.11 But the qualitative judgments were aided by empirical findings. Having an accurate account of what states were doing provided insights into what state decision makers were thinking. Conversely, the accounts of national leaders provided greater insight into empirically generated patterns of state behavior. The most interesting research questions often are those that require the integration of quantitative and qualitative methods. Before leaving this section, something should be said about the recent postmodernist challenge to scientific claims to objectivity. At a fundamental level, postmodernists claim that there is no objective reality; there are only the experiences and perceptions of individuals who relate stories based on those perceptions. How we describe the world and the actions of states, it is argued, is an unavoidable function of our experiences, biases, and perceptions. Without entering into the larger debate over the application of postmodernism to social science research, it would be remiss to fail to acknowledge the important warning imbedded in its sometimes obscure arguments.12 During the beginning phase of our research on the behavior of states in militarized crises, David Singer and I set out to construct an “atheoretical” typology of state actions. After some reflection, we realized that such an enterprise was impossible. We both were too grounded in presuppositions about the working of international politics to create anything that was truly atheoretical, so we settled on an attempt to develop a typology that was sufficiently “multitheoretical” to allow the greatest amount of flexibility as the research progressed.13 Even then we recognized that the typology rested on a realist theoretical foundation that was an inescapable part of our thinking about interstate conflict. That foundation, which postmodernists would call the “subtext” of the constructed typology, is evidence that descriptions of reality can never be entirely free of preconceptions. But this is not an “either-or” issue. If it is true that we ultimately approach a limit, in the mathematical sense, in the attempt to be completely objective, the degree to which we succeed matters. It is the conscious effort to narrow the gap between objectivity and accounts based on preconceptions that distinguishes scientific research from pamphleteering. As Weber argued, the first presupposition of science is that the scientific method itself is valid.14 Cumulative Knowledge The issue of how to obtain cumulative knowledge has been a part of the discourse among scientific researchers ever since James Rosenau brought together a group of “first generation IR scientists” in Ojai, California, in Page 424 → 1973 to discuss the state of the subfield.15 The debate over how to achieve accumulation continues with Starr's arguments in his chapter in this collection. The discussion here focuses on the progress to date. During the Ojai meetings, Harold Guetzkow referred to the subfield as having achieved no more than “islands of theories.”16 Three decades later we have progressed to where there are islands of findings. The findings are not trivial. Consider the paramount problem of world politics, war. We now know a great deal about war per se: when it has occurred; the magnitude, intensity, and severity of interstate and civil wars; war's evolution over time;17 and the relationship between war and changes in the structure of the interstate system.18 We also know that most modern wars have occurred between contiguous states;19 that democracies have rarely, if ever, fought wars with each other;20 that a relatively small number of states account for an inordinately high proportion of militarized disputes and wars;21 and that states with high recidivism rates tend to fight each other.22 We also know that the locus of crises in the past century has shifted from Europe to the Third World, to Africa in particular, as has the locus of most ethnopolitical conflict.23 There is strong evidence that wars are most likely to grow out of disputes over territory;24 that contiguity is a strong predictor of violence in militarized crises;25 that the most effective influence strategy in a militarized crises is a “firm-but-flexible” variant of tit for tat;26 that the lesson most commonly drawn by rival states from their

crisis experiences is to adopt a more coercive influence strategy in their next crisis;27 and that state power transitions—not civilizational clashes—have been the best predictors of ethnopolitical warfare.28 Several things stand out when one considers these findings. First, they are by no means marginal to the central questions of world politics. Second, none of them could have been obtained with confidence without employing scientific methods of inquiry. Third, in their present state they remain islands of findings in the sense that none qualifies as an explanatory theory of state or substate behavior; they exist as unfitted pieces of a much larger, multivariate puzzle. Elsewhere I have compared our findings regarding war to the much more heavily funded research on the relationship between diet and coronary artery disease (CAD).29 There, too, there have been starts and stops; correlational findings that resulted in runs on grocery supplies of items like oat bran have been refuted by later studies. But the frontier of science always looks uncertain and contentious as findings are replicated or refuted, and theories contend for acceptance. Behind the frontier of CAD research there is a base of accepted core knowledge, established theory regarding the relationship between serum cholesterol and the buildup of plaque on artery walls. The trend line of CAD research is wavy, but it is cumulative. Accumulation in CAD research has occurred through the integration of laboratory research with correlational population studies. The integration of the two approaches provides stronger explanatory power than is possible with just one of the tools, and sometimes the results are surprising. For Page 425 → example, a basic preventive strategy for CAD is to lower the blood serum levels of low-density lipoproteins (LDL), the bad cholesterol, while raising those of high-density lipoproteins (HDL), the good cholesterol. Laboratory research indicated that estrogen produces both of the desired effects. Population studies demonstrated that the risk of CAD in women rises after menopause, when women stop producing estrogen. Consequently, hormone replacement therapy, that is, adding estrogen, was presumed to aid in reducing the incidence of CAD in postmenopausal women. In fact, women who took estrogen did live longer and have less heart disease than those who did not take estrogen. It appeared to be a perfectly straightforward, easily explained relationship until a team of scientists conducted a blind study of at-risk postmenopausal women in which estrogen was matched against a placebo.30 Over a threeyear period, the study found no significant difference in the incidence of CAD within the group of women taking estrogen and the group taking a placebo. The most plausible biological explanation for the null finding is not that estrogen has no effect, but that it has several different effects that cancel out each other. Besides the beneficial effect of lowering serum levels of LDL and raising those of HDL cholesterol, estrogen promotes clotting and contributes to the inflammation of arteries, factors that increase the risk of heart disease.31 There also is a plausible societal explanation for the earlier positive correlational findings, which is that women who choose the hormone replacement therapy after menopause are likely to be more educated and more likely to take other steps to reduce the risk of heart disease. There is yet another contextual explanation. The participants in the Herrington study were all women who already had some evidence of CAD. It could be that estrogen does have predominantly beneficial effects, but only for women who have not yet accumulated any significant buildup of plaque in their arteries.32 The estrogen story illustrates the risk associated with relying on correlational evidence from uncontrolled population studies, an inescapable problem in most quantitative international politics research. But it also demonstrates the multiple effects that can be produced by a single variable. To turn to two examples closer to home, consider two impressive correlational findings regarding the war proneness of state dyads: the democratic peace and the positive association between contiguity and war. The democratic peace thesis grows out of the finding that there have been no, or virtually no, wars between two democratic states. There is considerable debate over the significance of the statistical finding, as well as over contending explanations for its occurrence. The inability to conduct a controlled study renders the debate regarding the significance of the statistical evidence endless. Critics point out that democracies historically have been less likely to share borders, more likely to be wealthy, and thereby content with the status quo, and more likely to share memberships in international institutions.

When it comes to explaining why democracies would be less likely to go to war with each other, the two most prominent contending explanations Page 426 → (cultural and structural) covary to such a degree that it is impossible to conduct a comparative test. Thus, researchers are left with attempting to find indirect indicators suggestive of one or another explanation.33 But suppose it were possible to conduct a controlled experiment like the estrogen test described earlier. Is it possible that we would find that the potential effects of democracy on foreign policy are as mixed as those of estrogen on the body? The two explanations mentioned in the previous paragraph are based on beneficial cultural and structural attributes presumed to be present in all democracies. But, like estrogen, democracy is a variable with many attributes, and different attributes can have different effects on relations with other states, and, as some have suggested in the case of estrogen, it also may be that the potency of particular effects are contextually dependent. Mansfield and Snyder argue, for example, that the early phase of a democracy in transition from an authoritarian regime may lead to a nationalistic, contentious foreign policy under popularly elected demagogues.34 It may be that statistically the more dangerous effects of majority rule and mass enthusiasm may be masked by a rapid transition to authoritarianism following the popular election of demagogues. Besides the special case of democratic dyads, the best statistical predictor of whether a pair of states will find themselves at war with each other is whether or not they share a common border. Wars tend to be fought over territory, and contiguity provides more opportunities for territorial disputes. But preliminary findings from a study by Starr using geographic information systems imaging indicates that among contiguous states, disputes are significantly less frequent in those border areas in which there are greater opportunities for cross-border interactions and transactions, that is, higher population density, better transportation facilities, and more economic transactions.35 Contiguity, it appears, can have contradictory effects. Depending on other variables, it can encourage border disputes or it can encourage community-building communication and transactions of the sort once described by Deutsch et al.36 As in the estrogen story, what appears to be a straightforward explanation based on an observed association turns out to be an oversimplification of the predictor variable's effects on the phenomenon of interest. When one considers the number of variables that are likely to exert a significant influence on complex social phenomena such as war, and the mixed effects of any one of the variables identified to date, it appears that we still are scratching at the surface of explanatory theory. We are all working at the frontier of scientific international politics research. The scientific frontier is a contentious place. One writer has described the process of scientific cumulation as “Darwinian evolutionary epistemology.”37 No new finding is accepted without challenge, and only the most robust survive. But the expectation in a mature science is that new findings ultimately will lead to the acceptance of new theories that will expand upon or supplant their predecessors. Those theories that meet with long-term acceptance will lead to the expansion of the cumulative core of knowledge. The question in our own field is whether there is any theory to date that Page 427 → has achieved sufficient consensus to be accepted as core knowledge. My own sense is that there is not. Consequently, we continue to work on the frontier of scientific research without the core knowledge that would provide irrefutable evidence of cumulative science. Policy Relevance Another difference between quantitative international politics research and that on CAD is that there has been nothing like the stampede to act upon findings in peace research that would be analogous to the explosion of sales of putative “heart-friendly” foods like oat bran or olive oil. People are most likely to act on those findings that are consistent with their existing beliefs, that serve their immediate interests, and that present few psychological or political costs and risks. Thus it is not surprising that American politicians have seized upon the democratic peace proposition. Conversely, as the war in Chechnya grinds to a close, it should come as no surprise that Huntington's “civilizational clash” hypothesis is viewed by Russian academicians as the most important new work in world politics, while Gurr's empirically based counterfindings are largely ignored.38 The policy relevance of our work, at least in the minds of political leaders, is dependent on their political agenda.

Methodology versus Content At the beginning of this essay, I noted how the intensity of the early debate between classicists and scientists was fueled by political stakes. Nowhere is that more evident than when curricular issues are considered. After all, which approaches graduate students, and even undergraduates, are taught will determine the future direction of world politics research. There is a limited amount of room in any curriculum, so when something new is added, something old must be removed. Bull's fear was that an emphasis on training in scientific methods would squeeze out more traditional subjects, such as diplomatic history, international law, and normative theory. His fears, insofar as the dominance of scientific approaches is concerned, have not been realized. Political science, and world politics in particular, lag behind their sister social science disciplines in the teaching of scientific approaches at the undergraduate level in all but a few, primarily Midwestern, universities. The pattern is more mixed at the graduate level, where at least some training in scientific method has become an established part of the world politics curriculum. Nevertheless, Bull sounded two notes of caution that are worth considering. The first is that an overemphasis on methodological training could lead to a generation of scientists who know little about the political or historical content of the subject that they are researching. I recently attended a professional meeting where a graduate student from a major university presented an empirical study of the origins of the Franco-Prussian War of 1870. At one point, the author referred to Napoleon III of France as Napoleon Page 428 → Bonaparte's grandson, or, he added, “maybe great grandson.”39 The enormous chronological error demonstrated such a lack of knowledge of the historical period that one wondered about the propositions that propelled the study, the research design, and the interpretation of the findings. There is such a thing as gaining a “feel” for the subject matter to guide one's thinking about it. Without that feel, one is likely to wander down research dead ends, or produce findings that rediscover the obvious. My experience in teaching undergraduates for more than thirty years has convinced me that we are best served by putting the content of world politics at the center of early undergraduate teaching, and supplementing it with training in method, with increased emphasis on theory and method, in upper-level courses. I do not underestimate the intellectual service that we provide to undergraduates by teaching scientific methods, but we should not do so at the expense of the joy of studying politics per se. We need to remember that students do not select courses in world politics because they want to learn more about mathematics. Second, there is the danger of dogmatism regarding what qualifies as acceptable knowledge. A lack of knowledge of the normative and philosophical issues debated by classicists, Bull argued, can lead to an unconscious or uncritical acceptance of a particular philosophical position. One does not have to be a social constructivist to recognize the problem to which Bull referred. A decade after the Bull-Singer debate, Vasquez published a study showing how extant scientific investigations were guided, perhaps unconsciously, by the underlying assumptions of realism.40 The problem is not only that realism may be outdated or incorrect in some contexts. When one works within the realist tradition there is a tendency to focus on conflict per se at the expense of the techniques for reducing, terminating, or avoiding it. It should be noted, however, that the unconscious acceptance of the realist paradigm has grown out of the ubiquity of the realist tradition in classical studies of world politics. If there is a cure for chronic realism it is likely to lie in the acceptance within the scientific community of findings and theories that challenge simplistic realist assumptions. The findings on the democratic peace may be one example. I would offer my own findings on conflict bargaining, escalation, and learning as another.41

Summary The final issue to consider is that of the relationship itself, that is, the current state of the cold war between the classicists and scientists. But before doing so, it may be useful to summarize the preceding points to provide some perspective on the evolution of quantitative world politics research. Certainly, Bull's charge of triviality can be laid to rest. Recent quantitative studies have focused not on the

margins of world politics but directly on the decisions of national leaders and the actions and interactions of states, the same events that represented the core of Thucydides' history. It should be noted, however, that an increasing number of those studies are augmented by qualitative case studies. Page 429 → The issue of attaining a cumulative science of world politics is no longer that of attainability but of how long it will take to move from islands of findings to an established base of core knowledge. I can offer no prediction with regard to how long that will take, nor do I think that there is any magic formula for reaching closure. Agreement will occur only when a combination of theory and findings is sufficiently robust to withstand the ready challenges of all comers on the frontier of science. There is some validity to Bull's third concern that the scientific study of world politics could lead to an emphasis on methods over content, particularly as we become more sophisticated in our methodology and more successful in our research. I am convinced that, as in so many other professional activities, meaningful empirical research requires a feel for the subject matter, which comes only through a deeper familiarity with the substance of politics. I am less worried about Bull's related concern, that a superficial understanding of classical theory could lead to a facile dogmatism. As I noted earlier, the rarified air of theoretical debate has not cured the classicists. The best antidote to theoretical dogmatism would appear to be robust counter-findings from scientific studies.

Moving from Competition to Integration The last issue is that of the contentiousness between the classical and scientific communities that is so evident in the early exchange between Bull and Singer. During much of the 1970s and 1980s the relationship could best be described as peaceful coexistence in the Soviet sense of the term: ideological, political, and economic competition short of violent hostilities. But over the past decade some of the competitiveness has lost its edge, and there is a growing tendency to combine classical and scientific approaches. Those working within the classical tradition are more likely to buttress their arguments with data-based analyses; those adopting a scientific approach are more likely to add qualitative case studies to add depth to their investigations. Despite his earlier criticisms of classical methods, Singer recently coauthored a book-length study that includes a number of qualitative case studies.42 I recently completed a study of learning in enduring rivalries that combines quantitative analyses of state behavior and qualitative descriptions of foreign policy-making in comparative case studies.43 The cold war between the classical and scientific approaches is more likely to end through an unconscious integrative solution that serves the interests of both sides than through the triumph of one research paradigm over the other.

Notes 1. Hedley Bull, “International Theory: A Case for a Classical Approach,” World Politics 18 (1966): 362. Bull's understanding of the classical approach is based on the methodologies of law, philosophy, and history, with particular emphasis on “the exercise of judgment.” Ibid., 361. 2. Ibid., 363.Page 430 → 3. J. David Singer, “The Incomplete Theorist: Insight without Evidence,” in Contending Approaches to International Politics, ed. Klaus Knorr and James N. Rosenau (Princeton, N.J.: Princeton University Press, 1969). 4. Ibid., 82. 5. See Harvey Starr, in this volume. 6. Thucydides, A Comprehensive Guide to the Peloponnesian War, ed. Robert Strassler (New York: Free Press, 1996), 15. 7. Ibid., 15. 8. Melvin Small and J. David Singer, Resort to Arms: International and Civil Wars, 1816–1980 (Beverly

Hills, Calif.: Sage, 1982). 9. Michael Brecher and Jonathan Wilkenfeld, A Study of Crisis (Ann Arbor: University of Michigan Press, 1997); Ted R. Gurr, Minorities at Risk: A Global View of Ethnopolitical Conflict (Washington, D.C.; U.S. Institute of Peace Press, 1993); Jacob Bercovitch and Allison Houston, “The Study of International Mediation: Theoretical Issues and Empirical Evidence,” in Resolving International Conflicts: The Theory and Practice of Mediation, ed. Jacob Bercovitch (Boulder, Colo.: Lynne Rienner, 1996). 10. Russell J. Leng and J. David Singer, “Toward a Multitheoretical Typology of International Behavior” (paper presented at the Events Data Conference, East Lansing, Mich., April 15–16, 1970); Russell J. Leng and J. David Singer, “Militarized Interstate Crises: The BCOW Typology and Its Applications,” International Studies Quarterly 32 (1987): 155–73. 11. Russell J. Leng, Bargaining and Learning in Recurring Crises: The Soviet-American, Egyptian-Israeli, and Indo-Pakistani Rivalries (Ann Arbor: University of Michigan Press, 2000). 12. See Pauline M. Rosenau, Post-Modernism and the Social Sciences: Insights, Inroads, and Intrusions (Princeton, N.J.: Princeton University Press, 1992); Michael J. Shapiro, “Textualizing International Politics, ” in International/Intertextual Relations: Postmodern Reading of World Politics, ed. James Der Derian and Michael J. Shapiro (New York: Lexington Books, 1989). 13. Leng and Singer, “Toward a Multitheoretical Typology of International Behavior.” 14. Max Weber, “Science as a Vocation,” in From Max Weber: Essays in Sociology, ed. H. H. Gerth and C. Wright Mills (1919; reprint, New York: Oxford University Press, 1946), 143. 15. James N. Rosenau, ed., In Search of Global Patterns (New York: Free Press, 1976). 16. Harold Guetzkow, “Sizing Up a Study in Simulated International Processes,” in Rosenau, In Search of Global Patterns, 91. Guetzkow's assessment assumes a rather loose definition of theory. “Islands of hypotheses” would have been more accurate. 17. Small and Singer, Resort to Arms. 18. J. David Singer, Stuart Bremer, and John Stuckey, “Capability Distribution, Uncertainty, and Major Power War, 1820–1965,” in Peace, War, and Numbers, ed. Bruce Russett (Beverly Hills, Calif.: Sage, 1972); Frank Wayman, “Bipolarity and War: The Role of Capability Concentration and Alliance Patterns among Major Powers, 1816–1965,” Journal of Peace Research 21 (1984): 25–42; Daniel S. Geller and J. David Singer, Nations at War: A Scientific Study of International Conflict (Cambridge: Cambridge University Press, 1998), ch. 6. 19. Peter Wallensteen, “Incompatibility, Confrontation, and War: Four Models and Three Historical Systems, 1816–1976,” Journal of Peace Research 18 (1981): 57–90; Stuart Bremer, “Dangerous Dyads: Conditions Affecting the Likelihood of Interstate War, 1816–1965,” Journal of Conflict Resolution 36 (1992): 309–41. 20. Zeev Maoz and Bruce Russett, “Normative and Structural Causes of Democratic Peace, 1946–1986,” American Political Science Review 87 (1993): 624–38; Michael Doyle, “Liberalism and World Politics,” American Political Science Review 80 (1986): 1151–61. 21. Charles S. Gochman and Zeev Maoz, “Militarized Interstate Disputes, 1816–1976: Patterns, Procedures, Insights,” Journal of Conflict Resolution 28 (1984): 585–615.Page 431 → 22. Zeev Maoz, “Pacifism, Dispute Proneness, and Addiction: Patterns and Correlates of National and Dyadic Conflict History, 1816–1992” (paper presented at the annual meeting of the Peace Science Society [International], Yale University, New Haven, Conn., October 27–29, 2000). 23. Brecher and Wilkenfeld, A Study of Crisis; Ted R. Gurr, “Peoples against States: Ethnopolitical Conflict and the Changing World System,” International Studies Quarterly 38 (1994): 347–77. 24. John A. Vasquez, The War Puzzle (Cambridge: Cambridge University Press, 1993). 25. Brecher and Wilkenfeld, A Study of Crisis. 26. Russell J. Leng, Interstate Crisis Behavior, 1816–1980: Reciprocity versus Realpolitik (Cambridge: Cambridge University Press, 1993); P. K. Huth, Extended Deterrence and the Prevention of War (New Haven, Conn.: Yale University Press, 1988). 27. Leng, Interstate Crisis Behavior; Leng, Bargaining and Learning in Recurring Crises. 28. Gurr, “Peoples against States.” For more complete accounts of recent findings related to international conflict and war, see Geller and Singer, Nations at War, and Vasquez, The War Puzzle. 29. Russell J. Leng, “Cumulation in Q.I.P.: 25 Years after Ojai,” Conflict Management and Peace Science

18 (2000): 133–47. 30. David M. Herrington et al., “Effects of Estrogen Replacement on the Progression of Coronary-Artery Atherosclerosis,” New England Journal of Medicine 343 (2000): 522–29. 31. Ibid., 527. 32. See Elizabeth Nabel, “Coronary Heart Disease in Women: An Ounce of Prevention,” New England Journal of Medicine 343 (2000): 572–73. 33. Bruce M. Russett, Grasping the Democratic Peace: Principles for a Post-Cold War World (Princeton, N.J.: Princeton University Press, 1993), 40–42. 34. Edward Mansfield and John Snyder, “Democratization and the Danger of War,” International Security 20 (1995): 5–38. 35. Harvey Starr, “The ‘Nature’ of Contiguous Borders” (paper presented at the annual meeting of the Peace Science Society [International], Yale University, New Haven, Conn., October 27–29, 2000). 36. Karl W. Deutsch et al., Political Community and the North Atlantic Area (Princeton, N.J.: Princeton University Press, 1957). 37. Ernst Mayr, This Is Biology: The Science of the Living World (Cambridge, Mass.: Belknap Press, 1998). 38. Samuel P. Huntington, The Clash of Civilizations: Remaking of World Order (New York: Simon and Schuster, 1996); Gurr, Minorities at Risk; Gurr, “Peoples against States.” 39. Napoleon III was Bonaparte's nephew. 40. John A. Vasquez, “Coloring It Morgenthau: New Evidence for an Old Thesis on Quantitative International Politics,” British Journal of International Studies 5 (1979): 210–28. 41. Leng, Interstate Crisis Behavior; Leng, Bargaining and Learning in Recurring Crises. 42. Geller and Singer, Nations at War. 43. Leng, Bargaining and Learning in Recurring Crises.

Page 432 →

QUALITATIVE METHODS IN INTERNATIONAL RELATIONS Jack S. Levy Three decades ago, Sartori complained about the number of “unconscious thinkers” in the field of comparative politics, and the same could be said of the study of international relations.1 Most qualitative analyses were idiographic rather than nomothetic,2 historically specific rather than theoretically driven, and too little concerned with the logic of inference and questions of generalizability. Scholars gave little attention to the problem of how to control for extraneous variables in situations in which the number of variables typically exceeded the number of cases, or to the question of whether there are alternative methods for validating causal inferences in a single case. Both critics and advocates of case study methods agreed that qualitative research in the 1950s and 1960s precluded the cumulation of knowledge across historical cases.3 This lack of scientific rigor created an image of qualitative analysis as highly subjective, pliable in fitting facts to theoretical arguments, nonreplicable, and essentially nonfalsifiable. Much has changed in the last two or three decades, both in terms of the growing body of literature on the methodology of qualitative analysis and the increasing social science orientation of most qualitative research in international relations. My aim in this essay is to survey the expanding literature on qualitative methods and to highlight some of its major themes. Qualitative analysis includes everything from interpretive ethnographic studies to macrohistories spanning millennia to microanalyses of particular events, and individual cases have been analyzed by quantitative as well as qualitative methods. Thus we should not equate qualitative methods with the case study method. The core of the literature on the methodology of qualitative research in the international relations field focuses on comparative and case study methods from a positivistic perspective, however, and that is the focus of this essay. I give particular attention to the meaning of Page 433 → the case concept, the different kinds and purposes of case studies, the different types of case study designs for testing theories, and the distinctive contributions and limitations of comparative and case study methods. A well-defined body of literature on comparative and case study methods was beginning to emerge just as Sartori lamented unconscious thinking in the field.4 These efforts to develop a methodology of comparative analysis were informed by comparable work in sociology,5 and were stimulated by the success of historical sociology.6 By the early 1990s, the methodology of comparative and case study analysis had a well-defined place in the fields of comparative and international politics.7 The publication of King, Keohane, and Verba's Designing Social Inquiry triggered another surge of interest in the methodology of qualitative research.8 This was in part a response to the “KKV” argument that there is a single logic of scientific inference, that qualitative analysis is not fundamentally different from large-N statistical analysis, and that the former needs to conform more closely to the latter. These arguments led to a major debate in the discipline. Symposia in the American Political Science Review (1995) and in the APSA-CP Newsletter (1998) and subsequent articles and books dealt with the “single logic” question, the advantages and limitations of case study approaches, the different types of case study approaches, and specific problems of case selection and other aspects of research design.9 One indicator of the expanding interest in qualitative methodology is the increasing number of graduate courses offered in this area. Another is the organization of the Inter-University Consortium on Qualitative Research Methods (CQRM), which includes instructional workshops, collection of syllabi, and on-line discussion groups.10 Advances in the methodology of qualitative research have been paralleled by an increasing methodological selfconsciousness and sophistication in applications of qualitative methods to empirical research. The current generation of qualitative international relations researchers are much more nomothetic in their general orientations, both in terms of their interest in using explicitly stated theories to guide historical interpretations and in their greater interest in developing theoretical generalizations valid across time and space. They are more attentive to the question of proper case selection and other aspects of research design, more sensitive to potential

threats to valid inference, and generally more social scientific in orientation.11

Definitions Although some define the case study method as distinct from the comparative method,12 nearly all case studies involve comparisons, whether they be explicit or implicit, across cases or within cases. In addition, the literature has evolved in such a way that the comparative method has come to be associated with the analysis of a small number of cases,13 further eroding any distinction between the case study and comparative methods. Some use the term case study to refer to a single case study and comparative method to refer to comparisons among a small number of cases, but the similar logic Page 434 → underlying both single and multiple case studies leads most scholars to use case study method to refer to both.14 Despite the pervasiveness of case study analysis in international and comparative politics, scholars are far from agreement on how to define either case or case study. Ragin and Becker have edited a book entitled What Is a Case? but provide no clear definition or answer.15 Precisely how one defines “case” depends in part on the purposes of inquiry. Those who are primarily interested in understanding and interpreting a particular historical episode, with little interest in constructing or testing more general theoretical propositions, tend to define case in terms of a set of events bounded in time and space. For such scholars, World War I or the cold war may each be a “case” to be explained. Given the shift toward a more theoretical orientation in the international relations field and an increasing interest in constructing and validating theoretical propositions, the vast majority of scholars conceive of a case as an instance of something else, of a theoretically defined class of events.16 The question is always, What is this a case of? In this sense World War I itself is not a case, but certain theoretically defined aspects of World War I may be cases of some broader phenomenon, such as deterrence, balance of power, power transition, diversionary action, or war termination. This is explicit in George's conception of the method of “structured, focused comparison,” which focuses on a particular analytically defined aspect of a set of events and uses a well-defined set of theoretical questions to structure empirical inquiry. This analytical orientation recognizes that observation is theory laden, and that researchers' theoretical preconceptions will lead them to decide which of the myriad of events constituting a historical episode are to be selected out, studied, and recast in terms of “variables.”17 From this perspective, “cases come wrapped in theories” and do not exist independently of the analytic framework a scholar brings to a particular subject.18 Cases are analytical constructions. They are made, not found; invented, not discovered.19 This does not imply that there are multiple, equally valid answers to our questions about the world, only multiple questions that we might ask. It is important to emphasize that a case generally includes many observations on the same variable. The July 1914 crisis, for example, includes a substantial number of observations of the use of coercive threats, domestic influences, and misperceptions, to name but a few variables. Indeed, one of the main strategies of case study analysis, just like large-N analysis, is to generate as many testable implications of one's hypotheses as possible in a given case.20 This standard view of multiple observations per variable per case is more useful than Eckstein's definition of a case as “a phenomenon for which we report and interpret only a single measure on any pertinent variable,”21 and most methodologists have moved away from Eckstein's definition.

Types of Case Studies There are many types of case studies and numerous ways of classifying them. Most classifications focus on the purpose or function of case studies Page 435 → and build on Lijphart's typology of atheoretical, interpretive, hypothesis-generating, theory-confirming, theory-infirming, and deviant case studies.22 Eckstein suggests a similar typology: configurative-idiographic, disciplined-configurative, heuristic, and crucial case studies based on most likely and least likely designs.23 Eckstein suggests the additional category of plausibility probe. These are ideal types, and in practice many case studies combine several of these aims. The following classification represents a combination of the Lijphart and Eckstein categories.

Atheoretical or configurative-idiographic case studies are traditional single-case analyses often associated with area studies. They are highly descriptive and aim to understand and interpret a single case as an end in itself rather than to develop broader theoretical generalizations. Idiographic case studies are inductive; they involve a minimum of a priori theoretical preconceptions, and the interpretation emerges from the case itself. The analyst generally attempts to create a gestalt or holistic picture of a historical episode, explaining all aspects of the episode and all of their interconnections. This is “total history,” which “cannot decide to leave out any aspect of human history a priori.”24 Interpretive or disciplined-configurative case studies also aim to explain/interpret a single case, but that interpretation is explicitly structured by a theory or well-developed theoretical framework that focuses attention on some theoretically specified aspects of reality and neglects others. This is analytic history rather than total history, but it is still idiographic in that it aims to explain a particular historical episode rather than develop or test theoretical generalizations. Although such “case-explaining” case studies are common in the international relations field, they are less highly valued than work that generates or tests theories,25 and many international relations theorists believe that historically specific research is best left to historians. This is understandable but unfortunate, because the explicit and structured use of theory to explain discrete cases often provides better explanations and understandings of those cases—or at least some aspects of those cases—than do less structured historical analyses. The more case interpretations are guided by theory, the more explicit their underlying analytic assumptions, normative biases, and causal propositions; the fewer their logical contradictions; and the easier they are to validate or invalidate empirically. Theoretically guided, interpretive case studies can significantly enhance our descriptive understanding of the world, and international relations scholars have much to contribute to this task.26 Whereas atheoretical and interpretive case studies are basically idiographic in their explanatory objectives, hypothesis-generating or heuristic case studies are more nomothetic in their aims. They examine a particular case or perhaps several cases for the purpose of developing more general theoretical propositions, which can then be tested through other methods, including large-N methods and perhaps alternative case study methods. Case selection is driven by theoretical considerations, not by the intrinsic interest or historical importance of the case itself, and only particular aspects of the case are investigated. Page 436 → It would be more correct to say that case studies can contribute to the process of theory construction than to theory itself, for the latter—defined as a logically interconnected set of propositions about empirical phenomena—requires a more deductive orientation than case studies provide. Thus Achen and Snidal argue, “the logic of comparative case studies inherently provides too little logical constraint to generate dependable theory,” and they complain that findings of case studies have “too often … been interpreted as bodies of theory.”27 The same can be said of statistical or experimental methods. The application of each of these methods can stimulate the imagination and suggest new hypotheses, but isolated hypotheses fall short of theory, and neither case study methods nor statistical methods can by themselves generate theory. Case studies permit an intensive examination of particular historical sequences, and in doing so they can contribute to the process of theory development by helping to clarify the meaning of key variables and the validity of empirical indicators used to measure them, and by suggesting additional causal mechanisms, causal variables, and interaction effects. They can also help to identify the contextual variables that affect hypothesized causal processes and to identify the scope conditions under which particular theories are valid. These are all important steps in the theory-building process. Thus Achen and Snidal argue that “because they are simultaneously sensitive to data and theory, case studies are more useful for these purposes [of developing analytic theory] than any other methodological tool.”28 The role of case studies in generating hypotheses, or at least in refining and sharpening them, is enhanced by the close interaction of theory and data in case study analysis. The analyst begins with a theory, uses it to interpret a case, and simultaneously utilizes the case to suggest important refinements in the theory, which can then be tested

on other cases or perhaps even on other aspects of the same case.29 George and Smoke's analysis of deterrence in American foreign policy, for example, is organized sequentially in terms of theory specification, application of the theory to historical cases, and reformulation of the theory based on the cases.30 The interplay between theory and evidence is also explicit in “analytic narratives” in which formal rational choice theory guides analytic histories, the anomalies of which are then used to refine the theory.31 Although hypothesis-generating case studies can sometimes contribute to the development of entirely new hypotheses, through unexpected discoveries in the process of investigating other phenomena, case studies are often more useful in helping the researcher to refine existing hypotheses, as in George's method of structured, focused comparison.32 The greater the theoretical structure guiding inquiry, and the better defined the researcher's hypotheses, the more efficient the hypothesis-generating process. But sometimes a full-fledged test of a hypothesis is premature. Plausibility probes, like pilot studies in experimental or survey research, are intermediate steps between hypothesis construction and hypothesis testing. They enable the researcher to refine the hypothesis or theory before engaging in a costly and Page 437 → time-consuming research effort, whether through the massive collection of quantitative data or through extensive fieldwork. Deviant case studies, which focus on empirical anomalies in established theoretical generalizations in order to explain them and refine existing hypotheses, are a particularly powerful strategy for theory refinement, much like examining the residuals in a statistical analysis. Deviant case studies can help to validate the measurement of key variables; identify omitted variables, interaction effects, or alternative causal paths; or identify the scope conditions under which a particular theory is valid. Revised hypotheses can then be tested over a broader set of cases or unexplored aspects of the same case. The analysis of “borderline” cases in the hypothesized absence of war between democracies is a form of deviant case analysis.33 Another way case studies can contribute to theory development is through what Lazarsfeld described as the “analysis of the dependent variable.” 34 The analyst examines cases that on the surface appear to be similar on the dependent variable, in order to identify its different subtypes. This facilitates the development of a more differentiated conception of the dependent variable and a more nuanced set of hypotheses. George and subsequently George and Bennett develop this idea in their discussion of “typological theory.”35 They emphasize that there may be a number of alternative causal paths to a given outcome (equifinality) and attempt to identify these alternative causal paths and, if possible, to specify the conditions under which each is most likely to occur. An example is George and Smoke's exploration of the various ways in which deterrence can fail and their elaboration of the different causal paths associated with each. Similar logic underlies Collier and Levitsky's effort to articulate various subtypes of democracy.36 Many scholars emphasize this role of comparative and case study methods in contributing to the process of theory building, in stimulating the imagination and generating hypotheses that can then be more rigorously tested. Stinchcombe suggests that “lots of facts” can be “good hard stones for honing ideas.”37 Lijphart argues that a comparative perspective (as distinct from the comparative method per se) can be a helpful element in discovery.38 He quotes Stretton, who argues that “the function of comparison is less to stimulate experiment than to stimulate imagination.… Comparison is strongest as a choosing and provoking, not a proving, devise; a system for questioning, not for answering.”39 In addition to playing an essential role in the explanation of individual historical episodes and a contributory role in the generation of hypotheses, case studies can also be used to test hypotheses and theories. This is Lijphart's theory-confirming and theory-infirming roles of case studies, which are best collapsed into a single theory-testing category.40 As Lijphart recognizes, however, the use of case studies for testing hypotheses faces a very serious problem: the existence of many variables in conjunction with a relatively small number of cases. As a consequence, outcomes are over-determined, and it is difficult if not impossible to be certain that changes in Page 438 → the dependent variable are due to changes in the hypothesized independent variables and not to the effects of extraneous variables. Much of the literature on the comparative method and on case study analysis deals with the problem of how to make causal inferences in small-N research, when the number of variables generally exceeds the number of cases.41

Lijphart recognized the many variables/small-N problem and offered a number of possible solutions.42 Three try to directly influence the ratio of cases to variables: increase the number of cases by expanding the domain of the analysis both spatially and temporally; reduce the number of variables, either by collapsing conceptual categories or by data reduction techniques such as factor analysis; or focus on a relatively small number of key variables by constructing more parsimonious theories. Lijphart's fourth solution focused on “comparable cases”—cases similar in terms of the control variables but different in terms of hypothesized explanatory variables that one wants to investigate. In emphasizing an increase in the ratio of cases to variables, Lijphart basically accepted the utility of large-N analysis and conceded that the comparative method was inferior to the experimental or statistical methods for the purposes of causal inference. He concluded that comparative analysis was a “first stage” devoted to the careful formulation of hypotheses, which are then tested in a “second stage” statistical analysis.43 He also acknowledged a certain tension between the goals of increasing the number of cases and focusing on similar cases, because the latter narrows the range of possible cases. In a subsequent article on the comparative method, Lijphart conceded that the maximizing-N strategy—which is based on controlling for extraneous variables through partial correlations—and the comparable-cases strategy—which is based on control through a carefully selected set of matched cases—involved different logics and were more fundamentally opposed than he had initially acknowledged.44 He retracted his earlier argument that the comparative method works best with a larger number of cases, and defined the comparative method as equivalent to the small-N analysis of comparable cases. In this way Lijphart moved closer to Eckstein who stressed the advantages of small-N analysis, including single-case analysis.45 Following Lijphart, scholars now conceive of the comparative method and case study methods as strategies for dealing with a relatively small number of cases. The comparative method is often defined as a strategy for conducting research on naturally occurring phenomena in a way that aims to control for potential confounding variables through careful case selection and matching rather than through experimental manipulation or partial correlations.46

Varieties of Case Selection Strategies The comparable-cases strategy is closely related to John Stuart Mill's method of difference, which focuses on cases that have different values on the dependent variable and similar values on all but one of the independent variables.47 In terms of the logic of inference, this facilitates the identification Page 439 → of causal factors that vary with the dependent variable by eliminating all variables that are constant over the similar cases. Mill's method of agreement focuses on cases that are similar on the dependent variable and different on all but one of the independent variables, in order to eliminate all factors that vary across cases on the independent variable and that therefore cannot account for similar outcomes across cases on the dependent variable. Mill's methods of agreement and difference are comparable to “most-different” and “most-similar” systems designs, respectively.48 A most-different systems design identifies cases that are different on a wide range of explanatory variables but not on the dependent variable, while a most-similar systems design identifies cases that are similar on a wide range of explanatory variables but different on the value of the dependent variable.49 The former eliminates extraneous causal variables that vary across cases, while the latter eliminates causal variables that do not vary across cases. Mill argued that the method of difference was more powerful than the method of agreement in establishing causation. Similarly, Lijphart and Smelser each preferred most-similar systems designs.50 Przeworski and Teune, on the other hand, preferred most-different systems designs, which maximize the number of extraneous variables that can be eliminated because they vary while the dependent variable does not.51 The basic logic of the two designs is the same—to identify patterns of covariation and to eliminate independent variables that do not covary with the dependent variable. It is rare that either strategy alone can fully eliminate extraneous variables, and the best strategy generally involves

the combination of most-similar and most-different systems designs. Mill recognized this and argued for a method of “concomitant variation” involving a combination of the methods of agreement and difference. One problem in the application of Mill's methods and of most-similar and most-different systems designs is the difficulty of identifying cases that are truly comparable—identical or different in all respects but one. In addition, because of the possibility that several different sets of conditions may lead to the same outcome—which Mill identified as the “plurality of causes” and which modern systems theorists refer to as “equifinality”—Mill's methods can lead to spurious inferences if they are used mechanically or not supplemented with the use of withincase methods like process tracing. Mill acknowledged this and concluded that for this reason the application of the methods of agreement and difference to political science was “completely out of the question.”52 Proponents of comparative case methods argue that Mill was far too cautious, and that insisting on precise and absolute comparability imposes “a too exacting scientific standard.”53 Comparative researchers emphasize that experimental and statistical methods themselves are imperfect, and they focus on the question of how best to overcome the acknowledged limitations of the comparative method. They give particular emphasis to strategies of case selection and to process tracing to supplement basic controlled comparisons. I return to process tracing in the next section and focus here on case selection strategies. Page 440 → Within-case comparisons of hypothesized relationships at different points in time within the same case are particularly powerful. Such longitudinal comparisons generally take the form of most-similar systems designs because they are able to hold so many variables constant: political history, culture, institutions, geography, and other variables that change only slowly (if at all) over time. This facilitates the identification of the small number of variables that vary with the dependent variable of interest. George, and George and Bennett, use the label of the “congruence method” for this kind of within-case comparison within the methodology of structured, focused comparison.54 Rosen's analysis of Indian strategic doctrine over time and how it changed as a function of changing strategic culture within a relatively static geopolitical context is a good example of a longitudinal mostsimilar systems design.55 Another useful case study design, which facilitates the control over additional variables, involves a combination of across-case and within-case comparisons. A good example is Snyder's study of imperial overextension, which combines comparisons of the behaviors of different states, different individuals within the same state, and the same individuals within a given state over time.56 While comparative researchers argue about the relative merits of alternative case selection strategies, one thing they agree on is that the strategy of random selection of cases, so useful in large-N statistical analysis, will often generate serious biases in small-N research.57 Scholars generally argue that the analysis of a limited number of cases is better served by a careful selection of nonrandom cases. One of the most serious dangers in the deliberate selection of nonrandom cases involves overrepresenting cases from either end of the distribution of a key variable. This is particularly serious when it involves cases with extreme values on the dependent variable, because it results in a reduction in the slope estimates generated by regression analyses (assuming linear relationships) and thus underestimates the strength of causal effects.58 Selecting cases with very high (or very low) values of the dependent variable are common as well as consequential. Precisely because of their historical significance and high salience, major wars and major revolutions are the kinds of cases that most attract scholarly attention. The selection of cases with extreme values of the independent variable does not have a comparable effect, and this asymmetry is the basis for warnings of the dangers of “selecting on the dependent variable.”59 These problems apply to small-N as well as large-N research.60 Research based on the selection of cases with no variation at all on the dependent variable (“no-variance” designs) are particularly problematic. In the study of the causes of war, for example, if the analyst were to examine only wars and observed a particular factor present in every case, she could

not infer that this factor systematically contributes to the outbreak of war because there might be countless other cases not observed in which the same factor were present but in which war did not occur. Page 441 → Although warnings to avoid selecting on the dependent variable have been useful in reminding students of wars, revolutions, and other phenomena to include cases in which wars or revolutions did not occur—to think about the “dogs that didn't bark”—the mechanical application of this basic rule obscures some important situations in which selection of observations on the dependent variable might be a useful strategy for research. One involves the strategy of studying deviant cases for the purpose of analyzing why they deviate from theoretical predictions. Another is situations in which the hypothesis posits necessary conditions for the occurrence of a particular outcome. For testing hypotheses that posit necessary conditions, the basic logic of inference requires the selection of cases on a particular value of the dependent variable because the only observations that can falsify the hypothesis in question are those in which a particular outcome of the dependent variable occurs in the absence of a condition that is hypothesized to be necessary for that outcome.61 If the hypothesis is deterministic in its assertion of necessary conditions, and if one is confident that no measurement error is present, the observation of a single case in which the posited condition is absent is in principle sufficient to falsify the hypothesis. If we allow for some measurement error, even a small number of anomalous cases in which the hypothesized causal factor is not present can significantly undermine our confidence in a hypothesis based on necessary conditions.62 Similarly, causal propositions positing sufficient conditions for a given outcome can usefully be tested through a case study research design, though here it is essential to select cases on a particular range of values of the independent variable posited to be sufficient for a given outcome.63 If the prediction is strong and if measurement error is negligible, hypotheses positing sufficient conditions can be seriously undermined by identifying a very small number of cases in which the hypothesized sufficient condition is not followed by the predicted outcome. The analysis gets more complicated if there is more than one necessary or sufficient condition, or if there are multiple causal paths that can lead to the same outcome. A particular condition might be necessary for one sequence to operate, and that sequence may be sufficient for a particular outcome to occur, but there may be other sequences that also lead to the same outcome but that do not involve the key condition in the first sequence. The impact of some variables may be contingent on the values of other variables, so that simple additive models will not work, and the analyst must examine the combinations or interaction effects of different sets of factors. Ragin refers to this general problem as “multiple conjunctural causation.”64 He argues that standard statistical methods cannot easily deal with this phenomenon,65 and develops “qualitative comparative analysis” based on Boolean algebra to identify and test combinatorial hypotheses.66 Another strategy for case selection for the purposes of testing theories involves what Eckstein called crucial case studies, which are related to the concepts of most-likely or least-likely case research designs.67 A most-likely Page 442 → case is one that almost certainly must be true if the theory is true, in the sense that all the assumptions of a theory are satisfied and all the conditions hypothesized to contribute to a particular outcome are present, so the theory makes very strong predictions regarding outcomes in that case. If a detailed analysis of a most-likely case demonstrates that the theory's predictions are not satisfied, then our confidence in the theory is seriously undermined. The logic of inference is Bayesian, in the sense that the marginal impact of the data on our confidence in the validity of a hypothesis depends on our a priori judgments of the validity of the hypothesis.68 The greater the a priori likelihood of the hypothesis, the lower the impact of confirmatory data and the greater the impact of disconfirmatory data. Similar logic applies to a least-likely case design, which selects “hard” cases in which the predictions of a theory are quite unlikely to be satisfied because few of its facilitating conditions are satisfied. If those predictions are nevertheless found to be valid, our confidence in the theory is increased, and we have good reasons to believe that the theory will hold in other situations that are even more favorable for the theory. Least-likely case research

designs follow what I call the “Sinatra inference”—if I can make it there I can make it anywhere. Most-likely case designs follow the inverse Sinatra inference—if I cannot make it there I cannot make it anywhere. Most-likely and least-likely case designs are often based on a strategy of selecting cases with extreme values on the independent variables, which should produce extreme outcomes on the dependent variable, at least for hypotheses positing monotonically increasing or decreasing functional relationships. Alternatively, a most-likely case design can involve selecting cases where the scope conditions for a theory are fully satisfied, while a leastlikely case design identifies cases in which the theory's scope conditions are satisfied weakly if at all. A good example of a most-likely case design is Lijphart's study of political cleavages and stability in the Netherlands.69 Pluralist theory, which was widely accepted in the discipline, argued that cleavages that cut across various social groups promoted social peace and political stability, while cleavages that were mutually reinforcing across various social groups contributed to social conflict and political instability.70 Because there were very few cross-cutting cleavages in the Netherlands, pluralist theory predicted high levels of social conflict and low levels of political stability. By demonstrating that the opposite was true, Lijphart's analysis contradicted the unconditional statement of the theory and went a long way toward refuting it. This is a good example of the power of a well-selected individual case study to seriously undercut a widely accepted theory.71 The power of most-likely and least-likely case analysis is further strengthened by defining most likely and least likely not only in terms of the predictions of a particular theory but also in terms of the predictions of leading alternative theories. This builds on the idea that a theory is falsified not by the data alone but by a “three-cornered test” involving the theory, the data, and a rival theory.72 The strongest support for a theory comes when a case is least likely for a particular theory and most likely for the rival Page 443 → theory, and when observations are consistent with the predictions of the theory but not those of its competitor. A good example here is Allison's application of three models of foreign policy decision making to the Cuban missile crisis.73 Allison argued that the missile crisis was a least-likely case for the organizational and bureaucratic models of decision making and a most-likely case for the rational-unitary model. We might expect organizational routines and bureaucratic politics to affect decision making on budgetary issues and on issues of low politics, but not in cases involving the most severe threats to national security, where rational calculations to maximize the national interest should dominate and where politics should stop “at the water's edge.” If Allison could show that bureaucratic and organizational factors had a significant impact on key decisions in the Cuban missile crisis, we would have good reasons to expect that these factors would be important in a wide range of other situations.74

Process Tracing The preceding discussion of case selection strategies suggests that certain kinds of individual case studies can contribute to hypothesis testing as well as to hypothesis construction. Most between-case and within-case comparisons are correlational in nature and examine whether a particular set of conditions is associated with hypothesized outcomes, while holding constant as many other factors as possible. Thus George, and George and Bennett, refer to within-case comparisons of hypothesized relationships at different points in time within the same case as the “congruence method” and include it within the methodology of structured, focused comparison.75 There is another approach to within-case analysis, one that is quite common in the practice of case study research but that is often neglected in attempts to formally describe case study methodology, and that is process tracing.76 Process tracing follows a different logic and tries to uncover the intervening causal mechanisms between conditions and outcomes through an intensive analysis of the evolution of a sequence of events within a case. The logic of inference is much more similar to what philosophers of history call genetic explanation77 than to explanations based on covering laws and deductive nomological logic.78 Process tracing provides several comparative advantages for testing many kinds of intervening causal mechanisms, particularly those involving propositions about what goes on inside the “black box” of decision

making and about the perceptions of actors.79 One of the implications of the democratic peace proposition, for example, is that democracies are perceived differently than autocracies and that these differences have a significant impact on behavior. These perceptions are often better explored through small-N case study methods than through large-N statistical methods. Case study process-tracing methods can also be extremely useful in the empirical analysis of nonlinear propositions involving critical inflection points. The testing of such propositions is extremely sensitive to the accurate identification of these inflection points. In order to avoid circular inferences, Page 444 → this must be done with indicators measured independently of the behavior predicted by the theory. It may be difficult to identify empirical indicators of these inflection points that are valid across a large number of cases for the purposes of a large-N analysis, and case study methods can be used to help identify these tipping points and why they occur.80 Similarly, process tracing can be extremely useful in the exploration of path-dependent macrohistorical processes that are extremely sensitive to patterns of timing and sequence.81 Many proponents of case study analysis argue that process tracing has a comparative advantage over large-N statistical methods in validating intervening causal mechanisms, because statistical methods are limited to establishing correlations while case studies can trace the steps in a causal chain.82 While there may be some truth to this argument—particularly for propositions that involve equifinality, complex and contingent interaction effects, or path dependencies—this argument goes too far and needs to be qualified. There is a tendency among some case study researchers both to underestimate the possible utility of statistical analysis for empirically differentiating among hypothesized causal mechanisms83 and to exaggerate the utility of process tracing for this purpose.84 Those engaged in a close process tracing of a causal chain still face the problem identified by Hume, the impossibility of establishing causality from empirical observation. We cannot know for certain that a particular outcome y is the causal result of a set of factors x rather than another set of factors z that have been omitted from the formal analysis. Our confidence in such an inference is greatest, however, if each link in the causal chain is based on a well-established empirical regularity (probabilistic or otherwise) that has been confirmed by large-N studies or possibly comparative case studies in other comparable empirical domains.85 It is true that we have few strong regularities in international relations, and that this limits our ability to infer causality based on covering laws. But this is a general problem of any approach that attempts to infer causality from a sequence of empirical observations, not just a problem with the covering law model. It is more useful to think of causality as an analytical construct, a component of our theories rather than something that can be inferred directly from empirical observation.86 All theories about the empirical world have testable implications.87 Many of these implications concern the relative frequencies or magnitudes of readily observable events and can be best validated by large-N statistical studies. Other implications deal with hypotheses that posit necessary or sufficient conditions, that fall within the rather opaque black box of decision making, or that are for various reasons difficult to measure with validity and accuracy over large numbers of cases in different historical and cultural contexts. These implications can often be effectively analyzed through process tracing. The greater the empirical validation of the testable implications of a theory, by whatever method, the more confidence we can have in a theory, and hence in the causal mechanisms posited by the theory. Page 445 →

Limitations of Case Study Methods I have argued that case studies are essential for the description, explanation, and understanding of particular historical episodes, and that they can also be useful in the development and refinement of more general theoretical propositions. Case studies can demonstrate that certain constellations of variables generate predictions with nonempty cells, even if they cannot establish the relative frequency with which the predicted event occurs. Case studies can also play a role in theory testing, particularly if the theory is very strong and makes point predictions or posits necessary or sufficient conditions. Few of our theories of international relations satisfy these criteria,

however, and for the purposes of testing most theories case study methods have a number of rather serious limitations.88 One is the large number of variables relative to the small number of cases, which is probably the central issue in the literature on the comparative method and case studies over the past three decades. In attempting to demonstrate that her hypothesized causal variables, and not other variables, explain various outcomes, the case study researcher can achieve some degree of control through careful case selection based on most-similar or mostdifferent systems designs, or preferably a combination of both. The researcher can gain additional leverage over her theory through most-likely and least-likely designs, defined both in terms of the theory and the leading rival theories. If a theory's scope conditions are fully satisfied, if its testable implications are precise, if measurement is valid and accurate, and if cases are carefully selected, the case study researcher can often make a plausible argument that the theory is either supported or disconfirmed for the cases under investigation.89 It is much less likely, however, that she will be able to convincingly demonstrate that her findings are valid for comparable instances of the same phenomenon beyond her immediate study. The case study researcher gains leverage on internal validity, but only at the expense of external validity.90 A related problem is that case study methods cannot easily get at “probabilistic” theories, whether those theories involve probabilistic causal mechanisms or whether the operationalization of more deterministic theories involves substantial measurement error. Both kinds of theories lead to probability distributions of predicted outcomes rather than to point predictions and can be falsified with confidence only with a fairly large number of cases. This is a major strength of statistical analysis and a serious limitation of small-N research.91 Many case study researchers acknowledge the limitations of their method for the analysis of probabilistic relationships. Unlike large-N researchers, case study researchers want to explain all variation and leave to chance no variation or anomalous results, whether because of omitted variables or measurement error. Ragin and Zaret, for example, argue that comparative methods “are logical and not statistical in nature because they are used to identify invariant relationships, not statistical or probabilistic relationships Page 446 → … to identify patterns of constant association, not to explain variation.”92 Ragin argues that “the comparative method does not work with samples or populations but with all relevant instances of the phenomenon of interest.… [Explanations] are not conceived in probabilistic terms because every instance of a phenomenon is examined and accounted for if possible…. The comparative method is relatively insensitive to the relative frequency of different types of cases.”93 Similarly, Becker argues that “narrative analysts … are not happy unless they have a completely deterministic result. Every negative case becomes an opportunity to refine the result, to rework the explanation so that it includes the seemingly anomalous case.”94 This emphasis on refinement and reworking is the key. Rather than stopping with a probabilistic relationship that explains a certain amount of the variation in outcomes, with omitted variables and measurement error captured by an error term, case study researchers continue to probe in an attempt to reduce further both sources of error and explain additional variation. In doing so they can generate more complete explanations, but of a smaller number of cases and with a loss of parsimony and generalizability. It may be possible to explain nearly all variation in a handful of variables of interest in a modest number of cases, and some phenomena we want to explain might involve a relatively small number of cases (hegemonic decline, for example), but most phenomena of interest to international relations theorists occur too frequently to conduct detailed case studies of all of them. We need some means of generalizing beyond our sample of cases. Statistical methods do this through a combination of control through partial correlations and randomization of other extraneous influences, but the latter works only if N is large. Although case study methods based on least-likely or most-likely system designs permit generalization, these generalizations are based more on deductive logic (the Sinatra inferences) than empirical demonstration and must be tested on other cases. Case study researchers face another problem: they have difficulty in assessing the relative causal weights of the various factors influencing a particular outcome, unless those factors are either necessary or sufficient for certain outcomes to result. Case study methods can be useful in determining the presence or the absence of a particular

variable and its impact on the presence or absence of outcomes. They might also be able to establish empirically the direction of a variable's impact, and perhaps provide a very rough approximation of its impact in terms of categories of high and low. Case studies are much more limited in their ability to determine empirically the relative magnitude of various causal influences when those factors are neither necessary nor sufficient for a given outcome.95 Because necessary or sufficient conditions are rare, this is another important limitation of case study methods. The ability to estimate different causal effects empirically is a major strength of statistical analyses of larger numbers of cases (assuming the functional relationship is correctly specified and the key variables are measurable across cases, which can be quite problematic). Regression analyses can estimate the amount of variance explained by each variable, the additional Page 447 → amount of variance explained when another variable is added to the model, and the proportional effects on the dependent variable of comparable changes in each of the independent variables.

Combining Case Study and Statistical Methods Statistical methods have their own limitations, of course, particularly concerning the validity of concepts and the operational indicators used to measure them across a large number of cases (the “unit homogeneity” assumption). Proponents of comparative case studies also worry about this problem. They talk at length about balancing the need for “conceptual traveling” (using concepts that are valid across time and space, which facilitates generalization) with the dangers of “conceptual stretching” (applying concepts in historical and cultural contexts in which they have a different meaning or are otherwise not appropriate).96 This is not the place for a detailed comparative evaluation of the advantages and disadvantages of statistical and comparative methods, but many analysts from each methodological perspective have increasingly come to the conclusion that by combining both statistical and case study methods, researchers can use the advantages of each to partially offset the limitations of the other. These methods can be combined in a single study or sequentially as part of a “research cycle” in a larger research program.97 Work that integrates statistical and case study methods in a single study includes Huth's analysis of the success and failure of extended deterrence, Ray's study of the democratic peace, Martin's analysis of the role of international institutions in multilateral economic sanctions, and Simmons's study of the politics of adjustment to international economic pressures of the 1920s and 1930s.98 Examples of the sequential integration of statistical and case study methods in a larger research program include the International Crisis Behavior Project, 99 Doyle's research program on the democratic peace,100 and the Mans- field and Snyder project on democratization and war.101 Although scholars agree on the utility of mixed-method approaches, and although we have begun to see more efforts of this kind, scholars have made few efforts to elaborate on exactly how different methods can be combined or the proper sequence for combining them. While Russett, Lijphart, and others suggest a sequence involving comparative methods for refining hypotheses followed by statistical methods to test them, reversing this sequence can also be useful.102 In the democratic peace research program, for example, statistical methods were first used to establish the extraordinarily strong empirical relationship between democratic dyads and the absence of war. Case study methods were then used to validate whether states were properly classified as democracies or nondemocracies; to explore the intervening causal mechanisms linking peaceful outcomes to the characteristics of democracies or possibly to alternative causal mechanisms; and to explore additional testable implications of the democratic peace hypothesis, including differences in leaders' perceptions of democratic and nondemocratic adversaries.103 Page 448 → Case study methods can also be usefully combined with formal rational choice theories, in part because of the difficulty of systematically measuring some of the key concepts in rational choice theory (preferences, utilities, probabilities, and informational environments, for example) across a large number of cases. In his study of the July 1914 crisis, for example, Levy empirically examined how political leaders perceived the set of feasible

outcomes of the crisis and rank-ordered their preferences over those outcomes, and used this framework to anchor an analytic case study of the outbreak of World War I.104 Examples of a more deductively structured use of rational choice theory to guide case study analysis include the “analytic narratives” research program105 and Bueno de Mesquita's analysis of church-state relations in medieval Europe.106

Conclusions In a volume on quantitative and qualitative methods it is natural to focus on questions of method. The utility of particular methods cannot be separated from questions of theory, however, and in many respects the greatest potential for advances in our knowledge about international relations remains theoretical. Developing better theories is particularly important for those who wish to test theories with qualitative methods, because the number of observations needed to test a theory is inversely related to the precision of a theory's predictions. The stronger the theory, the more specific its predictions, and the greater the divergence in predictions from those of a rival theory, the fewer the number of observations that are necessary to provide a satisfactory test of the theory, and thus the more valuable case study methods. This is why case study methods are so useful for testing theories that posit necessary or sufficient conditions. If theories are weaker, and if divergence between the predictions of competing theories is smaller, a greater number of observations are required for a meaningful empirical test, and case study methods are at a disadvantage. Building better theories is not the only solution here. Another is to think more carefully about the testable implications of existing theories, in terms of quality as well as quantity. Competing theories generate both overlapping and divergent testable implications. The former are irrelevant for testing competing theories, but the latter are critical. As a field we have probably done a better job building theories and developing methods for testing them than thinking creatively about identifying those testable implications of competing theories that are most divergent and consequently most conducive to providing definitive tests between rival theories. The application of all methods would benefit from more imaginative and clever thinking at this critical stage of research, at the juncture of theory and research design, but qualitative case study researchers have particularly strong incentive to do this.

Notes 1. Giovanni Sartori, “Concept Misinformation in Comparative Politics,” American Political Science Review 64, no. 4 (1970): 1033–53.Page 449 → 2. Idiographic inquiry aims to describe, understand, and interpret individual events or a temporally and spatially bounded series of events, whereas nomothetic inquiry aims to generalize about the relationships between variables and, to the extent possible, construct lawlike propositions about social behavior. See Jack S. Levy, “Explaining Events and Testing Theories: History, Political Science, and the Analysis of International Relations,” in Bridges and Boundaries, ed. Colin Elman and Miriam Fendius Elman (Cambridge: MIT Press, 2001), 39–83. 3. For example, Brecher, Steinberg, and Stein argued that most empirically oriented foreign policy analyses were single country in orientation and “usually devoid of theoretical value.” Michael Brecher, Blema Steinberg, and Janice Stein, “A Framework for Research on Foreign Policy Behavior,” Journal of Conflict Resolution 13, no. 1 (1969): 75. 4. Sidney Verba, “Some Dilemmas in Comparative Research,” World Politics 20, no. 1 (1967): 111–27; Adam Przeworski and Henry Teune, The Logic of Comparative Social Inquiry (New York: Wiley, 1970); Arend Lijphart, “Comparative Politics and the Comparative Method,” American Political Science Review 65, no. 3 (1971): 682–93; Arend Lijphart, “The Comparable Cases Strategy in Comparative Research,” Comparative Political Studies 8, no. 2 (1975): 133–77; Harry Eckstein, “Case Study and Theory in Political Science,” in Handbook of Political Science, vol. 7, ed. Fred I. Greenstein and Nelson W. Polsby (Reading,

Mass.: Addison-Wesley, 1975), 79–138; Alexander L. George, “Case Studies and Theory Development,” in Diplomacy: New Approaches in Theory, History, and Policy, ed. Paul Lauren (New York: Free Press, 1979), 43–68. 5. Arthur Stinchcombe, Constructing Social Theories (New York: Harcourt, Brace and World, 1968); Neil Smelser, “The Methodology of Comparative Analysis,” in Comparative Research Methods, ed. Donald P. Warwick and Samuel Osherson (Englewood Cliffs, N.J.: Prentice Hall, 1973), 42–86; Theda Skocpol and Margaret Somers, “The Uses of Comparative History in Macrosocial Inquiry,” Comparative Studies in Society and History 22, no. 2 (1980): 156–73. 6. Barrington Moore Jr., Social Origins of Dictatorship and Democracy: Lord and Peasant in the Making of the Modern World (Boston: Beacon Press, 1966); Charles Tilly, ed., The Formation of National States in Western Europe (Princeton, N.J.: Princeton University Press, 1975); Theda M. Skocpol, States and Social Revolutions (New York: Cambridge University Press, 1979). 7. Charles Tilly, Big Structures, Large Processes, Huge Comparisons (New York: Russell Sage Foundation, 1984); Alexander George and Timothy McKeown, “Case Studies and Theories of Organizational Decision Making,” in Advances in Information Processing in Organizations, ed. Robert Coulam and Richard Smith (Greenwich, Conn.: JAI Press, 1985), 43–68; Charles C. Ragin, The Comparative Method (Berkeley: University of California Press, 1987); Charles C. Ragin, “Introduction to Qualitative Comparative Analysis, ” in The Comparative Economy of the Welfare State, ed. Thomas Janoski and Alexander M. Hicks (New York: Cambridge University Press, 1994), 299–319; Barbara Geddes, “How the Cases You Choose Affect the Answers You Get: Selection Bias in Comparative Politics,” in Political Analysis, vol. 2, ed. James A. Stimson (Ann Arbor: University of Michigan Press, 1990), 131–50; Charles C. Ragin and Howard Becker, eds., What Is a Case? (New York: Cambridge University Press, 1992); David Collier, “The Comparative Method,” in Political Science: The State of the Discipline II, ed. Ada Finifter (Washington, D.C.: American Political Science Association, 1993), 105–19; David Collier and James Mahoney, “Conceptual Stretching Revisited: Adapting Categories in Comparative Analysis,” American Political Science Review 87, no. 4 (1993): 845–55. 8. Gary King, Robert Keohane, and Sidney Verba, Designing Social Inquiry (Princeton, N.J.: Princeton University Press, 1994). 9. David Collier and James Mahoney, “Insights and Pitfalls: Selection Bias in Qualitative Research,” World Politics 49, no. 1 (1996): 56–91; Douglas Dion, “Evidence and Inference in the Comparative Case Study,” Comparative Politics Page 450 → 30, no. 2 (1998): 127–46; Timothy J. McKeown, “Case Studies and the Statistical World View,” International Organization 53, no. 1 (1999): 161–90; Charles C. Ragin, Fuzzy-Set Social Science (Chicago: University of Chicago Press, 2000); Alexander L. George and Andrew Bennett, Case Studies and Theory Development (Cambridge: MIT Press, forthcoming). 10. For further information about the consortium, and for syllabi on qualitative methods, see http://www.asu.edu/clas/polisci/cqrm. 11. There has been an interesting reversal in the “balance” between the literature on qualitative methodology and applications of qualitative methods over the last three decades. In the early 1970s, the methodological literature on comparative and case study methods failed to adequately reflect the theoretical and methodological sophistication of some of the best applied research in the field, including work by George and Brecher. Today, applied qualitative research probably lags behind developments in the literature on the methodology of qualitative analysis. Alexander L. George and Richard Smoke, Deterrence in American Foreign Policy (New York: Columbia University Press, 1974); Michael Brecher, The Foreign Policy System of Israel: Setting, Images, Process (New Haven, Conn.: Yale University Press, 1972). 12. Lijphart, “Comparative Politics and the Comparative Method,” 682. 13. Collier, “The Comparative Method,” 105. 14. Andrew Bennett, “Case Study: Methods and Analysis,” in International Encyclopedia of the Social and Behavioral Sciences, ed. Neil J. Smelser and Paul B. Baltes (New York: Pergamon, forthcoming). 15. Ragin and Becker, What Is a Case? 16. See George, “Case Studies and Theory Development.” 17. If we think of cases as instances of broader theoretical categories, and if we acknowledge that historians are more idiographic in orientation than are political scientists, it should not be surprising that historians rarely use the term case. Levy, “Explaining Events and Testing Theories.”

18. John Walton, “Making the Theoretical Case,” in Ragin and Becker, What Is a Case? 122. 19. For further discussion of whether cases are made or found see Ragin and Becker, What Is a Case? 20. See King, Keohane, and Verba, Designing Social Inquiry. 21. Eckstein, “Case Study and Theory in Political Science,” 85. 22. Lijphart, “Comparative Politics and the Comparative Method,” 691. 23. Eckstein, “Case Study and Theory in Political Science,” 96–123. The first two in the list come from Verba, “Some Dilemmas in Comparative Research.” 24. Eric Hobsbawm, On History (New York: New Press, 1997), 109. Because of their “total” orientation toward the subject matter, atheoretical case studies do not fit the definition of a “case” as an instance of a broader class of phenomena. 25. Stephen Van Evera, Guide to Methods for Students of Political Science (Ithaca, N.Y.: Cornell University Press, 1997); Levy, “Explaining Events and Testing Theories.” 26. One example is Allison's use of three alternative theoretical frameworks to interpret the Cuban missile crisis (though this study had theoretical aims as well). Graham Allison, Essence of Decision: Explaining the Cuban Missile Crisis (New York: Little, Brown, 1971). 27. Christopher Achen and Duncan Snidal, “Rational Deterrence Theory and Comparative Case Studies,” World Politics 41, no. 2 (1989): 145. 28. Ibid., 169. 29. The widely accepted injunction against testing a theory or hypothesis against the same data from which it was generated does not preclude us from testing hypotheses generated in a case study against different data from the same case. If a case study of the origins of the cold war leads a researcher to formulate hypotheses about enduring rivalries, and if those hypotheses have implications for the termination of enduring or militarized rivalries, then those same hypotheses can be tested against evidence from the end of the cold war. Indeed, Page 451 → one of the most useful ways of validating a historical interpretation is to derive additional implications of that interpretation for other aspects of that particular historical episode. 30. George and Smoke, Deterrence in American Foreign Policy. 31. Robert Bates, Avner Greif, Margaret Levi, Jean-Laurent Rosenthal, and Barry Weingast, Analytic Narratives (Princeton, N.J.: Princeton University Press, 1998). 32. George, “Case Studies and Theory Development.” 33. See James Lee Ray, Democracies and International Conflict (Columbia: University of South Carolina Press, 1995); John Owen IV, Liberal Peace, Liberal War: American Politics and International Security (Ithaca, N.Y.: Cornell University Press, 1997). 34. Noted in Howard S. Becker, “Cases, Causes, Conjunctures, Stories, and Imagery,” in Ragin and Becker, What Is a Case? 209. 35. George, “Case Studies and Theory Development” George and Bennett, Case Studies and Theory Development. 36. David Collier and Steven Levitsky, “Democracy with Adjectives: Conceptual Innovation in Comparative Research,” World Politics 49, no. 3 (1997): 430–51. One caveat is worth noting. If the proliferation of subtypes and of the number of causal paths leading to them is taken to an extreme, the theory might explain all the anomalies but only at the cost of a loss of analytic power and parsimony. This problem is minimized if the variety of causal paths is linked to the conditions under which each is most likely to occur. 37. Arthur L. Stinchcombe, Theoretical Methods in Social History (New York: Academic Press, 1978), 5. 38. Lijphart, “The Comparable Cases Strategy in Comparative Research,” 159. 39. Hugh Stretton, The Political Sciences (London: Routledge and Kegan Paul, 1969), 245–47. 40. Lijphart, “Comparative Politics and the Comparative Method,” 692. 41. The many variables/small-N problem affects the use of case studies for theory testing but not for describing and explaining individual historical episodes or for generating or refining hypotheses to be tested by other means. If the universe of cases is relatively small (hegemonic wars, for example), and if the aim of case study analysis is to generate hypotheses, there may be advantages of using a small number of cases. This leaves a maximum number of cases for hypothesis testing, given the need to test hypotheses on data that are independent of the data from which the hypotheses are generated. 42. See also King, Keohane, and Verba, Designing Social Inquiry, 217–18.

43. Lijphart, “Comparative Politics and the Comparative Method,” 685. 44. Lijphart, “The Comparable Cases Strategy in Comparative Research,” 163. 45. Campbell also emphasized the utility of a single case, but only after retracting his earlier argument to the contrary. Donald Campbell, “Degrees of Freedom and the Case Study,” Comparative Political Studies 8, no. 2 (1975): 178–93. See also McKeown, “Case Studies and the Statistical World View.” 46. John Frendreis, “Explanation of Variation and Detection of Covariation: The Purpose and Logic of Comparative Analysis,” Comparative Political Studies 16, no. 2 (1983): 255. 47. John Stuart Mill, A System of Logic (1875; reprint, London: Longman, 1970). 48. Przeworski and Teune, The Logic of Comparative Social Inquiry; Theodore Meckstroth, “‘Most Different Systems’ and ‘Most Similar Systems’: A Study in the Logic of Comparative Inquiry,” Comparative Political Studies 8, no. 2 (1975): 133–77. 49. This terminology has generated some confusion. Mill defines agreement or difference in terms of the dependent variable, whereas most similar and most different are defined in terms of explanatory variables. Thus Mill's method of agreement is equivalent to a most-different systems design, and Mill's method of difference is equivalent to a most-similar systems design. 50. Lijphart, “The Comparable Cases Strategy in Comparative Research” Smelser, “The Methodology of Comparative Analysis.”Page 452 → 51. Przeworski and Teune, The Logic of Comparative Social Inquiry. 52. Mill, A System of Logic, bk. 6, ch. 7. 53. Lijphart, “Comparative Politics and the Comparative Method,” 688. 54. George, “Case Studies and Theory Development” George and Bennett, Case Studies and Theory Development. 55. Stephen Peter Rosen, Societies and Military Power: India and Its Armies (Ithaca, N.Y.: Cornell University Press, 1996). 56. Jack Snyder, Myths of Empire: Domestic Politics and International Ambition (Ithaca, N.Y.: Cornell University Press, 1991). 57. See King, Keohane, and Verba, Designing Social Inquiry, 126; Collier and Mahoney, “Insights and Pitfalls.” 58. King, Keohane, and Verba, Designing Social Inquiry, ch. 4. 59. Ibid., ch. 4. 60. Collier and Mahoney, “Insights and Pitfalls.” 61. An example involving large-N designs is Bueno de Mesquita's empirical analysis of his core hypothesis that a positive expected utility for war is a necessary condition for a state to initiate war. Bruce Bueno de Mesquita, The War Trap (New Haven, Conn.: Yale University Press, 1981). 62. Dion, “Evidence and Inference in the Comparative Case Study.” 63. The proposition that democracies rarely if ever go to war with each other posits that joint democracy is a sufficient condition for peace, and for the analysis of this proposition scholars focus exclusively on cases of democratic dyads. 64. Ragin, The Comparative Method. 65. Ragin argues that because the number of interaction effects necessary to capture combinatorial effects increases rapidly with the number of variables, serious degrees-of-freedom problems result. Ibid. For a new statistical approach to modeling multiple causal paths see Bear F. Braumoeller, “Modeling Multiple Causal Paths: Logic, Derivation, and Implementation” (manuscript, Harvard University, 2000). 66. For an application of Boolean models to the study of deterrence see Frank P. Harvey, “Practicing Coercion: Revisiting Successes and Failures Using Boolean Logic and Comparative Methods,” Journal of Conflict Resolution 43, no. 6 (1999): 840–71. 67. Eckstein, “Case Study and Theory in Political Science,” 113–23. 68. This assumes that measurement error is low. 69. Arend Lijphart, The Politics of Accommodation (Berkeley: University of California Press, 1968). 70. See David Bicknell Truman, The Governmental Process (New York: Knopf, 1951). 71. Ronald Rogowski, “The Role of Theory and Anomaly in Social-Scientific Inference,” American Political Science Review 89, no. 2 (1995): 467–70. Lijphart also contributed to theory construction by suggesting hypotheses on why the theory of cross-cutting cleavages broke down in the Netherlands. See

Lijphart, The Politics of Accommodation. 72. Imre Lakatos, “Falsification and the Methodology of Scientific Research Programmes,” in Criticism and the Growth of Knowledge, ed. Imre Lakatos and Alan Musgrave (Cambridge: Cambridge University Press, 1970), 91–196. 73. Allison, Essence of Decision. 74. Similarly, if the rational-unitary model cannot explain state behavior in an international crisis as acute as the one in 1962, we would have little confidence that it could explain behavior in situations of noncrisis decision making. 75. George, “Case Studies and Theory Development” George and Bennett, Case Studies and Theory Development. 76. Ibid. 77. Ernest Nagel, The Structure of Science (Indianapolis: Hackett, 1979), 564–68; W. B. Gallie, “The Historical Understanding,” History and Theory 3, no. 2 (1963): 149–202. 78. Carl G. Hempel, “The Function of General Laws in History,” Journal of Philosophy 39 (1942): 35–48.Page 453 → 79. Experimental methods may be superior for testing many of these hypotheses because of their ability to control for extraneous variables. It is often difficult, however, to generalize from laboratory settings that do not incorporate the relevant political variables and that cannot fully replicate the stakes and emotions inherent in the contexts of foreign policy decision making. 80. Sidney Tarrow, “Bridging the Quantitative-Qualitative Divide in Political Science,” American Political Science Review 89, no. 2 (1995): 474. 81. Paul Pierson, “Increasing Returns, Path Dependence, and the Study of Politics,” American Political Science Review 94, no. 2 (2000): 251–67; Ira Katznelson, “Periodization and Preferences: Contributions of Comparative Historical Social Science” (manuscript, Columbia University, 2000). 82. Andrew Bennett and Alexander L. George, “The Alliance of Statistical and Case Study Methods: Research on the Interdemocratic Peace,” APSA-CP Newsletter 9, no. 1 (1998): 6–9. 83. A good example of a large-N study that empirically distinguishes between alternative causal mechanisms within the black box of decision making is Schultz's statistical test of competing institutional constraint and informational signaling models of the democratic peace proposition. Kenneth A. Schultz, “Do Democratic Institutions Constrain or Inform? Contrasting Two Perspectives on Democracy and War,” International Organization 53, no. 2 (1999): 233–66. 84. Many of those committed to large-N analysis exaggerate in the opposite directions. 85. Clayton Roberts, The Logic of Historical Explanation (University Park: Pennsylvania State University Press, 1996). 86. Thomas D. Cook and Donald T. Campbell, Quasi-Experimentation (Chicago: Rand McNally, 1979), ch. 1. 87. If theories have no observable consequences, they cannot explain the variation in outcomes that is the primary task of social science, and hence those theories have no explanatory power. This does not imply that all theories are easy to test. 88. For an argument that hypotheses involving necessary conditions are more common, see Gary Goertz and Harvey Starr, eds., Necessary Conditions: Theory, Methodology, and Applications (Lanham, Md.: Rowman and Littlefield, forthcoming). 89. These are very demanding criteria. They lead Lieberson to conclude that Mill's methods cannot be applied if the number of cases is small because they do “not allow for probabilistic theories, interaction effects, measurement errors, or even the presence of more than one cause.” Stanley Lieberson, “Small N's and Big Conclusions,” in Ragin and Becker, What Is a Case? 105–18; Stanley Lieberson, “More on the Uneasy Case for Using Mill-Type Methods in Small-N Comparative Studies,” Social Forces 72 (June 1994): 1225–37. Multiple causes and interaction effects can probably be dealt with, but only by increasing the number of cases in order to encompass all possible interaction effects. This creates a practical problem for intensive case study methods. Some case study researchers acknowledge this problem and supplement Mill's methods with process tracing (George and Bennett, Case Studies and Theory Development), while others have adopted a Boolean-based “qualitative comparative analysis” (Ragin, The Comparative Method). 90. The large-N researcher makes the opposite trade-off.

91. Whether social science theories can have causal mechanisms that are truly probabilistic—as opposed to those that reflect measurement error, the misspecification of functional forms, or omitted variables that have yet to be understood and specified—is quite problematic. See Wesley C. Salmon, Four Decades of Scientific Explanation (Minneapolis: University of Minnesota Press, 1990), and George and Bennett, Case Studies and Theory Development. I prefer to speak of a theory's probabilistic testable implications rather than probabilistic causal mechanisms. My thinking (admittedly provisional) on this complex issue has benefited from conversations with Andrew Bennett.Page 454 → 92. Ragin and Zaret, 1983, 744. 93. Ragin, The Comparative Method, 15. Ragin's exploration of the relevance of the logic of fuzzy sets for social science analysis is an attempt to deal with the problem of measurement error and probabilistic testable implications. Ragin, Fuzzy-Set Social Science. 94. Becker, “Cases, Causes, Conjunctures, Stories, and Imagery,” 212. 95. Dorothy Vaughan, “Theory Elaboration: The Heuristics of Case Analysis,” in Ragin and Becker, What Is a Case? 184. Proponents of “offense/defense theory,” for example, theorize about whether military technology favors the offense or defense but fail to establish the magnitude of this effect relative to that of the balance of material capabilities and related variables. Sean M. Lynn-Jones, “Offense-Defense Theory and Its Critics,” Security Studies 4, no. 4 (1995): 660–91. 96. Sartori, “Concept Misinformation in Comparative Politics” Collier and Mahoney, “Conceptual Stretching Revisited.” 97. Skocpol and Somers, “The Uses of Comparative History in Macrosocial Inquiry.” 98. Paul Huth, Extended Deterrence and the Prevention of War (New Haven, Conn.: Yale University Press, 1988); Ray, Democracies and International Conflict; Lisa L. Martin, Coercive Cooperation (Princeton, N.J.: Princeton University Press, 1992); Beth A. Simmons, Who Adjusts? (Princeton, N.J.: Princeton University Press, 1994). 99. Brecher, The Foreign Policy System of Israel; Michael Brecher, with Benjamin Geist, Decisions in Crisis: Israel, 1967 and 1973 (Berkeley: University of California Press, 1980); Michael Brecher and Jonathan Wilkenfeld, A Study of Crisis (Ann Arbor: University of Michigan Press, 1997). 100. Michael Doyle, “Liberalism and World Politics,” American Political Science Review 80, no. 4 (1986): 1151–70; Michael Doyle, Ways of War and Peace (New York: Norton, 1997). 101. Edward D. Mansfield and Jack Snyder, “Democratization and the Danger of War,” International Security 20, no. 1 (1995): 5–38; Jack Snyder, From Voting to Violence: Democratization and Nationalist Conflict (New York: Norton, 2000). 102. Bruce M. Russett, “International Behavior Research: Case Studies and Cumulation,” in Approaches to the Study of Political Science, ed. Michael Haas and Henry S. Kariel (Scranton, Penn.: Chandler, 1970), 425–43; Lijphart, “Comparative Politics and the Comparative Method.” 103. Bennett and George, “The Alliance of Statistical and Case Study Methods.” 104. Jack S. Levy, “Preferences, Constraints, and Choices in July 1914,” International Security 15, no. 3 (1990–91): 151–86. 105. Bates et al., Analytic Narratives. 106. Bruce Bueno de Mesquita, “Popes, Kings, and Endogenous Institutions: The Concordat of Worms and the Origins of Sovereignty,” International Studies Review 2, no. 2 (2000): 93–118.

Page 455 →

CASE STUDY METHODOLOGY IN INTERNATIONAL STUDIES From Storytelling to Hypothesis Testing Zeev Maoz

Introduction Case study is arguably the most common methodology in international studies—including both international relations and comparative politics. It is also the most widely used across subfields of the discipline. Whereas other methodologies are more likely to be associated with certain substantive subfields than others, the case study methodology appears to be sufficiently flexible so as to apply to almost all subfields. For example, it is difficult to imagine the application of game theory or rational choice models to a constructivist analysis of change in international politics, but a case study employing the concepts and tools of constructivism can be easily imagined.1 Likewise, the analysis of singular events—such as the collapse of the Soviet Union—through the use of quantitative studies using a large number of cases is difficult to apply. Yet an in-depth case study of such events is easily conceivable.2 It is ironic that—compared with other disciplines where case studies have been used extensively, for example, psychiatry and clinical psychology, social work, social anthropology, ethnography, and history—the discussion of the foundations and considerations of this approach in political science is relatively sparse and surprisingly unsophisticated. This is reflected not only in the methodological discussions of political case studies. It is evident even more in the practice—or malpractice—of case study methodologies in empirical research. The present paper offers a general criticism of the way case study methodology has been used in research on international studies. I do not suggest that case study is an improper or somehow inferior methodology in such research. On the contrary, from several perspectives, case study research is not only a proper methodology in such research but offers several advantages that other methodologies do not. The problem is not in the method; it is in the way it was used. Consequently, I attempt to offer some elements of a general strategy that can serve the systematic use of case study research—mostly, but not exclusively—for confirmatory purposes.3 Page 456 → I focus on three aspects of the case study methodology that are symptomatic of the misuse of this methodology and suggest some principles and considerations for a more proper use of this approach: case selection, case design, and cross-case comparison. All three aspects of case study methodology have a major impact on the ability to make inferences from the study. This paper is designed as follows. First, I discuss some of the important properties of case study methodology for confirmatory purposes. I argue that case study may in many ways have significant advantages over other methods used in political research. However, to capitalize on those advantages, systematic considerations of selection, design, and comparative analysis must be built into the research projects. Second, I discuss the problems of case selection, case design, and cross-case comparison using examples of two projects that have relied on case study research: projects on deterrence theory4 and the case study examination of the democratic peace proposition.5 Finally, I discuss the implications of these issues for theory development and theory testing in international relations.

Some Advantages of Case Study Research Methodologies The literature on case study in political science,6 as well as more general treatment of this approach,7 suggests that case study approaches have some important properties that make them useful as a research methodology. Some of these points are important not only in justifying case study research, in general, but also in preferring this approach to other research methodologies in some cases.

Cost-Effective Exploratory Research Methodology Case studies enable exploratory analyses at relatively low cost of resources. Conducting surveys or data collection projects on large populations just to discover that the variables used in the study are not the central ones in the process under study may be an exercise in futility. Even the cost of a pilot study on these issues may be exorbitant. Process Tracing Good theories are more than a set of hypotheses each positing a bivariate relationship between variables. Good theories—especially in a social context—outline a process that tells a story; the more dynamic the story, the more aesthetically attractive the theory.8 However, capturing a story in a set of quantitative analyses is often a very complex operation (for example, requiring multiple equation models, dealing with endogeneity problems, and so forth). Due to both methodological and practical problems, quantitative analyses often break theories into sets of bivariate or multivariate hypotheses that do not capture the dynamic aspects of the stories. For that reason, the actual ability to match a hypothetical story that contains an explicit Page 457 → process derived from the theory with an actual story is a unique feature of the case study method. Just as the ability to match a theory-derived story with the actual case is important, so is the potential of uncovering deviations between the actual story and the theory. By uncovering differences between the expected story and the actual story, a researcher could identify factors that are outside of the theory but that account for important facts in the historical case. Unveiling Causal Processes in Nonevents: Dogs That Don't Bark and Dogs That Don't Growl An important attribute of in-depth case study research is the ability to uncover nonevents and their characteristics. Unfortunately, this attribute has received little attention in the literature, and—until quite recently—was underexplored in actual research. Briefly, certain theoretical approaches attempt to account for nonbehavior (alienation-related explanations of non-turnout in voting, democracy and peace, deterrence success, and so forth). In such cases, process tracing can attempt to uncover the dynamics of the propositions not only by examination of deviant cases but also by examining whether the proposition offers a viable explanation in “typical” cases that seemingly “match” the hypothesized nonbehavior. I will elaborate on this property later. Systematic Comparison of Processes A given variable x might be related to another variable y across a large number of cases, but the process by which these variables may be related can differ across cases. For example, education may be related to income across a large number of individuals, but the relationship may be different for different individuals. For example, a subset of the cases (people) fitting the rich-educated bracket may be actually old-money people, that is, people who were born into affluent families and whose education was determined by their family origins. Another subset of rich people may well be people who come from lower-income families and, due to their drive to obtain high education, acquire education-related wealth. Likewise, some poor people may be born into poor families who were unable to support their education, thus making for second-generation poverty. On the other hand, another subset of poor people may be people who come from affluent families who failed to adjust to structural changes in the economy (for example, rural families who lost their wealth as prices for agricultural products declined). Systematic comparison of processes may reveal unspecified features of the theory, even theories that were supported by quantitative evidence. As such, the ability of mixing exploratory and confirmatory analysis in the same design is an important feature of case study research. In light of these advantages, it is evident why case study approaches have been so commonly used in political research. Yet it is ironic that little cumulation has been accomplished in international studies due to the use of Page 458 → case studies. In the following sections, I discuss some basic features of case study use and misuse in international relations in order to suggest a more coherent general strategy for the use of such research in the future.

Case Selection There is a nearly complete lack of documentation of the approach to data collection, data management, and data analysis and inference in case study research. In contrast to other research strategies in political research where authors devote considerable time and effort to document the technical aspects of their research, one often gets the impression that the use of case study absolves the author from any kind of methodological considerations. Case studies have become in many cases a synonym for freeform research where everything goes and the author does not feel compelled to spell out how he or she intends to do the research, why a specific case or set of cases has been selected, which data are used and which are omitted, how data are processed and analyzed, and how inferences were derived from the story presented. Yet, at the end of the story, we often find sweeping generalizations and “lessons” derived from this case. We are also confronted by policy prescriptions from time to time. In other instances authors pay lip service to the methodology, citing typically outdated methodological discussions of case study approaches in political science, and jump straight into the story. The studies in which authors make a conscious effort to spell out what exactly they attempt to do and why they are doing what they are doing are extremely few. Consider the inferential leaps in the discussion of methodology in one of the—admittedly more insightful—studies of the cold war: We examine the impact of deterrence and compellence in detailed studies of the two most acute Soviet-American confrontations of the last quarter century: the Cuban missile crisis of 1962 and the crisis in the Middle East in 1973. We demonstrate the largely pernicious consequences of threatbased strategies in most cases…. In our reconstruction of both crises, we explore the contradictory consequences of nuclear threats and nuclear weapons. These two crises provide a window on the broader relationship between the United States and the Soviet Union during the Cold War…. We contend that the strategy of deterrence prolonged the Cold War and helped to extend the life of communism in Eastern Europe and the Soviet Union. We agree with George Kennan that the primacy of military over political policy in the United States during the Cold War delayed rather than hastened the great changes that finally overtook the Soviet Union. By looking through the window of these two crises at the broader superpower relationship, we can learn lessons that will be applicable to the prevention, management, and resolution of international conflict beyond the Cold War.9 Page 459 → The choice of the Cuban missile crisis needs little justification. It is universally recognized as the most acute confrontation of the Cold War…. We analyze a second Soviet-American crisis to control for the idiosyncrasies of Cuba. The most serious superpower crises of the postwar period were over Berlin in 1948–49 and 1959–62, and the crisis that erupted in the Middle East in 1973.… A comparative analysis of the missile crisis and Berlin would have been unrepresentative of Soviet-American interaction in crisis during the Cold War. The crisis in 1973 seemed a more appropriate choice. It reflects a different style of crisis management in Moscow and in Washington, but includes a strong nuclear threat, and in contrast to the Cuban missile crisis, took place in an era of strategic parity and superpower detente.10 As noted in the italicized portions of this text, a lot of heroic inferences are made from this study that uses two cases. Not only do the authors suggest important insights about the conduct and consequences of the cold war, but they also offer “insights” about crisis management and conflict resolution in more general domains. How credible are these insights? When authors discuss methodological considerations prior to the case study, they typically focus on case selection. The selection of cases is admittedly a very important element of a more systematic approach to case

study research. Case selection must be based on a basic understanding of the properties of the population from which the cases are drawn. This understanding requires—first and foremost—a clear and unbiased definition of the general population. In the examples discussed herein, authors in many cases make a sincere effort to define the population of cases; however, this definition is often biased. Therefore, the actual choice of cases—no matter how carefully done—tends also to be biased. Both qualitative deterrence literature and the qualitative democratic peace literature focus on crises. They offer elaborate explanations of why crises are important laboratories for testing propositions about deterrence or about the democratic peace. In both bodies of literature, however, crises represent only a small subset of the relevant cases for most types of investigation. The key problem in biased case selection lies in improper logical definition of the research problem. Often, it emerges out of a truncated specification of the problem. Consider a typical research problem posed by deterrence theory: do deterrence threats work, and if so, under what conditions are these threats effective? We expect, as do most deterrence theorists, that states resort to this strategy in situations where two parties confront each other in a conflict of interests. At least one party (the challenger) is perceived to want to change a certain political, territorial, or economic status quo. The other party (the defender) wishes to defend it. The deterrence situation consists of the defender making a series of threats that promise to impose an excessively high cost on the challenger should it act forcefully to change the Page 460 → status quo. We typically consider deterrence effective if it results in the peaceful (or nonviolent) preservation of that status quo. So, let us lay out the problem in terms of the relationship between deterrence and status quo preservation. Suppose we studied two hundred crises. In each crisis we identified the challenger and the defender. We examined whether or not threats were made by the defender, whether they were credible (conveying both capability and intent), whether they were clearly communicated and clearly understood by the challenger, and so forth. Table 1 seemingly suggests that if we coded two hundred crises, using systematic case studies or other data collection methods, we would conclude that there is fairly good evidence that deterrence is an effective warprevention strategy. But the context in which deterrence operates is not crises but conflict, of which crises are but one—fairly extreme—manifestation. Testing the effectiveness of threats in crisis is akin to studying determinants of election outcomes on the basis of exit polls; the information about nonvoters may be as important or more important in many respects as the information about voters' choices. Therefore a truncated definition of the population is apt to lead to biased inferences. Consider the example discussed previously. Suppose that the actual crises constitute a small proportion of all crisis opportunities, that is, situations that could potentially erupt into crises. Suppose, also, that once we examined those opportunities, we got the distribution shown in table 2. Now, the positive relationship between effective deterrence threats and status quo preservation turned into a negative one. This inferential fallacy stems from selection bias. It can arise out of quantitative research as well as from case studies. However, because case studies by definition represent a much smaller—and typically nonrandom—subset of all cases, the likelihood of inferential bias is substantially higher than in quantitative analysis. The same logic applies to the democratic peace case study literature. Simply replace the labels of the independent variable by Joint Democracy/No Joint Democracy, and one gets the same type of problem. Page 461 → How do we deal with these problems? First and most important, we must recognize that case selection entails the risk of selection bias and thus of resultant inferential bias. Second, it is important to recognize the context in which case selection—under any type of systematic confirmatory method—is conducted. This context defines to a large extent the range of solutions available to the researcher. The elements of the context are twofold: the logic of

experimental design and the concept of sampling. A step toward proper case selection process requires considering both elements. Without going into the logic of experimental design in great detail, the basic notion is that two processes operate simultaneously, one with an experimental treatment and one without. However, assignment of subjects to groups is based on systematic establishment of similarity among subjects. Specifically, the control and experimental groups should be identical (or similar) in terms of their distribution on a number of variables that are considered confounding (that is, likely to affect the outcome of the process). Where the two groups differ is in the type of experimental treatment they receive. Thus, selection of subjects is done on the basis of confounding variables, and assignment to groups determines the value of independent variable (no treatment, treatment Type I, Type II, etc.). The purpose of these features is to ensure control and variability in independent variables. Once the distribution of the overall population of subjects has been determined in terms of the potential confounding variables, the assignment of subjects to control/experimental groups is based on randomizing within strata. The implication is that cases are not selected on the basis of the distribution of the dependent variable but rather in terms of variability in a set of potentially confounding variables. More important, if a systematic selection of cases is conducted, it should be on the basis of the distribution of independent variables rather than on the basis of the dependent variable. Sampling requires a similar process. A certain part of the process must allow for systematic stratification of the population in terms of confounding Page 462 → variables, and in terms of the independent variables. If we wish to examine hypotheses about the relationship between foreign policy positions and presidential candidate preferences, and we have reason to believe that candidate preference is also affected by race and region, we want to make sure that the racial and regional makeup of the sample is proportional to the distribution of these variables in the general population. This requires systematic—or stratified—sampling. Another part of the process, however, requires randomization. Sampling and experimental design provide some clues for a rational case selection strategy in case study design. The following principles may be applicable: 1. Understand your theory. Lay out the hypotheses and spell out their status. Do your hypotheses state that X is a necessary condition for Y, a sufficient condition, or both? 2. Define what kind of variability is required to test your theory. In general, the greater variability you allow for in your design, the more credible your conclusions are likely to be. 3. Identify potentially confounding variables—that is, factors that are likely to affect the dependent variable/s that are not explicitly included in your theory. 4. Explicitly define your population, and assess the consequences of the criteria of inclusion and exclusion in your population. 5. Attempt, as far as possible, to determine how the population is distributed on the confounding variables.

Case Study Design The key difference between exploratory and confirmatory case studies lies in case design. An analogy might help explain this point. Consider a detective trying to solve a murder case. By the time the detective gets to the crime scene, the basic contours of the situation have been worked out by the crime scene police teams (forensics, pathology, etc.). The identity of the victim (if available) is known, the murder weapon (if available) is identified, and the proximate time of death is established. Other basic information is gleaned just at the outset, and this information forms the basis of the detective's investigation. The first stage of the investigation entails identifying suspects. If several suspects are identified, the detective tries to narrow the range of suspects by matching details about each of them with the characteristics and clues from the murder scene and other background material. Alibis are checked by matching the suspects' whereabouts with the proximate time of murder. Motives are investigated, and physical evidence from the murder scene or related material (e.g., fingerprints, DNA, blood stains, etc.) is matched with suspects and their belongings. A process of elimination ensues until one person emerges as the prime suspect.

At that point, a case for the prosecution is prepared. At this stage, the process consists of integrating the evidence into one coherent story that Page 463 → contains a logical and consistent explanation of the motive and the process leading up to the murder (as well as postmurder events that are connected to the crime scene, such as cover-up attempts—hiding the body, the murder weapon, destroying physical evidence, and so forth). A hypothetical reconstruction of the murder is conducted (assuming that the suspect does not confess). That reconstruction, then, is the case that the prosecutor presents in court. Just as in the analogy, by the time we start our case study, the basic contours of the case are known. We know the outcome (we have a body). We know the background of the case (we know something about the victim). And in the typical case study, we know the basic details of the story. Given this starting point, exploratory case study is equivalent to the first stage wherein the detective attempts to narrow down the range of suspects to one suspect. Confirmatory case study consists of preparing the case for prosecution in court once a prime suspect has been identified. The confirmatory advantages of the case study approach over other methods lie—as we have pointed out—in the ability to contrast the actual facts with an expected process. The case study can reveal the why—the causal mechanism—underlying a relationship of the type x is related to y specified in a hypothesis. However, if we are to take full advantage of this feature, we must specify a priori what kind of process we expect to unveil. The design problem in case studies consists of a backward-reasoning process—from an outcome to an antecedent explanation. This is especially the case with respect to those studies dealing with “familiar” cases, the general elements of which are known. The art of case study design is—to a large extent—the art of reconstruction. In exploratory design we typically work with a kind of “protocol” about the general class of problems within which the case falls. Just like the detective looking for suspects, the search is based on a general strategy, derived from cumulative past experience of the detective herself and of the general discipline of police investigation. But the general “protocol-based” search must gradually adapt to the specific case in point. In confirmatory research, the design of the case is theory driven. The theory lays out a hypothetical story. The researcher's role is to operationalize the story in the particular case. Specifically, the confirmatory case study design requires the a priori specification of a story that contains traceable elements, each of which is derived from the theory. Because research in the positivist tradition assigns greater significance to refutation than to confirmation, two types of elements are required, confirmatory and disconfirmatory. A prosecutor preparing a case for court must work on the assumption that she must prove guilt beyond reasonable doubt, whereas the defense's task is just to invoke reasonable doubt. Thus, physical evidence connecting the suspect to the murder scene (e.g., DNA from a hair sample found on the murder victim) is important confirmatory evidence. However, if the suspect has a firm alibi, this counts more in the final judgment. For that reason, the fact that there exists a plausible story and that the expected facts derived from that hypothetical story match the actual case is not sufficient Page 464 → to support the theory. The case study must be able to disconfirm other plausible explanations or to eliminate possible evidence that generates reasonable doubt about the theory in the case. Very often, the facts that match the story are not necessarily the opposite of the facts that invoke a reasonable doubt. Therefore, the design of the case must be based on both a priori specification of evidence for the theory and evidence against the theory. This requirement is especially important if what is under investigation is not one explanation but two competing ones, which is often the case in political research. In this case, the design requires not only specifying the evidence supporting each theory individually but also identifying the areas of converging expectations of both theories and—most important—the area of competing expectations. It is only in this area of competing expectations that evidence consistent with one of the theories is inconsistent with the other. The assumption that whenever two theories compete with one another all the expectations derived from them must also compete with each other often drops even the most experienced researcher into an inferential trap.11 Finally, as previously mentioned, one of the underused attributes of the case study methodology is the ability to overcome the problem of “the dog that didn't bark” (or even “the dog that didn't growl”).12 Consider the case of deterrence theory. One of the common complaints of scholars who use case studies to study deterrence theory is

that we can only study deterrence failures, because such instances are easily identifiable. Since successful deterrence induces inaction, it is impossible to actually test cases of deterrence success. This is analogous to claims that whereas the failure of legal sanctions in preventing criminal behavior is clearly detectable, the success of such sanctions cannot be fully and reliably established. But deterrence theory lays out a fairly straightforward process of war prevention. We first establish an initial condition of the presence of a conflict of interest. One side is said to harbor operational plans or general intentions to forcefully change a territorial and/or political status quo. The other side views that status quo as highly desirable and wants to preserve it. The defender then chooses to base its plan to preserve the status quo on deterrence, that is, on a set of credible threats promising to inflict disproportionate punishment on the challenger if it attempts to change the status quo by force. The threat is then communicated to the would-be challenger and, if effective, is supposed to alter the challenger's calculus of conflict initiation. What a good case study of deterrence success can do is define observable traces of such a process. Specifically: 1. Specify evidence for an intention by the challenger to upset the prevailing status quo by force (e.g., documentary or other indications of operational plans, political discussions, mobilization efforts, force design and deployment, and so forth). 2. Identify the deterrence effort and its operational manifestations. Specify what would constitute evidence that the defender identified either Page 465 → specific or general risk to the status quo emanating from the challenger: a decision process by the defender resulting in a deliberate policy of deterrence; a policy that communicates the deterrence threat and backs it up by concrete steps that are designed to establish or foster its credibility (e.g., military force-building and deployment, civilian mobilization, trip-wiring policies, boosting public support for the policy, etc.). 3. Identify evidence suggesting that the challenger has received and digested the deterrent threat and has incorporated the implications of the threat (the probability that the defender would carry it out, the cost of implementing the action subject to retaliation, and so forth). 4. Identify evidence suggesting that the decision-making process entailed specific consideration of the threat and its implications. In other words, evidence should suggest that there were references by the decision makers to the deterrent threat and to the intentions and capabilities of the defender. 5. Identify evidence suggesting that the calculus of the decision was altered due to the deterrent threat, so that the decision taken may well have been different from the one actually made either without the challenger's awareness of a deterrent threat or if the threat had been seen as less credible than it actually was. One approach to demonstrating the effect of the deterrent threat on the challenger's decision process is to conduct sensitivity analysis on elements of the credibility of the threat. For example, if decision makers referred to a certain level of probability that the defender would carry out the threat, assess how crucial this assessment was in determining decisional outcomes. Alternatively, if decision makers indicated that they might incur a certain cost as a result of the defender's implementation of its threatened response, consider what might have been the effect on the final outcome had the cost not been incorporated into the decision calculus, or if the extent of loss would have been lower.13 6. If the decision making of the challenger ends with a decision to refrain from violating the status quo or to delay decision on violation of the status quo, then we have a “candidate” for deterrence success. Otherwise, this is a fairly clear case of deterrence failure. 7. In a candidate case for deterrence success, we must make sure that our case is valid beyond a reasonable doubt.14 How do we do that? As noted earlier, two related operations must take place. First, we must make sure that the evidence for our theory is consistent with the theory's expectations as well as inherently logical. For example, we must show a correspondence between the nature of the threat issued by the defender and the way the threat was perceived by the challenger. We must be able to identify the impact of the deterrent threat on the challenger's calculus in prespecified terms, and so forth. Second, we must be able to identify potential evidence that may be consistent with alternative explanations and eliminate it. For example, if it is argued that public opinion or competing elites in the challenger's Page 466 → state opposed foreign policy adventurism, we must be able to identify how domestic considerations might have affected the challenger's calculus and examine how pivotal such evidence was in accounting for

7. the final decisional outcome.15 This could be done by employing a similar test to the one applied in order to determine how pivotal were deterrence-related considerations in the challenger's decision. 8. If possible, it is advisable to develop “crucial experiments” or discriminating tests, where competing theories offer contradictory predictions about a set of observable facts. This is a difficult operation, and not always feasible. Yet it is worth the effort because—if successfully conducted—it removes analytical ambiguities of the sort discussed previously. The significance of testing evidence for a theory against competing explanations cannot be overemphasized. Severe headaches are associated with migraines and hypertension, which are treated by painkillers, but such headaches are also associated with aneurysms or brain tumors that require highrisk surgery to treat. If neurosurgeons or general practitioners were to design their diagnostic processes as many of us design our case studies, the courts would have been overflowed with malpractice suits, and the number of unnecessary surgical operations would have been extremely high. Some examples from the case studies on the democratic peace can serve to demonstrate both the problem and avenues for dealing with it. Layne examines the Fashoda Crisis of 1895–98 between France and Britain.16 Layne also examines a series of crises that pitted France against Great Britain between 1830 and 1848.17 Rock examines U.S.-Anglo relations over the 1845–1930 era.18 These studies examine why crises between democracies did not escalate into war in an effort to test the normative and institutional explanation of the democratic peace proposition. In these studies, the authors lay out a set of hypotheses emerging, in their view, from the democratic peace and the political realist paradigms in order to examine whether democratic norms and institutions or Realpolitik considerations averted war in these crises.19 Because in none of these studies is there an a priori outline of a process that should follow from these sets of hypotheses, facts are presented in a manner suggesting that their selection and presentation were chosen deliberately to refute one theory and support another. For example, Layne's discussion of the 1830–32 “crises” between France and Britain over the Belgian question seems to ignore the seemingly negligible fact that in reality both France and Great Britain were on the same side of the issue: both supported Belgian independence from Holland, and both resisted Prussian attempts to reannex Belgium to Holland (and to involve Russia in the process). The debate between France and Great Britain in these crises was in fact over a British fear that a highly assertive action by France against Holland would provoke a combined Prussian-Austrian-Russian attack. Because of domestic reasons, the British feared that they would be drawn into such a war on Page 467 → France's side but against their own personal wishes. Examining the crises of 1830–32 strictly in a French-British context, without examining the broader context—wherein France used force successfully against the Dutch and both Britain and France joined forces to deter Prussia and Russia from aiding the king of Holland—suggests sloppy historical work apart from a weakly designed case study.20 But beyond creating the impression that history has become a victim of a certain theoretical inclination, the principal problem of these case studies is that none of the facts presented by the authors suggests that if one explanation is meaningful, the other is not. In fact, both domestic politics, institutional constraints, and hardcore realpolitik calculations may have played a role in each and every instance discussed by the authors. Because the two sets of hypotheses are spelled out such that they could both be additively valid (or additively false), they in fact support the findings of previous studies on the democratic peace that both realpolitik and normative or structural factors account for changes in the probability of dyadic conflict.21 Both case studies that attempt to study the phenomenon of peaceful termination of crises between democracies failed to specify a critical test, thus inhibiting any meaningful inferences about the validity or limitations of the democratic peace proposition.

Cross-Case Comparison Just as one must carefully design the research of an individual case, it is imperative to design a priori the comparison of cases. All those who have written about the use and strategy of case study research in political science have pointed out that case studies are useful confirmatory methods only if the results could be generalized beyond the cases. Let us return to our murder case analogy to illustrate this point. Each murder case is unique and

must be treated as such by an investigator. It has its individual victim, its specific circumstances, and of course, its murderer/s. As such it is seemingly ungeneralizable. However, if this were the case, either anybody could solve murder cases or nobody could. The fact that some people specialize in murder investigations implies that cumulative experience matters. The underlying assumption must be that no matter how unique each case may be, there are underlying characteristics of many murder cases that are similar across individual victims, circumstances, and perpetrators. The structure of such investigations is, in general, quite similar, and so is the structure of the prosecution's work once a suspect is identified. That cumulative experience may matter is a significant point in police work. But experience can make a difference only if an investigator is capable of inferring meaning from both similar and dissimilar features of a given case with respect to other cases. To be able to infer such meaning, a person must have in mind a “typical” structure of murder cases—so that each case can be matched against this structure.22 Thus, in comparative case study design, a necessary condition for generalization is a systematic design of individual cases and the execution of Page 468 → case studies in a manner that allows a fair comparison. There needs to be an underlying hypothetical story that guides the collection of data and the account of each case so that the final results of individual cases can be meaningfully compared with each other. The purpose of this underlying story is to enable the researcher to both match facts and detect mismatches. For example, in the outline of the deterrence process, we required that the challenger understand the deterrent threat in the terms intended by the defender. If the facts of an individual case suggest that this was not what happened in reality, we have a problem of communication. How this problem affects the challenger's decision process is something that becomes very significant as an indicator of the theory's propositions. An important part of the difficulty of drawing meaningful inferences from the case study literature on deterrence and the democratic peace stems from the fact that different case studies are based on different hypothetical stories (when such stories are given by the authors to begin with). This is typically the case when edited volumes seek to “test” a set of hypotheses derived from a given theory or research program. An examination of the Jervis, Lebow, and Stein volume on deterrence and the Fendius-Ellman volume on the democratic peace suggests that each author went pretty much on his or her own in terms of the methodology of individual cases.23 Without denying that many of these case studies offer important insights into the workings of deterrence theory, it is clear that any general inferences from these cases as a whole are virtually impossible. This is, first and foremost, because it is impossible to tell whether the inferences derived from an individual case—even if intrinsically valid in terms of the case itself—are due to the inherent characteristics of the case, the decision makers involved, the specifics of the dyadic or regional context, or the different design of the study by the author.24 Suppose we were to design an experiment on the effect of a certain drug on the rate of recovery of patients from pneumonia. Suppose we decided to replicate this experiment several times and assigned each experiment to another person. Suppose each person was instructed to divide the subjects into experimental and control groups, but we did not provide the experimenters with details of the experimental design. Now, suppose that once we collect the data from the experiments we notice that the age groups of the subjects varied across experiments: one experimenter used young subjects, whereas another experimenter used adults, and still another experimenter used older subjects. The experimental and control groups differed also in terms of the timing of the intervention: in one case the drug was given to subjects at an early stage of the disease, whereas in another it was administered at an advanced stage of the disease. Also, the dosages of the drug given to subjects varied substantially across experiments. Now, how are we to interpret variations in results across experiments? Are they due to the unstable effects of the drug or to the variations in the design in the experiment? Given the large number of factors that distinguished the experimental designs, it is impossible even to start assessing these questions. What is obvious is that this project can hardly constitute any type of evidence Page 469 → regarding the potential effect of the drug on the rate of recovery from pneumonia. When single authors develop comparative case studies, there is clearly more structure in the design than one can expect in an edited volume or a series of books, each written by a different person. The work of George and Smoke on deterrence suggests that there is an underlying theoretical structure that guides the cases studied therein.25 In that case, the specific design of the case study is rather loose and couched in a series of questions the

authors ask about the given case. Nevertheless, the ability to compare the cases of deterrence failure in this study is far more reliable than in the other examples cited herein. Comparative case study takes us into the realm of replication. For that reason we must bear in mind the principles guiding replicative research while designing such a project. Replication is not necessarily duplication. Rather it is a process where the principal characteristics of the original design are preserved, but in most cases, at least one characteristic is altered, so that the original experiment can be generalized under controlled conditions. To use our pneumonia drug example, if the original experiment was administered to patients in the 15-to-45-year-old age group, we may want to replicate the original experiment (using identical dosages, on subjects at the same stage of the disease, with a similar male-female mix) now employing subjects in the 46-to-70-year-old age group. The next replication may vary the dosage while keeping other characteristics constant, and so forth. This implies that unless we control some fundamental elements of the case study design, we may not be able to compare across cases. This brings us to some basic principles of comparative case design: 1. Case design must assume that all cases contain some fundamental similarities as well as fundamental differences. The purpose of a common design is to sort out, as much as possible, which elements of the process in each case match the hypothetical story that underlies the theory being tested, and which are unique to the case. 2. The most important principle of cross-case comparative design is transparency. Let the reader know the criteria by which the comparison is conducted and by which inference is derived in cross-case comparison. 3. Develop a cross-case comparative design. Specify what features of the process in each case are crucial for the ability to make inferences from the case with respect to the theory tested in the study. Identify at the outset which cross-case variabilities are not necessarily detrimental to your theory. (For example, in the case of deterrence theory, one may argue that a fairly accurate perception of the deterrent threat by the challenger is a pretty crucial element of the ability to test the effect of deterrence on the preservation or change of the status quo. On the other hand, the structure of the decision-making body in the challenger—small group, individual, or large cabinet—may be of less importance as long as all or most participants in that body knew that Page 470 → they were contemplating violation of a status quo against a deterrent threat.) 4. Define standards of inference from multiple cases in light of possible variability in results across cases. Note that a perfect fit between all cases and the theory arouses suspicion because history is seldom fully accountable by a single theory, no matter how good the theory. The typical situation is that some cases support the theory while others do not. More likely, the typical situation is that some cases fit a subset of the elements of a theory, while other cases fit another subset of the theory. For example, suppose that in our hypothetical study of deterrence, we have three cases that are fully consistent with a relationship between effective deterrence and preservation of the status quo. (Two were cases where deterrent threats were credible, were properly communicated, and were shown to have prevented the challenger from violating the status quo, and one was a case where the deterrent threat was not credible and the challenger ended up acting to change the status quo by force.) Suppose that two more cases were cases that provide rather clear evidence against the theory. (In both cases deterrent threats were credible and perceived as such by the challenger, but the latter nevertheless went ahead and launched an attack.) Finally, in three additional cases, evidence was mixed. (In one case the defender had the will and capacity to carry out the threat, but the challenger was not fully aware of the nature of the red line [or trigger] of the deterrent threat. In two other cases, the challenger's decision makers who had been aware of the threat and the consequences of violating the status quo were nonetheless forced into action by domestic public opinion or by the pressure of allies.) How are we to interpret deterrence theory in light of this kind of mixed evidence? Statistical tests have clear standards to determine whether a hypothesis is supported or refuted (e.g., the probability of Type I error). What are the standards for case study? Clearly, it is impossible to apply the rules of inference used in statistical analysis to case study research. No matter how many cases are compared, they probably do not make for a minimum number enabling statistical analysis-like inference.26 It is difficult to establish rules of thumb for such situations. Yet two principles are crucial. First, list the evidence—for, against, and mixed—explicitly in the comparison. Second, whatever the final verdict, clearly the qualifications need to

4. be stated explicitly and be consistent with the strength of evidence. 5. State the possibly confounding factors that your case studies revealed. If these factors reappear in different cases, they suggest possible omitted variable bias. Such bias requires redefinition or refinement of the theory. Principles of cross-case comparison are no less important than principles of case selection and case design, and more effort should be put into developing Page 471 → systematic processes of comparison in case study research than exists currently in the methodological work on qualitative research in general and case study research in particular.

Conclusion As noted, in recent years there have been some methodological improvements that are available to scholars who wish to use case study methodology to test theories. Unfortunately, these advances have gone largely unnoticed by most practitioners of this methodology in international studies. More often than not, readers of case studies get the impression that case selection and case design issues are presented as lip service to methodology and that crosscase comparison criteria are almost always left unspecified. This paper suggests some elements of a strategy that may help make case study research a more scientifically meaningful methodology than it has been in the past. By treating explicitly considerations of case selection, case design, and cross-case comparison, we may be able to capitalize on the many and important properties of case study research in the context of testing political theories systematically. Several general principles should guide the use of a given methodology for confirmatory purposes. Transparency For a case study to have any confirmatory value, the design must be transparent. Readers must understand what was studied in the case, what was included and what was excluded, how evidence was collected, and how it was processed. Finally, the considerations of the analysis and weighting the evidence in the process of making substantive inference must be explicit. Transparency serves several purposes. First, it allows detection of the underlying assumptions of the various decisions made in the research. Second, it allows replication. Third, it enables meaningful criticism. All these elements are crucial for scientific progress. Adequacy A methodology must fit the nature of the problem to be examined. This is less of a problem for case study research than for other methodologies because of the flexibility of the method, but it is important to understand that not all methods can be applied to all problems. On the other hand, a more common problem in science is that there are multiple methods that can adequately treat a certain problem. In such cases the researcher's choice of the method may be based on which one is the most effective in terms of the theory tested, in terms of the nature of the data, and in terms of practical considerations such as time, budget, and so forth. Page 472 → Limitations and Evolution of Methodologies There is no perfect methodology. In science we often think of experimental design as being the most “scientific” methodology because of the ability to achieve control in everything but the specific variable that is being manipulated. We tend to treat all other methodologies as somewhat inferior to an experimental design. It is important to note, however, two things about experimentation that are even truer about less “reliable” methodologies. First, experimentation is useful only if and when it is replicated. Replication establishes not only whether the findings in any given experiment hold up when repeated. Rather, replication allows identification of the limits of generalizability of a given set of experiments. That is, replication allows researchers to establish up to what level of variability in experimental conditions the effect of the independent variable holds up. In contrast to a

growing tendency to replicate findings in quantitative research on international problems, case study research shows little or no effort to replicate findings and approaches of previous research. Typically, the variability across sets of case studies testing given theories is so large that any comparison is meaningless. Standardization of case study research using more systematic and transparent criteria of case selection, case design, and cross-case comparison may move us a great distance toward replication in case-based research. Second, the experimental method itself is dynamic and constantly undergoing improvement. If we were to compare methodology textbooks in social psychology of the 1990s to their predecessors of the 1970s and 1960s, we would observe major changes in the methodology of experimental design. The same applies to the evolution of quantitative methodologies—even in the context of international relations research.27 The same applies to case study methodology. There are important innovations in case study research. Some of these innovations take place in other substantive domains that use case studies. It is important to be aware of them and see whether they could be incorporated into our studies. Openness to Multimethod Research Each methodology has some liabilities and weaknesses. Also, generalizing about a theory or making theoretical inferences requires testing for the theory's robustness. We have noted the importance of replication in establishing robustness of a given theory. However, it is important to note that robustness can and should be tested across methods. If inferences are based on a given methodology, then the theory's propositions should be more suspect than if those inferences hold across studies using different methodologies. For that reason, we must be open to using multiple and varied methodologies to test our theories. The significant advances in international studies over the years reflect constantly growing theoretical, substantive, and methodological pluralism in the field. In the process, we often make mistakes. Science learns both Page 473 → from the mistakes and from the accomplishments of previous research. This was an effort to reflect upon the use of case studies in international relations in order to help improve future research.

Notes 1. Alexander Wendt, Social Theory of International Politics (Cambridge: Cambridge University Press, 1999). 2. The notion that single events cannot be used to confirm or refute general theoretical arguments has long been rejected. See Gary King, Robert O. Keohane, and Sidney Verba, Designing Social Research: Scientific Inference in Qualitative Research (Princeton, N.J.: Princeton University Press, 1994), 11–12. 3. The present paper does not attempt to replace the extensive literature on case studies in general or in international studies in particular. There are several good texts on the subject. (King, Keohane, and Verba, Designing Social Research, is by far the best book on the subject.) Rather, it attempts to highlight some issues in case study research that have received relatively little attention in the methodological literature and almost no attention in actual research using case studies. 4. Alexander L. George and Richard Smoke, Deterrence in American Foreign Policy: Theory and Practice (New York: Columbia University Press, 1974); Robert Jervis, Richard Ned Lebow, and Janice Gross Stein, Psychology and Deterrence (Baltimore: Johns Hopkins University Press, 1985). 5. Miriam Fendius-Ellman, ed., Paths to Peace: Is Democracy the Answer? (Cambridge: MIT Press, 1997). 6. Arend Lijphart, “Comparative Politics and the Comparative Method,” Comparative Politics 65, no. 3 (1971): 682–98; Harry S. Eckstein, “Case Study and Theory in Political Science,” in Handbook of Political Science, vol. 1, Political Science: Scope and Theory, ed. Fred Greenstein and Nelson W. Polsby (Menlo Park, Calif.: Addison-Wesley, 1975); Alexander George, “Case Studies and Theory Development,” in Diplomacy, ed. Paul G. Lauren (New York: Free Press, 1979); Alexander George and Timothy J. McKeown, “Case Studies and Theories of Organizational Decision Making,” Advances in Information Processing in Organizations 2, no. 1 (1985): 21–58; King, Keohane, and Verba, Designing Social Research, 43–46; Stephen Van Evera, Guide to Methodology of Students of Political Science (Ithaca, N.Y.:

Cornell University Press, 1997); Andrew Bennett, “Causal Inference in Case Studies” (paper presented at the annual meeting of the American Political Science Association, Atlanta, 1999). 7. For example, John W. Creswell, Qualitative Inquiry and Research Design: Choosing among Five Traditions (Beverly Hills, Calif.: Sage, 1998); Jacques Hamel, Case Study Methods (Thousand Oaks, Calif.: Sage, 1993); Robert Stake, The Art of Case Study Research (Thousand Oaks, Calif.: Sage, 1995). 8. Charles Lave and James G. March, Introduction to Models in Social Science (New York: Wiley, 1974). 9. Richard N. Lebow and Janice Gross Stein, How We All Lost the Cold War (Princeton, N.J.: Princeton University Press, 1993), 4–5. Italics added. 10. Ibid., 7. Italics added. 11. See for example the interesting study of diagnosis of breast cancer in David M. Eddy, “Probabilistic Reasoning in Clinical Medicine,” in Judgment under Uncertainty: Heuristics and Biases, ed. Daniel Kahneman, Paul Slovic, and Amos Tversky (Cambridge: Cambridge University Press, 1982). 12. Fendius-Ellman, Paths to Peace, 48. 13. For example, for probability-based sensitivity analysis in case studies see the following: Janice Stein and Raymond Tanter, Rational Decision Making: Israel's Security Choices, 1967 (Columbus: Ohio State University Press, 1980); Zeev Maoz, “The Decision to Raid Entebbe,” Journal of Conflict Resolution 25, no. 4 (1981): 677–707. The application of a similar logic with respect to the Page 474 → relative effects of certain value dimensions is given in Astorino-Courtois. Astorino-Courtois conducts a value-based sensitivity analysis by sequentially removing dimensions of values that were present in a decision process to examine their effect on the stability of actual decisions. Beyond the usefulness of the technique in testing for effects of a given threat on decisions, it is a nice approach for dealing with counterfactuals in foreign policy research. Allison Astorino-Courtois, “Clarifying Decisions: Assessing the Impact of Decision Structures on Foreign Policy Choices during the 1970 Jordanian Civil War,” International Studies Quarterly 42, no. 4 (1998): 733–54. 14. The significance of going to extra lengths to “prove” deterrence success (or failure) is due to the risk of committing a “Type I” error, that is, of rejecting the null hypothesis when it is valid. This is considered a more severe violation than committing a “Type II” error (accepting the null hypothesis when it is false). On this issue as it relates to case study research, see a later part of this section. 15. In many cases both types of evidence may be present. For example, if we were to eliminate the domestic value dimension or the “deterrence” from the challenger's decision calculus, the decision may have changed. In such cases, the evidence for deterrence does not outweigh the evidence for alternative explanations. This suggests the need to develop more discriminatory tests. 16. Christopher Layne, “Kant or Cant: The Myth of the Democratic Peace,” International Security 19, no. 2 (1994): 87–125. 17. Christopher Layne, “Lord Palmerston and the Triumph of Realism,” in Fendius-Ellman, Paths to Peace, 61–100. 18. Stephen R. Rock, “Anglo-U.S. Relations, 1845–1930,” in Fendius-Ellman, Paths to Peace, 101–50. 19. I say in their view because most of the hypotheses attributed to the normative or structural explanations of the democratic peace propositions are in my view clearly unrelated to these explanations as outlined by democratic peace scholars. See Zeev Maoz and Bruce Russett, “Normative and Structural Causes of Democratic Peace,” American Political Science Review 87, no. 3 (1993): 624–38; Bruce Russett, Grasping the Democratic Peace: Principles for a Postwar World (Princeton, N.J.: Princeton University Press, 1993); and Zeev Maoz, “Realist and Cultural Critiques of the Democratic Peace,” International Interactions 24, no. 1 (1998): 3–89, for discussion of these explanations and their misrepresentations by critics of the democratic peace propositions. 20. For a more detailed analysis of this crisis both in the French-Prussian and French-Dutch context, see the case summaries in the Web site that accompanies the book by Zeev Maoz and Ben D. Mor: Bound by Struggle: The Strategic Evolution of Enduring International Rivalries (Ann Arbor: University of Michigan Press, 2002). The Web site is located at http://spirit.tau.ac.il/poli/faculty/maoz/iha.html. 21. Zeev Maoz, “The Renewed Controversy over the Democratic Peace,” International Security 23, no. 1 (1997): 193. 22. Of course, a risk entailed in such logic is that people would attempt to fit a specific case to a standard “structure,” thereby downplaying or ignoring its unique features. This touches upon a long-standing debate

in the literature on individual or social inference between those who claim that people typically use base rates to discount individual indicators and those who argue that people usually overestimate the significance of individual indicators at the expense of base rates. On this kind of debate as it applies to foreign policy decision making, see Zeev Maoz, National Choices and International Processes (Cambridge: Cambridge University Press, 1990), 201–2. 23. See Jervis, Lebow, and Stein, Psychology and Deterrence; Fendius-Ellman, Paths to Peace. 24. A similar problem characterizes the case studies conducted on crisis behavior by the International Crisis Behavior (ICB) Project, although the theoretical framework there is more developed than in the examples of deterrence studies or democratic peace studies. This is even more regrettable because most of the Page 475 → ICB case histories are full-volume studies. This affords the authors the opportunity to both recount in detail the specific events in the case as well as provide a more analytic account of the crisis behavior process in terms of the project's framework. The summary of the case study results provides important insights into crisis behavior, but these are questionable in places due to differences in specific case designs. (See Michael Brecher and Jonathan Wilkenfeld, A Study of Crisis [Ann Arbor: University of Michigan Press, 1997]). In many cases, the authors of the individual case studies upon which this project is based do not offer any noticeable design. Some of the full-volume studies of the ICB Project include Michael Brecher with Benjamin Geist, Decisions in Crisis: Israel in 1967 and 1973 (Berkeley: University of California Press, 1980), Alan Dowty, Middle East Crisis (Berkeley: University of California Press, 1984); Adeeb Dawisha, Syria and the Lebanese Crisis (New York: St. Martin's Press, 1980); and Avi Shlaim, The U.S. and the Berlin Blockade 1948–1949 (Berkeley: University of California Press, 1983). Despite these problems, it must be stated that Brecher and Wilkenfeld's study is actually an outstanding and unique example of carefully conducted cross-case comparison in international studies research. 25. George and Smoke, Deterrence in American Foreign Policy. 26. King, Keohane, and Verba, Designing Social Research, 213–17. King, Keohane, and Verba assert that the minimal number of cases to make any kind of meaningful inferences in qualitative research is probably more than five, but there is no need for a comparative case design that exceeds twenty cases. Ibid., 216. 27. One of the interesting side effects of the democratic peace research program is that it invoked a lively methodological debate among quantitative scholars and a number of major innovations in testing issues related to democracy and peace and more general issues concerning international conflict. See, for example, Nathaniel Beck, J. N. Katz, and Richard M. Tucker, “Taking Time Seriously,” American Journal of Political Science 42, no. 4 (1998): 1260–88; and Bruce Russett and John Oneal, Triangulating Peace (New York: Basic Books, 2000).

Page 476 → Page 477 →

FOREIGN POLICY ANALYSIS

Page 478 → Page 479 →

FOREIGN POLICY ANALYSIS Steady Progress and a Half-Empty Glass Yaacov Y. I. Vertzberger

The Domain of Foreign Policy Analysis Foreign policy analysis (FPA), broadly defined, is a field of study that describes and investigates the structures, processes, and outcomes of the purposeful policy initiatives and responses that are conceived by sovereign political entities and directed toward other political units (not necessarily sovereign states) beyond their borders. The boundaries of the FPA domain are porous and cannot be strictly drawn, and thus it is not uncommon for some of the literature to belong in more than one field of study (e.g., security studies and FPA). Furthermore, FPA as a field of study emerged over the last four decades unbound by a single paradigm, containing multiple and diverse approaches, and should not be associated with (as some do) and conceived to be narrowly wedded to one or another variant of statecentric realism. In a nutshell, it is claimed here that while FPA has been a lot more vibrant than some observers are inclined to believe,1 it has also delivered substantially less than one would and could have hoped for; hence the proverbial half-empty glass metaphor. The discussion in this paper does not intend to dwell on and review all aspects of FPA2 but to address critically the most salient weaknesses in the current state of FPA and their implications for the agenda, methods, and priorities of future directions that FPA studies should take. A panoramic reflective view of the FPA landscape since the 1950s leaves an observer with mixed feelings. There has obviously been steady progress that is expressed in extensive theoretical and conceptual development, in Page 480 → the growing empirical richness and availability of both case studies and aggregate data analysis, and in increasing methodological diversity, rigor, and sophistication. Yet, at the same time, there are reasons to believe that there is in FPA a lot less than what meets the eye and that in spite of progress we are still facing a half-empty glass. What is missing in the current state of our knowledge and understanding may occasionally outweigh, or at least equal, the weight of what we claim to know. It cannot even be taken for granted that faster and more dramatic improvement and progress in the state of the art are around the corner. Scientists, especially social scientists, are after all not only socialized by the norms of their trade to open-mindedness, curiosity, and changes; they are also products of their societies and are exposed and socialized to the norms of stability, risk aversion, and, in general, a preference for continuity over change even in the face of negative feedback. As such, they are driven by both cognitive and motivational biases and fallacies that too often have resulted in little, or skewed, learning; they tend to follow the dominant intellectual fashions of the time less critically than they should, and in the process overestimate the value and contributions of these fashions to actual progress in their respective fields. The following discussion identifies and elaborates on how and why FPA progress has been less than adequate. In brief, it is claimed that the causes for this state of affairs include overlearning by imitation, decontextualization, oversimplification, and biases toward precision, conservatism, and risk aversion. It is argued that these have seriously straitjacketed progress in the field of FPA. Two qualifying comments on this assessment of the state of FPA research should be made at this point. One is that an evaluation is not independent of the observer's initial expectations. Against lower expectations current achievements may look more impressive than against higher initial expectations.3 Second, some of the more general criticisms of FPA that follow are not unique to this research domain but apply to other fields as well; they represent a general malaise in the study of international politics and more broadly the political science discipline.4 The following sections set out the nature of progress and the multiple forms it could take, on the one hand, and discuss in detail areas of inadequacy of progress and missed opportunities in FPA research, on the other hand.

The Nature of Progress: Unpacking and Repacking the Research Enterprise A meaningful discussion of progress in a research domain requires that it should be ontologically unpacked and its

constituent elements individually examined, and then repacked for an overall progress assessment. In essence, a research domain consists of three constituent components, or building blocks: theory, which requires systematic conceptualizing and generalizing toward theory building; empirical data, which result from comprehensive search, collection, and organization of relevant data; and methodology, which requires applying a set of transparent and explicit rule-based methods Page 481 → and techniques for the transformation of data into theory-relevant information and then relating the transformed empirical data to the theory for hypotheses-testing and theoryvalidation. The premise underlying my argument is that progress does not necessarily occur simultaneously and to the same extent across all of the three constituent building blocks of the research enterprise, nor can progress be synchronized to achieve that effect. Hence, identifying where and to what extent the most progress took place, whether in theory, empirical data, or methodology, is important for a holistic assessment of progress. It is also important to recognize the trade-offs between progress in one constituent block compared with another and the cost-effectiveness of progress. I would argue that in the social sciences in general, and international politics research in particular, the most salient condition for substantial progress is advances in theorizing; the second most salient is improvement in data availability and quality; and the least prominent factors for progress are methodological innovations, which are also the most vulnerable to diminishing returns. Implied in this argument is the notion that progress in each of these constituent elements of research can have overall regressive effects, when more is less, for reasons that will be further explained. The logic of my argument for a three-tiered, ordinally-layered salience of improvement in knowledge is simple and self-explanatory. Knowledge and its qualitative improvement is first and foremost the product of ideas and concepts. In the absence of ideas there is no coherently meaningful and generalizable knowledge across time and space. Furthermore, since in the social domain information is often not readily available, theory plays an important role in going beyond the information given. Although data are indispensable stimulants of important research questions and puzzles, in the absence of theory data produce merely descriptions and at best ad hoc explanations. Whether theory is produced inductively or deductively is of little importance in this context and is mostly a matter of researchers' preferences. Advances in methodology improve the logical quality and validity of theory, allow for the efficient transformation of data into information, and verify the validity of the knowledge that results from the coupling of theory with data (history). But in and by itself methodology does not produce knowledge; it is merely an instrument for scientific progress and not an integral part of it. To put it concisely, in the social sciences important scientific innovations are more likely to be driven by the logic of discovery than by the logic of verification. Herbert Simon is correct in criticizing the preoccupation with the logic of verification in the social sciences,5 and he traces this bias to the practices of the philosophy of science, on which social science epistemology heavily depends. “The history and philosophy of science have been excessively formulated with the drama of competition between theories.… Researchers in psychology and some of the other behavioral sciences, where there is often a severe deficit of good candidate theories, have sometimes been misled by the literature on philosophy and methodology to exhibit in research and publication an exaggerated concern with verification, accompanied by uncritical Page 482 → application of statistical tests of hypotheses to situations where they don't properly apply.”6 Whether a scientific investigation is discovery-driven or verification-driven will have decisive ramifications for the research questions that should and can be asked—and which variables can be applied and will be salient to answer these questions. Taking into account the relative limitations of alternative methodologies across variables and data sets, a verification-driven approach to research seriously limits the type of explanation that can be offered, while a discovery-driven approach has a much more open and wide horizon for the range of what are considered testable and acceptable explanations. This makes progress more likely and more broadly based. In other words, preference for one orientation—discovery or verification—over the other will have substantial agenda-setting implications for research.7 Looking now at the substance of progress, it can heuristically be categorized as either horizontal or vertical. Horizontal progress expands the amount of knowledge and is largely incremental, but not exclusively so. Vertical progress represents a qualitative change at a deeper level and is more likely to be transformational. The crossover line between horizontal and vertical progress is not always sharp and clear, nor is it obvious when a quantitative

increase in available knowledge becomes a base for qualitative change in the state of knowledge. More specifically progress may take one or more of several forms. Expansion represents a quantitative increase in knowledge on particular subjects. Verification/falsification produces meta-knowledge that tells us how valid our knowledge is about an issue. Integration implies combining segments of knowledge that were originally produced independently into a whole, where the whole is more than the sum of its parts. Integration can take place within a single issue area and level of analysis boundary or across issue areas and levels of analysis. Contextualization goes beyond mere integration by providing contingent knowledge. Unlike unconditional propositional statements that tell us something that is supposedly unconditionally true about the social environment, contextual propositions tell us what is true (or false) on a conditional base, and what will be true should these conditions change. Extrapolation reflects going one step further in the integration and contextualization processes by extending knowledge beyond the porous boundaries of an issue area or a discipline and integrating it with knowledge from other disciplines to create interdisciplinary theories (e.g., important progress in this direction was made in theories of cognition, theories of markets, and theories of institutions). These five forms of progress are not necessarily mutually exclusive and may take place serially or simultaneously. As more of these forms of progress are experienced in any particular research field, it can be said that progress is more extensive. It is important to note that this ordinal scale of progress is not meant as an ordinal evaluation of the research produced by individual scholars (micro-behavior) but is intended to provide a long-term perspective and evaluation of the state of a discipline or subdiscipline. In other words, I am talking Page 483 → about the macroconsequences of microbehavior. This is then not a recommendation to individual scholars concerning what line of research they should pursue. In fact, only by allowing individual scholars to make the most efficient individual choice as to how best to apply their interest and talent is the aggregate research product likely to be maximized and progressive. It results from a self-regulatory mechanism that drives aggregation and complementarity of individual, narrowly focused research efforts to the end consequence of disciplinary progress. How and why this happens is well beyond the subject of this paper, and we shall therefore black box this mechanism without explaining it. To illustrate my point consider the contributions of Margaret Hermann on the role of personality in foreign policy behavior, those of Morton Halperin to the understanding of bureaucratic politics, the work of Charles Hermann and his associates on international crisis behavior, or the research of Stephen Walker and his associates on rolerelated foreign policy behavior.8 Each of these authors has been involved in expanding knowledge in a relatively narrow issue area. But it is unimaginable to have an integrative view of foreign policy behavior in crisis without integrating their building blocks into a comprehensive perspective that would take into account the cumulative pushes and pulls in coping with a foreign policy challenge, of personality attributes' effect on preferences, roleframed objectives, and bureaucratic structural constraints within which the former two motives have to be played out. Whether this level of integrative contextualized progress of individually produced knowledge and theory will actually take place depends on the incentives offered and impediments posed, such as the nature of the research agenda and the norms dominating the profession, as well as the effects that common motivational and cognitive biases have on scholars. As will be argued throughout this paper, FPA has not achieved the extent of progress that it should and could attain because the incentives for progress were not vigorous enough to cope with and overcome impediments to progress that are specified in the following sections. Consequently, innovative individual-level research was not regularly translated and integrated to result in the anticipated broad, progressive macroconsequences in FPA. It can be argued that most FPA research advances focused on the first two types of progress, expansion and verification. We obviously know now, by far, more and with a greater degree of confidence about an increasing variety of subjects and issues concerning international politics than we knew four decades ago. But much of this knowledge lacks coherence and can metaphorically be described as an archipelago consisting of islands of knowledge, between which travel and communication tend to be irregular and nonsystematic. On the other hand, the types of progress that are conspicuously less visible are integration, and even more so contextualization and extrapolation. By focusing on the parts, and so often missing a holistic vision, we often misread the salience of specific instances of progress by exaggerating it.9 In the absence of integrated contingent knowledge, small

advances in specific, sometimes even esoteric issues loom much larger than they would had they been perceived within a broad holistic background—against Page 484 → which the actual value of small advances is more correctly assessed.

The Emperor's III-Fitting Clothes and the Quality of Progress In the famous story about the emperor's new clothes, or rather lack of them, only one little boy is willing to acknowledge that the “emperor is naked.” Forty years on, the emperor, older and possibly wiser, is now fully dressed but has acquired a taste for borrowing and wearing his older cousins' and the court's fashion setters' (one an economist, the other a mathematician) ill-fitting clothes, which have not always been tailored to his size and needs. This small variation on the well-known story demonstrates one of the major quandaries that haunts FPA in proceeding toward better methods of verification: wholesale adoption of norms and standards of “good scientific practices” that were borrowed from other science disciplines. This results in relying on other sciences to serve as role models without investing enough thought in the relevance of these norms to the particular circumstances (problem types, available data types, nature of variables) found in international politics. To be sure, there have been some adjustments, but on the whole these imported standards of good scientific practices have acquired a difficult-to-challenge status and occasionally continue to obstruct necessary progress. Important reasons for the fallacies that will be identified in this essay can be traced to the stranglehold that these norms have had over the research culture in FPA (as well as over other domains of research in international politics). What is particularly difficult when dealing with these deleterious influences over knowledge development is that each of these norms makes sense on its own. They become an obstruction when they are taken too literally and to the extreme. The simultaneous application of several such norms results in a synergy of straitjacketing. Obviously a dose of these norms is a necessary condition for validity, but an overdose of the same measures is dangerous. The crux of this problem is that it is not so easy to determine how much is enough. In the absence of clear standards of sufficiency, there is a tendency to go to the extreme by applying a well-known fallacious heuristic: “If some is good a lot is much better.” What are these norms? I refer mostly to the norms of decontextualization and parsimony, precision, and the network of methods and rules that are derived from them. These have become part and parcel of the quest for “scientificism,” a term that I use to describe the exaggerated concern with how scientific our research enterprise is along with defining “scientific” in narrow, strict terms that are applied inflexibly and without distinction. Initially these norms and their derivations represented necessary conditions for the emergence of good science in the physical and mathematical sciences. From there these methods penetrated and then came to dominate the discipline of economics. Consequently the success associated with their application made them vivid and cognitively available to be considered for emulation during the formative years of FPA, when international politics became sensitive to the need to transform itself into a more scientific field Page 485 → of research. Once adopted, they acquired a life of their own and became difficult to unlearn. What we are facing is what social psychologists will recognize as an accumulation of a well-known set of cognitive biases.10 The attention to and choice of norms of scientificity was driven by the salience and vividness of readily available examples of success (the availability bias). These have come to represent, and to be associated with, what is exclusively considered properly executed research (representativeness bias). Although it made sense at the time to use such norms as a starting and anchoring point, their effectiveness should have been reassessed with the passage of time and accumulation of experience. Yet the opposite has happened. As is common, the anchoring point acquired an immovable stability (the anchoring bias) that has stood in the way of reassessment and adjustment as generations of scholars have become socialized into this set of beliefs about good science. In fact, one could argue that these norms have evolved into institutions—that is, “shared concepts used by humans in repetitive situations organized by rules, norms and strategies”11—and have acquired the same rigidity and resistance to change that is so often associated with other institutions. In the following sections we explore and explicate how these general broad fallacies are expressed in specific inadequacies from the setting of the research agenda and research strategies to soft spots of methodological progress.

Anchored in the Past: The Costs of a Conservative Attention Span

A major weakness of FPA research has been excessive conservatism. It tends to remain focused on the traditional fields of diplomacy and security, where FPA has its roots. Conspicuously absent are a number of more recent but central research domains to which FPA could have and should have made a contribution. Most salient among these is the field of international political economy (IPE). With the increase of interdependence and globalization and the consequential importance of economic diplomacy, economic conflict, and economic cooperation, the application and expansion of theories and methods that have been the backbone of FPA research in the security field to the analysis of state foreign economic behavior could be logically expected. This, however, has not been the case. IPE remains the domain of largely structural explanations where the actors are states, transnational actors, organizations, and institutions. Surprisingly, it seems as if people—individuals or groups—do not play any role in economic decisions and processes, at least not a consequential one. To be sure the literature of IPE made a mellow but brief overture to incorporate the role of ideas and beliefs that should have provided a window of opportunity for a creative dialogue between these two fields. But the opportunity was never seized and faded away to the detriment of both FPA and IPE research. When reading in the IPE literature references to expectations, interests, preferences, beliefs, and change thereof, one keeps wondering who carries these cognitions. Do organizations have cognitions? Do institutions have beliefs? Where do they come from? Are they always exogenous or could they be endogenous? But Page 486 → these issues have never truly been a subject of concern, even when such issues were occasionally raised12 only to be comfortably avoided and relegated again to blissful inattention. A related but separate area of neglect is what has become an influential research area in the social sciences, that of institutionalism/neoinstitutionalism.13 FPA could have contributed important insights to the literature by modifying the dominance of rational choice premises that underlay much of this literature.14 FPA has much to say on issues that are at the center of the discussion on institutions, such as the consequences of limited access to information and constraints on optimal information processing. Similarly FPA can make a salient contribution to the analysis of transaction costs, especially on the costs of sharing information among multiple actors and coordinating the interests, objectives, and behaviors of actors in mixed games with a view toward reducing transaction costs. No less important can be unveiling the linkage between institutional failures and organizational, group, and individual frailties.15 A third issue barely addressed in spite of its increasing importance with the spread of democracy is the role of accountability and sociopolitical policy acceptability, as a concern shaping both the process and outcome of foreign policy decision making and decision implementation. Fourth and no less puzzling is that in spite of the closeness of areas of interest, FPA and public policy have largely kept their distance from each other with few exceptions.16 The literatures of both FPA and public policy reveal very little dialogue between these two research areas that could have fruitfully cross-fertilized each other's insight into the domestic and international policy environments. Fifth, the combined effect of conservatism and complexity-aversion attitudes are at least in part responsible for the fact that FPA still lacks a theoretical framework that will allow us to move from the analysis of foreign policy as a two-level game (domestic and international) to a three-level game (domestic, international, supranational) and will account for the increasingly salient impacts of regionalization and globalization. More specifically, the emergence of the European Union (EU) as a major actor requires a theory of foreign policy analysis that looks at the threeway influences and interactions between the state's domestic arena, its international policymaking processes, and the role that the inputs of and interactions of the suprastate (the EU) institutions with individual national states have in shaping policy preferences, processes, and outcomes, as a consequence of the creeping devolution of authority from individual states to the EU.17 Sixth, systematic prescriptive efforts were relegated to a minor role and are often viewed with snobbish disdain in FPA in spite of the fact that one would expect foreign policy analysis to be deeply concerned with putting foreign policy to work effectively.18 Questions of policy relevance continue to play a relatively marginal role in the big debates within FPA. Only in this impoverished policy-relevance context can it be understood why important issues of policy relevance—such as debiasing or the quality of governance, and even the seeming intuitively important criteria and assessments Page 487 → of when policymakers' judgments and the resulting policies are a

success and when they are a failure19—have had a minimal role in theory construction. But FPA has not only been marred by avoiding, neglecting, and actively ignoring important research subjects and being biased toward attention to the most traditional issue areas in international politics. There is also a host of problematic epistemological issues concerning how subjects and issue areas that were successful in attracting attention and were subject to extensive scholarly discussion have been approached and treated. This is elaborated in the next section.

Biases in Research Design and Strategies Perhaps one of the most apparent epistemological deficiencies concerning the manner in which research has been conducted is what amounts for all practical purposes to decontextualization. It can take more than one form as will be clarified subsequently. But the crux of it is that research in foreign policy analysis, including foreign policy decision making, has tended to shy away from including context as integral to the research design and at best has incorporated only a very thin context. Serious attempts to incorporate systematically contextual variables in research design are sparse, because thick contextualization runs the risk of being treated as a violation of the tenets of what is considered “good, valid research”—parsimony. Theories and models are supposed to be generalizable across situations and to explain all or most contingencies, with as few variables as possible, thus being in fact too often context free. This is justified in the name of what has become one of the ultimate criteria for good hardhitting research—that is, parsimony, a criterion borrowed from the practices of the hard sciences. This criterion in essence states that (1) explanations should be simplified and/or (2) the process of explanation should be economized. There is of course nothing wrong with these principles, except when they are taken too literally and applied with indistinguishing zeal, lacking in judgment and common sense, which should temper and regularize the application of the parsimony principle.20 It is ironic of course that positivists, who were the strongest proponents of parsimony, neglected to apply their own standards of validity and failed to note that “again and again in the history of science what seemed to be hopelessly complex, before someone had a good theoretical idea, has turned out to be beautiful in its simplicity after that idea—and it is a necessary act of scientific humility to remember that again and again those ideas turned out to be wrong.”21 The tendency toward decontextualization has, in my view, produced some important deleterious effects and biases in two central theoretical foci of FPA: the nature of the debate on rationality and the treatment of bounded rationality. The debate on rationality is one of the more persistent unresolved conflicts of view in FPA.22 The disagreements have come to be framed in polarized terms that pit a view of human behavior as driven by rationally calculated motivations that conforms to the expected utility model against Page 488 → a view that behavior, even if intended to be rational, is driven by a range of cognitive and motivational constraints and biases that make rational choice impractical and an incorrect descriptive proposition. This debate is no longer productive as it is framed now. In fact, reality is much more complex than reflected by the debate. The same individuals, groups, and organizations may, within a single process of decision making, demonstrate different levels of (non)rationality, and this is even more true across time, issues, and separate decision opportunities. Practically no single individual, group, or organization is systematically and without fail rational or nonrational to the same extent all the time. It is, therefore, much more productive to perceive rationality in terms of a spectrum stretching between two poles, one of pure rationality and the other pure irrationality. In between there is a broad space that represents varying extents of nonrationality or bounded rationality (see figure 1). The debate should thus focus on where and under what circumstances actors are likely to be closer to one pole or the other. Yet rephrasing the research questions in those terms implies contingency and complexity, which are a lot less convenient to deal with than a more simple, polarized, either-or type choice between rationality and bounded rationality.23 Another facet of this issue is the static form that bounded rationality research has acquired. Since Simon coined the term, its conceptualization has not evolved significantly.24 Its simplicity seems to offer a riposte to the rivaling simple premises of rationality. Yet this has become an impediment to the development of a contingent theory of

bounded rationality, which will recognize that bounded rationality is neither a unitary concept nor a unidimensional phenomenon. Bounded rationality represents a spectrum of decision-related behaviors that can take more than one form. Therefore, to take full advantage of this useful and productive concept, what was and is still needed is a typological specification of the definition of boundedness that will reflect the various ways and different extents of boundedness and the potential consequences from each boundedness type. I would argue that here too the absence of such a typology has resulted from complexity-avoidance. Bounded rationality was structured to be as simple and unambiguous as rationality. That was done irrespective of and without taking into account that the two were different “beasts” by definition. Rationality was a relatively clearly delimited and closed behavior type, while bounded rationality is an open-ended definition of behavior that can contain multiple forms of behavior all fitting the boundedness condition. Hence, rationality requires mostly a process of calculation, while bounded rationality largely involves a process of judgment. Page 489 → That aside, a number of other cognitive and motivational biases are built into the parsimony imperative that result from cognitive constraints and shortcomings, to which scholars are as prone as other human beings. These biases have made context-free theorizing a tempting prospect, have become a difficult-to-shake-off habit, and have also taxed creativity and the process of scientific diversity. Limitations on Memory, or Scholars as Intellectual Misers It is a well-recognized fact that the human brain can handle only a limited information load, as so clearly indicated by the seminal study on “the magical number seven plus or minus two.”25 This is, of course, a powerful incentive for limiting the complexity of theorizing and explanations. By avoiding the complexity associated with context we have made a tacit decision that rather than expand our mental capabilities, we would narrow our insight and understanding. For all practical purposes we have turned our natural cognitive limits into a normative scientific merit. It is no wonder we are so obsessed with a search for a “do-it-all” paradigm that has few moving parts, and hence we have become attached to paradigms such as neorealism, rational choice theory, and their derivations that respond well to our biocognitive limitations. It should be noted that these paradigms have fewer than seven assumptions and seem to fit perfectly with available memory power. The Aggregation and Sample Bias Most methodologies that have a quantitative logic depend on sampling. The nature of information and its availability in many domains of foreign policy behavior does not provide readily available, sufficiently large samples of items or even allow for their artificial creation in an experimental context. Yet to be able to generalize we depend on a large enough sample size that (a) can be treated by applying one or more statistical methods and (b) has external validity. By allowing for context we might reduce the uniformity of sample populations and increase their diversity; consequently the sample has to be broken down into subsamples that may include such a small number of items that statistical treatment becomes irrelevant or impossible. Thus, incorporating context threatens in many cases to affect adversely the employment of quantitative methods that have come to play such an important role in theory building and especially in verification. Multiple Causations and Causation Loops The larger the number of relevant influences (variables) on the dependent variable, the more difficult and complex becomes the process of structuring the causal relationships among variables, and the more demanding is the tracing and mapping of causation directions. The larger the number of variables, the greater the chance for causation loops, which pose very serious Page 490 → problems of internal validity that act as impediments for the application of quantitative or quasi-quantitative research methods. The Actor's Motive Bias By eliminating context we are implicitly assuming that human behavior can be reduced to a relatively small

number of standardized actions and reactions. In a way we apply to individual behavior a limited inventory of standard operating procedures that we assume to remain unaffected by differing contingencies. But what might be valid for organizational behavior is not necessarily correct about individuals. True, people prefer to be cognitive misers as long as is possible. But people are not, at least not all of them, inherently stupid to the point of systematically defying their self-interest in valid knowledge although veridicality is not their only motive.26 They learn—sometimes sooner, sometimes later—that contingencies matter, that responses to problems that have some basic similarities but are also quite different have to be contingency-tailored, and that mindless, repetitive behavior may be ineffective and costly. In cases where accountability is important, it could serve as a powerful incentive for reflective rather than reflexive coping and response.27 In other words, eventually people learn to become more context-sensitive. Therefore, theories about human behavior that are not context-sensitive miss an important, even critical, aspect of human motivation and behavior. Slow Integration The preceding points are some of the reasons why a critical task of theory-building—the process of integrating the multiple and diverse theoretical strands and pieces of FPA hypotheses and theories into a coherent whole that will cut across levels of analysis—has been subjected to the consequences of complexity aversion. One can compare the state of FPA research to a set of LEGO pieces that becomes really meaningful, although each piece has a separate existence, only when the pieces are assembled and by the way that they are put together. Each LEGO piece can be integrated in more than one way. Only by assembling them so that all the parts fit together to create a whole, meaningful structure is the task accomplished. This principle is even more the case with the building of knowledge structures. Instead, FPA reveals a clear preference, which is embedded in the incentive structure, for research that is narrowly, or even very narrowly, focused on a single or small number of hypotheses and variables that explain a fraction of a broader phenomenon. This is not to say that there have not been studies that emphasized the grand picture in taking a multilevel integrative approach to important theoretical and empirical questions of international politics.28 But these are not the rule, and integrative studies were more often than not criticized for being nonparsimonious. These criticisms encouraged continued “fractionization,” which remains the dominant feature of research. Page 491 → Paradoxically, almost nobody denies the importance of integration, but at the same time there is a deeply ingrained reluctance among many to recognize and face trade-offs between theoretical integration and powerful prevailing methodological norms, and make the necessary choice between these norms and holistic theorybuilding by focusing on what is really important—the advancement of theory-building. This reluctance is not exceptional but part of a long-identified common bias: “Many people believe that their core values do not conflict and they need to be prodded, sometimes poked pretty hard into acknowledging that these values do conflict.”29 It is accompanied by an unstated assumption that somehow, someday these separate theoretical segments will come together and constitute a well-integrated theory that will be elegant, insightful, and of course parsimonious. How this is going to come about is rarely explained and specified.

Methodological Progress: How Real? How Much? There is little doubt that any systematic review of the literature over time will reveal steady but major improvements and increased diversity in the methodologies used. These are reflected in better understanding of the conditions necessary for validity, in qualitative research as well as in quantitative research, and in the increased sophistication in methods of data processing and hypothesis-testing that include advanced statistical methods, formal modeling, dynamic game theory, experimental methods, and a plethora of other even more exotic methods. Those methodological advances have so heavily penetrated the way research is conducted and also our classrooms that they are considered indispensable to the training of younger generations of researchers. Most of these research tools have their origins in other disciplines where often the type of data available and data requirements are different from those of foreign policy-related data. Coming straight off the shelf of other research arenas, their fit with foreign policy data formats is not necessarily perfect. The use of these methods imposes

premises that do not sit well with the realities of international politics (e.g., the nature of actors, their interests, their preferences, their motives, etc.). Consequently what is in many ways a truly admirable pace and scope of methodological progress has not come without leaving behind a trail of queries that have been either suppressed or inadequately answered. Rather than discuss particular methods in detail, I shall raise here three general conceptual issues concerning the quality and value of progress represented by methodological advancement in FPA. These apply to the valuation of numbers, precision, and the more general issue of the cost-effectiveness of methodological progress. The search for meaningful theoretical simplicity has become increasingly a dominant imperative of theorybuilding and explanation, as the policy environment has continued to evolve since 1945 and has become more complex to comprehend and to attend to. The management of complexity was given a boost by the behavioral revolution of the 1960s. This revolution was accompanied by rigorous positivism and the rise to prominence of quantitative Page 492 → methods and formal modeling, which borrowed heavily from the harder disciplines in the social sciences (such as economics) and the natural sciences, and which seemed to put parsimony and effective management of complexity within reach. As has been the case with other disciplines in the social sciences, the rise of behaviorism in the 1960s created a strong tendency toward identifying truth with numbers. The aphorism, “numbers don't lie,” has gained currency in FPA. The increased sophistication of research methods and their mating with improved technology as a result of the computer revolution has made it a lot easier to generate, store, access, and process numbers and has made numbers-related methodological advancement a preferred target for growth. Closely-associated with this impetus for progress was the precision imperative. This has been another feature of a persistent quest by the social scientists—the drive toward “scientificism” in line with the modality set by hard sciences and economics.30 Precision took the form of attempting to bring the rigor of mathematical logic to arguments about state behavior in foreign policy, as well as attempting to describe international politics in an easy-to-define, grasp, and compare numerical format. But the enterprise was not without its difficulties. State behavior, at least not all important aspects of it, is not necessarily a natural candidate for treatment by these methods. A great deal of creativity was invested in tailoring these instruments to fit the subject matter of international politics. The results have not been equally convincing across all issue areas and research projects. The tailoring processes involved the imposition of highly-restrictive, difficult-to-justify assumptions that in some cases represented too much of a deviation from reality and in other cases required data transformation that raised serious validity questions. As a consequence these creative efforts resulted more than once in “precise imprecision.” Both producers and consumers of this research product have to be aware of this and consider the question of whether the substitution of transparent imprecision (in the original raw data) for what are often hidden imprecisions (in the transformed data) is a worthwhile trade-off. It might be, but only if other features of the product compensate for it. Theodore Porter apparently has it right when he asserts that in “fields dominated by a relatively secure community, much of what we normally associate with the scientific mentality—such as an insistence on objectivity, on the written word, on rigorous quantification—is to a surprising degree missing. Scientific knowledge is most likely to display conspicuously the trappings of science in fields with insecure borders, communities with persistent boundary problems. That is, one has to look at a wider context for science to understand even the accepted forms of scientific production, the standards by which work is judged.”31 A prominent illustration of the trust in numbers and related biases was the central role played by the concept and measurement of explained variance. Supposedly, uncertainties would be cleared away by theories that provide crisp answers based on a small number of relevant explanatory independent variables whose exact contributions to the explanation of variance in the dependent variable can be precisely calculated. Identifying a small Page 493 → number of critical variables that account for a large share of or, even better, most of the variance seemed at hand. This is indeed a desirable state of theorizing and verification, but only if one is confident about the process of transforming the raw data into a format that can be used in a meaningful measurable form without losing much of the actual meaning of the raw data, creating instead in fact an alternative data set, easier to use, but at the same time not an accurate representation that corresponds to the reality from which it was originally sampled. Yet in the worthy effort to establish a sub-discipline that has all the right scientific credentials, these issues of whether the

data and variables are still really what they seem has become secondary and has been transcended by the grand objective of scientificism. Aside from the internal validity problem posed by quantitative data sets, what was central to this approach was the notion of the share and measurement of explained variance. It is essential to parsimony but misses the very realistic possibility that behavioral outputs would ultimately result from cumulative and interactive inputs, so that even a small increment of influence by a variable responsible only for a minor share of the variance could be the deciding cause for the particular output. In the absence of this small, even minute, influence, there would be no anticipated outcome no matter how important and extensive the weight of other variables is. The prominence of behaviorism, and especially the powerful statistical and other methodological tools that it generated, have also been closely associated with the triumph of the logic of verification over the logic of discovery. The combination of easy-to-access databases and the availability of statistical methods and the computerized means to apply them at low cost has resulted in an advance of studies that conformed to verification standards of what was considered acceptable for publishing, even if the substantive contributions of too many of these studies were marginal or even questionable. In other words, research products increasingly reflected the fact that they could be easily produced by sitting in front of a computer with access to multiple databases and statistical software packages, rather than being a reflection of thoughtful or original research questions and creative conceptualization and theorizing. Consequently, political science in general, and FPA in particular, evidenced the proliferation of industrialized science at the expense of carefully crafted science. The bottom line is that the new technical opportunities for aggregate studies produced on the one hand heavily invested, exciting, and very important insights into state behavior.32 But on the other hand, these same opportunities resulted occasionally in an incentive for creating substantial redundancies and an oversupply of scientific information of marginal value and even lesser utility. The excesses of this trend require that more attention and emphasis be given once more to the logic of discovery and the related “So what?” question in evaluating and publishing research products. Similarly, for many years now formal modeling has seemed to hold the hope that it offers an epistemology that can combine theoretical richness and logical consistency, as well as precision and rigor.33 But does the record justify and validate these expectations? It is highly debatable whether Robert Page 494 → Powell's statement that “formal models in international relations frequently are at least as rich causally as ordinary-language analyses” is a valid one.34 There are structural limits on the extent to which complexity can be captured by formal models that cannot be easily overcome. In fact, earlier in the same paragraph, as well as elsewhere, Powell admits that “a model may be too simple in some absolute sense and one may hope to relax some restrictive assumptions as the modeling enterprise continues.”35 Unfortunately the last forty years of experience has given little credence to such hope because progress in that direction was slow and achievements less than anticipated. The modeling dialogue turned out to be a lot less productive than we could be led to believe. Indeed, it is too often true that “models are very spare … certainly too spare to explain any particular outcome in any degree of specificity,”36 which raises salient questions about their suitability to generate useful foreign policy theories.37 Our argument does not claim that the cluster of approaches and methods generally described as formal theory has not made important contributions to FPA. But at the same time it takes the realistic view that formal modeling has delivered a lot less than it promised, and, extrapolating from the record so far, it remains an unlikely candidate to surprise us in the future with major breakthroughs. This is what makes the questions of cost-effectiveness so acute, especially when other competing but important and relevant work is practically ignored. To exemplify, Powell, in discussing the role of asymmetric information as a cause of war, points to the need to problematize where these asymmetries come from.38 He defines it as an important task for the future. There is little indication in the book of awareness that this is not virgin ground and that an extensive, multidimensional, and rich body of literature in political psychology produced over the last thirty years has already successfully addressed these questions.39 Herein is a potential impediment to progress, as defined earlier in the paper. When the stricture of a methodology, whatever its other merits,40 does not allow for the incorporation and integration of important and relevant knowledge into the explanation and interpretation of the world, merely because it cannot be formulated in a particular manner required by formal modeling, not because it is untrue, then the methodology becomes an

impediment to progress. In other words, the means determine the ends rather than the other way around. And I allow myself to be skeptical about whether in the future formal theory will be in a much better position to present a rich, convincing detailed analysis of these issues that will present competitive, different, and better answers than those already suggested by political psychology. This pessimistic view draws on the fact that after almost forty years, formal modeling's progress toward answering convincingly the type of important questions posed here is hardly astounding. To borrow the words of an old hand and chronicler of the Great Game in Central Asia: “I think, the Game really was a game with scores but limited substantive prizes.” This leads to my final point: scholars (even those who are adherents of the rational choice paradigm) seem to exclude scholarly work from the basic, rational, standard tests of value and utility they would consider necessary Page 495 → in practically every other domain of social activity requiring investment of resources—that is, a test of cost-effectiveness. There has not yet been a systematic “audit” comparing the research products that result from the application of more complex and more costly research methods with the products of less expensive research methods for quality or originality. We tend to assume that the costly methodologies pay off. It is time to consider criteria and auditing methods to evaluate the cost-effectiveness of methodological progress to find out whether what we are experiencing is actually significant progress or costly progress with relatively marginal returns on our investment. It is of course an irony that very often those who are the most committed believers in theoretical parsimony are also the most fervent practitioners of methodologies that are anything but parsimonious. We are so obsessed with methods that we do not stop and ask often enough and rigorously enough the “So what?” question regarding research findings. It would be wise not to fall prey to another instance of the representativeness bias by automatically associating highly-sophisticated, seemingly rigorous methods with a supposed substantive valueadded to the knowledge product from the application of those methods.

Conclusions Theory-building in the social sciences in general and in international politics in particular is the art of the possible. It also does not converge with the pure truth; at best it represents a practical truth. Evangelical scientificism and methodologism only stand in the way of creativity and are impediments to holistic theory-building, a stage that we are just beginning to reach and in which our fledgling efforts have resulted in some impressive and some less impressive results. More specifically, several propositions flow from the arguments and observations in this paper. First, FPA requires a substantive adjustment and modernization of its research agenda to fit the changing nature of the policy environment and actors in the twenty-first century and to avoid becoming increasingly less relevant to the concerns of our times. Second, excessive concern with lack of parsimony is the result of the tyranny of methodology over common sense and relevance. In some way it reflects parroting of other disciplines and crossdisciplinary cloning rather than nuanced and thoughtful cross-disciplinary learning. Third, we are somewhat too permissive in our embrace of the premise of “multiple paths to knowledge.” While in principle it is a correct and productive premise, it does not necessarily follow that all paths to knowledge are equally effective and have the same utility function. Furthermore, the different paths to knowledge may be incompatible.41 Fourth, a unity of theory and praxis has to take into account the requirements of praxis and its diversity. This can be achieved only by attention to context and the building of contingent theories that reflect context-dependence. Context can appear in various roles or even multiple roles: that of cause, that of constraint, that of provider of meaning, and that of provider of differentiation and nuance. Fifth, we have started on the road Page 496 → to extending the reach and improving the quality of our research by looking outside the confines of the political science discipline, attempting to learn from other success stories. This has been the right thing to do. Yet, for reasons that go beyond the scope of this article, the study of international relations has been a field affected by fashions42 mostly set in other disciplines concerning both what should be investigated and how research should be conducted. It is unfortunate that fashions have too often been adopted with a passion and sanctified by practice to the point that excludes the taking of a critical perspective to evaluate those fashions' contributions, their claims for originality, and the actual value added. Collective premature cognitive closure has made productive debates on the merits of one theory or method over another a ritual more than an actual process of learning. This is a reminder that effective learning, and hence progress, occasionally involves extensive unlearning or forgetting. There is still much that we need to

learn, but FPA also carries excess baggage that we would be better off by dropping overboard, in spite of sunk costs. Finally, there is an urgent need for an innovative institutional incentive structure that will stimulate creativity rather than quench it by rewarding conservatism, conformism, and risk aversion, as is generally the case regarding the manner in which the discipline of political science awards and withholds the benefits of tenure, salary, and prestige.43 This paper should not be construed as an argument against more rigor, better methods, higher standards of scientific validity, and the employment of sophisticated formal methods and quantitative tools. But it is surely an argument for being practical and cautious about employing relevant standards of scientific validity and for not being carried away with the maxim, “If some is good a lot is better”—in order that we may avoid the pernicious consequences of seeing new technologies and methods that worked in other fields as panaceas for solving problems that are inherent to the unique attributes of our field of inquiry and are of a different breed. It is an argument that we should recognize the constraints that make FPA a subject more of scientific approximations than of scientific precision. But even more important, it is critical that we recognize the dichotomy between the logic of discovery and the logic of verification and understand that the two may converge, but that is not necessarily always the case. In those cases where the latter is true, it would be wiser and more productive to allow the logic of discovery to predominate over the logic of verification. It might do us all good if in our quest for a scientific truth to which we are all passionately committed, we should keep in mind the following sagacious observation and advice: “Truth is both fragile and subtle.… science at its best is like a firm but gentle hand that holds a butterfly without crushing it.”44 It seems that in our fierce love for truth we occasionally embrace it so tightly that we crush it without even being aware of it, and then we feel righteous about it. Having said all that, and although this paper takes a harsh critical view of the study of FPA, I conclude by reiterating the theme with which this paper opened and that has been repeated throughout the text. In spite of the Page 497 → many criticisms, perhaps not all acceptable to everybody, there is no doubt that when all is said and done a balanced evaluation will also take note of the immense amount of knowledge that has accumulated since the late 1950s. Not all of this knowledge is consensual, and there are ongoing debates and sharp disagreements between the dominant schools in the field. Encouraging debates and disagreements is the path away from stagnation and toward a lively and meaningful process of scientific growth and progress.

Notes An earlier version of this paper was prepared for presentation at the forty-first annual convention of the International Studies Association held March 14–18, 2000, at Los Angeles, while the author was a fellow at the Netherlands Institute for Advanced Study (NIAS). I would like to thank NIAS for its generous hospitality and support. A revised version was prepared during the period I spent as a visiting scholar at the Peace and Governance Program, the United Nations University (UNU), Tokyo. I am grateful to UNU for this unique opportunity. The author also thanks Michael Brecher and Miles Kahler for useful comments and suggestions. 1. For example, C. Brown, Understanding International Relations (London: Macmillan, 1997), 79. 2. For useful reviews of many of the issue areas in FPA see C. F. Hermann, C. W. Kegley Jr., and J. N. Rosenau, eds., New Directions in the Study of Foreign Policy (Boston: Allen and Unwin, 1987); M. Light, “Foreign Policy Analysis,” in Contemporary International Relations: A Guide to Theory, ed. A. J. R. Groom and M. Light (London: Pinter, 1994), 93–108; L. Neack, A. K. J. Hey, and P. J. Haney, eds., Foreign Policy Analysis: Continuity and Change in Its Second Generation (Englewood Cliffs, N.J.: Prentice Hall, 1995). The literature referred to in the paper does not aspire to capture and reflect the full extent of the wealth of FPA literature. It is used to illustrate the main specific points made in the paper, with special emphasis on the more recent literature, rather than with the purpose of being all-inclusive.

3. Note, for example, the recent review of the state of political cognition research in which the author seems to be very pleasantly surprised by her findings, having started with very low expectations. K. M. McGraw, “Contributions of the Cognitive Approach to Political Psychology,” Political Psychology 21, no. 4 (2000): 805–32. 4. For example, M. Holden, “The Competence of Political Science: Progress in Political Research Revisited, ” American Political Science Review 94, no. 1 (2000): 1–19. 5. For example, B. Bueno de Mesquita and J. D. Morrow, “Sorting through the Wealth of Notions,” International Security 24, no. 2 (1999): 56–73; R. Powell, In the Shadow of Power: States and Strategies in International Politics (Princeton, N.J.: Princeton University Press, 1999). 6. H. A. Simon, Models of Discovery (Dordrecht, Holland: Reidel, 1977), xvi–xvii. 7. A clear but implicit dichotomy can be observed between journals that emphasize the logic of discovery and those that emphasize the logic of verification. A few illustrations will clarify this point: International Organization is clearly discovery-driven and has made this attribute an important aspect of its organizational culture, which has helped keep it on a consistent track in that sense for almost three decades in spite of regular changes of editors. In my view, this characteristic probably has made it the most consistently interesting and idea-rich journal in international studies. The Journal of Conflict Resolution, on the other hand, is verification-driven, while Political Psychology and International Studies Quarterly have an uneven record of volatility, tilting toward one side or the other depending on the composition of editorial teams and coincidence, but with a slight but obvious preference for criteria that emphasize the importance of verification. It should be noted that whatever a journal's main orientation, it does not exclude the other orientation but only gives it less prominence in editorial preferences. 8. M. G. Hermann, “When Leader Personality Will Affect Foreign Policy: Some Propositions,” in In Search of Global Patterns, ed. J. N. Rosenau (New York: Free Press, 1976), 326–33; M. G. Hermann, “Personality and Foreign Policy Decision-Making: A Study of 53 Heads of Government,” in Foreign Policy DecisionMaking: Perception, Cognition, and Artificial Intelligence, ed. D. A. Page 498 → Sylvan and S. Chan (New York: Praeger, 1984), 58–80; M. H. Halperin, Bureaucratic Politics and Foreign Policy (Washington, D.C.: Brookings Institution, 1974); C. F. Hermann, ed., International Crises (New York: Free Press, 1972); S. G. Walker, ed., Role Theory and Foreign Policy Analysis (Durham, N.C.: Duke University Press, 1987). 9. One example of overrating by decontextualization of seemingly convincing research findings is the case of the democratic peace hypothesis. On closer examination the generalized findings raised serious validity questions that have led to a much more restrictive and conditional view of the validity of the democratic peace hypothesis. See, for example, the review by Miriam Fendius-Elman, “The Never Ending Story: Democracy and Peace,” International Studies Review 1, no. 3 (1999): 87–103. 10. D. Kahneman, P. Slovic, and A. Tversky, eds., Judgment under Uncertainty: Heuristics and Biases (Cambridge: Cambridge University Press, 1982); R. Nisbett and L. Ross, Human Inference: Strategies and Shortcomings of Social Judgment (Englewood Cliffs, N.J.: Prentice Hall, 1980); J. Reason, Human Error (Cambridge: Cambridge University Press, 1990); Y. Y. I. Vertzberger, The World in Their Minds: Information Processing, Cognition, and Perception in Foreign Policy Decisionmaking (Stanford, Calif.: Stanford University Press, 1990). 11. E. Ostrom, “Institutional Rational Choice: An Assessment of the Institutional Analysis and Development Framework,” in Theories of the Policy Process, ed. P. A. Sabbatier (Boulder, Colo.: Westview, 1999), 37. 12. J. Goldstein and R. O. Keohane, eds., Ideas and Foreign Policy: Beliefs, Institutions, and Political Change (Ithaca, N.Y.: Cornell University Press, 1993); D. A. Lake and R. Powell, eds., Strategic Choice and International Relations (Princeton, N.J.: Princeton University Press, 1999). 13. For example, T. Eggertsson, Economic Behavior and Institutions (Cambridge: Cambridge University Press, 1990); E. G. Furubotn and R. Richter, Institutions and Economic Theory: The Contribution of the New Institutional Economics (Ann Arbor: University of Michigan Press, 1997); D. C. North, Institutions, Institutional Change, and Economic Performance (Cambridge: Cambridge University Press, 1990); B. G. Peters, Institutional Theory in Political Science: The “New Institutionalism” (London: Pinter, 1999); M. Rutherford, Institutions in Economics: The Old and the New Institutionalism (Cambridge: Cambridge University Press, 1996). 14. For example, L. L. Martin and B. A. Simmons, “Theories and Empirical Studies of International

Institutions,” International Organization 52, no. 4 (1998): 729–57. 15. For example, Y. Y. I. Vertzberger, Risk Taking and Decisionmaking: Foreign Military Intervention Decisions (Stanford, Calif.: Stanford University Press, 1998); D. Vaughan, “The Dark Side of Organizations: Mistake, Misconduct, and Disaster,” in Annual Review of Sociology, vol. 25, ed. K. S. Cook and J. Hagan (Palo Alto, Calif.: Annual Reviews, 1999), 271–305. 16. For example, M. Bovens and P. 't Hart, Understanding Policy Fiascoes (New Brunswick, N.J.: Transaction, 1996); P. 't Hart, Groupthink in Government: A Study of Small Groups and Policy Failures (Amsterdam: Zwets and Zeitlinger, 1990). 17. A. Prakash and J. A. Hart, “Coping with Globalization: An Introduction,” in Coping with Globalization, ed. A. Prakash and J. A. Hart (London: Routledge, 2000), 1–26; B. White, “The European Challenge to Foreign Policy Analysis,” European Journal of International Relations 5 (1999): 37–66. 18. See A. L. George, Bridging the Gap: Theory and Practice in Foreign Policy (Washington, D.C.: United States Institute for Peace Press, 1993). 19. D. A. Baldwin, “Success and Failure in Foreign Policy,” in Annual Review of Political Science, vol. 3, ed. N. W. Polsby (Palo Alto, Calif.: Annual Reviews, 2000), 167–82; D. L. Byman and M. C. Waxman, “Kosovo and the Great Air Power Debate,” International Security 24, no. 4 (2000): 5–38; D. Larson and S. Renshon, Good Judgment in Foreign Policy: Theory and Applications (Lanham, Md.: Rowan and Littlefield, 2001); P. E. Tetlock, “Good Judgment in International Page 499 → Politics: Three Psychological Perspectives,” Political Psychology 13, no. 3 (1992): 517–39. 20. It is worth noting that scholars rejecting theoretical complexity on the grounds of the lack of parsimony have not been in the least reluctant to accept technical (methodological) complexity. Perhaps they consider trading off theoretical simplicity at the cost of methodological complexity a justified transaction. From the same perspective those, like the author, less concerned with the costs of theoretical complexity perceive the costs of methodological complexity to be too high compared with the benefits from the gains of theoretical simplicity. 21. M. J. Levy, “Does It Matter If He's Naked? Bawled the Child,” in Contending Approaches to International Politics, ed. K. Knorr and J. N. Rosenau (Princeton, N.J.: Princeton University Press, 1969), 89–90; also C. Beed and C. Beed, “Is the Case for Social Sciences Laws Strengthening?” Journal for the Theory of Social Behaviour 30, no. 2 (2000): 131–53; A. O. Hirschman, “Against Parsimony: Three Easy Ways of Complicating Some Categories of Economic Discourse,” Economics and Philosophy 1, no. 1 (1985): 7–21; G. King, R. O. Keohane, and S. Verba, Designing Social Inquiry: Scientific Inference in Qualitative Research (Princeton, N.J.: Princeton University Press, 1994), 104–5; R. N. Lebow, “Beyond Parsimony: Rethinking Theories of Coercive Bargaining,” European Journal of International Relations 4, no. 1 (1998): 31–66; C. Tilly, “To Explain Political Processes,” American Journal of Sociology 100, no. 6 (1995): 1594–1610. 22. For some of the discussions and critical reviews of the promise and limits of the rationality paradigm see K. S. Cook and M. Levi, eds., The Limits of Rationality (Chicago: University of Chicago Press, 1990); J. D. Fearon, “Rationalist Explanations for War,” International Organization 49, no. 3 (1995): 379–414; D. P. Green and I. Shapiro, Pathologies of Rational Choice Theory: A Critique of Applications in Political Science (New Haven, Conn.: Yale University Press, 1994); M. Kahler, “Rationality in International Relations,” International Organization 52, no. 4 (1998): 919–41; S. Kelley, “The Promise and Limitations of Rational Choice Theory,” Critical Review 9, no. 1–2 (1995): 95–106; R. E. Lane, “What Rational Choice Explains,” Critical Review 9, no. 1–2 (1995): 107–26; L. L. Martin, “The Contribution of Rational Choice: A Defense of Pluralism,” International Security 24, no. 2 (1999): 74–83; Vertzberger, Risk Taking and Decisionmaking; S. M. Walt, “Rigor or Rigor Mortis? Rational Choice and Security Studies,” International Security 23, no. 4 (1999): 5–48; and C. H. Zuckert, “On the Rationality of Rational Choice,” Political Psychology 16, no. 1 (1995): 179–98. 23. For example, B. D. Jones, “Bounded Rationality,” in Annual Review of Political Science, vol. 2, ed. N. W. Polsby (Palo Alto, Calif.: Annual Reviews, 1999), 297–321. 24. H. A. Simon, Models of Man (New York: Wiley, 1957). 25. G. A. Miller, “The Magical Number Seven, Plus or Minus Two: Some Limits on Our Capacity for Processing Information,” Psychological Review 63, no. 2 (1956): 81–97. 26. Z. Kunda, “The Case for Motivated Reasoning,” Psychological Bulletin 108, no. 3 (1990): 480–98.

27. P. E. Tetlock, “The Impact of Accountability on Judgment and Choice: Toward a Social Contingency Model,” in Advances in Experimental Social Psychology, vol. 25, ed. M. P. Zana (San Diego: Academic Press, 1992), 331–76. 28. For example, A. Bennett, Condemned to Repetition: The Rise, Fall, and Reprise of Soviet-Russian Military Intervention (Cambridge: MIT Press, 1999); M. Brecher, Crises in World Politics: Theory and Reality (Oxford, England: Pergamon, 1993); R. N. Lebow, Between Peace and War: The Nature of International Crisis (Baltimore: Johns Hopkins University Press, 1981); C.A. Kupchan, The Vulnerability of Empire (Ithaca, N.Y.: Cornell University Press, 1994); R. N. Lebow and J. G. Stein, We All Lost the Cold War (Princeton, N.J.: Princeton University Press, 1994); Z. Maoz, National Choices and International Processes (Cambridge: Cambridge University Press, 1990); J. N. Rosenau, Turbulence Page 500 → in World Politics: A Theory of Change and Continuity (Princeton, N.J.: Princeton University Press, 1990); G. H. Snyder and P. Diesing, Conflict among Nations: Bargaining, Decision-Making, and System Structure in International Crisis (Princeton, N.J.: Princeton University Press, 1977); Vertzberger, The World in Their Minds; Vertzberger, Risk Taking and Decisionmaking. 29. P. E. Tetlock, “Coping with Trade-Offs: Psychological Constraints and Political Implications” (n.d., mimeographed), 45. 30. C. Argyris, The Inner Contradictions of Rigorous Research (New York: Academic Press, 1980); T. Mayer, Truth versus Precision in Economics (Aldershot, England: Elgar, 1993); T. M. Porter, Trust in Numbers: The Pursuit of Objectivity in Science and Public Life (Princeton, N.J.: Princeton University Press, 1995); M. N. Wise, ed., The Values of Precision (Princeton, N.J.: Princeton University Press, 1995). 31. Porter, Trust in Numbers, 230. 32. For example, M. Brecher and J. Wilkenfeld, A Study of Crisis (Ann Arbor: University of Michigan Press, 1997). 33. Bueno de Mesquita and Morrow, “Sorting through the Wealth of Notions”; R. B. Morton, Methods and Models (Cambridge: Cambridge University Press, 1999); Powell, In the Shadow of Power. 34. Powell, In the Shadow of Power, 37. 35. Ibid., 37 and concluding chapter. 36. Walt, “Rigor or Rigor Mortis?” 38. See also G. L. Munck, “Game Theory and Comparative Politics: New Perspectives, Old Concerns,” World Politics 53, no. 2 (2001): 173–204. 37. C. Elman, “Why Not Realist Theories of Foreign-Policy Theory?” Security Studies 6, no. 1 (1996): 13–16. 38. Powell, In the Shadow of Power, 202–3. 39. The writings on the many different aspects of the forms, causes, and consequences of information asymmetries are too many to be cited in full; hence I shall cite just an illustrative sample: Brecher, Crises in World Politics; J. F. Burke and F. I. Greenstein, How Presidents Test Reality: Decisions on Vietnam, 1954 and 1965 (New York: Russell Sage Foundation, 1989); A. L. George, Presidential Decisionmaking in Foreign Policy: The Effective Use of Information and Advice (Boulder, Colo.: Westview, 1980); O. R. Holsti, “Crisis Decision-Making,” in Behavior, Society, and Nuclear War, vol. 1, ed. P. E. Tetlock, J. L. Husbands, R. Jervis, P. C. Stern, and C. Tilly (New York: Oxford University Press, 1989), 8–84; R. Jervis, Perception and Misperception in International Politics (Princeton, N.J.: Princeton University Press, 1976); D. W. Larson, The Origins of Containment: A Psychological Exploration (Princeton, N.J.: Princeton University Press, 1985); Lebow, Between Peace and War; J. S. Levy, “Misperception and the Causes of War,” World Politics 38, no. 1 (1983): 76–99; J. S. Levy, “Learning and Foreign Policy: Sweeping a Conceptual Minefield,” International Organization 48, no. 2 (1994): 279–312. J. A. Rosati, “The Power of Human Cognition in the Study of World Politics,” International Studies Review 2, no. 3 (2000): 45–75; J. G. Stein and R. Tanter, Rational Decision-Making: Israel's Security Choices, 1967 (Columbus: Ohio State University Press, 1980); D. A. Sylvan and S. J. Thorson, “Ontologies, Problem Representation, and the Cuban Missile Crisis,” Journal of Conflict Resolution 36, no. 4 (1992): 709–32; P. 't Hart, E. K. Stern, and B. Sundelius, eds., Beyond Group-think: Political Group Dynamics and Foreign Policy-Making (Ann Arbor: University of Michigan Press, 1997); Vertzberger, The World in Their Minds; S. G. Walker, “The Interface between Belief and Behavior: Henry Kissinger's Operational Code and the Vietnam War,” Journal of Conflict Resolution 21, no. 1 (1977): 129–68. 40. E. M. S. Niou and P. C. Ordeshook, “Return of the Luddites,” International Security 24, no. 2 (1999):

84–96. 41. On the question of incompatibility, see the dialogue between S. Smith, “Foreign Policy Theory and the New Europe,” in European Foreign Policy: The EC and Changing Perspectives in Europe, ed. W. Carlsnaes and S. Smith (London: Sage, Page 501 → 1994), 1–20, and W. Carlsnaes, “In Lieu of a Conclusion: Compatibility and the Agent-Structure in Foreign Policy Analysis,” in ibid., 274–87. 42. M. Brecher, “International Studies in the Twentieth Century and Beyond: Flawed Dichotomies, Synthesis, Cumulation,” International Studies Quarterly 43, no. 2 (1999): 213–64. 43. Holden, “The Competence of Political Science,” 15. 44. K. S. Bowers, “Situationism in Psychology: An Analysis and Critique,” Psychological Review 80, no. 5 (1973): 333.

Page 502 →

BELIEFS AND FOREIGN POLICY ANALYSIS IN THE NEW MILLENNIUM Stephen G. Walker

Introduction A main axis of intellectual tension in the area of foreign policy analysis over the past forty years is the issue of the importance of “beliefs”1 in the explanation and prediction of foreign policy decisions and outcomes. The seminal decision-making approach to foreign policy articulated by Snyder, Bruck, and Sapin at midcentury was partly in reaction to a skewed emphasis by the realist tradition on external circumstances (e.g., the balance of power) and the omission of beliefs in explaining foreign policy decisions.2 Snyder and his colleagues argued that this strategy of explanation came at the expense of neglecting the definition of the situation represented by the decision maker's beliefs about the external and internal setting for decision. One of the consequences was to limit insights into the dynamics of international crises, an important area of foreign policy analysis during the cold war and crucial to realist concerns with the question of war and peace among the great powers in the international system. Subsequent research efforts by several scholars attempted to demonstrate the validity of the decision-making approach to international crises over the next several decades.3 In spite of such efforts, a debate has occurred over whether beliefs are merely epiphenomena or important causal mechanisms in foreign policy analysis.

Are Beliefs Merely Epiphenomena? As the subfield of foreign policy analysis evolved, the claims on behalf of the general approach and the focus on the decision maker's definition of the situation became increasingly qualified and diluted in several ways. To their credit, practitioners of the decision-making approach identified many of these limitations. Holsti concluded that beliefs were most likely to be influential in foreign policy decisions when the decision maker was faced with radical uncertainty, that is, situations that were ambiguous, Page 503 → nonroutine, novel, or unanticipated, prompting a reliance on old information in the form of preexisting beliefs rather than new information from the environment.4 George argued that even under those conditions beliefs were most directly and strongly related to the preferences of decision makers rather than to the actual behavior of states.5 The impact on a state's behavior was likely to be diluted, as the impact of individual preferences was modified by other variables such as stimuli and conditions in the international environment and limitations on choice imposed by organizational or domestic circumstances. More severe criticisms stemmed from the problems associated with aggregating from individual beliefs to collective decisions, which changes the decision-making unit in foreign policy analysis from the individual to larger entities such as small groups, bureaucratic organizations, special interest groups, or governmental institutions culminating in the state as the actor.6 If the power to decide does not reside in the individual, then it becomes even more difficult to trace the microfoundations of decisions to beliefs. In accounts of state decision making at higher levels of aggregation, beliefs are reduced to epiphenomena, that is, merely doctrinal expressions of more fundamental bureaucratic interests, organizational structure, or institutional history.7 They are important guides to action, but their contents are determined heavily by other antecedent variables, leaving little room for the autonomous impact of beliefs on behavior. While they are located close to the behavior to be explained, shared beliefs do not escape the “endogeneity trap.”8 In sum, the cumulative effects of these trends in an increasing order of severity include the following selfcriticisms and indictments by others of the impact and importance of beliefs on foreign policy and world politics: (1) beliefs are important and independent influences on foreign policy decisions only under a restricted range of

conditions; (2) beliefs are just one variable within a larger complex of explanatory variables; (3) the impact of beliefs on foreign policy behavior tends to be indirect and general rather than direct and specific; (4) beliefs tend to be an intervening or dependent variable in foreign policy analyses rather than an independent variable; (5) beliefs are extraneous variables in foreign policy and world politics and best ignored except for the possible explanation of deviant cases when other explanations fail.

Are Beliefs Important Causal Mechanisms? And yet “beliefs” of one kind or another still command a lot of scholarly attention and expenditure of intellectual resources. The transdiscipline of political psychology has emerged over the past twenty-five years with an international society, an annual meeting, a journal, and a summer institute that trains graduate students. Within the past decade a political psychology section has also been formed within the American Political Science Association. A preoccupation with the “agent-structure” problem and the recent Page 504 → “constructivist” turn in international relations theory has redirected international relations scholarship toward psychocultural phenomena as important causes of foreign policy behavior.9 During this revival, Goldstein and Keohane have suggested three types of causal powers associated with explanations of foreign policy and international politics that incorporate beliefs as mechanisms.10 First, beliefs can perform as “worldviews” that inform the leader about the nature of the political universe and exert mapping effects.11 That is, the leader infers from these beliefs who self and other are and what to do given this map of the social situation. Second, beliefs can also take the form of “principled beliefs” that exert ethical or normative power as steering effects on strategic decisions.12 Third, beliefs may serve as “causal beliefs” about cause-andeffect relations in problem representation and have implications for action in the form of solution effects.13 It is possible to account for this resurgence partly by reference to the tenor of the times. With the unexpected end of the cold war, scholars became concerned with their inability to anticipate this development without incorporating the impact of changes in the beliefs of decision makers.14 The vacuum that followed the end of the cold war was not simply a power vacuum with the demise of the USSR. There was also an intellectual vacuum, as both policymakers and scholars struggled to grasp the realities of the new world order. Without the precepts of a bipolar world in the post-cold war era to guide decisions, the impact of the beliefs of leaders in this uncertain environment was projected as likely to increase.15 The turn toward beliefs as causal mechanisms in one form or another may also be part of a wider reaction against rational choice theory that preceded the end of the cold war.16 Postmodernists, constructivists, and political psychologists tend to share distaste for rational choice models of decisions, especially those that restrict the autonomous ability of agents to select goals for political action. The beliefs and identities of individuals are not easily subsumed under the interests and preferences of rational choice models within any of these schools of thought.17 The arguments arrayed against rational choice models within the decision-making approach to foreign policy analysis range from rational choice theory's lack of descriptive power for how decisions are actually made to its inability to account for the actor's preferences or goals, which are often taken simply as givens in subjective utility and strategic models of rational choice.18 If one looks beyond the polemical exchanges to the research practices of scholars who employ some form of beliefs in their models, there is common ground shared by rational choice and other decision-making theorists. Beliefs are one explanation for “flawed” decisions, ones guided by misperceptions that make seemingly irrational choices to outside observers more “rational” once the flawed information on which they were based is taken into account.19 Even foreign policy successes that appear to follow the ends/means maxims of rational choice may turn out to be the result of perceptions whose flaws cancel out one another.20 More recent work by rational choice theorists also recognizes beliefs as important intervening Page 505 → variables between external circumstances and the foreign policy choices of decision makers.21 These examples suggest that the debate over rational choice is partly over “thin” versus “thick” models of rationality.22 A thin model assumes that beliefs are relatively unimportant and that decision makers are highly

influenced by circumstances. These external boundaries override the internal dispositions of leaders toward other courses of action. A thick model assumes that beliefs are relatively important and that decision makers have internal boundaries that limit the capacity for choice. These internal constraints may be either content-full or content-free biases, depending upon whether they refer to schemata or heuristics.23 If beliefs are shared or simply recognized as representing the expectations of others, they may act as epistemic or cultural boundaries in much the same way that more idiosyncratic or internalized beliefs do but from an external location.24

A Strategy for Synthesis The challenge facing theorists in the new millennium who begin with “beliefs” as the primordial concept in their approach to foreign policy analysis is to identify these internal and external boundaries of rational choice. Ironically, therefore, the best strategy may be to clarify and refine what it means to be “rational” even though the rational choice approach itself raises the hackles of political psychologists and others noted in the previous section. One example of this strategy is the introduction of prospect theory into foreign policy analysis.25 A second example is to bring rational choice and cognitive models of decision making together in a common application.26 A third example is to link psychological models of information processing with the insights of bounded rationality theorists operating out of other research traditions.27 More broadly, there are two general lines of inquiry associated with this strategy joined by a shared assumption and inspired by a common goal. A shared assumption is that beliefs are not merely boundaries but pivotal mechanisms for explaining choices and consequences in foreign policy and world politics.28 A common goal is to demonstrate the various kinds of causal powers associated with beliefs that support this shared assumption. One line of inquiry is to focus on beliefs as microlevel mechanisms, while the other is to focus on beliefs as macrolevel mechanisms of choice.29 This “two-lane” view of the research road ahead is consistent with a definition of political psychology as a branch of social psychology, which occupies an intellectual niche between the disciplines of psychology and sociology and has a focus on the links among individuals, social groups, and behavior at different levels of analysis. Much of the work on foreign policy decision making over the past forty years has focused on beliefs primarily as microlevel, individual mechanisms rather than as macrolevel, social mechanisms. In the parlance of rational choice theorists, the emphasis has been on the “subjective” in subjective rational utility models rather than the “strategic” in strategic rationality Page 506 → models.30 With the exception of some of the work on deterrence,31 the focus in foreign policy analysis has been largely on the one-sided actions of states rather than the two-sided strategic interactions between them.32 An emphasis on strategic interactions in international relations theory outside a crisis context has been dominated by the realist and liberal schools who use instrumental rationality calculations about costs and benefits to inform their balance-of-power or institutional models.33 The work of the constructivist school focuses on the creation of norms and institutions taken for granted by realists and liberals.34 None of these three schools—realists, liberals, and constructivists—seems to think that individuals matter so much as social forces or interests in the formation of beliefs. However, my sense is that individuals—particularly leaders—do matter in at least three ways: (1) the actions of leaders affect the formation, reinforcement, and alteration of shared beliefs and interests within a state; and (2) the beliefs of leaders in one state also affect the strategic interactions between states, (3) thereby having an impact on the beliefs and interests of leaders in other states. In other words, beliefs have steering effects on behavior and are subject to learning effects from stimuli in the form of cues from other leaders.35

A Research Agenda for Future Progress One way to capture the dynamics of these three processes and the pivotal impact of beliefs is to analyze initiatives toward peace as well as toward war within the conventional model of an international crisis. This analytical move is consistent with a conceptualization of crises as turning points rather than simply confrontations,36 while

retaining the insights of the conventional crisis model as (a) involving a precipitant event, (b) prompting a series of moves by the protagonists, (c) reaching a turning point, and (d) ending with a “resolution” of a crisis in one of four outcomes—(1) settlement, (2) deadlock, or (3) domination by one side or (4) the other side.37 In previous applications of the model, the transition from turning point to resolution may be peaceful or violent but usually follows the militarization of a conflict by the threat of force as the precipitating event.38 A reorientation of the model allows for the precipitating event to produce an initiative to de-escalate conflict or to escalate cooperation. Some examples include the announcement by Anwar Sadat that he would visit the Israeli Knesset to initiate a peace settlement with Israel regarding the Sinai Peninsula and the Gaza Strip, the initiative by Mikhail Gorbachev to reduce tensions between East and West in Europe, and Neville Chamberlain's adoption of the appeasement strategy in an attempt to reach a general settlement of issues between Great Britain and Nazi Germany. As these three cases indicate, respectively, the precipitating event may occur in the midst of a militarized dispute over a concrete issue, come during a protracted period of tension such as the cold war between the United States and the Soviet Union, or appear as a diplomatic initiative between states that have not yet become implacable antagonists. Page 507 → This reconceptualization of the conventional crisis model expands the universe of cases for analysis in new and interesting ways. One can now view military initiatives and diplomatic initiatives within the same general analytical framework. Employing the same concepts and propositions for analyzing the dynamics that occur within this framework creates the opportunity to subsume and explain within the same theoretical context what appear superficially to be disparate processes. A cogent theoretical and empirical account of the successes and failures from such divergent initiatives as Gorbachev's “New Thinking” that led to the settlement of the cold war and Hitler's “Ostpolitik” that resulted in the escalation to world war requires a two-sided analysis. The respondent's perspective and behavior must also be incorporated. A formal model that unites these processes is an expanded version of the international interaction game developed by Bueno de Mesquita and Lalman (see figure 1).39 This game begins with a “move by nature”—that is, a precipitating event in the form of a stimulus from such environmental sources as decisions by third parties, a change in leadership, or a technological breakthrough. Singly or together, they prompt an initiative in the form of a demand (D) by one state toward another state, defined as a claim Page 508 → accompanied by a threat of force (F) in the event that the claim is not met. The target of the demand may choose not to make a counterdemand (~D), terminating the game (and the conflict) by acquiescence, or extend the game by issuing a counterdemand. A counterdemand escalates the conflict to a military crisis defined as a confrontation in which the protagonists have exchanged threats to use military force. One of the parties in the conflict may choose force or not at this point, leading to a choice by the other. If the first does not choose force, then the other may either reciprocate, leading to negotiations, or choose force. If both sides choose force, the outcome of the conflict is war. If one chooses force while the other does not, then the conflict outcome is capitulation in favor of the state that chose force.40 The logic of this interaction game is fueled from the beginning by a threat to use force. Demands are the only initiatives fully specified in the game. A state's responses may be choices leading to acquiescence, capitulation, or negotiation, but the de-escalation process accompanying these choices is not modeled as an initiative. The interaction game begins either with a demand by state A or a demand by state B. If neither state initiates a demand, then the outcome is simply the status quo (SQ). However, this one-sided view of state interactions is easily expanded in figure 2 to include cooperative initiatives following a move by nature. Either Page 509 → state A or state B initiates a proposal (P) defined as a claim accompanied by a promise to commit the resources necessary to carry out the contents of the proposal. The target may choose not to respond with a counterproposal (~P), terminating the game with acquiescence, or extend the game by issuing a counterproposal.

A counterproposal escalates the interactions to a diplomatic crisis defined as a window of opportunity in which the protagonists have exchanged promises to use rewards. One of the parties may choose reward (R) or not (~R) at this point, leading to a choice by the other. If the first does not choose reward, then the other may reciprocate and continue negotiations, or choose reward. If both states choose reward, the outcome of their interactions is peace. If one chooses reward while the other does not, then the outcome is exploitation by the party that did not choose reward. While the logic of the general interaction game is straightforward in its extended form as a decision tree, lacking in previous rational choice applications is a systematic examination and theoretical presentation of both the psychological processes and the strategic interaction processes that shape the course and fate of cooperative initiatives. Conversely, foreign policy analyses of the beliefs and psychological processes experienced by leaders initiating either cooperation or conflict are often one-sided accounts that do not focus heavily on the actual interactions between states, or they do not examine systematically the corresponding decision-making processes experienced by the other side. Even accounts of strategic interactions that explicitly recognize the theoretical importance of psychological processes often stop somewhat short of both measuring the beliefs of leaders and integrating these observations into their analyses.41

Conclusion To sum up, an expanded focus on crises as turning points in international relations integrates threats of conflict and opportunities for cooperation, thereby synthesizing a focus on the respective explananda in the realist and liberal schools of international relations theory. Identifying beliefs as pivotal macromechanisms of bounded rationality within this framework brings the constructivist and rational choice schools into the analysis. Candidate theories with bounded rationality mechanisms include social identity theory and role theory at the macro level of analysis and prospect theory and cognitive consistency theory at the micro level.42 This approach also calls into question the criticisms of beliefs and the decision-making approach discussed earlier. Beliefs are not extraneous variables in foreign policy and world politics, which are best ignored unless they are needed to explain foreign policy fiascoes. They are pivotal variables with mapping, steering, and solution effects within a dynamic interactive causal model rather than being simply intervening or dependent variables in onesided foreign policy analyses. The conceptualization of beliefs as important causal mechanisms rather than merely as structural boundaries enhances the possibility that their impact is likely to be more specific and direct Page 510 → instead of general and indirect. Beliefs are no longer just another variable within a larger complex of variables. By combining a bidirectional, strategic approach with a two-level, psychological approach, the qualifications that beliefs are important and independent influences on foreign policy for only a narrow range of behavior or under a restricted range of conditions are also undercut. The restriction of explananda in cognitive models to a narrow range of behavior is relaxed by expanding the crisis framework to include both conflict and cooperation processes. Holsti's specified scope conditions, in which beliefs are relevant in a “weakly structured” situation characterized as ambiguous, nonroutine, novel, or otherwise unanticipated, are also expanded.43 Under these conditions beliefs were postulated as microlevel mechanisms for shaping individual decisions. The converse of these conditions is a “strongly structured” situation where external constraints shape behavior. In these latter situations, however, there are also likely to be roles or social identities, which act as macrolevel mechanisms that constrain behavior and thereby expand the importance of beliefs as causal mechanisms.44

Notes 1. I put the word beliefs in quotation marks to indicate that I use it here generically to refer to the various conceptualizations for the products of cognition in the human mind—for example, schemata, attitudes, perceptions, images. Although the various research scholars that employ these different conceptualizations

are arguably all part of the “cognitivist” research program in world politics, I also acknowledge that it is necessary and desirable to be more specific on the kind of “beliefs” employed when conducting a particular analysis. These cognitive constructs may also be shared with others, extending the scope of my usage to include the cultural beliefs of a group, society, or system of states. See P. E. Tetlock, “Social Psychology and World Politics,” in Handbook of Social Psychology, ed. D. G. Gilbert, S. T. Fiske, and G. Lindzey (New York: McGraw-Hill, 1998), 869–912; J. Goldstein and R. O. Keohane, eds., Ideas and Foreign Policy (Ithaca, N.Y.: Cornell University Press, 1993); P. J. Katzenstein, ed., The Culture of National Security (New York: Columbia University Press, 1996); A. Wendt, Social Theory of International Politics (Cambridge: Cambridge University Press, 1999); V. M. Hudson and M. W. Sampson, eds., Culture and Foreign Policy Analysis, special issue of Political Psychology 20, no. 4 (1999). 2. R. C. Snyder, H. W. Bruck, and B. Sapin, eds., Foreign Policy Decision-Making as an Approach to the Study of International Politics, Foreign Policy Analysis Project Series, no. 3 (Princeton, N.J.: Princeton University Press, 1954). 3. For example, G. D. Paige, The Korean Decision (New York: Free Press, 1958); O. R. Holsti, R. A. Brody, and R. C. North, “Measuring Affect and Action in International Reaction Models,” in International Politics and Foreign Policy, 2d ed., ed. J. N. Rosenau (New York: Free Press, 1969), 679–96; C. F. Hermann, “International Crisis as a Situational Variable,” in ibid., 409–21; C. F. Hermann and L. P. Brady, “Alternative Models of International Crisis Behavior,” in International Crises, ed. C. F. Hermann (New York: Free Press, 1972), 281–303; O. R. Holsti, Crisis, Escalation, War (Montreal: McGill-Queen's University Press, 1972); M. Brecher, Decisions in Israel's Foreign Policy (London: Oxford University Press, 1974); G. H. Snyder and P. Diesing, Conflict among Nations (Princeton, N.J.: Princeton University Press, 1977); R. N. Lebow, Between Peace and War (Baltimore: Johns Hopkins University Press, 1981); M. Brecher, B. S. Steinberg, and J. G. Stein, “A Framework for Research on Foreign Policy Behavior, Page 511 → ” Journal of Conflict Resolution 13, no. 1 (1969): 75–101; M. Brecher and J. Wilkenfeld, A Study of Crisis (Ann Arbor: University of Michigan Press, 1997). 4. O. R. Holsti, “Foreign Policy Decisions Viewed Cognitively,” in The Structure of Decision, ed. R. M. Axelrod (Princeton, N.J.: Princeton University Press, 1976), 18–54. 5. A. L. George, “The Causal Nexus between Cognitive Beliefs and Decision-Making Behavior,” in Psychological Models in International Politics, ed. L. S. Falkowski (Boulder, Colo.: Westview, 1979), 95–124. 6. G. T. Allison, Essence of Decision: Explaining the Cuban Missile Crisis (Boston: Little, Brown, 1971); S. D. Krasner, Defending the National Interest (Princeton, N.J.: Princeton University Press, 1978). 7. S. Van Evera, “The Cult of the Offensive and the Origins of the First World War,” International Security 9, no. 1 (1984): 58–108; D. D. Avant, “The Institutional Sources of Military Doctrine: Hegemons in Peripheral Wars,” International Studies Quarterly 37, no. 4 (1993): 409–30; J. L. Snyder, Myths of Empire (Ithaca, N.Y.: Cornell University Press, 1991); Goldstein and Keohane, Ideas and Foreign Policy. 8. R. O. Keohane and L. L. Martin, “Institutional Theory as a Research Program,” in Progress in International Relations Theory: Metrics and Methods of Scientific Change, ed. C. Elman and M. F. Elman (Cambridge: MIT Press, forthcoming). 9. A. Wendt, “The Agent-Structure Problem in International Relations,” International Organization 41, no. 3 (1987): 335–70; A. Wendt, “Anarchy Is What States Make of It,” International Organization 46, no. 2 (1992): 391–425; Wendt, Social Theory of International Politics; Katzenstein, The Culture of National Security. 10. Goldstein and Keohane, Ideas and Foreign Policy, 8–11. 11. See also A. L. George, “The ‘Operational Code,’” International Studies Quarterly 13, no. 2 (1969): 190–222; George, “The Causal Nexus between Cognitive Beliefs and Decision-Making Behavior” M. L. Cottam, Foreign Policy Decision-Making: The Influence of Cognition (Boulder, Colo.: Westview, 1986); K. L. Shimko, Images and Arms Control (Ann Arbor: University of Michigan Press, 1991). 12. See also M. N. Barnett, Dialogues in Arab Politics (New York: Columbia University Press, 1998); S. G. Walker, M. Schafer, and M. D. Young, “Systematic Procedures for Operational Code Analysis,” International Studies Quarterly 45, no. 1 (1998): 175–190; S. G. Walker, M. Schafer, and M. D. Young, “Presidential Operational Codes and Foreign Policy Conflicts in the Post-Cold War World,” Journal of Conflict Resolution 43, no. 5 (1999): 610–25.

13. See also D. A. Sylvan and J. F. Voss, eds., Problem Representation in Foreign Policy Decision Making (Cambridge: Cambridge University Press, 1998); G. M. Bonham, M. J. Shapiro, and T. L. Trumble, “The October War,” International Studies Quarterly 23, no. 1 (1979): 3–44; R. M. Axelrod, ed., The Structure of Decision (Princeton, N.J.: Princeton University Press, 1976). 14. Wendt, “Anarchy Is What States Make of It.” 15. R. Jervis, “Leadership, Post-Cold War Politics, and Psychology,” Political Psychology 15, no. 4 (1994): 769–78. 16. D. P. Green and I. Shapiro, Pathologies of Rational Choice Theory (New Haven, Conn.: Yale University Press, 1994). 17. R. K. Ashley, “The Poverty of Neorealism,” International Organization 38, no. 2 (1984): 225–86; Wendt, Social Theory of International Politics; N. Geva and A. Mintz, eds., Decision-Making on War and Peace (Boulder, Colo.: Lynne Rienner, 1997). 18. The Rational Deterrence Debate: A Symposium, special issue of World Politics 41, no.2 (1989); Geva and Mintz, Decision-Making on War and Peace. 19. For example, R. Jervis, Perception and Misperception in International Politics (Princeton, N.J.: Princeton University Press, 1976); I. L. Janis, Groupthink, rev. ed. (Boston: Houghton Mifflin, 1982); D. A. Larson, Origins of Containment: A Psychological Explanation (Princeton, N.J.: Princeton University Press, 1985); Page 512 → Y. F. Khong, Analogies at War (Princeton, N.J.: Princeton University Press, 1992). 20. Y. I. Vertzberger, The World in Their Minds (Stanford, Calif.: Stanford University Press, 1990). 21. For example, B. Bueno de Mesquita and D. Lalman, War and Reason (New Haven, Conn.: Yale University Press, 1992); see also B. Bueno de Mesquita, Principles of International Politics (Washington, D.C.: Congressional Quarterly Press, 2000). 22. D. Little, Varieties of Social Explanation (Boulder, Colo.: Westview, 1991). 23. S. T. Fiske and S. E. Taylor, Social Cognition, 2d ed. (New York: McGraw-Hill, 1991). Whereas schemata are hypotheses about the environment, heuristics are the rules by which individuals test the hypotheses in a schema. By extension, heuristics may also be viewed as “content-free” schemata, that is, as rules functioning to organize incoming information even in the absence of person, self, role, or event schemata. See also Jervis, Perception and Misperception in International Politics; J. G. Stein, “Building Politics in Psychology,” Political Psychology 9, no. 2 (1988): 245–72. 24. P. Haas, “Introduction: Epistemic Communities and International Policy Coordination,” International Organization 46, no. 1 (1992): 1–35; Wendt, “Anarchy Is What States Make of It.” 25. B. R. Farnham, ed., Avoiding Losses/Taking Risks (Ann Arbor: University of Michigan Press, 1994); J. S. Levy, “Prospect Theory, Rational Choice, and International Relations,” International Studies Quarterly 41, no. 1 (1997): 87–112; R. McDermott, Risk-Taking in International Politics (Ann Arbor: University of Michigan Press, 1998). 26. R. K. Herrmann and M. P. Fischerkeller, “Beyond the Enemy Image and Spiral Model: CognitiveStrategic Research after the Cold War,” International Organization 49, no. 3 (1995): 415–50; Bueno de Mesquita and Lalman, War and Reason. 27. M. G. Hermann and C. F. Hermann, “Who Makes Foreign Policy Decisions and How?” International Studies Quarterly 33, no. 4 (1989): 361–88; A. Mintz, “The Decision to Attack Iraq,” Journal of Conflict Resolution 37, no. 4 (1993): 595–618; B. R. Farnham, Roosevelt and the Munich Crisis (Princeton, N.J.: Princeton University Press, 1997). 28. P. Hedstrom and R. Swedberg, eds., Social Mechanisms (Cambridge: Cambridge University Press, 1998). 29. C. W. Kegley Jr., “Decision Regimes and the Comparative Study of Foreign Policy,” in New Directions in the Study of Foreign Policy, ed. C. F. Hermann, C. W. Kegley Jr., and J. N. Rosenau (Boston: Allen and Unwin, 1987): 247–68; M. J. Shapiro, G. M. Bonham, and D. Heradsveit, “A Discursive Practices Approach to Collective Decision Making,” International Studies Quarterly 32, no. 4 (1988): 397–419; Vertzberger, The World in Their Minds. 30. Little, Varieties of Social Explanation. The distinction between these types of rational choice models evokes echoes of a much deeper schism over methodological commitments regarding “inside-out” and “outside-in” explanations and interpretations of foreign policy and world politics. See J. D. Singer, “The Level-of-Analysis Problem in International Relations,” in The International System: Theoretical Essays, ed.

K. E. Knorr (Princeton, N.J.: Princeton University Press, 1961), 77–92; K. N. Waltz, Theory of International Politics (Reading, Mass.: Addison-Wesley, 1979); C. Elman, “Horses for Courses?” Security Studies 6, no. 1 (1996): 7–53; Wendt, “The Agent-Structure Problem in International Relations” D. Dessler, “What's at Stake in the Agency-Structure Debate,” International Organization 43, no. 3 (1989): 441–73; W. Carlsnaes, “The Agency- Structure Debate in Foreign Policy Analysis,” International Studies Quarterly 36, no. 3 (1992): 245–70; R. L. Doty, “Aporia: A Critical Exploration of the Agent-Structure Problematique in International Relations Theory,” European Journal of International Relations 3, no. 3 (1997): 365–92. 31. For example, A. L. George, D. K. Hall, and W. E. Simons, Limits of Coercive Diplomacy (Boston: Little, Brown, 1971); A. L. George and R. Smoke, Deterrence Page 513 → in American Foreign Policy (New York: Columbia University Press, 1974); R. N. Lebow and J. G. Stein, “Beyond Deterrence,” Journal of Social Issues 43, no. 4 (1987): 5–71; R. N. Lebow and J. G. Stein, “Deterrence: The Elusive Dependent Variable,” World Politics 42, no. 3 (1990): 336–69; P. K. Huth and B. M. Russett, “Testing Deterrence Theory,” World Politics 42, no. 4 (1990): 466–501; R. J. Leng, Interstate Crisis Behavior, 1815–1990 (New York: Cambridge University Press, 1993). 32. N. R. Richardson, “Dyadic Case Studies in the Comparative Study of Foreign Policy Behavior,” in Hermann, Kegley Jr., and Rosenau, New Directions in the Study of Foreign Policy, 161–77; J. Lepgold and A. C. Lamborn, “Locating Bridges: Connecting Research Agendas on Cognition and Strategic Choice,” International Studies Review 3, no. 3 (2002): 3–30. 33. For example, R. Powell, In the Shadow of Power (Princeton, N.J.: Princeton University Press, 1999); R. L. Schweller, Deadly Imbalances (New York: Columbia University Press, 1998); R. O. Keohane, After Hegemony (Princeton, N.J.: Princeton University Press, 1984). 34. M. Finnemore and K. Sikkink, “International Norm Dynamics and Political Change,” International Organization 4, no. 4 (1998): 887–918; Wendt, Social Theory of International Politics; Katzenstein, The Culture of National Security. 35. J. S. Levy, “Learning and Foreign Policy,” International Organization 48, no. 2 (1994): 279–312. 36. J. A. Robinson, “Crisis: An Appraisal of Concepts and Theories,” in Hermann, International Crises, 20–38; C. A. McClelland, “The Beginning, Duration, and Abatement of International Crises,” in ibid., 83–108; M. Brecher, J. Wilkenfeld, and S. Moser, Crises in the Twentieth Century (New York: Pergamon, 1988). 37. Snyder and Diesing, Conflict among Nations; Bueno de Mesquita and Lalman, War and Reason; Leng, Interstate Crisis Behavior, 1815–1990. 38. Brecher, Wilkenfeld, and Moser (Crises in the Twentieth Century) are a partial exception to this generalization. They recognize the possibility of a crisis as a turning point in the transition from war to peace. However, they do not extend this logic and conceptualize it as applying to the escalation of cooperation as well as the de-escalation of conflict. 39. Bueno de Mesquita and Lalman, War and Reason. 40. Ibid., 30–35. 41. Attempts to close this gap in varying degrees include R. M. Axelrod, The Evolution of Cooperation (New York: Basic Books, 1984); J. S. Goldstein and J. R. Freeman, Three-Way Street (Chicago: University of Chicago Press, 1990); G. W. Breslauer and P. E. Tetlock, Learning in U.S. and Soviet Foreign Policy (Boulder, Colo.: Westview, 1991); Bueno de Mesquita and Lalman, War and Reason; Leng, Interstate Crisis Behavior, 1815–1990; Herrmann and Fischerkeller, “Beyond the Enemy Image and Spiral Model” Lepgold and Lamborn, “Cognitive Processes and Strategic Choice” R. J. Leng, Bargaining and Learning in Recurring Crises (Ann Arbor: University of Michigan Press, 2000); D. A. Larson, Anatomy of Mistrust (Ithaca, N.Y.: Cornell University Press, 2000). 42. J. Mercer, “Anarchy and Identity,” International Organization 49 (1995): 229–52; S. Stryker and A. Statham, “Symbolic Interaction and Role Theory,” in Handbook of Social Psychology, vol. 1, 3d ed., ed. G. Lindzey and E. Aronson (New York: Random House, 1985), 311–78; Wendt, Social Theory of International Politics; Farnham, Avoiding Losses/Taking Risks; C. Taber, “POLI:An Expert System Model of U.S. Foreign Policy Belief Systems,” American Political Science Review 86, no. 4 (1992): 888–904; R. Little and S. Smith, Belief Systems and International Relations (New York: Blackwell, 1988); Walker, Schafer, and Young, “Systematic Procedures for Operational Code Analysis” Walker, Schafer, and Young, “Presidential Operational Codes and Foreign Policy Conflicts.”

43. Holsti, “Foreign Policy Decisions Viewed Cognitively.” 44. M. N. Barnett, “Institutions, Roles, and Disorder,” International Studies Quarterly 37, no. 3 (1993): 271–96; Wendt, Social Theory of International Politics.

Page 514 →

PUBLIC OPINION AND FOREIGN POLICY ANALYSIS Where We Were, Are, and Should Strive to Be Ole R. Holsti This essay examines perspectives on public opinion and foreign policy, first during the formative years of the International Studies Association and, second, at the beginning of the new millennium. Because of limited space, the essay necessarily focuses on some main currents of theory and research at the expense of drawing subtle distinctions between various perspectives on the subject. The concluding section identifies some areas that should be near or at the top of our research agendas during the coming years.

Where We Were When the International Studies Association was founded in 1958, the cold war was a dominant fact of international relations. To be sure, Stalin had died in 1953 and the next few years had witnessed the end of wars in Korea and Indochina, the Austrian State Treaty, and the Geneva summit conference, but the Soviet invasion of Hungary, the Berlin “deadline crisis,” and the collapse of the Paris summit conference suggested that the period of détente in the mid-1950s was but a brief interlude rather than a major turning point in the cold war. The dominant view of public opinion during that period can be stated succinctly. Ill-informed, emotion-driven, shortsighted, and self-absorbed publics constitute a major obstacle that threatens to prevent the United States and other Western democracies from playing a constructive role in an increasingly dangerous international system. This dolorous view of U.S. public opinion drew from and melded two quite different intellectual traditions, one European and the second American: classical realist political philosophy dating back to Thucydides, Machiavelli, and de Tocqueville, and the new “science” of public opinion polling that had emerged only during the decade prior to World War II. The primary contributors to twentieth-century realism as a theory of international politics were Europeans, including British diplomat-historian E. H. Carr and German political theorist Hans J. Morgenthau. Despite some Page 515 → differences among realists, they shared a very jaundiced view of public opinion. Carr reserved some of his most vitriolic words for skewering Woodrow Wilson and other liberal statesmen who had attributed to the general public both the interest and wisdom to serve as an effective constraint on the war-making proclivities of leaders. Woodrow Wilson's “plain men throughout the world,” the spokesmen “the common purpose of enlightened mankind,” had somehow transformed themselves into a disorderly mob emitting incoherent and unhelpful noises. It seemed undeniable that, in international affairs, public opinion was almost as often wrong-headed as it was impotent…. The breakdown of the post-War utopia is too overwhelming to be explained merely in terms of individual action or inaction. Its downfall involves the bankruptcy of the postulates [about public opinion] on which it is based.1 Although the intellectual roots of realism were largely European, it also flowered on campuses and elsewhere in postwar America. During the 1950s, diplomat-historian George F. Kennan and syndicated columnist Walter Lippmann, who had written two important books on public opinion during the 1920s, led the chorus of realist analysts who despaired of the ability of democracies to conduct effective foreign policies as long as publics, and their representatives in legislatures, had a significant voice in policy-making.2 Writing in the middle of the decade, Lippmann was especially somber about the prospects for the democracies. The unhappy truth is that the prevailing public opinion has been destructively wrong at the critical junctures. The people have impressed a critical veto upon the judgments of informed and responsible officials…. Mass opinion has acquired mounting power in this country. It has shown itself to be a

dangerous master of decision when the stakes are life and death.3

The development of scientific polling was the second primary contributor to the dominant skeptical view of public opinion. Whereas the realist perspective drew heavily on a long European tradition, many of the initial advances in survey research were made by American pollsters such as George Gallup and social scientists in several disciplines. During World War II and the years immediately following that conflict, a good deal of research focused on various facets of a single question: Would the United States play an active role in the postwar world, or would it eschew a leadership role, turning most of its attention and energies inward, as it had done following World War I? The pioneering students of public opinion on international issues—including but not limited to historian Thomas A. Bailey, sociologist Martin Kriesberg, and political scientists Gabriel A. Almond and James N. Rosenau—largely reinforced the skeptical perspectives of the realists.4 Their Page 516 → analyses of survey data generated by Gallup and other major polling organizations led them to conclude not only that the American public is indifferent to and very poorly informed about world affairs but also that whatever opinions they have are so poorly rooted in any broader conceptions about international affairs that they are subject to wide, emotion-driven mood swings. According to Almond, “for persons responsible for the making of security policy these mood impacts of the mass public have a highly irrational effect. Often public opinion is apathetic when it should be concerned, and panicky when it should be calm.”5 The proposition that public opinion lacks coherence and structure was reinforced by several pathbreaking studies of voters and the bases of their electoral decisions. In an essay drawing upon these voting studies, Philip Converse found that the political beliefs of the general public lack any “constraint” or underlying ideological consistency that might provide structure or coherence to political thinking and voting.6 His analyses led him to conclude that mass political beliefs are best described as “non-attitudes.” The worst nightmares of those who feared that the public would force the United States to withdraw into a mindless isolationism did not, however, come to pass. United Nations membership, the Truman Doctrine, the Marshall Plan, NATO membership, and resistance to aggression in Korea, sometimes described as “the revolution in American foreign policy,” would appear to have posed a puzzle for those most fearful of a powerful but feckless and isolationist public opinion. One answer seemed to emerge from a study of survey data by William Caspary.7 He concluded that contrary to Almond and Lippmann, the public has in fact been quite “permissive” toward a wide range of international undertakings, including some—notably the Vietnam War, according to Caspary—that had a rather tenuous relationship to vital national interests. A larger-scale study by Bernard Cohen, drawing heavily on interviews with foreign policy officials, came to a similar conclusion.8 Finding the evidence of public constraints on foreign policy decision makers to be less than compelling, Cohen challenged diplomatic historians and foreign policy analysts to demonstrate rather than merely assert that public opinion in fact has any significant impact on policy; he even cast doubts on the conventional wisdom that public opinion establishes “limits” or boundaries beyond which any administration would venture only at its own electoral peril. Moreover, an important study of Congress indicated that the “electoral retribution” hypothesis was more valid for domestic than foreign policy issues; legislators had considerably greater leeway on the latter than the former.9 Thus, although the view of the public as apathetic, ill-informed, and incapable of holding coherent views on world affairs had not materially changed by the early 1970s, the more apocalyptic fears expressed by Almond, Lippmann, and other realists had been assuaged by the sense that, in fact, public opinion was at best a residual category in explaining foreign policy. Even analysts with doubts about the realist depiction of the state as a “unitary actor” turned their attention elsewhere within the domestic political Page 517 → arena—for example, legislatures, interest groups, organizational processes, bureaucratic politics, and the like. If the impact of public opinion on foreign policy ranges from little to none and from rarely to never, why devote much attention to it? Thus, public opinion had become something of a backwater in foreign policy analysis.

Where We Are Just as World War II generated interest in and research about the nature of postwar public opinion, the Vietnam War served as the catalyst for reexamination of the prevailing view of public opinion as volatile, lacking structure, and largely irrelevant in the conduct of foreign affairs. As the costs of the war increased and the prospects for victory faded, the prescription espoused by Bailey, Almond, Kennan, Lippmann, and countless other realists—unshackle foreign policy decision makers from the dangerous grip of public opinion—came under increasing question. The decision makers responsible for American policy in Vietnam, derisively described by journalist David Halberstam as The Best and the Brightest, now seemed somewhat less than Olympian in their vision of the national interest.10 Ironically, Hans Morgenthau was transformed from the towering theorist of realism to a public figure by his frequent speeches aimed at influencing the public against U.S. policy in Vietnam, and even Walter Lippmann came to see draft card burners as more enlightened on the war than President Johnson. The Vietnam War also gave rise to a number of new survey projects on American foreign policy that attempted to go beyond some of the questions that had been the staples of existing polling organizations: internationalism versus isolationism, judgments of presidential performance, and levels of support for specific foreign policy undertakings. During the late 1960s the Verba-Stanford project probed deeply into opinions on the Vietnam War;11 the Chicago Council on Foreign Relations initiated quadrennial surveys on many aspects of American foreign policy in 1974;12 and two years later the Foreign Policy Leadership Project undertook the first of six surveys on the foreign and domestic policy views of opinion leaders.13 As a result of these projects, as well as extensive analyses of growing survey archives generated by the Gallup Organization, the National Opinion Research Center, and many others, public opinion analysts have mounted serious challenges to the views that had prevailed three decades ago. The previously cited study by Caspary was among the first to describe public opinion, at least through 1953, as stable rather than volatile.14 A fuller and more systematic analysis of public opinion toward the wars in Korea and Vietnam posed another challenge to the thesis of irrational mood swings among the general public. Mueller found that public support for the American war effort in both conflicts eventually declined, but it did so in ways that seemed explicable and rational rather than random and mindless.15 Increasing public opposition to U.S. participation in both conflicts followed a pattern that matched rising battle deaths, suggesting that the public used an understandable, if simple, rule of thumb to assess American policy. Page 518 → The most comprehensive challenge to the thesis that public opinion is mindlessly volatile emerged from a series of studies by Benjamin Page and Robert Shapiro.16 Using a database that included more than six thousand questions that had been posed by major polling organizations over a half century beginning in the mid-1930s, their analyses focused on the 20 percent of those that had been asked at least twice. Page and Shapiro found that public opinion in the aggregate is in fact characterized by a good deal of stability, no less on foreign policy than on domestic issues. Most important, when changes (defined as a difference of 6 or more percent from one survey to the next) took place, they appeared to be “reasonable, events-driven” reactions to the real world, even if the factual information upon which they were based was marginally adequate at best: “In particular, most abrupt foreign policy changes took place in connection with wars, confrontations, or crises in which major changes in the actions of the United States or other nations quite naturally affect preferences about what policies to pursue.”17 Similar conclusions about the importance of external events emerged from a study of the period spanning the Kennedy through the Reagan administrations.18 An interesting variant of the “rational public” thesis stipulates that public opinion serves as a counterweight against administration policies. Miroslav Nincic found public support for a conciliatory stance toward the Soviet Union during the hawkish Reagan administration while supporting more assertive policies when the more dovish Jimmy Carter was president.19 Similar findings emerged from another study.20 For at least the periods covered by Nincic and Mayer, these results not only challenged the thesis of volatility in public attitudes but turned that proposition on its head because the public was identified as a source of moderation and continuity rather than instability and unpredictability.

Philip Converse's chapter on the lack of coherence and structure in mass belief systems is a widely cited classic; it has also served as the springboard for a plethora of further methodological and empirical studies.21 That literature is far too vast to summarize here. Although the more recent research has yet to achieve a consensus on all aspects of the question, there does appear to be some convergence on two points of relevance to foreign policy analysis. First, even though the general public is rather poorly informed about the factual aspects of international relations—if any study has seriously challenged that generalization, it has escaped my attention—attitudes about foreign affairs are in fact structured in moderately coherent ways. Indeed, low information and an ambiguous international environment may actually motivate rather than preclude having some attitude structure. For example, the concept of “cognitive miser” suggests that individuals may use a limited number of beliefs to make sense of a wide range of facts and events. Second, there is growing evidence that the single internationalist-to-isolationist dimension that dominated much of the early research is inadequate to categorize beliefs about international affairs. An analysis of the first Chicago Council on Foreign Relations (CCFR) survey of elites and the general public used factor analysis and other statistical Page 519 → techniques to identify three distinct perspectives on international affairs: conservative internationalism, liberal internationalism, and noninter-nationalism.22 An important further step on ways of describing foreign policy attitudes emerged from Eugene Wittkopf's secondary analyses of the four CCFR surveys conducted between 1974 and 1986.23 His results identified “two faces of internationalism”: militant internationalism (MI) and cooperative internationalism (CI). Dichotomizing and crossing these dimensions yields four foreign policy belief systems: hard-liners (support MI, oppose CI), accommodationists (oppose MI, support CI), internationalists (support both MI and CI), and isolationists (oppose both MI and CI). Strong support for the MI/CI scheme has also emerged from studies of opinion leaders, although they used different sampling designs and questions to construct the MI and CI scales.24 It is worth noting that the two dimensions of the MI/CI scheme correspond rather closely to the most venerable theoretical perspectives on international relations. Realist concepts such as security dilemma, relative capabilities, and a zero-sum view of conflict are also basic to the MI dimension. There are comparable similarities between the CI dimension and liberalism, including a view of security that encompasses more than geopolitical and military balances; the potential for institution building to reduce uncertainty and fears of perfidy, as well as information costs; the possibilities of improved international education and communications to ameliorate fears; and the positive-sum outcomes that may result from such activities as trade. A further distinction between unilateralism and multilateralism as an important aspect of public opinion has also emerged from several studies.25 Because many of the debates about post-cold war activities, including military interventions abroad, have revolved around the relative merits of “going it alone” versus “burden sharing,” there is reason to believe that this distinction will become more rather than less important in describing public opinion. A somewhat different way of describing the structure of attitudes on foreign affairs emerged from a study by Jon Hurwitz and Mark Peffley.26 They proposed and tested a structure of foreign policy beliefs in which specific policy preferences are derived from broader postures—for example, militarism or isolationism—which are, in turn, constrained by a set of core values relating to international affairs. Thus, a few rather general beliefs, such as a “tough-minded” approach to international affairs, appear to have served as organizing devices that enable subjects to respond in a reasonably coherent manner to a wide range of issues, including defense spending, nuclear arms policy, policies toward the Soviet Union, and international trade. Clearly, the research of the past several decades has yet to produce a consensus on many important questions about the coherence and substance of foreign policy beliefs among the general public. Nor does any of it suggest that the ill-informed publics of yesteryear have been transformed into knowledgeable and sophisticated analysts of international affairs. Nevertheless, it is also clear that the earlier consensus depicting public attitudes as Page 520 → lacking any real coherence has been challenged from many quarters, only a few of which have been touched upon here. A good deal of credible evidence suggests that members of the general public use various rules of thumb, not necessarily confined to the traditional internationalist or isolationist blueprints, for organizing their political thinking. These rules of thumb may often appear to be somewhat simplistic, but it should be borne in mind that even top-ranking decision makers have been known to use such simplifying adages as “the lessons of Munich,” “the domino theory,” and “the lessons of Vietnam” to help them diagnose and cope with the

complexities of international affairs. Among the most important questions about public opinion are these: To what extent, on what kinds of issues, under what circumstances, and in what types of political systems, if any, does it have an impact on public policy? If it does have an influence, what are the means by which public attitudes become known so that they can have an impact on decision makers? As these are also the most difficult questions to answer, it is not surprising that we have many analyses describing the state of, or trends in, public opinion and far fewer compelling studies tracing out its impact on the foreign policy process. There is quite a bit of anecdotal evidence that many presidents have followed Theodore Roosevelt in thinking of the White House as a bully pulpit from which to convey diagnoses of world affairs and to drum up support for preferred policies. Despite the advantages that presidents enjoy in this respect, they have also differed in the degree of success that they have enjoyed. Although President Reagan, “the great communicator,” was unable to build a groundswell of public approval for his efforts to intervene in the civil wars in El Salvador and Nicaragua, his less verbally gifted successor, George Bush, generally gained strong support for his foreign policies. At least one noted public opinion analyst has asserted that the Vietnam War marked a major watershed because the public is no longer willing to accept the proposition that “the president knows best,”27 and some subsequent studies have provided supporting evidence for that proposition.28 The more difficult question concerns public influence on policymakers. Again there is anecdotal evidence of presidential concern about the state of public opinion. Franklin Roosevelt seemed fascinated by the possibilities of polling the public, and he secretly commissioned Hadley Cantril to provide him with regular reports on such questions as likely public attitudes toward America's appropriate international role after the guns of World War II had stopped firing. In contrast, his successor, Harry S. Truman, publicly expressed disdain for pollsters and the value of evidence that they might provide—including the confident prognoses by Gallup and other pollsters that Thomas Dewey would win the 1948 presidential election by a landslide. In more recent administrations, polling has become institutionalized. For example, the Reagan White House privately commissioned at least 184 surveys, an average of almost two per month.29 We have relatively few “insider” accounts from opinion analysts who served policymakers, but those that exist suggest that the general public isPage 521 → not viewed merely as an essentially shapeless lump that can readily be molded by public relations activities or a compliant media to meet the immediate policy needs.30 According to Beal and Hinckley, “opinion polls are at the core of presidential decision making.”31 There is also evidence, however, that recent presidents have become less responsive to public preferences; polling data have often been used to develop strategies for leading or manipulating the public rather than as guidance on policy decisions.32 In addition to anecdotal evidence, three methodologies have been employed to assess the public-policymaker relationship: quantitative-correlational analyses, intensive case studies, and interview-based research. Although the evidence has not given rise to a consensus, a significant part of it has raised questions about the universal validity of the proposition that foreign policy-makers and processes are impervious to public influences. The conventional wisdom that voting behavior is grounded in domestic rather than international issues—in Bill Clinton's phrase, “It's the economy, stupid”—was challenged by Aldrich, Sullivan, and Borgida in a systematic study of presidential campaigns between 1952 and 1984.33 Their analyses revealed that in five of the nine elections (1952, 1964, 1972, 1980, and 1984) foreign policy issues had “large effects.” Other studies of voting behavior have shown the importance of retrospective evaluations of performance, including the conduct of foreign policy, on voter choice, especially when the incumbent is a candidate.34 Two additional quantitative studies have measured the congruence between changes in public opinion and a broad range of policies over extended periods. The first, an analysis of public opinion and policy outcomes spanning the years 1960 to 1974, revealed that in almost two-thirds of 222 cases, policy outcomes corresponded to public preferences. The consistency was especially high (92 percent) on foreign policy issues.35 The second study of the opinion-policy relationship covered an even longer period (1935 to 1979), which included 357 significant changes

of public preferences.36 The results also revealed a high degree of congruence between opinions and foreign policy (62 percent); the figure for domestic policy was only slightly higher (70 percent). These findings are impressive, but even the best correlational study cannot wholly dispel all doubts about causality; the possibilities for erroneous interpretation of even impressive correlation coefficients are well known, especially when there is reason to wonder whether a simple, one-directional model of influence is sufficient to capture the complexities of the foreign policy-making process. One of the authors of these studies37 has acknowledged that “not even time-series analysis can provide a magic bullet that will kill all the demons of causal ambiguity.” The problem of causality, as V. O. Key pointed out, is this: “If one is to know what opinions government heed, one must know the inner thoughts of presidents, congressmen, and other officials.”38 For this reason, intensive case studies using archival research, elite interviews, or a combination of both are invaluable to supplement quantitative-correlational research. Page 522 → Not surprisingly, existing case studies have yielded somewhat mixed results. For example, Doris Graber found in an analysis of four cases during the Adams, Jefferson, Madison, and Monroe administrations that in each instance public opinion was “an important factor in decision making, but by no means the most important single factor.”39 In contrast, in a study encompassing the quarter century of the McKinley, Roosevelt, Taft, and Wilson presidencies, Hildebrand revealed that if public opinion even entered into executive decisions, it was only after policy decisions had already been made.40 Some recent case studies suggest that the impact of public opinion varies according to the type of issue, the stage in the policy process, the decision context, and policymakers' beliefs about and sensitivity to public opinion. Thomas Graham analyzed the impact of public opinion on four arms control issues—international control of atomic energy, the Limited Test Ban Treaty, the SALT I/ABM Treaties, and SALT II—spanning seven American administrations from presidents Truman through Carter.41 He distinguished between four stages in the policy process: getting the issue on the agenda, negotiations, ratification, and implementation. Although public opinion had an important impact in all four cases, it also varied according to the stage in the policy process. Public opinion had a direct impact on getting the issue on the agenda and on ratification of the agreements, but only an indirect effect on negotiations and implementation. Archival research on four defense and foreign policy episodes during the Eisenhower administration confirmed that distinctions between stages in the policy process as well as the decision context are important for understanding the ways in which public opinion may affect the policy process.42 Graham's cases included four decision contexts: crisis (the 1954 Formosa Straits confrontation with China), reflexive (the French defeat at Dien Bien Phu), innovative (response to the Soviet launch of the Sputnik space satellite), and deliberative (the “New Look” defense policy). The impact of public opinion varied across the four cases. Lawrence Jacobs and Robert Shapiro have undertaken the most ambitious analysis of the impact of public opinion in their study of the relationship between private presidential polling and presidential behavior in administrations since that of John Kennedy.43 In addition to archival research, they have interviewed top White House personnel as well as those engaged in conducting the surveys. By combining analyses of survey data, archival evidence, and interviews, their research design represents a major step toward providing a more detailed and nuanced answer to the questions posed at the beginning of this section. This description of the current state of theory and research on public opinion and foreign policy is necessarily cursory rather than comprehensive; it blurs rather than highlights some distinctions and controversies, and it largely overlooks methodological issues except at the very broadest level. It has attempted to convey the sea change in thinking about public opinion and foreign policy during the period since the founding of the International Page 523 → Studies Association. While there appears to have been considerable theoretical, empirical, and methodological progress during that time, there are still major steps that should appear on the agenda for future research.

Where We Should Strive to Be According to one hypothesis, the end of the cold war, the declining salience of foreign policy for much of the public, and the complexity of post-cold war issues will combine to reduce the impact of public opinion, thereby giving policymakers greater latitude in making decisions.44 An alternative line of reasoning suggests that the influence of public opinion in the conduct of foreign affairs may well increase during the post-cold war era. The case against public opinion is that effective diplomacy requires at least three features, none of which is enhanced by more active public participation: secrecy, speed, and flexibility; these are also most likely to be essential in times of war, crisis, and confrontation. Although security concerns remain vital, the top tiers of post-cold war foreign policy agendas are also likely to include issues on which publics and their representatives are likely to hold strong views and regarding which the thesis that “the president [or prime minister, chancellor, etc.] knows best” may seem less compelling than during World War II or the cold war. Many of these “intermestic” issues also cut across conventional international-domestic boundaries. They include trade, refugees and immigration, the international environment, and intrastate religious, ethnic, racial, and nationalist conflicts in which one or more parties may call for assistance or intervention from abroad. Both of these hypotheses may be valid, depending on the administration, issue, and circumstances, but whatever the case, it makes sense to consider ways of improving our understanding of public opinion and its relationship to foreign policy. As noted earlier, the opinion-policy link is less well-developed than other areas of inquiry. Although there have been important studies of this type in recent years, some questions that merit further consideration include these: Upon what indicators of public sentiments, if any, did leaders rely? Did leaders use information about public preferences primarily in order to align policies with them, or in order to develop strategies for influencing or manipulating the public? Did policymakers rule out certain courses of action because of a belief that a lack of public support would reduce or eliminate the prospects of success? Did policymakers decide upon certain foreign policy undertakings, even if the chances of success were deemed very slight, because of a belief that the public demanded some action? How, if at all, did expectations of future public reactions affect the appraisal of policy options by policymakers? Page 524 → How, if at all, did calculations about the electoral or other domestic political consequences of certain decisions restrain or motivate policymakers? Were the timing of foreign policy decisions and the choice of means to carry out an undertaking influenced by beliefs about what the public would or would not accept, would or would not demand? How accurately did policymakers gauge public sentiments? In cases of policymakers “misreading the public,”45 what were the reasons for the miscalculation? Did motivated or unmotivated biases enter into misreading public preferences? Cross-National Studies Cross-national studies constitute a second major area for further research. The role of public opinion in foreign policy is often a central issue in the flourishing contemporary debates about the “democratic peace” theory—the proposition that democracies do not go to war against each other. Even among democracies, the nature and impact of public opinion may vary across countries and political systems. Moreover, the expansion of democracies during the past decade has created opportunities for research on a much wider range of issues than could have been contemplated during the cold war in Eastern Europe, the republics of the former Soviet Union, Latin America, and

elsewhere. Such opportunities are even becoming available in China. Standard Questions Standard questions are also a high priority. There is ample evidence that the wording of questions and the context in which they are posed can significantly affect responses. Even minor variations in questions about a single issue, asked at the same time, can yield quite different responses. Consequently, drawing inferences from responses to a single question can be hazardous, but one can do so with greater confidence if identically worded questions appear regularly in surveys conducted over an extended period. Unfortunately, relatively few questions about foreign and defense policy have been posed with the same frequency or regularity as, for example, those asking about presidential performance in the United States. Even a question about the merits of an active U.S. role in the world, first used in several surveys during World War II, was asked only once during the 1960s. One of the many important contributions of the seven Chicago Council on Foreign Relations surveys since 1974 is that there has been a carryover of some key questions from survey to survey, providing opportunities for assessing trends. Richard Sobel has proposed a list of thirty questions for polling on foreign policy crises, but progress toward developing a list of standard questions has lagged.46 Cross-national research will present some special challenges in this respect because it cannot be assumed that simple word-for-word translations of questions from one language and culture will necessarily convey the same meaning elsewhere. Page 525 →

Conclusion Has research on public opinion and foreign policy fulfilled the hopes of the late 1950s? The question does not seem germane because at that time the subject did not seem especially interesting and thus hopes were not very high. The fears that a misguided public would prevent wiser leaders from pursuing policies needed to protect and promote vital national interests, while also avoiding a nuclear war, had begun to recede, at least in part because of the sense that public opinion was not all that important in coping with the crucial issues of the day. The next four decades witnessed significant challenges to the previous views, although a new consensus has yet to be achieved on many significant questions about the nature, structure, and impact of public opinion. No doubt the lack of consensus is one reason that the subject remains a vibrant area of inquiry. Today the general public is less frequently described in the frightening terms that dominated thinking during the 1950s. One of the classic studies of recent years is entitled “The Rational Public,” and another has coined the term “low information rationality” to describe the manner by which the public deals with a complex world.47 Moreover, opportunities for gaining information about the world have never been greater. Cheap international travel, dramatically rising graduation rates from secondary schools and universities, and new information technologies that even dedicated science fiction fans of the 1950s could hardly have imagined have contributed to these opportunities. Indeed, a distinguished political scientist, a former president of the International Studies Association, and onetime skeptic about the role that the public could play in foreign affairs, has written that the electronic information revolution has provided general publics with the information and analytical skills necessary to become vital players in global affairs.48 Nevertheless, there are also reasons for disquiet. The possibility that increasing survey data will merely be used by leaders to manipulate the public has already been mentioned. A closely related concern is the persistent poverty of international knowledge. Many studies reveal that levels of information about foreign affairs among publics in the industrial world remain abysmally low, most notably in the United States. One disturbing example illustrates the point. Gallup surveys commissioned by the National Geographic Society in 1948 and 1988 revealed that basic geographic knowledge—for example, identification of the largest country in the world, or location of Great Britain on a map—has actually declined during the forty-year interval. Considering that geography has virtually dropped out of school curricula and that many universities no longer house geography departments, perhaps those troubling results are not especially surprising. Only 10 percent of Americans could identify any of the nations involved in post-cold war expansion of NATO, and no doubt even some of those would not have been able to

locate Hungary, Poland, and the Czech Republic on a map. In those circumstances, could there have been a principled debate in the United States on adding new members to NATO? Page 526 → There are many reasons to applaud the more benign and realistic view of the general public that has emerged from several decades of research, but it is hardly a time for complacency, especially among educators. At what point does “low information rationality” become “no information irrationality”? Perhaps preventing the former from becoming the latter should be the most important item on our agenda concerning public opinion and foreign policy.

Notes 1. Edward Hallett Carr, The Twenty Years' Crisis, 1919–1939: An Introduction to the Study of International Relations (London: Macmillan, 1939), 50–51, 53. 2. George F. Kennan, American Diplomacy, 1900–1950 (New York: Mentor Books, 1951); Walter Lippmann, Essays in the Public Philosophy (Boston: Little, Brown, 1955). 3. Lippmann, Essays in the Public Philosophy, 20. 4. Thomas A. Bailey, The Man in the Street: The Impact of American Public Opinion on Foreign Policy (New York: Macmillan, 1948); Martin Kriesberg, “Dark Areas of Ignorance,” in Public Opinion and Foreign Policy, ed. Lester Markel (New York: Harper, 1949); Gabriel A. Almond, The American People and Foreign Policy (New York: Harcourt, Brace, 1950; reprint, with a new introduction, New York: Praeger, 1960); James N. Rosenau, Public Opinion and Foreign Policy: An Operational Formulation (New York: Random House, 1961). 5. Gabriel Almond, “Public Opinion and National Security,” Public Opinion Quarterly 21, no. 2 (1956): 371–78. 6. Philip Converse, “The Nature of Belief Systems in Mass Publics,” in Ideology and Discontent, ed. David E. Apter (New York: Free Press, 1964), 206–61. 7. William R. Caspary, “The ‘Mood Theory’: A Study of Public Opinion and Foreign Policy,” American Political Science Review 64, no. 2 (1970): 536–47. 8. Bernard C. Cohen, The Public's Impact on Foreign Policy (Boston: Little, Brown, 1973). 9. Warren E. Miller and Donald E. Stokes, “Constituency Influence in Congress,” American Political Science Review 57, no. 1 (1963): 45–56. 10. David Halberstam, The Best and the Brightest (New York: Random House, 1972). 11. Sidney Verba et al., “Public Opinion and the War in Vietnam,” American Political Science Review 61, no. 2 (1967): 317–33. 12. John E. Rielly, ed., American Public Opinion and U.S. Foreign Policy, 1975 (Chicago: Chicago Council on Foreign Relations, 1975). Also similarly titled monographs in 1979, 1983, 1987, 1991, 1995, and 1999. A report on the 2002 CCFR survey will be available in late 2002. 13. Ole R. Holsti and James N. Rosenau, American Leadership in World Affairs: Vietnam and the Breakdown of Consensus (London: Allen and Unwin, 1984). 14. Caspary, “The ‘Mood Theory.’” 15. John E. Mueller, War, Presidents, and Public Opinion (New York: Wiley, 1973). 16. Benjamin I. Page and Robert Y. Shapiro, The Rational Public: Fifty Years of Trends in Americans' Policy Preferences (Chicago: University of Chicago Press, 1992); Benjamin I. Page and Robert Y. Shapiro, “Changes in Americans' Policy Preferences, 1935–1979,” Public Opinion Quarterly 46 (1982): 24–42. 17. Page and Shapiro, “Changes in Americans' Policy Preferences, 1935–1979,” 34. 18. William G. Mayer, The Changing American Mind: How and Why American Public Opinion Changed between 1960 and 1988 (Ann Arbor: University of Michigan Press, 1992).

19. Miroslav Nincic, “The United States, the Soviet Union, and the Politics of Opposites,” World Politics 40, no. 4 (1988): 452–75; Miroslav Nincic, Democracy and Foreign Policy: The Fallacy of Political Realism (New York: Columbia University Press, 1992).Page 527 → 20. Mayer, The Changing American Mind. 21. Converse, “The Nature of Belief Systems in Mass Publics.” 22. Michael Mandelbaum and William Schneider, “The New Internationalism,” in Eagle Entangled: U.S. Foreign Policy in a Complex World, ed. Kenneth A. Oye, Robert J. Lieber, and Donald Rothchild (New York: Longman, 1979), 40–63. 23. Eugene R. Wittkopf, Faces of Internationalism: Public Opinion and American Foreign Policy (Durham, N.C.: Duke University Press, 1990). 24. Ole R. Holsti and James N. Rosenau, “The Structure of Foreign Policy Attitudes among American Leaders,” Journal of Politics 52 (February 1990): 94–125; Ole R. Holsti and James N. Rosenau, “The Structure of Foreign Policy Beliefs among American Opinion Leaders—After the Cold War,” Millennium 22 (summer 1993): 235–78. 25. Eugene R. Wittkopf, “On the Foreign Policy Beliefs of the American People: A Critique and Some Evidence,” International Studies Quarterly 30, no. 4 (1986): 425–45; Ronald H. Hinckley, “Public Attitudes toward Key Foreign Policy Events,” Journal of Conflict Resolution 32, no. 2 (1988): 295–318; William O. Chittick, Keith R. Billingsley, and Rick Travis, “A Three-Dimensional Model of American Foreign Policy Beliefs,” International Studies Quarterly 39, no. 3 (1995): 313–31; Bruce M. Russett, Controlling the Sword: The Democratic Governance of National Security (Cambridge: Harvard University Press, 1990). 26. Jon Hurwitz and Mark Peffley, “How Are Foreign Policy Attitudes Structured? A Hierarchical Model,” American Political Science Review 81, no. 4 (1987): 1099–1120. 27. Daniel Yankelovich, “Farewell to ‘the President Knows Best,’” Foreign Affairs 57, special issue (1979): 670–93. 28. Michael Clough, “Grass-Roots Policymaking,” Foreign Affairs 73 (January-February 1994): 2–7; Maxine Isaacs, “The Independent American Public: The Relationship between Elite and Mass Opinions on American Foreign Policy in the Mass Communication Age” (Ph.D. diss., School of Public Affairs, University of Maryland, 1994). 29. Robert Y. Shapiro and Lawrence R. Jacobs, “Public Opinion, Foreign Policy, and Democracy” (paper prepared for the Conference on Polls, Policy, and the Future of American Democracy, Northwestern University, Evanston, Ill., May 13, 2000). 30. Hadley Cantril, The Human Dimension: Experiences in Policy Research (New Brunswick, N.J.: Rutgers University Press, 1967); Ronald H. Hinckley, People, Polls, and Policy-Makers: American Public Opinion and National Security (New York: Lexington Books, 1992). 31. Richard S. Beal and Ronald H. Hinckley, “Presidential Decision Making and Opinion Polls,” Annals of the American Academy of Social and Political Science 472 (March 1984): 72–84. 32. Lawrence R. Jacobs and Robert Y. Shapiro, Politicians Don't Pander: Political Manipulation and the Loss of Democratic Responsiveness (Chicago: University of Chicago Press, 1999); Shapiro and Jacobs, “Public Opinion, Foreign Policy, and Democracy.” 33. John H. Aldrich, John L. Sullivan, and Eugene Borgida, “Foreign Affairs and Issue Voting: Do Presidential Candidates ‘Waltz before a Blind Audience’?” American Political Science Review 83, no. 1 (1989): 123–41. 34. Morris P. Fiorina, Retrospective Voting in American National Elections (New Haven, Conn.: Yale University Press, 1981); Paul Abramson, John H. Aldrich, and John Rohde, Change and Continuity in the 1988 Election (Washington, D.C.: Congressional Quarterly Press, 1990). 35. Alan D. Monroe, “Consistency between Public Preferences and National Policy Decisions,” American Politics Quarterly 7, no. 1 (1979): 3–19. 36. Benjamin I. Page and Robert Y. Shapiro, “Effects of Public Opinion on Policy,” American Political Science Review 77, no. 1 (1983): 175–90. 37. Benjamin I. Page, “Democratic Responsiveness? Untangling the Links between Public Opinion and Policy,” PS: Political Science and Politics 27 (March 1994): 17–21.Page 528 → 38. V. O. Key Jr., Public Opinion and American Democracy (New York: Knopf, 1961), 14.

39. Doris A. Graber, Public Opinion, the President, and Foreign Policy: Four Case Studies from the Formative Years (New York: Holt, Rinehart and Winston, 1968), 318. 40. Robert C. Hildebrand, Power and the People: Executive Management of Public Opinion in Foreign Affairs, 1897–1921 (Chapel Hill: University of North Carolina Press, 1981). 41. Thomas W. Graham, “The Politics of Failure: Strategic Nuclear Arms Control, Public Opinion, and Domestic Politics in the United States—1945–1980” (Ph.D. diss., Political Science Department, MIT, 1989). 42. Douglas C. Foyle, Counting the Public In (New York: Columbia University Press, 1999). 43. Lawrence R. Jacobs and Robert Y. Shapiro, “The Rise of Presidential Polling: The Nixon White House in Historical Perspective,” Public Opinion Quarterly 59, no. 2 (1995): 163–95; Jacobs and Shapiro, Politicians Don't Pander. 44. Shapiro and Jacobs, “Public Opinion, Foreign Policy, and Democracy” James M. Lindsay, “The New Apathy: How an Uninterested Public Is Reshaping Foreign Policy,” Foreign Affairs 79 (September-October 2000): 2–8. 45. Steven Kull and I. M. Destler, Misreading the Public: The Myth of a New Isolationism (Washington, D.C.: Brookings Institution Press, 1999). 46. Richard Sobel, “Polling in Foreign Policy Crises: Ascertaining the Questions to Ask” (unpublished manuscript, Princeton University, 1995). 47. Page and Shapiro, The Rational Public; Samuel L. Popkin, The Reasoning Voter (Chicago: University of Chicago Press, 1991). 48. James N. Rosenau, Turbulence in World Politics: A Theory of Change and Continuity (Princeton, N.J.: Princeton University Press, 1990).

Page 529 →

SIMULATION AND EXPERIMENTATION IN FOREIGN POLICY ANALYSIS Some Personal Observations on Problems and Prospects Jonathan Wilkenfeld While many in the field of foreign policy analysis associate me most closely with the systematic study of foreign policy behavior in general, and conflict and crisis decision making in particular—most notably, the International Crisis Behavior (ICB) Project1—I have chosen to highlight a somewhat different direction that my recent research has taken. During the past two decades, almost as a sideline, I have been heavily involved in the conceptualization and development of simulation approaches to the teaching of international politics and foreign policy. The network-based, distributed foreign policy simulations offered by the International Communications and Negotiations Simulations (ICONS) Project, centered at the University of Maryland, now reach about sixty university classrooms each academic year, and ICONS exercises have been run in more than thirty-five countries since their inception in 1982. More recently, ICONS has worked with high school teachers and curriculum experts, and the high school version is employed in about 150 high schools each year. In addition to the University of Maryland, there are now ICONS centers at the University of Connecticut and Whittier College in California, as well as in Israel.2 Page 530 → But even that is not what I really want to focus on here. My work in foreign policy analysis during the last decade has increasingly involved the use of experimental approaches to the study of foreign policy decision making. It is my adventures, and in many cases misadventures, in this new realm that I want to focus on in this paper. I begin with some background on experimentation in political science in general, and in international politics in particular. This will be followed by the story of how I got involved with this approach in the first place. Finally, in order to examine what experimentation can and cannot do for us, I present a research design and some preliminary data in which both cross-national and experimental approaches are used to extend our knowledge on one particular phenomenon in international politics, the process of mediation in international disputes.

Experimental Design in Political Science While experimentation remains outside the mainstream of international relations, and of political science in general, efforts have been made in recent years to bolster the acceptance of experiment-based research designs within the field. Kinder and Palfrey offer a treatise on why experimentation should be integrated into political science as a way to “supplement, not replace, traditional empirical methods.”3 What is it that experiments offer political scientists? Experimentation allows researchers to isolate a crucial independent variable by providing the means to control extraneous factors, and it allows analysts to look at the effect of that variable by comparing the results of those who were exposed to the variable with the results of an unexposed control group.4 Experimentation provides researchers a precise view of a certain phenomenon, a certain relationship. It is this quality that makes experimentation valuable: it is the most reliable means of examining and understanding the nature of a causal relationship, of demonstrating the internal validity of a theory in political science.5 Experiment-based research provides insights into relationships that are difficult, if not impossible, to observe in any other way. Snyder argues that the insights and depth of understanding that experimentation can uniquely provide make it an important heuristic tool in the “discovery phase of science-building.”6 Because the results of experiments can enlighten analysts' understanding of how and why a relationship exists and functions, this type of research allows for the refinement of general or preliminary theories in a way distinct from examinations of largeN (quantitative) or small-N (qualitative) empirical data.7 Without the empirical data that generate initial hypotheses and theories, there is no starting point for experimental research. But experimental research allows for

the continued evolution and the “better articulation” of existing theories.8 Some political scientists have come to appreciate the potential benefits of experimentation and have applied it to their research. Bositis records that while only nine articles using experimental research designs appeared in major political science journals between 1924 and 1950, 104 such articles appeared from 1980 to 1989 alone.9 McGraw, too, sees experimentation Page 531 → as a methodology “on the rise” in political science, especially in the areas of public opinion, collective action, public policy, and—relevant to our study—decision making.10 Among a few scholars, though, efforts have long been made to incorporate experiments into the core realm of international relations research. In 1970, the International Political Science Association convened a special roundtable conference to examine experimentation and event simulation as a means to build and test international relations theories. The conference generated a series of essays documenting experiment-based efforts to refine and enhance theories of international relations.11 This conference was, itself, a follow-up to a vast experiment launched by Guetzkow in the late 1950s and early 1960s.12 His Inter-Nation Simulation set a precedent—all too often forgotten by scholars of international relations—for how experimentation can be used to inform and improve international relations. For instance, the project generated results that illuminated understandings of how perceptions affect interstate relations. Given the technological advances made since Guetzkow's undertaking, we can only expect an increase in what we are able to learn from international relations-focused experiments today.

The Conflict and Negotiation (CAN) Research Group About twelve years ago, I began a lengthy collaboration with Sarit Kraus, a computer scientist specializing in distributed artificial intelligence. (I had barely heard of artificial intelligence in general at that point.) Kraus had been working on developing an automated agent capable of participating with other automated agents or with human players in the Diplomacy board game. With a strong interest in modeling the negotiation process from a bargaining and game-theoretic perspective, she was directed to me by her associates in the University of Maryland Institute for Advanced Computer Studies, some of whom had heard of my work on ICONS. It was apparent from the beginning that although we had very different training, we had a joint interest in better understanding how humans (in my case) and machines (in Kraus's case) negotiated in difficult (crisis) situations in order to reach mutually-beneficial agreements. Whereas my objective was to understand the conditions under which negotiations ended without the resort to violence, Kraus's objective was to understand how intelligent agents (machines) could be endowed with negotiation skills such that they could share scarce resources (tools like printers, Internet access, etc.) or cooperate to perform tasks. In both cases, negotiations are employed in situations where time is critical to some or all parties, resources are scarce, and tasks may require cooperative behavior. We wanted to minimize the amount of time spent on negotiation and maximize the likelihood of reaching mutually-beneficial outcomes.13 Sustained by significant support since 1992 from the National Science Foundation, the U.S. Institute of Peace, and Sun Microsystems, and with a research team consisting of professors and graduate students at the University of Maryland, the Hebrew University, and Bar Ilan University, we have Page 532 → produced a significant body of work on the dual fronts of automated agents and international negotiation. Our areas of expertise include negotiation and decision theory in social science in general and political science in particular, cognitive schema theory within psychology, and intelligent systems and distributed artificial intelligence in computer science. Our experimental work has been characterized by several features that tend to set it apart from other experimental work in foreign policy analysis.14 First, we have focused exclusively on behavior in crisis, with a particular emphasis on the effects shortness of time and a high level of threat have on crisis dynamics and outcomes.15 Second, our subjects had access to an elaborate decision support system (DSS), which, when properly used, could provide information on the utilities associated with various outcomes and hence enhance the subjects' ability to maximize expected utility, even under conditions of incomplete information.16 Third, all communications among the participants have taken place in controlled network environments, allowing for detailed analysis of the types of tools they consulted as well as the content and form of their interactions during the negotiation. Together, we hoped these features would help provide a rich environment in which experiments could be run and their results assessed.17

Summary of CAN Research Work to Date At the core of our experimental work to this point is the development of a strategic model of negotiation with an accompanying DSS based on three different scenarios: a hostage crisis (involving an Indian plane hijacked by Sikh separatists and forced to land in Pakistan),18 a fishing dispute (based on the Canada-Spain dispute of 1995),19 and the Ecuador-Peru Border Dispute of 1981 (see following).20 Five broad sets of research issues were addressed in our previous work. The Impact of the Use of DSSs on the Utility-Maximizing Behavior of Crisis Negotiators Experimental research led to the following conclusions: (a) average utility scores were higher for DSS users than for nonusers, and agreements were reached more frequently by DSS than by non-DSS users; and (b) in simulations that ended in agreement as opposed to opting out, the player most closely associated with the mediator role tended to send approximately the same number of messages to the other two players.21 The Impact of the Dynamics of Crisis Negotiations on Their Outcomes Experimental research led to the following conclusions: (a) DSS-supported negotiators are most strongly motivated by utility maximization, while non-DSS-supported negotiators tended to be motivated by a desire to uphold principles; (b) DSS-supported negotiators are more successful than non-DSS-supported negotiators in achieving high utility outcomes; (c) the existence Page 533 → of a DSS-supported negotiator among the adversaries is likely to result in higher overall utility scores than when no such negotiator is present; and (d) crisis negotiations in which one of the parties has access to a DSS are more likely to end in agreement than are nonDSS-supported negotiations.22 The Relationship between the Level of Cognitive Complexity of Crisis Negotiators and the Outcomes of Crisis Negotiations Experimental research led to the following conclusions: (a) negotiators at higher levels of cognitive complexity developed greater proficiency with the DSS; and (b) negotiators did not show an overall relationship between cognitive complexity and crisis outcome, either in terms of the achievement of higher utility scores or in terms of a greater tendency to reach agreement.23 The Impact of the Mix of Cognitive Complexity Levels among Crisis Negotiators on the Outcome of the Crisis Experimental research led to the following conclusions: (a) the more homogeneous the cognitive levels of the negotiators, the more likely they were to reach agreement; and (b) the more homogeneous the cognitive levels of the negotiators, the more quickly they reached agreement.24 The Impact of the Mix of Cognitive Complexity Levels among Crisis Negotiators and Crisis Mediators on the Outcome of the Crisis A set of experiments involving a two-party fishing dispute led to the following preliminary finding: agreements were reached with greatest frequency when the cognitive complexity level of the mediator was higher than that of the disputants.25

Merging Empirical and Experimental Approaches: Searching for More Robust Explanations Thus far, this paper has focused exclusively on recent contributions made by experimental work to our understanding of international phenomena: utility-maximizing behavior in crisis negotiation, levels of cognitive complexity and the process of negotiation, and most recently, cognitive complexity and receptivity to mediation. While this research has led to better understanding of experimentation, it has taken place with relatively limited reference to larger research questions in international politics and foreign policy. The main point of this paper is to

demonstrate how a merged research design can take advantage of data derived from both large-N empirical studies and studies based on experimental methods to extend the range of questions we can ask (and presumably answer) about the dynamics of international politics. For this purpose, ongoing research at the University of Maryland on mediation in international crises provides the backdrop for discussion. Data recently collected on the role of mediation in international Page 534 → crises for the International Crisis Behavior Project will be juxtaposed with experimental work on mediation in international crises.

Mediation in International Crises The United Nations defines mediation as “a method of peaceful settlement of an international dispute where a third party intervenes to reconcile the claims of the contending parties and to advance his own proposals aimed at a mutually-acceptable compromise solution.”26 Mediation generally occurs and is most relevant when the conflicting parties have reached a stalemate in fighting or in the negotiation process, or in both, and the parties usually accept mediation when they feel that a mediator's presence will facilitate an agreement that is more favorable than can be achieved via dyadic negotiations.27 The list of potential mediators is quite extensive: a single state (small-, medium-, or large-sized powers), a group of states, an international governmental organization, a regional governmental organization, a private transnational organization, a private individual, or a mixture of any of these. A number of scholars, including Bercovitch, Princen, Young, Susskind and Babbitt, and Rubin discuss the characteristics of the parties and the disputes themselves that are likely to enhance the success of mediation.28 While there is no general consensus, the literature on mediation seems to have converged on three basic roles that mediators can play in a negotiation, and most authors agree that the roles change as the mediator and disputants “gain information and skill.”29 First is the mediator as communicator or facilitator,30 serving as a channel of communications between two parties when a stalemate has been reached or when communications have broken down and face-to-face negotiations are not possible. This type of mediation is also referred to as third-party consultation,31 good offices, or process facilitation.32 The second role defined by Touval and Zartman is mediator as formulator. Here the mediator provides a substantive contribution to the negotiations in order to assist the conflicting parties in conceiving of a way to resolve their dispute when the parties reach a rigid impasse in the negotiation process. Hopmann suggests that this role is particularly effective when the parties' emotions are running high, and he highlights certain processes that the mediator as formulator may use to propose solutions: asking the parties to brainstorm, suggesting that issues be fractionated or linked together, inventing new solutions, and so forth. The third and final role that Touval and Zartman discuss is that of the mediator as manipulator. In addition to formulating solutions, this type of mediator uses its domestic, regional, or international position and its leverage—“resources of power, influence, and persuasion”—to “manipulate the parties into agreement.”33 The mediator as manipulator becomes a party to the solution, if not the dispute itself,34 and often needs to augment the appeal of its solution by using carrot-and-stick measures (adding and subtracting benefits to and from the proposed solution).35 Hopmann indicates that only a powerful mediator can play this role, and he cites two modes of mediator involvement as manipulator: influencing the direction of the negotiation Page 535 → process by raising the costs of disagreement and the rewards for agreement; manipulating the international environment to push the parties toward agreement. Burton, Carnevale and Pegnetter, Fisher, Keashly and Fisher, and Kelman argue that the mediator as facilitator /communicator is the most effective role, while Bercovitch, Bercovitch and Houston, and Touval and Zartman feel that the mediator in a formulator and manipulator role will be most effective because of the amount of leverage that such a mediator possesses.36 Princen notes that while it is often necessary to employ directive strategies to bring about an agreement between parties, the use of threats so closely associated with the directive/manipulative role can damage the “atmosphere of good will, trust, and joint problem solving” between the parties, which, if exacerbated enough, can break down negotiations.37 In the preliminary analysis presented here, two sets of questions pertaining to mediation are explored using the International Crisis Behavior data set on 412 international crises between 1918 and 1994: (1) under what circumstances is mediation most likely to occur? and (2) what types of outcomes are most closely associated with

mediation? From the experimental perspective, the question to be explored is how different types of mediation, as opposed to no mediation, affect the utility-maximizing behavior of negotiators, the pace of crisis abatement, and the satisfaction with mediation. Let us turn first to the empirical findings based on ICB.38 Empirical Findings on Mediation in International Crises First, some descriptive findings regarding mediation in twentieth-century international crises.39 We should note at the outset that 29 percent of all international crises between 1918 and 1994 exhibited some form of mediation. Threat. If we classify crises by level of threat perceived by the crisis actors, it turns out that mediation occurs at the highest rate in crises characterized by territorial threat. Polarity. Instances of mediation appear to be increasing. The data show the following trend for frequency of mediation: multipolarity (1918–39) 27 percent; World War II (1939–45) 9 percent; bipolarity (1945–62) 20 percent; polycentrism (1963–89) 34 percent; and post-cold war unipolarity (1989–94) 62 percent. System level. Mediation occurred with the following frequency at various system levels: subsystem 35 percent; mainly subsystem 23 percent; mainly dominant system 14 percent; and dominant system 8 percent. Ethnic origins of crisis. Mediation occurred as follows: secessionist crises 34 percent; irredentist crises 51 percent; nonethnic crises 22 percent. Geographic proximity. The pattern of mediation occurrences was as follows: contiguous actors 34 percent; nearneighbors 22 percent; distant actors 14 percent. The second set of questions has to do with the types of crisis outcomes most closely associated with mediation in the twentieth century. As was Page 536 → the case above, we are concerned only with whether or not mediation occurred, leaving questions concerning type of mediation and type of mediator to a later paper. Mediation and content of outcome. In those cases in which mediation occurred, outcomes involving ambiguous terminations—where one or more parties experienced compromise or stalemate—were prevalent in 60 percent of the cases. Conversely, when mediation did not occur, the majority of cases, 54 percent, showed definitive outcomes—involving a mix of victory and defeat. Mediation and form of outcome. Crises in which mediation occurred were more likely to terminate in agreement—formal, semiformal, tacit—than were nonmediated crises—63 percent versus 27 percent. Conversely, most nonmediated crises terminated through unilateral acts—52 percent compared with only 18 percent for mediated crises. Mediation and extent of satisfaction with outcome. Mediated crises produced more satisfying outcomes for the actors involved—41 percent of mediated crises showed all or most parties satisfied with the outcome, compared with only 25 percent for the nonmediated crises. Mediation and escalation or reduction of tensions. While we find a trend toward long-term tension-reduction among mediated crises—60 percent versus 50 percent for nonmediated crises—the results are not statistically significant. Clearly, then, mediation makes a difference. With mediation, outcomes are more likely to involve compromise, agreements are more likely to be reached, the parties are more likely to be satisfied, and there appears to be a tendency toward reduction of tensions over time. Experimental Findings on Mediation in International Crises The empirical findings derived from ICB provide us with a general picture of the effects that mediation—writ large—has on crisis and crisis negotiations. But these data cannot provide the full story of how mediation affects

crisis processes and outcomes, nor do they reveal either the dynamics of interaction between the crisis actors and the mediator or the effects of different styles of mediation. An experimental research design, though, can build upon these general empirical findings and allow us to pursue more extended research questions. The experimental environment in which mediation in international crisis was studied consists of a generalized decision support system for individual decision makers, a crisis scenario, and a communications system for both negotiators and mediators.40 For purposes of these very preliminary experiments, a scenario loosely based on the Ecuador-Peru Border Dispute of 1981 was chosen. This case, part of an ongoing protracted conflict between 1935 and 1994 (ultimately terminating in 1998), which produced a total of four international crises,41 involved a historical instance of successful mediation by the Organization of American States, which called upon the United States, Argentina, Brazil, and Chile to intervene. Four possible Page 537 → outcomes were built into the simulated negotiation: territorial division, cease-fire, arbitration, and status quo. In part, we sought to establish the reliability of our experimental environment through our ability to replicate in broad general terms the outcome of this historical case. The original research design called for three groups of subjects drawn from a population of political science majors at the University of Maryland currently taking courses in aspects of international politics. Pairs of students were to take on the roles of Ecuador and Peru under three different circumstances: no mediation, mediator as facilitator, and mediator as manipulator.42 Ultimately, only fifteen simulations were run, and we dropped the nomediation control groups. While the small N and the lack of a control group severely limit our ability to generalize from these preliminary results, the findings do point in some interesting directions that are consistent with our main point: that experimental designs are able to address questions that we have difficulty getting at in more conventional empirical designs. Some highlights of these preliminary experimental results follow. Mediation and expected utility. Among simulations that ended in agreement, mediator as manipulator produced mean utility point outcomes that were 20 percent higher for Ecuador—the weaker of the two actors in this crisis—and 9 percent higher for Peru, as compared with the outcomes for mediator as facilitator. However, mediator as facilitator produced mean utility point differences between the two parties that were 29 percent lower than for mediator as manipulator. That is, mediation as facilitation seemed to produce outcomes characterized by greater mutual benefit. Mediation and pace of crisis abatement. Among all outcomes, simulations with mediator as manipulator terminated in 34 percent less time than simulations with mediator as facilitator. However, among the simulations that terminated in agreement between Ecuador and Peru, termination time was 11 percent shorter for facilitative than for manipulative mediation. The perceptions of participants in the simulation mirror these empirical results. One hundred percent of the respondents who worked with the mediator as manipulator believed that the mediator sped up the pace of crisis abatement. Results for the mediator as facilitator were more ambiguous: 40 percent believed that the mediator sped up the negotiation process; another 40 percent believed that the mediator actually slowed the abatement process down; while 20 percent believed that the mediator had no effect on the pace. Expectations for mediation. Prior to the simulation, 58 percent of subjects anticipated that the involvement of a mediator in the negotiations would make it easier or much easier to achieve a desired outcome. Following the simulation, 67 percent of respondents who had been subject to mediator as manipulator continued to believe that the mediator made crisis abatement easier. A noticeable change occurred among those who were subject to mediator as facilitator: within that group, only 20 percent of respondents believed that the mediator had made it easier to reach an outcome, while 40 percent believed that the mediator had actually made it more difficult. Satisfaction with outcome. No significant relationship was found between how satisfied a negotiator was with the outcome of the simulation Page 538 → and the type of mediator involved in the negotiation. Among those who worked with the mediator as manipulator, 67 percent achieved an outcome they considered to be satisfactory or very satisfactory, while 60 percent of those who worked with the mediator as facilitator found the outcome of their negotiation to be satisfactory. Interestingly, all of the respondents who indicated that the outcome for their country was unsatisfactory also indicated that mediator involvement made the negotiations more difficult.

Satisfaction with mediation. “The grass is always greener on the other side”: of those involved in the mediator-asfacilitator negotiations, 60 percent noted that they would have preferred the mediator to have been more forceful or much more forceful. No respondents wished that this type of mediator had been less forceful. On the other hand, 100 percent of respondents from the mediator-as-manipulator simulations reported that they wished the mediator had been less, or much less, forceful. Even those who achieved an outcome they considered to be satisfactory believed this mediator played too forceful a role. In sum, mediator as manipulator produced higher utility outcomes for both parties, while mediator as facilitator produced greater mutual benefit (smaller gap between the utility values for the two parties). Mediator as manipulator generally resulted in shorter elapsed time to outcome in general, but mediator as facilitator produced shorter elapsed time to agreements. Most negotiators who experienced mediator as manipulator believed that mediation made it easier to reach a desired outcome, while only a small portion of negotiators who experienced mediator as facilitator felt that mediation had helped reach a desired outcome. However, all parties were dissatisfied with mediation: most of those who experienced facilitation would have preferred a more forceful mediator, while all subjects who experienced manipulation expressed a desire for less forceful mediation. Finally, there were no discernible differences in satisfaction with the outcome resulting from the two forms of mediation.

Prospects and Limitations for Experimental/Simulation Work in Foreign Policy Analysis At one time or another, we have all bemoaned the fact that foreign policy analysis is plagued by a combination of relatively small Ns and relatively large numbers of variables potentially contributing to the foreign policy decision-making process. On the surface, experimental work can successfully address both of these issues. With a more or less unlimited supply of willing undergraduates as subjects, our N is virtually limitless. By carefully constructing our simulation models and controlling our experimental environments, we can limit the number of variables we have to consider in any given experimental run. If this is the case, why are there not more eager scholars flocking to the simulation laboratories, and why are those of us who are there having so much trouble? First, I would argue that the transition from simulation environment, whose primary application in political science is for instruction and training, Page 539 → to the experimental laboratory is not a simple one. First and foremost, most of us lack any real training in experimental design (how many graduate programs in political science offer courses in this area?). What to the experimental psychologist seems natural requires of the political scientist a very real investment of time and energy, and a lot of trial and error. Thus, even with the prospects of great advancements in knowledge seemingly at our fingertips, too few of us venture into the unknown terrain of the experimental laboratory. Moreover, once in the lab, we are faced with a host of other seemingly-insurmountable obstacles. Consider our subjects: typically college freshmen (a few months out of high school) or, if we're lucky, upperclassmen. Cluttering their brains on any particular day are the aftermath of the previous evening's indiscretions and plans for tonight, swarming hormones, extreme hunger, thirst, and a strong need for a cigarette (if not worse), not to mention their at best neutral positions—if not downright hostility—toward having to participate in the experiment in the first place. Combine that with their relatively uninformed position on the nature of the international system in general, and foreign policy decision making in particular, and you are lucky if you can get anything that resembles replicable results.43 That is not to say that real-world decision makers don't have extraneous things on their minds, but at least we know that some of their attention is focused on policy issues. Even if these two problems could somehow be neutralized, the issue of the simulation itself remains. We know from the simulation literature that a model should be detailed enough to represent the important aspects of the reality it is meant to represent but not so detailed as to overwhelm the participant with information (even though we know that that is exactly what happens to real-world decision makers, particularly in crisis). This turns out to be an extremely difficult undertaking. In our own case, we have some evidence to suggest that in our previous round of experiments involving a model of a fishing dispute roughly resembling the dispute between Canada and Spain in 1995, we not only packed the model with excessive detail but then superimposed an elaborate mediation process, and the results showed that the subjects were not coping and were therefore making decisions based on only partial use of the decision support tools they had available to them. If you add to all this the expense and time

involved in the design and execution of a series of experiments, it is not surprising that relatively few political scientists venture into this domain. Given this rather bleak picture, where do I see simulation and experimental design fitting into the array of approaches now dominating the field of foreign policy analysis? At the level of political psychology, there are simply some characteristics of decision makers and decision making that we are never going to be able to study systematically with anything approaching a large enough N to allow for careful statistical work. It is here, under controlled laboratory environments, that we can begin to make progress in understanding how decision makers cope with and process information in crisis versus noncrisis environments, how third parties can affect Page 540 → the course of and outcomes to contentious situations, how actor attributes such as power (of various sorts) affect their negotiation strategies and behavior, and the circumstances under which violent strategies are chosen. While nothing can substitute for the data we can glean from in-depth case studies, the experimental environment—when properly specified—can help us fill in the gaps in our knowledge and ultimately allow us to generalize beyond a limited experimental environment. But to do this right, we must begin to properly train the next generation of experimental foreign policy analysis scholars. This is not everyone's cup of tea, nor should it be. Nevertheless, it does offer an important alternative to the cross-national and case study approaches that dominate the field of foreign policy analysis today. But we lack the experimental infrastructure, the culture of the laboratory, and perhaps most important, the channels by which experimental results can become important sources of information as we seek to better understand the behavior of political decision makers. And it is critical that we be able to demonstrate more convincingly that results obtained in experimental environments can help us better understand the behavior of real decision makers in critical situations. Much work remains to be done on this front.

Notes While I take full responsibility for the contents of this paper, a number of people have made very significant contributions to my thinking on these issues over the past several years. The most important are my collaborators in simulation and experimental environments. Key ICONS collaborators have been Betsy Kielman, Brigid Starkey, Mark Boyer, Joyce Kaufman, Doreen Bass, Patty Landis, Judith Torney-Purta, Harold Guetzkow, Beth Blake, David Crookall, Robert Noel, Barry Hughes, Leopoldo Schapira, and Richard Brecht. In the experimental phase of my work, I owe much to Sarit Kraus, Tara Santmire, Joe Oppenheimer, Kim Holley, J. Joseph Hewitt, Michael Harris, Toni Santmire, Phillip Schrodt, Charles Taber, Alex Mintz, Nehemia Geva, Michael Lebrun, and Kristian Gleditsch. The Conflict and Negotiation (CAN) Research Group at the University of Maryland—David Andersen, Victor Assal, Christopher Frain, Amy Pate, David Quinn, Tara Santmire, and Kathleen Young—has provided able assistance in both empirical and experimental data collection and analysis. Over the years, this work has been generously supported by the National Science Foundation (currently under grants SES9905575 and IIS9820657), the U.S. Institute of Peace (under grants SG-35-95 and SG-52-00), the U.S. Department of Education (Title VI and FIPSE), IBM, and Sun Microsystems. 1. M. Brecher and J. Wilkenfeld, A Study of Crisis, paperback with CD-ROM ed. (Ann Arbor: University of Michigan Press, 2000). 2. B. Starkey, M. Boyer, and J. Wilkenfeld, Negotiating a Complex World: An Introduction to International Negotiation (Boulder, Colo.: Rowan and Littlefield, 1999); see also http://www.icons.umd.edu. 3. D. Kinder and T. Palfrey, “On Behalf of an Experimental Political Science,” in Experimental Foundations of Political Science, ed. D. Kinder and T. Palfrey (Ann Arbor: University of Michigan Press, 1993), 1. 4. D. T. Campbell and J. C. Stanley, Experimental and Quasi-Experimental Designs for Research (Chicago: Rand McNally, 1963).

5. J. A. Laponce, “Experimenting: A Two-Person Game between Man and Nature,” in Experimentation and Simulation in Political Science, ed. J. A. Laponce and P. Smoker (Toronto: University of Toronto Press, 1972), 3–15; D. A. Bositis, Research Designs for Political Science: Contrivance and Demonstration in Theory and Practice (Carbondale: Southern Illinois University Press, 1990); K. Mc-Graw, “Political Methodology: Research Design and Experimental Methods,” in A New Handbook of Political Science, ed. R. E. Goodin and H. D. Klingemann (New York: Oxford University Press, 1993), 769–86. 6. R. C. Snyder, “Some Perspectives on the Use of Experimental Techniques in the Study of International Relations,” in Simulations in International Relations, ed. H. Guetzkow et al. (Englewood Cliffs, N.J.: Prentice Hall, 1963), 7. 7. R. A. Brody, “The Study of International Politics qua Science: The Emphasis on Methods and Techniques,” in Contending Approaches to International Politics, ed. K. Knorr and J. N. Rosenau (Princeton, N.J.: Princeton University Press, 1969), 110–28; Kinder and Palfrey, “On Behalf of an Experimental Political Science.” 8. A. Franklin, Experiment, Right or Wrong (New York: Cambridge University Press, 1996). 9. Bositis, Research Designs for Political Science, 65. 10. McGraw, “Political Methodology.” 11. Laponce and Smoker, Experimentation and Simulation in Political Science.Page 541 → 12. H. Guetzkow et al., Simulations in International Relations. 13. S. Kraus, J. Wilkenfeld, and G. Zlotkin, “Multiagent Negotiation under Time Constraints,” Journal of Artificial Intelligence 75 (1995): 297–345; J. Wilkenfeld, S. Kraus, K. Holley, and M. Harris, “GENIE: A Decision Support System for Crisis Negotiations,” Decision Support Systems 14 (1995): 369–91. 14. See N. Geva and J. Mayhar, “The Cognitive Calculus of Decisions on the Use of Force” (paper presented at the annual meeting of the American Political Science Association, Washington, D.C., August 28–31, 1997); A. Mintz, N. Geva, S. Redd, and A. Carnes, “The Effect of Dynamic and Static Choice Sets on Political Decision Making,” American Political Science Review 91, no. 3 (1997): 553–66; P. Tetlock and A. Belkin, eds., Counterfactual Thought Experiments in World Politics: Logical, Methodological, and Psychological Perspectives (Princeton, N.J.: Princeton University Press, 1996). 15. For a more formal definition of crisis, see Brecher and Wilkenfeld, A Study of Crisis. 16. For a discussion of decision support systems in general and their application to negotiation in particular, see Wilkenfeld et al., “GENIE: A Decision Support System for Crisis Negotiations.” 17. A secondary, though no less important, interest was the creation of a controlled environment in which negotiators could be trained for work in conflict management and resolution. 18. See Kraus, Wilkenfeld, and Zlotkin, “Multiagent Negotiation under Time Constraints” J. Wilkenfeld, S. Kraus, and K. Holley, “The Use of Decision Support Systems in Crisis Negotiations,” in Encyclopedia of Library and Information Science, ed. A. Kent (New York: Dekker, 1999). 19. J. Wilkenfeld, S. Kraus, Tara Santmire, and C. Frain, “The Role of Mediation in Conflict Management: Conditions for Successful Resolution” (paper presented at the annual meeting of the International Studies Association, Washington, D.C., February 16–20, 1999); J. Wilkenfeld, S. Kraus, Tara Santmire, and C. Frain, “The Role of Mediation in Conflict Management: Conditions for Successful Resolution,” in Methodology in International Relations, ed. Z. Maoz et al. (forthcoming). 20. For details on our experimental procedures, see Tara Santmire, J. Wilkenfeld, S. Kraus, K. Holley, Toni Santmire, and K. Gleditsch, “The Impact of Cognitive Diversity on Crisis Negotiations,” Political Psychology 19, no. 4 (1998): 721–48. 21. K. Holley and J. Wilkenfeld, “The Use of Decision Support Systems in International Education” (paper presented at the annual meeting of the International Studies Association, Washington, D.C., March 28–April 1, 1994); S. Kraus, J. Wilkenfeld, M. Harris, and E. Blake, “The Hostage Crisis Simulation,” Simulation and Gaming 23 (1992): 398–416. 22. Wilkenfeld et al., “GENIE: A Decision Support System for Crisis Negotiation.” 23. J. Wilkenfeld, S. Kraus, Tara Santmire, K. Holley, and Toni Santmire, “Cognitive Structure and Crisis Decision Making” (paper presented at the annual meeting of the International Studies Association, San Diego, April 16–20, 1996). 24. Santmire et al., “The Impact of Cognitive Diversity on Crisis Negotiations.” 25. Wilkenfeld et al., “The Role of Mediation in Conflict Management.”

26. United Nations, Handbook on the Peaceful Settlement of Disputes between States (New York: United Nations, 1992), 40. 27. S. Touval and I. W. Zartman, “Introduction: Mediation in Theory,” in International Mediation in Theory and Practice, ed. S. Touval and I. W. Zartman (Boulder, Colo.: Westview, 1984), 7–17; P. T. Hopmann, The Negotiation Process and the Resolution of International Conflicts (Columbia: University of South Carolina Press, 1996); I. W. Zartman and S. Touval, “International Mediation in the Post-Cold War Era,” in Managing Global Chaos: Sources of and Responses to International Conflict, ed. C. A. Crocker and F. O. Hampson, with P. Aall (Washington, D.C.: U.S. Institute of Peace Press, 1996), 445–61. 28. J. Bercovitch, “International Mediation: A Study of the Incidence, Strategies, and Conditions of Successful Outcomes,” Cooperation and Conflict 21, no. 3 (1986): 155–68; J. Bercovitch, “Mediation in International Conflict: An Overview Page 542 → of Theory, a Review of Practice,” in Peacemaking in International Conflict: Methods and Techniques, ed. I. W. Zartman and J. L. Rasmussen (Washington, D.C.: U.S. Institute of Peace Press, 1997), 125–53; T. Princen, Intermediaries in International Conflict (Princeton, N.J.: Princeton University Press, 1992); O. R. Young, The Intermediaries: Third Parties in International Crises (Princeton, N.J.: Princeton University Press, 1967); L. Susskind and E. Babbit, “Overcoming the Obstacles to Effective Mediation of International Disputes,” in Mediation in International Relations, ed. J. Bercovitch and J. Z. Rubin (New York: St. Martin's Press, 1992), 30–51; J. Z. Rubin, “Conclusion: International Mediation in Context,” in ibid., 249–72. 29. Princen, Intermediaries in International Conflict, 65. 30. Touval and Zartman, introduction to Touval and Zartman, International Mediation in Theory and Practice; J. W. Burton, Global Conflict: The Domestic Sources of International Crisis (Brighton, England: Wheatsheaf Books, 1984); Hopmann, The Negotiation Process and the Resolution of International Conflicts. 31. L. Keashly and R. J. Fisher, “A Contingency Perspective on Conflict Interventions: Theoretical and Practical Considerations,” in Resolving International Conflicts: The Theory and Practice of Mediation, ed. J. Bercovitch (Boulder, Colo.: Lynne Rienner, 1996), 235–61; H. C. Kelman, “Informal Mediation by the Scholar/Practitioner,” in Bercovitch and Rubin, Mediation in International Relations, 64–96. 32. Hopmann, The Negotiation Process and the Resolution of International Conflicts. 33. Touval and Zartman, introduction, 12. 34. Hopmann qualifies the notion of the manipulative mediator being a party to the dispute by stating that this mediator “almost” assumes the role of a disputant in the negotiation process. Hopmann, The Negotiation Process and the Resolution of International Conflicts. 35. Touval and Zartman, introduction; Zartman and Touval, “International Mediation in the Post-Cold War Era,” in Crocker and Hampson, Managing Global Chaos. 36. J. W. Burton, Conflict and Communication: The Use of Controlled Communication in International Relations (New York: Free Press, 1969); P. J. Carnevale and R. Pegnetter, “The Selection of Mediation Tactics in Public Sector Disputes: A Contingency Analysis,” Journal of Social Issues 41, no. 2 (1985): 65–81; R. J. Fisher, “Third Party Consultation: A Method for the Study and Resolution of Conflict,” Journal of Conflict Resolution 16, no. 1 (1972): 67–94; R. J. Fisher and L. Keashly, “Distinguishing ThirdParty Interventions in Inter-group Conflict: Consultation Is Not Mediation,” Negotiation Journal 4, no. 4 (1988): 381–93; Kelman, “Informal Mediation by the Scholar/Practitioner” Bercovitch, “International Mediation” J. Bercovitch and A. Houston, “The Study of International Mediation: Theoretical Issues and Empirical Evidence,” in Bercovitch, Resolving International Conflicts, 11–35; and Touval and Zartman, introduction. 37. Princen, Intermediaries in International Conflict, 58. 38. Since the International Crisis Behavior data sets did not originally contain variables focusing explicitly on mediation, coding was undertaken in 2000 on five new variables pertaining to mediation: whether or not mediation took place, who served as mediator, impact of mediation on type of outcome, impact of mediation on pace of abatement, and timing of mediation. Coding was performed by Amy Fields and Heather McQueeney, under the supervision of David Quinn. It should be noted that this coding was based on information contained in the ICB crisis summaries (Brecher and Wilkenfeld, A Study of Crisis). All of the following conditions had to be present for the case to be considered one of mediation: A third party intervenes in a new or ongoing negotiation process.

Mediator is not a direct party to crisis.Page 543 → Mediator's results are nonbinding. Crisis actors determine the outcome of the crisis. The mediator does not have decision-making power. Mediation is a nonviolent form of intervention. Mediator's presence is voluntary; that is, mediator can leave the negotiations. Mediation begins once a stalemate between the adversaries has been reached. 39. For the sake of brevity, detailed statistical findings will not be presented in this paper. 40. The experimental environment is described more fully in Santmire et al., “The Impact of Cognitive Diversity on Crisis Negotiations” Wilkenfeld et al., “GENIE: A Decision Support System for Crisis Negotiations” J. Wilkenfeld, K. Young, V. Assal, and D. Quinn, “Mediating International Crises in the 20th Century” (paper presented at the International Studies Association convention, Hong Kong, July 26–28, 2001). 41. For a discussion of these crises in detail, see Brecher and Wilkenfeld, A Study of Crisis. The scenario and decision support systems, developed by Kathleen Young, David Quinn, and Chris Frain, are available from the author upon request. A more extended discussion of this research project appears in Wilkenfeld, Young, Assal, and Quinn, “Mediating International Crises in the 20th Century.” 42. Due to limitations on the number of subjects available at this stage of our research, the role of “mediator as formulator” was not examined in this research design. 43. I am reminded of one of the very first works I became familiar with in simulation, a study by Hermann and Hermann that used the Inter-Nation Simulation (INS) to attempt to replicate the conditions leading to the outbreak of World War I. They checked for seemingly everything, yet curiously they couldn't get war to break out. As it turned out, a student playing one of the key roles was an avowed pacifist and simply wouldn't go to war under any circumstances. C. F. Hermann and M. G. Hermann, “An Attempt to Simulate the Outbreak of World War I,” American Political Science Review 61, no. 2 (1967): 406–16.

Page 544 → Page 545 →

INTERNATIONAL SECURITY, PEACE, AND WAR

Page 546 → Page 547 →

SECURITY THEORY Six Paradigms Searching for Security Edward A. Kolodziej This survey of security theory is divided into two parts. The first defines security as a political concept and phenomenon. This discussion provides a point of departure for reviewing prevailing security theories. The second, and longest, section briefly examines and evaluates the claims of six competing research programs concerned directly or indirectly with security. These include realism, neorealism, economic liberalism, liberal institutionalism, behaviorism, and constructivism. These research programs can be distinguished on the basis of their ontological, epistemological, methodological, and evidentiary assumptions about actors and their behavior. Given space constraints, the discussion will identify only the principal differences between these research paradigms and their implications for security theory. To forge links between this theoretical discussion and hard facts, I assess the capacity of each paradigm to explain the implosion of the Soviet Union and the end of the cold war. The enormous shifts in power that these changes in world politics represent are, arguably, a plausible, if provisional, test of these research programs and their relevance to security theory. No attempt will be made to exhaustively critique each school of thought, a task well beyond the scope of this evaluation. This overview argues that in greater or lesser measure, each has something worthwhile to contribute to our understanding of these tectonic changes in the structure and processes of global politics. This optimism challenges the pessimism adopted by many scholars. They severely criticize these research paradigms for their failure to explain, much less to predict, the collapse of the Soviet system and the postwar bipolarity.1 Conversely, this evaluation resists the temptation of some analysts who insist that their approach, however beset now, will eventually be validated in the post-cold war era, if enough time is allotted for their predictions to be realized.2 Based on this survey, four conclusions are drawn about the state of security theory today. First, there exists no one paradigm capable of explaining the security behavior of relevant actors. Second, there exists no one methodology capable of providing a definitive theory of actor behavior. Third, Page 548 → there exists no one policy prescription available to actors to solve their security dilemmas.3 Fourth, and following from these generalizations, there is no agreement across these research programs about which actors in world politics have decisive influence and impact on security. Competing actors, aside from the state, include all forms of economic association—notably multinational corporations—the rise, enlargement, and increasing pervasiveness of global capitalist markets conceived as a political and socioeconomic institution; the expansion in the number, scope, and power of transnational actors; and the growing significance of individuals and small groups and national and international public opinion. While the state remains central to most security research programs, the many competitors identified by these research programs seriously limit the development of a comprehensive theory of security in an increasingly globalized world. In contrast to much conventional wisdom, concerned about the proliferation of security theories and their contesting claims,4 the overall conclusion to be drawn from this brief survey is that depending on the research or policy problem posed, these paradigms provide useful, if not fully satisfactory, theoretical tools to pursue important questions bearing on the use, threat, control, and elimination of force relevant in explaining the outcomes of increasingly interdependent exchanges of the world's populations and states. As Imre Lakatos suggests, the present pluralism and the competition of these research programs are very likely the best we can do for now until a comprehensive theory of security can be developed that either surmounts these specialized perspectives or incorporates them into a common framework of analysis.5 None of the research programs outlined in this essay can be conclusively tested in the sense of being confirmed or rejected on the strength of a single event or chain of events, like the end of the cold war. As William Wohlfort suggests, “Strictly

speaking, no particular finding about the Cold War's end will suffice to ‘falsify’ an entire research program…. For a single series of events to constitute a critical test of a theory, it must not only be inconsistent with the theory but be unambiguously ruled out by it.”6 Conversely, a theory of international relations or security that has little or nothing to say about the enormous political changes ushered in by the implosion of the Soviet Union and the collapse of global bipolarity would be hard pressed to retain its credibility and viability as a research program.

What Would Security Look Like If You Bumped into It in a Dark Room? There are few more contested terms or concepts than “security.” For many it is associated with the preservation of the state and the nation-state system. For others, security is primarily concerned with the physical security of individuals in civil society and their reliance on the state to arbitrate their differences and to defend them against foreign aggression. These Hobbesian notions have since been enlarged by those who would include economic welfare and the material well-being of populations as central security concerns. Page 549 → As economic development, industrialization, and modernization relentlessly progress around the globe, security for many observers is transformed into threats to the global environment or commons, evidenced by widespread pollution of the earth's air, water, and arable land and the rising problem of global warming.7 Still others conceive security to be a fundamentally psychological phenomenon. Whatever causes anxiety or angst to the individual or social psyche—say the threat of nuclear war or terrorism—becomes a central security issue. It is readily apparent that if security were understood to extend to so broad a range of human experience, it would soon lose much of its defining power. Security would have no clear, objective referent to discipline-systematic study. Security would simply change with the eye of the beholder. These contrasting and conflicting definitions of security seriously hamper the development of a theory of security. If the actors and the domains of investigation as well as the behavior of the actors within each domain change according to the interests, values, and perceptions of the observer, then no general theory of security is likely possible. The fear of many analysts that security as a term will be so broadly defined that it will be useless as a basis for describing, explaining, and predicting actor behavior would appear to be justified.8 Security would explain too much and too little. To provide some consistency and coherence to this analysis, this discussion, as a first move, invokes a traditional understanding of security, dating to the Melian Dialogue and to the Hobbesian conceptualization of security. The Melian Dialogue posed the issue of security in terms of power, principally material power and coercive threats.9 Hobbes generalized this condition of conflict between societies by rooting the problem of security in the very makeup (i.e., ontology) of individuals. Their infinite and conflicting preferences and their unswerving will to get their way in their exchanges with homologues created a social condition of permanent and perpetual conflict. Since getting one's way was a function of power, all preferences could be standardized. If politics is simply a struggle for power and not a search for the ideal polity, as classical Greek and Christian thought conceived the problem of politics, a science of politics and of political behavior is possible. The precondition for such a science was the survival or security of the actors engaged in this ceaseless struggle. From a Hobbesian perspective, the political phenomenon of security arises whenever and wherever an exchange between actors occurs in which the desired outcome from the transaction is to be resolved by force or the threat of force. Even Hobbes's provisional solution of the Leviathan or the modern state, possessed of awesome power to resolve conflicts between actors, is problematic. However greater the capacity for survival and security of the state may be compared with that of actors, its existence and its interests are no less subject to human preferences and to the countervailing power of actors opposed to its authority, power, and policies.10 This understanding of security has the virtue of being generalized across all transactions of individuals and their humanly constructed agents. From all such exchanges the properties of force and threats of violence are abstracted Page 550 → from the preferences in which they are embedded. A theory of security can be isolated for examination from the more daunting task of assessing the human worth of the contesting values in play. This permits the analyst to develop a pure theory—or Weberian ideal—of security. Hobbes, who was arguably the first to conceptualize this condition of force as a permanent, underlying determinant of human exchanges, drew the conclusion that humans were perpetually at war of all against all. Unless arrested by mutual consent (impossible by what was real about humans) or by some awesome power (the Leviathan), life would be “solitary, poore, nasty,

brutish, and short.”11 Carl von Clausewitz simply projected this individual condition to the relations of states.12 Pure war between states was conceived as the “ideal” or pure endgame toward which armed conflict was directed. Real war, the historical representation of the pure case, typically fell short of its “ideal” end-point. Either a host of “frictions” prevented the rivals from eliminating each other or, and the more likely and generalized explanation, their political aims fell short of pure war for its own sake. This understanding of security and the tendency of violent or force-incipient exchanges to move toward a pure case or Weberian “ideal” provide theorists with a potentially definable set of transactions around which a theory of security might be developed. However, as some theorists argue (notably economic and institutionalist liberals and constructivists), the Hobbesian assumption of perpetual conflict between egoistic actors or between states is not necessarily an insurmountable ontological limitation. This claim widens the theoretical lens to include not only the determinants of those conditions and incentives inducing actors to threaten and use violence but also those factors and forces prompting actors to eschew coercion in resolving their differences. Security theory must then embrace not only all those transactions that are within the domain of violence or threats associated with a transaction—the position of realists and neorealists and some liberal institutionalists—but also those transactions and power determinants that limit violence and force. A theory of security would entail potentially empiricallyverifying observations about the incentive structures and processes of actor decisions and actions, inducing them to resolve their exchanges by peaceful or violent means. Security theory as a subfield of international relations theory would be principally concerned with explaining both peace and war, since one could not be understood in the absence of the other. At the risk of simplification, security theory may begin with Hobbes, but it must end with Kant and the possibility of perpetual peace if it is to be comprehensive and coherent.13 A final introductory clarification is also necessary. A theory of security is not a theory of strategy. The latter, however worthwhile, is a narrowly-conceived understanding of security. It focuses principally on the important and practical question of what ways are likely to be cost-effective in compelling an adversary by force or threats to accede to one's will. There is much to be said for this conceptualization of security when opponents appear to be clearly engaged in threatening their rivals to get their way as Page 551 → World Wars I and II and the cold war suggest. Hitlers, Stalins, or Maos are not willed away by wishes. However, it would be a mistake to identify a theory of security with a theory about strategies of security, narrowed to only those transactions between actors in which the actors seek advantage by reference to countervailing, reflexive force. To equate this form of analysis with a theory of security risks trapping the analyst in a vicious circle. This move isolates security theory from the broader and ultimately-determining considerations associated with international relations theory. Such conceptual isolation is healthy neither for the broader field of international relations theory nor for strategic theory. Both need to inform the other if either is to be validated.14 At a practical policy level, closure may be premature; actors may well be misled to believe that strategies to resolve differences with rivals are ruled out by the assumptions made by the strategist, whether advising peace or war. At a theoretical level, both conflict and cooperation in human exchanges have to be included in any theory of security if the behavior, interests, aims, and values of actors are to be adequately explained.

Explaining the End of the Cold War: The Clash of Paradigms Realism and Neorealism Realism has a long and venerable history. Its roots can be traced to Thucydides's account of the Peloponnesian War. While realist theorists differ across a spectrum of assumptions about actor behavior, they converge on the notion that power, notably material power and principally violence, is the final arbiter of differences between actors. Power is central to a theory of security; it is the driving force behind actor behavior in international relations. At a state level, realists stress that the claim of states to sovereignty and to a monopoly of legitimate violence in settling disputes at home and in defending the state and its population abroad privileges security over all other competing values. States implicated in the quest for power mutually confront a security dilemma from which none can escape. That dilemma can be relaxed, but it cannot be surmounted, as long as states dispose ultimate military power under conditions of anarchy. The state has no choice but to pursue power as an overriding imperative. Failure to do so imperils its existence and its dependent populations.

Whereas classical realism works principally at the unit or state level,15 neorealism focuses narrowly at the system level. It stipulates the system of states as a causal force operating on all of the units of the system. Units are constrained by the anarchy of this structure. That constraint is materially defined by the distribution of capabilities, most prominently military force, across the system. In this self-help system of functionallyundifferentiated units, states are driven by a security imperative; they are obliged to concentrate their resources in ceaseless efforts to improve their position within the system and their relative power with respect to competitors. All exchanges Page 552 → between states, even those that appear benign such as trade or investments,16 are matters of high politics if the material power of each state is changed by these transactions. Anarchical systems may be multipolar or bipolar. With respect to the cold war, Kenneth Waltz argued that bipolar systems were inherently more stable than multipolar systems. The overwhelming material power of the United States and the Soviet Union allegedly decreased uncertainty about the power implications of shifting alliances, since the power of the alliance leader was less dependent on allies than under a multipolar system. Superpower calculations about their relative power vis-à-vis their rival were, arguably, less prone to error. Alliances were also supposedly more stable; the superpowers relied less on their partners to balance their adversary than would have been the case in a multipolar system. They also possessed sufficient capabilities to keep their allies from defecting, and if defections nevertheless arose, they counted for less in the balance.17 Realism and neorealism provided what appeared to be a compelling explanation for the superpower struggle and the long peace after World War II. Classical realism placed the superpower struggle in a historical context traceable to centuries of competition between the European states for hegemony over Europe and, through their empires, over the world. The United States and the Soviet Union were depicted as simply the latest in a succession of contender powers for world domination. Their contrasting domestic regimes, economic systems, and ideologies, however significant to explain their behavior along noncoercive axes of exchange, were subordinated to what were portrayed as the objective conditions of power characterizing their armed competition. Peace depended on a superpower equilibrium, tilting often precariously on a delicate balance of nuclear terror. Neorealists reached the same conclusion about the long peace as their realist colleagues. Where they differed was their contrasting explanations of state behavior. Whereas realists understood conflict between states as arising from the dyadic power relations of the states, neorealists identified the system itself as the principal source of conflict, characterized by the decentralized distribution of violent capabilities, viz, anarchy. They claimed to have identified a scientific principle of conflict, independent of its historical representation either in the European state system or in the cold war or in any other historical system. Units within a decentralized system—states today—that failed to address anarchy as an imperative of their survival would be penalized by other states and selected out in the evolutionary struggle for relative power and position. The historical and shifting explanatory power of classical realism was purportedly placed on an ahistorical conceptual foundation. Critics have been unsparing in their attacks on realism and neorealism.18 These research programs are dismissed by many theorists for their failure to explain, or even to anticipate, the peaceful end of the cold war, the abrupt breakup of the Soviet Union, or the rise of an ascendant Western coalition of liberal, market democratic states. Neither would have predicted the collapse of the Soviet state into its constituent republics as long as the coercive Page 553 → arms of the state—the military, the police, and supportive civil bureaucracies—were intact and functioning. Under realist-neorealist assumptions, neither the state nor its leadership could be expected to permit the dissolution of the state. Nor was the relinquishment of the Soviet sphere of influence, principally in Eastern Europe, supposedly consistent with realist theory. Conversely, the Western coalition was able to surmount the security dilemma in constructing a security community premised on peaceful change. The economic and technological success of the West, resting on democratic regimes and market rules, prompted the economic and political reforms launched by the Gorbachev government. These are explained more by reference to internal or domestic pressures for enhanced economic performance (perestroika) and openness (glasnost and democratization) than by external power shortfalls vis-à-vis the United States and its allies.19 These unit-level adaptations appear particularly telling for neorealism, which implicitly rejects unit-level changes as a decisive determinant of system change.

Realist partisans have retorted to these criticisms. Given their more flexible and plastic understanding of power, realists have fought back to reclaim the theoretical high ground. They argue that the real and widening disparities between economic and technological capabilities of the West and East and, accordingly, the rising military superiority of the United States compelled internal Soviet reform at home and retrenchment abroad. Rather than a repudiation of realist assumptions, these shifts in Soviet policy and behavior purportedly vindicated realist expectations. For realists, in contrast to their neorealist allies, the perception by Soviet leaders of this growing imbalance in power and their response to this gap as well as to domestic pressures for improved economic growth and performance are compatible with realist theory. The détente sought by the Gorbachev regime with the West reflected Soviet weakness. It was a logical adjustment to Western power. Détente promised to slow the arms race, narrow the East-West military gap, shift scarce resources from military to economic priorities, and bolster the Soviet Union's long-term competitive position by attracting Western capital and technology. If the West temporarily succeeded in relaxing the security dilemma in their relations as they allied to meet the Soviet challenge, realists still expected (and some Soviets, too) that balancing against the ascendant power—the United States—would commence again. By positing an indeterminate future in which the incentives of states to seek relative gains would persist, the claims of realists and especially neorealists are precluded from being subject to conclusive invalidation.20 Economic Liberalism Classical and neoclassical liberal economic theory moves from a state-as-unit or systemic level of analysis to the choices of individual consumers and producers. Liberal economic theory bases its theory of behavior on individuals rather than on states or systemic structures. Many, including some Page 554 → liberal economic theorists, may well argue that liberal theory has little or nothing to say about security. That would be a mistake for several reasons. First, liberals have always stressed the peaceful effects of commerce and economic exchange if individuals are free to register their choices in global markets. Dating at least from the nineteenth century, with the rise of the Manchester School and British adoption of free-trade policies, liberals have insisted that increasing economic interdependence and the costs of dissolution of these mutually-beneficial transactions—ever expanding, deepening, and thickening—made wars less likely, even impossible to contemplate by state leaders or their populations. This expectation and the bolstering evidence of the increasing economic interdependence of the European states and peoples on the eve of World War I persuaded Norman Angell and liberal theorists to reject the possibility of a general war.21 Even the wholesale carnage of World War I did not dissuade liberals from identifying military bureaucracies and states, putatively moved by atavistic or national passions, as the primary sources of war rather than free markets and economic pursuits. Joseph Schumpeter's response to Lenin's theory of imperialism pivoted on Schumpeter's contention that those engaged in commerce preferred peace to war to advance their interests.22 Resonances of these lines of analysis persist in several, seemingly discrete, areas. There are those who, following Angell, restate the costs of a general war as the principal obstacle to its eruption, even if nuclear weapons had not been invented.23 There are also those who contend that the Atlantic security community and specifically the European Union, nested within that community, can be explained partially, but critically, by the widening and deepening of economic interdependence that all but precludes a resort to force by any state to ensure its security and welfare interests.24 There are more profound, theoretical reasons undergirding liberal theory than these historical and empirical referents, however compelling their surface persuasion. The pure or ideal model of “economic man” animating liberal theory is an economic maximizer. The pursuit of wealth and welfare dictates the efficient allocation of human and material resources over time. The model of perfect competition predicts the long-term tendency of individuals and other economic units, notably firms, to maximize economic gains by relying on the signaling of relative prices for goods and services and the factors of production in increasingly globalized markets. Alternative economic institutions, particularly command economies dependent on bureaucratic and state ownership of the means of production, are inherently incapable of responding—efficiently, effectively, and innovatively25—to the ceaseless and rising demands of individuals and nations for “more now.”26 The form and subjectively-determined properties of these economic demands could only be known by the individuals possessing them. Only free

markets, unfettered in theory by the coercive constraints of states, bureaucrats, or monopolists, are capable of registering these subjective demands objectively to orient the allocation of inherently scarce resources in maximizing wealth for the greatest number. Much the same way that neorealists argued that states were selected out by the evolutionary struggle for power Page 555 → if they did not maximize their relative position, so liberal economists argued that alternative economic systems to free markets would be selected out of the evolutionary process. They were incapable of solving the knowledge problem posed by the assumed subjective value of individuals, ever striving for material well-being.27 Economic man would always prefer more than less economic gain. Only free markets could satisfactorily address these informational and valuational problems through market pricing. Liberal theory has much to say about the forces pressing on Soviet decision makers in opting for economic and political reform, however disastrous it proved for their particular personal positions. It helps to explain, too, Communist Chinese market reforms since 1979 and the determination of the Beijing regime to become integrated into Western-dominated global markets even at the risk of losing Communist party political control of the Chinese mainland. What market theory is unable to explain is the coercive determination of property and contractural rights. These inherent elements of a theory of order fall outside the scope of liberal economic theory. Hence, neoclassical economic theory has no theory of power. Power is assumed away by stipulating free choice in a perfect market under conditions of full information to guide the choice of actor strategies to maximize their material welfare. Populations are reduced to individual, uncoerced buyers and sellers expressing their subjective values by registering their preferences through relative prices in potentially homogeneous and integrated world markets. Institutional economic theory has attempted to partially fill this void in neoclassical liberal theory, but to the satisfaction neither of many liberal economic theorists nor, and much more so, their realist critics.28 Left begging in liberal economic theory is why the Soviet Union collapsed into its constituent national republics and why many of these remain unstable as a consequence of continuing ethnic, religious, and national strife. Nor has the Russian Federation been able to install free markets as many liberal economists expected because it has yet to define the property and contractual rights needed to make Russian markets work. If the world now claps with one hand in applauding the ascendancy of markets as the provisional solution to global welfare, markets still produce inequalities of wealth. Those drawing on Marxist theory insist that these inequalities remain the principal sources of conflict in the world society. Non-Marxists argue, too, that the state must be brought back into economic theory. It is rapidly emerging as the principal mechanism to address both the problem of income inequality between the North and South and between individuals and groups within states of the developed North.29 The state is also expected to produce order and to redistribute wealth through taxation and borrowing to meet collective goods claims for education, infrastructure, environmental protection, or health and material assistance to the lame, the halt, the young, the infirm, and the old.30 The pricing mechanisms of the market may well be useful in attacking some of these problems, evidenced by the Kyoto Protocol and its application of cost and market rules to control greenhouse emissions, but the sanctioning of those practices remains keyed to state cooperation and their enforcement capacity. American refusal to Page 556 → ratify the protocol underscores the subordination of market mechanisms to political power and national priorities achieved more by feat than playing by market rules. Market theory per se has little to say about these environmental concerns and the value placed by individuals, groups, and states on ways to address them. These shortcomings of liberal theory and market practices may be insufficient to resurrect currently discredited Marxist predictions of coming capitalist crisis and war, but they do suggest that liberal thought is unable to account for continued political conflict within national civil societies and between states.31 Liberal Institutionalism Liberal institutionalism draws on the preceding theories rather than advancing an alternative theory of behavior. It adopts the realist and economic liberal beginning point of the rational egoist maximizer, whether applied to the individual actor or the state.32 If institutionalists have difficulty explaining the end of the Soviet state, they have less difficulty explaining progressive superpower cooperation. Drawing both on microeconomic theory and game theory, Thomas Schelling laid the theoretical foundation to explain cooperation between armed rivals. They had a mutual interest in avoiding unwanted wars, of limiting the human and material losses of hostilities, and of

bringing such conflicts to a swift close on terms congenial to their preferred, if differentially defined, interests. They might not surmount their rivalry, but they had powerful incentives to cooperate on arms control and disarmament policies for mutual if differing advantage.33 Robert Axelrod's groundbreaking The Evolution of Cooperation broadened the evidentiary basis of Schelling's insights and deepened his analysis of why egoists could be expected to cooperate.34 Based on the results of a simulation study, Axelrod contended that tournament results revealed that egoists could learn through repeated exchanges to cooperate to limit the costs of defection and to maximize the gains of cooperation. A simple tit-fortat strategy proved to be the winning strategy. It proved to be “nice, provocable, forgiving and clear”35 to the players; it signaled each that all of the players could mutually gain from cooperation rather than attempting to improve their positions by defecting. The incentive to cooperate was increased further if players were certain that their exchanges and their mutual dependency would continue. According to Axelrod, the “shadow of the future” constrained them from cheating since defection would be countered by mutual defection and loss. Once cooperation as a strategy gained ground among egoists, its gains became apparent. First through evolutionary trial and error and progressively through accelerated learning, cooperation could be installed and institutionalized. Liberal institutionalists built on Axelrod's theoretical insight to reinforce their claim that institutions, arising from the repeated exchanges of egoists, could both constrain their behavior and foster opportunities for cooperation. The security dilemma could be relaxed by the collective, absolute benefits produced by institutions and shared rules, norms, and principles of behavior. Page 557 → Institutions introduced greater transparency to actor motivation and behavior. Actors could increasingly rely on their partners to adapt to and comply with institutional rules. Greater predictability ipso facto decreased uncertainty and reduced the incentive of actors to hedge or compensate for possible defection by taking out insurance in the form of arms races and expansionist policies. These moves would presumably encourage a spiral of defections to the mutual disadvantage of all parties. To increased transparency, decreased uncertainty, and greater predictability were added the progressively lowered transaction costs of cooperation. Building on the economic theory of the firm and social cost theory, liberal institutionalists pointed to the incentive of decreased transaction costs that institutions afforded their adherents.36 If firms and multinational corporations could be partially explained by their supposed greater efficiency in cutting market transaction costs, so also could international organizations and new and powerful transnational actors, with their own aims and interests and agendas, explain what appeared to be the growing cooperation of states.37 The realist picture of incessant state conflict and the pursuit of relative gains in power and position was replaced by the notion of complex interdependence. By focusing on absolute rather than relative gains, egoists had incentive to cooperate across interdependent domains of interaction. Moreover, complex interdependence produced multiple international actors—corporations, intergovernmental and nongovernmental organizations, and individuals. They worked through multiple channels to bypass or to constrain state behavior. Their agendas for action were no less complex. The growing number of actors and their overlapping and conflicting agendas blurred the traditional distinction between high and low politics. Nonsecurity problems necessarily rose in importance on the list of state priorities as zero-sum security concerns declined in importance. The growing number of problems posed by interdependence were progressively less amenable to unilateral appeals to force and threat than before. In this expanding setting of increasing, deepening, and thickening interdependence, the role of military force, however much still useful, was being downgraded by other preferred and more relevant noncoercive policy instruments.38 If institutionalist theory is unable to explain the end of the cold war and the breakup of the Soviet Union if national and ethnic factors are included in the equation, institutionalists are able to include observations within their research programs about the behavior of the state and international actors ignored or slighted by realist and neorealist theorists. The benefits of détente pursued by Gorbachev and his reformers could be accommodated by liberal institutionalist expectations. The multilateralization of security interests through arms control agreements, such as the elimination of intermediate-range ballistic missiles between the Soviet Union and the United States, could be harmonized, if not fully explained, by reference to institutionalist notions of cooperation. Similarly, domestically-driven economic reforms and the determination of command economies to adapt to the superior

decision-making properties of free markets are easily incorporated Page 558 → into institutionalist thinking. Even the role of ideas and values could be assigned marginal explanatory power, as an intervening variable in orienting actor behavior, to reequilibrate the preponderant weight assigned by realists and liberal economists to exogenous, material factors as determinants of actor behavior.39 Liberal institutionalists do a better job than their competitors in describing actor behavior. Parsimony is sacrificed to complexity and, consequently, explanations and predictions based on the liberal institutionalist paradigm are less coherent and more problematic. Hard facts and observation are preconditions for assessing the validity of a theory. By insisting on the complexity of international exchanges, the influence of transnational actors, and the constraints of international institutions, institutionalists play a very useful role in theory development. The richer descriptions they provide of international actor behavior guard against too early closure by suppressing observations of actor behavior at odds with a security theory's explanations and predictions. It also problematizes reliance on convenient ceteris paribus assumptions to avoid unsettling contradictory empirical observations. This conceptual defense mechanism may save theories to fight another day, but at the expense of addressing empirical data that contests the explanatory and predictive power of the theory. Behaviorism Behaviorism is not a theory of security. Rather, it is a collective noun embracing a wide range of research programs that address security issues, most notably war. These research programs are loosely associated with each other by their commitment to a set of shared methodological approaches to the study of security and international relations. Questions of ontology and epistemology, central to realist or constructivist paradigms, respectively, are subordinated to prevailing methodological and evidentiary rules associated with quantitative measures of actor behavior. Mathematical and statistical measures and objectively-discernible and potentially-replicable evidence are used to test propositions about actor behavior. Whereas classical realism, whether viewed from a Hobbesian individual or state perspective, makes assumptions about the actors' makeup, most researchers practicing a scientific approach to explaining actor behavior and, specifically, to behavior having a security content are neutral with respect to the composition of the state, that is, whether states are inherently or essentially disposed to cooperation and conflict. Instead, these researchers attempt to measure the behavior of actors under different but clearly specified conditions to establish correlations between observed behavior and these conditions.40 These theorists reduce the cold war to a data point.41 The same methodological and evidentiary rules are applied to all applicable cases of interest to the researcher. The cold war and the implosion of the Soviet Union do not count for more or less than other relevant data or cases assembled by the researcher to test a proposition posed by the research design. The ideographic approach of historians or many realist theorists is eschewed, however rich and textured their detailed descriptions of conflict and war. Actor behavior is subject Page 559 → to empirical verification across time and space in terms of the expected behavior of the actor(s) under examination, measured typically at statistically-significant levels of occurrence. It is difficult to generalize across the vast and growing scholarly output associated with behavioral research programs. If researchers concerned with explaining war are representative of those focused on security studies, it would appear appropriate to adopt Manus Midlarsky's characterization of the variety and richness of these theoretical perspectives as “the search for theories of wars” rather than for a unified theory applicable to all cases of armed conflict.42 Midlarsky divides his useful overview of war studies into systemic, nation-state dyadic, and state-centered theories of war. Each category embraces a wide number of studies appropriate to the level of analysis and the actor—typically the state—under scrutiny. At a systemic level, the debate among behaviorists is between those who contend that either a multipolar or a bipolar system is more stable, that is, less prone to armed hostilities and war. Rather than deduce whether one or the other is more war-prone, the approach favored by Waltz,43 behaviorists attempt to identify all relevant cases in which differently configured multipolar systems, including the limiting bipolar model, can be identified and categorized to test their respective dispositions to induce actors to use force and threats and to resort to war. War is also operationalized by universally applied counting rules. It is typically defined as a hostile exchange between states resulting in a specific level of deaths and casualties. This standardizes war statistically and ostensibly frees

the war event from historical circumstances not directly pertinent to the model or statistical test being applied. Such a science of international relations and, by extension, of security purports to go beyond what J. David Singer terms “the selective recall of historical anecdotes or, worse yet, … folkloristic hunches about how individuals allegedly behave….”44 Systems might assume a hierarchic structure or a decentralized system of units of approximately equal material power or some kind of ordering principle between these extremes. To date, systemic research studies of war proneness have scarcely been conclusive. However much behavioral researchers may share a common methodology and positivist assumptions about valid and reproducible data, they continue to differ over how they measure their specified variables, over the space and time and domains of actor behavior, and over the statistical models applied to often noncomparable data sets. These differences are no more acutely apparent than the choices made by researchers of what level of analysis of actor behavior they wish to measure and test. In contrast to systemic theorists, many behaviorists prefer to focus at a unit or state level of analysis. This does not preclude the incorporation of systemic research results into their depictions of actor behavior. These are assumed to be taken into account in the calculations made by the actors in their transactions. Borrowing directly from microeconomic expected utility theory, these theorists substitute the pursuit of power and its relevance to war as the utility to be maximized rather than the economic liberal stipulation of material wealth or welfare. States as rational actors are Page 560 → expected to estimate the probability of achieving their objectives in light of the structure of the international system and, most relevantly, of the countervailing power likely to be arrayed against them by other states. Decision makers are expected to always choose the strategy most likely to yield “the highest expected utility.”45 The utility of that strategy is demonstrated over time by how it performs. The predictions of actor behavior depend supposedly on objective determination, and not on the analyst having to guess or interpret the ideational values or psychological makeup of the actor or its capacity to process information about the real world. The strength and weakness of expected utility theory is that it is almost flawless in its post-prediction capacity since whatever results arise from strategic action are confirmed by these winning strategies. Still other researchers prefer to go below the state or unitary actor level of analysis to explain or predict war. Much attention has been devoted in the past decade to the proposition that democracies do not fight each other. This Kantian expectation, resurrected by Michael Doyle,46 has been bolstered by a proliferation of empirical studies that, despite their individual reservations, converge on this proposition. Three principal causes are assigned as explanation for this pattern of behavior. First, democratic ideology values cooperation and consensus over force and threats in making collective choices. Second, democratic institutions mutually restrain the struggle for power internal to the state. Legislative, executive, and judicial branches balance each other. These checks and balances are assumed to be projected in the exchanges of democratic polities since they are presumably socialized to accept limits on their power. Finally, pluralistic politics and frequently changing coalitions and alliances preclude a zerosum game in which contestants are permanently put at a disadvantage or are excluded from participating in the making and application of rules for the society. Ideology, institutions, and pluralistic politics create an open society where decisions are more transparent and where other states, most notably other democracies, are able to identify those responsible for what decisions a state makes and the constraints under which those choices are being made.47 All of these behavioral approaches reconceptualize actor behavior in terms of those methods and rules of evidence that can meet empirical and replicable tests. For behaviorists, unless cases or data can be reduced to universallyapplied methods and rules of evidence, testable generalizations across observations are precluded. From this research viewpoint, the cold war and the implosion of the Soviet Union are instances of a larger set of actor actions and reactions relevant to a pattern of behavior of interest to the researcher. The conceptualizations of actors as participants in a conflict of global proportions are subordinated to these methodological protocols and standardized rules for collecting, organizing, and measuring data appropriate to the propositions to be validated or rejected. Behaviorists expect not only that persistent work will identify propositions that hold over time and space but also that such knowledge will accumulate to form a body of testable, scientific knowledge about actor behavior. Few among the most ardent partisans of behaviorism would argue Page 561 → that the expectation of accumulated

knowledge has been realized or that it will be achieved any time soon.48 This shortcoming, accentuated by failure of a behavioral approach to anticipate the end of the bipolar system and the ascendancy of Western liberal market democracies, exposes behavioral approaches to the criticism of relevance and salience. The research designs of interest to the analyst are privileged over those of the actors engaged in a deadly struggle and over what these actors believe they need to know about the political conditions and constraints under which they act to succeed. However much simplification and research controls are necessary for a science of security to progress, this proposition-testing approach opens behaviorists to charges of having no theory to guide their research programs. Realist and constructivist critics reject the idea that this approach can produce a general theory of actor behavior of international relations or security. For neorealists, like Waltz, no list of propositions can sum up to a theory of state behavior. Rules and principles of behavior can well be identified, but they are to be understood as derivatives of a general condition of power and its distribution across units under conditions of anarchy.49 Similarly, the correlation between democracies and nonwar can well be observed and the properties of democracies, as noted earlier, can be cited as the causes for peace. But this association of democratic properties with nonwar is a correlation, not a theory explaining either why democracies allegedly do not fight or why a world society of democratic regimes would necessarily be peaceful in resolving differences. Constructivists argue that, whether explaining war or peace, a behavioral approach—no less than realists and neorealists—assumes as given what is to be demonstrated. This is true whether the focus is on why populations prefer democratic regimes over alternatives or on why states have arisen as central actors in world politics. Security as a value ostensibly lies in how the identities of social actors are constructed. From the perspective of this constructivist critique, international relations is fundamentally not about the material conditions of power under which actors behave but about the ideational and normative determinants of these actors and the meaning, significance, and value they attach to these material conditions, partially or fully as extensions of their identities. These ideational determinants are reflected in the culture, religion, and language of the actors. They infuse meaning into what most other theorists assume are the material and exogenous determinants of actor behavior. Constructivism Constructivism, like behavioral research programs, represents a broad array of theorists and scholars rather than a coherent school of thought. Constructivists challenge the narrow scope of the actors and factors associated with the rival paradigms sketched above. This challenge assumes two forms. The first is based on epistemological, methodological, and evidentiary grounds. Constructivists stipulate a “cultural-institutional environment” that shapes and shoves actor behavior.50 The norms attributable to actors by constructivist Page 562 → researchers are argued to have causal effect. Rejected is the assumption of realists and neorealists that reduces these justifications to rationalizations of coercive power and material capabilities. Nor do constructivists accept the notion of neoliberal scholarship that the norms arising from international regimes and institutions are purely functional and instrumental, principally intervening variables in the exchanges of egoistic actors and subject to the shared, if differentially-enjoyed, utilities that institutions provide. Both material and rational instrumental reductionism are asserted to be inconclusive in explaining state behavior or the state's conception of its interests, most prominently those bearing on its survival and security. As three prominent theorists explain: “it makes little sense to separate power and culture as distinct phenomena or causes: material power and coercion often derive their power from culture…. The issue is what accounts for power, not whether power is present.”51 The second line of constructivist attack is more fundamental and rests on ontological grounds. Most of the researchers already discussed assume the state and its interests as given. State interests, whether in pursuit of relative (realists/neorealists) or absolute (economic and institutionalist liberals) gains, are exogenously determined. Constructivists problematize both the state and its interests. Stress is placed on the formation of actor identities. Background cultural-institutional norms are alleged to define the identity of an actor or state. That identity then defines the interests of the state-actor. These interests are not uniform over time and space, nor is the state necessarily unitary. Why is the United States indifferent, for example, to the possession of nuclear weapons by France or Great Britain, both with substantial nuclear striking power, but feels threatened by the possibility that Iraq or North Korea might possess even one nuclear weapon? Why do the states of NATO form a Deutschian security community, dedicated to the peaceful resolution of conflicts even after the targeted rival, the Soviet

Union and its satellites, has passed from the scene?52 Thomas Risse-Kappen argues that the democratic norms of these states defined these states as sharing a collective identity and interest in the peaceful resolution of differences.53 States may exist under conditions of anarchy, but “Anarchy Is What States Make Out of It,” as Alexander Wendt asserts.54 The cultural-institutional environment, essentially nonmaterial and ideational in composition, can purportedly reconstitute and transform the identities of states, the interests they believe important, and what forms of power they employ in pursuit of these interests. Constructivists may be divided into two broad groups. “Light” constructivists concede the force of material factors in driving actor behavior. They do not reject outright the assumptions of other research paradigms. Rather, they insist on ideational factors, too, to explain outcomes in world politics.55 Case studies and process-tracing studies are relied upon to offset the rationalist-deductive and large-N statistical studies underwriting behavioral research or the ad hoc presumptions of realist and neorealist theory.56 “Heavy” constructivists, on the other hand, go further. The material conditions within which actor exchanges are embedded are radically subordinated Page 563 → in their causal effect to the meaning and value attributed by actors to these factors, whether military capabilities or economic and technological resources. The identities constructed by actors determine their interests, policies toward each other, and behavior. This subjective conception of social identities and the endogenous determination of actor interests that it implies all but cuts off heavy constructivists from mainstream social science and its attempt to describe, explain, and predict actor behavior. Erklären, or explanation, is totally sacrificed to Verstehen, or the subjective understanding of the components of actor makeup, thought, and action. This position risks burning all conceptual bridges to prevailing and competing social science research programs. Adopting a postmodernist position, more philosophical than social scientific, heavy constructivists assert the primacy of norms as power and the derivation of these informing and controlling norms as a product of the discourse and discursive practices of actors.57 No tree falls unless it is heard by human ears; humans are the measure of what they see and value. Constructivists of different persuasions and levels of commitment offer a range of explanations for the end of bipolarity and the self-destruction of the Soviet Union as a state. Some focus on the cognitive reformulation of the external environment by key elite decision makers, specifically Soviet premier Mikhail Gorbachev. They suggest that the political learning of the Soviet leader and his partisans induced a fundamental change in their understanding of the military power arrayed against the Soviet state (détente over unremitting conflict), the technological and economic determinants of sustained growth (perestroika), and the popular supports for exterior and interior reform programs (glasnost and democratization).58 Ideas count as critical intervening variables that direct the strategic choices of decision makers in adapting to their material environment or in their attempts to impose their preferences on other actors and the power structures within which they act. Still others focus on structural constraints, like the reassertion of nationalism and the delegitimation of the Communist regime throughout Eastern Europe and within the Soviet Union itself, to explain the collapse of the Soviet empire.59 Constructivism and its adherents are strong on identifying hard facts at odds with the assumptions and findings of other security paradigms. These researchers are less helpful in providing a coherent theory either of international relations or of security. Case studies provide rich, thick, and textured interpretations of actor behavior that would appear to highlight and validate the causal force of identities, the impact of identity on interests, and background and fundamentally subjective cultural-environmental structure or ideas, values, and norms in directing actor thought, decisions, and actions. However, these studies are not easily replicable in the same sense that large data sets can presumably be assembled in like manner, if all researchers use the same rules for categorization and counting relevant to the collection process and to the proposition being tested. Nor can the findings of a case or several cases be generalized over a large number of cases or actor decisions. Constructivism has no clear theory of Page 564 → agency in the way that rational actor or realist or liberal economic theory possesses. The former can assume that actors engaged in surviving and in pursuing their security interests will always prefer more power or gain than less or, if a liberal institutionalist, more absolute power and gain than less. Similarly rational economic decision makers always prefer more material goods than less. They are aided in this pursuit by markets that provide them with relative prices to allocate scarce resources either as

suppliers or consumers of goods and services. Constructivists do not have a fixed point from which to develop their research designs and findings. They do not posit a maximizing utility interest that orients actor behavior and dictates strategies and choices. The notion of utility itself is a property of identity and the interests flowing from identity. These criticisms aside, the value of constructivist contributions should not be underestimated or dismissed. The constructivist research program has widened the scope of theoretical and empirical research and underscored the discrepant outcomes of actor behavior that need to be reconciled with the countervailing research assumptions and paradigms of other approaches. Constructivists reintroduce history into the study of international relations. They problematize power, interests, identities, and ultimately the state itself. As Peter J. Katzenstein concludes, “The historical evidence compels us to relinquish the notion of states with unproblematic identities.”60

Conclusion So where does the tour d'horizon leave us? On the one hand, we might despair over this disarray of paradigms and competing research programs. Clearly, there is no one approach, one methodology, or one set of policy prescriptions at the disposition of actors to solve their varied, overlapping, and conflicting security problems. Indeed, the very notion of what actors are is critical for the study of security, and the forces shaping their behavior remain open questions. This brief overview reveals these shortcomings of prevailing security theory and research. On the other hand, and following Lakatos, pluralism serves us better than we might believe—or wish. Depending on the security question to be addressed, each of these paradigms, in greater or lesser measure, has something potentially worthwhile to say about the behavior of relevant actors. They can, arguably, assist actors in understanding their security problems and in solving them. In retrospect, it would appear that a realist and neorealist perspective had relevance during the interwar period and even thereafter as the cold war unfolded insofar as the interests of the Western liberal market democracies and their survival were concerned. What, for example, does liberal institutionalism or liberal economic theory have to say about genocide in Rwanda or ethnic cleansing in the former Yugoslavia or racial and religious civil war in Sudan or continuing Middle East hostilities between Arabs and Jews? Conversely, as the logic and power of market mechanisms and democratization have grown apace around the globe, formerly persuasive realist Page 565 → and neorealist paradigms have lost force. The creation of a security community in the northern hemisphere between the liberal democracies would appear to have relaxed if not definitively surmounted the security dilemma besetting nation-states and the nation-state system. These trends in the globalization of market and democratic behavior, and the exogenous and endogenous factors explaining actor choices that their decisions and actions imply, must still be reconciled with the continuing force of national, ethnic, and religious identities at odds with expected actor behavior. The notion of the sovereign, unitary state within a self-help system, however attractive as a logical construct, is progressively less persuasive and impressive if the domestic veil is lifted to reveal how state policy is made. It is increasingly less persuasive when the welfare demands of populations everywhere for “more now,” for larger and sturdier social safety nets, or for a greater say in constructing their security arrangements are taken into account in describing and explaining state behavior. Nor do these democratizing forces—economic and political—appear to be fully explicable solely by reference to the exchanges of isolated, egoistic actors. There is no golden age, as some would argue,61 to which analysts can repair when security theory and actor practice were seemingly in harmony. Nor is such a golden age clearly discernible on the horizon today. Yet never, too, have security studies been more healthy and potentially more relevant to the differing and contesting needs of a proliferation of major actors in world politics.

Notes

1. See Pierre Allan and Kjell Goldmann, eds., The End of the Cold War: Evaluating Theories of International Relations (Dordrecht, Netherlands: Nijhoff, 1992); John Lewis Gaddis, “International Relations Theory and the End of the Cold War,” International Security 17. (1992–93): 5–58; Charles W. Kegley Jr., ed., Controversies in International Relations Theory: Realism and the Neoliberal Challenge (New York: St. Martin's, 1995); Richard Ned Lebow and Thomas Risse-Kappen, eds., International Relations Theory and the End of the Cold War (New York: Columbia University Press, 1995). 2. See John J. Mearsheimer, “Back to the Future: Instability in Europe after the Cold War,” International Security 15 (1990): 5–57; John J. Mearsheimer, “The False Promise of International Institutions,” International Security 19 (1994): 5–49; Stephen M. Walt, “The Renaissance of Security Studies,” International Studies Quarterly 35, no. 2 (1991): 211–39; Kenneth N. Waltz, “The Emerging Structure of International Politics,” International Security 18 (1993): 44–79. 3. These points are developed at greater length in Edward A. Kolodziej, “Security Studies for the Next Millennium: Quo Vadis?” in Critical Reflections on Security and Change, ed. Stuart Croft and Terry Terriff (London: Cass, 2000), 18–38. 4. See Lawrence Freedman, “International Security: Changing Targets,” Foreign Policy 110 (spring 1998): 53ff. 5. Imre Lakatos, “Falsification and the Methodology of Scientific Research Programmes,” in Criticism and the Growth of Knowledge, ed. Imre Lakatos and Alan Musgrave (Cambridge: Cambridge University Press, 1970), 91–195. 6. William C. Wohlforth, “Realism and the End of the Cold War,” International Security 19, no. 3 (1994–95): 95. 7. See the symposium in Arms Control for these divergent perspectives. Stuart Croft, ed., Symposium on Security Studies, special issue of Arms Control 13 (1992).Page 566 → 8. Freedman, “International Security.” 9. Thucydides, The Peloponnesian War (New York: Modern Library, 1951), 330–37. 10. Hedley Bull, The Anarchical Society: A Study of Order in World Politics (London: Macmillan, 1977). 11. Thomas Hobbes, Leviathan (New York: Dutton, 1950), 104; see also Dennis H. Wrong, The Problem of Order: What Unites and Divides Society (Cambridge: Harvard University Press, 1994). 12. Carl von Clausewitz, On War (Princeton, N.J.: Princeton University Press, 1976). 13. Immanuel Kant, “Perpetual Peace: A Philosophical Essay,” in The Theory of International Relations, ed. M. G. Forsyth et al. (New York: E.P. Dutton and Company, Inc., 1950), 200–44. 14. David A. Baldwin, “Security Studies and the End of the Cold War,” World Politics 48, no. 1 (1995): 117–41. The distinction between a theory of security and strategy is central to the debate over what is security studies. See Edward A. Kolodziej, “Renaissance in Security Studies: Caveat Lector!” International Studies Quarterly 36, no. 4 (1992): 421–28; and Walt, “The Renaissance of Security Studies.” For a recent restatement of the centrality of strategy over security theory, see Colin S. Gray, Modern Strategy (Oxford: Oxford University Press, 1999). 15. The leading statement of the traditional realist argument is, of course, Hans Morgenthau, Politics among Nations: The Struggle for Power and Peace (New York: Knopf, 1948). 16. Robert Gilpin, The Political Economy of International Relations (Princeton, N.J.: Princeton University Press, 1987); Joseph M. Grieco, Cooperation among Nations (Ithaca, N.Y.: Cornell University Press, 1990). 17. Kenneth N. Waltz, Theory of International Politics (Reading, Mass.: Addison-Wesley, 1979). For a contrasting view see Karl W. Deutsch and J. David Singer, “Multipolar Power Systems and International Stability,” World Politics 16, no. 1 (1964): 390–406. 18. For a review of these criticisms and a thoughtful defense of realism, see Wohlforth, “Realism and the End of the Cold War” William C. Wohlforth, ed., Witnesses to the End of the Cold War (Baltimore: Johns Hopkins University Press, 1996); William C. Wohlforth, “Reality Check: Revising Theories of International Politics in Response to the End of the Cold War,” World Politics 50, no. 4 (1998): 650–80. 19. Marshall I. Goldman, What Went Wrong with Perestroika (New York: Norton, 1991). 20. Mearsheimer, “Back to the Future” Waltz, “The Emerging Structure of International Politics.” 21. Norman Angell, The Great Illusion (London: Heinemann, 1909). 22. Joseph Schumpeter, Imperialism (New York: Meridian, 1955). 23. John Mueller, Retreat from Doomsday (New York: Basic Books, 1989).

24. John Gerard Ruggie, “International Regimes, Transactions, and Change: Embedded Liberalism in the Postwar Economic Order,” International Organization 36, no. 2 (1982): 379–415; John Gerard Ruggie, ed., Multilateralism Matters: The Theory and Practice of an Institutional Form (New York: Columbia University Press, 1993). 25. János Kornai, The Socialist System: The Political Economy of Communism (Princeton, N.J.: Princeton University Press, 1992). 26. Adam Smith, The Wealth of Nations (Oxford, England: Clarendon, 1976). 27. Friedrich von Hayek, Individual Freedom and Economic Order (Chicago: University of Chicago Press, 1948); Friedrich von Hayek, The Fatal Conceit: The Errors of Socialism (London: Routledge, 1988). 28. Robert Gilpin, War and Change in World Politics (Cambridge: Cambridge University Press, 1981); Gilpin, The Political Economy of International Relations; Douglass C. North, Institutions, Institutional Change, and Economic Performance (Cambridge: Cambridge University Press, 1990).Page 567 → 29. Lester Thurow, The Future of Capitalism (New York: Morrow, 1996). 30. Peter Evans, Dietrich Rueschemeyer, and Theda Skocpol, eds., Bringing the State Back In (Baltimore: Johns Hopkins University Press, 1985). 31. Space constraints preclude an extensive review of the relevance of Marxist theory to world politics. 32. David A. Baldwin, ed., Neorealism and Neoliberalism (New York: Columbia University Press, 1993). This volume brings together realists, neorealists, and institutionalists who accept the assumption of rational egoists as a common beginning point for theory. They diverge along several dimensions, among the most important being the explanatory power of institutions in shaping state behavior. 33. Thomas Schelling, The Strategy of Conflict (New York: Oxford University Press, 1960). 34. Robert Axelrod, The Evolution of Cooperation (New York: Basic Books, 1984). 35. Ibid., 176. 36. Ronald H. Coase, “The Nature of the Firm,” Economica 4 (1937): 386–405; Ronald H. Coase, “The Problem of Social Cost,” Journal of Law and Economics 3 (1960): 1–44. 37. Robert O. Keohane, After Hegemony: Cooperation and Discord in the World Economy (Princeton, N.J.: Princeton University Press, 1984). 38. Joseph S. Nye Jr., Bound to Lead (New York: Basic Books, 1990). 39. Judith Goldstein and Robert O. Keohane, eds., Ideas and Foreign Policy: Beliefs, Institutions, and Political Change (Ithaca, N.Y.: Cornell University Press, 1993). 40. For example, see the Correlates of War Project, J. David Singer and Melvin Small, Resort to Arms: International and Civil Wars, 1815–1980 (Beverly Hills, Calif.: Sage, 1982). 41. Wohlforth, “Reality Check.” 42. Manus I. Midlarsky, ed., Handbook of War Studies (Boston: Unwin Hyman, 1989), xx. Midlarsky's emphasis. 43. Waltz, Theory of International Politics. 44. J. David Singer, “System Structure, Decision Processes, and the Incidence of International War,” in Midlarsky, Handbook of War Studies, 20. 45. Bruce Bueno de Mesquita, “The Contribution of Expected-Utility Theory to the Study of International Conflict,” in Midlarsky, Handbook of War Studies, 144. 46. Michael W. Doyle, “Liberalism and World Politics,” American Political Science Review 80 (December 1986): 1151–69. 47. Bruce Russett, Grasping the Democratic Peace: Principles for a Post-Cold War World (Princeton, N.J.: Princeton University Press, 1993). 48. See, for example, Singer, “System Structure, Decision Processes, and the Incidence of International War,” 20–21. 49. Waltz, Theory of International Politics. 50. No attempt is made to cover the vast and growing literature that can be associated with the constructivist research program. Valuable as a starting point is the edited volume by Peter Katzenstein, notably chapters 1 and 2, which cite most of the relevant literature up to its publication. Peter J. Katzenstein, The Culture of National Security (New York: Columbia University Press, 1996). For more recent analysis see John Gerard Ruggie, “What Makes the World Hang Together? Neo-Utilitarianism and the Social Constructivist Challenge,” International Organization 52, no. 4 (1998): 855–85; and Alexander Wendt, Social Theory of

International Politics (Cambridge: Cambridge University Press, 1999). 51. Ronald L. Jepperson, Alexander Wendt, and Peter J, Katzenstein, “Norms, Identity, and Culture in National Security,” in Katzenstein, The Culture of National Security, 40. 52. Karl Deutsch, Political Community and the North Atlantic Area (Princeton, N.J.: Princeton University Press, 1957). 53. Thomas Risse-Kappen, “Collective Identity in a Democratic Community,” in Katzenstein, The Culture of National Security, 357–99.Page 568 → 54. Alexander Wendt, “Anarchy Is What States Make Out of It: The Social Construction of Power Politics,” International Organization 46, no. 2 (1992): 391–425. Milner also develops this line of analysis. Helen Milner, “International Theories of Cooperation among Nations,” World Politics 44, no. 2 (1992): 466–96. 55. For example, see Michael Barnett and Emanuel Adler, eds., Security Communities (Cambridge: Cambridge University Press, 1998). 56. For examples, consult Martha Finnemore, National Interests in International Society (Ithaca, N.Y.: Cornell University Press, 1996); and Audie Klotz, Norms in International Relations: The Struggle against Apartheid (Ithaca, N.Y.: Cornell University Press, 1995). 57. Richard K. Ashley, “The Poverty of Neorealism,” International Organization 38, no. 1 (1984): 225–86. 58. Robert G. Herman, “Identity, Norms, and National Security: The Soviet Foreign Policy Revolution and the End of the Cold War,” in Katzenstein, The Culture of National Security, 271–316. 59. See the essays in Lebow and Risse-Kappen, International Relations Theory and the End of the Cold War, for examples of these interpretations of why the cold war ended, which putatively contradict competing explanations. 60. Katzenstein, The Culture of National Security, 23. 61. Walt, “The Renaissance of Security Studies.”

Page 569 →

SECURITY AND PEACE Understanding, Production, and Work Style Davis B. Bobrow The most attractive accomplishment of security and peace studies and policies would be for them to become historical curiosities akin to alchemy, Victorian-era plumbing, or vanquished diseases. A second best would be signs of progress on that road marked by improved understanding and early diagnosis, and—even better—more available and less costly means for prevention and treatment, containment and cure. To extend the medical metaphor, we would then see in the present or in confident prospect reductions in the incidence and severity of insecurity, destruction, casualties, and deaths, and in the opportunity costs of measures to achieve such reductions. From the second-best perspective, these are in some ways the best of times for security and peace studies and policies. In other ways, they are far from that and leave a lot to be desired. These two aspects of our current intellectual and policy situation suggest lessons that are sobering in their implications for the limits on our understanding of security and peace and even more so for the application of our understanding to produce those collective accomplishments. Of no less importance, this duality challenges us to adopt a work style appropriate both to what we have learned and produced and to our continuing limitations—and to socialize successor generations of security and peace experts into it. The following pages begin with a crude summary of what seems to me to be “good news” about the point to which we have come and then turn to a “darker side” suggesting how far we have to go. These lead me to identify some persistent realities that should be squarely faced up to and to offer some suggestions on how we should deal with our unfinished agenda.1

The Good News We can readily recognize the abundance of intellectual and policy effort devoted to understanding and producing security and peace over the last half century and six now widely accepted expansions in the scope and resources of the field. Page 570 → First, even the most cursory review of our major journals shows in recent decades a cornucopia of “prisms” for considering security and peace that apparently go beyond traditional power-centered and thus competitive (realism) and normatively centered and potentially cooperative (idealism) perspectives. Prominent entrants in the security and peace theory sweepstakes now wear rational choice, institutionalism, critical security, gender, power transition, domestic politics, failed state, democratic peace, and ideational perspective colors. The proliferation of prisms has been accompanied, particularly in the last two decades, by five additional expansions of scope and increases of attention and resources: issue domains, generic types of actors, relevant specific actors with asymmetric agendas and assets, technology-enabled interactions, and information and information-processing tools. Discussions of security and peace have come to regularly encompass issue domains other than the politicalmilitary such as economic, environmental, demographic, public health, sociocultural continuity, and fundamental political forms. That expansion has brought with it attention to a widening range of generic types of actors inside and outside of governments. Consider the now rather accepted inclusion of central government civil ministries, subnational public authorities, firms and in particular multinational corporations, nongovernmental organizations (domestic

and transnational), expert (“epistemic”) communities, international governmental organizations (regional and global, functionally specific such as international financial institutions or multipurpose such as the G-7). The end of the cold war has also brought increased attention to a larger if changing number of specific actors (with proper noun names) with more asymmetric agendas and capabilities for pursuing security and peace. Those players and purposes often were not, at least in prevailing American frames, important in the bipolar construct or were only considered in relation to it with considerable distortion. Asymmetric agendas now and for the future combine with asymmetric assets and vulnerabilities derived from the expansion discussed next. Technology innovation and diffusion have changed the world in terms of damage-inflicting capacity, communication, mobility, transparency, and financial and commercial globalization. Their joint effects have massively increased the variety, speed, reach, visibility, and relevance of interactions across borders. As a consequence, the behaviors and the perceptions, the independent and intervening variables if you will, that matter for actual and recognized states of security and peace have proliferated. Barriers to entry into and competition in security and peace arenas have been lowered while time to adapt to ease of entry in analytic or policy terms has shrunk. A final type of expansion involves resources to help in coping with or even being aware of the sets of expansions just noted. Compared with several decades ago, researchers and analysts of security and peace matters are enormously advantaged in terms of the amount and timeliness of information. The costs required for exploiting and processing that information have been Page 571 → sharply reduced. Retrieval and pattern identification have been facilitated and with them the feasibility of promptly flagging anomalies and changes. And surely the expert cadres and skill repertoires applicable to problems of security and peace have grown enormously within and across functional disciplines and geographic area specializations. In sum, those who would research, manage, and even teach security and peace matters have entered in a period of “riches” with respect to concepts, phenomena, information, and human capital. All that bodes well for the secondbest accomplishment suggested earlier.

The Darker Side One surely can recognize those riches and still have serious reservations about the extent to which our capacities to understand and produce security and peace have substantially improved. There are in general two reasons for such reservations. First, some of the expansions may be more matters of rediscovery or relabeling (“branding”) than of genuine, progressive innovation. Second, the improvements may not have kept pace with the challenges posed by the six expansions. The first reservation implies that we may not be as absolutely better off in comprehending and dealing with security and peace problems as many think. The second implies that even if we are, those achievements may in a relative sense lag behind ramifying challenges to security and peace. I will briefly address the first reservation and then turn to the second. To a significant extent, the first five expansions involve resurrection and relabeling rather than genuine discovery and innovation if we extend our recollections into (a) a more extended past than the several most recent decades and (b) a broader set of interpretations and policy emphases than those of the cold war superpowers. To deny this amounts to confusing cold war emphases in the American security and peace studies community with earlier and even contemporaneously different foci of attention. Earlier periods and contemporary policy intellectual communities were not devoid of recognitions of the role of institutions and political forms national and international, ideas and problem-framing constructions, critiques of state power and its intellectual servants, or applications of economistic rationality. Security and peace were held to involve matters of economics and commerce, natural resource scarcity and control, identity-based aspirations and sociocultural survival concerns, natalist and public health issues, and domestically-oriented political opportunism. In such frames, generic actors other than central government military departments were also treated as securityand peace-relevant subjects ranging from arms merchants to transnational financiers to religious faith movements

and institutions. There surely was attention to multilateral institutions such as leagues and concerts. Specific, named actors other than the largest and most powerful states did not escape notice including networks. They were recognized to have on occasion asymmetric agendas and assets for inflicting harms on the conventionally more powerful involving resolve, Page 572 → agility, and guile. After all, use of the slingshot against the heavilyarmed warrior goes back a long way. Finally, the current significance of new technologies and their applications is but the most recent of many such transformative waves. Whatever one's views on their degree of genuine innovation, recent expansions or resurrections are to be applauded. Nevertheless, they do not necessarily outweigh the evolution of intellectual and policy challenges to security and peace. More and different schools of thought make it less likely that important elements of the security and peace problematique will go unrecognized. They do not, however, guarantee that a wise synthesis will occur. Instead, they may defer it by focusing effort on deepening each school and finding grist for its narrow-purpose mill by treating only particular aspects of the other five expansions. That would not be unhelpful for the production of security and peace if the schools have harmonious implications for all or even most relevant parties, means, and ends—but that is questionable. Consider, for example, two of the most intriguing, potentially powerful, and empirically relatively well-based contemporary perspectives—power transition and democratic peace theories. The first prescribes unipolarity globally and regionally and thus policies to achieve and maintain single-state dominance alone or through coalition leadership. It has an obviously realist core emphasizing shares of traditional power assets, and it focuses on security and peace at the level of states and combinations of states. It raises questions about the willingness of others to provide the leading state with the benefits it would want, or perhaps even need, to play such a role. Power transition theory leaves open the possibilities of genuine conversion to the conventions espoused by the prevailing leading state. It also allows for the possibilities that others might be willing to do so only with a sigh so long as they see no attractive choice, or even adopt a deliberate stratagem of compliance to secure the assets to catch up with and eventually surpass the leading state globally or regionally. The first may well call for the leading state to be notably reliable and generous in its provision of security and peace to current and potential followers. That, however, poses a requirement of the domestic politics of the leading state, but why and how that requirement will be met lies outside of power transition theory. The second, as with the contemporary engagement of China, poses eventual security and peace risks. If a power transition is eventually inevitable, we are left to ponder the feasibility of transforming the probable successor prior to transition. Why should a realist-minded successor informed about power transition theory believe that transformation efforts by the current leading state are primarily intended to smooth a transition rather than avoid or postpone one? Democratic peace theory (with its idealist core) in contrast prescribes creation of a world of essentially compatible, if somewhat competitive, states based on sufficiently common norms, interests, and domestic politics to damp security threats and buttress peace preferences. A “community of democracies” is held to be at least a latent security community for its members. Page 573 → While the focus is on state political forms, those recommended do offer in a probabilistic fashion conditions conducive to (but not guaranteeing) security and peace for subpopulations and even individual citizens. For security and peace, there are several significant problems since we begin from a world of substantial power inequalities and less-than-full democracies. One is that even though established democracies seldom war with each other, powerful democracies do on occasion coerce (threaten at least the nonmilitary security of) less powerful ones. How severe can class disparities between democracies be before ostensible security communities are rather imperial and tributary state systems? A second is the less than short and linear path to the sorts of democracy the theory envisions. Whatever one labels such partof-the-way-to- (or from-) democracy cases, their elites may have incentives to engage in actions and rhetoric whose immediate consequences are to lessen the national, regional, and international degree of security and peace. Indeed, such a nonprogressive course may become more rather than less likely with the initial onset of plebiscitarian democracy.

In the current world, the policy and policy intellectual communities of the American leading state seems to simultaneously subscribe to the ideas of both power transition and democratic peace theory. That combination does not inherently lead others to attribute full respect for their security and peace to those who hold and act on such a vision. International audiences may instead see a realist wolf draped in an idealist democratic cloak. Even if the illustrative theories were more compatible than seems to be the case, they—like most of our other candidate conceptual perspectives—provide quite incomplete prescriptions. Imagine our reaction to a doctor who told us to take a particular medication without saying how much of it, for how long a period of time, and without warning and countermeasures about potential adverse side effects. Our likely demurs would not be trivial, and might easily turn to anger and pursuit of compensation if the treatment turned out to at most partially deal with symptoms but not the root pathology that is at the source of our troubles. If “prisms” are equivalent to diagnostic specializations, that undesirable possibility may well be increased. Of course, our “patient” may find herself given conflicting diagnostic advice (as with the “too fast” and “too slow” postmortems on the recent breakdown of the Israeli-Palestinian peace process). That situation poses risks of “compromise” mixing of emphases depriving each of efficacy, and not just of a mistaken choice between alternatives. What of the other five previously mentioned expansions? That in issue areas surely is warranted as it comes to more closely approximate the stocks and flows that pose threats to individual, group, national, and international safety, and that singly or in combination can generate movement toward conflict or cooperation. Yet that expansion may pose a particular danger, and certainly poses some very difficult questions. The danger is of course that of securitization with its basic features of threat from the behavior and intentions of foreign others, assertion of privileged claims on resources, and warrant for waivers from the checks and balances institutional and normative of otherwise appropriate political processes. Surely a multidimensional Page 574 → conception of threats (and thus of threat agents) to security in the first instance only multiplies rather than eliminates the classic security dilemma. It does indeed make cooperative security and (impure) public goods provision more needed, but the domain expansion does not in itself guarantee increased provision in general or in a particular domain. As security and peace come to be conceived of in a multifaceted way, some difficult questions become more pressing. What are the dynamics of cause and effect for each facet? What about interactions between facets? Consider the difficulties of sorting out the cause-and-effect thicket of relationships between economic, environmental, demographic, public health, and democratic governance domains. Even if the general sign of relationships is understood, that does not suffice to direct intellectual and policy resource allocation at the margin. We still lack an encompassing metric across security facets to weigh trade-offs between them. Simply adding to the mercantilist duo of power and prosperity enlarges our agenda more than it clarifies how to pursue it in real time, actual space, and available budgets. Expansion of the types of actors relevant to security and peace brings with it not only the benefits of checks and balances and multiple point-of-entry systems but their disadvantages as well. After all, the expansion of types rests on the view that no single one should be expected to fully control what happens about security and peace. What are some of the darker-side issues? One is that of whether similarity of security-and peace-related preferences (of outcomes, means, and processes) within a generic type of actor exceeds differences. To the extent it does not, focusing on the variety of types may be misleading. Suppose instead that each type is rather homogeneous for our concerns. We then need to consider the possibilities for, transaction costs of, and likely content of deals struck between types. Are such arrangements easier for some facets of security, and if so which ones? If so, do we need to guard against unduly focusing on them and slighting others more crucial for security and peace? Under what circumstances are the deals likely to have desirable characteristics of anticipation, specificity, bindingness, and sustainable commitments that would significantly enhance security and peace? The warranted awareness of more and more varied particular specific actors with the associated increases of asymmetries in agendas, assets, and vulnerabilities imply that several kinds of problems for security and peace are more rather than less difficult. Analytically, identifying a security and/or peace Pareto frontier or near-frontier for the relevant parties becomes harder. Even if identifiable, the requisites for its pursuit mount in terms of temporally-synchronized internal policy climates, acceptable transaction costs, equivalent risk-taking, common perceptions of precedents, and compatible interpretations of events and moves in terms of confidence-building or

lessening. Accordingly, our needs for understanding and prediction ramify in general and in terms of “local knowledge” about what may for many have been obscure actors. In logic, those challenges could be lessened by defining some actors, assets, and vulnerabilities as exogenous to peace and security systems, a sort of intellectual nonrecognition policy. That would confine intellectual and Page 575 → policy challenges to what amount to preferred others and preferred policy means and strategies based on prior investments and achievements. In reality, that choice may not be available. Cognitive and affective stances by some parties to security and peace matters cannot by themselves forestall others from having and pursuing their particular security agenda and claims, using whatever seems to them to be their most advantageous means, or adhering to culturally rooted visions of a good peace or a good war. That leaves us with needs for informed and creative adjustment and adaptation (not to be confused with appeasement or total relativism). Meeting those needs calls for extending knowledge and insight about a substantially increased number of specific actors and about the offensive, defensive, and deterrence possibilities associated with the universe of asymmetrically-available assets and pertinent vulnerabilities. Interactions enabled by the new technologies surely can facilitate security and peace in important ways. They also challenge established intellectual and policy conceptions about security and peace. That applies to the importance of particular general variables such as geographic distance and the mass of an actor's economic and population base. That also applies to the degree of overlap between the set of sovereign states and the universe of securityand peace-relevant generic and specific actors. In each of these examples, declines in relative importance are for more stable, slow-to-change elements of security and peace problems with more established management repertoires. We then face more imminent and significant obsolescence of intellectual and policy structures unless we find ways to generate less fixed, routinized, and inertial institutions and analytic frameworks. Even the sixth expansion, that in information availability and professional expertise, is not necessarily an unmixed blessing. The possibilities of professional tribalism in sociology of knowledge terms versus synoptic judgment may well be increased analogously to the darker prospect from prism proliferation. The availability of larger and larger volumes of information may be met primarily by gatekeeping selection and compression mechanisms (individual, organizational, and software) unrepresentative of the world marked by the other expansions. The users of information may be unaware of the biases built into those mechanisms, let alone able to hold gatekeepers accountable. Lower barriers to information dissemination may lend themselves to advertent (disinformation and information management) or inadvertent reductions in information quality.

Some Lessons (We Should Have) Learned Any selection of lessons from the previous good news and darker-side observations is, of course, substantially personal and subjective, but then so are most efforts at brief retrospective and prospective treatment of a highlyactive and socially-important basic and applied research field. With that caveat, here are mine. Beyond an aura of desirability, security and peace are and will remain contested or captured concepts. There will not be consensus on their operational Page 576 → content, sufficient conditions, or an encompassing, umbrella metric to measure the degree to which security and peace exist or the extent and rate of movement toward or away from them. As Wolfers suggested some decades ago, the basic conceptions are “ambiguous.” That ongoing contestation has two major foci. The first involves disagreement about what aspects of human affairs individual and collective must be present to label a situation as one of security or peace. The second involves disagreement about for whom those chosen aspects held to amount to security or peace have to be present. The second dispute goes on about levels of analysis categories (e.g., individual, group, nation) and sociological and economic markers (e.g., race, religion, cohort, and class)—as well as about the divisibility or indivisibility of security and peace in relation to different nation-states and groups. Security and peace concepts can of course be captured at particular points in space and time. Such local and temporary consensus reflects and for a while may reinforce success by some fraction of actors and institutions in

privileging their constructs of security and peace (and thus their claims for attention and resource priority). The resulting conventional wisdom givens for academic and policy specialist discourse amount to political outcomes more than natural laws, as is also the case for security-and peace-linked prevailing distributions of attention and resources. Political competition and distributional controversy domestically and internationally are and will be present far more often than not. Accordingly, so too will be efforts to contest and capture the operational meaning and the indicative leading, concurrent, and lagging indicators of security and peace. At the same time, there is and will be widespread agreement that a necessary condition for security and peace is not met for those who are currently or in prospect targeted for physical coercion. Yet even that agreed principle has substantial ambiguity and controversy with regard to differentiating less-than-unconditional, unqualified, and interstate situations. Does the necessary condition include absence of physical coercion and coercive threats by governmental and movement elites against their own citizens and members or placing them in harm's way? Does posing threat to others tend to stimulate beliefs that the poser is also a target that has yet to attain security or peace? Whatever the answers to these questions, we have yet and are unlikely to achieve agreed rules to assess the degree to which the necessary conditions for security and peace are unmet—descriptive or normative accounting rules to weigh the present versus the future, the probability of pain being inflicted and its likely severity, or the shortfall for self relative to that for others. As with positive conceptions of security and peace, those about their absence also become subjects for capture, a quest in which the contenders rarely start from a level playing field. The views and partisans advantaged tend to follow from interpreted historical legacies and recent experiences. Absent gross failures, dramatic anomalies, or pervasive fatigue, established interpretations tend to prevail in academic and policy circles. Already-accepted views and their resource-advantaged custodians can dominate interpretations. Page 577 → They are well positioned to both frame the agenda for security and peace studies and analyses, and then to draw certification from that frame. Established conventions about insecurity and nonpeace are extraordinarily resilient and evidenceresistant. While they may eventually lose credibility, they can for an extended period deprive alternatives of the momentum that would make them persuasive. The odds work against timely recognition of grounds for change and favor elaboration and interpretive bolstering and tuning of previously established insecurity and nonpeace conceptions. The six expansions cannot be relied on, and may on balance interfere with, the achievement of domestically and internationally widely-agreed and timely understandings about the determinants of security and peace and the implications of policies to produce them. Rather than relying on help from those expansions, we should accept a continuing need to adapt to them. More specifically, functional types of issues are not as readily separated from each other as tractability for analysis or policy bargaining would have us desire. Decomposability and issue-isolation is a matter of fragile and tenuous consent by multiple, varied parties (for example, dual-use technology transfer as a military or economic matter, or emissions reduction as an environmental or economic security matter). Specific problem-solving between a pair or small number of states is hard to contain from becoming a precedent in the minds of others (for example, U.S. “rewards” to influence North Korean nuclear programs). A good fit between signals intended and signals interpreted becomes less likely with a potential security and peace negative spiral. That is hardly a new problem, but it may well already have been worsening with the multiplication of generically- and specifically-different actors and accelerated signal transfer and interpretation. Without subscribing to a thesis of civilizational conflict, one can accept one of civilizational and subcivilizational miscommunication. The signals an actor in the security and peace problematique sends and believes it has received are substantially affected by prior judgments about where it sits, how and by whom it was placed there, and the agenda and capabilities of the other parties in the communication chain. All too readily what is for one party a defensive measure will for another be an offense-enabling one; an effort at compromise based on strength be a sign of weakness or a trap; an act of acceptance and empowerment be a move to dominate. Whatever tenuous confidence we have had in the fidelity of signaling systems needs to be downgraded and signal strength and interpretation clarification given more attention.

The technology and information expansions warrant reduced confidence in the feasibility of segregating issue domains, narrowing precedents, and clearly signaling intentions, incentives, and disincentives. Even before the recent expansions, history provides substantial cautions about overoptimism in those respects. It would make security and peace studies easier if highly general theories could be readily operationalized and specified for specific situations. It would also be welcome for security and peace basic research and applied Page 578 → analysis if there was some standard inventory of readily-known holdings that would allow us to calculate or simulate reliably (as distinct from statistically significantly) the outcomes and payoffs in particular cases of security conflict and competition and of pacts to transition along a hot war-cold war-cold peace-warm peace continuum. Such full or near clairvoyance would have moved us much closer to the most complete achievement with which this essay began than we actually have come. For now at least, we should bear in mind that the acid tests of security and peace understandings and production (or insecurity reduction) take place in specific relational situations with varying contexts. There is then an imperative for situational specification. The evolution of those situations cannot be relied upon to conform to standard, easily measured portfolios of assets. Such assets are often neither universally applicable or actually employed. “Davids” can and do on occasion frustrate, penalize, and get at least some of their way over “Goliaths” for at least a while. Strategic theories and the security and peace policies and capabilities derived from them necessarily assume that some imaginable relational situations are relevant and that others are not. Those assumptions may be warranted on historical or statistical grounds or on shares of assets already thought to be crucial. They do, however, leave their adherents open to unexpected and only belatedly-recognized developments. The cold war record is replete with instances that, at least for the United States and other advanced industrialized countries, were not even on their strategic-planning radar screens and whose local dynamics were not well understood in a timely fashion, or were even denied significance while insecurity mounted. If anything, the expansions make timely awareness and understanding of security- and peace-relevant specific relational situations less rather than more likely for several reasons. The first is that there are many more and more varied possibilities that compete for attention and quick responses. Not surprisingly, all that makes reliance on general diagnoses and prescriptions rather than case-specific ones more appealing. Yet those general approaches may not fit well with particular relational situations however benign the intentions with which they are followed (e.g., the independence referendum for East Timor or democratization of Indonesia), and may even if only for a while trigger a loss in security and peace in being and in prospect. There are nontrivial temptations for the defenders of any prevailing strategic or scholarly fashion to defend it against anomalous relational situations, or at least to attend primarily to those more readily accommodated within such fashions. If anomalous situations can no longer be ignored, processes go to work to treat them as matters to be dealt with briefly so that one can soon get back to core business, or to contend that they lack substantial security and peace relevance (“no vital national interest,” no cumulative knowledge for an academic sect). In an internationalized, and perhaps globalized, era of security and peace, outsiders to local situations may even assume that apparent progress in the aspect salient to them can be equated with progress in security and peace from the perspective of each local party. The appropriateness of such judgments needs the most careful Page 579 → scrutiny especially for what amount to multiparty, nested relationships. Consider the Oslo process for various Palestinian and Israeli groupings or the recent regime change in Belgrade for Kosovar Albanians and Montenegrins. Suppose we accept the basic loss-aversion stance of prospect theory with its associated propensities for risk-taking and cost-bearing. It follows that we need to understand what loss prospects are at work for the parties in particular relational situations rather than assuming them to conform to the roster envisioned by outsiders. We of course inhabit a world where adherents and mission custodians of prevailing strategic theories, policy packages, and academic schools have something to lose in material and nonmaterial (status and deference) terms. We should not expect such stakeholders “to go quietly” in the face of what are for them isolated and transient security and peace anomalies.

Finally, we should have learned to have serious reservations about the likelihood that even well-resourced governments or security and peace expert communities will be very good at anticipating and adjusting to discontinuities and asymmetries—especially ones that pose greater complications for security and peace be it as opportunities or threats. Several examples at different levels of generality are indicative. Behavior during the period between World Wars I and II is not encouraging, and that period seems to me relatively if imperfectly comparable to the present. Like the present, it had radical changes in military technology in destructive capacity, spatial reach, and operational domain (air); attempts at multilateral arms limitation; embryonic supranational governance institutions; a communications revolution (radio); communal assertion after the collapse of imperial systems; new and fragile nation-states; democratic hopes; major nations seeking to regain what they held to be their rightful international standing and redress historical grievances; and economic shocks with contagious international effects. Reliance on extrapolation from trends, essentially reliance on continuity, had at best a mixed record even during what some now think of as the relatively comprehensible and manageable cold war period. Contrary to claims for extrapolation (as in the Hart-Rudman U.S. Commission on National Security/21st Century), a projection from 1956 to 1981 would have missed or at best tardily recognized the Sino-Soviet split, nuclear parity with the Soviet Union, the economic rise of East Asia, great-power intervention failures (Vietnam, Afghanistan), and strategic arms control. Even for the actor(s) receiving enormous American attention, the Soviet Union and associated Warsaw Pact, most analysts missed the dynamics leading to collapse and tended to overestimate military and military technoindustrial performance. Nor in retrospect was our after-the-fact comprehension of hallmark, critical events for academic learning and policy (and policy process) prescriptions about crisis management all that impressive. Consider the often-discussed Cuban missile crisis of 1962. For several decades, the national governmental parties had only an imperfect understanding of each other; American nongovernmental researchers knew little of significant Page 580 → U.S. actions of which the Soviet leadership was probably aware; and the highest American officials were probably unaware of some orders not complied with and some steps they had not authorized toward a general war posture. We can be misled by postmortems and not just by extrapolations. It is important to recognize that the phenomena underlying cautions of the sorts just illustrated are far more intractable than is suggested by allegations of deliberate deception. Government officials do of course fail to tell the truth on occasion (e.g., the Tonkin Gulf incident and perhaps the more recent bombing of the Chinese embassy in Belgrade). Security and peace scholars are humanly prone to focus on evidence supporting what they sincerely view as conceptually, empirically, and normatively desirable interpretations. The more fundamental difficulty is that the truth often is hard to know about recent and current security and peace matters, and there were and are abundant incentives not to admit that given security and peace stakes.

Some Suggested Emphases The previous good news, darker-side, and lessons learned opinions do not yield a complete set of recommendations for “best practices” for understanding and producing security and peace. They do, however, suggest several elements that should be emphasized. The most general is humility and modesty of claims. Those principles should be put into practice by conceptual pluralism, attention to situational specifics, and acknowledgment of uncertainties and conditionalities. While in some ways banal, this suggestion stands in contrast to the often excessive claims of partisans in current “security culture wars.” A modest stance involves less attention to establishing the dominance of any single theoretical prism, issue domain, or generic type of actor. Instead, we would recognize that at some times and some places each of the diverse elements of those categories has some relevance—while no single element tells the whole story before or after the fact. Rather than advocating any one underlying strategic conception, we would recognize the situationally dependent implications of, for example, offensive dominance, robust defense, deterrence, coercive

diplomacy, preventive diplomacy, and the cooperative avenues of public goods provision and unilateral and multilateral self-restraint. Rather than arguing about exclusivist superiority, we would start from the possibility of simultaneously relevant insights from, for example, realist, liberal, and constructivist perspectives. We would not opt for the primacy in general of any one of rational unitary actors, organizational routines, domestic politics, or intraelite relationships. Nor would we pursue contentions that for the present and future predominant importance lies with states or firms or networks; unilateralism or multilateralism; or regionalization or globalization. We would instead pursue the characteristics and incidence of situations with different mixes of the elements of each set. Another aspect of modesty and pluralism calls for rejection of disciplinary dominance. That is, we should not equate the factors that explain, predict, Page 581 → or produce particular security and peace problems or outcomes with the turf of any single discipline. This enjoinder applies within and between the several social science disciplines and those of natural science and engineering. Substantial emphasis should go to bridging across types of expertise and informed appraisal of the limits of knowledge and consensus in each of them. Beyond humility and modesty, three aspects of professional practice merit more emphasis than they often get. The first, science and technology (S&T) literacy, requires serious departures in the education of security and peace specialists, especially as conducted in most political science departments and international affairs professional schools. The other two are more immediately feasible in those venues. Many possibilities for security and peace have an important S&T element be it in independent/intervening /dependent variable terms, or those of danger/safety/reassurance. While obvious with respect to weaponry in the so-called “old” military security agenda, it is no less true for the so-called “new” security agenda of information warfare, terrorism, trafficking, and peacemaking. S&T has substantial relevance for economic, environmental, demographic, and public health security. For each of these, S&T literacy consists at a minimum of four sorts of understandings. The first consists of the vulnerabilities of advanced technologies to primitive technologies and their adaptive use. The second involves differences between technology performance in a controlled or favorable environment (the laboratory, prototypes, managed tests, conducive theaters or adversaries) and performance in the far messier and suboptimal real world. The third emphasizes time in terms of obsolescence rates relative to how long it takes to procure, field, and countermeasure innovations and trends. Finally, account needs to be taken of global diffusion of S&T competencies and capabilities to and through nonstate as well as state actors. More optimistically, there are two emphases arguably well within our reach with more encouraging implications for understanding and producing peace and security or at least curtailing insecurity. While neither has been without its advocates, each has often been ignored or received only lip service. The most obvious area for greater emphasis would have us make more use of available intellectual resources and legacies. With regard to work in the past (a past of both centuries and decades), more awareness of legacy contributions would at least reduce energy consumed with results in which revival exceeds discovery. Greater attention should also be directed to security and peace analyses from experts outside of the American-centered research and policy communities and, in particular, their work primarily intended for and consumed by audiences outside of the United States, United Kingdom, Canada, and Australia.2 Finally, several themes in the legacy stock of security and peace studies seem to me to have particularly helpful uses for the future of security and peace studies. These themes, featured in the applied systems analysis approach of Albert Wohlstetter, merit greater attention because they face up to and engage the sorts of complexity provided by the expansions discussed earlier rather than trying to eliminate or evade it. Page 582 → First, the systems approach treats a given security or peace problem as a relational situation involving subsystems interacting with each other over time. Each subsystem has both material (economic, geographic, S&T) and

sociopolitical (human and organizational including belief systems) elements. Convenient as it might be, analyses are prone to misjudgments if they ignore or take for granted similarity or permanence of particular subsystems or their elements. International security and peace studies are not replete with lawlike regularities although there are some probabilistic relationships, rules of thumb, and heuristics. Those are not to be ignored, but their applicability to a particular system or subsystem needs to be established rather than assumed. One requirement for doing that involves careful situational operationalization of their terms. Applications of the applied systems approach should feature “opposed systems design” and “end-to-end analysis.” That holds for explanation and prediction about, as well as for prescription of interventions to affect, particular systems. The world of opposed systems design is not necessarily one of conflict, but it is one of differing and less than fully compatible preferences. Otherwise there would not be a security or peace problem. As a system operates over time, the parties in it engage in a series of interactions based on recognitions and expectations about themselves and each other as seeking to achieve nonidentical preferences. Each resorts to instrumental strategies and moves that have an international element of blocking, countering, diverting, or shaping those of others. No party is assumed to be inherently passive or generally indifferent to outcomes or to the behavior of others in the system. Each party is assumed to have some latitude about what it can do by way of measure-countermeasure moves over time. That latitude can include types and arenas of behavior quite different in degree and kind from one party to another. The parties' benefit-to-cost or effectiveness-to-cost calculations need not be identical in terms of components, weightings of components, or discounting of the future relative to the present. Each party will try to adapt its instrumental behavior in light of its interpreted information about the past, current, and impending behavior of others and material factors in the system. End-to-end analysis delineates the path or paths from the starting point taken for a security or peace relational situation to a historical or future outcome. An appropriate analogy is a process flow chart. End-to-end analyses recognize the intervening factors that can affect outcomes and examine their content and probabilities. As a work style, it guards against some of the distortions in choice pointed out by behavioral decision theory—unwarranted extrapolation from the ease or difficulty of initial steps along a path to chances of achieving later steps and then reaching a particular end point. In combination, the applied systems approach, opposed systems design, and end-to-end analysis are well suited to accommodating and indeed being strengthened by the six widely-recognized expansions summarized previously (including that of multiple prisms). While initially developed for security Page 583 → and peace problems of a military and nation-state character, the recommended ensemble is not limited to such problems. To conclude, the good news then is that of a style of work appropriate to the challenges of understanding and producing security and peace—if we choose to adopt it. The darker side is that the recommended style of work is very demanding in several ways. It calls for pooling knowledge and expertise across established borders of “schools,” disciplines, and organizations. It does draw on and exploit general conceptual perspectives and historical patterns. Yet it treats them less as sources of conclusions and more as agenda setters for mustering, comprehending, and manipulating the empirical details of particular relational situations and changes in them. Given the variety, “hardiness,” and resilience of security and peace problems, there is little reason to expect them to yield further to a less demanding professional work style.

Notes 1. The brevity of these reflections makes it only appropriate to note several caveats. My views are necessarily expressed in the form of oversimplified, stylized assertions lacking in desirable qualification and amplification. I have had to depart from the canon of extensive citations and thus scant giving credit where credit is due. Such omissions may be less serious because my concern lies more with a continuing and

unfinished agenda than with compiling a roster of accomplishments. Finally, I in no way mean to imply that anything that follows is a fresh discovery, but only that the views presented tend for the most part to recommend a departure from one or another currently fashionable emphasis. 2. This de facto expert security community curiously is identical to the national members of a club of intensely cooperating government intelligence agencies—perhaps a recipe for compounding security parochialism.

Page 584 →

CONVERGENCES BETWEEN INTERNATIONAL SECURITY STUDIES AND PEACE STUDIES Louis Kriesberg International security studies and peace studies are not a single subfield of international relations. Analysts in security studies and those in peace studies have generally viewed themselves and been viewed by others as working in quite different domains. Some persons in each area have been critical or dismissive of the efforts of those in the other. Nevertheless, many persons across both areas actually share significant concerns and questions, such as how to avoid or to limit wars and other violent conflicts. Furthermore, the work being done in each of these domains is increasingly overlapping. To enhance the possibilities of beneficial cooperation among analysts in these domains, the past relations and the current movements toward convergence should be examined. After doing so, I will discuss promising options for the future.

Earlier Relations In previous decades, many persons working in security studies and many persons working in peace studies differed in several significant ways. For example, they tended to draw from different international relations approaches. Persons doing international security studies tended to draw from the realist approach and those doing peace studies tended to draw from a liberal idealist or constructivist approach.1 Realists generally assume that states are unitary actors seeking to maximize power. Liberal idealists, however, generally stress the importance of diverse domestic actors, transnational organizations, and normative factors. Realists emphasize the anarchic nature of the world system, and other approaches stress the varying degree of integration and the shared ways of thinking in the world or regions of it.2 Page 585 → Peace studies analysts have generally drawn from a wider variety of fields of inquiry than have analysts working in international security studies. Many workers in peace studies have been more receptive, for example, to the perspective and insights of feminist thinking in international relations.3 The feminist attention to the manifold roles of women in sustaining social life fits well with the concern of many in peace studies about the ways people at the grass roots affect and are affected by so-called high politics. The emphasis among many feminists upon the distinctive qualities of women has contributed to studying the roles women have played in peace movements and the roles they might play in countering wars and overcoming large-scale violence.4 Some peace studies analysts have also been relatively attentive to the possible impact of religion and of culture on war and peace.5 Analysts have shown how religious beliefs, organizations, and leaders have contributed to mitigating as well as to exacerbating violent conflicts. Additionally, some workers in the peace studies domain have been relatively attentive to transnational social movement organizations and to critical analyses of global political economy. The many trends constituting globalization have increased the power of multinational corporations, and the persons who control those organizations increasingly shape and use the global market for their benefit.6 They also provide new opportunities for resistance and more egalitarian relations.7 Other differences are also noteworthy. International security analysts have tended to assume the perspective of one primary actor in an international conflict, typically their own country, while peace studies analysts tended to take a more global or systemic perspective. Persons in international security studies generally focused on military means while persons in peace studies stressed nonviolent methods. International security analysts have examined in great detail, for example, the nature of nuclear warheads and delivery systems and their impact on military strategy.8 Issues relating to deterrence and to nuclear proliferation have drawn great attention since the end of World War II. More recently, attention has been given to chemical and

bacteriological weapons and to terrorism. Finally, international security analysts tended to concentrate on avoiding war (negative peace), while peace researchers often stressed issues of justice and equity (positive peace) as well. Members of each camp also tended to differ in their institutional bases: those working in international security often were employed in institutes receiving foundation and government funds, while those in peace studies often were employed in colleges and in universities, and sometimes in nongovernmental social movement organizations. These differences were particularly strong in the United States in the early decades after World War II. The differences arose in large part from varying career origins, intellectual traditions, and network associations. In the United States, an important tradition in peace studies, emphasizing nonviolence and social justice, initially developed in church-related colleges. Page 586 → The first peace studies program was established in 1948 at the Church of Brethren-affiliated Manchester College in Indiana. In the late 1950s, research-oriented centers began to be established, notably the Center for Research on Conflict Resolution at the University of Michigan. Many of the people engaged in security studies flourished in college and university departments and international relations programs. They also were associated with academic and nonacademic institutes associated with the U.S. government and its armed forces—for example, RAND.9 As I discuss later, the divergence in theory, research, and practice between people engaged in these two domains has hampered the members' work in each camp. Some people, of course, did research and examined policy alternatives that in some ways bridged these differences. One arena of shared interests was the causes of war and of peace; indeed, analyses on this topic by Wright, Richardson, and Deutsch et al. were part of the important quantitative tradition in peace research and hence peace studies.10 Work in this tradition has continued among both peace researchers and international security analysts, as illustrated by the work of Singer, Isard, Leng, and Vasquez.11 They have tended to analyze factors that might account for variations in warfare over time and among different countries, often using quantitative data. Those factors include systemic features such as the number of major powers in the system, state characteristics such as type of governance, and relationship factors such as trading interdependency. Among other topics, considerable writing about crisis management and foreign policy decision making constituted areas where peace studies and international security studies have overlapped, with some variance in emphasis. Thus, some persons in the international security studies domain tended to assume that officials generally acted in terms of rational calculations of relatively fixed national interests, while analysts in the peace studies domain often emphasized group, normative, and emotional factors. Nevertheless, some analysts identified with each domain read and critiqued each other's work and influenced each other. This may be noted in work on the way officials from antagonistic states interact with each other in crises.12 Many professional associations also provided settings for informal and formal exchanges of ideas. For example, the International Studies Association has long included an International Security Studies Section and a Peace Studies Section, and members from each section sometimes participated together on the same panels. Within the American Sociological Association, the Section on the Sociology of World Conflicts was established in the early 1970s and its membership always has included students of peace and of military forces. The name of the section currently is Peace, War, and Social Conflict. The International Sociological Association included a Research Committee on Armed Forces and Society (RCoI), and in 1980 that was reorganized to include sociologists studying international relations, peace, and conflict resolution; the name of the research committee was changed to Armed Forces and Conflict Resolution.13 Page 587 → The differences between the peace and security domains have not been as great in Europe as in the United States. Thus, many European peace institutes included work related to alternative military doctrines. In varying degrees this was the case for the International Peace Research Institute Oslo (PRIO), established in 1959; the Department of Peace Studies at the University of Bradford, England, established in 1973; the Peace Research Institute

Frankfurt (PRIF); the Copenhagen Peace Research Institute (COPRI); and the Tampere Peace Research Institute (TAPRI). In such European centers, ideas were developed, for example, about nonoffensive defense, which many peace researchers argue contributed to changes in Soviet thinking and hence to the ending of the cold war.14 Thus, the Soviet military posture was modified so that it would be recognized as defensive and not regarded as threatening offensive actions. In many other countries in the world, international security studies and peace studies emerged only in the 1980s or 1990s. In the contemporary context, for reasons that I will discuss, the domains are becoming less sharply differentiated. Consequently, work in the relatively new centers of study reflects some of the changes that emerged after the cold war ended, including greater convergence between international security studies and peace studies. From 1948 until the end of the 1980s, Americans both in security studies and in peace studies typically concentrated their attention on the cold war. The former generally viewed it as a military contest between two camps consisting of governments.15 Persons working in international security studies emphasized military relations, particularly concerning nuclear deterrence and extended deterrence. They gave attention to the military hardware as well as the doctrines about their use. Some of them also studied foreign policy decision making and crisis management, particularly as conducted by high officials.16 Analysts working in peace studies examined social movements, such as peace movement organizations opposing nuclear weapons or engaged in people-to-people diplomacy.17 They also analyzed the role of the militaryindustrial complex, particularly in the United States.18 In addition, many of them examined the processes of socialization in families and schools and the influences of the mass media.19 Some of these researchers were themselves associated with peace movement organizations. Interestingly, analysts in both camps generally failed to anticipate the sudden transformation and ending of the cold war. Neither camp undertook intensive analyses about the possible explanations and implications of that failure. However, some persons in each domain offered reasons for the transformation in terms of the policies they had previously argued; those policies, they said, worked to prevent an escalation of the cold war and eventually to transform it.20 Thus, security analysts contended that the long-standing militant containment policy of Democratic and Republican governments or the arms buildup and proactive anti-Soviet policy of President Ronald Reagan's first administration forced the Soviet leaders to yield and accept a new accommodation. Page 588 → On the other hand, peace studies analysts pointed to other factors to account for the transformation of the Soviet Union's foreign policy, the ending of the cold war, and then the dissolution of the Soviet Union. They stressed the importance of the assurances given to the Soviet Union in the 1975 Helsinki Accords about the permanence of the borders established at the end of World War II. They also stressed the growing social exchanges and human ties fostered by détente, which continued despite the Reagan administration's lack of support for such activities.21 In short, many differences existed in past decades between persons working in peace studies and security studies. They differed in their research methodologies, in the importance accorded to particular values, in the questions they posed, and hence in the kinds of answers they tended to provide. These differences limited the work of persons in each domain, as they talked past each other. They too often ignored each other's ideas and findings, or when they did give them attention, it was to argue against them. If analysts in each domain gave more consideration to each other's work, they might have complemented each other's efforts and each enriched the other, as discussed in this chapter.

Contemporary Developments In the post-cold war era, the domains of peace studies and international security studies have moved much closer together. Thus, security is now understood to refer to much more than military matters. It is widely thought to include concerns such as environmental degradation, refugee flows, and economic issues. Furthermore, large-scale

violence directed against ethnic, religious, and other communal groups became matters of broadly shared international attention.22 The roles of international governmental organizations and of nongovernmental organizations in preventing and stopping highly destructive conflicts rapidly expanded. Many of those organizations also increasingly became engaged in peace-building after such conflicts.23 These and related developments are due to profound global changes and associated changes in ways of interpreting the transformed world. Partly because of these new circumstances, the analysts in both domains have also undergone many changes in the questions they pose and the answers they offer. Global Changes In recent decades, as is widely noted, the world has been undergoing profound and rapid changes. This is evident in the increasing economic integration, ever more swift and intensive communication, and many other global trends. Consequently, the world is becoming more homogenized, but with various people resisting and some reacting violently against that very development.24 Furthermore, multinational business corporations, transnational social movement organizations, and international governmental institutions Page 589 → are increasing in number, size, and resources.25 All this contributes to reducing state sovereignty. These interacting developments have profound significance for peace and security studies. Conflicts marked by large-scale violence within states greatly increased in the 1990s but show some signs of slackening since then.26 These conflicts have often been between peoples mobilized by appeals to different ethnic, linguistic, religious, or other communal identities. Although these conflicts may seem to be domestic, external connections generally are significant; they are manifested in the growing engagement of diaspora groups, the impact of the violence on neighboring countries, and the support given to one side by officials in other governments who seek to advance their own interests. As a consequence of such developments and the other global trends previously noted, international norms about human rights are increasingly shared. The demand to take action to stop gross violations of human rights has become widespread and insisted upon by many states and international organizations, nongovernmental as well as governmental. The visibility of such human rights violations and the normative insistence to stop them have resulted in numerous interventions by international organizations such as the United Nations. These interventions, unlike earlier UN interventions, were not contingent upon the concurrence of the combating parties.27 This has given a new legitimacy to the use of force in international settings.28 Yet the costs in deaths and protracted engagement have diminished the willingness of many governments to undertake such interventions. In any event, the increased relative significance of large-scale internal conflicts has raised the salience of domestic conflicts for security studies workers and violence management for peace studies workers.29

Changes among Peace and Security Analysts Many changes in the activities, institutional settings, and ways of thinking among persons in the security studies and peace studies domains have occurred since the 1970s and especially since 1989. These changes have tended to bring security analysts and peace analysts into closer association. The development of problem-solving conflict resolution ideas and practices has affected greatly both security and peace studies.30 This has been particularly so in the United States since the early 1970s.31 By the end of the last century, the ideas and applications of problem-solving conflict resolution had spread and evolved in most parts of the world. This development provides a newly-shared language and way of thinking for persons working in peace studies and international security studies. Even tough-talking diplomats are now likely to assert that they are striving for win-win outcomes. The growth of the field of conflict resolution, in conjunction with the other previously noted developments, has affected the practice of people in both international security and in peace studies. Peace researchers are much more likely themselves to conduct and to associate with people who engage in peacemaking and peace-building work. This includes arranging and facilitating Page 590 → conflict resolution workshops and dialogue groups,

providing conflict resolution training, and consulting about conflict management systems. Such work is undertaken by a wide variety of organizations, such as the Carter Center of Emory University, CDR Associates, the Fellowship of Reconciliation, the Mennonite Conciliation Service, Search for Common Ground, and the United States Institute of Peace (USIP). In a complementary fashion, persons engaged in international security studies are increasingly engaged in studying how to prevent large-scale violence and to build stable peace.32 They strive to do so using a wide array of nonviolent methods, including economic sanctions and diplomatic mediation. The Institute for Defense Analysis in Alexandria, Virginia, in conjunction with the United States Institute of Peace, has designed a computer-based peace-building/postsettlement transition simulation to help examine alternative strategies. Persons in international security studies increasingly work with their own and foreign governments to foster reconciliation and institutions to manage conflicts, as they are involved with activities relating to peacekeeping, nation-building, and peace-building. Furthermore, they increasingly associate with persons in nongovernmental organizations—for example, those providing peace-building services, advocating the protection of human rights, and giving humanitarian assistance.33 As a result of all these developments, more people are engaged in the practice of peace-building, and persons with backgrounds in the international security and the peace studies camps more often interact with each other. In addition, more venues for such interactions have been established—for example, the United States Institute of Peace, founded in 1984. Intellectual exchange is fostered by journals such as Security Dialogue and International Security. Conflict resolution courses and programs are increasingly being introduced in traditional international relations programs. Furthermore, programs in civilian institutions provide training for military and other national defense executives and managers, as in the national security studies courses provided by Syracuse University's Maxwell School of Citizenship and Public Affairs and Johns Hopkins University's Paul H. Nitze School of Advanced International Studies. Finally, the curricula at military service academies, the command and general staff colleges, and the war colleges increasingly reflect a broader range of perspectives and of security tasks, such as peacekeeping.34

Future Developments In anticipating the future, I will stress likely developments that I regard as desirable. I anticipate that new subfields will develop that cut across the domains of security and peace studies. Some of these subfields will focus on substantive issues. For example, they are likely to relate increasingly to different conflict stages: prevention, limiting and interrupting violent conflict escalation, terminating violence, and postconflict peace-building. In addition, they are likely to focus on particular ways of conducting and terminating conflicts; such methods include mediation and negotiation, constructive Page 591 → ways of conducting a struggle, contributing to conflict transformation, and fostering domestic conditions and institutions that are conducive to peace. Finally, some issues relating to general theoretical and analytical approaches will affect the work of most people in peace studies and security studies. One substantive subfield that has rapidly grown relates to early warning and preventive diplomacy. At first, the emphasis was on forecasting the outbreak of intense violence and providing information officials might use in anticipating their responses to such violence and its consequences. This work drew on the international security approach. Subsequently, the emphasis shifted to developing policies for averting and limiting such violent outbreaks. This emphasis upon preventive diplomacy drew relatively more on the peace studies approach.35 A major subfield focuses on interrupting and stopping violence. The matter of particular interest here is the role of external actors, whether international governmental organizations, nongovernmental actors, or national governments. Thus, there is important new work on sanctions. For example, targeted, or “smart,” sanctions have recently been stressed, but even their effectiveness is in question for certain goals.36 Another expanding substantive subfield relates to peacekeeping and postconflict peace-building. Several issues are currently receiving considerable attention, and that is likely to increase among peace and security analysts.

Peacekeeping by the UN and by regional international organizations has expanded greatly after the end of the cold war.37 Military personnel are increasingly trained in peacekeeping operations and how to avert conflict escalations.38 Currently, increasing attention also is being given to sustaining agreements that have been reached.39 Particular attention is more and more given to building institutions that would reduce intercommunal conflicts and provide channels for their nonviolent management.40 This includes various forms of federal structures, autonomy, and power sharing. It also includes democratic processes providing basic political rights. Increasingly, too, attention is being given to the civil society needed to sustain democratic processes and institutions. The finding that democracies do not make war against each other has become a matter of great interest and the subject of considerable research. Workers in peace studies and security studies are examining the nature of that relationship and of the relationship between democratic polities and foreign intervention.41 Nongovernmental organizations that provide humanitarian relief, medical care, and other services are proliferating. They are working in some of the same areas as do national governmental organizations and international governmental organizations. Consequently, inefficient use of resources and inappropriate competition sometimes arise. Ways to achieve a more effective division of labor and better coordination are increasingly being discussed.42 Practitioners and analysts alike are giving great attention to attempts at reconciliation between antagonists, and this attention is likely to grow. The Page 592 → issues are complex, since the very concept is multidimensional. In moving toward reconciliation, one or more party in a fight is likely to seek justice, truth, security, and kindness, perhaps even forgiveness.43 Various groups may emphasize one or another of these dimensions at different times. Consequently, balancing the achievement of various dimensions of reconciliation is difficult and occurs in varying sequences and degrees among different groups in the opposing communities.44 Aside from various substantive subfields within which persons working in peace studies and in security studies will act in some degree cooperatively, some general issues will affect people working in these domains. For example, policy-relevant work is likely to be increasingly done for and with nongovernmental actors and not directed only at government officials. Academics may well spend some time engaged in nonacademic work, while persons engaged in applied work are likely for a time to take academic appointments. This engagement will most likely affect not only the questions asked but also the kind of answers sought and found. Thus, attention is likely to be given to relatively malleable factors, to agency rather than to structural conditions. In addition, the increased interaction and overlap between persons in these domains may have a salutary effect on theory, research, and practice in each. Much international relations theorizing has been directed at what others have written, with too little reference to evidence relevant to the ideas. Advocates of particular approaches may overstate the case in order to get a hearing for matters that they feel have been given too little attention. But that can result in having the ideas dismissed, and that further escalates the debate. So driven, the debates become difficult to resolve. Attention to the shared questions is a way to transform such debates. Attention also needs to be given to the information that would help answer the questions that are posed. The information may be qualitative or quantitative, but in either case, it has an external referent about which experientially-based consensus is possible. Thus, the discussions about the finding that democracies do not make war on each other have relied heavily on a variety of empirical analyses. Finally, some persons in these domains are likely to undertake analyses that help synthesize the ideas and practices of international security studies and of peace studies. Some such work is being done, and more is needed.45 The synthesizing work obviously requires focusing on policy-relevant issues. One such focus relates to the dilemmas associated with advancing human rights and justice while also settling fights and ending violence. Not all values can be maximized at the same time, and this poses dilemmas as the pursuit of a good goal may result in many bad consequences. Improving knowledge about the likelihood of various consequences of different strategies will contribute to resolving the dilemmas.

Such knowledge, however, cannot be precise for any single case. We mortals will have to rely on good judgment in each case.46 Yet, insofar as good information and understanding is widely shared, good judgment is more likely to occur and be able to be followed. Page 593 →

Conclusions I anticipate that convergence between security studies and peace studies will continue as the range of the peacemaking and peace-building agenda grows broader and more complex. Work in each domain has already benefited by the increasing overlap between the domains. The policy questions addressed and the repertoire of answers offered by persons in each domain have expanded. Peace studies workers increasingly recognize that atrocities occur, that violence takes many forms, and that there are times for acting with urgency. Security studies workers increasingly recognize the growing importance of cooperating with nongovernment actors and of expanding reliance on nonmilitary options. Progress is being made. One important sign of and contributor to that progress is the increased range of voices joining the conversations about peace and security. As previously noted, this is the case for the increasing engagement of women in peace and security studies and the growing application of feminist thought in those domains. They are providing new insights and important perspectives. In addition, work by persons in and from regions of the world previously excluded or ignored from conversations are now included. Thus, many centers of peacemaking and security studies are being established in more and more countries in the developing world. In South Africa, for example, some centers have done significant work for many years and are now broadening their range of work. The Centre for Intergroup Studies, for instance, was founded in Capetown in 1968, under the leadership of Hendrik W. van der Merwe; it focused on conflict resolution within South Africa, helping to develop channels of communication among the major groups there. Now, as the Centre for Conflict Resolution, with Laurie Nathan as the executive director, the scope of activities has been expanded to include facilitation and mediation among disputing parties in Burundi and elsewhere in Africa. Although the closer relations between workers in international security and in peace studies have many benefits, there also are risks. One risk for peace researchers is that they become co-opted by high officials and lose their special relations with and perspectives derived from persons at the grass roots. A risk for security studies workers is to adopt words from the peace studies field but without incorporating the overall context. Many differences in orientation are likely to persist, as they should. Thus, some people will work more closely with high-ranking officials in governments and international governmental organizations. Others, however, will work more closely with nongovernmental and grass-roots organizations as, for example, do Lederach and Galtung, Jacobsen, Brand-Jacobsen, and Tshudie.47 The time frames of workers in these fields are also likely to continue to differ. Some persons, frequently they are political leaders, operate in terms of the short term, dealing with the crisis at hand. Others, frequently working in religious institutions or advocacy organizations, work in particular communities and countries for decades. Persons in peace studies often stress Page 594 → long-term structural factors, including education, culture, and class structure. Such differences sometimes will be a source of opposition and tension, but they are needed. Unfortunately, the division of labor may be manifested in a highly competitive and destructive manner. To minimize such conduct, certain truths need to be reiterated. The matters of concern in peace studies and international security are so complex that no single approach can provide a full treatment and a comprehensive understanding. No one can do everything, so a division of labor can be quite useful. The division of labor, however, can and should function so workers in each domain recognize the insights and knowledge of workers in the other domain. That can be constructive and helpful for members of each community in performing their tasks.

Notes I thank the following persons who read earlier drafts of this paper and kindly gave me their comments: Eileen Babbitt, James Bennett, Volker Franke, and Nils Petter Gleditsch. I, of course, remain responsible for the observations and judgments made in this chapter. 1. Peter L. Berger and Thomas Luckman, The Social Construction of Reality (New York: Doubleday, 1966); Hans J. Morgenthau, Politics among Nations (New York: Knopf, 1948); Kenneth N. Waltz, Theory of International Politics (New York: Random House, 1979); Alexander Wendt, “Anarchy Is What States Make of It: The Social Construction of Power Politics,” International Organization 46, no. 2 (1992): 391–425. 2. Kenneth N. Waltz, “Structural Realism after the Cold War,” International Security 25, no. 1 (2000): 5–41. 3. Cynthia Enloe, Bananas, Beaches, and Bases (Berkeley and Los Angeles: University of California Press, 1989); Elise Boulding, Cultures of Peace: The Hidden Side of History (Syracuse, N.Y.: Syracuse University Press, 2000); Betty A. Reardon, Sexism and the War System (Syracuse, N.Y.: Syracuse University Press, 1996). 4. Harriet Hyman Alonso, Peace as a Woman's Issue: A History of the U.S. Movement for World Peace and Women's Rights (Syracuse, N.Y.: Syracuse University Press, 1993); Tamra Pearson d'Estree and Eileen F. Babbitt, “Women and the Art of Peacemaking: Data from Israeli-Palestinian Interactive Problem-Solving Workshops,” Political Psychology 19, no. 1 (1998): 185–209; Simona Sharoni, Gender and the IsraeliPalestinian Conflict (Syracuse, N.Y.: Syracuse University Press, 1995); Carolyn M. Stephenson, “Gender Differences in Conflict Resolution,” in United Nations Expert Group Meeting on Political Decision-Making and Conflict Resolution: The Impact of Gender Difference, UN Document EGM/PDCR/12996/EP4 (Santo Domingo, Dominican Republic: United Nations, 1996); Anna C. Snyder, “Conflict and Consensus-Building at the 4th UN World Conference on Women: A Study of International Nongovernmental Cooperation” (Ph.D. diss., Syracuse University, Syracuse, N.Y., 1998). 5. Douglas Johnston and Cynthia Sampson, eds., Religion, the Missing Dimension of Statecraft (New York: Oxford University Press, 1994). 6. William I. Robinson, Promoting Polyarchy: Globalization, U.S. Intervention, and Hegemony (New York: Cambridge University Press, 1996). 7. Jackie Smith, Charles Chatfield, and Ron Pagnucco, eds., Transnational Social Movements and Global Politics: Solidarity beyond the State (Syracuse, N.Y.: Syracuse University Press, 1997). 8. Bernard Brodie, ed., The Absolute Weapon (New York: Harcourt Brace, 1946). 9. Fred Kaplan, The Wizards of Armageddon (New York: Simon and Schuster, 1983). 10. Quincy Wright, A Study of War (Chicago: University of Chicago Press, 1942); Lewis F. Richardson, Statistics of Deadly Quarrels (Pittsburgh: Boxwood Press, 1960); Karl W. Deutsch, Sidney A. Burrell, Robert A. Kann, Maurice Lee Jr., Page 595 → Martin Lichterman, Raymond Lindgren, Francis L. Loewenheim, and Richard W. Van Wagenen, Political Community and the North Atlantic Area (Princeton, N.J.: Princeton University Press, 1957); Carolyn M. Stephenson, “Peace Studies, Overview,” in Encyclopedia of Violence, Peace, and Conflict, ed. Lester Kurtz (San Diego: Academic Press, 1999), 809–20. 11. J. David Singer, Stuart A. Bremer, and John Stuckey, “Capability Distribution, Uncertainty, and Major Power War, 1820–1965,” in The Correlates of War I: Research Origins and Rationale, ed. J. David Singer (New York: Free Press, 1979), 265–97; Walter Isard, Understanding Conflict and the Science of Peace (Cambridge, Mass.: Blackwell, 1992); Russell J. Leng, “Influence Strategies and Interstate Conflict,” in Singer, The Correlates of War, 124–57; John Vasquez, The War Puzzle (Cambridge: Cambridge University Press, 1993). 12. For example, see Glenn H. Snyder and Paul Diesing, Conflict among Nations: Bargaining, Decision Making, and System Structure in International Crises (Princeton, N.J.: Princeton University Press, 1977); Richard Ned Lebow, Between Peace and War (Baltimore: Johns Hopkins University Press, 1981); Michael

Brecher, Crises in World Politics (Oxford, England: Pergamon, 1993); Margaret G. Hermann and Charles F. Hermann, “Who Makes Foreign Policy Decisions and How: An Empirical Inquiry,” International Studies Quarterly 33, no. 4 (1989): 361–87; Robert Jervis, Perception and Misperception in International Politics (Princeton, N.J.: Princeton University Press, 1976). 13. In the late 1970s, the idea of broadening the scope of RCOI was raised by some International Studies Association members interested in finding a hospitable research committee for their activities. Exploring options for this group, I was encouraged by Morris Janowitz, a leading figure in RCoI, to join that research committee and broaden its scope. Gwyn Harries-Jenkins, then chair of RCoI, agreed, and the scope and name of the committee was formally changed. Sociologists studying military forces predominate in the membership of the committee, but those studying conflict resolution and nonviolence also belong and participate in panel sessions and other committee activities. 14. Jorgen Dragsdahl, “How Peace Research Has Reshaped the European Arms Dialogue,” in Annual Review of Peace Activism, 1989, ed. John Tierman (Boston: Winston Foundation for World Peace, 1989), 39–45. 15. Many peace researchers, particularly in Europe and in the Third World, examined imperialism and economic dependency. For example, see the frequently cited article by Johan Galtung, “A Structural Theory of Imperialism,” published originally in the Journal of Peace Research in 1971; it is reprinted in Johan Galtung, Peace and World Structure: Essays in Peace Research (Copenhagen: Ejlers, 1980). 16. Snyder and Diesing, Conflict among Nations; Graham T. Allison, The Essence of Decision: Explaining the Cuban Missile Crisis (Boston: Little, Brown, 1971). 17. John Lofland, Polite Protestors: The American Peace Movement of the 1980s (Syracuse, N.Y.: Syracuse University Press, 1993); Sam Marullo and John Lofland, eds., Peace Action in the Eighties (New Brunswick, N.J.: Rutgers University Press, 1990). 18. Dieter Senghass, Rüstü ng und Militarismus (Frankfurt: Suhrkamp, 1972). 19. James William Gibson, Warrior Dreams: Violence and Manhood in PostVietnam America (New York: Hill and Wang, 1995). 20. Louis Kriesberg, International Conflict Resolution: The U.S.-USSR and Middle East Cases (New Haven, Conn.: Yale University Press, 1992); Richard Ned Lebow and Thomas Risse-Kappen, eds., International Relations Theory and the End of the Cold War (New York: Columbia University Press, 1995); William Curti Wohlforth, The Elusive Balance: Power and Perceptions during the Cold War (Ithaca, N.Y.: Cornell University Press, 1993). 21. Mary Kaldor, “Who Killed the Cold War?” Bulletin of the Atomic Scientists (July-August 1995): 57–60. 22. Walker Conner, Ethnonationalism: The Quest for Understanding (Princeton, N.J.: Princeton University Press, 1994); Donald L. Horowitz, Ethnic Groups in Conflict (Berkeley: University of California Press, 1985).Page 596 → 23. Larry Dunn and Louis Kriesberg, “Mediating Intermediaries: Expanding Roles of Transnational Organizations,” in Studies in International Mediation: Essays in Honour of Jeffrey Z. Rubin, ed. Jacob Bercovitch (London and New York: Macmillan and St. Martin's, 2002). 24. Benjamin Barber, “Jihad vs. McWorld,” Atlantic Monthly, March 1992, 53–63. 25. Smith, Chatfield, and Pagnucco, Transnational Social Movements and Global Politics. 26. Ted Robert Gurr, Peoples versus States: Minorities at Risk in the New Century (Washington, D.C.: U.S. Institute of Peace Press, 2000). 27. Otara A. Otunna and Michael W. Doyle, eds., Peacemaking and Peacekeeping (New York/Oxford: Rowman and Littlefield, 1998). 28. Andrew J. Goodpaster, When Diplomacy Is Not Enough: Managing Multinational Military Interventions (New York: Carnegie Commission on Preventing Deadly Conflict, 1996), 1–38. 29. Noted in personal communication by Eileen Babbitt. 30. Persons working in the problem-solving conflict resolution approach tend, in varying degree, to emphasize ways in which adversaries in a conflict shift from seeing each other as engaged in a struggle in which what one side gains is at the expense of the other to seeing a mutual problem for which they seek a solution that has some mutual benefit. 31. Louis Kriesberg, “The Growth of the Confict Resolution Field,” in Turbulent Peace, ed. Chester A. Crocker, Fen Osler Hampson, and Pamela Aal (Washington, D.C.: U.S. Institute of Peace Press, 2000),

407–26. 32. Carnegie Commission on Preventing Deadly Conflict, Final Report of the Carnegie Commission on Preventing Deadly Conflict (New York: Carnegie Corporation, 1997); Arie M. Kacowicz, Yaacov BarSiman Tov, Ole Elgstrom, and Magnus Jerneck, eds., Stable Peace among Nations (New York/Oxford: Rowman and Littlefield, 2000). 33. Alice Ackermann, Making Peace Prevail: Preventing Violent Conflict in Macedonia (Syracuse, N.Y.: Syracuse University Press, 2000). 34. Volker Franke pointed out some of these developments in a personal communication. 35. David Carment and Patrick James, eds., Peace in the Midst of Wars (Columbia: University of South Carolina Press, 1998); Michael S. Lund, Preventing Violent Conflicts (Washington, D.C.: U.S. Institute of Peace Press, 1996). 36. David Cortright and George A. Lopez, eds. Economic Sanctions: Panacea or Peacebuilding in a PostCold War World (Boulder, Colo.: Westview, 1995); David Cortright and George Lopez, with Richard W. Conroy, Jaleh Dashti-Gibson, and Julia Wagler, The Sanctions Decade: Assessing UN Strategies in the 1990s (Boulder, Colo.: Lynne Rienner, 2000). Evidence against the effectiveness of economic sanctions has been reported for a long time by peace researchers and others. For example, see Peter Wallensteen, “Characteristics of Economic Sanctions,” Journal of Peace Research 5, no. 3 (1968): 248–67. 37. Robert A. Rubinstein, “Cultural Aspects of Peacekeeping: Notes on the Substance of Symbols,” Millennium 22, no. 3 (1993): 547–62. 38. David Last, “Organizing for Effective Peacebuilding,” in Peacekeeping and Conflict Resolution, ed. Tom Woodhouse and Oliver Ramsbotham (London: Cass, 2000), 80–96. 39. Fen Osler Hampson, Nurturing Peace: Why Peace Settlements Succeed or Fail (Washington, D.C.: U.S. Institute of Peace Press, 1996). 40. John McGarry and Brendan O'Leary, eds., The Politics of Ethnic Conflict (London: Routledge, 1993). 41. Nils Petter Gleditsch and Havard Hegre, “Peace and Democracy,” Journal of Conflict Resolution 41, no. 2 (1997): 283–310; Margaret G. Hermann and Charles W. Kegley Jr., “The U.S. Use of Military Intervention to Promote Democracy: Evaluating the Record,” International Interactions 24, no. 2 (1998): 91–114; Bruce Russett, “The Democratic Peace: And Yet It Moves,” International Security 19 (spring 1995): 164–75; Bruce Russett and John R. Oneal, Triangulating Page 597 → Peace: Democracy, Interdependence, and International Organization (Princeton, N.J.: Princeton University Press, 2001). 42. Louis Kriesberg, “Coordinating Intermediary Peace Efforts,” Negotiation Journal 12 (October 1996): 341–52; John Paul Lederach, Building Peace: Sustainable Reconciliation in Divided Societies (Washington, D.C.: U.S. Institute of Peace Press, 1997). 43. Michael Henderson, The Forgiveness Factor (London: Grosvenor Books, 1996); Martha Minow, Between Vengeance and Forgiveness (Boston: Beacon Press, 1998); Eugene Weiner, ed., The Handbook of Interethnic Coexistence (New York: Continuum, 1998). 44. Louis Kriesberg, “Paths to Varieties of Inter-Communal Reconciliation,” in From Conflict Resolution to Peacebuilding, ed. Ho-Won Jeong (Fitchburg, Md.: Dartmouth, 1999). 45. Hugh Miall, Oliver Ramsbotham, and Tom Woodhouse, Contemporary Conflict Resolution (Cambridge, England: Polity, 1999). 46. Isaiah Berlin, “My Intellectual Path,” New York Review of Books, May 14, 1998, 53–60. 47. Lederach, Building Peace; Johan Galtung, Carl G. Jacobsen, Kai Frithjof Brand-Jacobsen, and Finn Tschudi, Searching for Peace (London and Sterling, Va.: Pluto, 2000).

Page 598 →

ACCOUNTING FOR INTERSTATE WAR Progress and Cumulation J. David Singer This turn of the millennium certainly offers us another opportunity—and incentive—to get more serious about trying to explain, and perhaps reduce the incidence of, that brutal, stupid, and destructive form of collective behavior known as war. In almost every epoch and almost every corner of human civilization, we find the politicians and the priests calling for an end to such barbarous behavior. But time after time, their prescription is to demand that the other party give up its aggressive behavior and meet the demands of our reasonable and accommodating brethren. This recurrent pattern is nicely captured by Mark Twain in “The War Prayer,” penned in response to the U.S. intervention in the Philippines in 1898, rejected by Harper's Bazaar in 1905, but finally published in November 1916 as the drumbeat for war against Germany had begun to accelerate. In praying, “Grant us the victory, O Lord our God,” Twain then laid out the corollary, unspoken prayer: “Oh Lord our God, help us to tear their soldiers to bloody shreds with our shells; help us to cover the smiling fields with the pale forms of their patriot dead; help us to drown the thunder of their guns with the shrieks of their wounded, writhing in pain. … We ask it, in the spirit of love, of Him Who is the Source of Love.”

Some Promising Auguries Can we perhaps do better than this sanctimonious and self-serving hypocrisy, which hardly seems to move us closer to the day when war begins to lose its legitimacy, when its counterproductivity becomes fully evident, and when human groups back away from lending psychic and material resources to its preparation? Some might say that such a process is already under way, thanks to the devastation wrought by the two world wars, the advance of weapons technology, the repugnance with which we witness the genocidal conduct associated with the last half century's wars in the Balkans, sub-Saharan Africa, and Southeast Asia. Along with Mueller and Levi we might Page 599 → reasonably adopt so sanguine an outlook, but the indicators to the contrary are difficult to ignore.1 The land mines still planted along boundaries from Korea to Kosovo taking scores of casualties every month, the steadily-increasing worldwide trade in light weapons, the obscene U.S. military budget, and the presence of more than thirteen thousand strategic warheads on high alert among the nuclear powers all remind us that the preparedness for war, the willingness to threaten war, and the ease with which political elites mobilize and carry out military actions are very much with us. As I write these lines at the beginning of a new year and a new millennium, the pope in Rome and some of his counterparts elsewhere are calling on us to desist from armed violence, but as Mark Twain reminded us in his prayer for our soldiers, more of the world's religious leaders are busily calling the faithful to the barricades and encouraging their propensities for yet another holy war. Another of the weak reeds that some of us cling to is that of the “Kantian peace,” the proposition that as international trade continues to increase, and as more societies are governed by democratic/republican regimes, such tendencies will conduce to a diminishing incentive to the mobilization for and conduct of war. The problem with this particular hope is twofold. First, even though world trade is increasing, a closer look shows that the worldwide level of intrastate commerce continues to grow at a somewhat faster rate. Second, the data-based evidence suggests that—even if intersocietal trade were on the increase—the pacifistic consequences are far from clear. While Polachek, Russett, and O'Neal all interpret the data in this fashion, scholars such as Barbieri conclude otherwise.2 When we shift from trade to regime types, we find the same uncertainties. The U.S. foreign policy establishment likes very much the idea that the fraction of the world's national regimes has become increasingly democratic, and that this reflects a tide of the democratic peace. But if we look closely at the coding criteria, there is not much hard

evidence to support the alleged expansion of democracy. Yet another slim reed is that known as the “evolution of cooperation.” In a computerized tournament attracting some of the top game theorists, the most successful strategy for inducing cooperative behavior on the part of refractory members of a social system turned out to be “tit for tat.” In these simulations, most players start off with largely cooperative moves and continue to reinforce such behavior all around. But since any human system will contain some noncooperative members (“rogue states”?), the problem is one of modifying their behavior, and in the tit-for-tat strategy, we punish them by responding with similar noncooperative behavior. The charming premise is that such conflictful responses will show them the error of their ways and lead our “defectors” into more cooperative behavior patterns. The problems are threefold. First, how often in history have we seen most states in the global or a regional system behaving in a cooperative fashion? Perhaps the Congress of Vienna system held for a while, but by 1854 Europe had experienced the Franco-Spanish, Russo-Turkish, Austro-Sardinian, Page 600 → Schleswig-Holstein, Roman Republic, and Crimean Wars. Second, every social organization is a coalition, and there will almost always be some members of the coalition with incentives to initiate or perpetuate sharp conflict, thus not exactly contributing to the evolution of cooperation. Third, and most critical, it is almost axiomatic that the way to escalate a dispute is to react to hostile and uncooperative behavior by responding in kind; at the very least, the strategy might call for tit minus one for tat. This approach to war reduction might—given these problems—attract no attention at all were it not for the highly desirable characteristic that it requires no strengthening of international institutions, thus avoiding the state sovereignty issue, and no need to generate a worldwide shift in norms. So much for peace on the cheap.

Retrospect and Prospect Given this pessimistic reading, let me step back from the contemporary scene for a moment and look both backward and forward in our effort to get a better handle on the explanation of war. We might think of four stages in the history of our effort to explain—and perhaps limit—the frequency of war. The first runs from the earliest appearance of armed, organized combat up through the end of World War I, in which we saw a growing body of literature that was interesting, diverse, and insightful, but resting largely on assertion, speculation, and anecdote. The second stage—marked largely by the work of Richardson and Wright in the interwar period3—illustrated the possibility of using scientific method in our quest. The third stage, perhaps best exemplified by the Correlates of War Project beginning in the early 1960s, went the next step, combining a rigorous methodology with an explicitly agnostic, multitheoretical orientation, and might be called the “natural history” stage.4 The dominant concern was/is to ascertain what goes with what: under what conditions do which behaviors culminate in which outcomes on the war-peace spectrum? This current stage is already beginning to blend into the fourth, in which the results of our systematic historical empiricism might permit the articulation and testing of more formal and mathematized models that will rest less and less upon the untested and often naive premises found in their opening rounds.

Premises: Epistemological and Empirical Bearing in mind that we are now well into the third stage—the natural history stage—but barely into the fourth stage in our search for the explanation of war, it might be the propitious moment to stop and evaluate some of the empirical and epistemological assumptions that ought to guide our investigation. Leaving aside for the moment our ethical assumptions, let us look more closely at the assumptions from which we might want to work on the epistemological and empirical dimensions. Let me begin with an epistemological assumption that will cause only a modest amount of grief in these precincts but could lead to howls of anguish Page 601 → in the Modern Language Association; the reference is, of course, to the syndrome known as “deconstruction,” which is usually associated with those schools of thought that could be called “the posties,” embracing the post-modern, poststructural, postbehavioral, and most germane to us, postpositivist outlook. I make no pretense of being on intimate terms with colleagues of those persuasions or

familiar with the distinctions among them. As I understand this fairly recent orientation—and it may go back to Karl Mannheim5—developing a science of human behavior just is not possible for several reasons: (a) there may not even be any empirical social reality; it is nothing more than imaginings given credibility by the language we use to describe these fictitious happenings; (b) even if there is social reality, we can never apprehend it because every human being is the product of his or her cultural identity, gender, age, profession, education, and accumulated personal experiences; and (c) the representations of reality that we generally accept are “socially constructed,” created and imposed on societies by the elites as an instrument of control. A milder—and less nihilistic—version of the “mission impossible” school doesn't deny the existence of an empirical reality but argues that we positivists define it too narrowly. For example, Alker suggests that our “data” must include not only the alleged historical and empirical phenomena that we attempt to describe and explain but also the reports about alleged observations of such phenomena, sometimes referred to as a call for more hermeneutics and less exegesis, as in biblical studies.6 Coming closer to the perspective of our small but growing “epistemic community”7 are those in the tradition of the late Hedley Bull who—with a touching faith in their own observational accuracy and perfect recall—question the need for all this scientific paraphernalia and assure us that the classical methods of the historian will do just fine.8 These traditionalists might want to listen to Mannheim more attentively; a good scientist attends to rigorous methods of observation, measurement, analysis, and inference simply because we do take some of the latter's warnings seriously. In order to rectify all these failings of the contemporary scientific scene, we are offered quite a variety of dubious epistemologies. Already familiar to the quantitative world politics community is Dessler's effort to draw from the physical sciences such platitudes as the need to “identify the mechanisms through which specified outcomes occur” or “the real structures that produce the observed phenomena.”9 And at the end of this thoughtful essay, we are left with no guidance toward this end beyond the strictures of “scientific realism”10 and the belief that causality does indeed obtain in the referent world. This line of reasoning will of course lead to the articulation of increasingly ambitious “theories” that are, in truth, little more than speculative models that rest on problematic and far from operational and thus testable premises. One variant of this approach is the “analytic narrative” in which one combines a case study with a formal—if preoperational—model and then moves back and forth in the hope of making the facts and the model increasingly congruent.11 This would not be such a bad idea if it were applied to a population of cases that constituted an explicit class of Page 602 → cases, and then it sought to bring the model and most of the cases into harmony. But doing it a case at a time is virtually guaranteed to be a decades-long exercise in futility. Of more ancient vintage is the “structured, focused comparison,”12 which, on closer examination, may be focused but typically fails on the comparison dimension because the model is not expressed in terms that are sufficiently precise and operational to permit their application to a population of putatively comparable cases. The net effect is a series of individual case studies each examined via models that are roughly similar but will end up applying rather different criteria and coding rules, resulting in something only weakly structured, focused, and comparative. So much, then, for the ambitious enterprise by which we might eventually hope to bring the ideographic and the nomothetic into some sort of harmony. It is exactly because of the inevitable tension between the immanent vagaries of our individual crises, wars, alliances, treaties, institutions, and so on on the one hand and the scientific aspiration to find a general explanation of some class of phenomenon on the other that all of these proposed methods are nothing more than the thoughtful cogitation that all scientists engage in as they mull over the separate cases about which they hope to generalize and the model that they hope will permit such a generalization. And, to put a realistic spin on this, there will almost always be a fair number of cases that just do not fit—at least in the earlier stages of the investigation. We may have quasi-deterministic models in our minds, but the natural world most certainly will defy us by its stochastic indeterminacy and ubiquitous equifinality! Closely related is the doomed effort to ascertain/demonstrate “causality.” In my view, this is a chimera invented to help persuade us that we really can know how the world truly works, and it might best be replaced with the concept of explanation. That is, as we uncover more and more regularities and covariations—and thus generate an increasingly rich and credible description of the ways in which all sorts of events and conditions go together empirically, we will arrive at increasingly-accepted explanations, and as these persuasive “stories” are combined

with what we think we know about individual and group behavior, the closer we will come to legitimate theories—in the sense of codified knowledge that commands the assent of our most capable colleagues. Considerably more unnerving, and further down the self-styled “scientific realist” path, is an approach that is called “evolutionary epistemology” intended to resolve the realist-versus-relativist impasse that arose out of the alleged passing of positivist and empirical methodologies. These recurrent debates over epistemology in the social sciences need not surprise us, given how poorly scientific method is taught and how dubious is the status of the social sciences in the academic world. Further, when we read those philosophers who tend to support and illuminate the scientific approach to social phenomena, we find that many of them are as confused as they are inaccessible; I find Hempel and Popper, not to mention Harre and Lakatos, not very helpful.13 Like religion, philosophy of science should be kept out of reach Page 603 → until we are sufficiently mature and experienced to be adequately immune to their all-too-plausible blandishments! Turning from the doubters to the doers, let me articulate some of the more relevant empirical-epistemic assumptions about the “nature of” the universe and the social systems—sometimes called “artificial” for reasons that still baffle me14—of concern to us, and ways of comprehending these systems. A central axiom is that the global system and the social groups it comprises evince highly regular and thus recurring patterns, but we have only begun to discover them and confirm their existence. There are lawlike regularities, but we still have only a dim grasp of them. And despite the existence of all these regularities, they are far from deterministic; while certain types of states respond over and over to the same stimuli in the same context in the same way, there are plenty of exceptions, giving us statistical distributions rather than deterministic and perfectly uniform law. Further, when we find extraordinary regularities, it is worth noting that the “causal path” between background conditions and behavioral outcomes can vary considerably; there are several ways to go from a given set of initial conditions to a well-specified outcome. Such regularities can take several forms, of which the simplest is a trend line in the magnitude of a given attribute of a given system; inasmuch as every social system is an evolving one, rates of change in such an attribute are also subject to change over time. Moreover, cyclical patterns are often found, usually not in fixed intervals of real time offering true periodicity, but in the form of recurring sequences or the order in which events and conditions follow one another over and over. In my judgment, we are not likely to uncover fully-deterministic regularities not only because all the systems and subsystems that we study are evolving, and in sequences that may well be irregular, but also because there is an inherent randomness in their behaviors and interactions vis-à-vis one another. Of course, the magnitude of such stochastic phenomena will diminish as we reduce (inter alia) measurement error, erroneous simplifying assumptions, misspecifications of our models, and the factors subsumed under the error term. But it will never be reduced to zero; as asserted previously, this residual randomness is inherent in the universe. In sum, our search for the lawlike regularities in matters of war and peace will advance erratically as a result of both empirical findings and theoretical insights; we will not get very far relying on only one or the other, despite pendulumlike fashions; systemic, data-based investigations are rarely “barefoot” or “dustbin” empiricism, nor is the construction of formal models merely “intellectual gymnastics.” In our search for an explanation of interstate war, we may discover empirically that two, three, or more subtypes arise out of systematically different conditions and processes, but we can always increase the goodness of fit by adding more differentiating variables; if not careful, we end up with a different model for each of the eighty-odd interstate wars since 1816, looking no different from our colleagues of the “No two wars are alike!” school. Page 604 →

Extension and Recapitulation While most of the foregoing discussion is addressed to the study of interstate war, it should come as no surprise that a good part of it might be extended to the general field of world politics.

Semantic Problems And a good place to start that extension might be with the semantic side of things. Note, first, that I label our field “world politics” rather than international relations, or worse yet, IR. These initials are also widely used to denote such fields of investigation/speculation as information retrieval, infrared, interocular response, interrogatory reaction, and so forth. More seriously, given the variety of actors that we study—ranging from individuals and agencies to regions and the global system—“international” misses a lot. Equally important, we cannot be satisfied with nothing more than “relations” in the sense of bonds, links, connections, and relationships. How about the behavior of the many entities of interest as well as the interactions among them? Relations might be trade partners, alliance members, neighbors, rivals, and so on, whereas it is interaction patterns that help to create such relations, and these relations, in turn, help generate such interaction patterns. Other semantic carelessness can stand in the way of cumulative research. Consider, for example, how often we still use the archaic concept of nation-state, when few such can be found in today's system of territorial states; rather, some of them are multinational and perhaps more are subnational, in which a given “nation” is divided into several states. That short-lived Wilsonian ideal of self-determination for each nation organized and legitimized as a coterminous state that arose out of the dismemberment of the pre-World War I empires began to crumble on the eve of World War II, and was barely visible at the dawn of the nuclear age. Or consider the traditional concept of bipolarity or multipolarity, clarified and examined in the post-Napoleonic epoch to reflect the number of discernible clusters of states, typically bound together by military alliance, trade relations, shared diplomatic representation bonds, or perhaps certain cultural or political similarities.15 But with the emergence of the “cold war” rivalry, many of us16 began to forget history and to think of polarity in terms of simply the number of major powers in the system, on the dubious if inarticulate premise that around each single major or superpower we will always find a cluster of bound-together states. What happens when—as on the eve of most European wars from the Congress of Vienna to the Rome-Berlin-Tokyo axis—we find several major powers in the same pole? Another case is that of the traditional distinction between collective defense and collective security, with the former designating a typical military alliance arrayed against another coalition or even a single identified “revanchist” state. Collective security, on the other hand, was used to describe Page 605 → regional or nearlyuniversal organizations designed to preserve the status quo, regardless of which of its members might turn out to be the ultimate threat to order.17 Either out of ignorance or an effort to enhance the legitimacy of the cold war alliances, such arrangements as NATO, the Warsaw Pact, and the UN were all increasingly described as collective security systems even though the first two are collective defense, and only the UN deserves to be called a collective security organization. Another example of more recent vintage is “ethnic conflict.” Whether the opposing groups are different in terms of language, religion, ethnicity, social class, or region, and whether they reside in the same territorial state or are more or less separated by one or two boundaries, the journalists are calling the hostility, enmity, or violence “ethnic.” At the least, we should use interethnic, interreligious, and so on, and thus avoid the foolish and simpleminded implication that it is the difference between the groups that accounts for the conflict. Finally, it is difficult to exaggerate the mesmerizing effect of calling military spending a “defense budget” that label reduces the citizen's or journalist's interest in asking where that money goes. Do the following U.S. military adventures qualify as “defense”: Spanish-American War, Cuban invasions, Vietnam, Panama, Chile, Guatemala, Nicaragua, Afghanistan? In this context of how we understand military preparedness as defense, worth noting is the recently created U.S. Pentagon “Agency of Threat Reduction,” perhaps recognizing that most of the Defense Department's resources go for threat enhancement, threat enlargement, and intimidation. Moving back from the specific to the general, one of the more debilitating semantic failures is that of how we understand two of our most central concepts: (a) theory and (b) data. Looking first at the notion of data, most of the writing in our field can barely differentiate between data and what might more clearly be called anec-data. In the more disciplined works of social science, data refers to a series of observations on years, places, events, or conditions generated by the application of a careful and precise and operational set of procedures, permitting us to compare the value or magnitude of that factor across these several observations. But among laymen, politicians, and prescientific scholars, almost any or all impressions, facts, narratives, reports, or recollections are honored

with the label data. Seeing this careless usage suggests that quite a few of us have arrived at our anecdotage even before leaving graduate school! Similarly, all too many of us seem quite willing to use the word theory for any effort to generalize ranging from a piece of folklore, a charming hunch, an off-the-wall hypothesis, or an axiomatic equation reflecting nothing more than an assumption that seems plausible. This is especially unfortunate because it gives equal status and privilege to these rather casual propositions and to a body of knowledge that (a) rests upon a set of generalizations that (b) have been operationally articulated, (c) put to a fairly stringent empirical test, (d) are logically consistent, and (e) can provide a reasonably coherent story that explains why a certain empirical regularity occurs across a particular class of comparable cases. Why, one might ask, Page 606 → would we use the same word to describe the vague ideas with which we begin an investigation and the disconfirmable explanatory scheme with which the investigation is concluded? The Ethical Conundrum Turning from the ways in which our semantic carelessness and disagreements have hobbled the effort to build a coherent and cumulative science of world politics, we might next address a number of ethical issues. Let me begin with the pernicious distinction between positive theory and normative theory. In the preceding paragraphs, I suggest what seems to be a reasonable understanding of theory in the scientific—or “positive”—sense to mean a coherent body of knowledge that helps to explain a well-defined set of empirical/historical regulations and patterns. What, then, is “normative theory”? In common usage, the distinction is between what is and what ought to be, and while this is clearly a distinction worth making, it hardly makes sense to use the same word to describe both. When we address normative—and thus, ethical—issues, it seems to me that we need to rely heavily on “positive theory,” while avoiding the tendency to conflate the two. Most scholarly writing on matters ethical tends to begin—and end—with an elaboration as to what kinds of behavior are acceptable under which empirical conditions, leading to two problems. First, and most obvious, is the inability of our ethical specialists—or in the current jargon, diplomatic “ethicists”—to operationally define the specific acts or the specific conditions under which such acts should be proscribed. Second, and perhaps even more disabling, is the reluctance to examine the consequences of a given class of behavior under a given class of conditions. For example, some would condemn and others would condone the use of military force by an outside actor in response to a serious violation of the human rights of a citizenry by its own government or a faction thereof.18 Typical of this dilemma were the situations in Yugoslavia and Rwanda. As the Serb forces laid siege to Sarajevo and Dubrovnik, Western governments vacillated and quarreled while Russia and China argued for nonintervention. Pacifist elements called for negotiation while the assault continued, and in the absence of meaningful intervention, the Milosevic forces grew in strength and brutality such that the “ethnic cleansing” campaigns against Moslems, Albanians, and Croatians inter alia escalated until the U.S. and the NATO allies finally and belatedly intervened with massive bombing attacks that lasted nearly three months. The result was several thousand casualties, a near-rupture in Chinese relations, “ethnic cleansing,” and continued genocide against Moslems in Kosovo. In my judgment, much of this tragedy could have been avoided had the Western big three pushed the UN Security Council to authorize a small ad hoc air strike against the Serb forces in early 1991. Even had the United Kingdom and United States (along with Turkey, Israel, and Sweden each providing fewer than five aircraft) taken out the Serb artillery sites in the first days, my prediction was that the ethnic cleansing campaign would never have been instituted. The point, of course, is that Page 607 → foreign intervention—by itself—could not be evaluated in moral terms, except in light of the likely consequences, and to make such a contingent forecast required more than a series of op-ed arguments. The history of world politics can be traced through such amateurish debates, with predictably inconclusive results. The dynamics of world politics are inevitably so complex that easy and consistently accurate predictions will be agonizingly rare. Thus, we need to combine norms as to (a) what actions fall where on an ethically acceptable-unacceptable scale, and (b) what are the most probable consequences of such actions, along with (c) where those predicted consequences fall on the ethically acceptable scale. In other words, the ethical problem cannot be separated from the scientific problem, and the latter rests largely on the extent to which we employ scientific method to examine such classes of historical cases and ascertain the “lessons of history.”19

The Level-of-Aggregation Question Since the appearance of the level-of-aggregation issue as a major concern among our more visible theorists,20 we have debated which level of social aggregation is most likely to lead to the more powerful explanations of war and other crucial dynamics in world politics, and with the beginning of the cumulative quantitative research programs in the mid-1960s, the issue has taken on increasing salience. While both of us converged around the significance of the system level at that time, it may have been less an awareness of the implications than suspicion that the autonomy of the state—tending to ignore the powerful constraints of the system's structure and outline—was seriously exaggerated. Whereas Waltz continued to assert that the most potent factor in world politics was system structure—by which he often meant simply the distribution of material capabilities21—a good many of us, having once asserted the system's importance, moved on to a more nuanced position. For instance, I began to look in a fairly systematic way at the national security and foreign policy decision process,22 concerned with the ways in which both the system-dominant attributes and the subsystem-dominant ones interacted23 to perpetuate the menacing state of anarchy. But the more we attended to the state level of aggregation and the decision process, the more evident it became that our research priority was premature. Many of the early “peace research” scholars took too literally the UNESCO dictum that it is in the minds of men that wars begin, failing to appreciate how unlikely we were to get at the decision rules until we had a far fuller grasp of the kinds of behavior patterns that were generated by certain classes of decision makers in response to certain classes of conditions and events. In other words, it soon became clear that in order to comprehend the “throughput” of decision making, we needed to know much more about the ways in which the input covaried with the output. This is because the myriad sets and combinations of decision rules would offer endless possible models, whereas knowing which inputs were correlated with which outputs could drastically reduce that variety by ruling out a good many of the possibilities. Page 608 → Following that reasoning, the scientific research community moved rapidly away from the more psychologicallyoriented projects and attended more and more to system level/macrolevel studies, producing a rather impressive range of findings, nicely summarized in Geller and Singer.24 But not surprisingly, this emphasis gradually came under criticism, not only because it could tell us rather little about the decision rules that might link up certain regional or global conditions and events with the incidence of war or other interaction patterns of interest at the several levels of aggregation. Furthermore, this particular paradigm inhibited the examination of process-oriented or dynamic models, with the Bremer and Cusack volume bringing together much of that criticism and promising examples of the more dynamic paradigm.25 There was, of course, another and equally critical concern: the widelyshared belief that any interesting set of international outcomes could barely be accounted for by variables at any one or two levels of aggregation, no less using only bivariate statistical analyses. Worth noting here is that my 1961 article on the “level of analysis problem” was intended only to emphasize the distinctions and not to say that a given model needs to be restricted to a single level of aggregation.26 Not surprisingly, then, the priority today is toward the multilevel and multivariate model, reflecting the consensus that outcomes of theoretical or policy interest just cannot be explained by one or two factors at one or another level of social complexity. Too many psychological, domestic, regional, dynamic, and system-level phenomena are unfolding and interacting within the same “causal” sequence, and it is essential that we keep an eye on most of them at the same time, while some of them can be held constant and others are “allowed” to change their values and magnitudes sequentially or simultaneously.

Conclusion As I consider the six profound questions that our editors asked us to address, I hope that this essay does a responsible job. But space and time preclude a truly thorough and original treatment of that assignment. Perhaps, then, the editors will permit me, in place of an original conclusion, to reprint a short—and allegedly humorous—summary of where our field stands scientifically, and how come. It was written nearly three decades ago, appeared in Political Science in 1975 under the title “Cumulativeness in the Social Sciences: Some Counter-

Prescriptions,” and with permission of the original publisher, reappears here to illustrate the extent to which our field is still plagued and stymied by some of the same bad habits that seemed back then to stand in the way of cumulative progress. Are we hardwired for failure? In every social science, there tends to be a recurrent and cyclical preoccupation with the lack of cumulativeness. Some attribute this to the familiar “absence of theory” and lay it at the doorstep of “barefooted empiricism.” Others might see the culprit lurking in the conceptual morass that often Page 609 → passes for theory and would suggest that grand schemata that are not—and usually cannot—be tested will hardly make for greater cumulativeness. There seems to be more than a germ of truth in both of these suspicions, but let me suggest a third possible source of our disappointment. I refer to certain norms and practices found among both the theorizers and the empiricists: those folkways that we pick up in college and graduate school and are seldom able to shake in the postdoctoral years. On the assumption that an awareness of them and their implications may lead to their gradual extinction, I itemize here a few of what may be our less attractive foibles. While some of them may be peculiar to the field of world politics, most seem to be found all across the discipline.

Terminology 1. If our discipline is concerned with global politics, be sure to call it “international relations” then we will all know that we're interested only in the relations among nations, and not the attributes, relationships, and interactions of all sorts of entities. 2. As long as you precede your efforts with the phrase “It doesn't matter what we call it, provided we define our terms,” feel free to ignore all conventional definitions. 3. Any time you happen across an isolated fact or a welter of verbiage, be sure to label it “data.” 4. If we already have a well-accepted word for a generally understood concept, be sure to coin a new one; don't be transparent when you can be prismatic or refractory. 5. When you're not sure which dimensions of a phenomenon you're trying to describe, refer to “the nature of ———.” 6. If you're referring to any observed or hypothesized regularity, call it “structure.” 7. If you have trouble differentiating among hunches, suppositions, convictions, preferences, and findings, just call them “theories.” 8. If you're using the word “relationship,” don't let on whether you mean covariation of variables, similarities between entities, bonds and links between entities, or something else. 9. When using the word “paradigm,” don't let on whether you mean model, research strategy, a set of axioms, epistemological criteria, or something else. 10. When using the word “parameter,” don't let on whether you refer to the isomorphism between the sample and the population, a constant, a slowly changing variable, a nonmeasured variable, a boundary condition, or something else. 11. If you are unable to articulate an idea clearly, begin to crank out a large number of examples; we'll eventually figure out what you're driving at.Page 610 → 12. Whenever you refer to physics, chemistry, or biology, call them the natural (or exact) sciences; then it'll be clear that we're in an unnatural or inexact science.

Taxonomy and Typology 1. As you shift from one research problem to another, be sure to change your taxonomy in subtle and unreported ways; this reduces the probability of integrating the results. 2. When discussing roles, relationships, or interactions, avoid identifying the social entities that play these roles or experience these relationships and interactions; otherwise, you'll be taken for a stodgy “institutionalist.” 3. When constructing a typology, don't use categories that are mutually exclusive, logically exhaustive, or rest on explicit dimensions.

4. When focusing on conflictful interactions, be sure to say that you're dealing with “the exchange system.” 5. If you're concerned with the impact of unemployment on national foreign policies, emphasize that you're examining “the economic system,” and if unemployment happens to fall unevenly among different ethnic groups, note that this requires study of “the cultural system” the more “systems” you examine, the less you'll worry about putting your findings together. 6. When one or two attributes of a system show change over time, insist that “it's a completely new system.” 7. If the folklore has it that the mitrailleuse or satellite surveillance or public diplomacy has affected the course of world politics, insist that it has led to different and successive systems.

Epistemology 1. If you disagree with a colleague's epistemology, tell him how they “do it in physics.” 2. If it is suggested that certain attributes of a social system can be described by observing the distribution of certain attributes among its subsystems, mutter something about “the ecological fallacy in reverse.” 3. If a given piece of work doesn't spell out—in mathematical form of course—all the possible relationships among variables that might obtain, observe laconically that “barefooted empiricism remains far from dead.” 4. If a colleague's work strikes you as too deductive, remind him of the importance of all those chemists in their labs, and if too inductive, quote Einstein or Bohr. 5. If a colleague's work shows a strong preoccupation with reproducibility and precision, alert others to his or her “indifference to theory.”Page 611 → 6. If, after years of urging that insight and intuition have no place in science, you discover the limits of hyperpositivism, announce that science is a failure. 7. If a colleague is not persuaded by the logic and evidence you adduce, invoke a carefully selected metaphor from everyday life; any discontinuities between child rearing and strategic deterrence, or between driving a car and running a foreign office, will be graciously overlooked.

Research Strategy 1. When undertaking a new investigation, don't bother to read prior studies in that area; until you came along, no one else did it right. 2. If your methodological repertoire is limited, emphasize that you only employ those methods “appropriate” to the specific inquiry at hand. 3. When you're stymied on the measurement of one of your critical variables, put that project aside and write another essay on what Thucydides really meant. 4. When another's study explicitly focuses on the interaction effects of predictor variables A and B, quickly note his or her “indifference” to variable C. 5. When beginning a new set of investigations, don't be misled by the plausibility of alternative models; pick one that you like and get on with the derivation of “nontrivial deductions.” 6. If your model has no recognizable similarities to the referent world, remind your critics that models are supposed to be “useful, not truthful.” 7. When a colleague strays into other disciplines for new concepts or models, alert others immediately to these dilettantish tendencies. 8. If you find a colleague working on a given class of problems for several years, point out his or her “lack of breadth.” 9. When drawing analogies from interpersonal to international relationships, never go to the findings of psychologists; their experimental evidence might be inconsistent with your argument. 10. If another project has invested considerable effort in identifying a population of nations, intergovernmental organizations, or other actors, be sure to either ignore their listing or quibble about that South Pacific island that was omitted. 11. When our critics point out the absence of cumulative knowledge about world politics, press the button that says, “After all, we're an infant discipline.” 12. If several people, at the same institution or not, are unoriginal enough to be working on the same problem

12. (such as the causes-of-war problem), be sure to observe that they are mere replicas of one another. And if they're working within a common paradigm, ask whatever happened to creativity. Page 612 →

Reportage and Communication 1. When publishing more than one paper on a given problem, don't indicate where it falls in your sequence of reports, or whether it represents an extension, refinement, or revision of earlier reports; a little mystery is good for your scholarly reputation. 2. Since it's too much trouble to revise a manuscript to incorporate the comments and suggestions you solicited from others, just tack on the necessary footnotes; the tightness of your reasoning is less critical than the illusion of exhaustive scholarship. 3. If your paper is on the casual side, don't worry about identifying sections and subsections in it; merely insert a numeral—preferably Roman—on every fifth page. 4. In writing up a nice and tight empirical study, don't waste time putting it into a larger context; get right down to the matter at hand. 5. Be sure that the title of your article or book promises considerably more than is delivered. 6. If you've picked up a good idea or insight from someone else's work, be sure to write it up as something brand new and creative. 7. If you're a panel discussant, paraphrase all the authors' caveats and self-criticisms, being sure to imply that they were oblivious to such problems. 8. Never specify or reiterate the spatiotemporal domain to which you hope to generalize; all propositions are universal. 9. Avoid the rigid tendency of making your opening and closing paragraphs consistent with each other, and in any event, don't let either of them be consistent with the actual operations you carried out. 10. Be sure that the major query of a study is carefully camouflaged, and that the outcome variable remains shrouded in mystery; otherwise, another researcher might be able to refute your findings. 11. Rather than spell out the case for the validity of your indicators, allude to your “auxiliary theory”27 and get on to the important matters. 12. When your student turns in a paper that is incoherent, disorganized, grammatically improper, and replete with errors of spelling and punctuation, ignore such trivial weakness and assure him or her that we're scientists, not literary critics. 13. To use your time efficiently, never consider undertaking the construction of a data set, but follow others' work closely, and as soon as you hear of a potentially useful set, request it from the drone who put it together; that kind of scut work is inappropriate for creative scientists. 14. If, with the support of a public or private funding agency, you generate a useful and high-quality data set, refuse it to others until you've milked it dry; you can always say that the set is not yet complete or clean. 15. If someone proposes the creation of a new journal, or a section in an existing journal, devoted to data making and index construction, cite Page 613 → Conant to the effect that empiricism is not science but is merely a poor substitute for good theory. 16. If someone suggests a different way of printing and distributing journals, remind him or her that we've always done it this way. 17. If you've read a colleague's paper and can't figure out what the hell he did or why, be sure to praise its “heuristic value.”

Ideology and Policy 1. When a colleague's work is addressed to some minor social inconvenience, such as war, hasten to note his or her indifference to poverty or injustice. 2. If a colleague's empirical findings are inconsistent with your ideological premises, point out that “it is no accident” that his or her research is supported by———, or that his nation is the richest (or poorest) in the world, or that he is not a she, an underdog, an African, or untenured.

3. When a colleague uses the same coding categories to describe both Soviet and American diplomatic actions, tell others that you don't worry about his patriotism, but … 4. When you're challenged as to the accuracy of your facts on the Chinese ABM, tell the critics, “If you knew what I know, and had a Q clearance, you'd believe me.” 5. If a taxpayer asks whether your research will lead to an improvement in U.S. policy, remind him that you're a scientist, not a politician. 6. If the question of policy implications should arise, put on your other (citizen's) hat and note that social scientists are not different from barbers. 7. When you're trying to account for the difference between the infantry's and the cavalry's share of the defense budget, invest a minimum of twelve man-years, and when you're trying to account for the differences between crises that end up in war and those that do not, write a learned essay over the Easter holidays. 8. When examining the ratio between Yale and Harvard men in the Navy Supply Corps, break out your entire methodological armamentarium, but in examining the ratio of military to civilian fatalities, rely heavily on an exegesis of the “just war” doctrine; that's what “normative theory” is for. 9. If, after extolling the virtues of “value-free” science for years, you discover that ethical considerations might just be relevant, denounce scientific method as a hoax perpetrated by the establishment in order to preserve the status quo; allusions to the Karls (Marx and Mannheim) should wrap up the argument.

The Clincher Finally, if by some remote chance, we begin to gather scientific momentum and there is a real danger that cumulativeness and codification might get out of hand, allude to Kuhn and call for a new paradigm. Page 614 →

Notes 1. John Mueller, Retreat from Doomsday: The Obsolescence of Major War (New York, Basic Books, 1989); Werner Levi, The Coming End of War (Beverly Hills, Calif.: Sage, 1981). 2. Solomon Polachek, “Peace Economics: A Trade Theory Perspective,” Peace Economics, Peace Science, and Public Policy 1, no. 2 (1994): 12–15; Bruce Russett, Grasping the Democratic Peace: Principles for a Post-Cold War World (Princeton, N.J.: Princeton University Press, 1993); O'Neal et al., “The Liberal Peace: Interdependence, Democracy, and International Comfort, 1950–1985,” Journal of Peace Research 33 (1996): 11–28; Katherine Barbieri, “Economic Interdependence: A Path to Peace or a Source of Interstate Conflict?” Journal of Peace Research 33 (1996): 29–49. 3. Lewis Richardson, Arms and Insecurity: A Mathematical Study of the Causes and Origins of War (Pittsburgh: Boxwood Press, 1960); Quincy Wright, A Study of War (Chicago: University of Chicago Press, 1942). 4. J. David Singer, “The Etiology of Interstate War: A Natural History Approach,” in What Do We Know about War? ed. John Vasquez (Boulder, Colo.: Rowman and Littlefield, 2000). 5. Karl Mannheim, Essays on the Sociology of Knowledge (New York: Oxford University Press, 1952). 6. Hayward R. Alker, Rediscoveries and Reformulations (Cambridge: Cambridge University Press, 1996). 7. Peter Haas, “Do Regimes Matter? Epistemic Communities and Mediterranean Pollution Control,” International Organization 43, no. 3 (1989): 377–403. 8. Hedley Bull, “Strategic Studies and Its Critics,” World Politics 20, no. 4 (1968): 593–605. 9. David Dessler, “Beyond Correlations: Toward a Causal Theory of War,” International Studies Quarterly 35 (1991): 343, 345. 10. Roy Bhasker, A Realist Theory of Science, 2d. ed. (London: Routledge and Kegan Paul, 1986). 11. Robert H. Bates, Analytic Narratives (Princeton, N.J.: Princeton University Press, 1998). 12. Alexander L. George, “Case Studies and Theory Development: The Method of Structured, Focused

Comparison,” in Diplomacy: New Approaches in History, Theory, and Policy, ed. Paul G. Lauren (New York: Free Press, 1979), 43–68. 13. Carl G. Hempel, Philosophy of Natural Science (Saddle River, N.J., Prentice Hall, 1966); Karl Popper, Logic of Scientific Discovery (New York, Basic Books, 1959); Rom Harre, The Method of Science (London: Wykeham, 1970); Imre Lakatos and Alan Musgrave, Criticism and the Growth of Knowledge (Cambridge: Cambridge University Press, 1970). 14. Herbert Simon, Science of the Artificial (Cambridge: MIT Press, 1969). 15. William L. Langer, European Alliances and Alignments, 1871–1890 (New York, Knopf, 1950); Edward V. Gulick, Europe's Classical Balance of Power (New York: Norton, 1955). 16. Glenn Snyder and Paul Diesing, Conflict among Nations (Princeton, N.J.: Princeton University Press, 1977). 17. Inis L. Claude, Power and International Relations (New York: Random House, 1962). 18. Laura W. Reed and Carl Kaysen, eds., Emerging Norms of Justified Intervention (Cambridge, Mass.: American Academy of Arts and Sciences, 1993). 19. J. David Singer, “Conflict Research, the Security Dilemma, and Learning from History,” in Behavior, Culture, and Conflict in World Politics, ed. William Zimmerman and Harold Jacobson (Ann Arbor: University of Michigan Press, 1993). 20. J. David Singer, “The Level-of-Analysis Problem in International Relations,” World Politics 14, no. 1 (1961): 77–92; Kenneth N. Waltz, Man, the State, and War: A Theoretical Analysis (New York: Columbia University Press, 1959).Page 615 → 21. Waltz, Man, the State, and War; Kenneth N. Waltz, Theory of International Politics (Reading, Mass.: Addison-Wesley, 1979). 22. J. David Singer, “Inter-Nation Influence: A Formal Model,” American Political Science Review 57, no. 2 (1963): 420–30. 23. Morton Kaplan, System and Process in International Politics (New York: Wiley, 1957). 24. Daniel S. Geller and J. David Singer, Nations at War: A Scientific Study of International Conflict (Cambridge: Cambridge University Press, 1998), chs. 5 and 6. 25. Stuart A. Bremer and Thomas R. Cusack, eds., The Process of War: Advancing the Scientific Study of War (New York: Gordon and Breach, 1995). 26. Singer, “The Level-of-Analysis Problem.” 27. Herbert M. Blalock, Social Statistics (New York: McGraw-Hill, 1979).

Page 616 →

NOTES FROM THE UNDERGROUND A Tale of Three Perspectives Linda B. Miller The invitation to participate in a millennial reflections panel on international security and peace studies and to publish these remarks arrived at a propitious moment. The public opportunity to reappraise my own academic career as I was already doing privately meant a chance to ruminate in a way that might have value for younger scholars at earlier stages of their careers. And since my own trajectory closely parallels that of the International Studies Association (ISA) in terms of time, such an overview should be of general interest to the membership. In that spirit, I offer what I have titled “Notes from the Underground.” Such a literary title signifies not only my origins as an undergraduate student of literature and history before I embarked on the professional study of political science and international relations (IR) in graduate school, but also my nearly four decades of teaching undergraduates, primarily women, in liberal arts colleges, “the underground” in relationship to major research universities, where much of the research and writing on international security originates. I have held research appointments at a number of U.S. institutions (Princeton, Harvard, Columbia, Brown), which has kept me up-todate with major findings and alleged paradigm shifts. In many ways, then, I have had the best of a variety of worlds, with the usual balance of successes and disappointments, of recognition and underappreciation in roughly equal measure. More recently, I have added a third perspective to the first two of student and teacher, that of editor of an official ISA journal, International Studies Review. I am struck by the ways in which these three perspectives—student, teacher-researcher, and editor—have built on each other, in part, as Bob Jervis reminds us, because we tend to fit new data into existing categories, and, in part, because we do construct our own autobiographies, perhaps at the risk of not being able to get out of them. In this supposedly new age of “narratives” and “discourses,” we should examine our perspectives periodically to see not only what needs to be discarded but also what should be retained, whatever its origins. I am not surprised Page 617 → that much of what I learned as an undergraduate in the late 1950s at Harvard is relevant in 2000, especially the need for multidisciplinary insights, for knowledge of people and places as the essence of politics. More surprising, given the pedestrian quality of graduate education at Columbia in the early to mid-1960s, some of what I learned there is equally vital, especially the idea that national and by extension international security is, in Arnold Wolfers's phrase, “an ambiguous symbol.”1 The important role of history as the database for international relations rather than the hard sciences was a self-evident truth. How else could patterns in human behavior over time and space be discerned and analyzed? Over the decades of teaching and writing, I think I have learned as much from my students as I have from my own research, at least when I reflect on the matter away from the drudgery of paper grading! In my years in the classroom at Barnard, Wellesley, Harvard, and Brown, I have repeatedly stressed the importance of clear thinking. I have urged students to find an individual voice, shorn of the slavish obeisance to higher authority that mars so much of the scholarship that comes my way now as an editor. The students' demand that theory must explain something, not just parade itself as categorical, especially if it is parsimonious, has attained the stature of a congenial commandment. Since many of these students come to the study of international relations from nonAmerican backgrounds, the idea that their experiences and their governments and societies must be part of any informed conversation is assumed, though it still seems to be controversial in some of the parochial professional literature. Similarly, students instinctively weigh the interplay of domestic and international factors in their analyses of security and insecurity. Classical political theory, Western and non-Western, as a rich source of complex questions is sometimes a hard sell in the classroom, but it is a superior antidote to the current events sprawl that may have brought students to the class in the first place. Recently, path-breaking work in comparative

political theory2 has rejuvenated the prospects for giving students a richer understanding of human nature on which theories of security rest, albeit uneasily. So embedded in my own approach are these bedrock views that debates about their validity seem strained and artificial. Yet as an editor I confront daily the lengthy efforts to reinvent these wheels before getting to the heart of a manuscript's arguments. Particularly persistent are the endless acknowledgments or references to texts like Kenneth Waltz's Theory of International Politics that have shaped the field but threaten to distort it by narrowing the agenda of acceptable questions to investigate more than twenty years after publication.3 In a changed world, this tendency is a serious impediment in a field that likes to think of itself as “cutting edge” rather than derivative. Students who are able to cultivate a toleration for ambiguity are comfortable with the notion that the state may be both “obsolete” and “obstinate,” to use Stanley Hoffmann's trenchant formulation.4 Often, they have little difficulty understanding that globalization may actually stimulate national identity or, better, a plethora of local identities.5 They are frequently Page 618 → able to distinguish between tribalism that stems from a lack of diversity within societies or little contact with the outside world and the fear of the “other” that demagogues exploit as a means of obtaining and perpetuating political power. Although they are intrigued with the possibility that democracies might not go to war against each other now, they are persuaded that this observation might be limited to the North Atlantic-North American region rather than constitute a law of global politics in the twentyfirst century. At the best of times, the models or guideposts we employ may lead to “generic knowledge,” in Alexander George's phrase,6 if we are alert to politics on the ground as the true test of IR theorizing. This organic connection between theory and practice is often buried in the self-styled “great debates” that have roiled the field of security studies periodically. Strikingly, my students majoring in the natural sciences have been the most critical of efforts to apply “scientific” methods to the social universe of social science. They are equally surprised that there is any controversy over whether ideas as well as material factors matter in the drama of large events like the end of the cold war. Just as they accept the blurred lines between comparative politics and international relations, they take constructivism on board in the same critical fashion they bring to realism and idealism, to neorealism and liberal internationalism. Eclecticism in the pursuit of truth or as a way of ensuring the legitimacy of “multiple pathways”7 is not a cop-out. It is not necessary to romanticize the common sense of undergraduates, to insist on the wisdom of babes, as providing a sounder basis for advanced study than the serious work of scholars who have asked the right questions and never been satisfied with easy answers or sloppy methods. No, I cite these examples and others that follow because we often go astray when we insist on the primacy of theory, its supposed elegance and coherence, when we know that these intellectual constructs are just that. Theories do not and cannot on their own delineate the range of choice and consequence, let alone the sequence of cause and effect, of international relations in its many and wondrous parts. Theory alone cannot tell us when it is appropriate to speak truth to power, though it may make such speech more penetrating when it does take place. Clearly, attention to other disciplines has its hazards as well as its advantages. Political scientists who have struggled to incorporate the dictums of economics into their toolkits, to insist on rational choice assertions as most appropriate to the study of political phenomena, have been too slow to accept revised thinking about the limits of this technique. Economists who now see the benefits of psychology in understanding rationality's limits are leaving political scientists behind the curve, although “emotion,” along with “cognition,” is now attracting some IR scholars as a fruitful, if still tenuous, research topic. Economists who have been investigating the utility of biologically driven evolutionary theories for several decades are now meeting equally skeptical political scientists, so the infatuation with natural science models may be diminishing somewhat. Page 619 → We have been here before. Political scientists clung to obsolete hard science methods long after physicists had left the Newtonian world far behind. What keeps political scientists in thrall to the theories of other disciplines for too

long is not the superior wisdom of these approaches so much as the dogmatic belittling of those social scientists who are inherently suspicious of disciplinary imports. That unfortunate deformity, coupled with a tendency of American graduate education in political science to embrace fads, produces the next generation of a professoriate that is increasingly specialized and doctrinaire. The implications of this dreary trend are explored at length in two recent issues of International Security, the leading U.S. journal. What is striking about the exchange between Stephen Walt and a variety of other scholars is the debate about whether formal models should be judged in part by their utility in analyzing the “real world” as opposed to such criteria as logical consistency.8 Unwilling to concede any ground on this point, despite the wealth of examples that show formal models as dramatically flawed with respect to empirical tests, the quantifiers are condemned to ex post facto interpretations of events like the end of the cold war that nonformal scholars have already explored and incorporated into their own less rigid theories. What is the way forward? Not surprisingly, my preferences reflect the three perspectives I have outlined. Since indiscriminate borrowing from other disciplines continues to be such a dubious undertaking, I recommend that we encourage cross-fertilization based on linkages between political science and other social sciences that respect the core of those subjects. Expressed more concretely, large events in the contemporary political world, like the collapse of territorial empire, the upsurge in ethnic conflict or genocide, or the insistence on identity, might be analyzed by groups of scholars brought together by a belief that each of their disciplines has something needed to complete an empirically-rich picture of the phenomena, especially when the events themselves take place in areas of the world less well-known to U.S.-based security researchers. Fortunately, such efforts are already under way and reaching publication in book form, notably the Borderline series of the University of Minnesota Press. In volume 14, for example, anthropologists, steeped in cultural studies, explore the state in conjunction with political scientists who investigate culture in ways reminiscent of earlier scholarship on war and peace before the cold war and its restricted research agenda took over mainstream IR scholarship. A fruitful result of this collaboration appears in chapter 11, where Hugh Gusterson argues that authors in International Security not only failed to prepare their readers for the end of the cold war, a tall order to be sure, but did so by insisting that debates over weapons systems were more worthy of discussion and financial support from foundations than mere speculation about transformative historical change that they could not imagine.9 A final reason for proposing group research where scholars cross disciplinary boundaries with their training and intuition intact, rather than projects where some concepts must be ranked lower due to their disciplinary origins, Page 620 → reflects the increasing interest in normative “theories” that promise not policy relevance but serious reconsideration of accepted practices in world politics and domestic affairs associated with the imperatives of hegemony derived from five decades of cold war theory and practice. In security studies, this hegemonic bent has crippled U.S. scholarship much as the same tendency has produced some tarnished outcomes in American foreign policy. Renewed interest in ethics and international affairs, long treated as irrelevant to the main struggles of world politics, has revealed the poverty of much previous scholarship, with its pretensions and self-referential discourse. Only when we acknowledge that “foreign policy is an inherently moral activity” and that “the national interest is an inherently moral category”10 will we begin to reposition the study of international security in a more secure intellectual context. Only then will the construction and assimilation of norms move to the center of our vision from the periphery where it languished during the cold war. Only then will the newer debates about globalization and governance as security issues attract the sustained interest they so richly deserve. Even as the first fruits of this intellectual journey are reaching publication, a growing insistence that our studies must try to repair the world—tikkun olam—is apparent. What such a reconnection between theory and practice will accomplish is admirable: a restoration of security studies so that it fulfills the overarching goal of my three perspectives, a liberal art in the fullest meaning of the term.

Notes 1. Arnold Wolfers, Discord and Collaboration (Baltimore: Johns Hopkins University Press, 1962), ch. 10. 2. Roxanne Euben, Enemy in the Mirror (Princeton, N.J.: Princeton University Press, 1999). 3. Kenneth N. Waltz, Theory of International Politics (Reading, Mass.: Addison-Wesley, 1979). For a spirited defense of structural realism, see Kenneth N. Waltz, “Structural Realism after the Cold War,” International Security 25, no. 1 (2000): 5–41. 4. For a restatement and reappraisal of this term, see Stanley Hoffmann, “A Retrospective on World Politics,” in Ideas and Ideals, ed. Linda B. Miller and Michael J. Smith (Boulder, Colo.: Westview, 1993), ch. 1. 5. See G. Pascal Zachary, “Get over It,” Foreign Policy (September-October 2000): 62. 6. Alexander L. George, “Knowledge for Statecraft: The Challenge for Political Science and History,” International Security 22, no. 1 (1997): 44–52. 7. Donald Puchala, “Marking a Weberian Moment: Our Discipline Looks Ahead,” International Studies Perspectives 1, no. 2 (2000): 141. 8. Michael Brown et al., eds., Rational Choice and Security Studies (Cambridge: MIT Press, 2000). 9. Jutta Weldes et al., eds., Cultures of Insecurity (Minneapolis: University of Minnesota Press, 1999). 10. David Welch, “Ethics and Foreign Policy,” Georgetown Journal of International Affairs 1, no. 1 (2000): 79.

Page 621 →

INTERNATIONAL POLITICAL ECONOMY

Page 622 → Page 623 →

REFLECTIONS ON THE FIELD OF INTERNATIONAL POLITICAL ECONOMY Helen Milner

Introduction The academic field of international political economy (IPE) is a relatively young one. As an established part of international relations (IR), it rose to prominence in the late 1960s and early 1970s. Scholarly writings that would now be called IPE date back much further,1 but they were not widely-recognized as part of a distinct field until the 1970s.2 This means that the field is only about thirty years old. The period of oil shocks in the developed countries beginning in the early 1970s provided a strong impetus to the organization of the field. It brought to the fore a set of questions that either had not been addressed in IR before or had been dismissed as unimportant. Concern with these “new” issues energized scholars and prompted the founding of the field. The key questions at that time dealt with five issues central to IR: the use of economic instruments of statecraft, especially relative to military force; hegemonic stability theory and the decline of the United States, or more broadly of the West; the role of domestic politics; dependency theory and development; and the importance of international institutions. The questions that animated the field were ones such as the following: (1) was military force still useful given the new (or renewed) importance of economic resources? (2) had the United States lost its hegemony, especially in becoming so dependent on foreign oil from a group of less-developed countries? (3) why did the advanced industrial countries, which were fairly similar in their domestic structures and international positions, respond so differently to the oil shocks? (4) were the lesser-developed countries all bound to remain peripheral, dependent economies trapped in low-growth trajectories? and (5) did this period of monetary and real economic shocks elevate the importance of international institutions, such as OPEC, the International Monetary Fund, the World Bank, and the International Energy Agency, in world politics? Since then, research in the area has grown greatly, and the field has become well-established within American and European Page 624 → political science departments.3 However, a number of the issues occupying the heart of the field have changed. As I argue below, hegemonic stability theory has largely faded, and the other issues identified earlier in this paragraph have taken on new and different casts in light of the field's current preoccupation with the causes and consequences of globalization.

Definition of the Field What is IPE? This is less clear than ever. Defining the field is not obvious and is often done in two distinct ways. On the one hand, it is defined as everything that is not part of security studies in IR; on the other it is more narrowly defined (as at its inception) as dealing with either economic independent or dependent variables—e.g., economic factors as causes or economic outcomes as consequences. In this latter, narrower sense IPE is defined as only those issues related to the interaction of politics and economics, or states and markets. To be part of IPE, a study must have either an economic independent or dependent variable. The definition of the field matters since it helps to adumbrate the central issues that demarcate the enterprise. In the first, more expansive definition, IPE includes basically all issues not in security studies. Fundamentally, this seems to mean any issue where the use of military force either does not occur or is not a likely event or a central preoccupation. Many interpret the field in this broader sense and include a vast array of issues and approaches in IR. IPE tends in this version then to become the study of a huge range of phenomena that often have very little to do with economic factors, either as causes or outcomes. These include the study of international institutions—whether or not they involve economic issues—environmental issues, human rights, the role of women and minorities, and international cooperation of any sort. Some go even further and extend IPE far beyond the confines of IR or even these broad issues. Tooze, for

instance, argues that “international political economy as a focus of inquiry extends beyond the problematic of conventional international relations…. An IPE problematic will initially ask questions about assumptions and values. IPE developed as a critique of existing orthodoxy through exposing the implicit assumptions and values in accepted approaches.”4 Or as Strange claims, “The whole point in studying IPE rather than international relations is to extend more widely the conventional limits of the study of politics, and the conventional concepts of who engages in politics, and of how and by whom power is exercised to influence outcomes. Far from being a subdiscipline of international relations, IPE should claim that international relations are a subdiscipline of IPE.”5 Others desire to broaden the field even further. Amoore et al. call not only for the “historicization” of IPE but also for considering “how time and history condition the mode of analysis appropriate to the social sciences, and how human subjectivity infuses the theory/practice relationship that in turn shapes our understanding of the world. In other words, we have to go beyond questions of methodology to embrace a substantive concern with the historicity Page 625 → of knowledge itself before we can genuinely construct a historicized IPE. Such a concern raises the issue of ‘reflection’ as a key problematic to consider in any attempt to historicize IPE.”6 Is this expansion of the field useful? While all of the topics mentioned above are important, they do not seem to be relevant solely or even largely to IPE. The role of norms and values, “critical” theory, reflectivity, methodology, epistemology, and ontology all are relevant to IR, political science, and social science more generally. When researching the relationship between economics and politics, all of these issues can be important; and when done in the context of this relationship, they are part of IPE. But they are not intrinsically or largely the domain of IPE. Neither are other important areas of IR. IPE should not be conflated with research into international institutions or cooperation. IPE is not just about cooperation or about international institutions. The field of IPE may include these if they relate to economic factors or issues, but not all institutions or cooperation involve economic dimensions. Studies of the UN, for example, are not intrinsically part of IPE; they are only if they include some economic component. Furthermore, IPE also involves the study of conflict, both militarized and not; it is not just concerned with cooperation. The economic causes of war and peace or the economic consequences of military conflict, for example, certainly fall within the realm of IPE. Moreover, much of IPE does not necessarily involve international institutions; a great deal of economic interchange has little to do with formal international institutions or cooperation. IPE and the study of international institutions and/or cooperation are different subjects, although they do overlap when they involve economic issues. The narrower definition of IPE is more heuristic. It assumes that economic factors are an inseparable part of the field. The interaction of politics and economics, or more narrowly markets and states, becomes key in this approach. It involves showing how political factors like government policy choices influence economic outcomes, especially the operation of markets; and conversely, it entails showing how economic phenomena may alter the way politics operates, often by changing the preferences and/or capacities of the actors. The mutual interaction of these two then is the central distinguishing feature of the field. Without an economic component, I would argue, the phenomena are not properly part of IPE. So political studies about topics that have little or no economic aspect should not be considered IPE.7 How then is IPE related to IR? Is IPE just a subfield of IR? Or does its reach extend into IR but also beyond? I maintain here that the latter is the case. IPE overlaps with important parts of IR, but it also branches out into other fields, drawing both methodological tools and substantive questions from them. It certainly has strong links, for instance, with economics and comparative politics. Indeed, the early work in IPE was very much geared as a response against economists: it consisted largely in showing economists that they had to add political variables to their models. The scholarship of Robert Gilpin on multinationals, that of Benjamin Cohen on imperialism, and the work of Susan Page 626 → Strange on the nature of the international monetary system are all examples of research that rejected economists' explanations of these phenomena and emphasized the effects of politics. They challenged economists' explanations of international trade, production, and finance that ignored politics. Gilpin's book demonstrated that one had to consider the role of power in shaping markets and the operation of multinational corporations.8 Cohen's book on imperialism also challenged economists by arguing that the balance of power among leading states was the central cause of this

phenomenon.9 Susan Strange highlighted the role of American and British power in shaping the international monetary system.10 These authors challenged economists' arguments about the international economy, showing that one could not adequately understand these topics without knowledge of political factors. What is the state of this relationship between the two fields now? Today parts of the field are tightly linked to the discipline of economics. This is especially true in studies of economic policy-making. Some, like Strange herself, have argued that there is too much dependence—and perhaps even an uncritical reliance—in IPE on economics and its models. Nevertheless, a wide range of scholars use economic models as a basis for understanding the preferences or likely behavior of actors, and in so doing fuse the two fields.11 Others continue the earlier tradition of showing economists that they must add political variables to their explanations.12 And others, of course, reject “economic approaches” out of hand as being ahistorical, decontextualized, and conservative, as some of the earlier authors cited claim. So what should be the relationship between IPE and economics, especially international economics? Some argue that we should jettison economic theorizing entirely, that rational choice models simply cannot give us much understanding of important political (and social) phenomena. Strange, echoing many others, claims that economics “as a discipline tends to exaggerate the rationality in human behaviour…. How much more has international economic history shown that political choices on economic policies have seldom been motivated by carefully reasoned assessments of quantifiable economic costs and benefits, but rather by political aims and fears, and sometimes by totally irrelevant considerations and irrational emotions.”13 This complete rejection seems an extreme position for a field that wants to explain many of the same things that the economics profession has been working on for a century or more. If one takes a less extreme view, however, what are the possibilities? Should we just be adding political variables to economists' models? Or should we be rewriting their models by adding political content to economic theories? Interestingly, the latter is what economists themselves are starting to do, as I shall discuss later. How are IPE and comparative politics related? More and more IPE work is comparative and domestic in flavor. Although much IPE scholarship earlier was comparative in spirit,14 it is now very hard to separate the two areas. Integrating the two fields seems to be an important avenue for future progress. Studies in IPE could benefit from more use of the theory and knowledge of domestic politics developed in comparative politics. It is imperative to a better understanding of IPE topics, like trade, money, and development, Page 627 → to use models of domestic politics that explain how policy is made. Understanding where actors' preferences come from and how they are “aggregated” formally or informally into actual policy choices, for instance, is a fundamental element of comparative politics that bears strongly on IPE questions. Incorporating variables from traditional comparative politics, such as domestic political institutions, could provide fruitful new ways to explore IPE issues. Some of this has occurred as scholars explore variables such as the nature of electoral rules, the party system, the legislative process, and the number of veto players for explaining policy choices, as Rogowski and Simmons have done.15 Fruitful collaboration between comparative politics and IPE has been and will continue to be a source of progress for the field. As the diagram in figure 1 suggests, IPE has links with a number of fields. Its reach extends into and beyond IR. And it certainly has links with comparative politics and economics. As I elaborate later, the field is ever more tightly linked to comparative politics. It is difficult today to distinguish the fields of comparative and international political economy. In part because of globalization, much of the most interesting work involves the crossover area where scholars look either at the way in which domestic politics affects the international economy or at how international economic factors shape domestic politics.

Key Issues Why does the definition of the field matter? Defining the field is important because it focuses attention on a common set of critical questions and issues. The narrower definition proposed earlier has the advantage that it draws researchers to concentrate on a well-defined series of topics. In the Page 628 → field of security studies, for

example, most researchers agree that the central topic is war; the search for its causes, consequences, and prevention provide a unifying set of issues in the field. In IPE, there is probably no such single focus of attention. But there are two or three topics that have long dominated the field's agenda: economic development and its cognate, differential rates of economic growth; the effect of domestic politics on the international economy and vice versa; and conflict and cooperation among states in the international economy. First, the issue of economic development and differential growth rates has long been a central one. Why some countries (or regions) grow faster than others has been a motivating research question for many studies since the field's inception. Political scientists have tended to focus on the political determinants of economic growth and development. This is limited not just to studies of current developing countries16 but also to studies of how the West became the center of the industrial revolution and the consequences of its economic superiority,17 as well as to studies of the political prerequisites of economic growth for developing and transition countries.18 In addition, many studies of economic policy-making by political scientists have sought to understand policy choices largely because of their impact—often indirect and untested—on the differential economic growth rates of countries.19 A second traditional issue that defines the field is the impact of the international economy on domestic politics and vice versa. The older literature on interdependence and the more recent work on globalization are examples.20 Dependency theory also examined this question as well as the former one about economic growth.21 The interaction between domestic and international political economy is a major area of interest that persists today. A third topic that has long defined the field's range of interests has been the examination of conflict and cooperation among states in the international economy. Attention to the role of international institutions in maintaining good economic relations among states and in fostering greater cooperation has been a central issue.22 The literature on trade wars, competitive devaluations, and other forms of economic conflict has also been prominent.23 This research area also includes studies of economic sanctions and statecraft,24 as well as research on the ways in which economic assets can help countries achieve their political goals against the will of other countries.25 The central questions all relate to the interaction of politics and economics among states in the international system. The political, economic, and social causes and consequences of differential economic growth rates for nations are the central focus when the field of IPE is more narrowly defined. What are the main research agendas in IPE today? They continue to concern these three central questions by which the field is defined, but they have also been refashioned over time, both as a result of accumulating knowledge in the field and changes in the world. The main research agendas have altered since the 1970s, although many of the new themes have evolved Page 629 → from older ones. Previously, in the 1970s and early 1980s the main themes, as pointed out earlier, concerned economic versus military sources of power; hegemonic stability theory; the role of domestic politics in policy-making, especially the role of state-society interaction in this process (e.g., strong versus weak state and societal arguments); dependency theory and development; and the role of international institutions in promoting economic cooperation, especially among the G-7 countries. Changes have occurred in the field's focus of attention. Some themes, like dependency theory and hegemonic stability theory, have been overtaken by world events. Others, like the domestic politics and international institutions ones, have been transformed from their earlier incarnations into new ones associated with the rise of globalization. World events have affected the focus of the field, although they have not completely altered it. Despite these realworld changes, a number of the older themes are still (indeed, even more) relevant and pursued by scholars in IPE. The wave of democratization and economic liberalization that began in the mid-1980s and accelerated after the end of the cold war has helped to remake the agenda in the development area. Combined with the problems that dependency theory had in dealing with the increasing heterogeneity in economic performance and status of the once-developing countries, these changes drew attention away from the earlier issues. Currently the impact of democratization, and regime type more generally, on development is a key topic, and the interaction of economic and political reform has become central. Dependency theory has largely been discarded, but concerns over the impact of capitalist world markets on small developing countries—as well as on large countries—persist. Although the logic of dependency theory is no longer accepted by many, worries about how small countries can achieve stable and high levels of economic development, especially when buffeted by very powerful global

economic pressures, are still at the forefront of research and policy concerns. Hegemonic stability theory (HST) has suffered a similar decline in attention within the field. A dominant research agenda in much of the 1970s and 1980s, the theory eventually ran into both empirical and theoretical difficulties. While never empirically “falsified,” HST proved to have numerous difficulties explaining the more specific changes in stability and openness in the international economy.26 In addition, the revival of the United States and the relative decline of many other potential challengers, like the Soviet Union and Japan, since the late 1980s alleviated worries over the loss of hegemony. Finally, the challenges to the theory's logic were never successfully rebutted.27 Since the late 1980s scholars have been therefore less interested in the theory, in spite of the newer concerns over American “hyper-power” by non-Americans. The issue of globalization has captured the heart of the field's attention over the last decade. The phenomenal growth of world trade and capital flows has prompted renewed attention to the whole topic of international economic integration. Both the causes of and consequences of a global market for nation-states have become hotbeds of research. This new focus has Page 630 → transformed several earlier themes pursued by the field; research on the relationship between domestic and international political economy, on the role of international institutions, and on the foundations of power in world politics has been modified over time. Globalization as a process has cast increasing doubt on the traditional separation of the domestic and international levels of analysis. The mutual interaction between the two is dramatically evidenced today, and research on this interaction—often in the form of two-level games—has burgeoned.28 Very few studies of comparative political economy these days can avoid including research on the impact of global markets, as they might have a decade ago.29 In addition, scholars have begun examining more carefully the impact of domestic politics on the prospects for international integration. In particular, issues about how regime type (democracy versus autocracy) affects economic reform, growth, and liberalization have gained prominence.30 Moreover, research about the roles of international institutions has also been altered perceptibly. The growing need for these institutions to stabilize and organize the increasingly-integrated world economy has put notable pressure on them. Research into what role they can hope to play, whether they act as independent agents, and what structures and procedures will be most efficacious is of great interest today.31 The literature on economic cooperation and conflict has also been subtly transformed. International institutions have gained in importance both in scholarship within the field and in policymakers' eyes. The great momentum in European integration that began in the early 1980s, for instance, has generated renewed attention to this area but in a different guise than earlier. Applications of theories of international cooperation and domestic political institutions to explain European Union integration are booming areas of research today.32 These recent themes, then, centered on globalization—e.g., the interaction of domestic and international factors, two-level games in foreign economic policy-making, regime type and economic policy (effects of democracy on policy, crises, political and economic reform, and development), and the political economy of European integration—constitute some of the main, new centers of gravity in the field.

Methodological Debates What is the state of methodological debates in IPE? I would argue that they tend to mirror those in IR and political science generally. The two debates of central importance involve the use of rational choice techniques similar to those in economics and the value of large-N versus small-N (or case) studies. Many scholars question the value of rational choice techniques that derive from economics, as already noted.33 Even more doubt arises over the utility of formal rational choice models, although in truth there are only a few IPE scholars who use formal models.34 Again the definition of IPE matters; using formal rational choice techniques is not enough to make you a scholar of IPE in the more narrow definition of the field proposed before. Moreover, as I claim elsewhere, ironically IPE is less formalized than security studies for better or worse.35 Page 631 → Is the assumption of rational choice in general useful in IPE? The issue here relates to both rationality and choice.

Some scholars, like Strange, argue against the likelihood of rational behavior, while accepting that actors in IPE may have some latitude for choice. Others object to the formulation as one of choice; they see constraints on actors (usually from the system or existing norms) as preeminent. Others, of course, reject both tenets. It seems harder to abandon or argue against the assumption of (some) rationalism in IPE than in other areas not concerned with the economy. Given the tight connection to economics and the fact that actors here are often making decisions about economic variables—where costs and benefits are frequently measurable—the assumption of rationalism seems more tenable. This is perhaps why there exists much less work on psychology in IPE than in other areas of IR; and, I would argue, less in the constructivist mode—at least in IPE defined more narrowly. Consideration of the costs and benefits in taking actions as impressed by the operation of markets means that agents here are often calculating with some element of reason; and the only reason to calculate is if they have some degree of choice. Does formalism help? Sometimes it may, and sometimes it may not. If one wants to communicate or collaborate with the economists, it is essential. And if one believes that being very clear about the assumptions one uses, and how they logically must lead to one's conclusions, then the formal statement of arguments is imperative. If one cares less about communicating with economists and if the clear statement of one's precise assumptions is not possible, then formalism may be less useful. In areas where our theories of politics are poorly developed, there is likely to exist little understanding of the causal relationship between variables, let alone any precise knowledge of how different assumptions lead to different causal relations. In these situations formal models may not be very helpful, and may even be misleading. The theorist may not comprehend the restrictive nature of her specific assumptions and hence see the particular formal model as more general than it is. The utility of this approach, as with others, depends on the question at hand and the state of the field. Does rational choice explain everything about IPE? Certainly not. There is plenty of room for other approaches, and more to be gained from combining different approaches than in limiting them. Most economists, let alone political scientists, recognize this. As one volume assessing the contributions of Nobel Prize-winning economists to political science has pointed out, “one theme pervades [the Nobelists' reflections]: a chafing dissatisfaction with the standard neoclassical paradigm of economic analysis…. There is no dissatisfaction evident with the fundamental postulate of neoclassical economics—that the unit of analysis on which all else is built is the individual choice…. Nevertheless … the neoclassical paradigm … is being not just picked at but massively reworked by those dissatisfied with its assumptions as a basis for analyzing human behavior.”36 As economists are discovering, there are interesting models of nonrational behavior that can be applied. The essential element is that these models themselves have to be rigorous, formally stated, and/or empirically testable for the economists Page 632 → to find them useful. The work on prospect theory begun by Kahneman and Tversky has been one of the most notable diversions from rationalist models.37 Movement away from rational choice to more cognitive models seems increasingly likely. But movement away from the postulate of agent's choice will clearly not be led by economists or psychologists; it is more likely to come from scholars in the fields of political science and sociology, and especially from either systems theory or constructivism, where choice and agency are less important. In terms of empirical methods, what are the debates in the field? Again, I think they are the same as elsewhere in political science. The central issue seems to be how valuable are case studies versus large-N analysis. What is the value of large-N studies that use statistical techniques to analyze data? Several objections to them reappear: scholars wonder about the noncomparability and uniqueness of cases, the inadequacy of quantitative measurement of variables, and the overestimation of the precision of findings. Their contributions include greater confidence in the external validity of claims, and less concern about selection biases. How useful are case studies? Objections involve their low external validity and weak generalizability, as well as serious worries about replicability and selection bias. The contributions of case studies lie in their ability to help scholars think through a theory and its causal mechanisms. Often they can offer some evidence when no better data can be found or provide some critical test of a new idea. They also can be used to show the internal validity of the argument explored. Both methods then are useful for different tasks. When it is especially hard to adequately quantify some variables and/or our predictions from theory are clear, cases may be very helpful. But

case studies alone may have serious biases in them. It is probably ideal to combine the two approaches. One can then flesh out the logic of the argument in the cases and determine if it has some external validity in the large-N analysis.

Some Conclusions Is there consensus in the field of IPE on substance and methods? This depends at least on how one defines the field. The more narrowly it is defined, the more strongly one probably can answer this in the affirmative. Finally, how extreme are the cross-national differences in conceptions of the field of IPE? Are there significant differences at least in European and American views of the field? Waever in the International Organization retrospective volume argues that there are, and that Europeans will never adopt the social “scientific” approaches common in the American community.38 I would dispute this. It is true that the European academic community in IR has been less enamored with such approaches, although this is not true in other fields. The methods of economics and the natural sciences are as “scientific” in Europe as in the United States, so any differences in political science have nothing to do with the European academic culture. Indeed, I Page 633 → would argue that the field of political science in general in Europe has many of the same kinds of divisions over definitions of the field, its methods, and its theories as in the United States. As a concluding point, it is apparent that economists are increasingly taking over areas of IPE and comparative political economy. They are adding political variables to their models, combining international and domestic levels, and moving into the study of international institutions.39 If IPE scholars do not respond, economists shall monopolize the field for better or worse. There are at least two ways the field could respond to this. One is to move away as quickly and dramatically as possible from the central tenets of economics and thereby establish the autonomy, if not antinomy, of IPE to economics. The other way is to engage in more economically-sophisticated work in the field, taking on the economists on their own ground and demonstrating once again that to understand IPE requires more than just knowledge of economics. This path requires that political scientists have a thorough knowledge of the economics of their research area and integrate their findings with those of economists; it fosters an interdisciplinary brand of IPE. These paths are not mutually exclusive, of course. Both may serve to contribute to our knowledge of the international political economy.

Notes 1. For example, Albert Hirschman, National Power and the Structure of Foreign Trade (Berkeley: University of California Press, 1945); Jacob Viner, “Power and Plenty as Objectives of Foreign Policy in the 17th and 18th Centuries,” World Politics 1, no. 1 (1948): 1–29. 2. This claim is not without dissenters: “the unwillingness of [IPE's] foremost American advocates to seriously acknowledge that political economy as a rigorous discipline goes back to the opening decades of the seventeenth century (generally attributed to Montchretien in 1615), rather than beginning in 1968 with Richard Cooper, is confirmed when perusing most standard texts.” Peter Burnham, “Open Marxism and Vulgar International Political Economy,” Review of International Political Economy 1, no. 2 (1994): 221. 3. The organization of political science departments in other areas of the world, such as China or Russia, is less well-known. Whether they have established IPE positions and research centers would be an interesting question to address. 4. Roger Tooze, “Perspectives and Theory,” in Paths to IPE, ed. Susan Strange (London: Allen and Unwin, 1984), 7. 5. Susan Strange, “Wake Up Krasner, the World Has Changed,” Review of International Political Economy 1, no. 2 (1994): 218. 6. Louise Amoore, Richard Dodgson, Randall Germain, Barry Gills, Paul Langley, and Iain Watson, “Paths to a Historicized International Political Economy,” Review of International Political Economy 7, no. 1

(2000): 54. 7. One could of course argue, in terms dear to economists' hearts, that all phenomena are “economic” in that they involve costs and benefits and the operation of markets. 8. Robert Gilpin, U.S. Power and the Multinational Corporation (New York: Basic Books, 1975). 9. Benjamin Cohen, The Question of Imperialism (New York: Basic Books, 1973). 10. Susan Strange, Sterling and British Policy (London: Oxford University Press, 1971).Page 634 → 11. For example, Helen V. Milner, Resisting Protectionism (Princeton, N.J.: Princeton University Press, 1988); Ronald Rogowski, Commerce and Coalitions (Princeton, N.J.: Princeton University Press, 1989); Jeffry Frieden, “Invested Interests: The Politics of National Economic Policies in a World of Global Finance,” International Organization 45, no. 4 (1991): 425–51. 12. For example, Beth Simmons, Who Adjusts? (Princeton, N.J.: Princeton University Press, 1994); Stephen Haggard and Robert Kaufman, The Political Economy of Democratic Transitions (Princeton, N.J.: Princeton University Press, 1995); Sylvia Maxfield, Gatekeepers of Growth (Princeton, N.J.: Princeton University Press, 1997). 13. Susan Strange, “International Economics and International Relations: A Case of Mutual Neglect,” International Affairs 46, no. 1 (1970): 310. 14. For example, Peter Katzenstein, ed., Between Power and Plenty (Madison: University of Wisconsin Press, 1978); Peter Katzenstein, Small States in World Markets (Ithaca, N.Y.: Cornell University Press, 1985); Peter Gourevitch, Politics in Hard Times (Ithaca, N.Y.: Cornell University Press, 1986). 15. Ronald Rogowski, “Trade and the Variety of Democratic Institutions,” International Organization 41, no. 2 (1987): 203–24; Simmons, Who Adjusts? 16. For example, Stephen Haggard, Pathways from the Periphery (Ithaca, N.Y.: Cornell University Press, 1990); Haggard and Kaufman, The Political Economy of Democratic Transitions; Michael Shafer, Winners and Losers (Ithaca, N.Y.: Cornell University Press, 1994). 17. For example, Douglass North and Robert Thomas, The Rise of the Western World (New York: Cambridge University Press, 1973); Robert Gilpin, The Political Economy of International Relations (Princeton, N.J.: Princeton University Press, 1987); Immanuel M. Wallerstein, The Modern World System (New York: Academic Press, 1974); Charles Tilly, Coercion, Capital, and European States (Cambridge, England: Blackwell, 1990). 18. For example, Adam Przeworski, Democracy and the Market: Political and Economic Reforms in Eastern Europe and Latin America (New York: Cambridge University Press, 1991); Joel Hellman, “Winners Take All: The Politics of Partial Reform in Postcommunist Transitions,” World Politics 50, no. 2 (1998): 203–34. 19. For example, Katzenstein, Small States in World Markets; Simmons, Who Adjusts?; Robert Wade, Governing the Market (Princeton, N.J.: Princeton University Press, 1990); Geoffrey Garrett, Partisan Politics in the Global Economy (New York: Cambridge University Press, 1998). 20. For example, Robert O. Keohane and Joseph Nye Jr., Power and Interdependence (Boston: Little, Brown, 1977); Edward Morse, Modernization and the Transformation of International Relations (New York: Free Press, 1976); Robert O. Keohane and Helen Milner, eds., Internationalization and Domestic Politics (New York: Cambridge University Press, 1996); Garrett, Partisan Politics in the Global Economy. 21. For example, James Caporaso, ed., Dependence and Dependency in the Global System, special issue of International Organization 32, no. 1 (1978); Wallerstein, The Modern World System; Fernando Henrique Cardoso and Enzo Faletto, Dependency and Development in Latin America (Berkeley: University of California Press, 1979). 22. For example, Ernst Haas, The Uniting of Europe (Stanford, Calif.: Stanford University Press, 1958); Keohane and Nye, Power and Interdependence; Robert O. Keohane, After Hegemony (Princeton, N.J.: Princeton University Press, 1984); Stephen Krasner, ed., International Regimes (Ithaca, N.Y.: Cornell University Press, 1983). 23. For example, John Conybeare, Trade Wars (New York: Columbia University Press, 1987); Kenneth Oye, Economic Discrimination and Political Exchange (Princeton, N.J.: Princeton University Press, 1992); Barry Eichengreen, Golden Fetters: The Gold Standard and the Great Depression, 1919–1939 (New York: Oxford University Press, 1992); Simmons, Who Adjusts?; Marc Busch, Trade Warriors (New York: Cambridge University Press, 1999).Page 635 →

24. For example, Hirschman, National Power and the Structure of Foreign Trade; David Baldwin, Economic Statecraft (Princeton, N.J.: Princeton University Press, 1985); Lisa Martin, Coercive Cooperation (Princeton, N.J.: Princeton University Press, 1992); Dan Drezner, The Sanctions Paradox (New York: Cambridge University Press, 1999); Jonathan Kirschner, Currency and Coercion (Princeton, N.J.: Princeton University Press, 1995). 25. For example, Gilpin, U.S. Power and the Multinational Corporation; Gilpin, The Political Economy of International Relations; Paul Papayoanou, Power Ties: Economic Interdependence, Balancing, and War (Ann Arbor: University of Michigan Press, 1999). 26. For example, Stephen Krasner, “State Power and the Structure of International Trade,” World Politics 28, no. 3 (1976): 317–47; David Lake, Power, Protection, and Free Trade (Ithaca, N.Y.: Cornell University Press, 1988). 27. For example, John Conybeare, “Public Goods, Prisoners' Dilemma, and the International Political Economy,” International Studies Quarterly 28, no. 1 (1984): 5–22; Arthur Stein, “The Hegemon's Dilemma,” International Organization 38, no. 2 (1984): 355–86; Duncan Snidal, “The Limits of Hegemonic Stability Theory,” International Organization 39, no. 4 (1985): 579–614; Joanne Gowa, Allies, Adversaries, and International Trade (Princeton, N.J.: Princeton University Press, 1994); David Lake, “Leadership, Hegemony, and the International Economy,” International Studies Quarterly 37, no. 4 (1993): 459–89. 28. For example, Helen V. Milner, Interests, Institutions, and Information (Princeton, N.J.: Princeton University Press, 1997); Lisa Martin, Democratic Commitments (Princeton, N.J.: Princeton University Press, 2000). 29. For example, Alexander Hicks, Social Democracy and Welfare Capitalism (Ithaca, N.Y.: Cornell University Press, 1999); Torben Iversen, Contested Economic Institutions (New York: Cambridge University Press, 1999). 30. For example, Haggard and Kaufman, The Political Economy of Democratic Transitions; Dani Rodrik, “Democracies Pay Higher Wages,” Quarterly Journal of Economics 114, no. 3 (1999): 707–38; Edward Mansfield, Helen V. Milner, and B. Peter Rosendorff, “Free to Trade,” American Political Science Review 94, no. 2 (2000): 305–21. 31. For example, Lloyd Gruber, Ruling the World: Power Politics and the Rise of Supranational Institutions (Princeton, N.J.: Princeton University Press, 2000); B. Peter Rosendorff and Helen V. Milner, “The Optimal Design of International Institutions: Why Escape Clauses Are Essential,” International Organization (forthcoming). 32. For example, Matthew Gabel, Interests and Integration (Ann Arbor: University of Michigan Press, 1998); Geoffrey Garrett and George Tsebelis, “An Institutional Critique of Inter-Governmentalism,” International Organization 50, no. 2 (1996): 269–300. 33. For example, for some debates see Jon Elster, ed., Rational Choice (New York: New York University Press, 1986); James Coleman and Thomas Fararo, eds., Rational Choice Theory: Advocacy and Critique (Newbury Park, Calif.: Sage, 1992); Donald Green and Ian Shapiro, Pathologies of Rational Choice Theory (New Haven, Conn.: Yale University Press, 1994). 34. For example, Stephen Walt, “Rigor or Rigor Mortis?” International Security 23, no. 3 (1999): 5–48. 35. Helen V. Milner, “The Analysis of International Relations: Formal Models in International Political Economy,” in The Analysis of International Relations, ed. Detlef Sprinz and Yael Wolinsky (Ann Arbor: University of Michigan Press, 2001). 36. James Alt, Margaret Levi, and Elinor Ostrom, eds., Competition and Cooperation (New York: Russell Sage Foundation, 1999), xv. 37. Daniel Kahneman and Amos Tversky, “Prospect Theory,” Econometrica 47, no. 2 (1979): 263–91. 38. Ole Waever, “The Sociology of a Not So International Discipline,” International Organization 52, no. 4 (1998): 687–728.Page 636 → 39. For example, Kyle Bagwell and Robert Staiger, Domestic Policies, National Sovereignty, and International Economic Institutions, NBER Working Paper Series, no. 7293 (Cambridge, Mass.: National Bureau of Economic Research, 1999); Giovanni Maggi, “The Role of Multilateral Institutions in International Trade Cooperation,” American Economic Review 89, no. 1 (1999): 190–214; Torsten Persson and Guido Tabellini, Political Economics (Cambridge: MIT Press, 2000).

Page 637 →

SOME THOUGHTS ON INTERNATIONAL POLITICAL ECONOMY IN THE CONTEXT OF PUBLIC POLICY EDUCATION Robert T. Kudrle In the early 1980s I tried to develop an approach to international economic relations for Humphrey Institute public policy students that would combine parsimony with comprehensiveness; I was also serving on the editorial team of International Studies Quarterly. Both activities led me to think more systematically about the connections between economics and politics than I had done before. I saw no advantage in abandoning the basic theory that seemed reasonably successful in my own discipline of economics, and I had to consider the students. Public policy students with an interest in international affairs care little about the disciplinary origin of a satisfying approach, still less whether insights come from American government, comparative politics, international politics, foreign policy, or public policy. Here economics holds a distinct—some would argue, meretricious—advantage: microeconomic subfields employ the same tools and a common language, and our students must learn microeconomics for policy analysis. My natural starting point was “political economy” as William Riker rather narrowly defines it: the application of microeconomics to political phenomena.1 In other words, instead of using two completely different approaches to politics and markets, insofar as possible, the same approach would be used. Elsewhere in this volume, Helen Milner has noted two other common uses of the term when preceded by “international,” ones that focus on subject matter rather than method. The broadest usage sweeps up virtually all of international relations not connected with military force; the one she prefers is restricted by economic considerations as dependent or independent variables.2 By stressing method rather than substance, I sidestep that dispute. As a practical matter, however, the greatest traction from the approach outlined here comes where economic variables loom large; I have applied it mainly to explore the sources of foreign economic policy. None of the prevailing theories of international policy at that time was grounded in the methodological individualism upon which virtually all of economics is built. Instead, there was a consistent attempt to link foreign economic policy to broad prevailing international relations approaches, almost Page 638 → invariably Marxism, liberalism, and realism.3 With Davis Bobrow, I sketched the barest outline of what an analysis of foreign economic policy might look like when approached from a standpoint based on microeconomics.4

The Basic Approach The economist confronting political phenomena asks what actors want and what incentives they have to act collectively. Most public policy products (law and administration) are either pure (e.g., defense) or impure (e.g., law enforcement) public goods based on the criteria of nonexhaustibility and nonexcludability. This starting point naturally treats foreign policy as a species of public policy rather than as something utterly different. In terms of linkage to the political science literature, this implies immediate attention to comparative politics and policy. In international affairs, citizens appear concerned about public policies that serve certain fundamental goals that must be sought as collective goods. These goals may be regarded as guiding individuals' policy preferences in the absence of the special personal stakes that drive domestic policy preferences in the simplest public choice models. The inductively identified goals should not be capable of further reduction or summarization and they should be universal among modern states. The most fundamental appear to be autonomy, security, and prosperity. Autonomy deserves first consideration because some desire to be mâitre chez nous lies at the heart of an internally stable state. A desire for autonomy is very closely akin to common definitions of nationalism.5 At the nation-state level ambitions may vary widely from, at minimum, a desire to maintain a sense of national distinction to, at maximum, an attempt to avoid any substantial foreign influence.6 Protection from foreign violence is sought within that context. A state's security challenge can really only be considered in the context of autonomy because various approaches to security pose varying autonomy costs. The weighing of such costs depends on how the

group or dominant elements in it view other nation-states. Prosperity can be considered simply the maximization of national wealth because modern nation-states command redistribution policies.7 A fourth goal, standing, appears to be less universally important, yet it clearly drives much foreign policy behavior, especially by large states. Standing refers to international rank and influence upon, or regard from, those in other states.8 Some might question the sufficiency of the goals. Concern about the environment is included with prosperity because the high-income countries, at least so far, value the natural world primarily as contributing to income, broadly construed, rather than for its own sake. Others might assert that just as altruism plays an important role within states, an international counterpart should be recognized. Foreign assistance as a percentage of national product in most countries declined in the 1990s from an already small base, however.9 Page 639 → By making multiple goals explicit, the investigator is forced to consider how governments pursue trade-offs and synergies among them. Fungible resource allocation in the service of nonprosperity goals may be paid as an ordinary price in reduced consumption of other goods and services, but depending on the circumstances, some policies directly barter contributions to one goal for diminished service to another, while other policies promise positive payoffs for more than one goal. The broad outlines of difference in the foreign economic policies of a number of high-income countries seemed satisfactorily interpretable in these terms during the cold war.10 The emphasis in economics on maximization and efficiency focuses attention on the opportunity costs involved in the pursuit of national goals. It also forces a clear distinction between preferences and constraints.11 The relation between them involves consideration of (temporary, moving) equilibria—however imperfectly achieved from some national interest standpoint—and predicts changed policy in response to shifts in either the effective preferences or the constraints. Constraints include the structure of the international system and the place of the state in it, the levels of income and technology of system members, and the degree of congruence of purpose in various dimensions of the state in question with that of other states. Unpacking the latter involves consideration of strategic interaction and bargaining for large states (“price makers”) and adjustment for others (“price takers”). Cartel theory suggests hypotheses about the importance of size distribution of states for stability.12 One can envision a rational benevolent dictatrix considering her utility function with its several national goals and pursuing an optimal policy by bringing multidimensional indifference and (not necessarily everywhere concave) production surfaces together to produce an optimum. Such a seemingly absurd abstraction can be valuable if it contributes to a clearer understanding of the slippage between various constructions of the “national interest” and existing policy. This highlights not just the role of special interests but also broader problems of the gathering and processing of information and the aggregation of preferences. It also raises the traditional concern in economics for efficiency. How and why are changes in internal and external conditions failing to register in policy change? How much of the observed situation can be explained by bargaining or strategic interaction at two or more governance levels and by previous commitments? Most fundamentally, how and why are preference aggregation and choice mechanisms skewing policy away from the apparent desires of the typical citizen? Comparing actual foreign economic policy with some vision of ideal policy serving some definition of the “public interest” requires a conception of the latter, and that implies judgments, not just about political economy and the traditional autonomy and security concerns of political science, but about economic issues per se. It also demands some guesses about the distribution of preferences because the valuation of autonomy, standing, and security cannot be directly observed. A comprehensive public policy curriculum for IPE must address all of these subjects. Page 640 → Rational ignorance also plays a central role in the positive theory of policy determination from this perspective.13 The private good of information processing and evaluation by the individual compared with its perceived value presents a formidable barrier to intelligent choice, especially where the transparency of a policy's impact is low. In

such circumstances political leadership channeled by existing domestic structures plays a particularly important role in determining policy stasis and change.14 Interests tend to loom in intensity with stakes relative to agent numbers, as suggested in the broad-brush collective action literature. A satisfactory economic approach must carefully consider both collective action and specific mechanisms of collective choice.15 Until quite recently, economists paid little attention to the origin of preferences. Economic historians, however, cannot avoid the issue, and Douglass North sketched the rudiments of a theory of ideology and ideological change congenial to the economic model.16 Such ideology includes both a set of cause-and-effect beliefs and some standards of fairness. It serves to overcome otherwise rational antisocial behavior (free riding), and the state's enormous investment in legitimacy reflects its importance. Individuals' ideologies change when experience becomes sufficiently inconsistent with the previously accepted ideas. The same logic makes institutional change (in the sense of formal structures of governance) viscously endogenous.

Recent Research in IPE I applied this approach in writing only informally and intermittently over the years after it was first developed.17 Much international relations scholarship of the past fifteen years or so, however, appears highly congenial to it. The role of policy ideas and their diffusion across time and space has been explored as never before.18 The research agendas in liberal institutionalism emphasize rationalism in a way readily-recognized and appreciated by economists.19 Recent writing that links domestic and international politics often employs creative models that economists understand and admire.20 And whatever the mix of qualitative and quantitative analysis, the newer writing typically pays great attention to formal scientific argument.21 Game theory changed the landscape in political science because it was formal and in economics because it was formally strategic.22 Both disciplines now acknowledge the importance of the subject, but there is intense disagreement in both about its value relative to other research activities.23 Game theory plays a significant but modest role in the curriculum of public policy schools, and this is likely a stable equilibrium for the foreseeable future. In a tightly packed curriculum, game theory's payoff appears mainly heuristic. The professional value of training beyond first principles remains to be demonstrated. One very simple application of game theory links domestic and international policy-making by construing decision makers as simultaneously involved in “games” at both levels at the same time. As Putnam so memorably puts it, “Any key player at the international table who is dissatisfied with Page 641 → the outcome may upset the gameboard and conversely, any leader who fails to satisfy his fellow players at the domestic table risks being evicted from his seat.”24 This approach provides an important approach for the development of explicit, high-level international agreements with large attentive domestic constituencies. Yet other important agreements take a very different form, as Slaughter has stressed.25 Large parts of international economic life are now governed by “policy networks” that develop formulas for international cooperation on a broad range of issues based on common experience or best-practice technique, completing bypassing negotiations at levels above the most directly affected parties and their immediate regulators. Policy networks suggest a mode of international governance far removed from the world of realism (despite the typically greater influence of those with the highest stakes and greatest experience). More generally, the economist considering IPE will likely conclude that realism's focus on defending the primacy of the security goal leaves an ever-greater part of the policy space for persuasive consideration by others. Economists' instincts admittedly push toward excessive attention to prosperity issues, but opportunity cost is always on their minds. They will therefore always want to know when, how, and why security concerns are traded off against other goals. They will also want to know when a quest for relative economic gain can be clearly distinguished from a desire for a greater share of an absolute gain, the policy preferences of special interests, or sector-specific security concerns. As has been widely noted, the increased attention to economic variables and methods by political scientists has taken place at the same time that economists have begun routinely addressing political issues.26 Explicitly

interdisciplinary work has sometimes produced impressive results. A notable contribution to IPE includes a sophisticated survey of factor specificity as a source of protectionist politics.27 The findings stress answers that turn on the structure of the economy but, even more importantly, the structure of the political system. Many articles in the major economics journals now focus on arenas other than markets. Although much of this work is presented even less accessibly than most models in political science, some of it can be easily explained. An example is the tenacity of trade protectionism, which is a highly-inefficient societal transfer. Fernandez and Rodrik offer a persuasive explanation in some circumstances: despite confident estimates of aggregate gains, even by dominant groups, uncertainty about the identity of gainers can prevent reform.28 Another important contribution is the recent revisionist account of reciprocity in the General Agreement on Tariffs and Trade provided by Bagwell and Staiger.29 In sharp contrast to the emphasis on mercantilist impulses stressed by most commentators, this interpretation develops the possibility that reciprocity is embraced to avoid the terms of trade deterioration that uncoordinated liberalization can produce.30 Marxism has never been much appreciated by economists on theoretical or empirical grounds, and most recent critical theory also appears uncongenial Page 642 → (as is its intention). Much of constructivism, however, which has enlisted some fine scholars from the Marxist tradition, appears to complement the rationalist scholarship or to contest it in promising ways (as argued by some of its leading practitioners).31 Moreover, some of the same empirical territory is now being addressed from viewpoints grounded in economics by directly including the views and behavior of others in utility functions.32 However innovative the latter may be, the results will doubtless appear meager to those from other traditions. The parsimony that economists prize can survive only limited contextualization.33 As in the past, problems of affect and cognition will be reinterpreted, ignored, or minimized. Persuasion, for example, will typically be treated as information rather than arousal. The economic approach does not deny some importance for factors not explicitly considered but simply asks how much is lost from their exclusion. For example, constructivists claim that changes in underlying preferences, as opposed to strategies in their pursuit, cannot be endogenized.34 The ratio of the latter to the former may be higher than it appears, however. One example can illustrate the claim. Current changes in thinking about organized labor in many countries reflect concern that prevailing institutions limit general prosperity by reducing the adaptability of the economy as well as its level of employment and output. Policy changes that would increase efficiency by lodging benefits directly with individuals rather than through the workplace—if those changes were bolstered by other policy modifications that preserved social minima and the overall level of material equality—would satisfy those who saw the historic norms of “the rights of labor” as instruments of those objectives, while leaving others highly dissatisfied. More generally, the economic approach will track international norm transmission largely as a combination of information, experience, rational ignorance, and the concentration and diffusion of stakes—whether the latter be material or ideal. No review of recent scholarship relevant to the approach outlined at the beginning of this paper can ignore the burgeoning rational choice literature in comparative politics. Although it treats issues broader than IPE, the work is “based on microfoundations, in which the collective outcome is derived, albeit often perversely, from choices made by individuals.”35 While much of this work focuses on purely domestic political developments, Bruno Frey has developed an application of public choice that traces individual aspirations within international organizations.36 Another large body of writing relevant to the approach outlined here grows from the work of Oliver Williamson and the “New Institutional Economics,” which is characterized by the twin claims that institutions matter and that they can be analyzed by the tools of economic theory.37 Williamson has developed a four-level scheme for analyzing social institutions that embed political economy phenomena in constraints with varying adjustment periods.38 Much of the research discussed in this article can be usefully considered in this framework. Most fundamental are customs, traditions, norms, and religion that have historically had alteration periods that vary between one hundred and one thousand years. Much of the constructivist Page 643 → project and parallel work on norms by lawyers and economists explore this first level. Next are the broadest institutional rules of society in which change usually takes place between a decade and a century. Positive political theory in both political

science and the law39 can address issues at this level of analysis, which includes the functioning of both domestic and international institutions. The last two levels deal with most of the phenomena of economics. The third level deals with private governance structures that adapt over a decade or less and includes developments such as integration and new forms of contracting. Some or all of the development of publicly sanctioned international policy networks belongs here too. Fourth and finally, there is the continuous operation of markets. Those employing rational choice in political science and the law sometimes have difficulty mastering phenomena in the last two of Williamson's categories, while naive economic approaches to public policy determination often ignore institutional variation and sometimes never move beyond simple analogies suggested by the last two levels alone. Recent research by Irwin and Kroszner employs several elements of the scheme sketched at the beginning of this paper to analyze a major reorientation of American foreign economic policy.40 After identifying private interests, ideology, and institutions as the principal contending (and obviously potentially-complementary) approaches to policy change, the authors consider each to interpret the growth of Republican support for the Reciprocal Trade Agreements Act during the 1940s.41 While Irwin and Kroszner focus on the broad determinants of policy, Bates, De Figueiredo, and Weingast employ methodological individualism and goals similar to those noted at the beginning of this paper to analyze popular Serbian support for Milosevic during the 1990s.42 Although security trumps other concerns in a way that is familiar and reassuring for realists, explicit trade-offs among security, autonomy, and prosperity at the level of the citizen are not. In this study, rationality is attributed to behavior that is treated elsewhere as atavism.

IR and IPE as Social Science Eichengreen suggests that there has not been enough “systematic, standardized” use of data with theory in the international relations literature by comparison with economics and that case studies are too idiosyncratic to cumulate to greatest advantage.43 But consider the relative difficulties. Most standard scholarship in microeconomics deals with fairly straightforward measures as dependent, independent, and endogenous variables. In sharp contrast, broader social phenomena are fraught with complex and often invisible connections. As Olson has stressed, indivisibilities in the consumption of nonmarketed goods sharply separate available measures from changes registered in utility functions.44 Moreover, governments typically produce goods that are either nonexclusive or have large economies of scale in production. In both cases, the output eludes satisfactory measurement, is produced with un-known Page 644 → levels of efficiency, and is consumed with a valuation that can only be roughly estimated. These problems would remain even if organized interests did not participate in the political process and voters with single-peaked preferences decided issues in referenda. Such considerations imply that the sometimes clear boundaries between markets that justify exploration with partial equilibrium analysis in economics frequently blur when political phenomena are confronted. And even when a political issue appears neatly-bounded, the number of potentially-relevant variables about which there is adequate information, and especially appropriate measurement, is usually small. There is frequent resort to proxy. The number of policy environments, typically states, and the dangers of uncontrolled but perhaps critical changes in the relevant environment over time sharply limit the number of data points in either cross sections or time series (usable panel data are sometimes available). Light specification invites a poor fit, omitted variable bias, or both, while more complete specification increases multi-collinearity and spends usually meager degrees of freedom. Causation is often contested, and the data seldom permit satisfactory tests for endogeneity.45 Conceptual, measurement, and data considerations mean that largely-qualitative analysis remains essential in careful political analysis, including IPE, and it will necessarily play a relatively larger role there than in economics for the indefinite future. There is also the ever-present danger of a loss of important elements of a conceptual approach through premature or excessive formalization.46 This does not suggest that a qualitative research project need be less scientific; it may, however, be less conclusive. I say may because, when political and economic research are evaluated as description, explanation, and prediction on major issues, the marked superiority of

economics can often be questioned.47

Looking to the Future The previous discussion treated the nation-state as a given and assumed the collective pursuit of all goals through its agency. This considerably oversimplifies the past and provides a poor guide to future politics and policy. The economic approach to policy expects changing behavior in responses to changing constraints. Self-conscious groups not coterminous with the boundaries of the nation-state have always produced internal autonomy tensions. The reductions of external security threats in many parts of the world have doubtless eased central government fears about responding with accommodation. On the other hand, modern technology has generated economies of scale and learning effects that have pushed the optimum market size ever upwards for much economic activity. Hence the quest for prosperity propels autonomy-compromising coordination—if not heavier governance—as seen in the World Trade Organization as well as myriad economic integration schemes.48 Some elements of autonomy may be strengthened as others are ceded in regional ventures. This can be seen as a trade of less “internal” for more “external” autonomy in a multistate effort.49 For example, Page 645 → the European Union (EU) certainly exercises more bargaining power in the world than could any of its nation-states acting alone. Europe also provides an example of identification-pooling for the pursuit of standing.50 Europe's rank in the many dimensions of human activity was widely recognized to exceed what any single state could achieve alone. Finally, alliances and collective security arrangements have always implied a trade-off of autonomy for security. Although I have expressed skepticism about the severity of globalization's current threats to central areas of national policy such as growth and taxation,51 globalization will increasingly affect attitudes and identities relevant to the fundamental goals. This will complicate spheres of autonomy demand as well as the group identification for which the psychic income of standing is relevant. The focus of security efforts will migrate upward. Domestic support for altruistic policy beyond the state's borders will likely expand. These and other changes will likely lead to a sharing of policymaking authority by the Westphalian state with both “lower” and “higher” levels of governance. Shifts in political power among levels of government are well-advanced in Europe, but every serious integration initiative forces a rethinking of the relation of the goals with respect to types of policy and levels of governance. More generally, the forces of globalization challenge the linkages between the goals and the governance structures that have served them in the past. I also argue that the direction and speed of integration, with its melding of foreign and domestic public policy, cannot be understood without attention to trade-offs as they are perceived by individual citizens. This suggests the seminal importance of the approach taken to Serbian politics by Bates, De Figueiredo, and Weingast.52 Parallel analysis appears well-suited to individuals facing political choices in less dramatic circumstances. Europeans are currently experiencing a double dose of economic and social change as globalization and European integration interact with each other. But the two sets of changes should not be lumped together. Europe is engaged in an experiment with true economic integration based on a special history containing elements of both common identity and hideous conflict. This has legitimized a centralizing political counterpart to European economic integration that lacks a parallel in most of the rest of the world. Dani Rodrik has recently provided a schematic approach to the development of the global economy that can be used to illustrate the qualitatively distinct characteristics of the European project, although that was not his purpose.53 Rodrik claims that we now face a political economy “trilemma” of the nation-state, mass politics, and integrated national economies.54 By the third term, Rodrik means the kind of integration in which Europe is currently engaged: goods, services, and all factors of production, especially labor. He claims that one of three choices must be made. The state must severely limit its policy activities so that much authority is focused above, where mass politics can deal with it directly. Alternatively, mass politics must be prevented from effectively addressing much of what really matters—one way of casting the “democratic deficit.” If neither of the first two Page 646 → paths is chosen, the gains of complete integration must be sacrificed. The last point needs a careful look. Outside of Europe, one sees little popular interest in those gains. Rodrik cites John Helliwell's research that

the U.S.- Canadian border reduces economic efficiency much more than might be guessed from a look at tariff rates or residual investment controls after the North American Free Trade Agreement (NAFTA).55 As Rodrik suggests, much of this inefficiency doubtless stems from the difficulty of international contract enforcement and could be overcome only with greater legal centralization. But the example also suggests why the trilemma may be more apparent than real outside of Europe. There is no significant Canadian opinion favoring joint governance with the United States to gain that last quantum of economic efficiency—however surprisingly large it is alleged to be (and the Canadians are paying most of the inefficiency price). The autonomy cost of increased prosperity is so high as to block consideration of significantly stronger common governance.56 On the other hand, the trilemma does well-characterize the current European situation, created as it has been by both economic and noneconomic forces. Its challenges offer cautionary lessons to the rest of the world. Rodrik's model treats economic integration as a multidimensional phenomenon with no apparent qualitative breakpoints. It can be considered from another perspective, one that I think highlights how goal conflicts will limit the embrace of full international integration. I have argued elsewhere that the importance of the state can be largely gauged in one major dimension: the extent to which it can control its membership and can effectively claim tax authority over their income and wealth.57 To the extent that the state cannot substantially control permanent population movements across its borders and that it cannot tax the income and wealth of its residents, either because of lack of transaction transparency or jurisdictional mobility, the state withers. If unskilled immigration cannot be controlled, state redistributional efforts can be neutralized. In addition, if the state cannot effectively tax human and nonhuman capital, its redistributive capacity drops sharply. Both possibilities threaten a critical element of national sovereignty: the ability to shape the state's after-tax distribution of income. The long-run impact of policy innovations can often scarcely be described. The current level of European economic integration appears to lag very far behind the legal structure that enables it. The rest of the world should watch with special care how Europe deals with the closely-linked issues of internal migration and fiscal policy. Cultural differences continue greatly to inhibit labor mobility within Europe. Yet other factors are also at play. Although many of the most successful politicians of the left as well as the right claim allegiance to markets, some unknown part of current European labor immobility must be assigned to the restricted role of labor and housing markets. If market forces in these dimensions become much more important, interstate mobility could accelerate. The almost certain accession of more low-income countries will put some downward pressure on earnings, particularly of the unskilled, even if the populations do not move. That tendency would be considerably magnified Page 647 → by migration because the nontradable sector would be directly affected. Forecasting here begins to take on the character of science fiction. Nevertheless, the U.S. experience suggests a very impressive example of path dependence: the remarkable potential of chain migration to cause rapid influx of culturally-alien newcomers from minor initial causes.58 Such developments could raise both prosperity and autonomy threats. And it appears that any control over migration from new states into the EU will be limited to a fixed period after accession. There is no provision in EU doctrine or policy for the migration equivalent of the “escape clause” in trade law. Fiscal tensions will loom much larger in Europe's future as well, and the present level of decentralization lacks long-run viability. The European tax base remains overwhelmingly national, and some member-states have never been effective at direct tax collection. The challenge has now been greatly increased. EU citizens can now easily shift themselves and their tax liabilities from one jurisdiction to another. And if they don't actually move, tax havens can hide their income anyway. Even if current tax transparency problems were solved,59 some possible dynamics can alarm. Imagine a future attempt by a high-income EU nation-state to ameliorate an income distribution or an unemployment problem born of unfavorable technological change, the departure of industry for cheaper labor, or unskilled immigration. It might try to raise tax rates on local “winners”—those who own high amounts of human and nonhuman capital. But the “winners” can legally abandon that state, thus exacerbating its problems. Of course, cultural viscosity prevails now, but how strong will it remain over time? Only greater tax coordination within the union—so far resisted in the name of state autonomy—appears as a solution. Regional integration schemes outside of Europe have a very different character. They have been built almost entirely on prosperity aspirations although external autonomy and security advantages have sometimes been

claimed. Political integration of any kind has usually been rejected from the outset, and the uncontrolled movement of persons across borders has not been embraced. NAFTA, for example, specifically eschewed political innovation and was sold to the American public partly as a means to prevent unwanted immigration, not to open labor markets.60 The motivations for the more than thirty regional trading blocs that sprang up in the 1990s were doubtless quite varied, but the aspirations of citizens and leaders alike were certainly parallel to those underlying NAFTA more than those of the EU. Enlarging goods and services markets appeared to increase a state's commitment to a stable environment for business and hence to increase attractiveness to direct investors (although the level of binding commitment that has been achieved so far in most such arrangements has been modest). Citizens in most such schemes almost certainly prefer increased prosperity at the lowest possible price in autonomy foregone. In no case does a powerful parallel political agenda propel continuing integration, as was the case in Europe. Therefore, if we assume attentive constituents, integration proponents will succeed as they can argue for efficient trade-offs among the goals. This emphasis on opportunity cost inherent Page 648 → in the economic approach, as well as its emphasis on maximization and equilibrium, may grow in value for IPE scholarship. Nearly all international economic interaction and the policies that condition it can be considered in three categories: exchange, penetration, and constraint. Economists and political scientists have studied the first category mainly by focusing on the way by which trade—in both goods and capital—changes the wealth and security positions of various states. This has been the focus of most of international economics and much of international politics and political economy. In some of the formulations of research reviewed earlier, the interactions have no effect beyond changes in wealth positions. In sharp contrast, penetration issues stress a broader range of interactions that are epitomized by the literature on the multinational corporation. The social impact of a permanent foreign presence generates issues well beyond those of mere commerce, and much of that presence has been regarded with ambivalence or hostility. Finally, many of the concerns associated with “globalization” involve international factors mainly as they constrain the operation of domestic policy. Concerns about the changing incidence of certain taxes in response to increased capital and labor mobility illustrate those issues. The broader agenda reduces the overall importance of increased scholarly understanding of the domestic politics of purely exchange issues and increases the need to examine much more of the comparative politics and public policy literature in political science. For the same reasons, additional subfields of economics are rising in importance for the study of the policy aspects of IPE. Public finance, labor economics, and environmental economics are growing in importance as international taxation, immigration, and the global environment have joined trade, investment, and finance issues as foci of major attention. Nothing in the approach sketched above should be interpreted as denigrating insights from approaches to IPE not considered in this paper. I have merely reported on a framework devised largely for the non-political scientist and nonspecialist, whose interest is typically in policy analysis and who already knows the fundamentals of microeconomics. It aims to order existing knowledge and to probe its limits. I am encouraged that the approach appears increasingly consistent with much of the best social science scholarship.

Notes 1. William Riker, “Political Science and Rational Choice,” in Perspectives on Positive Political Economy, ed. James E. Alt and Kenneth Shepsle (New York: Cambridge University Press, 1990), 180–81. 2. As Milner undoubtedly recognizes, there is innate ambiguity in her sensible approach. For example, she appears to exclude environmental and human rights issues. But the externalities from general economic production and their amelioration look like IPE to me, especially given the enormous domestic and international distributional stakes involved. Similarly, while all of international human rights may not be IPE, rights assigned to labor, including child labor, seem to me just as inseparable from the economic

system as most externalities. Northern solicitude is widely seen in the South as largely motivated by protectionism.Page 649 → 3. Robert Gilpin, U.S. Power and the Multinational Corporation (New York: Basic Books, 1975); Stephen D. Krasner, Defending the National Interest: Raw Materials, Investments, and U.S. Foreign Policy (Princeton, N.J.: Princeton University Press, 1978). In her essay in this volume, Lisa Martin notes the increasing rejection of this overall framework in recent years. Marxism may have faded as a compelling approach as both its global political importance and its general academic attraction diminished, the concerns of domestic pluralism and international liberalism became increasingly suffused with rational choice, and realism's preoccupation with security reduced its apparent relevance. 4. Robert T. Kudrle and Davis B. Bobrow, “U.S. Policy toward Direct Foreign Investment” World Politics 34, no. 3 (1982): 353–79. 5. A nation is defined by Yael Tamir as “a community whose members share feelings of fraternity, substantial distinctiveness, and exclusivity…. They thus seek to secure for themselves a public sphere where they can express their identity, practice their culture, and educate their young.” Yael Tamir, “The Enigma of Nationalism,” World Politics 47, no. 3 (1995): 425. 6. A typical citizen's physical safety can often be assumed after capitulation to a foreign power, and such defeat need not lead to a decline in the material standard of living. 7. As a practical matter, however, states have not consistently pursued wealth maximization even when such behavior cannot be rationalized as service to the other goals. In contemporary policy this can typically be construed as a distortion impelled by special interests. Some current and much past policy, however, has been guided by a faulty understanding of the efficacy of various measures for generating greater wealth. 8. Autonomy and standing are thus quite distinct; the former focuses on a group's view of its own situation, independent of the views of others. 9. Even in the 1980s only Norway consistently made an aid transfer of more than 1 percent of its national product. 10. Davis B. Bobrow and Robert T. Kudrle, “Mid-Level Power Strategies for Changing International Niches: Experience in the Old World Order,” Journal of East Asian Affairs 5, no. 2 (1991): 237–70. For example, Finland paid a clear autonomy price with the Soviet Union for security gains during the cold war. The United States traded prosperity for security during that period, while Japan appears to have pursued policies resulting in synergy between them at modest cost in autonomy. 11. Edward P. Lazear, Economic Imperialism, NBER Working Papers Series, no. 7300 (Cambridge, Mass.: National Bureau of Economic Research, 1999). 12. See Duncan Snidal, “The Limits of Hegemonic Stability Theory,” International Organization 39, no. 4 (1985): 579–614; Joanne S. Gowa, “Rational Hegemons, Excludable Goods, and Small Groups: An Epitaph for Hegemonic Stability Theory?” World Politics 41, no. 3 (1989): 307–24. 13. Anthony Downs, An Economic Theory of Democracy (New York: Harper and Row, 1957); Mancur Olson, The Logic of Collective Action: Public Goods and the Theory of Groups (Cambridge: Harvard University Press, 1965). 14. Peter J. Katzenstein, “International Relations and Domestic Structures: Foreign Economic Policies of Advanced Industrial States,” International Organization 30, no. 1 (1976): 1–45. 15. See Robert H. Bates, “Comparative Politics and Rational Choice: A Review Essay,” American Political Science Review 91, no. 3 (1997): 703. 16. Douglass C. North, Structure and Change in Economic History (New York: Norton, 1981). 17. Robert T. Kudrle, “Good for the Gander? Foreign Direct Investment in the United States,” International Organization 45, no. 3 (1991): 399–424; Robert T. Kudrle, “Markets, Governments, and Policy Congruence across the Atlantic,” in Balancing State Intervention: The Limits of Transatlantic Markets, a Rand Study, ed. Roger Benjamin, Richard Neu, and Denise Quigley (New York: St. Martin's, 1995), 101–43; Robert T. Kudrle and Stefanie Ann Lenway, “Progress for the Rich: The U.S.-Canada Free Trade Agreement,” in Progress in International Page 650 → Relations, ed. Emanuel Adler and Beverly Crawford (New York: Columbia University Press, 1991). 18. For example, Judith Goldstein and Robert O. Keohane, eds., Ideas and Foreign Policy: Beliefs, Institutions, and Political Change (Ithaca, N.Y.: Cornell University Press, 1993). 19. For example, Lisa L. Martin, Coercive Cooperation: Explaining Multilateral Economic Sanctions

(Princeton, N.J.: Princeton University Press, 1992). 20. For example, Helen V. Milner, “Industries, Governments, and the Creation of Regional Trading Blocs,” in The Political Economy of Regionalism, eds. Edward D. Mansfield and Helen V. Milner (New York: Columbia University Press, 1997), 77–106; Walter Mattli, The Logic of Regional Integration: Europe and Beyond (Cambridge: Cambridge University Press, 1999). 21. This concern is treated in Gary King, Robert O. Keohane, and Sidney Verba, Designing Social Inquiry (Princeton, N.J.: Princeton University Press, 1994). 22. For a fine introduction to the former see James D. Morrow, Game Theory for Political Scientists (Princeton, N.J.: Princeton University Press, 1994). 23. See Sam Peltzman, “The Handbook of Industrial Organization: A Review Article,” Journal of Political Economy 99, no. 1 (1991): 201–17; Stephen M. Walt, “Rigor or Rigor Mortis? Rational Choice and Security Studies,” International Security 23, no. 4 (1997): 5–48. 24. Robert D. Putnam, “Diplomacy and Domestic Politics: The Logic of Two-Level Games,” International Organization 42, no. 3 (1988): 434. 25. Anne-Marie Slaughter, “Governing the Global Economy through Government Networks,” in The Role of Law in International Politics: Essays in International Relations and International Law, ed. Michael Byers (Oxford: Oxford University Press, 2000). 26. See Helen Milner, ch. 42 in this volume. 27. James E. Alt, Jeffry Frieden, Michael J. Gilligan, Dani Rodrik, and Ronald Rogowski, “The Political Economy of International Trade: Enduring Puzzles and an Agenda for Inquiry,” Comparative Political Studies 29, no. 6 (1996): 689–717. This article extends James E. Alt and Michael Gilligan, “The Political Economy of Trading States: Factor Specificity, Collective Action Problems, and Domestic Political Institutions,” Journal of Political Philosophy 2, no. 2 (1994): 165–92. 28. Raquel Fernandez and Dani Rodrik, “Resistance to Reform: Status Quo Bias in the Presence of Individual-Specific Uncertainty,” American Economic Review 81, no. 5 (1991): 1146–55. 29. Kyle Bagwell and Robert W. Staiger, “An Economic Theory of GATT,” American Economic Review 89, no. 1 (2000): 215–48. 30. The fact that this rationale is broadly consistent with the operation of the GATT, of course, does not itself overturn the traditional explanation. 31. See Martha Finnemore and Kathryn Sikkink, “International Norm Dynamics and Political Change,” International Organization 52, no. 4 (1998): 887–917. 32. For example, B. Douglas Bernheim, “A Theory of Conformity,” Journal of Political Economy 102, no. 5 (1994): 841–77; Gary S. Becker and Kevin M. Murphy, Social Markets: The Marriage of “Economic” and “Social” Forces (Cambridge: Harvard University Press, 2000). 33. Lawrence Lessig, “Social Meaning and Social Norms,” University of Pennsylvania Law Review 144 (1996): 2181–87. 34. Finnemore and Sikkink, “International Norm Dynamics and Political Change,” 914. 35. Bates, “Comparative Politics and Rational Choice,” 702. 36. Bruno S. Frey, “The Public Choice of International Organizations,” in Perspectives on Public Choice: A Handbook, ed. Dennis C. Mueller (New York: Cambridge University Press, 1997). 37. See Oliver E. Williamson, Markets and Hierarchies: Analysis and Antitrust Implications (New York: Free Press, 1975); Oliver E. Williamson, The Economic Institutions of Capitalism (New York: Free Press, 1985); Oliver E. Williamson, The Mechanisms of Governance (New York: Oxford University Press, 1996). Page 651 → The quotation is from R. C. O. Matthews, “The Economics of Institutions and the Sources of Economic Growth,” Economic Journal 96, no. 4 (1986): 6. See also Oliver E. Williamson, “The New Institutional Economics: Taking Stock, Looking Ahead,” Journal of Economic Literature 38 (September 2000): 595–613. A “new” institutionalism can be found in a variety of other social science disciplines beyond economics and political science. For an extended discussion, see Robert E. Goodin, ed., The Theory of Institutional Design (Cambridge: Cambridge University Press, 1998), ch. 1. 38. Williamson, “The New Institutional Economics,” 597. 39. For the latter see Daniel A. Farber and Phillip P. Frickey, “Positive Political Theory in the Nineties,” Georgetown Law Journal 80 (1992): 452–76. 40. Douglas Irwin and Randall Kroszner, “Interests, Institutions, and Ideology in Securing Policy Change:

The Republican Conversion to Trade Liberalization after Smoot-Hawley,” Journal of Law and Economics 42, no. 2 (1999): 643–73. 41. It concludes that the interactions of institutional change with changing interests explain shifts in the Republican position, while ideologically-related explanations do not. 42. Robert H. Bates, Rui J. P. De Figueiredo Jr., and Barry R. Weingast, “The Politics of Interpretation: Rationality, Culture, and Transition,” Politics and Society 26, no. 2 (1998): 221–56. 43. Barry Eichengreen, “Dental Hygiene and Nuclear War: How International Relations Looks from Economics,” International Organization 52, no. 4 (1998): 1012. 44. Mancur Olson, “Toward a Unified View of Economics and the Other Social Sciences,” in Perspectives on Positive Political Economy, ed. James E. Alt and Kenneth Shepsle (New York: Cambridge University Press, 1990), 212–31. 45. The amount of really formal work in most of political science remains quite modest. For some content analysis, see Lisa L. Martin, “The Contributions of Rational Choice: A Defense of Pluralism,” International Security 24, no. 2 (1999): 74–81. 46. For a discussion in the context of the new institutional economics, see Williamson, “The New Institutional Economics,” 604ff. 47. The failure of economists to predict such major recent developments as the stagnation of the Japanese economy in the 1990s or the U.S. rate of growth and deficit reduction in the late nineties comes immediately to mind. 48. Walter Mattli, “Sovereignty Bargains in Regional Integration,” International Studies Review 2, no. 2 (2000): 148–80. 49. Davis B. Bobrow and Robert T. Kudrle, “Hedging: A Neglected Aspect of Regional Integration” (prepared for the eighteenth world congress of the International Political Science Association, Quebec City, August 1–5, 2000). 50. As perhaps first famously noted by Jean-Jacques Servan-Schreiber, Le défi américain (Paris: Editions Denoel, 1967). Of course, subnational groups also pursue international standing, as soccer competition illustrates. This simply suggests the obvious point that a single individual may identify with groups of varying aggregation for different activities. 51. Robert T. Kudrle, “Does Globalization Sap the Fiscal Power of the State?” in Coping with Globalization, eds. Aseem Prakash and Jeffrey Hart (New York: Routledge, 2000). 52. Bates, De Figueiredo, and Weingast, “The Politics of Interpretation.” 53. Dani Rodrik, “How Far Will International Integration Go?” Journal of Economic Perspectives 14, no. 1 (2000): 177–89. 54. Rodrik calls it the “augmented trilemma” by analogy to the economist's “impossible triple” of fixed exchange rates, capital mobility, and independent monetary policy. 55. John F. Helliwell, How Much Do National Borders Matter? (Washington, D.C.: Brookings Institution Press, 1998). 56. NAFTA does have an as-yet unrealized agenda of greater commonality in some areas, notably competition policy, but that agenda is very modest by European standards, and the failure to address it so far suggests lack of urgency.Page 652 → 57. Kudrle, “Does Globalization Sap the Fiscal Power of the State?” 58. From an economic viewpoint, each newcomer's contribution to the welfare of persons already in the jurisdiction turns on economies and diseconomies of scale and net fiscal impact: the contribution to the provision of pure public goods and the difference between the individual's tax payments and the value of exhaustive goods and services provided by governments to that person. By this measure, unskilled immigration appears to lower welfare, at least in the United States. In nonmarginal amounts, it also increases inequality in the before-tax distribution of income. 59. For an exhaustive discussion of the problem and a review of possible policies, see Reuven S. AviYonah, “Globalization, Tax Competition, and the Fiscal Crisis of the Welfare State,” Harvard Law Review 113 (May 2000): 1573–1646. 60. Although the United States does not control immigration effectively, this appears to result more from special interest pressures and a widespread view of illegal immigration as a victimless crime than from technical impossibility. Mercosur claims to envision open labor markets as a goal, but nothing has been

done to implement the rhetoric. Given the dramatic differences in living standards and differences in size, it is impossible to imagine open migration from Brazil to Argentina for the foreseeable future.

Page 653 →

INTERNATIONAL POLITICAL ECONOMY From Paradigmatic Debates to Productive Disagreements Lisa L. Martin In their instructions to the millennial reflections panel participants, the organizers of these panels suggested that they were necessary, in part, because of the increasing diversity of subfields and sections in the International Studies Association. They wrote that this diversity “has made increasingly difficult the crucial task of identifying intrasubfield consensus about important theoretical and/or empirical insights.” In this essay, I take issue with the organizers' claim. International political economy (IPE), perhaps in contrast to the field of international relations (IR) more broadly, is today characterized by growing consensus on theories, methods, analytical frameworks, and important questions. This is not to say that scholars in IPE all agree with one another—disagreement is as prevalent as ever. However, disagreement today generally takes the form of productive, theoretically and empirically motivated claims, rather than the paradigmatic clashes that characterized IPE as an early field of study. Scholars typically see alternative approaches as complementary or applying under different, specifiable conditions. Or disagreements might take the form of debate about relative weights that we should attribute to alternative explanatory variables. These types of disagreements are the hallmark of a mature, productive field of research. In this essay I elaborate my claim, illustrating it with examples drawn from the study of international trade.

The Development of IPE as a Field of Study IPE is a relatively young field. In its early years, it developed as a subfield within IR. Its questions, methods, and theories were drawn from the study of international relations, not from economics. Thus, Robert Gilpin's characterization of IPE as divided into three paradigms—realism, liberalism, and Marxism—emphasized the links between IPE and IR more generally.1 These three paradigms characterized the teaching and study of IR quite accurately, and IPE was seen as fitting into the mold defined by IR. Books published in Page 654 → the major series in IPE, at Cornell University Press, as well as articles published in the major outlet for IPE, International Organization, confirm the utility of Gilpin's designation until recent years. Yet, even while the threefold paradigmatic framework proved useful, hints of its inadequacy were apparent. Many authors and arguments did not fit neatly into any of the categories. Certain kinds of arguments did not fit into this categorization at all. Consider, for example, an argument about IPE that focused on the role of domestic institutions. Such an argument would not be liberal, since it concentrates on the state more than domestic interests. But it would not be realist unless national security concerns were paramount in those institutions, and it would not be Marxist unless the institutions were captured by capitalist interests. Thus, one of the more promising lines of research in IPE could not be captured by Gilpin's typology. Over time, an alternative organizing device has emerged for the study of IPE. It takes the familiar form of a 2 × 2 table.2 On one dimension, we ask whether the argument concentrates on the explanatory role of interests or institutions. On the other, we ask whether the major explanatory factors in the argument are located at the domestic or international level. This framework has proven more useful for arranging various works in IPE relative to one another, allowing us to identify precisely where various arguments differ from one another and where they agree. Nearly all major approaches in IPE fit fairly well in this framework, with the exception of works that consider the interaction of domestic and international levels, or the interaction of interests and institutions. I will discuss these points shortly. An additional valuable aspect of this organizing framework is that it has facilitated dialogue between political scientists and economists, which in itself is one of the more remarkable and progressive developments in IPE over the last decade or so. If IPE had its roots in IR, over time IPE scholars have recognized that they must have a firmer grasp of the

economics underlying the phenomena they study. At the same time, economists have begun adding political factors such as collective action and government institutions to their models. The result is an increasing overlap in the work of political scientists and economists, and the emergence of scholars who can talk comfortably to both groups. This dialogue with economists, and the fact that works in IPE must now confront the economic foundations of the policies they study, has made IPE a highly productive area of research in recent years. Increasingly, the theoretical apparatus used to study IPE has less in common with IR and more in common with political economy more generally.

Economics and Politics in International Trade These general points can be illustrated by work on the politics of international trade. IPE scholars initially approached this topic with a focus on either the domestic or international level. On the domestic, E. E. Schattschneider's famous study of the Smoot-Hawley Tariff provided the framework for thinking about free-trade and protectionist interests.3 Protectionist Page 655 → interests—primarily import-competitors—were concentrated and politically powerful. Free-trade interests, consumers and to some extent exporters, were more difficult to organize and not politically powerful. Those working on domestic explanations of trade policy also noted that U.S. institutions seemed to matter a great deal for trade policy. Until the 1940s, according to this analysis, trade policy was made by Congress, which was responsive to narrow protectionist interests and encouraged logrolling. The result was inefficiently high levels of protection that hurt U.S. consumers and had negative effects on the international level. Only after Congress delegated trade policy to the president, through the Reciprocal Trade Agreements Act (RTAA) of 1934, could policy reflect diffuse free-trade interests.4 Thus, another important aspect of domestic-level analysis was the emphasis on institutions. However, more recent institutionalist analyses have noted the weakness of earlier analyses of the RTAA. The logic of why Congress delegated authority, and why this delegation made a difference, often did not stand up to close scrutiny, relying on assumptions of naive constituents, members of Congress who acted against their own interest, or ideologically driven presidents. Douglas Irwin and Randall Kroszner (economists) have shown that congressional voting patterns do not fit the predictions of these explanations of the RTAA.5 On the international level, two sorts of analyses of trade dominated. Comparativists such as Peter Katzenstein and Peter Gourevitch contributed to the growing IPE literature by concentrating on how a state's role in the international economy influenced its trade policies.6 In work that has largely stood the test of time, they demonstrated how and why small, trade-dependent states adopted a package of policies that combined economic openness with high levels of government spending to protect individuals who were put at risk by exposure to the international economy. Another influential line of analysis on the international level, hegemonic stability theory, identified the distribution of power as the key factor explaining levels of openness in the international economy. Stephen Krasner elaborated the argument that the presence of a hegemonic state was a necessary condition for openness, resting his argument on an analysis of the interests of states of different sizes in maintaining an open international economy.7 He presented some empirical evidence to back up this claim but also argued that the rather weak fit between economic openness and patterns of hegemony suggested a certain amount of inertia in the international economy. Economic historians, notably Charles Kindleberger, were also active in developing versions of hegemonic stability theory.8 It has continued to have an impact today, for example in David Lake's more complex analyses of the incentives facing states with different levels of power.9 Economic interests did play a role in these important contributions to the development of IPE as a field. Yet, when one reads these works today, it is difficult not to be struck by the lack of contact between the new IPE field and the field of economics. With some exceptions, few references to the economics literature appear. Economic models or standard bodies of economic theory do not make an appearance. Methods commonly used by economists, Page 656 → such as statistics, are largely absent. While the economic interests of actors play an important role in the IPE literature during this period, few authors found it necessary to delve very deeply into

economics to determine exactly what these interests were. The result was a deep divide between IPE and international economics, crossed only in part by figures like Kindleberger. Economists had little use for debates about realism versus liberalism, and political scientists had little interest in how assumptions about factor mobility mattered. In stark contrast, the field of IPE today is characterized by intense interaction between economics and political science. Undergraduate IPE readers and graduate syllabi in IPE feature nearly as many publications by economists as by political scientists. New research in IPE has little chance of publication or influence if it does not draw appropriately on economic theory and results. In the area of international trade, the introduction of economics to IPE has been dramatic. Beginning in the late 1980s, studies of the politics of trade began drawing explicitly on economic theory. Ronald Rogowski's application of the Stolper-Samuelson model to explain the pattern of coalitions through history stimulated both work building on his insights and vigorous responses drawing from alternative economic models.10 Stephen Magee's famous “Three Simple Tests of the Stolper-Samuelson Theorem” illustrated the convergence between economic and political approaches to IPE.11 In order to determine whether Stolper-Samuelson provides an accurate model of coalitions on trade policy, Magee examined patterns of lobbying before Congress on trade bills. This approach is identical to that used by political scientists in testing other models of interests on trade, for example, Helen Milner's study of how exporters and multinational corporations have changed trade policy.12 Economists increasingly draw on political data to test their arguments, and political scientists use economic models of actor interests. Much of the debate today over interests on trade policy revolves around applications of Stolper-Samuelson versus the Ricardo-Viner model and questions the conditions under which these different models apply. The key difference in assumptions between the models is the level of factor specificity assumed. Stolper-Samuelson assumes that factors are mobile, easily shifted from one use to another. Ricardo-Viner, in contrast, assumes specific factors, where the costs of moving to the next-best use can be prohibitive. The two models predict very different patterns of coalitions on trade policy and demands for policy. Determining whether trade politics is best explained by one model or the other has generated an active research agenda. Even more productive, perhaps, are efforts to develop measures of factor specificity that will allow us to specify the conditions under which one or the other model will fit best.13 On the analysis of domestic institutions, we again see substantial improvement in the sophistication of the models of institutions being used, as well as contributions from both political scientists and economists. The analysis of interaction between Congress and the president in the United States has moved away from the problematic assumptions underlying earlier Page 657 → assessments of the RTAA. Rather than vague invocations of “learning,” self-abnegation, or inattentive constituents, recent analyses of institutional choice allow for strategic politicians and parties, and active, demanding constituents. For example, Susanne Lohmann and Sharyn O'Halloran argue that a combination of partisanship and economic shocks explains patterns of congressional delegation to the president.14 Michael Bailey, Judith Goldstein, and Barry Weingast argue that the institutions for trade policy reflect electoral considerations and the benefits that can be realized from negotiating with other states.15 Michael Gilligan argues that institutions matter to the extent that they “empower exporters.”16 On the international level, some work on the effects of the distribution of power continues. For example, Edward Mansfield has subjected hegemonic stability theory to more rigorous empirical tests.17 Joanne Gowa works along complementary lines, analyzing how security relations influence patterns of trade.18 However, the most active areas of research today focusing on interests on the international level draws from new economic theories. Strategic trade theory, models of intrafirm trade, and analysis of how investment and trade interact are setting the new agenda for research in this area. Studies of the role of international institutions in trade have also grown, again exemplifying contributions by both economists and political scientists. One particularly active area of research is in regional trade agreements, where economic and political factors interact in especially intriguing ways. Kyle Bagwell and Robert Staiger have

developed a series of models of international trade negotiations and institutions, beginning to provide economic microfoundations for studies of these areas.19 Research on the World Trade Organization is also taking off, although this is one area in which the contributions of economists remain modest. Instead, legal scholars have made more of a contribution to the study of the WTO's institutional features, such as its dispute settlement mechanism.

The Future: Getting to Interactions As this brief survey of research on trade indicates, the interests/institutions and domestic/international organizing framework continues to prove useful for categorizing active research agendas. Productive research is taking place in all four of the boxes defined by this framework. It also illustrates the extent to which paradigmatic clashes have disappeared from the bulk of research, and the degree to which a good understanding of economic principles is today essential for scholars working on IPE. All of these developments are, in my opinion, positive ones. IPE is a mature, exciting field of inquiry. However, being mature does not imply that the field is becoming dry or played out. Our understanding of interests and institutions is still quite primitive, with massive room for new work. Current research agendas provide the tools and point the way; but a multitude of challenges remain. Page 658 → One of the major remaining challenges will involve breaking down the barriers between interests and institutions, and between the domestic and international levels of analysis. In other words, as the field progresses, our fourfold typology may become less useful. Research on the interaction between interests and institutions, and between the domestic and international levels, is just beginning and is essential. Helen Milner, for example, has developed a model of the interaction between domestic interests and institutions that provides a rich set of testable hypotheses.20 She also brings another state into the model, suggesting a way to begin integrating the international and domestic levels of analysis. In the area of monetary politics, Lawrence Broz's analysis of exchange rate regimes is an example of how domestic and international factors interact in determining actors' interests.21 Studying interactive effects is complex and difficult, but the building blocks seem to be in place. Overall, our assessment of the development of IPE over the last forty years has to be highly positive, indicating an exciting and tractable research agenda for the future.

Notes 1. Robert Gilpin, The Political Economy of International Relations (Princeton, N.J.: Princeton University Press, 1987). 2. This framework is identical to that suggested by Jeffry Frieden and David Lake in the introduction to the newer editions of their IPE reader, except that I refer to “institutions” instead of the “state.” Jeffry A. Frieden and David A. Lake, International Political Economy: Perspectives on Global Power and Wealth, 4th ed. (Boston: Bedford/St. Martin's, 2000). 3. E. E. Schattschneider, Politics, Pressures, and the Tariff: A Study of Free Private Enterprise, as Shown in the 1929–1930 Revision of the Tariff (New York: Prentice Hall, 1935). 4. I. M. Destler, Making Foreign Economic Policy (Washington, D.C.: Brookings Institution Press, 1980). 5. Douglas Irwin and Randall Kroszner, “Interests, Institutions, and Ideology in Securing Policy Change: The Republican Conversion to Trade Liberalization after Smoot-Hawley,” Journal of Law and Economics 42, no. 2 (1999): 643–73. 6. Peter Katzenstein, Small States in World Markets: Industrial Policy in Europe (Ithaca, N.Y.: Cornell University Press, 1985); Peter Gourevitch, Politics in Hard Times: Comparative Responses to International Economic Crises (Ithaca, N.Y.: Cornell University Press, 1986).

7. Stephen Krasner, “State Power and the Structure of World Trade,” World Politics 28, no. 3 (1976): 317–27. 8. Charles Kindleberger, The World in Depression, 1929–1939 (Berkeley: University of California Press, 1973). 9. David A. Lake, “Beneath the Commerce of Nations: A Theory of International Economic Structures,” International Studies Quarterly 28, no. 2 (1984): 143–70. 10. Ronald Rogowski, Commerce and Coalitions: How Trade Affects Domestic Political Alignments (Princeton, N.J.: Princeton University Press, 1989). 11. Stephen Magee, “Three Simple Tests of the Stolper-Samuelson Theorem,” in Issues in International Economics, ed. Peter Oppenheimer (London: Oriel, 1980), 138–53. 12. Helen V. Milner, Resisting Protectionism: Global Industries and the Politics of International Trade (Princeton, N.J.: Princeton University Press, 1988).Page 659 → 13. See, for example, James E. Alt et al., “Asset Specificity and the Political Behavior of Firms: Lobbying for Subsidies in Norway,” International Organization 53, no. 1 (1999): 99–116; Michael J. Hiscox, “Class versus Industry Cleavages: Inter-Industry Factor Mobility and the Politics of Trade,” International Organization 55, no. 1 (2001): 1–46. 14. Susanne Lohmann and Sharyn O'Halloran, “Divided Government and U.S. Trade Policy: Theory and Evidence,” International Organization 48, no. 4 (1994): 595–632. 15. Michael Bailey, Judith Goldstein, and Barry Weingast, “The Institutional Roots of American Trade Policy,” World Politics 49, no. 3 (1997): 309–38. 16. Michael Gilligan, Empowering Exporters: Reciprocity, Delegation, and Collective Action in American Trade Policy (Ann Arbor: University of Michigan Press, 1997). 17. Edward Mansfield, Power, Trade, and War (Princeton, N.J.: Princeton University Press, 1994). 18. Joanne S. Gowa, Allies, Adversaries, and International Trade (Princeton, N.J.: Princeton University Press, 1994). 19. Kyle Bagwell and Robert Staiger, “An Economic Theory of GATT,” American Economic Review 89, no. 1 (1999): 215–48. 20. Helen V. Milner, Interests, Institutions, and Information: Domestic Politics and International Relations (Princeton, N.J.: Princeton University Press, 1997). 21. J. Lawrence Broz, “The Domestic Politics of International Monetary Order: The Gold Standard,” in Contested Social Orders and International Politics, ed. David Skidmore (Nashville: Vanderbilt University Press, 1997).

Page 660 → Page 661 →

CONCLUSIONS: AND ASSESSMENT OF MILLENNIAL REFLECTIONS ON INTERNATIONAL STUDIES Frank P. Harvey and Michael Brecher The questions posed to each contributor at the outset of the Millennial Reflections Project were designed to elicit self-critical reflections on the state of international studies (IS) at the dawn of a new millennium. We asked more than forty distinguished scholars to evaluate the current status and future prospects of the discipline, to assess the extent to which we have acquired significant theoretical insights, to tell us where we stand on unresolved debates and why we have failed to resolve them, to evaluate the standards we should use when selecting and applying appropriate methodologies, and to address the broader question of progress and cumulation in international studies. The academic credentials of the contributors combine to make this group uniquely qualified to answer these important questions. The most difficult challenge for us as editors was to draw together a coherent set of conclusions from such a diverse collection of think-pieces. Although each contributor was given the same set of questions, some authors addressed all of them, others dealt with only one or a few, and several refused to answer any question that privileged the standard social science framework, since that would bias the results in favor of identifying progress. Interpretations of the questions were as varied as the answers, and when interpretations were similar the approach was different—“progress,” for example, was assessed in terms of specific theories, subfields, paradigms, epistemic communities, the IS discipline more generally, and, in a few cases, all of the above. This diversity made it very difficult to select an appropriate framework to organize our impressions and to formulate our overall conclusions. To avoid a disjointed discussion of progress and a somewhat superficial assessment of the state of international studies, we decided to conclude by revisiting our starting point. Recall that we began the project by recognizing the virtue of “dialogue” across epistemic and paradigmatic communities, but we argued that such interepistemic dialogue would remain premature if there is insufficient intraepistemic consensus within communities to justify the conversation in the first place. If those who share common interests and Page 662 → approaches cannot agree on what they have accomplished to date, where would we find the reference points to facilitate creative, substantive, and meaningful dialogue across perspectives? Our call for self-critical, state-of-the-art “reflection” was meant to challenge members of the international studies community to search for intraepistemic consensus, engage in self-critical debate about significant accomplishments and limitations within their own subfield or paradigm, and, most important, avoid the tendency to define accomplishments with reference to the failures and weaknesses of other perspectives.

Expansive versus Selective Cumulation in International Studies One of two mutually-exclusive views about progress and cumulation in international studies is implicit in each of the forty-four contributions. One view maintains that the health of IS, or of any academic discipline for that matter, is derived from its capacity to grow and expand—the more epistemic communities, paradigms, subfields, frameworks, theories, methodologies, questions, puzzles, and data there are, the healthier and more progressive the discipline. Expansion signifies openness, mutual respect, accommodation, and maturity. The other view maintains that the health and maturity of any academic discipline is not solely a function of its capacity to expand but of its capacity to learn and select . Maturity, in other words, is demonstrated through emerging consensus on important questions, solutions to key puzzles, appropriate methodologies, validated theories, and agreement on the utility of specific frameworks and paradigms for describing and/or explaining international behavior. In other words, any discipline can grow and expand, but only a mature and progressive discipline can select by resolving core theoretical and methodological debates.

The overwhelming impression derived from the forty-four contributions to the Millennial Reflections Project is that the field of international studies is very healthy in the expansive sense but is far less progressive (and becoming increasingly digressive) in the selective sense. For those who believe selective cumulation represents the hallmark of a mature academic discipline, the lack of agreement within epistemic communities and paradigms should be particularly disconcerting. We develop our argument with reference to selected illustrations from the chapters. Our starting point is Vertzberger's observation in “Foreign Policy Analysis: Steady Progress and a Half-Empty Glass” that any research domain (i.e., epistemic community) “consists of three constituent components, or building blocks: theory, which requires systematic conceptualizing and generalizing toward theory-building; empirical data, which result from comprehensive search, collection, and organization of relevant data; and methodology, which requires applying a set of transparent and explicit rule-based methods and techniques for the transformation of data into theory-relevant information and then relating the transformed empirical data to the theory for hypotheses-testing and theory-validation.” A meaningful discussion of progress in IS requires that a research domain be “ontologically un-packed Page 663 → and its constituent elements individually examined, and then re-packed for an overall progress assessment.” This is precisely how we intend to proceed.

Epistemic Complexity, Expanding Division, and the Proliferation of Unresolvable Debates Contributions to this volume will be unpacked and categorized according to where they fit within core theoretical and methodological debates in IS. The outline in figure 1 represents a snapshot (or road map) of these debates and is included here to facilitate discussion. Two points should be noted before proceeding. First, not all of the chapters fall neatly into one or another of these categories—some authors, for example, address multiple debates in IS. Second, the list (and associated illustrations) is not meant to be exhaustive; there are many other debates and points of contention within and across the two dominant epistemic communities that are not included in the figure, which only confirms our central point about the extent of complexity and division in IS. The figure is meant to serve two important functions: it helps to organize the contributions, and, more important, it illustrates the multiple impediments to selective cumulation and progress. Intraepistemic Debates: Mainstream International Studies (figure 1, section 1) Epistemological debates and divisions within IS fall into two broad categories: inter- and intraepistemic. These, in turn, can be further subdivided into inter- and intraparadigmatic (theory and methodology) and inter- and intrasubfield debates. Interepistemic debates represent the highest level of division in the field—disagreements between proponents of mainstream IS (section 1 of the outline) and critical or alternative perspectives (section 2) are the most obvious examples. These divisions are least surprising and most difficult to resolve because they encompass so many points of contention. Intraepistemic debates, on the other hand, occur among scholars who share a common set of assumptions about the nature of knowledge and how we should go about generating it. For example, the standard social science model assumes, among other things, that facts and values are distinguishable, that a positivist-empiricist methodology is essential for progress, that patterns and regularities in the social world can be uncovered and are meaningful, and that cumulation and progress are essentially linear. Although this particular model is shared by virtually every scholar in the mainstream of international studies, there are nevertheless several intra- and interparadigmatic divisions over theory and methodology. New and unresolved debates within and across realism, institutionalism, mainstream constructivism, liberal feminism, and Marxism/radicalism (or globalism) are examples. All these disagreements take place within the same epistemic community, yet they are no less damaging to the search for selective cumulation. Page 664 → Page 665 → Intraparadigm Debates: Theory (1A)

Contributions to the section on realism provide clear illustrations of ongoing intraparadigmatic debates in the mainstream literature. Each author in this section offers a different interpretation of the paradigm's core tenets and assumptions, with no clear consensus emerging even among those who claim to accept realism as a potent explanatory framework for contemporary international relations. Proponents continue to define realism in ways that facilitate confirmation, and critics continue to apply very narrow definitions that facilitate refutation. There is as yet very little consensus on the precise requirements for a fair and balanced test of realism and associated theories. Page 666 → Realism, according to K. J. Holsti (“Performance and Perils of Realism in the Study of International Politics”), has four functions: (1) it is a compressed, or shorthand, description of the essential characteristics of international politics; (2) it is a diagnostic tool for analyzing the sources of international conflict and derivatively for identifying the conditions for peace and security; (3) it is a set of prescriptions that derives from the diagnostic exercise; and (4) it is a normative theory that posits the survival and security of the political community as the ultimate political value. Any evaluation of the tradition must distinguish between these four functions. But realists do not all agree that these four characteristics encompass the epistemological foundation or identity of the paradigm. Classical, contemporary, offensive, defensive, and neorealists differ widely on what they consider to be truly “essential” characteristics of the international system. Realists do not uniformly agree on the diagnostic relevance of the paradigm or the circumstances under which the paradigm is more or less likely to apply. Nor do they uniformly conceptualize the paradigm in normative terms. Similarly, almost every realist agrees that modifications and adjustments are likely to improve the paradigm's contemporary relevance, but there is extensive disagreement about why adjustments are needed and what changes should be made to accommodate emerging international and transnational trends. James (“Systemism and International Relations: Toward a Reassessment of Realism”) calls for modification because “realism has a vast, un-tapped potential” and existing treatments “are highly self-limiting”—his recommendations apply to theoretical adjustments at the systems level. Grieco (“Modern Realist Theory and the Study of International Politics in the Twenty-first Century”), on the other hand, recommends “internal development” and “external dialogue” with other approaches like social constructivism and domestic structuralism. In contrast, Mearsheimer's forceful and un-apologetic defense of neorealism in “Realism, the Real World, and the Academy” implies that few if any adjustments are required to enhance the paradigm's relevance—its potency as an explanatory framework remains as strong today as it ever was. According to Mearsheimer, the sweeping dismissal of realism by critics today has less to do with the paradigm's capacity to explain contemporary international politics and more to do with the political and ideological agendas of critics within the discipline. There are also fundamental disagreements over the paradigm's capacity to explain conflict and war. Grieco's conclusion that “modern realism will continue to provide the surest starting point for our understanding of the problem of war and peace among nations” directly contradicts Vasquez's (“Realism and the Study of Peace and War”) view that a complete break with the realist paradigm will probably produce a better guide to research … as well as more theoretical insights…. Page 667 → War is brought about by more than just differences in capability. Peace cannot be preserved and built simply because one side is more powerful than another. Politics, like everyday life, does not consist solely of intimidation and coercion. Until the field has a paradigm that examines more than just the single factor of power, it will not understand what causes war or how peace can be built.

Although Vasquez's work continues to be a standard reference for critics, there are at least three reasons why his findings are far from sufficient to establish closure on these debates. First, few if any classical or contemporary realists have ever focused exclusively on a single factor of “power” or “capabilities,” nor have they ever argued that intimidation and coercion encompasses everything there is to know about international behavior. Second, most realist definitions of power and capabilities are far more nuanced than Vasquez allows, and their explanations of war have been decidedly more complex than he acknowledges. Third, while Vasquez is correct that realists have an obligation to explain anomalies that challenge their expectations about international behavior, a more balanced assessment of the paradigm's contemporary relevance requires that the number and significance of anomalies within realism be compared to those within institutionalism, feminism, constructivism, or any other paradigm. The same standards of verification must be used when evaluating alternative perspectives. Consider Singer's (“Accounting for Interstate War: Progress and Cumulation”) observations from the point of view of theory-damaging anomalies: How often in history have we seen most states in the global or a regional system behaving in a cooperative fashion? Perhaps the Congress of Vienna system held for a while, but by 1854 Europe had experienced the Franco-Spanish, Russo-Turkish, Austro-Sardinian, Schleswig-Holstein, Roman Republic, and Crimean Wars…. every social organization is a coalition, and there will almost always be some members of the coalition with incentives to initiate or perpetuate sharp conflict, thus not exactly contributing to the evolution of cooperation. It is not surprising that Vasquez's conclusions are inconsistent with those of K. J. Holsti, Grieco, Mearsheimer, and many other realists. But they also appear to be inconsistent with positions put forward by other well-known critics of realism, including Keohane (“Institutional Theory in International Relations”), Nye (“Transnational Relations, Interdependence, and Globalization”), most institutionalists, and several other contributors to this collection, all of whom conclude that realism continues (and will continue to) provide an important part of any complete explanation of interstate (or intergroup) conflict. It has never really been a question of right or wrong, but of time, place, and circumstance. Page 668 → Of course, realism is not the only paradigm to produce internal debates—institutionalists, feminists, and constructivists continue to question the extent of consensus within their own communities. Keohane argues in “Institutional Theory in International Relations,” for example, that there are good reasons to be self-critical about progress within the institutional paradigm: “The most general point is that we have been better at extending our descriptions to new issue areas, and incorporating new trends into them, than we have been at improving the explanatory power of our theories. That is, our record is better at description, and even descriptive inference, than at explanation.” Among the problems cited by Keohane are that “distributive issues have been recognized as important, but institutional theory has been slow to integrate them into its analysis; “we have been slow to develop systematic databases of institutions, which could be used for sophisticated quantitative work” and “methodologically, we face the usual problem of strategic interaction: rarely are there unique equilibrium solutions. Since international institutions are characterized by multiple actors and audiences, the range of possible outcomes is enormous.” Kolodziej (“Security Theory: Six Paradigms Searching for Security”) highlights debates within constructivism: Constructivists may be divided into two broad groups. “Light” constructivists concede the force of material factors in driving actor behavior. They do not reject outright the assumptions of other research paradigms…. “Heavy” constructivists, on the other hand, go further. The material conditions within which actor exchanges are embedded are radically subordinated in their causal effect to the meaning and value attributed by actors to these factors, whether military capabilities or economic and technological resources. The identities constructed by actors determine their interests, policies toward each other,

and behavior.

With respect to feminist research, Zalewski (“Feminism and/in International Relations: An Exhausted Conversation? or Feminists Doing International Relations: The Cut(ting) Edge of Contemporary Critical Theory and Practice?”) notes how difficult it is to make authoritative statements about the impact of work on gender in international relations, because “it relies on the assumption that feminist work in international relations can be fitted into a coherent framework upon which such judgments can then be made. This is not the case as feminism is a vast and sometimes contradictory body of work that makes it therefore difficult (and perhaps inadvisable) to make sweeping and authoritative generalizations about” (emphasis added). In sum, no paradigm is immune to internal debates, divisions, and contradictions, and none constitutes a solid, well-defined, coherent body of research and findings. Leaving aside the very difficult task of resolving disagreements across Page 669 → these diverse perspectives (see 1C in figure 1), the more relevant point here is that there are no clear indications that debates within these communities are likely to be resolved any time soon. This lack of closure, and the exponential increase in the number of new debates, does not bode well for progress in a “selective” sense. Intrasubfield Debates (1A.1) Further confounding the search for selective cumulation is the fact that proponents of each of the previously mentioned paradigms can be further separated into specific subfields, each encompassing unique debates and disagreements. Examples from foreign policy analysis, international security studies, and international political economy are described in the following paragraphs to illustrate the point. In his assessment of the research program on foreign policy analysis (FPA) in “Foreign Policy Analysis: Steady Progress and a Half-Empty Glass,” Vertzberger concludes that the field “still lacks a theoretical framework that will allow us to move from the analysis of foreign policy as a two-level game (domestic and international) to a three-level game (domestic, international, supranational) and will account for the increasingly salient impacts of regionalization and globalization.” The emergence of the European Union is cited to illustrate his point—the EU requires a theory of foreign policy that accounts for the three-way interaction between the state's domestic arena, its international policy-making processes, and the EU's supranational institutions, all in the context of a creeping devolution of authority away from nation-states. In addition to questions of where FPA should be going as a subfield, there are also differences within FPA regarding the variables and techniques we need to improve our overall understanding of foreign policy. Walker (“Beliefs and Foreign Policy Analysis in the New Millenium”) claims that “beliefs” are important causal mechanisms—“Identifying beliefs as pivotal macro-mechanisms of bounded rationality within this framework brings the constructivist and rational choice schools into the analysis.” Moreover, a focus on “crises” as turning points in international relations “integrates threats of conflict and opportunities for cooperation, thereby synthesizing a focus on the respective explananda in the realist and liberal schools of international relations theory.” The hope, shared by other contributors to the FPA section, is that by refocusing, retooling, and reemphasizing key variables, progress, synthesis, and cumulation are just around the corner—it's simply a question of time. In “Simulation and Experimentation in Foreign Policy Analysis: Some Personal Observations on Problems and Prospects,” Wilkenfeld makes a similar argument with reference to “simulations” and the use of controlled laboratory environments. These new techniques, according to Wilkenfeld, will allow us to Page 670 → “make progress in understanding how decision makers cope with and process information in crisis versus noncrisis environments, how third parties can affect the course of and outcomes to contentious situations, how actor attributes such as power (of various sorts) affect their negotiation strategies and

behavior, and the circumstances under which violent strategies are chosen…. the experimental environment—when properly specified—can help us fill in the gaps in our knowledge and ultimately allow us to generalize beyond a limited experiential environment.”

Ole R. Holsti (“Public Opinion and Foreign Policy Analysis: Where We Were, Are, and Should Strive to Be”) reviews new insights from research on public opinion and foreign policy, focusing specifically on the implications of new and emerging technologies. He offers several very important suggestions for improving opinion research and addressing long-standing problems with validity and reliability. New variables (e.g., “beliefs” as macro-mechanisms of bounded rationality and “crises” as turning points), new techniques (e.g., simulations and controlled laboratory experiments), and new public opinion technologies combine to create new approaches for analyzing and explaining foreign policy. Each author offers a slightly different take on the same theme—namely, it's just a question of time before our models and methods improve our capacity to generate consensus about foreign policy behavior. But consider the underlying implications of this position—old techniques are considered obsolete, not because they are inherently wrong or inappropriate but because they haven't succeeded in establishing closure on the subfield's core questions and puzzles. The result is a never-ending search for new techniques, different questions, and new debates and, ultimately, more division. Expansive rather than selective cumulation becomes the norm, not necessarily by choice but because we really have no other option. Similar examples can be found in Kriesberg's discussion in “Convergences between International Security Studies and Peace Studies” of progress in international security studies. He anticipates consensus and “convergence” between security and peace studies “as the range of the peacemaking and peace-building agenda grows broader and more complex.” According to Kreisberg, “one important sign of and contributor to that progress is the increased range of voices joining the conversations about peace and security. As previously noted, this is the case for the increasing engagement of women in peace and security studies and the growing application of feminist thought in those domains. They are providing new insights and important perspectives.” But new insights and new perspectives, by definition, render old insights and old perspectives suspect. If contemporary security and peace studies are indeed characterized by complexity and an ever-expanding number of perspectives and approaches, then the field is likely to produce more, not fewer, intellectual debates and disagreements. Again, expansive rather than selective cumulation represents the default position. Page 671 → Like Kriesberg, Kolodziej (“Security Theory: Six Paradigms Searching for Security”) identifies similar points of contention within security studies but tries to conclude on a diplomatic note: “depending on the security question to be addressed, each of these paradigms, in greater or lesser measure, has something potentially worthwhile to say about the behavior of relevant actors. They can, arguably, assist actors in understanding their security problems and in solving them” (emphasis added). While these “middle-ground” sentiments are always welcomed by both sides in these debates, the more relevant (and usually unanswered) question is how to define words such as “greater,” “lesser,” or “potentially worthwhile.” Any attempts to render these statements more precise inevitably produce the unresolvable debates that continue to plague this (and other) subfields. With respect to international political economy, Martin concludes in “International Political Economy: From Paradigmatic Debates to Productive Disagreements” that consensus within IPE on substance and methods “depends at least on how one defines the field. The more narrowly it is defined, the more strongly one probably can answer this in the affirmative.” Notwithstanding this caveat, however, Martin makes the strongest case that selective cumulation has occurred within IPE: International political economy … perhaps in contrast to the field of international relations (IR) more broadly, is today characterized by growing consensus on theories, methods, analytical frameworks, and important questions. This is not to say that scholars in IPE all agree with one

another—disagreement is as prevalent as ever. However, disagreement today generally takes the form of productive, theoretically and empirically-motivated claims, rather than the paradigmatic clashes that characterized IPE as an early field of study. Scholars typically see alternative approaches as complementary or applying under different, specifiable conditions. Or disagreements might take the form of debate about relative weights that we should attribute to alternative explanatory variables.

But Martin never really specifies where the “growing consensus” on theories, methods, and analytical frameworks has led, or whether we have in fact established closure on any of the core IPE questions and puzzles—which questions? which puzzles? which methods? Nor does she clearly define what “productive” or “specifiable conditions” actually mean. Debates within IPE may have moved on from inter-to intraparadigm disagreements (“the hallmark of a mature, productive field of research,” according to Martin), but maturity in this case is defined almost entirely in terms of expansive, not selective, cumulation. If the “growing consensus” in IPE is actually a growing acceptance that alternative approaches deserve an equal voice, without a clear and easily identifiable consensus on when and under what conditions these alternatives do and do not apply, then IPE, like every other subfield in IS, remains stuck at the expansion stage of progress. Page 672 → Subdebates over Levels, Theories, and Cases (1A.2, 1A.3, and 1A.4) Subfields also encompass additional level of analysis (1A.2), inter- and intratheory (1A.3), and inter- and intracase (1A.4) debates. Disagreements over “appropriate” levels of analysis have a long pedigree in the literature and will not be covered at length here. It suffices to point out that foreign policy behavior, security and peace studies, international political economy, and most other subfields continue to be explored from the perspective of different levels of analysis, each producing a diverse collection of causal variables and competing claims about explanatory relevance. The search for selective cumulation is further confounded by ongoing debates within levels of analysis between proponents and critics of specific theories. Perhaps the most obvious example is deterrence theory—notwithstanding decades of research on whether deterrence constitutes a valid theory of human social and /or military strategic behavior, we have arrived at two mutually exclusive conclusions: yes and no. Both conclusions are based on findings from the same list of deterrence encounters, yet the prevailing and unchanging opinion defended by both sides is that their approach to testing is more appropriate and their findings are more than sufficient to establish closure on the subject. In “Alternative and Critical Perspectives” Smith offers a similar illustration from research on globalization theory—“What if the answers are ultimately dependent on which ‘facts’ one selects and what one sees as the main issues in the debate? Think for a second about the concept of globalization: how might one resolve a debate about its impact on world politics if one cannot agree on who is the relevant unit for assessing this impact, or if one sees it as having both positive and negative effects?” The point is that debates over deterrence and globalization theories may be impossible to resolve. In addition to uncompromising debates over the validity of individual theories, there are also a host of inter- and intracase debates that produce mutually-exclusive interpretations of single events. Consider for example the mirror-image interpretations of NATO's intervention into Kosovo. Critics maintain that NATO officials miscalculated and misunderstood the negative effects of letting politics influence their decisions and strategies. Excluding ground troops from the outset of the conflict (in order to limit casualties) prolonged the war, because Slobodan Milosevic knew that damage would be limited to air strikes alone. In other words, NATO lacked the commitment and resolve to mount a credible deterrent threat. The lesson: never exclude ground troops as an option, and if you do, never give your enemy a clear indication of your strategy by making this explicit. The mirror-image interpretation of the same case is that Milosevic misunderstood the benefits to NATO of letting politics guide their air-strike strategy. Excluding ground troops from the outset, and making this preference

explicit, actually helped to shorten what would have been a long and protracted war. Milosevic was denied any opportunity to use NATO casualties to demonstrate to his military and domestic constituencies that they Page 673 → were winning, or at least not the only ones losing, the war. Consequently, Milosevic was limited to only a few small and insignificant victories—e.g., the capture of three Macedonian peacekeepers, the downing of the U.S. stealth bomber, and the Chinese embassy bombing. According to this perspective, casualties are not bad because they threaten to increase domestic pressure in NATO capitals to prematurely exit the war but because they convince the losing side (Serbia) to stay the course in hopes that NATO officials will collapse under the pressure of negative public opinion. Two interpretations of the same case lead to mutually-exclusive theoretical and policy-relevant findings about the use of ground troops. Even the “fact” that the Kosovo war lasted seventy-eight days produces distinct interpretations—critics cite this number as evidence that NATO miscalculated, because the war lasted as long as seventy-eight days; proponents cite the same number as evidence that NATO succeeded, because the war ended after only seventy-eight days. Now, multiply this particular example by the thousands of facts, observations, and operational decisions in a typical database. If different interpretations of one fact (seventy-eight days) from a single case (Kosovo) can affect our assessment of a simple theory (deterrence) from one subfield (security studies) within the same paradigm (realism), and if all of these differences exist within the mainstream epistemic community, what are the prospects for selective cumulation or meaningful dialogue across diverse perspectives? Intraparadigm Debates: Methodology (1B.1 through 1B.3) With all of the foregoing in mind, consider the multiplier effects of uncompromising debates within mainstream IS (regardless of paradigm or subfield) over the best available methods to compare cases, compile “facts,” analyze data, test hypotheses, and select theories. The field remains saturated with efforts to criticize and falsify by pointing out the flaws inherent in the methods used. Note the following list of typical criticisms: too many historically-distinct cases were aggregated and not subjected to sufficient in-depth examination; an insufficient number of cases were used to permit valid generalizations; the study lacked the rigor of deductive theorizing; mathematical models and game theory are too abstract and require an empirical referent, preferably informed by statistical analysis; statistical analysis lacks the rigor of formal theorizing; the statistical procedures were flawed, given the type of variables in the model; the level of analysis is inappropriate; the level of analysis is appropriate, but the inferences drawn are not; and so on. All of these criticisms and associated methodological preferences confound the search for consensus and selective cumulation in IS. As Singer (“Accounting for Interstate War: Progress and Cumulation”) correctly observes, “priority today is toward the multilevel and multivariate model, reflecting the consensus that outcomes of theoretical or policy interest just cannot be explained by one or two factors at one or another level of social complexity. Too many psychological, domestic, regional, dynamic, and system-level phenomena are unfolding and interacting within the same Page 674 → ‘causal’ sequence, and it is essential that we keep an eye on most of them at the same time while some of them can be held constant and others are ‘allowed’ to change their values and magnitudes sequentially or simultaneously.” Leng (“Quantitative International Politics and Its Critics: Then and Now”) makes perhaps the strongest argument in the volume that progress has occurred within the quantitative research program and that, collectively, the findings are statistically and substantively significant. However, as Zinnes (“Reflections on Quantitative International Politics”) clearly illustrates in her contribution, there is an important intraparadigmatic divide within the quantitative school, between proponents of statistical and formal analysis, that raises serious doubts about Leng's conclusion. Her argument is based on an impressive survey of the field's leading journals (American Political Science Review, American Journal of Political Science, Journal of Conflict Resolution, Conflict Management and Peace Science, Journal of Peace Research, and International Studies Quarterly) from 1990 through 1999. She finds that quantitative research in IR is heavily dominated by statistical analyses; with the exception of the American Political Science Review, “the use of statistical analyses has increased from the 1990–94 period to the 1995–99 period, while the use of mathematical models, with the exception of Journal of Peace Research, has held constant or decreased slightly.” While the field appears to be moving in the quantitative direction, “we are doing so largely through the use of statistical

analyses.” The results are disturbing to Zinnes for several reasons: The typical statistical study begins with an argument, formulated verbally, from which one or more hypotheses are generated. Data are collected and statistics are used to assess the likelihood that the hypothesis is false, or conversely that the data are adequate to conclude that the hypothesis is supported by the data. But what is the relationship between the verbal argument and the hypothesis? Without the use of a formal language—that is, a mathematical model—it is next to impossible to demonstrate that the hypothesis actually flows from the argument. The statistical analysis of the hypothesis thus becomes, by default, the “theory” that is being evaluated. The argument that supposedly produced the hypothesis is essentially irrelevant. But the argument that was meant to produce the hypothesis is usually where the interesting ideas are found, the place where the researcher is explaining why certain things happen. We have reduced what may have been a rich interpretation of the international world to an often un-interesting if-then statement. Consequently if IR research is becoming increasingly quantitative through the ever greater use of statistics, we are replacing possible explanation with far less interesting statements about what is associated with what. The same concerns are shared by Nicholson (“Formal Methods in International Relations”)—“it is hard to suppose that deductive modeling is going to be less needed. As theory advances, it is more likely to become more Page 675 → deductively complex rather than less.” But the debate does not end here; even if one accepts the call by Zinnes and Nicholson to apply formal theory rather than focus exclusively on statistical or correlative analyses, there are still important disagreements among formal theorists over appropriate methods. Brams (“Game Theory in Practice: Problems and Prospects in Applying It to International Relations”), for example, identifies several common errors in typical game-theoretic applications, including misspecifying rules, confusing goals with rational choice, arbitrarily reducing the multiplicity of equilibria, and forsaking backward induction. Each of these points of contention produces its own set of debates among game theorists, none of which appears to be on the verge of resolution. Notwithstanding these disagreements, Bueno de Mesquita (“Accomplishments and Limitations of a Game- Theoretic Approach to International Relations”) concludes that by any yardstick, game theory has matured markedly over the past six decades, but much remains to be done. Game theory endures significant limitations, some of which are being overcome as the state of knowledge improves, others of which are inherent to the approach and are unlikely to be removed, though they can be improved upon by borrowing from other perspectives…. … As the tools for dealing with computational complexity improve, we may perhaps look forward to evolutionary models or dynamic games replacing the comparative static analysis prevalent in current game theory. Consider the similarities between these observations and those put forward by contributors to the FPA section—it's just a question of time before new tools and techniques deliver the goods. What we need, according to Bueno de Mesquita, Brams, Nicholson, and Zinnes, then, is more rigor and better formal and game-theoretic models. In contrast to the call for more formal theory, Vertzberger (“Foreign Policy Analysis: Steady Progress and a HalfEmpty Glass”) offers a very different view, one that represents a growing sentiment in IS that formal theory is moving the field in the wrong direction: “When the stricture of a methodology … does not allow for the in-corporation and integration of important and relevant knowledge into the explanation and interpretation of the world, merely because it cannot be formulated in a particular manner required by formal modeling, not because it is untrue, then the methodology becomes an impediment to progress. In other words, the means determine the ends rather than the other way around.” And in contrast to Vertzberger's arguments, Starr concludes in “Cumulation, Synthesis, and Reasearch Design for

the Post-Fourth Wave” that the Page 676 → problem has less to do with the application of formal theory (or statistical analysis) and more to do with the failure “to match theory and logic to research design and method…. Formal models translate theory into mathematical and logical constructs. As such, they ensure the internal logical consistency of the formalized translation of theory. The challenge then confronting the researcher is somehow to capture these formalized relationships in empirical representations and evaluations.” Progress and cumulation require that we accommodate all elements of the “research triad”—theory, logic, and method. Debates over methodology, of course, are not limited to the quantitative school(s). Both Levy and Maoz (“Qualitative Methods in International Relations” and “Case Study Methodology in International Studies: From Story-telling to Hypothesis Testing” respectively) summarize new and emerging debates over qualitative methods and provide excellent reviews of the many contributions and limitations of comparative case study techniques. The point, again, is that if intraquantitative (1B.1) and intraqualitative (1B.2) debates are not likely to be resolved soon (and, indeed, appear to be multiplying), what are the implications for resolving intermethodology debates (1B.3) between these two schools? Interparadigm Debates: Theory (1C) Of course, even if disagreements over methods are resolved, say, by finding the optimum mix of approaches and techniques (e.g., structured focused comparison, structured empiricism, or Boolean methods), consider how complicated the search for selective cumulation becomes when one contemplates the unresolved debates (and multiple avenues of division) across paradigms. Young's contribution (“Are Institutions Intervening Variables or Basic Causal Forces?: Causal Clusters versus Causal Chains in International Society”), for example, highlights the ongoing interparadigm debates between realists and neoliberals. He compares neorealists like Susan Strange and John Mearsheimer, “who regard institutions as epiphenomena that have little or no independent impact on the content of collective outcomes,” with neoliberals like Keohane and Nye, “who assign greater causal significance to institutions, even though they, too, adopt an instrumental perspective on the roles that institutions play.” Other neoliberals go even further by “drawing a distinction between the provisions of individual regimes and the deeper, more encompassing constitutive features of international society.” These views alone represent at least three distinct schools of thought on a single issue—the relevance of international institutions and regimes. K. J. Holsti (“Performance and Perils of Realism in the Study of International Politics”) also addresses the realistneoliberal divide with reference to the European Union: Page 677 → Numerous studies have demonstrated beyond doubt that a realist-type analysis of the relations between the European Union members would be incomplete at best and totally misleading at worst. Other approaches, including institutionalism and constructivism, do a better job of describing and explaining the political dynamics of this region. So, contrary to realist characterizations of international politics much has indeed changed in the quality of international life. The EU and other international and supranational institutions (and associated regimes) are constantly cited as “crucial” anomalies for realists to explain, but even these examples are far from sufficient to establish closure on the subject of relevance, for two reasons. First, liberal internationalists face their own difficult questions about relevance—as Kolodziej (“Security Theory: Six Paradigms Searching for Security”) pointedly asks, “what, for example, does liberal institutionalism or liberal economic theory have to say about genocide in Rwanda or ethnic cleansing in the former Yugoslavia or racial and religious civil war in Sudan or continuing Middle East hostilities

between Arabs and Jews?” Second, realism continues to serve as a useful guide when explaining behavior within and among so-called anomalous nonstate actors, institutions, and corporations; much of this activity takes place in the context of self-interest, global competition, intergroup conflict, and protracted rivalry, and all of this behavior is driven by the “evolutionary” motive to survive. As such, references to realist theories about alliance formation, power balancing, and coercive diplomacy are entirely appropriate for explaining, for example, corporate mergers, the success and failure of international regimes, and even interactions among nongovernmental organizations. Finally, just as Mearsheimer (“Realism, the Real World, and the Academy”) mounts a persuasive defense of realism against the charge of obsolescence, Michael Cox (“The Continuing Story of Another Death Foretold: Radical Theory and the New International Relations”) offers a compelling defense of the radical (or Marxist) paradigm, with specific reference to its contemporary and emerging relevance in the context of globalization. The point is that these interparadigmatic debates over theory and methods, all of which take place largely within the same epistemic community, are not likely to end very soon, if at all. Intraepistemic Debates—Critical Theories and Alternative Perspectives (figure 1, section 2) Critical theorists, postmodernists, feminists, social constructivists, and proponents of alternative perspectives all have their own internal paradigmatic debates—Smith (“Alternative and Critical Perspectives”) cites fundamental disagreements between Robert Cox, Rick Ashley, and Rob Walker to illustrate the point: Page 678 → Within the alternative camp, there is some debate between rival perspectives, but frankly there is little agreement on epistemological matters. Thus critical theory requires exactly the kind of foundational epistemology that postmodernism denies is possible, just as standpoint feminism stands in contrast to postmodern feminism…. Within each of the alternative perspectives there is lively debate around a series of research questions, but these cannot be expected to add up to one alternative perspective to challenge the mainstream. Just as it is naive to ask for the feminist perspective on an issue, so is it misguided to ask for the alternative perspective on an issue. Pettman (“Critical Paradigms in International Studies: Bringing It All Back Home?”) makes a similar argument regarding feminist theory: Feminism appears to have had comparatively little effect on IS judging by most course offerings, articles or chapters, or references in many IS texts. But while the center holds, feminism does have a remarkably robust, diverse, and productive presence in and around IS.… Feminism in IS is becoming more international and more representative. Indeed, it is broad and rich enough to raise questions about its own centers and margins. There are important questions here about who “we” are, including in terms of background, generation, and current location, both geographic and institutional (each shorthand for very complicated feminist identity politics). This range, diversity, and at times discomfort provide the context within which we ask what it means to read, write, converse, and organize as feminists, in and around IS. Zalewski (“Feminism and/in International Relations: An Exhausted Conversation? or Feminists Doing International Relations: The Cut(ting) Edge of Contemporary Critical Theory and Practice?”) expands on these same observations— “The key debates [in mainstream IS] are well defined and indeed well protected. Scholars may disagree over the most appropriate approach to dealing with interstate relationships and have vigorous debates about this. But the value and centrality of the debate itself is well protected by the very selfdefinition of the discipline. This is not the case for those who work on feminism and gender.” Scholars within critical and alternative traditions share a common objective in challenging mainstream

assumptions, theories, and power. Yet housed within this epistemic community are theoretical and methodological debates that are as intractable and irreconcilable as those in the mainstream and, likewise, show no clear signs of resolution. Page 679 → Interepistemic Debates—Mainstream versus Critical Perspectives (figure 1, section 3) To the extent that these competing epistemic communities (mainstream and critical perspectives) continue to expand rather than select, the prospects for meaningful interepistemic dialogue are not encouraging. How exactly do scholars in the mainstream engage in a dialogue with critics who interpret everything the mainstream produces as either trivial, time specific, gender biased, or socially constructed, regardless of paradigm, subfield, theory or methodology? Peterson's contribution (“On the Cut(ting) Edge”) includes reference to many of these irreconcilable differences: Critiques of positivist empiricism have altered our understanding of agents, identity, and subjectivity. In contrast to the modernist conception of a unitary rational actor, contemporary social theory illuminates the multiplicity of subject locations (implying multiple identifications) and their dynamic interaction “within” the self and in relation to the self's environment. In short, identities are socially constructed as ongoing processes: they are embedded in and interact with historically specific social contexts composed of intersubjective meaning systems, practices, institutional structures, and material conditions. And, Statecentric political identity increasingly shares the stage with non-territorial claimants. Sub-and transnational social movements transgress territorial boundaries in favor of identities “grounded” on ecological, antinuclear, ethnic, feminist, religious, and other non-state-based commitments. Uneven flows of information, culture, people, technology, and capital involve human migrations and diasporic identities that complicate territorial frames of reference. More specific (but no less significant) differences between mainstream IR and feminist research are covered in excellent contributions by Pettman (“Critical Paradigms in International Studies: Bringing It All Back Home?”), Ling (“The Fish and the Turtle: Multiple Worlds as Method”), and Zalewski (“Feminism and/in International Relations: An Exhausted Conversation? or Feminists Doing International Relations: The Cut(ting) Edge of Contemporary Critical Theory and Practice”). Kolodziej (“Security Theory: Six Paradigms Searching for Security”) addresses debates between the mainstream and social constructivism: Constructivism has no clear theory agency in the way that rational actor or realist or liberal economic theory possesses. The former can assume that actors engaged in surviving and in pursuing their security interests will always prefer more power or gain than less or, if a liberal Page 680 → institutionalist, more absolute power and gain than less…. Constructivists do not have a fixed point from which to develop their research designs and findings. They do not posit a maximizing utility interest that orients actor behavior and dictates strategies and choices. The notion of utility itself is a property of identity and the interests flowing from identity. … The constructivist research program has widened the scope of theoretical and empirical research and underscored the discrepant outcomes of actor behavior that need to be reconciled with the countervailing research assumptions and paradigms of other approaches. Perhaps the most puzzling feature of many of these interepistemic debates is the assumption held by critics, such as Smith (“Alternative and Critical Perspectives”), that key debates “have been very largely solved within the mainstream.” Positivist methods, he argues, “become the standards for assessing work and are never seen as

themselves open to critique or question. They are presented as if they were beyond reproach, beyond debate, and beyond the realm of politics. They are the standards of the academy, and that is that.” There are at least three obvious problems with this line of argument. First, it ignores (perhaps intentionally) the many complex and unresolved debates within the mainstream. Worse, it assumes these intraepistemic debates are trivial, superficial, or entirely irrelevant to the search for “real” answers to fundamental questions about the international system. The default “critical” position, then, is simply to assume the mainstream is characterized by overwhelming consensus (and, by extension, intellectual laziness), not because it is true (it is obviously not) but because that assumption serves to justify the need for an “alternative” school. As Ray (“Reflections on Millennia, Old and New: The Evolution and Role of Quantitative Approaches to the Study of International Politics”) correctly observes: Somehow, second-wave specialists have convinced third-wave specialists … that we consider ourselves capable of being entirely objective, above prejudice, that what it means to be scientific, in our view, is to be sufficiently detached to allow the data to speak for themselves. Perhaps some second-wave specialists even believe that. Perhaps in the coming new millennium specialists making up the second wave can be more clear about the ultimate aims of their research endeavors and dispel this notion that to be scientific means somehow to be “above” emotions, commitment, or ideological convictions. Many of the most important, constructive, and potent criticisms of mainstream work are made by mainstream theorists, though acknowledging this fact is not particularly conducive to creating demand for a “critical” approach. Second, and perhaps more important, the epistemological preferences and goals of critical theorists are often indistinguishable from those of mainstream Page 681 → theorists. Smith's arguments in “Alternative and Critical Perspectives” help to illustrate this point: The problem here for alternative and critical approaches is the familiar one of not being willing to buy into [the mainstream] model of assessment. This is clearly a fundamental difference between the mainstream and the alternative and critical literatures, and I see no easy way to bridge it. My own view, for what it is worth, is that the acid test for the success of alternative and critical approaches is the extent to which they have led to empirically grounded work that explores the range and variety of world politics,” (emphasis added) The standards Smith prefers “tend to revolve around the categories of coherence, tight argumentation, knowledge of the relevant literature, the use of empirical evidence, persuasiveness of the major analytical moves, innovative scholarship, and the standard of writing and expression.” But how exactly are any of these objectives distinct from those in the mainstream? Unfortunately, Smith provides no guidance with respect to requisite criteria for making these judgments: “I think the way forward is to insist only on research being carried out to the highest standards possible as de- fined within the research area.” Smith, like any editor of a major academic series, wants to publish “the best postmodern feminist work, the best neorealist, the best quantitative work, and so on…. I hope that theoretical work, of whatever persuasion, should be informed by, and in turn inform, empirical material.” All of this, of course, begs the following questions: Which realist, neorealist, feminist, postmodern, institutionalist, or constructivist work is the “best” in that tradition? How does one go about making these judgments, especially when those within these traditions have such a hard time deciding what the best work is? Which positivist, formal, statistical, qualitative, or postpositivist methodologies are the best suited to generate empirically grounded research, and how do we know? How do we evaluate different standards for empirically (as opposed to nonempirically) grounded truth claims, or the importance of the findings we derive from any of this work? And, finally, how does one answer any of these questions without becoming entangled in the hundreds (thousands?) of methodological and theoretical debates that continue to play out in mainstream literature and journals, the same debates that have confounded international theorists since Thucydides?

Conclusions We conclude with five general observations about progress in international studies and the Millennial Reflections Project more generally. First, new debates, perspectives, theories, and approaches are proliferating much faster than old debates are being resolved—indeed, few if any of the “old” debates have ever been resolved. To the extent that consensus exists at all, it usually emerges in the context of narrowly-defined research Page 682 → programs encompassing small communities of scholars who focus on issues at lower ends of the epistemological ladder (see figure 1). Second, if we haven't yet achieved closure on key theoretical and methodological debates, we never will. On this point it is interesting to note that contributions to a 1972 symposium on cumulation in IS arrived at the same conclusion—expansive, not selective, cumulation is the defining characteristic of the discipline. Thirty years of research in IS has done little to change this fact, and another thirty years of work will have the same effect (see James Rosenau's In Search of Global Patterns [New York: Free Press, 1976]). Third, for those who remain convinced that constructive interepistemic dialogue and consensus is still possible (notwithstanding the arguments outlined herein), our most discouraging observation is that there are no solutions. Consider for example the recommendations by Haas and Haas (“How We Learned to Escape Physics Envy and to Love Pluralism and Complexity”) and Lapid (“En Route to Knowledge: Is There a “Third Path” (in the Third Debate)?”) on how to make the debates more constructive and move the field forward. The most common suggestion is to search for areas of agreement, establish and encourage “middle-ground” positions, and avoid what Lapid refers to as “flabby pluralism.” For Haas and Haas, progress implies the legitimacy of a plurality of approaches [and] tolerance for competing ontologies and epistemologies. However, we are not being tolerant for the sake of being tolerant, of being nice to our intellectual antagonists….The achievement of modest progress depends on continuing the conversation among all of us.… Nobody finally wins or loses in this perspective. The grand theoretical stakes are low because any cumulating that may be achieved is temporally and spatially limited and unlikely to challenge any deep commitments. But a few puzzles get solved along the way. Similarly, Lapid suggests that the IR scholarly community would be well advised to … proceed with a serious and reflective investigation of the “middle” of the third debate. Unfortunately, with few exceptions, this is not the direction in which the rising IR interest in the median space is currently evolving. International relations' via media may be “under construction,” but there is little evidence of a plausible and imaginative architectural design to guide a productive implementation of such a critical disciplinary project. But the call to proceed with serious reflections on middle-ground positions is based on a somewhat simplistic interpretation of the nature and complexity of core debates within IS. These and similar recommendations assume, for example, that there is sufficient agreement within perspectives, subfields, paradigms, and epistemic communities to facilitate meaningful and constructive dialogue in the first place. But without sufficient agreement within these communities, where exactly does one begin the search Page 683 → for the “middle”? Even if we assume sufficient consensus within communities, there is no guarantee that reflective engagement of the middle will move us in the right direction—logically, concessions by proponents of any two perspectives will, by definition, produce more distance between them and a third perspective. For example, concessions between realists and neoliberals on the structural determinants of “state” versus “institutional” power will create more distance between this middle position and, say, constructivism or feminism. The question is not how to identify a middle, but how to facilitate simultaneous movement toward multiple middle grounds across multiple debates at multiple levels of the epistemological ladder. That is the daunting challenge we face. Fourth, self-critical reflection does not come easily to most scholars. Although we asked all contributors to avoid

the tendency to define accomplishments with reference to the failures and weaknesses of other perspectives, few contributors were prepared to do this. Lake's explanation in “Progress in International Relations: Beyond Paradigms in the Study of Institutions” is informative here: Paradigm warriors are particularly prone to argue about the truth status of assumptions, to defend islands at the core of their approach rather than build bridges to those occupied by others, and to neglect research design in hopes of scoring debating points. In the continuing debate with neorealists on whether institutions “matter,” institutionalists have continued in this not-so-grand tradition.… We are stuck in a convention “trap,” where however much we may individually abhor some of the standards and practices in our field we are nonetheless compelled to perpetuate them for our own success and that of our graduate students. Perhaps the best approach, as Lake points out, is to “continuously ask paradigm warriors the important question, ‘What have you helped us understand lately?’” The norm, however, is to focus on what other perspectives have failed to accomplish in this regard. This more than any other fact explains the propensity toward expansive rather than selective cumulation in IS. Our final observation addresses Rosenau's criticisms (“The Globalization of Globalization”) of the entire Millennial Reflections Project: As one whose professional life exceeds that of the International Studies Association, I have mixed feelings about the invitation to reflect critically on the accomplishments, failures, debates, standards, approaches, and future agendas of international relations (IR) and its sub-fields. On the one hand, I have no doubt that our shared past is marked by enormous growth and progress. The conceptual and methodological equipment with which IR is probed today is far more elaborate, incisive, and diverse than was the case at the outset…. On the other hand, for all the progress that has marked IR's evolution, I am plagued Page 684 → with doubts about the field's capacity for adapting to the transformations at work in the world today. And my doubts extend to questioning the wisdom of engaging in introspective assessments of inter-and intraparadigmatic debates…. To articulate such concerns struck me as needless ritual that diverted valuable time and space from focusing on what was transpiring outside academe in what students call “the real world.” … Put differently, exercises like the present one can be a substitute for thinking about the nature of the world and how it should be studied, and I think that too often we resort to this substitute when the course of events seems too complex and obscure to comprehend…. In sum, I find myself persuaded that the future of our field, its capacity to confront the huge conceptual and methodological challenges that have eluded our generation, lies in the training of those who will enter the field in the future…. We need to acknowledge our own limitations and alert those we train to the necessity of their breaking with past paradigmatic assumptions and finding new ways of understanding and probing the enormous challenges posed by the dynamics of change, fragmegration, and micro-macro interactions. Rosenau's assertions beg the following questions: What precisely do we tell our graduate students to keep or discard, and how do we make those decisions? What is “the real world,” and how should it be studied? The debate continues.

Page 685 →

ABOUT THE CONTRIBUTORS Davis B. Bobrow Davis B. Bobrow is a professor of public and international affairs and political science at the University of Pittsburgh. He has been president of the International Studies Association, a member of the Defense Science Board, a councillor of the Peace Science Society (International), and an official in the Office of the Secretary of Defense. His recent writing deals with international futures, alternative security perspectives, regional integration arrangements, East Asia in international affairs, and policy design.

Steven J. Brams Steven J. Brams is a professor of politics at New York University. He is the author or coauthor of fourteen books that involve applications of game theory and social choice theory to voting and elections, bargaining and fairness, international relations, and the Bible and theology. His most recent books are Theory of Moves (1994); Fair Division: From Cake-Cutting to Dispute Resolution (1996, with Alan D. Taylor); and The Win-Win Solution: Guaranteeing Fair Shares to Everybody (1999). He is a fellow of the American Association for the Advancement of Science and the Public Choice Society, a Guggenheim fellow, a past president of the Peace Science Society (International), and in 1998–99 was a visiting scholar at the Russell Sage Foundation.

Michael Brecher Michael Brecher is the R. B. Angus Professor of Political Science at McGill University. Educated at McGill and Yale (Ph.D., 1953), he is the author or coauthor of eighteen books and eighty-five articles on India-South Asia, international systems, foreign policy theory and analysis, international crises, conflict and war, and the IndoPakistani and Arab-Israel protracted conflicts. Since 1975, he has been director of the International Crisis Behavior Project. Page 686 → His most recent books are Crises in World Politics (1993) and A Study of Crisis (1997, 2000, with Jonathan Wilkenfeld). He has received two book awards: the Watumull Prize of the American Historical Association in 1960 for Nehru: A Political Biography (1959) and the Woodrow Wilson Award of the American Political Science Association in 1973 for The Foreign Policy System of Israel: Setting, Images, Process (1972). Among his other awards are the Fieldhouse Award for Distinguished Teaching, McGill (1986); the Distinguished Scholar Award of the International Studies Association (1995); the Leon-Gerin Prix du Québec for the human sciences (2000); and the Award for High Distinction in Research, McGill (2000). He has been a fellow of the Royal Society of Canada since 1976 and has held fellowships from the Nuffield, Rockefeller, and John Simon Guggenheim foundations. He has been a visiting professor at the University of Chicago, the Hebrew University of Jerusalem, University of California, Berkeley, and Stanford University. In 1999–2000 he served as president of the International Studies Association.

Bruce Bueno de Mesquita Bruce Bueno de Mesquita is a senior fellow at the Hoover Institution at Stanford University and visiting professor of politics at New York University. Bueno de Mesquita holds a Ph.D. in political science from the University of Michigan and an honorary doctorate from the University of Groningen. He is president of the International Studies Association for 2001. His most recent books include Principles of International Politics (1999); Governing for Prosperity (2000); and the novel The Trial of Ebenezer Scrooge (2001).

Michael Cox

Professor Cox is currently editor of the Review of International Studies. He is an associate fellow of the Royal Institute of International Affairs in London and an executive member of the National Committee for the Study of International Affairs at the Royal Irish Academy in Dublin. The author, editor, and coeditor of more than twelve books, his most recent publications include Rethinking the Soviet Collapse (1998); The Interregnum: Controversies in World Politics, 1989–1999 (1999); E. H. Carr: A Critical Appraisal (2000); and American Democracy Promotion (2000).

Robert W. Cox Robert W. Cox is a professor emeritus of political science at York University, Toronto, Ontario. He is author of Production, Power, and World Order: Social Forces in the Making of History (1987) and Approaches to World Order (1996, with Timothy J. Sinclair). His recent work has been about civilizations. Page 687 →

Joseph M. Grieco Joseph M. Grieco is a professor of political science at Duke University. He is the author of Cooperation among Nations: Europe, America, and Non-Tariff Barriers to Trade (1990) and Between Dependency and Autonomy: India's Experience with the International Computer Industry (1984). Articles and notes by him have appeared in International Studies Quarterly, Security Studies, Review of International Studies, American Political Science Review, International Organization, Journal of Politics, and World Politics. His teaching interests include theories of international relations, issues of international political economy, international business-government relations, the relationship between international economics and international security, and the rise of the European nation-state. During the summer of 1994 he was the Karl W. Deutsch Visiting Professor at the Wissenschaftszentrum Berlin, and since May 1996 he has been a visiting professor at the Post-Graduate School of Economics and International Relations at the Catholic University of Milan.

Ernst B. Haas Ernst B. Haas is Robson Research Professor of Government Emeritus at the University of California, Berkeley, where he is also a professor in the graduate school. His most recent publication is the two-volume study Nationalism, Liberalism, and Progress (1997, 2000). His current research is concerned with the institutionalization of international norms.

Peter M. Haas Peter M. Haas is a professor of political science at the University of Massachusetts at Amherst. He received a Ph.D. in political science from the Massachusetts Institute of Technology. He has published widely on international environmental subjects, including pollution control in the Mediterranean, pollution control in the Baltic and North Seas, UNEP's regional seas programs, stratospheric ozone protection, and international environmental institutions. He has also published works on international relations theory, focusing on the interplay between knowledge and power in international policy coordination. His recent work has focused on the interplay between international institutions and scientific involvement in the creation and enforcement of international regimes addressing transboundary and global environmental risks. He is writing a book on the evolution of multilateral environmental governance since 1972. He is the author of Saving the Mediterranean: The Politics of International Environmental Cooperation (1990), editor of Knowledge, Power, and International Policy Coordination (1997), and coeditor of Institutions for the Earth: Sources of Effective International Environmental Protection (1993). He has written thirteen articles published in journals such as International Page 688 → Organization, Millennium, Global Governance, Environment, Environmental Conservation, and Global Environmental Change and fourteen chapters appearing in

eleven edited volumes. He has consulted for the Commission on Global Governance, the United Nations Environment Programme, the U.S. Department of State, the U.S. Environmental Protection Agency, the U.S. National Academy of Sciences, the American Association for the Advancement of Science, and the World Resources Institute. He has received grants from the National Science Foundation, German Marshall Fund, Rockefeller Brothers Fund, Institute for the Study of World Politics, and the Gallatin Foundation.

Frank P. Harvey Frank P. Harvey is director of the Centre for Foreign Policy Studies at Dalhousie University. He is also a professor of political science at Dalhousie. His current research interests include ethnic conflict in the former Yugoslavia, NATO military strategy and peacekeeping, and national missile defense. His books include The Future's Back: Nuclear Rivalry, Deterrence Theory, and Crisis Stability after the Cold War (1997); Conflict in World Politics: Advances in the Study of Crisis, War, and Peace (1998, coedited with Ben Mor); and Using Force to Prevent Ethnic Violence: An Evaluation of Theory and Evidence (2000, with David Carment). He has published widely on nuclear and conventional deterrence, coercive diplomacy, crisis decision making, and protracted ethnic conflict in such periodicals as International Studies Quarterly, Journal of Conflict Resolution, Journal of Politics, International Journal, Security Studies, International Political Science Review, Conflict Management and Peace Science, Canadian Journal of Political Science, and several others. Professor Harvey is currently working on his next book, Coercive Diplomacy and the Management of Intrastate Ethnic Conflict.

Kal J. Holsti Kal J. Holsti is Killam Professor of Political Science Emeritus at the University of British Columbia, Vancouver, Canada. He is presently a research associate with the Institute of International Relations at the same university. He is the author of seven books in the field, including The State, War, and the State of War (1996). He has contributed regularly to scholarly periodicals and has chapters in twenty-two edited volumes. He has been a visiting professor in the United States, Australia, Israel, and Japan. He is a past editor of International Studies Quarterly and the Canadian Journal of Political Science. In 1984–85, he was president of the Canadian Political Science Association and, two years later, was president of the International Studies Association. He was elected a fellow of the Royal Society of Canada in 1983. Page 689 →

Ole R. Holsti Ole R. Holsti has been on the faculties of Stanford University, the University of British Columbia, and, since 1974, Duke University as George V. Allen Professor of International Affairs. He is a former president of the International Studies Association and has received career awards from the ISA, the International Society for Political Psychology, and the American Political Science Association. He has also received two university-wide teaching awards. Holsti's most recent book is Public Opinion and American Foreign Policy (1996).

Patrick James Patrick James is a professor of political science at the University of Missouri. He is the author of eight books and more than eighty other scholarly publications. Among his honors and awards are the Louise Dyer Peace Fellowship of the Hoover Institution at Stanford University, the Milton R. Merrill Chair in Political Science at Utah State University, the Lady Davis Professorship in the Department of International Relations at the Hebrew University of Jerusalem, the Thomas O. Enders Professorship in Canadian Studies in the Department of Political Science at the University of Calgary, and a listing among the “2000 Outstanding Scholars of the 21st Century”

produced by the International Biographic Centre in Cambridge, England. James currently serves as editor of International Studies Quarterly and is on the governing councils of the Association for Canadian Studies in the United States and the International Studies Association.

Robert O. Keohane Robert O. Keohane received a Ph.D. from Harvard University. He is the James B. Duke Professor of Political Science at Duke University. He has taught at Swarthmore College, Stanford University, Brandeis University, and Harvard University, where he was the Stanfield Professor of International Peace. He is the author of After Hegemony: Cooperation and Discord in the World Political Economy (1984), for which he was awarded the second annual Grawemeyer Award in 1989 for Ideas Improving World Order. He is also the author of International Institutions and State Power: Essays in International Relations Theory (1989); coauthor of Power and Interdependence: World Politics in Transition (1977, 1988, with Joseph S. Nye Jr.); and coauthor of Designing Social Inquiry: Scientific Inference in Qualitative Research (1994, with Gary King and Sidney Verba). He is editor or coeditor of, and contributor to, nine other books. Between 1974 and 1980 he was editor of International Organization. He was president of the International Studies Association in 1988–89. He is a fellow of the American Academy of Arts and Sciences and has held a Guggenheim Fellowship and fellowships at the Center for Advanced Study in the Behavioral Sciences and the National Humanities Center. Page 690 →

Edward A. Kolodziej Edward A. Kolodziej is a research professor of political science and former head of the Political Science Department at the University of Illinois, Urbana-Champaign; director of the Ford Foundation Interdisciplinary Seminar on the World Society; and cofounder and first director of the Program in Arms Control, Disarmament, and International Security. He received a Ph.D. from the University of Chicago. He is the recipient of grants and awards from the Ford, Rockefeller, and MacArthur foundations, the National Science Foundation, the National Endowment for the Humanities, NATO, the Woodrow Wilson Center, Fulbright, and the U.S. Institute of Peace. He has written or edited ten books on security and foreign policy including The Uncommon Defense and Congress: 1945–63 (1966); French International Policy under De Gaulle and Pompidou: The Politics of Grandeur (1974); and Making and Marketing Arms: The French Experience and Its Implications for the International System (1987). The latter was cited by Choice as one of the major contributions to political science in 1987. Professor Kolodziej has also contributed more than one hundred articles to professional journals as well as chapters to edited volumes. His latest major publication is an edited volume, Coping with Conflict after the Cold War (1996), which develops global strategies for reducing armed conflict. He is currently working on three projects: a theory of global governance, a critical evaluation of contemporary security theory, and an edited volume on the global politics of human rights.

Louis Kriesberg Louis Kriesberg (Ph.D., University of Chicago, 1953) is currently doing comparative research on reconciliation and the changing accommodations between ethnic, religious, and other communally identified groups. He is a professor emeritus of sociology, the Maxwell Professor Emeritus of Social Conflict Studies, and the founding director of the Program on the Analysis and Resolution of Conflicts at Syracuse University. His writings include Constructive Conflicts (1998); International Conflict Resolution (1992); Social Conflicts (1973, 1982); and numerous chapters and articles. He coedited Intractable Conflicts and Their Transformation (1989) and Timing the De- Escalation of International Conflicts (1991).

Robert T. Kudrle Robert T. Kudrle is a professor of public affairs and law at the Hubert Humphrey Institute of Public Affairs and

the Law School, University of Minnesota. He studies industrial organization, public policy toward business, and international economic policy. He has served as associate dean for research at the Humphrey Institute and as director of the master's degree in public affairs. Kudrle has consulted for many organizations including the Antitrust Page 691 → Division of the U.S. Department of Justice, the Internal Revenue Service, the Urban Institute, the Canadian Department of Consumer and Corporate Affairs, the UN Center on Transnational Corporations, the Overseas Private Investment Corporation, and the Agency for International Development. His work has appeared in such journals as World Politics, International Organization, Journal of Conflict Resolution, Transnational Corporations, Public Finance, Journal of Law and Economics, and Canadian Journal of Economics. Kudrle is past coeditor of International Studies Quarterly and serves on its editorial board. He also has been on the editorial boards of Journal of Health Politics, Policy and Law, International Interactions, International Political Economy Yearbook, and Minnesota Journal of Global Trade. A Rhodes scholar, Kudrle holds a master of philosophy degree in economics from Oxford University and a doctorate in economics from Harvard University.

David A. Lake David A. Lake (Ph.D., Cornell University, 1984) is a professor of political science at the University of California, San Diego, chair of the department, and a coeditor of the journal International Organization. He taught at UCLA from 1983 to 1992. From 1992 to 1996, he served as research director for international relations at the UC-wide Institute on Global Conflict and Cooperation. Lake's areas of interest include international political economy, conflict studies, American foreign policy, and global governance. He has published widely in the field of international relations, including Power, Protection, and Free Trade: International Sources of U.S. Commercial Strategy, 1887–1939 (1988) and Entangling Relations: American Foreign Policy in Its Century (1999). He is also editor or coeditor of International Political Economy: Contending Perspectives on Global Power and Wealth (four editions); The State and American Foreign Economic Policy (1988); The International Political Economy of Trade (1993); Regional Orders: Building Security in a New World (1997); The International Spread of Ethnic Conflict: Fear, Diffusion, and Escalation (1998); and Strategic Choice and International Relations (1999). His current research focuses on democracy and its policy consequences within and between states.

Yosef Lapid Yosef Lapid is a professor of government at New Mexico State University. His research interests include international relations theory, culture and identity, and border studies. He is coeditor of The Return of Culture and Identity in International Relations (1996, with F. Kratochwil) and Identities, Borders, Orders: Rethinking International Relations Theory (2001, with M. Albert and D. Jacobson). Page 692 →

Russell J. Leng Russell J. Leng is the James Jermain Professor of Political Economy and International Law at Middlebury College. His recent works include Bargaining and Learning in Recurring Crises: The Soviet-American, EgyptianIsraeli, and Indo-Pakistani Rivalries (2000) and Interstate Crisis Behavior, 1816–1980: Realism vs. Reciprocity (1993).

Jack S. Levy Jack S. Levy is a Board of Governors Professor of Political Science at Rutgers University. He is the author of War in the Modern Great Power System, 1495–1975 and numerous articles on the causes of war and on foreign policy decision making. His current research projects include preventive war, great-power balancing behavior, the

relationship between economic interdependence and international conflict, the militarization of commercial rivalries, diversionary theory and politically motivated opposition to war, prospect theory and international relations, and the methodology of qualitative research.

L. H. M. Ling L. H. M. Ling is a senior lecturer in international studies at the Institute of Social Studies, The Hague. Her work applies a postcolonial-constructivist model of international relations to a wide range of issues: democratization, globalization, media images, urban design, peace and conflict studies, development and politics in East Asia, and transnational scholarship in the Western academy. Dr. Ling is the author of Postcolonial IR: Conquest and Desire between Asia and the West (2001).

Zeev Maoz Zeev Maoz is a professor of political science at Tel Aviv University. He is the author of several books, including National Choices and International Processes (1990); Paradoxes of War: On the Art of National Self-Entrapment (1990); Domestic Sources of Global Change (1996); and Bound by Struggle: The Strategic Evolution of Enduring International Rivalries (2002, with Ben D. Mor). His articles on international conflict, decision making, and international bargaining have been published in American Political Science Review, World Politics, Journal of Conflict Resolution, International Studies Quarterly, and other journals.

Lisa L. Martin Lisa L. Martin is a professor of government in the government department at Harvard University. She is a member of the executive committee of the Weatherhead Center for International Affairs. Her areas of teaching and research Page 693 → include the impact of institutions upon states' abilities to make credible commitments, the European Union, U.S. foreign policy, and legislatures in international cooperation. She has recently presented the following papers: “Agency in Delegation in IMF Conditionality” (with Devesh Kapur); “Governance Patterns in Tourism: The Leverage of Economic Theories” and “An Institutionalist View: International Institutions and State Strategies.”

John Mearsheimer John Mearsheimer teaches at the University of Chicago. He received a Ph.D. from Cornell University in 1980. He specializes in international relations theory, national security policy, war and international systems, and deterrence theory. He is the author of Conventional Deterrence (1983) and Liddell Hart and the Weight of History (1989). He is the director of the Workshop on International Security Policy.

Manus I. Midlarsky Manus I. Midlarsky is the Moses and Annuta Back Professor of International Peace and Conflict Resolution at Rutgers University. His most recent book is The Evolution of Inequality: War, State Survival, and Democracy in Comparative Perspective (1999), and he recently edited the volumes Inequality, Democracy, and Economic Development (1997) and Handbook of War Studies II (2000). His most recent articles have appeared in American Political Science Review, Journal of Politics, International Studies Quarterly, Journal of Conflict Resolution, and Journal of Peace Research. Currently he is working on a book-length analysis of genocide titled The Killing Trap: Genocides and Other Mass Murders of the Twentieth Century.

Linda B. Miller Linda B. Miller is a professor of political science at Wellesley College and a senior fellow of the Watson Institute at Brown University, where she edits International Studies Review under the sponsorship of both institutions. She previously taught or held research appointments at Barnard College, Harvard University, Princeton University,

Columbia University, and the Woods Hole Oceanographic Institution. She has published widely on American foreign policy, international organizations, European politics, and the Middle East in a variety of U.S., British, and Israeli journals and monograph series on grants from the Council on Foreign Relations, the Rockefeller Foundation, NATO, the Sloan Foundation, the Pew Charitable Trusts, and the Tami Steinmetz Center for Peace Research at Tel Aviv University. Her books include World Order and Local Disorder: The United Nations and Internal Conflict and Ideas and Ideals: Essays on Politics in Honor of Stanley Hoffmann. Page 694 →

Helen Milner Helen Milner has been a professor of political science at Columbia University since 1995. She was a research fellow at the Brookings Institution, 1983–1984, and held an advanced research fellowship in foreign policy studies at the Social Science Research Council, 1989–1991. She has been editor of Review of International Political Economy from 1996 to the present. Her research interests include international politics, international political economy, and West European politics. Her publications include Interests, Institutions, and Information: Domestic Politics and International Relations (1997); The Political Economy of Economic Regionalism (1997); Internationalization and Domestic Politics (1996); Resisting Protectionism: Global Industries and the Politics of International Trade (1988); “Democratic Politics and International Trade Negotiations: Elections and Divided Government as Constraints on Trade Liberalization,” Journal of Conflict Resolution, 41, no. 1 (1997); “Trade Negotiations, Information, and Domestic Politics,” Economics and Politics, 8, no. 2 (1996) (with B. Peter Rosendorff); “Resisting the Protectionist Temptation: Industry and the Making of Trade Policy in France and the U.S. during the 1970s,” International Organization, 41, no. 4 (1987); and “International Theories of Cooperation among Nations: A Review Essay,” World Politics, 44, no. 3 (1992).

Michael Nicholson Editors note: It is with sincere regret that the editors acknowledge the passing of Professor Michael Nicholson. A biography of Dr. Nicholson's work was compiled for this volume, but the editors felt it would be a fitting, and certainly more permanent, reminder of Professor Nicholson's contributions to international studies to include the following obituary written by lifelong friend and colleague Professor Amartya Sen. The obituary was published in The Independent newspaper October 17, 2001, and is reprinted here with their kind permission. When the European Consortium of Political Research awarded to Michael Nicholson the first Lewis Fry Richardson Award for Lifetime Achievement (Fall, 2001), he was already very ill; he died on 4 October (2001). The date of the award, 11 September 2001, would have appeared to Nicholson to be oddly significant, as he had worked all his life on conflicts and wars, and on the possibility of peaceful resolution of disputes. From his first book, Conflict Analysis (1971), through his classic monograph, Rationality and the Analysis of International Conflict (1992), to his more recent writings on international relations, he made wonderful use of his skills and expertise as a social scientist to explore how conflicts can be avoided or defused, without resorting to violence and bloodshed. The events of 11 September would have depressed him (as would Page 695 → the developments that followed), but, despite his great belief in the power of rationality, he would not have been altogether surprised by them. He had written about the limited use of reasoning and communication both in practical lives and in the theory of rational choice. Indeed, in a joint paper published earlier in 2001, Nicholson discussed the limitations from which the widely used instrumental view of rationality suffered. Choosing the format of a self-deprecating illustration (a style of reasoning that used to be more common in England than it is now), he argued, “Given a belief that I shall go to Paradise if I kill the

enemies of England, the act of killing can be construed as rational when otherwise it would not be.” He went on to discuss other ways of seeing rationality, and how their relevance can be made more widely appreciated and pursued, and how rationality can be used to promote peace rather than war. Nicholson's ideas were becoming even more relevant to the contemporary world just as he was physically leaving us. He was born in Yorkshire in 1933, son of Ernest and Irene Nicholson, and educated at Beverley Grammar School. I first met him in the autumn of 1953 shortly after we had both arrived as undergraduates at Trinity College, Cambridge, to do Economics. I came to know him very well only in my second year. By then, not only had he established himself as a brilliant student, but also his long-run interests (including rationality and conflict resolution) were already beginning to take shape. He was particularly interested in rational decision-making under uncertainty—a subject in which he would, later on, do his Ph.D. thesis and publish his earliest papers. Apart from his warmth and kindness as a friend, I was struck by his combination of extraordinary intelligence, healthy scepticism of received doctrines, and humane commitment to use his professional training to make a positive difference to the world. I got a clearer idea of the origin of that commitment when I went with him to his home, in Beverley, near York, and met his parents. It was wonderful to hear the unalloyed voice of Christian humanism (aside from discovering how delicious Yorkshire pudding can be). Nicholson absorbed and developed that humanism, shorn of religion (he remained an atheist throughout his adult life). Indeed, over the years he came to contrast reliance on reason with dependence on religious faith. “Once something is confined to being an Act of God,” he wrote five years ago, “then there is nothing much one can do about it except pray.” “As I am sceptical of the power of prayer,” he went on to say, “I put my hopes for some improvement in the human condition on the rigorous development of social scientific theory, and its testing by statistical methods, as a means of illuminating and hence perhaps influencing the future of society.” That exacting programme summarises Nicholson's basic commitment rather better than any other description I can provide. Page 696 → After Cambridge, he went to teach at Manchester University and then abroad—to Massachusetts Institute of Technology, Carnegie Institute of Technology (now called Carnegie-Mellon University) and the University of Stockholm. On his return to England, he took up a post at Lancaster University, and then one at University College London, before becoming in 1970 the Director of the Richardson Institute for Conflict and Peace Research, based in London and later in Lancaster. This was an excellent fit, given his strong commitment to this area of research. Aside from developing a fine institution, Nicholson also published several of his major works during his 12 years at the institute. Then he went abroad again, to the University of Texas at Austin, to Carleton University in Canada, to the Netherlands Institute for Advanced Study (NIAS) in Holland, and to Yale University. In 1990, he became Professor of International Relations at the University of Kent at Canterbury, and finally joined Sussex University as its Professor of International Relations in 1993, and remained attached to that university until his death. Nicholson married Christine Love in 1958. Christine, who too was a brilliant student, did English and Philosophy at Bristol University, and then did a master's degree on Kantian philosophy at Cambridge. She became involved in adult education and had a long academic career at the City Lit in London, until 1996; she died two years later. Although not particularly homebound, she was fascinated by the idea, against the current, of being a housewife, perhaps a “philosopher-housewife.” In the author

description that went with one of her papers in Philosophy, she insisted on being described, rather spectacularly, simply as “a Housewife.” (Michael wrote in his obituarial note on her that, “though ‘housewife’ was never much of a theme for Christine in practice,” she “bought many books on the subject.”) Christine and Michael had a wonderfully close and intellectually interactive married life. Having known Michael Nicholson closely from his undergraduate days, I am struck by the extent to which he managed to combine his intellectual discipline with his moral and political commitments. His was a life dedicated to humanity, but also lit by intelligence and guided by analysis. He remained actively engaged in his work till his death. In a joint paper with Karin Fierke (with whom he developed a personal relationship after Christine's death), published this year, he outlined an ambitious programme of integrating the formal demands of rationality with “the larger argument about rules, language and context.” A poet from Nicholson's college, John Dryden, has talked about the man who “trudged along unknowing what he sought, /And whistled as he went, for want of thought.” Michael Nicholson's life is a lasting invitation to stop whistling and start thinking. Amartya Sen Page 697 →

Joseph S. Nye Jr. Joseph S. Nye Jr., the Don K. Price Professor of Public Policy, is dean of the Kennedy School. He returned to Harvard in December 1995 after serving as assistant secretary of defense for international security affairs, where he won two Distinguished Service Medals, and as chair of the National Intelligence Council. Nye joined the Harvard faculty in 1964, where he has served as director of the Center for International Affairs and associate dean of arts and sciences. From 1977 to 1979, Nye was deputy to the undersecretary of state for security assistance, science, and technology and chaired the National Security Council Group on Nonproliferation of Nuclear Weapons. A member of the editorial boards of Foreign Policy and International Security, he is the author of numerous books and more than a hundred articles in professional and policy journals. His most recent books are Understanding International Conflicts, 3d ed. (2000); the third edition of his classic study cowritten with Robert O. Keohane, Power and Interdependence (2000); and a coedited volume, Governance in a Globalizing World (2000). Nye received his bachelor's degree summa cum laude from Princeton University in 1958. He did postgraduate work at Oxford University on a Rhodes scholarship and earned a Ph.D. in political science from Harvard.

V. Spike Peterson V. Spike Peterson is an associate professor in the Department of Political Science with courtesy appointments in women's studies, comparative cultural and literary studies, and international studies at the University of Arizona, Tucson. She has held research fellowships at the Australian National University, University of Bristol, and Goteborg University. She is a past recipient of the MacArthur Foundation Research and Writing Grant and the author of numerous chapters and journal articles on the topics of feminist international relations theory, global political economy, nationalism, and critical postmodernist and feminist theory. She is the coauthor of Global Gender Issues (1993, with Anne Sisson Runyan) and editor of Gendered States: Feminist (Re)Visions of International Relations Theory (1992, 1999). She has published articles in Millennium, Women's Studies International Forum, and International Feminist Journal of Politics and numerous chapters in edited volumes. She is currently working on a book-length project, tentatively titled Rewriting (Global) Political

Economy as Reproductive, Productive, and Virtual (Foucauldian) Economies.

Jan Jindy Pettman Jan Jindy Pettman is a professor of international relations at Australian National University and director of the Women's Studies Centre. She is coeditor of the International Feminist Journal of Politics. She is the author of Worlding Women: A Feminist International Politics. She writes extensively on international domestic work, sex work, and globalization in Southeast Page 698 → Asia. Her current research project is “A Feminist Perspective on ‘Australia in Asia.’”

James Lee Ray James Lee Ray is a professor of political science at Vanderbilt University. He is the author of Democracy and International Conflict (1995, 1998) and Global Politics, 7th ed. (1998). He is also the author of articles in such periodicals as Annual Review of Political Science, British Journal of Political Science, Conflict Management and Peace Science, Journal of Conflict Resolution, Journal of Democracy, Journal of Peace Research, International Interactions, International Organization, International Studies Quarterly, and Political Research Quarterly, and in numerous edited volumes. He currently serves as the editor of the journal International Interactions.

James N. Rosenau James N. Rosenau holds the distinguished rank of university professor of international affairs at the George Washington University. Professor Rosenau has held a Guggenheim Foundation Fellowship and is a former president of the International Studies Association. His scholarship has focused on globalization, the dynamics of change in world politics, and the overlap of domestic and foreign affairs, resulting in more than thirty-five books and 150 articles. Some of his recent publications include Thinking Theory Thoroughly: Coherent Approaches to an Incoherent World (1995, 2000); Stability, Stasis, and Security: Reflections on Superpower Leadership (2000); Along the Domestic-Foreign Frontier: Exploring Governance in a Turbulent World (1997); Global Voices (1993); Governance without Government (1991); and Turbulence in World Politics: A Theory of Change and Continuity (1990).

J. David Singer J. David Singer teaches in the Department of Political Science at the University of Michigan. Much of Professor Singer's work falls within the Correlates of War Project, which he began in 1964 and which continues to the present day. His research interests include world politics/international relations, peace research, research methods, public policy and administration, and comparative government and politics. He has many publications including Nations at War: A Scientific Study of International Conflict, 1816–1990 (1998, 2000, with Daniel Geller); “Militarized Interstate Disputes, 1816–1992: Rationale, Coding Rules, and Empirical Patterns,” Conflict Management and Peace Science (1997, with Stuart Bremer and Daniel Jones); “The Correlates of War,” in Encyclopedia of Violence, Peace, and Conflict (1998); “Armed Conflict in the Former Colonial Regions: From Classification to Explanation,” in van de Goor et al., eds., Between Development and Destruction (1996); “Patterns of Alliance Commitment and the Page 699 → Risk of Armed Conflict Involvement,” in Schneider and Weitsman, eds., Enforcing Cooperation (1997, with Volker Krause).

Steve Smith Steve Smith is pro vice chancellor (academic affairs) and professor of international politics at the University of Wales, Aberystwyth, which he joined in 1992. He was previously a professor of international relations at the University of East Anglia and has also taught at the State University of New York, Albany. Since 1986 he has been the founding editor of the Cambridge University Press Studies in International Relations series. From 1995 to 2000 he was a member of the SSRC/MacArthur Committee on International Peace and

Security. He has been a member of the executive committee of the British International Studies Association since 1980 and has served on the executive committee (1991–92) and was then vice president (1992–93) of the International Studies Association (United States). He has recently been appointed to the governing council of the American Political Science Association section on International Politics and International History. In 1999 he was awarded the Susan Strange Award by the International Studies Association for his contribution to intellectual diversity in the study of international relations. He is the author of some eighty academic papers and chapters in major international journals and edited collections, and he is the author/editor of thirteen books, including Explaining and Understanding International Relations (1990, with Martin Hollis); International Relations Theory Today (1995, edited with Ken Booth); International Theory: Positivism and Beyond (1996, edited with Ken Booth and Marysia Zalewski); and The Globalization of World Politics (1997, edited with John Baylis 2001). His research interests involve the theory of security studies, contemporary international theory, and foreign policy analysis.

Harvey Starr Harvey Starr received a Ph.D. from Yale University in 1971. His research and teaching interests include theories and methods in the study of international relations, international conflict and war, geopolitics and diffusion analyses, and domestic influences on foreign policy (revolution; democracy). His current research activities involve the two-level analysis of security management, various approaches to the democratic peace and democratic diffusion, geopolitics and the use of geographic information systems for the study of proximity and contiguity. In 1991 he became editor of International Interactions, and he has served on numerous editorial boards. He has served as president of the Conflict Processes section of the American Political Science Association (1992–95) and as vice president of the American Political Science Association (1995–96). He also served on the National Science Foundation Advisory Panel for Page 700 → Political Science (1992–93). He previously taught at Indiana University, serving as chairman of the Department of Political Science from 1984 to 1989. In addition, he has been a visiting fellow in politics at the University of Aberdeen, Scotland (1971–72, 1978–79), as well as visiting lecturer (1985) and visiting researcher (1990) at the Centre for Defence Studies at the University of Aberdeen. In 1996 he was a visiting fellow in the Department of International Relations, Research School of Pacific and Asian Studies, Australian National University. He is currently interim chair of the Department of Government and International Studies and Dag Hammarskjöld Professor in International Affairs at the University of South Carolina. He is author or coauthor of eleven books and monographs and more than sixty journal articles or book chapters. His most recent books are Anarchy, Order, and Integration: How to Manage Interdependence (1997); Agency, Structure, and International Politics (1997, with Gil Friedman); The Understanding and Management of Global Violence: New Approaches to Theory and Research on Protracted Conflict (1999, edited volume); and World Politics: The Menu for Choice, 6th ed. (2000, with Bruce Russett and David Kinsella). In 1998 he won the University of South Carolina Russell Award for Research in Humanities and Social Sciences in recognition of outstanding research and scholarship.

Christine Sylvester Christine Sylvester is a professor at the Institute of Social Studies in The Hague. Her books include the forthcoming Feminist International Relations: An Unfinished Journey and Feminist Theory and International Relations in a Postmodern Era. Recent articles appear in Alternatives, Third World Quarterly, International Feminist Journal of Politics, Body and Society, and a variety of Africanist journals.

J. Ann Tickner J. Ann Tickner is a professor of international relations and director of the Center for International Studies at the

University of Southern California. She is the author of Self-Reliance versus Power Politics: The American and Indian Experiences in Building Nation States (1987); Gender in International Relations: Feminist Perspectives on Achieving Global Security (1992); and Gendering World Politics: Issues and Approaches in the Post-Cold War Era (2001); as well as numerous articles and book chapters on feminist international theory.

John Vasquez John Vasquez is a professor of political science at Vanderbilt University. He is the author or editor of numerous books and articles on conflict studies. Current publications include What Do We Know about War? (2000); The Power of Power Politics: From Classical Realism to Neotraditionalism Page 701 → (1998); Classics of International Relations, 3d ed. (1996, editor); Beyond Confrontation: Learning Conflict Resolution in the PostCold War Era (1995); “Uncovering the Dangerous Alliances, 1495–1980,” International Studies Quarterly (1998, with Douglas M. Gibler); “War Endings: What Science and Constructivism Can Tell Us,” Millennium (1997); “The Realist Paradigm and Degenerative versus Progressive Research Programs,” American Political Science Review (1997).

Yaacov Vertzberger Yaacov Vertzberger is a professor of international relations at the Hebrew University of Jerusalem. He held visiting teaching and research positions at Stanford University; McGill University; the Institute of International Studies at the University of California, Berkeley; the East-West Center, Honolulu; Tel-Aviv University; the IDF College of National Defense; the Rockefeller Bellagio Center; the Swedish Institute of International Affairs, Stockholm; the Graduate School of International Relations and Pacific Studies at the University of California, San Diego; the Netherlands Institute for Advanced Study; and the United Nations University, Tokyo. He was also a member of the Mershon Center's Research and Training Group on the Role of Cognition in Collective Decisionmaking at Ohio State University. Vertzberger is the author of numerous articles and books on decision making, policy formation, international conflict, and Asian security and politics. Among his books are Misperceptions in Foreign Policymaking: The Sino-Indian Conflict 1959–1962 (1984); The World in Their Minds: Information Processing, Cognition, and Perception in Foreign Policy Decisionmaking (1990); and Risk Taking and Decisionmaking: Foreign Military Intervention Decisions (1998).

R. B. J. Walker R. B. J. (Rob) Walker is a professor in the School of Politics, International Relations, and the Environment at Keele University in the United Kingdom. He has written widely on the implications of contemporary global transformations for modern accounts of sovereignty, subjectivity, and political practice. He is best known for his work Inside/Outside: International Relations as Political Theory (1993), and he is currently completing a new book, After the Globe/Before the World (2003). He is editor of the journal Alternatives: Global, Local, Political.

Stephen G. Walker Stephen G. Walker received a Ph.D. from the University of Florida in 1971. He has held academic appointments at the University of Florida, University of Minnesota, and Arizona State University, where he is presently a professor of political science. He has served as vice president of the International Society of Political Psychology, coeditor of International Studies Quarterly, Page 702 → and on the editorial boards of Political Psychology and International Interactions. His research on the operational codes of political leaders as they relate to problems of crisis management has been funded by the National Science Foundation. He has published in several academic journals, including World Politics, International Studies Quarterly, Journal of Conflict Resolution, Journal of Peace Research, International Interactions, Political Psychology, and Journal of Politics.

Jonathan Wilkenfeld Jonathan Wilkenfeld (Ph.D., political science, Indiana University, 1969) is professor and chair, Department of Government and Politics, University of Maryland. He is director of the ICONS International Simulation Project and the Conflict and Negotiation Project at the University of Maryland and codirector with Michael Brecher of the International Crisis Behavior Project. He is a specialist in foreign policy decision making and crisis behavior, as well as simulation and experimental techniques in political science. He is the author of numerous articles and seven books, the most recent A Study of Crisis (1997, 2000, with Michael Brecher) and Negotiating a Complex World (1999, with Brigid Starkey and Mark Boyer). His research has been supported over the years by the National Science Foundation, the U.S. Department of Education, the U.S. Institute of Peace, Sun Microsystems, NCR, and IBM, among others. His current work focuses on the combined use of aggregate and experimental techniques to study the mediation process in international crisis negotiations and how decision makers learn from previous crisis experience.

Oran Young Oran Young is the author or coauthor of more than twenty books and numerous scholarly articles. Dr. Young is a professor of environmental studies, director of the Institute of Arctic Studies, and director of the Institute on International Environmental Governance at Dartmouth College. He is also an adjunct professor of political science at the University of Troms⊘ in Norway. Dr. Young served for six years as the founding chair of the Committee on the Human Dimensions of Global Change of the National Academy of Sciences in the United States and is now chair of the Scientific Steering Committee of the international project on the Institutional Dimensions of Global Environmental Change (IDGEC) under the auspices of the International Human Dimensions Programme on Global Environmental Change. In addition, he served for six years as vice president of the International Arctic Science Committee and is currently a leader in the development of a decentralized University of the Arctic. Dr. Young's scientific work encompasses both basic research focusing on collective choice and social institutions and applied research dealing with Page 703 → issues pertaining to international environmental governance and to the Arctic as an international region. Among his recent books are Governance in World Affairs (1999); Creating Regimes: Arctic Accords and International Governance (1998); International Governance: Protecting the Environment in a Stateless Society (1994); Arctic Politics: Conflict and Cooperation in the Circumpolar North (1992); and International Cooperation : Building Regimes for Natural Resources and the Environment (1989).

Marysia Zalewski Dr. Marysia Zalewski joined the Centre for Women's Studies at Queen's University of Belfast in September 1999. Her teaching and research interests include feminist theory, international relations theory, the politics of theory (particularly as related to practice), reproductive technologies, and masculinity studies. She has published widely in the area of gender and international relations. Her most recent book is Feminisms after Postmodernism: Theorising through Practice (2000).

Dina A. Zinnes Dina A. Zinnes teaches at the University of Illinois at Urbana-Champaign. She received a Ph.D. from Stanford University in 1963. Professor Zinnes joined the department in 1980. She has received grants from the National Science and Ford foundations and the Advanced Research Projects Agency. She has been vice president and president of the International Studies Association, vice president of the American Political Science Association, president of the Midwest Political Science Association, and president of the Peace Science Society. She was a member of the council of the Inter-University Consortium for Political and Social Research. She is past editor of the American Political Science Review and has served on the editorial boards of the American Journal of Conflict Resolution, International Studies Quarterly, World Politics, and Political Methodology. She has also served as a

member or chair of numerous professional organization committees.