By Jeffrey T. Checkel *
Martha Finnemore. National Interests in International Society . Ithaca, N.Y.: Cornell University Press, 1996, 149 pages.
Peter Katzenstein, ed. The Culture of National Security: Norms and Identity in World Politics. New York: Columbia University Press, 1996, 562 pages.
Audie Klotz. Norms in International Relations: The Struggle against Apartheid . Ithaca, N.Y.: Cornell University Press, 1995, 183 pages.
For the past decade a central locus of contention within international relations has been the neorealist-neoliberal debate. This exchange has been fruitful and cumulative, allowing proponents of the two research programs to sharpen arguments while simultaneously shedding light on key issues of world politics, for example, the conditions under which relative or absolute gains-seeking behavior occurs. 1
By and large, the constructivists under review would concur with such a characterization. Their critique of neorealists and neoliberals concerns not what these scholars do and say but what they ignore: the content and sources of state interests and the social fabric of world politics. Reaching back to earlier theoretical traditions (the English school, [End Page 324] some versions of liberalism) and reaching out to new disciplinary foundations (sociology), constructivists seek to expand theoretical discourse.
Regarding both the books under review and constructivism more generally, this essay advances three claims. First, I argue that constructivism has succeeded in broadening the theoretical contours of IR. By exploring issues of identity and interest bracketed by neoliberalism and neorealism, constructivists have demonstrated that their sociological approach leads to new and meaningful interpretations of international politics. Moreover, constructivists have rescued the exploration of identity from postmodernists. By arguing for its importance using methods accepted by the majority of scholars, they have been able to challenge mainstream analysts on their own ground. Second and more critically, I show that constructivism lacks a theory of agency. As a result, it overemphasizes the role of social structures and norms at the expense of the agents who help create and change them in the first place.
Third, I argue that constructivism remains a method more than anything else. The central challenge for these scholars is theory development. Having demonstrated that social construction matters, they must now address when, how, and why it occurs, clearly specifying the actors and mechanisms bringing about change, the scope conditions under which they operate, and how they vary across countries. To accomplish this task, constructivists must integrate their insights and assumptions with middle-range theory. Otherwise, the empirical ad hocism that plagues their current work will remain.
The essay is organized as follows. It begins by defining constructivism and its approach to the study of global politics. Next, it considers the empirical contribution of constructivists, focusing on the three books under review. Finally, the review explores several issues constructivists must address if they are to mount a sustained challenge to their competitors in contemporary IR.
The Social Construction of International Politics
The constructivist critique of neorealism and neoliberalism reaches well beyond the level-of-analysis argument of either Image I (individual) or Image II (domestic politics) theorists. Constructivism is concerned not with levels per se but with underlying conceptions of how the social and political world works. It is not a theory but an approach to social inquiry based on two assumptions: (1) the environment in which agents/states take action is social as well as material; and (2) this setting can provide agents/states with understandings of their interests (it can [End Page 325] “constitute” them). Put differently, these scholars question the materialism and methodological individualism upon which much contemporary IR scholarship has been built.
The first assumption reflects a view that material structures, beyond certain biological necessities, are given meaning only by the social context through which they are interpreted. Consider nuclear weapons–the ultimate material capability. Constructivists argue that it is not such weapons themselves that matter. After all, the United States worries very little about the large quantity of nuclear weapons held by the British; however, the possibility that North Korea might come into possession of even one or two generates tremendous concern. 2
The second assumption addresses the basic nature of human agents and states, in particular, their relation to broader structural environments. Constructivists emphasize a process of interaction between agents and structures; the ontology is one of mutual constitution, where neither unit of analysis–agents or structures–is reduced to the other and made “ontologically primitive.” This opens up what for most theorists is the black box of interest and identity formation; state interests emerge from and are endogenous to interaction with structures. 3
Constructivists thus question the methodological individualism that underpins both neoliberalism and neorealism. This agent-centered view asserts that all social phenomena are explicable in ways that involve only individual agents and their goals and actions; the starting point of the analysis is actors (states) with given properties. Ontologically, the result is to reduce one unit of analysis–structures–to the other–agents. 4
Also implicit in many constructivist accounts is a model of human and state behavior where rule-governed action and logics of appropriateness prevail. Such logics involve reasoning by analogy and metaphor and are not about ends and means. Under them, agents ask “What kind of situation is this?” and “What should I do now?”–with norms helping to supply the answers. Norms therefore constitute states/agents, providing them with understandings of their interests. 5 [End Page 326]
Scholars of rational choice, by contrast, use a behavioral model based on utility maximization: when confronted with various options, an agent picks the one that best serves its objectives and interests. Much rational choice research (“thick” rationalism) also makes assumptions about the content of these interests, typically that they are material goods such as power or wealth. State (agent) interests are given a priori and exogenously. Norms and social structures at most constrain the choices and behavior of self-interested states, which operate according to a logic of consequences (means-ends calculations). 6
It is important to note that constructivists do not reject science or causal explanation; their quarrel with mainstream theories is ontological, not epistemological. The last point is key, for it suggests that constructivism has the potential to bridge the still vast divide separating the majority of IR theorists from postmodernists. With the latter, constructivists share many substantive concerns (role of identity and discourse, say) and a similar ontological stance; with the former, they share a largely common epistemology. Constructivists thus occupy a middle ground between rational choice theorists and postmodern scholars. 7
To illuminate these differences between constructivists and other schools, it is helpful to explore their understanding of central terms. Consider “norms,” a concept that has gained much currency in IR scholarship over the past decade. While realists see norms as lacking causal force, neoliberal regime theory argues that they play an influential rule in certain issue-areas. However, even for neoliberals, norms are still a superstructure built on a material base: they serve a regulative function, helping actors with given interests maximize utility. Agents (states) create structures (norms and institutions). 8 For constructivists, by contrast, norms are collective understandings that make behavioral claims on [End Page 327] actors. Their effects reach deeper: they constitute actor identities and interests and do not simply regulate behavior. As explanatory variables, their status moves from intervening to independent (Finnemore, chaps. 3, 4; Klotz, chap. 6, for example). Norms are no longer a superstructure on a material base; rather, they help to create and define that base. For constructivists, agents (states) and structures (global norms) are interacting; they are mutually constituted. 9
Taken together, these moves by constructivists–their questioning of methodological individualism and materialism, along with a continuing commitment to the scientific enterprise–have brought a breath of fresh air to thinking about world politics, in ways accessible to nearly all scholars. A key issue, however, is whether such new perspectives allow these researchers to explain important international puzzles
and phenomena and thereby demonstrate the empirical value of their approach.
Puzzles and Anomalies in World Politics: The Constructivist Contribution
The books under review seek to make empirical contributions in three areas: the role of international institutions and organizations (Finnemore); international security (Katzenstein volume); and the effects of international norms (Klotz). To evaluate their success, it is necessary to establish a baseline for comparison.
On international institutions, the dominant school for well over a decade has been neoliberal institutionalism. Since the publication of Keohane’s After Hegemony, these scholars have shown increasing sophistication in exploring the conditions under which institutions are created in the first place and the various roles they play in world politics. 10
Partly out of a concern for theoretical parsimony, neoliberal institutionalists have purposely bracketed several issues, including the sources of state interests, which are given by assumption. These scholars also grant only a limited role to institutions, considering them to be the creation of self-interested states that at most constrain choices and strategies. [End Page 328] Virtually ignored is the possibility that the effects of institutions reach deeper, to the level of interests and identity.
The baseline for the second issue-area–international security–is difficult to establish with precision, given the turbulence stirred up within this subfield by the end of the cold war. Certainly realism and rationalism have been and remain dominant here, but scholars have refined their analyses by paying more attention to domestic politics.
Important studies have enriched our understanding of security by exploring the role of ideology and threat perception, coalition politics, cognitive variables, and perceptions. While some accuse these scholars of smuggling into their analyses sociological and cultural variables emphasized by constructivists, they are nonetheless still united in a common commitment to rationalism and materialism. On the former, key actors (elite decision makers or groups within the state) make cost/benefit calculations and choose strategies designed to maximize certain interests; on the latter, perceptual, ideational, and cultural factors are ultimately parasitic on a material base. 11
Research on international norms, the third area addressed by the books under review, has been heavily influenced by regime analysis. These scholars have typically demonstrated that regime norms constrain the behavior of states; they are an explanatory variable that intervenes between underlying power distributions and outcomes. 12
Work on epistemic communities and, more recently, on transnational policy networks has brought research on international regimes closer to the insights offered by constructivists. It does so by suggesting that regime norms have deeper cognitive effects. Studies of this sort are arguably still a minor current within regime theory; they are also beset by a number of problems. Moreover, these scholars, especially those working on epistemic communities, embrace a largely agent-centered view, where state decision makers calculate and reason in response to a changing material environment. 13 [End Page 329]
Constructing National Interests
With this background, the task is to assess the contribution of the constructivists, beginning with the book by Finnemore. She questions two assumptions upon which most work on international institutions and IR more generally rests: the definition of state interests and rational means-ends calculations as the dominant mode of human interaction (p. x). In ontological terms, she seeks to move scholarship away from agent-oriented approaches (neoliberalism, for example) by paying more attention to the structure side of the agent-structure debate (p. 7).
In an excellent opening chapter, Finnemore argues that a constructivist logic of appropriateness is just as plausible a predictor of human and state behavior as the rationalists’ logic of consequences. When one makes actor and state interests the dependent variable, as she does, such logics of appropriateness can be key in determining their content. From where do such logics come? Systemic norms propagated by international organizations are one possible answer; they provide states with direction and goals for action.
The core of the book is three case studies of how international institutions (and, in one case, an international nongovernmental organization) were able to reconstitute state interests. These not only make for fascinating reading, but they also offer fresh insights into how institutions matter in world politics. They are also carefully argued, typically using two streams of evidence: (1) correlations between the emergence of new systemic norms and changes in state interests and practice; and (2) analysis of discourse to see if actions are justified in ways consistent with the values and rules embedded in the norms. These data, along with attention to alternative explanations, allow Finnemore to build a plausible case.
Her study of the United Nations Educational, Scientific and Cultural Organization (UNESCO) is representative. Finnemore’s puzzle is to explain why so many countries–developing and developed–created similarly stuctured science policy bureaucracies in a relatively brief period. She begins with a rigorous consideration of alternative explanations for their creation, for example, that they were established at the behest of powerful domestic constituencies. After testing these quantitatively and finding them lacking, Finnemore advances her own norms-based argument.
She starts at the international level, documenting how a norm prescribing the creation of national science units initially took hold at UNESCO and was later consolidated there. On the latter, part of the evidence is a careful study of the changing discourse within UNESCO [End Page 330] and among its member states. In particular, she notes how, over time, the notion that such units were needed took on a prescriptive status and came to be taken for granted. 14
Finnemore then turns to the state level, establishing correlations between the norms promoted by UNESCO and the creation of science bureaucracies by a number of states. To move beyond correlations, however, she considers several cases (Lebanon, East Africa) in more detail, analyzing the personal and organizational pathways through which the UNESCO norms diffused to these states. While the evidence here is a bit weaker (Finnemore conducted no fieldwork in the respective countries), it is nonetheless sufficient to allow her to make a plausible case that the norms were causally important for the change in science policy. Put differently, norms embodying certain logics of appropriateness had provided states with a new understanding of their interests (chap. 2).
Analysis of this sort moves one beyond the understanding of institutions provided by neoliberal institutionalists in at least two ways. First, by endogenizing interest formation, Finnemore sheds much- needed light on a crucial issue ignored by neoliberals: how states come to define their interests in certain ways. International organizations can teach states to value certain goals: national science bureaucracies in the case of UNESCO and poverty alleviation as a policy objective in the case of the World Bank. Finnemore carefully argues that these new interests arose in the absence of domestic constituencies or powerful countries favoring them. Instead, they were diffused to states by systemic norms, from the outside, as it were. Materialist and rationalist explanations cannot account for such value and behavioral change.
Second, the book demonstrates that international organizations are not empty vessels that simply reduce transaction costs, as portrayed by neoliberals. They are purposive entities that are able, in some cases, to trump states and their power. Indeed, Finnemore’s rich source material at the international level gives her cases a sense of dynamism and history in the making that is typically absent from neoliberal accounts of institutions. She has thus provided a theoretically informed and empirically substantiated argument for how institutions not only constrain but also constitute states and their interests, solving what is a puzzle for other theorists. 15 [End Page 331]
The book also fills a gap in constructivism: failure to tell us why certain norms arise at particular times. Finnemore provides an answer by exploring the role of moral entrepreneurs: committed individuals who happen to be in the right place at the right time to instill their beliefs in larger global social structures (pp. 24-28, chap. 4, pp. 137-39). 16
Finnemore’s account is not without weaknesses, however. Most important, it is not clear what one does with her argument, with so much resting on contingencies and idiosyncratic variables. While Finnemore has demonstrated that social construction is causally important, she has failed to specify systematically when, how, and why this occurs. To be fair, one book cannot do everything. All the same, the critical next step should be the development of a specifically constructivist theory of international institutions, one that would elaborate such scope conditions.
A second weakness is the degree to which Finnemore’s analysis is consistent with constructivism’s mutual constitution of agents and structures. Now, exactly how one operationalizes mutual constitution is a dilemma for all empirical constructivists. Finnemore’s solution is a bracketing strategy, where she first brackets agency and then, structures; her case studies are broadly faithful to this approach (pp. 24-25, chaps. 2-4).
The problem is the wrong choice of agents: the entrepreneurs who are responsible for the creation of norms in the first place. To analyze the process of mutual constitution that led to a change of national interests within particular states (her dependent variable), the agents she should be exploring, especially given her emphasis on global norms as the structures, are groups and individuals in those same states. If Finnemore had focused on these agents, it would have led her to explore several important issues, for example, the feedback effects of state (agent) behavior on the norms themselves.
A final difficulty is unavoidable given Finnemore’s emphasis on systemic social structures: the neglect of domestic politics. A question that immediately comes to mind when reading her analysis is why norms diffuse differentially, that is, why they have so much greater impact in some countries than in others. Through what mechanisms do global norms work their effects domestically? Finnemore alludes to these issues at several points but provides no clear answers (pp. 125, 137). This is odd, since it is the constructivists, with their attention to practice and interaction, who should be keying upon process and mechanisms. 17 [End Page 332]
Culture and Security
In a curious way, the Katzenstein volume is both very ambitious and very cautious. The former is seen in its willingness to question, from a sociological perspective, the very microeconomic disciplinary foundations of IR, and to do so on empirical issues that realists will recognize as their own. At the same time, Katzenstein and his contributors do not advance an alternative theory of national security; in contrast to many of the better edited volumes, this one does not even provide a common theoretical framework used by all contributors. 18
It does offer extraordinarily fresh thinking about security, however, along with richly detailed case studies. Among the familiar security questions explored in a new way are the proliferation of conventional weaponry, the role of deterrence in the nonuse of nuclear and chemical weapons, the sources of military doctrine, the Soviet cold war endgame, and alliance dynamics in both the North Atlantic and the Middle East.
Chapter 1 (Katzenstein) and especially chapter 2 (Jepperson, Wendt, and Katzenstein) should be required reading in any graduate seminar on security or IR more generally. This is not because Katzenstein et al. have decisively trumped mainstream theorists or because they have everything right. Rather, the essays are extremely helpful in explaining how the theoretical schools (neorealism, neoliberalism, constructivism) differ and why it matters (chap. 1) and for making sense of a sociological approach to national security (chap. 2). Moreover, these scholars are interested in dialogue; the goal is not to demonize existing approaches but to note their limitations. 19
The volume’s sociological approach to national security involves relaxing the two core assumptions of neorealism and neoliberalism, which are (1) that the environment of states can be conceived solely in terms of physical capabilities and (2) that institutions and structures only constrain the behavior of states with fixed interests. Relaxing the first assumption opens the possibility of social structures being causally important in world politics, while relaxing the second suggests that the effects of these structures may reach beyond behavioral constraint to identity and interest formation. In other words, just like Finnemore, this is a challenge to the materialism and methodological individualism that dominate the discourse in mainstream IR (Katzenstein, 16-17). [End Page 333]
Given this stance, the volume needs to address two key questions: (1) the content and properties of the social structures having such profound effects on agents; and (2) the causal mechanisms through which these structures have their affects. The social structures doing the explanatory work are norms and, to a lesser extent, culture. The former are defined as collective expectations about proper behavior for a given identity (Jepperson, Wendt, and Katzenstein, 54). That this is the same definition as used by Finnemore and Klotz is one indicator, among others, that a constructivist research program is beginning to consolidate itself in IR. 20
The presence of these normative structures is established through a variety of well-established and standard methodological techniques, for example, interview data, qualitative content analysis of primary sources, statistical studies. The research strategy is broadly similar to Finnemore’s: document the presence of the social structures; note a correlation between these and new state interests; examine changing discourse as further evidence of these normative effects; and, finally, strengthen the case by considering alternative explanations, usually drawn from neorealist and neoliberal theories.
Risse-Kappen’s chapter on the North Atlantic Treaty Organization is a good example of the general approach. His puzzle is to explain NATO’s initial formation and endurance–events that are anomalous, he argues, from the standpoint of both traditional and more sophisticated realist theories of alliances. In the first part of his essay Risse-Kappen discusses these likely alternative explanations and carefully documents their shortcomings.
Next, he develops his own liberal constructivist approach, where the norms that govern the domestic decision-making process within liberal systems come to regulate the interactions among democracies in international institutions such as NATO. Democracies, Risse-Kappen argues, “externalize their internal norms when cooperating with each other. Power asymmetries will be mediated by norms of democratic decision-making among equals emphasizing persuasion, compromise and the non-use of force or coercive power.” He then deduces four different ways such norms will influence the interaction process among democratic allies (pp. 268-71). 21 [End Page 334]
Risse-Kappen illustrates the argument by showing how the interests in play, both in the formation of NATO and during several key crises (Suez 1956, Cuba 1962), were shaped by the democratic normative context in which they evolved. In other words, the interests of states and alliance decision makers (the agents) were being constituted by these democratic norms (the structures). His evidence is carefully culled from secondary and, especially, primary sources, for example, the U.S. government’s Foreign Relations of the United States series and materials in the National Security Archive. This allows him to dissect the decision-making process, showing how norms affected the preferences and interests of various alliance partners. 22
The essay by Risse-Kappen is not at all atypical for the Katzenstein volume, which contains a number of carefully argued studies documenting the impact of norms. Unfortunately, the volume is much weaker at theorizing the causal mechanisms that give these social structures such powerful constitutive effects. This is a fair criticism to make, as the authors clearly commit themselves to a largely causal epistemology (Katzenstein, 4-5, 7; Jepperson, Wendt, and Katzenstein, 52-53, 65-68). However, as Katzenstein himself admits in the book’s concluding essay, structural theories such as sociological institutionalism, which is accorded a central role in the volume, neglect important processes that translate structural effects (pp. 512-13). 23
One result is that the role of agency, while highlighted empirically in many of the chapters, is neglected theoretically. The volume short-circuits one loop in the constructivist method: the causal arrows flow primarily from structures to agents. Mutual constitution, however, implies they also flow from agents to structures. Some constructivists might object that such sequential (structures to agents then agents to structures) causal language misconstrues the essence of their ontology: the simultaneous, mutual constitution of agents and structures. However, the empirical application of mutual constitution by these scholars follows precisely the sequential logic outlined here. 24
Despite such shortcomings, this is a very important volume. Its commitment to causal analysis and standard methodologies contributes to a productive dialogue with neorealists and neoliberals; for the most part, these scholars are all talking the same language. In addition, the case [End Page 335] studies (chaps. 3-11) offer new and meaningful insights. For some authors, this means demonstrating that particular security outcomes can be explained only when realist analyses are supplemented with constructivist approaches (Herman’s chapter on Soviet foreign policy under Gorbachev; Risse-Kappen’s essay on NATO).
Other contributors, however, go a step further and argue that their constructivist approach supplants rationalist and materialist accounts. For example, in a superbly argued essay, Alastair Johnston shows that the persistence of China’s realpolitik over several centuries can be understood only in terms of a constructivist explanation that subsumes structural realism (Katzenstein, chap. 7).
Finally, in an innovation rare in any edited volume, Katzenstein has included an essay (chap. 12 by Kowert and Legro) that reflects critically on the book as a whole. This excellent chapter provides the sense of cumulation and summary that is missing when one reads across the various contributions. It achieves this not by championing the constructivist cause but by critically evaluating the volume’s shortcomings. For developing a more coherent constructivist research program, this is precisely what is needed. Katzenstein et al. are to be applauded for including such a chapter.
Global Norms and the Demise of Apartheid
The puzzle Audie Klotz seeks to explain is why a large number of international organizations and states adopted sanctions against the Apartheid regime in South Africa despite strategic and economic interests that had fostered strong ties with it in the past. Klotz argues that the emergence of a global norm of racial equality is at the heart of the explanation: it led states to redefine interests even though they had material incentives not to do so. This demonstrates a constitutive role for norms, she argues, where they affect state identity and do not simply regulate behavior (chaps. 1-2).
The case studies on the United States, Britain, and Zimbabwe (chaps. 6-8) make for especially fascinating reading. Klotz’s extensive empirical research and attention to domestic politics allow her to explore how this global norm first reached the national level and the effects it then had on the interests of various groups and individuals. In contrast to Finnemore and many contributors to the Katzenstein volume, Klotz offers much more process-level evidence on how norms actually reconstituted state interests.
The book thus fills in important gaps in both regime theory and constructivism. Concerning the former, Klotz demonstrates in a nicely [End Page 336] argued section that neoliberal regime analysis shortchanges the role norms play in international politics. This is not to argue that the neoliberals have it all wrong (Klotz does not say this); rather, their view of norms as constraints on states, as opposed to constituting them, is only half the story. Empirically, she shows how this theoretical move can actually be carried out (pp. 13-33). In an important sense, Klotz is empiricizing the abstract critiques of regime theory advanced by Friedrich Kratochwil and John Ruggie over the past decade. 25
The author is also to be commended for flagging an issue that constructivist research must address. As Klotz puts it: “The crucial question is then how a contested norm . . . becomes institutionalized, both globally and domestically” (pp. 24-25). Indeed, after reading enough of this work, one senses that there are all too many norms floating around “out there” that somehow insinuate themselves “in here,” that is, in the domestic arena. (While Finnemore furthers our understanding of how norm institutionalization works in international institutions, she neglects the question of domestic diffusion mechanisms and processes.)
Klotz addresses this issue by elaborating three transmission mechanisms that link norms and policy choice: community and identity; reputation and communication; and discourse and institutions. While these are ultimately underspecified (one would want to know what mechanisms under what conditions are likely to be at work in a particular national setting), this is nonetheless a foundation upon which other scholars should build. By elaborating causal mechanisms that specify diffusion pathways, constructivists will move away from the correlational analyses too often evident in their work; process tracing of this sort is a method whose time has come for constructivism.
Three weaknesses limit the impact of Klotz’s book, however. First, the ontology is not one of mutual constitution, not even in its bracketing form–comments to the contrary notwithstanding (Klotz, 168-69, 172). Instead, like both Finnemore and many of the Katzenstein case studies, this is a study of how social structures, a global norm of racial equality, reconstituted agents.
Second, the analysis is too often correlational (pp. 158-61, for example). In part, this results from a failure to specify more systematically the causal mechanisms operating at the domestic level (Klotz, 24-33). However, it is also an artifact of the source material, which is primarily secondary. Given the arguments that Klotz wishes to make about the [End Page 337] effects of global norms on various groups within states, archival, memoir, or interview data would seem essential.
For the congressional representatives in her U.S. study, to take one case, it matters tremendously for the argument whether their views, in the presence of the global norm, were changing because they feared adverse electoral consequences (the rationalists’ means-ends calculations) or because they learned new values and beliefs (the constructivists’ logic of appropriateness) (Klotz, chap. 6). Klotz’s correlations tell us that the views were changing but not why this occurred; the necessary process tracing is never fully carried out in the substantive chapters.
Third, the theory-building potential inherent in the book’s ambitious cross-national design goes unrealized. Klotz presents no theory that might predict her results or explain similar dynamics in other countries, if one wished to extend the study. This is unfortunate: in the end, one is still left wondering why regimes and norms have such powerful constitutive effects in some states but not in others. 26
Summary
Two conclusions follow from the above. Most important, constructivists have convincingly shown the empirical value of their approach, providing new and meaningful interpretations on a range of issues of central concern to students of world politics. At the same time, constructivist theorizing is in a state of disarray. These researchers, much like the rational choice scholars they criticize, have made too rapid a leap from ontology and methods to empirics, to the neglect of theory development. This matters tremendously. As a central architect of constructivism has recently put it: “If parsimony is over-rated as a theoretical virtue . . . cumulation is surely underappreciated.” And cumulation, it might be added, if it is to be efficient and productive, requires theory. 27
Agency, Theory Building, and the Constructivist Enterprise
My purpose in this last section is twofold. I begin by highlighting three issues that should be easy for constructivists to fix. Two, more difficult [End Page 338] questions are then explored: the role of agency and the need for theory. Without more sustained attention to agency, these scholars will find themselves unable to explain where their powerful social structures (norms) come from in the first place and, equally important, why and how they change over time. Without theory, especially at the domestic level, constructivists will not be able to explain in a systematic way how social construction actually occurs or why it varies cross nationally.
The Three Easy Fixes
Constructivists need, first, to pay greater attention to research design. As noted, much of the empirical work examines single countries or issues. Cross-national or longitudinal designs would help reduce the problem of overdetermination that is evident in many constructivist analyses, where social structures, usually norms, are invoked as one of several causal variables with little or no insight given on how much of the outcome they explain (Katzenstein, chaps. 4, 8, 10; and Klotz, 114, 162, passim). It would also be useful to consider cases when the “dog doesn’t bark,” that is, where state identity/interests, in the presence of a norm, do not change. 28
Second, these scholars should give equal attention to the bad things in world politics that are socially constructed. There is a tendency in the recent work to consider only ethically good norms, such as those imposing a stigma on the use of nuclear or chemical weapons, those that helped bring the cold war to an end, or the global norms that facilitated the demise of Apartheid. Some constructivists are aware of this problem (Finnemore, 6, 31-32; Kowert and Legro, in Katzenstein 485-86), but future work must address it. It will not only protect these scholars from getting caricatured as peaceniks by theoretical opponents, but it will also direct their attention to important unexplored issues such as the role of social construction in ethnic conflict and war. 29 [End Page 339]
Third, constructivists must take greater care in defining key terms, for example, institutionalization. This word is invoked in nearly every analysis of norms (Finnemore, 126; Katzenstein, 56, 96-97, 129, 143, 161, 472, 484; Klotz, 24-26), but the reader is given no explanation of what the process entails. In what institutions–or individuals–do norms reside? Must norms be internalized first by individuals through a socialization and learning process? If so, constructivists should pay greater attention to developing the often implicit cognitive models in their analyses. Or, does institutionalization occur at a higher level of aggregation, through bureaucratic and legal processes that affect society as a whole. If this is the level under examination, constructivists could benefit from the insights of historical institutionalists and of those in the ideas literature who have studied such dynamics. 30
The Challenges Ahead
Ontology and theory building are the central challenges for constructivists.
Bringing Agency Back In
This move is necessary if mutual constitution is to be taken seriously as a way of thinking about the social world. I appreciate the reasoning of some that a neglect of agency is legitimate, at present, as a corrective to the extreme agent orientation of most mainstream IR (Finnemore, chap. 1). Moreover, it has proved very difficult to apply mutual constitution in empirical research.
All the same, constructivists should want to avoid the charge that they are reducing one unit of analysis–agents (states, decision makers)–to the other–structures (norms). One result of this reduction is a failure to explore how norms arise in the first place (and the role of agency and power in this process), and how, through interactions with particular agents, norms change over time. 31
An example clarifies the importance of the last point. Post-cold war Europe has witnessed the emergence of norms advancing more inclusive conceptions of national membership (citizenship laws, rights of national [End Page 340] minorities). Promoted initially by nongovernmental actors and more recently by the Strasbourg-based Council of Europe, the content of these norms has now been modified significantly as a result of Russia’s instrumental exploitation of them in a bid to reassert its dominance among the former Soviet states. The constructivists’ normative structures are themselves being reshaped by the activities of purposeful agents. 32
Three reasons explain why agency has fallen through the ontological cracks for constructivists. First, many constructivists rely upon the insights of sociological institutionalism for their thinking about the social world. Those insights, however, are based upon a particular branch of organization theory that systematically excludes questions of agency, interest, and power. 33
Second, because of their focus on collectively held, intersubjective understandings (norms), most constructivists, not surprisingly, are less interested in questions of individual agency. Yet the evolutionary development of norms suggests that, at some early point in their life histories, they may not be collective in any meaningful sense; particular individuals (Finnemore’s moral entrepreneurs, for example) may play key roles at early stages. Thus, social construction at the level of agents is–or rather, should be–a relevant concern for these scholars. 34
Finally, Wendt, who has been so influential in developing constructivism, has explicitly bracketed individual agency as a factor to be explained by mutual constitution. For Wendt, a key distinction is between the corporate and social identity of states, with the former deemphasized because “its roots [are] in domestic politics.” Since he assumes a unitary state, corporate identity includes and subsumes that of the individual. The result is that social construction at the level of individual agents or, more generally, at any domestic level is neglected. While several theorists have criticized Wendt for this stance, no clear understanding of how to rectify it has emerged. 35 [End Page 341]
It is ironic that constructivists therefore find themselves in a predicament all too familiar to rational choice scholars: their ontology has led them to neglect key issues. The agent-centered approach of rational choice provides a clear perspective on the microfoundations of human behavior, but much less clarity on how this connects with the broader institutional and social context. The dilemma then is how to get from microfoundations to outcomes. 36
Constructivists, despite their arguments about mutually constituting agents and structures, have advanced a structure-centered approach in their empirical work. Moreover, Wendt’s theoretical stance has led to a neglect of domestic agency. The result is that constructivism, while good at the macrofoundations of behavior and identity (norms, social context), is very weak on the microlevel. It fails to explore systematically how norms connect with agents. 37
Social Construction and Theory Building
To explore such connections, constructivists will need to engage in theory development. At present, constructivism is, like rational choice, nothing more than a method. It leads one to ask certain questions and make certain assumptions. However, constructivists should surely want more. In fact, in the volumes under review, there is a persistent call precisely for greater specification of constructivism (Finnemore, 130-31; Kowert and Legro, in Katzenstein, 469-83; Klotz, 26-33). 38
The missing element is substantive, middle-range theory, which would provide constructivists with a set (or better, competing sets) of research questions and hypotheses that could be tested in various cross-national and longitudinal studies. The need for theory is especially evident at the domestic level, where the constructivist “norm” is empirical ad hocism with all sorts of implicit models of domestic politics and key actors being invoked. 39 [End Page 342]
If constructivists are to theorize at the domestic level, they will need to address three issues. How deep within a policy does one need to go with a constructivist analysis? How is such an analysis actually carried out? Under what conditions is a constructivist approach, as opposed to a rationalist one, even necessary to explain the effects of social structures?
Progress on the first issue requires specification of political actors, that is, some model of domestic politics within the state. There are all sorts of domestic frameworks available (pluralist, institutional, and so on), but these are well known and need not be discussed. Rather, I wish to suggest that constructivists have already uncovered abundant evidence that the state-level penetration of international social structures varies cross nationally. The how deep question clearly matters. 40
A few examples will clarify the point. Although Finnemore is not explicit on this score, one can infer from her empirical chapters that normative effects are limited to state bureaucrats (Finnemore, chaps. 2, 4). In the Katzenstein volume, some authors find norms held broadly within a polity (Berger on postwar Germany and Japan), while others see their effects confined to political and academic elites (Herman on the USSR) or to state decision makers (Risse-Kappen on NATO; Katzenstein, chaps. 9, 8, 10). Klotz’s cross-national design uncovers evidence of normative effects at the level of political elites in one instance (the U.S.); in her British case, however, such influences are partly blocked by deeper, historically constructed national discourses (Klotz, chaps. 6, 7).
To make sense of and explain such diversity, constructivists will need to theorize the varying processes through which social construction occurs. The insights gained from Klotz’s partial move in this direction indicate its importance. Here, constructivism would benefit greatly from utilizing methods developed by IR scholars seeking to place greater emphasis on process. 41
Having established that social construction occurs at various levels within the state, the second question can be addressed: how does one conduct such an analysis? For present purposes, assume three domestic levels: society, state institutions, and individual decision makers. Furthermore, due to space limitations, consider only the individual level. What does it mean to explore the social construction of individual decision makers? Theoretically, it is to explore how social structures interact [End Page 343] with and fundamentally affect the identities of these agents, how certain logics of appropriateness come to govern their behavior.
For constructivists, this means being able to explain how the interests and identities of particular agents, in the presence of norms, change–or, equally important, do not change over time. Despite its centrality, this issue, which directly addresses the cognitive microfoundations of constructivism, has not received the attention it should, especially in the empirical literature. However, a review of this work suggests three possibilities.
One is a learning argument drawn from cognitive psychology. Just such a dynamic is implicit in Finnemore’s book, where agents (state elites), through exposure to norms, are taught new identities and interests. Because interests are changing, one can infer that this is a constructivist claim about complex, rather than simple, learning. (In the latter, new information allows actors to pursue given interests more effectively; it can be accommodated within a rationalist framework.) 42
The problem for such arguments is that when one introduces the reality and friction of domestic politics, complex learning typically breaks down. Absent such processes, one is back in the rationalists’ world of simple learning. This politics-learning tension is well established both theoretically and empirically, with the basic insight being that learning becomes less likely as the circle of actors grows. 43
Symbolic interactionist theory in sociology provides a second possible way to probe these constructivist microfoundations. Here, individual identities and interests are formed through a process of interaction, with two mechanisms being key: imitation and social learning. Since imitation does not involve interaction (and, thus, mutual constitution), it is the social learning dynamic that plays a more central role in the constructivist accounts. Social learning, much like the cognitive/individual sort just discussed, can be simple or complex, but given the constructivist emphasis on identity change, the focus is again on the latter. Specifically, complex social learning occurs when identities and interests are learned in response to how actors are treated by significant others. 44 [End Page 344]
While intriguing, this line of reasoning has not yet been integrated with empirical research. When and if this occurs, the same problem as discussed above will confront constructivists: how to maintain complex learning in settings where the static created by domestic politics hinders it.
Social psychology provides a final possible tool for exploring social construction at the individual level. Here, the theoretical foundations are provided by Turner’s self-categorization theory, where the focus is on individual-group interactions. For constructivists, the key process in Turner’s work is depersonalization, for this is how individual identities and interests change through interaction with a larger social group. 45
Unfortunately, this process is so context dependent and unclear (does depersonalization occur through social learning? through coercion?), it is not at all certain how constructivists might integrate its insights into their work. Nor surprisingly, when these scholars have used variants of self-categorization theory, it has led to unresolved theoretical disputes, as well as to sloppy empirical work. 46
The criticisms and questions raised above should not be viewed as dismissive. In addressing an issue of central importance–how to connect social structures to agents–these scholars are building much-needed bridges to other literatures. In fairness to constructivists, scholars of rational choice have been harshly criticized in similar ways for their attempts at the reverse process: connecting their sparse microfoundations to broader social and normative structures. 47
These last comments lead directly to the third question constructivism needs to address more systematically at the domestic level: when is such an approach, as opposed to a rationalist one, even necessary to explain the effects of social structures? Because most of the constructivist work to date has been method driven, these scholars have failed to appreciate that the domestic effects of norms are at times best captured and explained by rational choice. 48
Klotz’s U.S. study, for example, suggests that global norms were not so much transforming the identities of congressional representatives as [End Page 345] creating constraints on their behavior (Klotz, chap. 6). In other words, one is back in the rationalist’s world of means-ends calculations (in this instance, a political survival calculus of how best to secure reelection). Now, Klotz, as well as many contributors to the Katzenstein volume, does recognize that norms can have instrumental effects such as these. Nonetheless, one would want clear indicators of when one dynamic or the other is likely to prevail. The challenge, then, is to develop scope conditions. 49
One is temporal. This is the division-of-labor argument briefly mentioned in the Katzenstein volume (Jepperson, Wendt, and Katzenstein, 70; Kowert and Legro, 490-91). Constructivism might be best at explaining identity and interest formation, but as some later time, when interests were stable, rationalism might be the right method. Such a solution would have the benefit of making everyone happy: there would be a legitimate place and time for all approaches. However, the devil is in the details. Empirically, how does one know a priori when a state is likely to be in a period of identity formation, where constructivism is appropriate, as opposed to a time when identities and interests are already fixed?
A second scope condition is a density-of-interactions argument, which has been applied primarily to international bargaining. At some stage in this process, actors may switch from the rationalists’ consequential, means-ends logic to a situation in which their preferences are in genuine flux and open to change through persuasion and communication. However, the key question is how one predicts such a switch. What needs to happen and when? Cognitive uncertainty by individual negotiators? The establishment, through communication and speech, of some level of collective trust among them? Lacking this specification, the same problems arise as with the division-of-labor argument. 50
A final scope condition explores the role of domestic institutions. “Institution,” in this case, refers to the bureaucracies, organizations, and groups that channel and define policy-making within states. In the three books under review, one sees two very different normative effects at the domestic level. In some instances, decision makers and elites are [End Page 346] essentially taught (Finnemore) or learn (Herman, in Katzenstein, chap. 
new beliefs and values in the absence of any obvious domestic pressures; that is, new (constructivist) logics of appropriateness come to govern their behavior. At other times, norms do not have individual effects; instead, they mobilize domestic groups that pressure elites to change policy in ways consistent with the norms (for example, Klotz, chap. 6). That is, normative effects are operating through (rationalist) means-ends calculations.
Perhaps this variation is explained and predicted by differences in political institutions across states. In liberal polities such as the U.S., where decision makers have little autonomy from societal groups, the rationalists’ instrumental logic more often captures the domestic effect of systemic social structures. In states with greater autonomy and insulation from society (say, the former USSR), constructivist logics may more often capture the unit-level affects of norms. 51
Conclusions
An IBM ad in a recent issue of the shows a well-heeled executive holding his head and shaking it in despair: “Oh no, another paradigm shift,” he laments. The good news for IR theorists is they face no such threat from the constructivists reviewed in this essay. However, this attests not to their failures but to the nature of their goals: dialogue, a widening of disciplinary foundations, and a commitment to causal analysis. These scholars are out not to colonize and deconstruct IR but to revitalize and expand its conceptual lenses.
That one can make so many critical observations about this work suggests, paradoxically, its achievements. The publication of the books discussed here, along with the work of scholars such as Wendt, Ruggie, and Kratochwil, has for the first time given constructivism a critical mass of research that is both theoretical and empirical. This allows a reviewer to probe for lacunae and tensions, as well as synergies in it.
At this point, instead of summarizing, a broader issue needs to be raised: what kind of constructivism do we want? Some constructivists might feel this review “mainstreams” them too much. The criticisms on research design, better specification of key terms, developing middle-range [End Page 347] theory, taking domestic politics and agency seriously, after all, sound like a primer for building a more coherent research program.
There are two reasons for constructivists to move in this direction. First, judging by many comments to this effect, it is the direction in which they wish to move. Their emphasis on dialogue and causal analysis suggests a fairly standard concern with building a rigorous and coherent body of research that speaks to and plays off other literatures within IR.
Second, in its present form, it is not clear what one does with constructivism. How could Finnemore’s insights be applied to other international institutions–NATO, for example? Why do the transnational norms, which figure so prominently in Klotz’s study, have seemingly no impact in contemporary China? Answers to such puzzles will come only when constructivists specify more clearly the actors–structures and agents–and causal mechanisms bringing about change, the scope conditions under which they operate and how they vary cross nationally. Absent this theorizing, the “what do we do with it” question will remain.
Jeffrey T. Checkel is Senior Researcher and Coordinator, Research on European Identity Change, arena (Advanced Research on the Europeanization of the Nation State), University of Oslo. He is the author of Ideas and International Political Change: Soviet/Russian Behavior and the End of the Cold War (1997). His current project is a cross-national study exploring the relation between norms and identity change in contemporary Germany, Ukraine, and Russia.
* Earlier versions of this essay were presented at the 1996 annual convention of the American Political Science Association, and at the workship on “Structural Change in International Politics,” sponsored by the German Political Science Association, February 1997. For comments, I thank Andrew Cortell, Aaron Hoffman, Jeff Legro, Thomas Risse, and Alex Wendt. The financial support of the Alexander von Humboldt-Stiftung and German Marshall Fund is gratefully acknowledged.
Notes
1 See Robert Keohane, After Hegemony: Cooperation and Discord in the World Political Economy (Princeton: Princeton University Press, 1984); David Baldwin, ed., Neorealism and Neoliberalism: The Contemporary Debate (New York: Columbia University Press, 1993); Robert Powell, “Anarchy in International Relations Theory: The Neorealist-Neoliberal Debate (Review Article),” International Organization 48 (Spring 1994); and “Promises, Promises: Can Institutions Deliver?” International Security 20 (Summer 1995).
2 Alexander Wendt, “Constructing International Politics,” International Security 20 (Summer 1995), 73.
3 For an excellent discussion of this black box for neoliberals and neorealists written by a theorist sympathetic to their enterprise, see Powell (fn. 1), 317-24.
4 On neoliberalism’s methodological individualism, see Volker Rittberger, Andreas Hasenclever, and Peter Mayer, “Interests, Power, Knowledge: The Study of International Regimes,” Mershon International Studies Review 40 (October 1996), 183-87. For that of neorealism, see Alexander Wendt, “The Agent-Structure Problem in International Relations Theory,” International Organization 41 (Summer 1987), 340-44.
5 On logics of appropriateness, see James March and Johan Olsen, Rediscovering Institutions: The Organizational Basis of Politics (New York: Free Press, 1989).
6 On the last point, see Barry Weingast, “A Rational Choice Perspective on the Role of Ideas: Shared Belief Systems and State Sovereignty in International Cooperation,” Politics and Society 23 (December 1995); and Dennis Chong, “Rational Choice Theory’s Mysterious Rivals,” in Jeffrey Friedman, ed., The Rational Choice Controversy: Economic Models of Politics Reconsidered (New Haven: Yale University Press, 1996). Useful introductions to rational choice are Jon Elster, “The Market and the Forum,” in Elster, ed., Foundations of Social Choice Theory (Cambridge: Cambridge University Press, 1986); James Morrow, Game Theory for Political Scientists (Princeton: Princeton University Press, 1994), chap. 2; and Donald Green and Ian Shapiro, Pathologies of Rational Choice Theory: A Critique of Applications in Political Science (New Haven: Yale University Press, 1994), chap. 2.
7 See, among others, Alexander Wendt, Social Theory of International Politics (New York: Cambridge University Press, forthcoming), chaps. 1-2. There is a good bit of confusion regarding these central tenets of constructivism; see, for example, John Mearsheimer, “The False Promise of International Institutions,” International Security 19 (Winter 1994-95), 37-47.
8 For example, Stephen Krasner, ed., International Regimes (Ithaca, N.Y.: Cornell University Press, 1983). My comparisons here are limited to mainstream ir, since it has been vastly more influential than postmodern work in shaping the field.
9 Strictly speaking, my discussion of norms as intervening or independent variables is not correct, as constitutive effects (A enables or makes possible B) are not captured by standard causal terminology (A causes B). See Wendt (fn. 2), 72. In practice, however, empirical constructivists use the terms interchangeably; see, for example, Mlada Bukovansky, “American Identity and Neutral Rights: From Independence to the War of 1812,” International Organization 51 (Spring 1997).
10 See, among others, Keohane (fn. 1); Lisa Martin, Coercive Cooperation: Explaining Multilateral Economic Sanctions (Princeton: Princeton University Press, 1992); and Beth A. Simmons, “Why Innovate? Founding the Bank for International Settlements,” World Politics 45 (April 1993).
11 See, among others, Stephen Walt, The Origins of Alliances (Ithaca, N.Y.: Cornell University Press, 1987); Jack Snyder, Myths of Empire: Domestic Politics and International Ambition (Ithaca, N.Y.: Cornell University Press, 1991); and William Wohlforth, “Realism and the End of the Cold War,” International Security 19 (Winter 1994-95).
12 Mark Zacher, Governing Global Networks: International Regimes for Transportation and Communications (New York: Cambridge University Press, 1996), for example. An excellent, synthetic review of the regime literature is Rittberger, Hasenclever, and Mayer (fn. 4).
13 See Peter Haas, Saving the Mediterranean: The Politics of International Environmental Cooperation (New York: Columbia University Press, 1990); and Kathryn Sikkink, “Human Rights, Principled Issue-Networks and Sovereignty in Latin America,” International Organization 47 (Summer 1993). For critiques, see Checkel, Ideas and International Political Change: Soviet/Russian Behavior and the End of the Cold War (New Haven: Yale University Press, 1997), chaps. 1, 7; and Helen Milner, “International Theories of Cooperation: Strengths and Weaknesses,” World Politics 44 (April 1992).
14 The documentation and data come chiefly from archives at unesco’s Paris headquarters.
15 For a similar argument, see David Strang and Patricia Mei Yin Chang, “The International Labor Organization and the Welfare State: Institutional Effects on National Welfare Spending, 1960-80,” International Organization 47 (Spring 1993).
16 On moral entrepreneurs and the development of norms, see also Ann Florini, “The Evolution of International Norms,” International Studies Quarterly 40 (September 1996).
17 Indeed, Wendt himself stresses the importance of mechanisms and process in causal constructivist theorizing. Wendt (fn. 7), chap. 2, 91-96.
18 In addition to the edited volume, Katzenstein has published a monograph that makes many similar sociological claims. See Peter Katzenstein, Cultural Norms and National Security: Police and Military in Postwar Japan (Ithaca, N.Y.: Cornell University Press, 1996).
19 See also ibid., chap. 2.
20 Definitional congruence in key concepts of a research program is often seen as a sign of its growing maturity. See Milner (fn. 13).
21 Students of the democratic peace literature will recognize this as a constructivist extension of their domestic norms argument. See Zeev Maoz and Bruce Russett, “Normative and Structural Causes of the Democratic Peace,” American Political Science Review 87 (September 1993).
22 Risse-Kappen has elaborated these arguments in a separate monograph; Risse-Kappen, Cooperation among Democracies: The European Influence on U.S. Foreign Policy (Princeton: Princeton University Press, 1995).
23 In his own book, Katzenstein pays much greater attention to mapping such processes, although a lack of explicit theorizing about them is still evident. Katzenstein (fn. 18), chaps. 3-6.
24 See Finnemore’s bracketing strategy (p. 25); Wendt (fn. 4), 364-65; and fn. 9 above.
25 Friedrich Kratochwil and John Ruggie, “International Organization: A State of the Art on an Art of the State,” International Organization 40 (Autumn 1986); and Friedrich Kratochwil, Rules, Norms and Decisions: On the Conditions of Practical and Legal Reasoning in International Relations and Domestic Affairs (Cambridge: Cambridge University Press, 1989).
26 By “theory” I mean middle-range theory and its development, which should be the goal of problem-driven empirical research. See, for example, Thomas Risse-Kappen, ed., Bringing Transnational Relations Back In: Non-State Actors, Domestic Structures and International Institutions (Cambridge: Cambridge University Press, 1995).
27 For the quote, see Wendt (fn. 7), chap. 1, 15. A central message of one recent and influential critique of rational choice is precisely its neglect of theory development, particularly of the middle-range sort. See Green and Shapiro (fn. 6), 188; and idem, “Pathologies Revisited; Reflections on Our Critics,” in Friedman (fn. 6).
28 On the last point, Klotz’s cross-national focus is an important step in this direction. For additional constructivist research utilizing single-country/issue designs, see Ray Koslowski and Friedrich Kratochwil, “Understanding Change in International Politics: The Soviet Empire’s Demise and the International System,” International Organization 48 (Spring 1994); Richard Price, “A Genealogy of the Chemical Weapons Taboo,” International Organization 49 (Winter 1995); Thomas Biersteker and Cynthia Weber, eds., State Sovereignty as Social Construct (New York: Cambridge University Press, 1996); Bukovansky (fn. 9); Nina Tannenwald, “The Nuclear Taboo: The Normative Basis of Deterrence” (Manuscript, Watson Institute for International Studies, Brown University, April 1996); and Jutta Weldes, “Constructing National Interests,” European Journal of International Relations 2 (September 1996).
29 On the last point, see Lars-Erik Cederman, “From Primordialism to Constructivism: The Quest for Flexible Models of Ethnic Conflict” (Paper presented at the annual meeting of the American Political Science Association, San Francisco, September 1996). A particularly egregious example of the caricaturing is Mearsheimer (fn. 7).
30 See Frank Longstreth et al., Structuring Politics: Historical Institutionalism in Comparative Analysis (New York: Cambridge University Press, 1992); and Judith Goldstein, Ideas, Interests and American Trade Policy (Ithaca, N.Y.: Cornell University Press, 1993). Not surprisingly, it is Katzenstein, the comparativist, who has offered the most careful constructivist account of domestic norm institutionalization. See Katzenstein (fn. 18), chaps. 1-3, 5, 7.
31 Dessler’s transformative model of international structure should be especially relevant to constructivists as they rethink the role of agency in their analyses. See David Dessler, “What’s at Stake in the Agent-Structure Debate?” International Organization 43 (Summer 1989).
32 Checkel, “Norms, Institutions and National Identity in Contemporary Europe” (Manuscript, October 1997).
33 See Paul DiMaggio, “Interest and Agency in Institutional Theory,” in Lynne Zucker, ed., Institutional Patterns and Organizations: Culture and Environment (Cambridge: Ballinger Publishing, 1988); Paul DiMaggio and Walter Powell, eds., The New Institutionalism in Organizational Analysis (Chicago: University of Chicago Press, 1991), chaps. 1, 4; Frank Dobbin, “Cultural Models of Organization: The Social Construction of Rational Organizing Principles,” in Diana Crane, ed., The Sociology of Culture: Emerging Theoretical Perspectives (Cambridge: Blackwell Publishers, 1994).
34 See also the discussion of norm reproduction in Florini (fn. 16), 374-75, 377-80.
35 Alexander Wendt, “Collective Identity Formation and the International State,” American Political Science Review 88 (June 1994); and, for the quote, idem, “Identity and Structural Change in International Politics,” in Friedrich Kratochwil and Yosef Lapid, eds., The Return of Culture and Identity in IR Theory (London: Lynne Rienner, 1996), 50-51. For critiques, see Sujuta Pasic, “Culturing International Relations Theory: A Call for Extension,” in Kratochwil and Lapid, 87-90; and Cederman (fn. 29), 13-19.
36 Rational choice institutionalism represents an effort to address this dilemma. See Norman Schofield, “Rational Choice and Political Economy,” in Friedman (fn. 6), 192-93, 207-8; and Peter Hall and Rosemary Taylor, “Political Science and the Three New Institutionalisms,” Political Studies 44 (December 1996), 958-62.
37 On the micro versus the macrofoundations of behavior and identity and the tensions between the two, see “Symposium: The Role of Theory in Comparative Politics,” World Politics 48 (October 1995), 13-15.
38 After earlier confusion, Wendt also now argues that constructivism is not a theory. Wendt (fn. 7), chap. 1.
39 All the books reviewed are strongest, theoretically, at the systems level, in large part because they draw upon an already well developed sociological literature that is systemic in orientation. See Martha Finnemore, “Norms, Culture and World Politics: Insights from Sociology’s Institutionalism,” International Organization 50 (Spring 1996).
40 Milner’s (fn. 13) advice to mainstream ir theorists on how to conceptualize domestic politics is relevant here as well.
41 Peter Evans, ed., Double-Edged Diplomacy: International Bargaining and Domestic Politics (Berkeley: University of California Press, 1993); Haas (fn. 13); Sikkink (fn. 13); and Risse-Kappen (fn. 26), among others.
42 Personal communication, Martha Finnemore, September 1996. See also Thomas Risse-Kappen, “Democratic Peace–Warlike Democracies? A Social Constructivist Interpretation of the Liberal Argument,” European Journal of International Relations 1 (December 1995). On the learning literature more generally, see Jack Levy, “Learning and Foreign Policy: Sweeping a Conceptual Minefield (Review Article),” International Organization 48 (Spring 1994).
43 On the learning theory–politics connection, see Richard Anderson, “Why Competitive Politics Inhibits Learning in Soviet Foreign Policy,” in George Breslauer and Philip Tetlock, eds., Learning in U.S. and Soviet Foreign Policy (Boulder, Colo.: Westview Press, 1991).
44 See Wendt (fn. 7), chap. 7; and idem, “Anarchy Is What States Make of It: The Social Construction of Power Politics,” International Organization 46 (Spring 1992). Symbolic interactionist theory has been developed primarily at the individual level, which is why I discuss it here. Wendt, unconvincingly in my view, argues that it can be applied at the level of (unitary) states as well.
45 See John Turner, Rediscovering the Social Group: A Self-Categorization Theory (Oxford: Basil Blackwell, 1987), chap. 3; and Penelope Oakes et al., eds., Stereotyping and Social Reality (Oxford: Blackwell, 1994), chaps. 1, 4.
46 On the former, compare Jonathan Mercer, “Anarchy and Identity,” International Organization 49 (Spring 1995); and Wendt (fn. 7), chap 7. For the sloppy empirical work, see Glenn Chafetz, “The Political Psychology of the Nuclear Nonproliferation Regime,” Journal of Politics 57 (August 1995).
47 For example, Robert Lane, “What Rational Choice Explains,” in Friedman (fn. 6).
48 For details, see Checkel, “International Norms and Domestic Politics: Bridging the Rationalist-Constructivist Divide,” European Journal of International Relations 3 (December 1997).
49 For other constructivist accounts portraying similar rationalist logics, see Price and Tannenwald, in Katzenstein, 138, 148-50; and Bukovansky (fn. 9), 21-51. Very similar questions of scope and domain are now being asked by several rational choice analysts. See the discussion of “segmented universalism,” in Green and Shapiro (fn. 6), 192-93, 204; Michael Taylor, “When Rationality Fails,” in Friedman (fn. 6), 230-33; and Powell (fn. 1), 324.
50 Thomas Risse, “The Cold War’s Endgame and German Unification” (A Review Essay), International Security 21 (Spring 1997). This constructivist conception of communication thus extends well beyond the rationalists’ “cheap talk.” For an excellent discussion, see James Johnson, “Is Talk Really Cheap: Prompting Conversation between Critical Theory and Rational Choice,” American Political Science Review 87 (March 1993).
51 For a full theoretical elaboration, see Checkel, “Between Norms and Power: Identity Politics in the New Europe” (Book manuscript in progress), chap. 2. Recent work on the role of international norms in U.S. policy-making is consistent with the argument made here. See Andrew Cortell and James Davis, “How Do International Institutions Matter? The Domestic Impact of International Rules and Norms,” International Studies Quarterly 40 (December 1996).
and phenomena and thereby demonstrate the empirical value of their approach.
new beliefs and values in the absence of any obvious domestic pressures; that is, new (constructivist) logics of appropriateness come to govern their behavior. At other times, norms do not have individual effects; instead, they mobilize domestic groups that pressure elites to change policy in ways consistent with the norms (for example, Klotz, chap. 6). That is, normative effects are operating through (rationalist) means-ends calculations. 
