Translating Terminologies

Recognizing that a shared ontology of mutual constitution points to common research questions is only possible if we move beyond the epistemological divide that riddles constructivist debates. Yet, we do need to understand alternative terminologies rather than simply privilege one set of assumptions over the other. Two key issues lead researchers to talk past each other: (1) the putative distinction between explanation and understanding and (2) inevitable differences that arise across levels of analysis. In an attempt to overcome these barriers, we offer a translation of key concepts. Then, we elaborate upon some of the methodological implications, concentrating on historical narrative at the macrolevel, process tracing at the mesolevel, and participant observation at the microlevel. Identifying complementarities and trade-offs between various tools of analysis will, we hope, provide a basis for building bridges across the prevailing epistemological divide.

questions often become categorized as representing ''modern'' versus ''postmodern'' or as denoting ''positivist'' versus ''post-positivist'' positions. Constructivists on the modernist or positivist side tend to see intersubjective understandings as stable, remaining unchanged by the perspective of the interpreter. They typically analyze the world in multicausal terms, treating components of structure and agency as variables to be tested. These researchers often frame questions in terms of policy and behavior, looking for patterns in the formation and operation of international regimes or foreign policymaking to demonstrate how norms matter. For example, contributors to The Culture of National Security: Norms and Identity in World Politics (Katzenstein 1996) draw on sociological concepts associated with hypothesis testing and generalization. The title alone signals the epistemology: ''culture'' and ''identity,'' both in the singular, with ''norms'' implying that meanings can be stable and knowable, independent of the researcher. Culture and identity, therefore, can be isolated from other characteristics of social life and treated as variables. In a similar vein, Jeffrey Checkel, in his contribution to this Forum, advocates the use of causal process tracing to test hypotheses about norms and to delineate ''scope'' conditions for processes of socialization.
Post modernists or post-positivists, in contrast, favor terminology that captures the instability of meanings. They insist that we work in a hermeneutical circle without any objective standpoint for analysis. Less concerned with generalization, they typically eschew the goal of causal explanationFthus rejecting the vocabulary of ''variables''Fin favor of tracing ''historical conjunctures.'' For example, contributors to Cultures of Insecurity: States, Communities, and the Production of Danger (Weldes et al. 1999) draw explicitly on anthropological insights. Cultures provide all-encompassing contexts within which we can appreciate nuances in meaning. Again, the choice of title signals an epistemological perspective: the ''production'' of multiple ''cultures'' suggests fluidity and malleability in social understandings, including the researchers' own biases. Rather than being an outcome to be explained, policies signal notions of security. In his contribution to this Forum, Kevin Dunn characterizes this anthropological approach as asking ''how possible'' rather than ''why'' questions because culturesFcomprised of dominant but evolving discoursesFset the parameters for action rather than determine particular choices.
Yet, these two edited volumes and this Forum also illustrate how alternative theoretical vocabularies can hide common research concerns. For instance, some talk about the ''production of representations,'' whereas others explore the ''functions of norms.'' But both groups see ''security'' and ''identities,'' among other core concepts, as conditioned by cultural context rather than determined by any objective characteristics. As Colin Wight (1999:122, emphasis in the original) points out, ''we can expect both determinacy and indeterminacy'' in meanings and practices. By looking at ''representations,'' ''norms,'' ''discourses,'' ''rules,'' and other types of intersubjective understandings, constructivists demonstrate that certain ideas get taken for granted or dominate while others remain unspoken or marginalized. Positivists and post-positivists agree, in other words, that agents or subjects create meanings within structures and discourses through processes and practices.
Because our common research agenda focuses on how and why meanings become institutionalized or change, we should reject any stark distinction between ''explanation'' through causal analysis versus ''understanding'' through interpretation (see, for example, Hollis and Smith 1990). In practice, there is considerable overlap. Although researchers concentrate more on one or the other, ''how'' and ''why'' questions are not easily separated. Those who say they explain behavior also interpret meaning, and those who focus on understanding language also explain action to some degree. As Hidemi Suganami (1999:371-372) points out, ''we explain something to someone [to make it understandable] . . . the act of 'explaining' assumes a common set of prior 'understandings.' '' In the next three sections, we will map out some methodological tools available to constructivists falling anywhere along this epistemological spectrum. We start by comparing historiographic techniques used to characterize changes in the international system. But these studies tend to privilege structure over agency, so we turn to works on international regimes to examine how actors frame and reframe ideas. Disagreement between constructivists over whether to incorporate the role of the researcher leads us, finally, to query the place of individuals in processes of socialization.

System Change
Constructivists analyze how institutions, comprised of rules and norms, establish habitual practices and procedures. The resulting social structures vary across historical periods, leading scholars to distinguish between systems, such as between the ancient, the medieval, and the modern. When feudalism gave way to absolutist monarchy, for example, sovereignty legitimized territorial political units while it undermined the authority of the Catholic Church (Hall 1999). In ancient Greece, city-states abided by diplomatic rules of arbitration (Reus-Smit 1999). Through these studies, constructivists challenge the common assumption that states have always been the primary actors in an anarchical international system, opening up the possibility for both agency and change.
Variations in the components of social structure, such as the norm of sovereignty and the practice of arbitration, offer one way to delineate systems and what drives macrohistorical change. One reason why constructivists disagree over the details of periodization is that they do not necessarily apply the same concepts in organizing source materials. Christian Reus-Smit (1999), for example, uses the notion of moral purpose to question Ruggie's (1986) emphasis on the seventeenth century as the transition to the modern era. By probing in detail the differences between Italian city-states, absolutist monarchies, and nation-states, he asserts that the multilateral practices of the nineteenth century are sufficiently distinct from those of the seventeenth to merit labeling it a new era. Rodney Hall (1999), in contrast, emphasizes moral authority in the decline of the Catholic Church, paying less heed to variations in how states interacted among themselves. Looking at the mythology of ancient civilizations through the Christian era and up to contemporary times, James Der Derian (1987), in turn, builds on the concept of alienation as a basis for tracing cultures of diplomacy.
Despite their differences, all these studies use historical evidence to tell a more coherent and complete (''better'') narrative. These researchers use some combination of secondary sources and archival materials to draw out new insights or to argue for the broader significance of particular events. They insist, for example, that classic texts, including legal documents such as the Treaty of Westphalia, need to be read within their historical context. Otherwise, we may superimpose contemporary concepts and concerns back in time, distorting the meaning of these documents. Thus, Reus-Smit (1999) insists that realists have misinterpreted The Peloponnesian War and The Prince because they take snippets out of context, leading to inaccurate inferences about Thucydides and Machiavelli's original arguments. Placing these texts in historical context, including contemporaneous primary and secondary sources, enables us to assess just how much the Treaty of Westphalia (or other manifestations of social structure) reflect new practices. Only then can we delineate historical junctures and characterize erasFbased on what we consider to be significant shifts.
Because conceptual choices lead to alternative demarcations of temporal (or geographical) boundaries, how should we determine which of these approaches is better? Those leaning toward positivism value the coherence or completeness of a narrative based on an objective search for new facts. But post-positivists stress that narratives create truth claims not objective knowledge. Consciously or not, these histories reflect the scholar's own interpretive framework because we derive fundamental assumptionsFincluding concepts like moral authority or aliena-tionFfrom underlying ideologies or ''meta-narratives'' (Lynch 1999). These assumptions, in turn, guide the formulation of research questions. The predominance of realist thinking, in this view, reflects US power, both in world politics and in the academy, whereas Marxism tends to be marginalized as an analytical framework because it challenges capitalism, a fundamental component of US hegemony (Agathangelou and Ling 2004).
Consequently, we need to be conscious of how our key concepts (such as social purpose, moral authority, and alienation) reflect underlying ideologies and shape the formulation of our research questionsFand answersFbecause we look for new evidence when our theoretical assumptions change or are challenged. For example, regardless of whether they seek to promote US power, realists are unlikely to explore shifts in the fundamental nature of the international system because they assume the persistence of anarchy. Constructivists, in contrast, show that states emerged as key actors during a particular period in European history. At issue is whether to accept states and anarchy as core theoretical assumptions and not as historical claims. Given a set of core assumptions, we can evaluate the accuracy and completeness of historical narratives, but history will not provide the basis for assessing whether realism is ''right.'' Because constructivists demonstrate that system transformation is possible, they are often portrayed (especially by critics) as championing agency. This charge shifts the terrain from the past to the future, turning theory into a prescriptiveFideo-logicalFguide to action. Even though we accept that theories can have prescriptive value, we reject the position that theories only reflect political agendas. Even when they serve as guides for action, theories still need to be evaluated on the basis of empirical analysis. Regardless of whether one champions or resists change, some analytical lessons prove more useful than others. For example, are we now experiencing the transition to a new era of globalization? Answers to this question can aid those who seek to prevent change as much as those who foster it.
The challenge, then, is to figure out why international systems sometimes change. The macrohistorical works surveyed above offer only a starting point because their primary goal is to identify differences between the contemporary and previous systems. UnderstandingFand perhaps shapingFthe processes of system change require more than static comparisons. As the following section on regime evolution demonstrates, methodologies that highlight interactions between structures and agents offer us complementary insights into mechanisms of change.

Regime Evolution
Constructivists agree that no institutions, interests, or identities are immutable. Prevailing ideas change over time because some people modify them. For instance, the influence of the Catholic Church diminished relative to the secular state, wage labor prevailed over slavery, imperialism gave way to decolonization, and the Bretton Woods economic system is under challenge. Protestant princes, abolitionists, and liberation movements (among others) transformed their worlds; today, anti-globalization activists seek to do the same. But if actors are conditioned (''constituted'') by the prevailing sociohistorical contexts of their times, why do some adopt new ways of thinking and behaving rather than follow habitual practices?
One set of debates about these questions centers on hegemony and the mechanisms that maintain it. For example, we might characterize contemporary capitalism as a cluster of social understandings based on a concept of money valued through exchange rates rather than gold and the legitimacy of wage labor rather than indentured servitude. Within this context, Ruggie (1992) argues, the postwar economic order, manifest in a particular multilateral form of formal organizations, reflected and reinforced US predominance. Robert Latham (1997) goes further to critique the role of liberal values and practices in the construction of that particular global order. He links capitalist ideology, liberal economics, and militarism as inseparable components of US power. Feminists, such as Ann Tickner (2001), make similar claims, highlighting how gendered assumptions explain this mutual constitution of liberalism, capitalism, and militarism.
People act within this capitalist framework of meaning (what some call ''regimes of truth''), but we cannot assume that its norms and practices always ''naturalize'' to create ''collective recognition of them as structural properties'' (Doty 1997:371). Hegemony, after all, does not eliminate ideological tensions or preclude challenges to the policies of multilateral organizations as is evident in debates over globalization. If ideologies do not automatically produce social facts, how do individual beliefs become collective understandings about the world? Why do some norms and practices attain sufficient autonomyFepistemic powerFto constrain actions? When should we treat them as social structures?
To answer these questions, constructivists explore the degree to which actors reinforce or transcend institutionalized practices, suggesting processes such as legitimation, learning, reasoning, and other forms of communication. By demanding or apologizing, for instance, people reinforce or recreate the meanings institutionalized in social structures. Research examining these issues uses a range of methodological techniques to capture various aspects of the language people use, how they use it, with what effects, and why. Often, these tools complement each other, but there can also be trade-offs between them. These methodological choices are rooted in the language that researchers use in their core concepts. A comparison of causal and conjunctural process tracing illustrates this point.
Positivist-leaning scholars often look for patterns in the language of international agreements, such as treaties, to observe norms. They then analyze structural constraints, collective action, or shifting interests as variables leading to testable hypotheses. By examining the diffusion of norms through European integration, for example, Checkel (2001) seeks to explain why some countries adopt pan-European conceptions of citizenship for minorities by contrasting coercion and social learning. Because he sees norms as stable intersubjective understandings that remain unchanged by the perspective of the interpreter, he treats discourse as objective knowledge to be discerned through interviews with participants and in meeting memoranda (for example, in the Council of Europe). Researchers' interpretations of policymakers' preferences and motives can also be verified with media reports and other observers' assessments. Tracing a chain of decisions about policy outcomes produces a causal claim that can, in turn, be the basis for generalization.
Eschewing causal claims in favor of historical contingency as noted above, postpositivists see language as more fluid, insisting that we work in a hermeneutical circle, which precludes any objective stand point. Mapping discursive conjunctures and disjunctures over time, with what is frequently called a genealogical approach, the researcher identifies ''chance occurrences, fortuitous connections, and reinterpretations,'' underscoring that structures or practices ''often change in such a way that they come to embody values different from those that animated their origins'' (Price 1995:86). For example, Dunn (2003) traces images of the Congo in Western public discourse from colonization through contemporary civil war. Recurring tropes, notably ''the heart of darkness,'' portray the Congolese as irrational savages. Such dehumanization legitimates foreign intervention by making it conceivable as a policy option. Rather than seeking to explain when or why intervention then occurs, Dunn concentrates on the effects of such policies on subsequent images of the Congo, noting the ebb and flow of particular representations that comprise Western and Congolese identities.
Both causal and conjunctural forms of process tracing enable us to grasp specific components of structural change. Indeed, they complement each other, with genealogy providing an explanation for the range of possible policies and causal process tracing focusing on the particular selection that occurs among these possibilities. Yet, both still leave unanswered many questions about when and why some people challenge prevailing norms and practices. Why do only some rules become internalized? Why do some people rebel? How do decision makers resolve conflicting social and self-interested expectations? Do we need to understand actor motivations? If so, can constructivists fit individuals into their analyses without ''negating the indissoluble unity'' that Roxanne Doty (1997:373) suggests is fundamental to the notion of mutual constitution?

Cognitive Transformation
Neither causal nor conjunctural process tracing sufficiently captures human cognition. The traditional array of interviews, news reports, memoirs, biographies, and archival records of meetings that delve into the elusive thinking of key decision makers can never reveal the ''real'' reasons for their choices. Despite the insights that can be gained from these sources, their limitations are also well known. People rewrite their views with their legacies in mind; records can be incomplete or inaccessible; and interviewees may not be forthcoming. Alternatively, public discourse captures how socially constructed subjects act within the confines of shared language. Yet, as Wight (1999:137, n. 4) points out, privileging these observable linguistic practices downplays ''the inner life'' of actors.
Participant observation offers an underutilized complement to these private and public sources of evidence, because it offers exceptional insights into the transformative potential of language at the social-psychological level. In particular, Carol Cohn's (1987) path-breaking assessment of her experiences in a nuclear weapons bureaucracy during the Cold War demonstrates how language constructs and reconstructs agents. She suggests four stages in this socialization process: listening, speaking, dialogue, and ''terror,'' a term she uses to indicate alienation from her former beliefs. Michael Barnett's (1999) more recent experiences as a political officer for the US mission to the United Nations confirm this trajectory.
When Cohn and Barnett arrived at their respective workplaces, they realized that their academic training and research experience did not prepare them for their new roles. The sexual imagery of the techno-strategic language of the weapons scientists, for example, starkly showed Cohn that she was an outsider. And the scientists ignored even sophisticated questions when she spoke in ordinary English. Colleagues, furthermore, teased and cajoled both Cohn and Barnett; personal acceptance required an ability to fit into the local culture. Both learned to speak these new languages and abide by the rules of the game. They describe, furthermore, how the use of these specialized languagesFengaging in dialogueFtransformed their thinking. Previously sympathetic to disarmament activists, for example, Cohn almost forgot how to articulate her long-standing critiques of strategic thinking. Barnett, in turn, found himself expressing a version of US interests that best meshed with the values voiced in the UN. In less than a year, both participant-observers had been socialized into their respective languages and bureaucracies.
Despite what appears to be these overarching similarities, we can also see differences in their analyses. Cohn shows us a world populated by imagery that distances the user from any impact that nuclear weapons might have on human lives. Nuclear strategists happily discuss war-fighting strategies and weapon survivability without any attention to the potential for nuclear holocaust. We cannot dismiss this language, because it affects resource allocation and the manufacture of actual bombs. Cohn's work thus highlights post-positivist claims that discourse creates new realities. Reflecting a positivist's inclination to explain policy outcomes, in contrast, Barnett concentrates on the discursive framing of ''interests'' that are produced and reproduced through institutionalized decision making. A few particular words, notably ''genocide,'' hold the power to define situations and demand policy responses. Hence, policymakers refused at the time to characterize the killings in Rwanda as genocide and are reacting similarly to the mass killings in Sudan.
Both writers also draw out the moral implications of discourse. For instance, Barnett, having participated in the UN decision not to send peacekeepers to Rwanda at the height of the genocide in 1994, now decries the sense of indifference that his role taught him; the reputation of the UN proved more important than saving Rwandan lives. As a result, these studies call on us to reject the tendency to see only those people capable of changing structures (that is, exhibiting epistemic power) as having agency. Constructivism can, therefore, lead to a broader basis for ethics if all individuals are held accountable for replicating or challenging the rules of the game.

Conclusion
Our purpose here has been to move beyond an abstract formulation of the agentstructure debate by drawing attention to the processes of mutual constitution and some techniques for researching them. We urge constructivists of all persuasions to recognize our shared aim to produce scholarship that carefully uses and evaluates evidence. If we lose sight of our commonalities, we miss an opportunity to learn from each other about how to understand and explain international relations. Choices among these (and other) methodologies certainly deserve greater attention because no single set of ''best'' tools could possibly apply to the diverse questions that arise in the societies in which we live, write, and teach. We welcome Jeffrey Checkel's and Kevin Dunn's contributions to this Forum as a step in this direction.

Department of Political Science, University of Oslo, Norway
''This argument is too structural. It's under-determined and based on unrealistic assumptions. Moreover, it tells us little about how the world really works.'' Among many scholars (the present author included), these are an oft-heard set of complaints. Consider two examples. The central thesis of the democratic peace literatureFthat democracies do not fight other democraciesFis hailed as one of the few law-like propositions in international relations. Yet, as critics rightly stress, we know amazingly little about the mechanisms generating such peaceful relations (Rosato 2003:585-586, passim;Forum 2005;Hamberg 2005). And in Europe, scholars have for years debated the identity-shaping effects of European institutions. One claim is that bureaucrats ''go native'' in Brussels, adopting European values at the expense of national ones. But here, too, critics correctly