Neither of us ever intended to help construct constructivism, much less hoped to win the Skytte Prize for it. Indeed, in graduate school we had never heard the term, until Nicholas Onuf introduced it to International Relations (IR) in 1989. But as things turned out, starting in 1992 our work and that of many others dramatically raised the profile of constructivism in IR. Just five years later Emanuel Adler (Reference Adler1997) could claim it as the “middle ground” in IR theory, between the positivist, neorealist-neoliberal orthodoxy on the Right and the more radical social constructivism of post-modernists,Footnote 1 critical theorists, and feminists on the Left. Indeed, the speed and spread of the constructivist insurgency between 1989 and 1999, and of its subsequent institutionalization in IR were remarkable.
Also remarkable was its bottom-up, generational aspect. This was not a movement of established academics who could easily afford to take intellectual and professional risks. It was a movement of mostly early-career researchers struggling to get tenure and graduate students hoping to even get a job. It was not easy getting our elders to take the ideas seriously, and sometimes it was downright unpleasant. Fortunately, there were exceptions to the generational rule, particularly Peter Katzenstein, prominent political economy scholar and former editor of International Organization (IO). After a post-Cold War epiphany, he grasped the needs of the moment and put together The Culture of National Security (Katzenstein Reference Katzenstein1996), which as we will discuss later proved to be a pivotal intervention in the debate. But Katzenstein will be the first to tell you that it was his graduate students who were leading the way.
In this light consider one data point and one trend. The data point is that as of 2024, seven of the top eleven all-time most cited articles in IO were constructivist papers.Footnote 2 The trend is from several “Teaching and Research in International Politics” (TRIP) surveys of over 1,000 IR scholars between 2004–2017.Footnote 3 Respondents were asked which of an array of theories best described their own approach to IR.Footnote 4 In 2004 constructivism came in fourth with 15%, behind liberalism, realism, and Marxism. By 2011, however, it had jumped to second place with 22%, putting it ahead of all others except “non-paradigmatic.”Footnote 5 This “changing of the guard” was highlighted in a 2012 article in Foreign Policy magazine surveying changes in IR’s Ivory Tower (Avey et al. Reference Avey, Desch, Long, Maliniak, Peterson and Tierney2012), which noted that “the surge in constructivist work, which focuses on the role of ideas and identity in shaping state preferences and international outcomes, started in the 1990s and shows no signs of leveling off.” Constructivism’s share increased to 25% by 2017, the last year for which TRIP data are available, and on a global level there are reasons to think that might be an under-count (Zarakol Reference Zarakol2017). Moreover, a full twenty-five years since the 1990s, constructivism is well represented in 14 of 15 recent undergraduate IR textbooks we surveyed. That means most IR students now enter graduate school knowing what constructivism is and some are already interested in doing such work themselves. Contrary to claims that it was a flash in the pan, fantasy theory, fad, mood, evil (Lake Reference Lake2011), or just dead, constructivism seems here to stay—in short, an academic success story.
In business schools, success stories often become case studies for students eager to come up with the next big thing. We can’t promise any such wonders here, but trying to understand constructivism’s success may shed light on how or whether it might be replicated. To that end, we take inspiration from Benjamin Cohen’s (Reference Cohen2008) intellectual history of International Political Economy (IPE) but scale it down to article size and add a constructivist twist. Like him, we are interested in understanding the emergence of a new body of scholarship. But while Cohen’s narrative foregrounds the agents or “intellectual entrepreneurs” who worked purposefully and skillfully to create IPE, ours highlights anonymous social structures that made it possible to construct constructivism in the first place.
Our structural starting point reflects our holistic, constructivist worldview, but we think it also makes good sense of the case. Even though many IR scholars quickly embraced “constructivism,” initially it did not have a clear or settled meaning. This may be because the word had little history in the social sciences. Most talk of “social construction” (two words) originated in sociology, where it is so widely accepted that it defines aspects of the discipline itself (e.g., Berger and Luckmann Reference Berger and Luckmann1966). But “constructivism” as a one-word “ism” does not seem to be used there or in any other social science except education theory. So, constructivism entered IR as something of an orphan and quickly became a mongrel of selective borrowing from philosophers like Wittgenstein (see especially Onuf Reference Onuf1989 and Karin Fierke Reference Fierke1996, Reference Fierke2002), and sociologists like Margaret Archer, Roy Bhaskar, Pierre Bourdieu, Anthony Giddens, Margaret Gilbert, Erving Goffman, and John Meyer—and going back further, Durkheim, Weber, and Mead. Constructivism is, indeed, what constructivists make of it, so everyone’s version of it is a little different. Nonetheless, we think a rough consensus had formed by the late 1990s around a structure-facing meaning that became “mainstream” constructivism.
Time marches on, and mainstream constructivism is now apparently “old” constructivism (e.g., McCourt Reference McCourt2022). Despite, or perhaps because of, its popularity, multiple variants of constructivism quickly emerged alongside the mainstream and started generating literatures of their own—securitization theory (Waever Reference Waever1993, Reference Waever and Lipschutz1995); ontological security theory (Kinnvall Reference Kinnvall2004; Steele Reference Steele2005; Mitzen Reference Mitzen2006), the practice turn (Adler and Pouliot Reference Adler and Pouliot2011), relationalism (Jackson and Nexon Reference Jackson and Nexon1999; Kurki Reference Kurki2020), and more.Footnote 6 All were recognizably “constructivist,” but they were also autonomous and to varying degrees favored a more processual over structural interpretation of the approach. In our view the distinction is less of a binary than a difference in emphasis or research question, since even the most structural constructivists recognized the importance of agency and process—or else they would not be constructivistsFootnote 7—but the difference seems important enough to justify two forms, old/structural and new/processual. Given the negative valence of “old” and being of a certain age ourselves, we will use “early” as a descriptor of the constructivism constructed in the 1990s and argue in the conclusion for the moniker “classical” instead.
In this article we are interested in making sense of the surprising rise of constructivism between 1989–1999,Footnote 8 especially in the United States (which until the mid-1990s was our primary experience of the field). Naturally, our approach will be a constructivist one. As such this article is not a review essay; there have been several good ones on constructivism already.Footnote 9 Nor will we try to defend our work against its critics; that, too, appears elsewhere. Our aim here is simply to use constructivism to help us understand the development of constructivism, and thereby shed some light on how we reached such a surprising point several decades on.
That is a long time, and we should note that our own views have evolved since this journey started. Wendt “got out of the constructivism business” around 2004, and in 2006 wrote an “auto-critique” of his earlier work from a quantum perspective. That led to a book on quantum social science (2015), which he thinks provides a physical basis for a quantum constructivism. For her part, Finnemore was always a reluctant paradigm warrior. Her concerns were, and continue to be, pragmatic. She just wants to make sense of the weirdness of the world and the puzzling politics that pervade it. Not having to fight all the time about whether social construction exists is an enormous relief.
Given the occasion of this Skytte Prize, we want to begin with those who came before constructivism, without whom none of us—new or old—would be here today. In the first section we therefore briefly acknowledge the scholars whose work most influenced us (and many others) in graduate school. Then, in the spirit of early constructivism, we first take a structural perspective, identifying several structural changes that created openings for constructivist scholarship. We then turn to the agency of constructivist IR scholars, who advanced scholarship on two fronts simultaneously, one social theoretical and the other empirical, which we take up in turn. From this we distill some of the theoretical gains that constituted the mainstream definition of constructivism, focusing on three core propositions that made constructivism new in the 1990s and distinguish it from other IR theories. We next examine how researchers using constructivsm “in the field” to understand real-world puzzles developed and refined these ideas. The conclusion considers the institutionalization of constructivism after 1999 and makes the case for a “classical” moniker for the constructivism of the 1990s.
Before Constructivism
The five years between 1979–1984 saw a veritable “Cambrian Explosion” of IR theory, the results of which are still visible today. Most of the major theoretical forms and fault lines in the field today emerged during that time. The explosion was triggered by the introduction into IR of a new kind of theory, social theory, and its philosophical concerns with epistemology and ontology. The introduction took place in two waves, depositing different and perpetually contending forms of social theory in their wake. The first was economistic or rationalist social theory,Footnote 10 which was followed by a slower, second wave of sociological and critical social theory. The growth of both kinds of social theory after 1979 brought IR scholars into regular contact with foundational conversations taking place in other social sciences and the humanities too. That eventually turned IR from a theoretical backwater in political science into a major entrepot and producer of diverse streams of theory—a larger success story of the whole field, of which constructivism is just one part.
In graduate school (1982–1989) Wendt processed everything through classmates and future constructivists Michael Barnett (Reference Barnett1993, Reference Barnett1995) and Jutta Weldes (Reference Weldes1989, Reference Weldes1996; Weldes and Saco Reference Weldes and Saco1996), all of us working with Raymond (Bud) Duvall (see Barnett and Duvall Reference Barnett and Duvall2005), the father of the “Minnesota School” (for which he will take no credit). Besides sociologists Roy Bhaskar (Reference Bhaskar1979) and Anthony Giddens (Reference Giddens1979), four IR scholars loomed largest for Wendt: Kenneth Waltz, Robert Keohane, Richard Ashley, and John Ruggie.
Waltz’s Theory of International Politics (Reference Waltz1979) started it all, reconstructing classical political realism as a fusion of rationalism, materialism, and positivism. The power of neorealism found expression most clearly in the book’s influential definition of international system structure. The “distribution of capabilities under anarchy” was not just constructivism’s originating Other, but neoliberalism and post-structuralism’s Other as well.
Keohane’s landmark After Hegemony (Reference Keohane1984) pushed back on Waltz’s materialist view of structure, carving out some space for non-material factors like ideas and international institutions to matter (also see Goldstein and Keohane Reference Goldstein and Keohane1993). However, Keohane’s challenge to realism was partly blunted by the rationalism he shared with Waltz.Footnote 11 While that agreement facilitated productive debates between neorealists and neoliberals, it also narrowed considerably the scope of IR theorizing.
For those who prefer their IR theory as a full-contact sport, 1984 also saw Ashley’s “The Poverty of Neorealism,” at 61 pages one of the longest and surely most intellectually challenging articles ever published in an IR journal.Footnote 12 Drawing on Foucault and Habermas, Ashley went after both pillars of Waltz’s structuralism, its rationalism and its materialism. In its place he offered a highly original, performative conception of structures like the balance of power as “grammars” for practice, as opposed to mechanical equilibria. Ashley’s anti-positivism alarmed many, however, that IR was under attack by relativist post-modernists and other “dissidents” (Ashley and Walker Reference Ashley and Walker1990), including Gramscian Marxism (Cox Reference Cox1981, Reference Cox1983), and radical feminism, just entering the field at that time as well (Cohn Reference Cohn1987; Enloe Reference Enloe1989; Peterson Reference Peterson1992; Tickner Reference Tickner1992).Footnote 13 Critical and post-modern theorizing is ontologically constructivist (and of the “new” variety), but based on an anti-foundational epistemology that gave their development a distinct trajectory from “constructivism.” We return to this divergence later.
Finally, there was Ruggie, who first pioneered the middle ground that gave rise to constructivism such that, were he here, he should be sharing our Skytte prize today.Footnote 14 Early in his career, Ruggie (Reference Ruggie1975) introduced IR to two key concepts, epistemic communities and international regimes. Both highlighted the power of human consciousness and meanings, especially shared, intersubjective ones. Ruggie (Reference Ruggie1983) developed his ideational sociological approach further in a long review of Waltz’s book, in which Durkheim took center stage; and in his classic article on “embedded liberalism” (1982), which defined the structure of international political authority as the “fusion of power and legitimate social purpose”—with emphasis on legitimacy, long neglected in IR in favor of power.
He followed up with an influential article with Friedrich Kratochwil (Reference Kratochwil and Ruggie1986) highlighting the tension between a positivist epistemology and an inter-subjectivist ontology (cf. Lapid Reference Lapid1989). Their solution was an interpretivist approach that takes inter-subjectivity seriously, but without rejecting positivism wholesale, an epistemological moderation that became a defining feature of constructivism.
In grad school, Finnemore (1986–1992) was fed the same diet of Waltz and Keohane as Wendt (Ashley was assigned as an afterthought), most of which left her puzzled. She came to the study of IR from the policy world, having worked on Capitol Hill. None of these academic accounts of politics remotely resembled the lived world of politics and policy. Sociology, next door to political science in those days at Stanford and, crucially, sharing the same basement café where grad students congregated, offered more useful tools.Footnote 15 Macro-theoretical arguments like John Meyer’s about world culture offered quantitatively testable hypotheses that competed directly with those of the IR “neos,” and the organization theory pioneered by Jim March offered attractive tools for understanding practical problems of governing and institutional change (see Finnemore Reference Finnemore1996a, Reference Finnemoreb).
In short, IR theory of the 1980s was heady stuff, informed by a level of social theorizing, both economistic and sociological, almost completely unseen in IR before.Footnote 16 That did not make the emergence of constructivism inevitable, but it did prepare the ground.
Social Structures Conducive to Constructivism
We turn first to the structural side of our story, which highlights the role of three kinds of social structure in the rise of constructivism: political, disciplinary, and intellectual. As social structures, their effects were constitutive before they were causal, and non-deterministic. Structural causation is not mechanical or kinetic. It is about constituting subjects and affordances for their actions.
Political Structures
Two major changes in the structure of world politics in the 1990s created intellectual needs that constructivism was able to help satisfy. Both are common knowledge so we will just register them here.
The end of the Cold War. The end of the Cold War (1989) and subsequent collapse of the Soviet Union were both hugely consequential, but in different ways. The first temporarily suspended the security dilemma, the second changed the distribution of power. Neither was anticipated by IR scholars (including proto-constructivists), and so both came as a complete surprise. As Stefano Guzzini (Reference Guzzini2000, Reference Guzzini155) points out, however, the real problem was not the failure to predict, which was universal, but that the orthodoxy “did not recognize the possibility that [such things] could even happen in the first place.” Gorbachev’s New Thinking was clearly at the center of how all these changes started, providing constructivists with some of their first and biggest empirical opportunities.Footnote 17
The end of the Cold War ushered in two new realities in the 1990s: Great Power peace and comity, and unipolarity and neoliberal hegemony in the Gramscian sense (see Cox Reference Cox1981; Reference Cox1983). Ideational factors like norms, identity, culture, and legitimacy were central to these new realities, and constructivists were quick to explore how such factors might work. In addition to Katzenstein’s volume (Reference Katzenstein1996) discussed later, there was Biersteker and Weber (Reference Biersteker and Weber1996) on state sovereignty; Hall on moral authority (Reference Hall, Lichbach and Zuckerman1997) and on national identity (Reference Hall1999); Reus-Smit (Reference Reus-Smit1997, Reference Reus-Smit1999) on the moral foundations of the state and international system; Adler and Barnett (eds. Reference Adler and Barnett1998) on security communities; Bukovansky (Reference Bukovansky1999) on international political culture; Tannenwald (Reference Tannenwald1999) on the nuclear taboo and more. All were talking about structure, but nothing like the material structure of neorealism.
Global civil society. A second structural change creating opportunities for constructivist thinking was the dramatic expansion of global civil society and its diverse agents, most prominently NGOs. Often based in the North, most NGOs were focused on citizen concerns neglected by governments, particularly weak or repressive governments in the Global South. Relatively little studied in American IR before 1990, by the end of the 1990s NGOs and other non-state actors had proliferated in both numbers and influence and became a new and lively focus of scholarship. They presented IR scholars with diverse new actors; new, identity-based interests like development and human rights; and new, non-violent tactics like “naming and shaming.” Transnational non-state actors successfully framed many new problems for states which IR scholars could study. Sociologists had been studying identity politics and social movements domestically for decades and offered an array of relevant concepts that could be adapted to global politics. Indeed, so good was the fit that much of the best early constructivist scholarship was about non-state actors rather than states, like Margaret Keck and Kathryn Sikkink’s Reference Keck and Sikkink1998 classic and best-selling book, Activists Beyond Borders.
Disciplinary structures. Though nothing like the massive social and political transformations just discussed, we believe two changes in the structure of IR as an academic discipline also played roles in the emergence and consolidation of constructivist ideas.
IO. International Organization’s widely accepted structural position since the 1980s as the #1 journal in IR meant that publishing new ideas there had an outsized or “influencer” effect on the rest of the field. If an idea was taken seriously at IO, then it was good enough to be taken seriously anywhere. Not only did the journal help new ideas reach a wide audience, but through its imprimatur it gave them epistemic credibility as well.
Then as now, there was intense competition to publish in IO, which historically has had one of the lowest manuscript acceptance rates in the field (10% in the 2000s, decreasing to less than 5% today). An abundance of good papers gives editors discretionary power, which they can hypothetically use either conservatively, in defense of incremental improvements to extant theory, or more progressively, by taking more intellectual risks. We might expect conservative forces to be especially powerful, particularly at the top journal in any field. Those editors have less to gain from placing bets on controversial ideas or unfamiliar ways of thinking than would editors at lower ranked journals looking to pull off a coup. Social dynamics reinforce this caution. Editorial boards of top journals embody the mainstream in any field, since members are chosen for being excellent mainstream scholars. Boards at top journals therefore tend to be intellectually homogeneous and conservative, especially on fundamentals and what counts as “good work.” Non-mainstream manuscripts often do not conform to those expectations and struggle in the review process. This is not unique to IR; in the history of science, it is often the orthodoxy that resists change the most because it has the most to lose.Footnote 18
At IO, IR has been fortunate to have flagship journal editors who were willing to test boundaries. Robert Keohane (1972–1980) led a complete re-founding of the journal, creating an outlet for serious social science research on world politics. Peter Katzenstein (1980–1986) published Ruggie on embedded liberalism (Reference Ruggie1982), Ashley on neorealism (Reference Ashley1984),Footnote 19 and Kratochwil and Ruggie’s challenge to positivist IR (Reference Kratochwil and Ruggie1986). Stephen Krasner (1986–1992) is responsible for Wendt—just a grad student—on the agent-structure problem (Reference Wendt1987), David Dessler’s response (Reference Dessler1989), Peter Haas on epistemic communities (Reference Haas1989), Ruggie (Reference Ruggie1992) on multilateralism, Wendt on anarchy (Reference Wendt1992), and a special issue of IO on epistemic communities edited by Peter Haas (Reference Haas1992). John Odell (1992–1996) gets credit for Finnemore (Reference Finnemore1993) on norms, Guzzini (Reference Guzzini1993) on structural power, Hall and Kratochwil (Reference Hall and Kratochwil1993) on culture and medieval IR, and Ruggie (Reference Ruggie1993) on territoriality; Barkin and Cronin (Reference Barkin and Cronin1994) on identity and the nation-state, Crawford (Reference Crawford1994) on Iroquois security community, Kozlowski and Kratochwil (Reference Kozlowski and Kratochwil1994) on the Soviet collapse, and Risse-Kappen (Reference Risse-Kappen1994) on the end of the Cold War, and so on. Reasons for rejecting all these papers could have been found, but they were accepted instead.Footnote 20
As far as we can tell, IO’s early openness to constructivism was not due to any special sympathy for the approach,Footnote 21 or even for intellectual diversity for its own sake. But it did involve the editors accepting that “good work” could take forms different than their own. This enabled IO to play a unique role in bringing constructivism into being. Did that leadership in turn benefit the journal, by making it a site for ensuing debate? Looking back, that case is easy to make: just recall the seven constructivist articles among the top eleven. But that pay-off could not have been known beforehand.
The rise of European IR. Footnote 22 Although IO was important, it also mattered for early constructivists that in the 1990s a coherent European IR community emerged much more fully onto the intellectual and institutional scene. The first Pan-European IR conference was held in 1992, and intellectually it turned out that European IR scholars were considerably more sympathetic to constructivism than to the rationalism dominant in the United States. This was due in part to Hedley Bull (Reference Bull1977) and the English School,Footnote 23 which had long recognized the “social” quality of international politics, as well as to the rise of securitization theory. A further impetus came out of Germany, where scholars had been debating for a decade (in German) the relevance of Jürgen Habermas’ work on norms and legitimacy for IR. In a breakthrough article, Thomas Risse (Reference Risse2000) synthesized this debate for English-speakers around the “logic of arguing,” which led to many further contributions in this vein.
Initially there were relatively few European IR journals for all this new output. Millennium and the Review of International Studies had anchored British IR for decades, the former an early supporter of critical IR theory especially, and the latter open to constructivism from the start. Less well-known to Americans were Cooperation and Conflict and Security Dialogue, which came out of Scandinavian peace research, as well as Journal of International Relations and Development, though the profile of all these journals rose considerably in the 2000s.
Thus, it was a big change when in quick succession European-led teams launched two major new journals, Review of International Political Economy (1994) and European Journal of International Relations (1995). These journals provided a friendly home for constructivist papers when there were few openings in American venues. More importantly, European IR journals collectively constituted a conversation space with enough breadth and depth for readers over time to really dig into the many issues that constructivism raised—to develop the theory rather than just deflecting attacks. That improved the quality of constructivist work and fostered the growth of an intellectual community in which constructivism was increasingly taken for granted as a starting point for conversation, not just as a footnote. Globally IR might still be an “American discipline,” but there is also now a “European discipline,” the story of which is partly about constructivism in its different forms.
Scholarly Debates and Constituting the “Middle Ground”
In addition to political and disciplinary structures, in the 1990s the contours of debate in IR theory itself were conducive to constructivism. With the neorealists and neoliberals on the right and the exiled post-structuralists on the left, there was a wide-open middle ground to claim.
The neo-neo debate and generational change. During much of the early 1990s mainstream IR was preoccupied with whether states are more concerned with relative or absolute gains, an issue Joseph Grieco (Reference Grieco1988) had raised in a neorealist critique of neoliberalism. It was an important question that spoke to the causal powers and durability of international institutions. But it was also very narrow, especially as a focal point for years of theoretical effort by some of the best minds in the field. It never caught on outside the United States.Footnote 24
The narrowness of American IR’s theoretical concerns contrasted starkly with the sweeping transformations going on at the time in the real world. In those same years (1989–1994) world politics changed more dramatically than it had at any point since 1945. That apparently did not change many minds among senior American IR faculty,Footnote 25 but in our experience many graduate students thought the moment demanded a theoretical, even paradigmatic response. If the theory we were being taught (or were teaching) could neither foresee nor explain the biggest global political change in decades, then what was it good for? With a lumbering, unengaged IR mainstream at such a critical juncture, in retrospect the insurgent appeal of constructivism is not surprising, particularly for younger scholars.
The exile of post-structuralism. As we have seen, constructivists were not the first to challenge the neorealist-neoliberal axis. That distinction belongs to three forms of critical theory in the 1980s: Gramscian Marxist, post-structural, and feminist, each in its own way social “constructivist” in the broad sense. Post-structuralists seemed to be the largest of these communities and reached an American high-water mark in the early 1990s. International Studies Quarterly (ISO) did a special issue that year on “dissident” IR edited by Ashley and Walker (Reference Ashley and Walker1990), and two years later Campbell’s (Reference Campbell1992) Writing Security came out, which became an almost instant classic.Footnote 26
In the aftermath of these successes, however, post-structural IR scholarship in the United States did not receive the kind of grudging, guarded, but ultimately mostly accepting reception that constructivism got. At IO, Ashley’s pioneering 1984 article was not followed by other theory papers,Footnote 27 nor (with the exception of ISQ) did other American journals take up the slack. Indeed, the two trends may have been connected: constructivism perhaps rose in popularity in part because it was not post-structuralism but was seen as a tamer version of it. Whatever the cause the contrast was striking since, ontologically speaking, both were social constructivist, differing only in degree.Footnote 28
Epistemologically, however, post-structuralists and constructivists had different responses to the fundamental tension in social science identified by Kratochwil and Ruggie (Reference Kratochwil and Ruggie1986). Positivist epistemologies assume that a clear distinction between subject and object can be drawn, allowing the researcher to take themselves out of the picture and treat actors as if they were just objects (or machines). Inter-subjectivist ontologies, in contrast, propose that social scientists too are subjects, and in the case of IR, we are members of the very objects that we are studying, like states, and the international system or community.
Whereas both constructivists and post-structuralists rejected old-fashioned positivism and naive falsificationism, constructivists retained a belief that their work was scientific (broadly defined), and as such should strive for objectivity, rigor, and all the rest. It might be a realist social science in search of causal mechanisms (Dessler Reference Dessler1991; Wendt Reference Wendt1999; Wight Reference Wight2006) or a more interpretivist one looking for meanings (Andrews Reference Andrews1975; Kratochwil Reference Kratochwil1989; Onuf Reference Onuf1989)—or both (Hollis and Smith Reference Hollis and Smith1990). Either way, constructivists took the modernist view that truth claims in social science should be constrained by reality, which is best discovered by pitting competing hypotheses against each other in a three-cornered fight with the evidence. Post-structuralists—or maybe post-modernists is more appropriate here on this epistemic terrain—did not accept this model of social inquiry. Rather than hang onto illusions of objectivity or test competing hypotheses, they embraced a participatory epistemology, in which social scientists are partly responsible for constituting the reality they are studying. That does not mean “anything goes” in empirical work, as the extended methodological reflection about discourse theory attests (e.g., Milliken Reference Milliken1999; Hansen Reference Hansen2006; Banta Reference Banta2012). But it was a rejection of “rigor” in the positivist sense of methods that either create or presuppose a dualism of subject and object (cf. Klotz and Lynch Reference Klotz and Lynch2007).
These distinct responses to the epistemological tensions inherent in social science in turn had sociological implications for potential conversations. Because they accepted the idea of a social science, constructivists had a lot of common ground methodologically at least with those positivists (call them neo-positivists) who tried to bring ideas into their work. That is not to minimize the differences between, for example, seeing ideas as material causes, hooks, or switches (e.g., Goldstein and Keohane1993) versus as constitutive of social reality (e.g., Onuf Reference Onuf1989). But in practice, empirically, there was much that neo-positivists and constructivists could productively talk about.Footnote 29 This was less true of neo-positivists and post-modernists, neither of which seemed to see any point in talking at all. There was more potential between constructivists and post-modernists, but looking back, narcissisms of small differences, and a few big real ones, ensured that their relationship in the 1990s would be more one of suspicion and neglect than collaboration.Footnote 30
The upshot is that through a combination of exclusion and choice, post-structuralists in the 1990s were essentially “exiled” from American IR theory, mostly to Canada and Europe. There they found many conversation partners among fellow exiles in feminist theory, critical theory, post-colonial theory, and international political theory. But conversation with the U.S. IR mainstream was dead.
Enter constructivism, which in Adler’s (Reference Adler1997) classic synthesis fills the vacuum or middle ground created by the polarization of neo–neo insularity and post-structural exile. Constructivists in effect became the loyal opposition to the governing positivist mainstream, committed epistemologically to the idea of IR as a social science, but challenging its typical materialist and individualist ontology. Given the imbalance in numbers, framing constructivism as a “middle” position was brilliant and probably facilitated its growing acceptance. Almost by definition it was the reasonable progressive stance—playing Mensheviks to Bolsheviks on the left and Tsarists on the right.
Up to this point we have been telling the story of early constructivism from its characteristically structural perspective, highlighting a variety of contextual factors that afforded opportunities for constructivist ideas to prove their worth and gain a seat at the IR table. Yet all the structures in the world will not matter if agents—in this case IR scholars—do not act on them. To go further we will therefore flip the narrative and address the construction of constructivism from a more agentic perspective. Recall that constructivists advanced their ideas in IR along two main fronts. There was the social theory front, which we discuss in a moment. There was also the empirical front, which showed how constructivist and sociological concepts could be used in empirical research on real world politics, which we examine subsequently. Progress on both fronts was essential for creating constructivism as we know it today.
Constructing ‘Constructivism’
Constructivism in IR was not born fully formed but constructed through diverse, often serendipitous, discussion in the 1990s. What resulted was thus more inadvertent than intended, and its coherence has always been a work in progress. Constructivism has no canonical statement. It morphed quickly into a variety of binaries—causal versus constitutive, conventional versus critical, thin versus thick, and state-centric versus non-state-centric—not to mention the original modern versus post-modern binary. Here we are more interested in what constructivists produced in common than in how they differed. As such, we will focus on the outcome of the construction process, which on the theory front was a rough consensus definition.Footnote 31 Our discussion is organized around three “postulates,” two ontological and one epistemological, shared by most constructivists. But first a caveat.
Although frequently placed alongside realism and liberalism in IR, constructivism is incommensurable with them in two important respects. First, it has no clear normative agenda and thus no political theory associated with it, in the way that realism and liberalism do. In our view constructivism brings to the table a sociological counterpart to rationalism, not a normative theory.Footnote 32 Second, also unlike realism and liberalism, constructivism is not a theory of politics at all, but a social theory that IR scholars put together from the work of various philosophers and sociological theorists. As such, the elements of constructivism are portable abstractions like agents and social structures, or ideas and material conditions in general, not specific entities like states in the international system. The parallel here with rationalism—also a social theory (Satz and Ferejohn Reference Satz and Ferejohn1994)—is strong. Just as rational choice theorizing needs to specify agents (states, voters, leaders), their beliefs and interests, and institutional contexts before it can proceed, so too does constructivist theorizing (Finnemore Reference Finnemore1996a, Reference Finnemore27). When IR scholars put the three theories side by side as equivalent offerings, they obscure both the shared rationalism of the neos and the qualitative differences with constructivism.Footnote 33
First Postulate: The Constitutive Power of Consciousness and Ideas
In both the political economy and security domains, through the 1970s materialism was the dominant postwar ontology in IR theory, particularly realist or, less often, Marxist. Primary explanatory importance was given to objective, seemingly material realities like economic and military capabilities, technology, human nature or biology, and geography. Conversely, the role of non-material factors like identities, culture, and ideology was discounted or ignored. To our knowledge Waltz never described neorealism as a “materialist” theory (much less as an “ontology”), but in his 1979 book materialism in IR reached its apogee, with its almostFootnote 34 single-minded focus on the distribution of material power defining the structure of the system. Indeed, in Marxist parlance neorealism seems a “vulgar” materialism, because it did not give even “relative autonomy” to ideational factors. It was tanks and ICBMs all the way down.
Against this, since Ruggie’s first suggestion in 1975, constructivists have foregrounded human consciousness and its contents, which might be generically called ideas or meanings. By “consciousness” we mean subjectivity or experience, the feeling of “what it’s like” to have one’s own, unique point of view, without which we would just be robots or zombies. From such a phenomenological perspective, what matters are the ideas that give meaning to events or objects, not those events or objects per se.Footnote 35 In politics shared or inter-subjective ideas are of particular importance, including most (all?) genuinely social phenomena: language, identities, norms, rules, cultures, institutions, social structures, discourses, relations, practices, even (if tendentiously) strategic interaction. Importantly, the significance of these shared ideas is not that material conditions no longer matter. It is that how they matter depends on what they mean. This is a constitutive effect of ideas, not a causal one (Wendt Reference Wendt1998; cf. Onuf Reference Onuf1989; Goldstein and Keohane Reference Goldstein and Keohane1993).
A simple, widely cited example of the constitutive effect of ideas is that five North Korean nuclear weapons were (and still are!) more threatening to the United States than 500 British ones (Wendt Reference Wendt1995). A more significant example is the end of the Cold War, which provided a real-world demonstration of the power of ideas. The West’s growing technological lead clearly mattered, but Gorbachev’s “New Thinking” understood that the reason it mattered as a threat was its implications for the Cold War. Take that shared belligerent understanding away, and you take away the material threat.Footnote 36 A final example of the power of ideas is the state itself, the elementary unit of most IR scholarship. Although with modern technology states can wreak unprecedented material destruction, ultimately their power depends on the strength of their Idea of the state, which holds especially the security forces together as a cohesive monopoly of force. Break that cohesion, and even the strongest states may quickly collapse or “fail,” as the USSR did in 1991 while hardly firing a shot. Ultimately the state is a state of mind, not a material object.
More generally, ideas and meaning have had perhaps their biggest influence in the large swath of constructivist scholarship dealing with the construction and effects of identities in world politics. One of the most important effects of identities is to constitute national and other interests, since actors cannot know what their interests are until they know who they are (Finnemore Reference Finnemore1996a, Reference Finnemore29). There are different kinds of identities, however, four in Wendt’s (Reference Wendt1999, Reference Wendt224) typology, including corporate, type, role, and collective identities, to which might be added a fifth, social, in the sense of Social Identity Theory (see Mercer Reference Mercer1995). These kinds are constituted in different ways and play different roles in social interaction, and deal with power in different kinds of ways (see especially Bially Mattern Reference Bially Mattern2001), all making a quick summary impossible. But this conceptual and empirical complexity provided fertile ground for constructivists, who remain actively interested in the construction and effects of identities.
Second Postulate: The Mutual Constitution of Agency and Structure
In addition to assumptions about consciousness and ideas (or the lack thereof), arguably every piece of social science also makes tacit ontological assumptions about the agent-structure problem, or how the actors in a system relate to the social structures in which they are embedded. This is the social science version of the part-whole problem in philosophy, where there are three broad categories of approaches on offer. Call one agent-first or bottom up, where the properties of the whole are expected to be reducible to the social interactions of independently existing parts (varieties of individualism). The second is structure-first or top-down, where the identity of the parts exists only in relation to an emergent social whole (varieties of holism and structuralism). And the third, arguably the hallmark of all forms of constructivism (both modern and post-modern) tries to have it both ways: (what are called) agents and social structures are “co-determined and mutually constitutive” (Giddens Reference Giddens1979), and their priority handled methodologically either by “bracketing” one and then the other (early constructivism), or by trying to dissolve or collapse the dualism altogether into pure processes (new constructivism, along with post-structuralism before). This is an abstract debate for sure, but also a political one, with real distributive stakes. Margaret Thatcher, that noted social ontologist, justified her deep cuts in the British welfare state with her famous assertion that “there is no such thing as society,” only individuals.Footnote 37 And where there is no society there is no social structure, and thus no possibility of structural violence or injustice either.
This is not the place to review the substantial literature in IR on the agent-structure problem.Footnote 38 Suffice it to say that the “mutual constitution” formula is not a simple matter of giving both individualism and holism their due and then declaring victory and going home. Indeed, despite the intuitive attractiveness of a balanced approach, individualism remains the dominant working ontology of mainstream social science. The burden of proof has always been on holism (and today relationalism too) to go beyond individualism, which only a minority think has been met.Footnote 39
The reason for the asymmetric burden is the materiality of human biology versus apparent immateriality of social structure, as evidenced by its invisibility. Being born into the world is a process of biological individuation, and it is in that material state—separated by our skins—that as agents we live our lives. Moreover, even though consciousness and meanings cannot be directly observed, it feels like they are inside our brains, which are inside our skin as well. In short, biology seems to give ontological priority to “rump” material bodies—to humans as animals—over social structures. This is not to deny that most of the thoughts in our heads have their origins in a lifetime of socializing with others. No one denies the role of socialization, not even Waltz (Reference Waltz1979: 74-7). The problem for holism is constitutional. At any given moment, our bodies and minds seem to be constitutionally independent of others and society as a whole. If society is nothing but individuals, then how could we constitute (as opposed to cause) thoughts inside each other’s heads?
Yet countless examples from contemporary relationalism to externalist philosophy of mind (Wendt Reference Wendt1999, Reference Wendt173-76) all the way back to Hegel’s master and slave, show us that society does constitute us as agents, even inside our heads (cf. Meyer and Jepperson Reference Meyer and Jepperson2000). So the question is how is that possible physically if social structures are invisible? That is a hard question to answer, so hard that Wendt (Reference Wendt2015) thinks it is worth exploring the possibility that consciousness and society are macroscopic quantum phenomena. This would enable us to say that, physically speaking, our minds are entangled non-locally in a web of meaning that penetrates our skins, supporting the “holographic” suggestion above that wholes are in their parts (see especially Pan Reference Pan2020). Although gaining traction in cognitive science, in IR the idea of a quantum social science has been met with considerable skepticism—and occasional enthusiasm—from positivists and critical theorists alike, so its fate remains to be seen.Footnote 40 Failing quantum, however, it remains unclear how else holism and relationalism could be physically possible.
Third Postulate: Taking Both Science and Inter-Subjectivity Seriously
We will be briefer with the third postulate, since we covered related ground previously. This postulate is epistemological rather than ontological, and appropriately two-sided. On one hand, there was a commitment among early constructivists to doing social science, broadly defined. This would be unremarkable except that there was not such a clear commitment among post-structuralists, making this one place where the fellow travelers parted ways. In the constructivist view the general principles governing its own knowledge production are the same as those for social scientists anywhere: the goal of inquiry is to make sense of the world around us; there are social facts in the world, which constrain the kinds of knowledge claims we can legitimately make; empirical research is important to establish those facts and as such should be as methodologically rigorous as possible; and rival explanations and interpretations should be engaged rather than ignored.
On the other hand, epistemology should follow ontology, in the sense that how we try to know our objects of study should reflect what kind of objects they are (Kratochwil and Ruggie Reference Kratochwil and Ruggie1986, Reference Kratochwil and Ruggie764-69). In the case of social science, our objects are in important part human experiences and the ideas that help constitute them and give them meaning, which is different than the objects of chemistry or geology. Consciousness has an “inside” to it that rocks do not, and so there are always “two stories to tell” about social life, from the inside and from without (Hollis and Smith Reference Hollis and Smith1990). As such, while positivism may be a great epistemology for the physical sciences, it is not obviously so great for the human sciences. Bringing knowing and being into alignment does not require throwing the positivist baby out with the bathwater, however, but complementing it with interpretivist research that seeks out shared meanings and their constitutive effects on all social life—a task that constructivists have taken on and advanced significantly.
What’s In a Name?
We like to think that the take-off of constructivism had something to do with the content of its ideas, their attractiveness, cogency, and portability—and of course with the skill of constructivists scholars in arguing for them, both empirically and theoretically. However, in writing this essay we have come to wonder about another, more unusual factor. That is the aesthetic and normative pull of the word “constructivism” itself, which summarized all these ideas under a single name.
In the late 1980s alternative names for the nascent ideas were less than ideal. Keohane’s (Reference Keohane1988) navel-gazing “reflectivism,” Giddens’ (Reference Giddens1979) unwieldy “structurationism;” Abbott’s (Reference Abbott2001) inelegant “constructionism;” even “social constructivism” are all inferior to the aesthetic appeal, economy and rhetorical force of simple “constructivism.” Some of that force comes from the positive valence of “construction,” which connotes building or creating, not tearing things down as in de-construction. Similarly, if the opposite of “constructed” is “natural,” then comparatively speaking constructivism embodies greater optimism about the possibility of social and political change. If an unjust institution was once constructed, then it can be un-constructed—if not right away, then eventually. “Natural,” in contrast, suggests inevitability and thus encourages acceptance of the status quo. To that extent, a belief in the power of human beings to change the world is built into the “normative slope” of constructivism (see Taylor Reference Taylor, Laslett and Runciman1969, 40ff).Footnote 41 That makes saying “I am a constructivist” in a way empowering, a political statement, but one sufficiently general to attract almost anyone. For which of us, through our work, does not hope to leave the world a little better than it was before?
All of which raises an interesting counterfactual. All other things being equal, would constructivism (sic) have coalesced and become as successful as it did, if Onuf had not given us the word? We are not so sure.
As a further illustration of the power of words, in the same article that popularized Onuf’s term, constructivism also got a punchy slogan in “anarchy is what states make of it.” The title was so transparent that students did not even have to read the article to know what it was about, but not just that. Like the normative slope of “constructivism,” consider which image of world politics is more empowering: a deterministic world in which states cannot escape the vise grip of the “logic of anarchy,” or a contingent world in which there is no single logic of anarchy, just self-fulfilling prophecies between which states are free (jointly) to choose? Yet the slogan was fortuitous, saved only by Ruggie’s agency as a reviewer (as it later turned out). The paper was originally submitted under its eventual sub-title, but Ruggie suggested making one of the section headings the title of the whole paper—and the rest, as they say, is history. Would “The Social Construction of Power Politics” have had the same impact? We are not so sure about that one either.
We turn now to the second, more empirical track along which constructivists advanced. Ultimately it was this more than the social theory that was decisive in constructivism’s success.
Constructing a Research Program
In his 1988 ISA Presidential Address, Bob Keohane laid down a challenge: new theoretical ideas about human subjectivity (sic) were “making telling points about rationalistic theory but, have so far failed to develop a coherent research program of their own” (Keohane Reference Keohane1988, Reference Keohane379). Being taken seriously as a competing theory required more than social theoretic arguments. After all, social theorists had been floating ideas about intersubjectivity and human consciousness for at least two decades. The challenge Keohane correctly identified was how to put these ideas to work understanding real-world empirical problems. That was where the constructivist “rubber” hit the road, metaphorically speaking. Constructivist research had to explain problems that rationalists could not, or they had to provide better, fuller explanations than existing rationalist alternatives.
Constructivism met this challenge in the early 1990s. Within a decade of Keohane’s challenge, constructivists were publishing empirical work in top outlets on topics ranging from nuclear deterrence and military doctrine to European integration, monetary policy, and human rights. Their embrace of these new ideas is, perhaps, not surprising. Constructivism was attractive in this environment when the world was changing but IR theorizing was stuck in narrow debates among the “neos.” It was not just the name, “constructivism,” that was attractive; it was also the substance. Constructivism was a relatively plastic social theory. It could be applied to all kinds of politics—international, but also national, regional, and local; interstate, but also substate and non-state. In a world where statist ontologies of billiard balls were being discredited, this plasticity was a big plus. It meant researchers could grapple with diverse types of actors across diverse levels of analysis (Checkel Reference Checkel1997). This, in turn, facilitated more IR engagement with comparativists and area specialists who for the most part had never abandoned their concern with culture, norms, and identities as important political phenomena.
IR, in its rationalist “neo” incarnations, had prided itself on being a domain apart. Anarchy was understood to create unique conditions in international politics that required unique theories. One reason Wendt’s Reference Wendt1987 and 1992 papers were so important was that they exploded that claim. Another attraction: constructivism was compatible with many different research methods. Much of the constructivist work done over the past thirty years has been qualitative, but there is nothing inevitable or necessary about this. The sociologists who pioneered constructivism’s underlying social theory used (and continue to use) quantitative methods of diverse types (e.g., Meyer, Ramirez, and Soysal Reference Meyer, Ramirez and Soysal1992; cf. Barkin and Sjoberg Reference Barkin and Sjoberg2017.)Footnote 42
These early empirical researchers were doing more than simply applying extant theory in some cookie cutter fashion. Each new deployment of constructivist ideas expanded, refined, and developed those ideas. Each generated new constitutive and causal claims upon which others could build. This was the central task that skeptics had identified. Initial constructivist theory formulations often looked vague. How do you know a norm (or identity or idea) when you see one in evidence? How do you know it is the norm (or identity or idea) that is doing the work that creates outcomes? Those were the tough research nuts to crack, but they were the questions empirical researchers always must answer to persuade reviewers and colleagues. Empirical constructivist research needed methods that could be defended, both among themselves and to the rationalists. These early papers began developing those methods (or cracking those nuts, so to speak). They did it in different ways, across different problems, but that diversity created rich soil for future work.
One happy outcome of intense rationalist skepticism was to force constructivists to be much more explicit and sophisticated in their methods, particularly qualitative methods. The Institute for Qualitative and MultiMethod Research, founded by Colin Elman and others in 2002, has trained two decades of young scholars. Jeff Checkel and others championed these qualitative methods issues at the big professional associations (ISA and APSA). The resulting array of publications, workshops, seminars (and later webinars) put constructivists at the vanguard of qualitative and interpretive methods in IR. By focusing on how to read and interpret texts or how to conduct and interpret interviews in rigorous ways, constructivists raised the bar for the entire IR field.
We cannot do justice to the diversity of empirical constructivist work emerging in the 1990s, but we briefly highlight three empirical research programs begun in that period that continue to flourish and expand today: Katzenstein’s frontal assault on security issues and the research on norms and identities that followed; constructivist IPE with its links to especially British IR via Susan Strange; and global governance, which has been both a topic of research and an impetus for constructivist theory “going global” in the twenty-first century. Each of these research programs figured out innovative ways to frame new questions or offer different answers to old questions. Some constructed new research methods: others used old methods in new ways or to new purposes. The conversations these research programs initiated still inform our doctoral training and our theoretical debates today.
Norms, Identity, and The Culture of National Security
Peter Katzenstein’s crucial role in constructivism’s launch as a research program is hard to overstate. He saw the potential in constructivist ideas in the early 1990s and organized a cadre of young scholarsFootnote 43 to challenge realists on their security home turf. The result was the volume, The Culture of National Security (1996). Both substantively and collegially, this project was crucial. In a series of three multi-day meetings over eighteen months, funded by the SSRC, that group thrashed out a shared understanding of what constructivism meant for them, defining terms and concepts. The volume then offered a collection of empirical papers demonstrating constructivism’s utility, either for explaining security phenomena that realists could not, or for challenging realist explanations and exposing their weaknesses. While the product was important, so, too, was the writing process. It forged a community of scholars, subsequently employed at diverse institutions, who continued to propagate constructivist ideas. This group gave a shot in the arm to two lines of research already underway, one focused on norms, another on identity.
Research on norms quickly became a staple of constructivist empirical work but constructivists had no monopoly on norms, which makes it a slightly odd arena for constructivists to claim terrain. Rationalists, drawing on economics, already offered influential accounts of norms and how they may influence political behavior, mostly through incentives (positive or negative), reputation, and information (Axelrod Reference Axelrod1986; Milgrom, North, and Weingast Reference Milgrom, North and Weingast1990). For constructivists interested in the power of changing ideas and social values, these formulations were thin and insufficiently social. They ignored the crucial constitutive work norms did, reconfiguring interests and identities of actors, and constructing new understandings of both the possible and desirable. For understanding wholesale social transformation, which seemed common in the 1990s, constructivist tools were more useful.
Just as rationalists frequently looked to economics for their theories, constructivists often found theoretical inspiration and tools in sociology. Indeed, much of the constructivist project can be seen as a rediscovery of sociological concerns, long neglected in political science and IR. Particularly important to constructivist norms research were Jim March and Johan Olsen’s (Reference March and Olsen1989, Reference March and Olsen1998) arguments about “logics of appropriateness” guiding decisions inside organizations. Unlike instrumentalist “logics of consequences” underpinning economic understandings of action, where actors ask “how can I get what I want”, logics of appropriateness shaped behavior by prompting questions about identity and social context. Actors don’t always know what they want and they often ask other questions like “what am I supposed to do now?” and “what kind of situation is this?” Answers to those questions depend on identity (“who am I?”) and social norms of proper behavior. Both logics shape behavior, as March and Olsen emphasize (and most constructivists would agree), but ignoring appropriateness, as rationalists did, left big explanatory power on the table, which constructivists scooped up. The common constructivist definition of a norm as “collective expectations for proper behavior of actors with a given identity” (Katzenstein Reference Katzenstein1996, Reference Katzenstein5) has clear roots in this logic of appropriateness formulation.
Finnemore and Sikkink’s Reference Finnemore and Sikkink1998 article on norm life cycles turns out to have done important work in facilitating the uptake of norms as a research topic. That article grew out of an ongoing argument between its authors. Finnemore’s early work drew on the macrostructural arguments from sociology and effects of “taken-for-grantedness” of social understandings, but Sikkink chafed at the structural focus of early constructivist arguments, insisting that the creative, often transgressive, agency of her transnational activists was crucial to social construction in the human rights field. The norm life cycle was their effort to connect the agency of “norm entrepreneurs,” like Sikkink’s activists, with institutionalized social structures in one conceptual frame using Wendt’s notions about mutual constitution. The life-cycle argument has been applied, refined, challenged, and amended many times over the years in applications across all kinds of political issues and beyond.
In addition to norms, the Katzenstein volume highlighted identity as a necessary focus of constructivist research. Indeed, identity was embedded within the volume’s definition of a norm. Logically, the two were intimately connected. “What should I do” is always depends on “who am I” and both depend on context. Figuring out how those connections worked—how actors understood themselves, their place in the world, who should do what in specified circumstances, and how those understandings changed over time—was a major task for early constructivists, both conceptually and empirically.
Early research on both norms and identity quickly showed the power and versatility of constructivist theory for tackling political transformation in different domains. Constructivists challenged conventional IR understandings of core concepts like “national interests” (Finnemore Reference Finnemore1993, 1996a; Klotz Reference Klotz1995; Bukovansky Reference Bukovansky1997) and sovereignty (Barkin and Cronin Reference Barkin and Cronin1994; Barnett Reference Barnett1995, Reference Barnett1998) showing empirically and theoretically how norms and identities constitute and transform both—and not necessarily only for the better. Janice Bially Mattern (Reference Bially Mattern2001, Reference Bially Mattern2004) highlighted the coercive role of “representational force” exercised through language in stabilizing the identity of security communities (cf. Fierke and Wiener Reference Fierke and Wiener1999). Understanding local effects of global norms was a major focus of work across a huge range of issues (Cortell and Davis Reference Cortell and Davis1996; Checkel Reference Checkel1999; Finnemore and Sikkink Reference Finnemore and Sikkink2001; Wiener Reference Wiener2004). Other work, much by the Katzenstein group, drilled down into specific security topics. Johnston demonstrated the cultural and ideational roots of realist theory itself with his analysis of Chinese realpolitik (Johnston Reference Johnston1995). Price (Reference Price1995) and Tannenwald (Reference Tannenwald1999) constructed quite different theoretical arguments about the creation of chemical and nuclear weapons bans rooted in the distinct social contexts of the two technologies. Bukovansky showed the power of “international political culture” to shape revolutions (Reference Bukovansky2002) and Kier (Reference Kier1995) demonstrated the power of culture and identity to explain dysfunctional (even disastrous) military doctrine. All this early work on identity and norms provided a foundation for other constructivist research programs going forward that built upon or reacted to it.
International Political Economy
Constructivist ideas are perhaps even more well suited to political economy questions than to security. Money is, after all, a social construction. So are markets and core market features like “confidence,” “trust,” and “credit-worthiness.”Footnote 44 Indeed, many of Ruggie’s original constructivist ideas were developed as tools to understand the “embedded liberal compromise” and its “social purpose” that underpinned postwar global economy (Ruggie Reference Ruggie1982). Constructivist research in IPE was assisted in part by connections with comparative political economy where scholars like Kathleen Thelen and Peter Hall were developing historical institutionalist arguments with strong ideational components that resonated with constructivism (Steinmo, Thelen, and Longstreth Reference Steinmo, Thelen and Longstreth1992; Thelen Reference Thelen1999; Hall Reference Hall, Lichbach and Zuckerman1997).
That many early constructivist IPE scholars focused on Europe was surely no accident. European integration has been one of the most consequential social construction projects of our time, involving the creation of new norms, new identities (like “European-ness”), and new institutional structures. Kathleen McNamara’s (Reference McNamara1998) work on the role of ideas in driving European monetary politics, and ultimately the creation of the euro, builds explicitly on constructivist ideas. Craig Parsons (Reference Parsons2002) used the EU construction process to refine arguments about ideational causes. Mark Blyth (Reference Blyth2002, Reference Blyth1997) built out the “embeddedness” of economic ideas in his examination of transformation of economic institutions. All of these political economy scholars faced skepticism, not just from rationalists but from economists for whom constructivist notions about intersubjectivity and constitutive causality were often alien.
Europeans were not just objects of study for IPE scholars. They increasingly became partners. This is particularly true of British international political economy where the towering figure was Susan Strange. Strange was not a constructivist. By her own account, she was not even a theorist—her interests were primarily empirical and analytical. Those interests, along with her background in policy and journalism, however, made her an astute analyst of power, and her insistence that power was fundamentally a relational and structural phenomenon dovetailed nicely with constructivist notions (Strange Reference Strange1988; Cohen Reference Cohen2008, ch.2). The institutions she helped create in Britain, notably the Review of International Political Economy in 1994, provided an important outlet for the growing body of IPE research generated by constructivist scholars, many influenced by her ideas.
Global Governance
A central question for IR has always been: where does global order come from? Standard rationalist arguments claimed that order required strong state power and direction. Hegemonic stability theory, which dominated IR thinking for two decades, would be a prime example. But being beholden to hegemons for global order was profoundly uncomfortable, both normatively and politically, for a great many scholars and policymakers. This created an opening for other theoretical arguments about order-creation and constructivists were quick to take advantage. Some like Reus-Smit (Reference Reus-Smit1997, Reference Reus-Smit1999), drawing on English School notions about a “society of states,” directly tackled the moral underpinnings of world politics, foregrounding culture’s role in its governing institutions. Ian Hurd fleshed out robust understandings of legitimacy and authority, both social phenomena, and showed how these can be used by IOs like the United Nations (Hurd Reference Hurd1999, Reference Hurd2007). Others, like Barnett and Finnemore (Reference Barnett and Finnemore1999, Reference Barnett and Finnemore2004), drew on organization theory to explain the “politics, power, and pathologies” of those international organizations that had become the go-to solution for so many problems in the 1990s.
The substance of the global governance agenda almost certainly helped attract a more internationally diverse group of IR scholars to constructivism and has helped constructivist theory “go global” in the ways documented by the TRIP surveys in subsequent decades. In addition to Europeans, Global South scholars saw both intellectual and normative opportunities in constructivist thinking. Amitav Acharya was an influential early mover here. His work on identity and ideas shaping ASEAN was central to this effort in the 1990s and he has continued to push for a truly “global” IR (Reference Acharya1997, Reference Acharya2000, Reference Acharya2014).
Continued Global Crises Help Constructivism Grow
One hallmark of constructivism has always been its dynamism and expectation of change. This stands in contrast to rationalist theories. Throughout the 1980s and into the 1990s, IR scholarship was focused on “equilibrium” models and outcomes or structures that would be stable. Neither of the “neos” was good at explaining change and change in the real world continued to challenge them. Just as constructivist tools had helped scholars understand sweeping change at the end of the Cold War, constructivism continued to be useful as the world presented crisis after crisis in the twenty-first century, upending IR scholars’ understandings. Here, we mention only two.
Understanding the attacks on New York and Washington in 2001 and subsequent “war on terror”, for all their horrors, required new theoretical approaches to the world. The attacks, themselves, violated core expectations or assumptions of realist theory. A handful of non-state suicidal terrorists armed with box cutters carried out a massive direct attack on the U.S. homeland. This was hardly the type of security threat rationalist IR (or the U.S. defense establishment) had spent time worrying about. After the attacks, suddenly, a statist ontology was inadequate to deal with the most pressing security problems the United States and its allies faced. “Interests” assumed by neorealism to be both material and obvious became hazy and “ideational.” Radical Islamic jihadism was hard to shoehorn into neorealist boxes. But for constructivists, comparativists, and others not wedded to the Waltzian rationalist model, 9/11 and the “war on terror” (itself an idea), offered a host of urgent empirical and theoretical puzzles. Constructivists were quick to respond, developing new tools to investigate the new politics. “Ontological security” arguments challenged core understandings of what security is and how it shapes state policy (Mitzen Reference Mitzen2006; Steele Reference Steele2005; Kinnvall Reference Kinnvall2004). Understanding identity, how it varies and how it changes, moved from a theoretical interest of constructivists to a pressing policy concern (Abdelal et al. Reference Abdelal, Herrera, Johnston and McDermott2006). The power of narrative and language to shape social realities and political action prompted scrutiny and analysis (Krebs and Lobasz Reference Krebs and Lobasz2007; Krebs and Jackson Reference Krebs and Jackson2007). Again, the depth and breadth of real-world political change energized constructivist researchers.
The 2007/2008 financial crisis similarly shook up the scholarly world of political economy with its dramatic exposure of the deeply social and socially constructed nature of global markets. It clarified (painfully!) the *ir*rational nature of these putatively rational market structures. Again, for constructivists and sociologists, the crisis offered an embarrassment of research puzzles and opportunities. Katzenstein and Nelson (Reference Katzenstein and Nelson2013, Reference Nelson and Katzenstein2014) investigated the role played by social structures of knowledge and risk assessment in the crisis. Blyth (Reference Blyth2013) explored “generative social learning” from crises and their implications for policy responses, particularly the enthusiasm for austerity. With each new crisis, constructivist ideas have been extended and refined to explain new problems.
Conclusion
Like norms, research programs have life cycles, building momentum and then cresting before eventually giving way to new waves of scholarship, as new worldly problems emerge, and new generations of students rise to tackle them. But what happens to old programs when they reach the end of “life”? Consider three possibilities, not mutually exclusive. They could be terminated, whether as dead ends or because there is nothing more to say. Alternatively, they could generate progressive cumulation, pocketing what has been learned and building on it nearby. Or relatives of the old ideas—first or second cousins—could take independent root and then achieve the critical mass to sustain autonomous literatures.
We see little evidence of imminent termination, with roughly 25% of IR scholars still self-identifying as constructivist, their research still being published, undergraduate textbooks still teaching it, even in the constructivism-skeptical United States. The eventual “death” of research programs might be inevitable, but in the meantime, constructivism has become part of the IR canon. Moreover, we also see evidence of cumulation and expansion, for example in the literature on norms, which is still growing as old norms die or change and new ones continually emerge. Although we cannot make the case here, we believe that within IR this literature, and constructivism overall, constituted progressive problem shifts in the Lakatosian sense (Lakatos Reference Lakatos, Lakatos and Musgrave1970).
Finally, we are most excited to see so many new constructivist relatives taking root in the field. The evolution goes beyond the usual suspects (practice theory, relationalism, ontological security, securitization) to include enough “turns” to make one’s head spin—the emotions turn (Crawford Reference Crawford2000), the temporal turn (Hutchings Reference Hutchings2013), the psychoanalytic turn (Cash Reference Cash2020), the quantum turn (Der Derian and Wendt Reference Der Derian and Wendt2020), and more. And with the emergence of new materialism from post-structuralism (Lundborg and Vaughn-Williams Reference Lundborg and Vaughan-Williams2015), why not revisit the original 1980s epistemological split between modern and post-modern constructivism? In all these areas old constructivist thinking has played or could play an important role, suggesting to us that the difference between new and old—or processual and structural constructivisms, if you prefer—is more one of emphasis or research question than of kind. Be that as it may, it seems important to take a capacious view of constructivism, since one never knows how or when an old idea could become new again.
But to facilitate traffic back and forth, something needs to be done about the designation “old.”Footnote 45 McCourt (Reference McCourt2022) writes generously about the continuing value of old constructivism, so he does not seem to intend anything derogatory by it. However, in our view in this context the term is inescapably problematic. So mindful of the power of names and in a punchy mood, we propose to end this essay with a baptism, from old constructivism to classical constructivism for two reasons.
A minor one is that it is more appropriate ontologically, if Wendt (Reference Wendt, Guzzini and Leander2006) is right that like rationalism and positivist social science more generally, 1990s constructivism is also based ultimately on the worldview of classical physics. We can see this in the choice of math. Although most constructivists do not use formal theory or statistics, if they had a project where such methods would be useful, they would reach for the same mathematical tools that rationalists do—classical logic, classical probability theory, classical decision theory, classical game theory, and so on—rather than their quantum counterparts, which almost none of us have been taught. However, almost nobody agrees with Wendt that IR should become a quantum social science, so we will just register the point and move on.
The real reason to prefer “classical” over “old” is their connotations. Consider three bodies of knowledge: classical physics, classical economics, and classical realism. Each one is older than its successor, and less adequate scientifically overall—but all are widely considered to retain value as parts of a larger truth; think “neo-”classical economics. Now replace “classical” with “old.” The connotations are completely different. Although in some contexts being old is prized or venerated, this is not generally true in science. Modifying physics, economics, or realism with “old” would signify ideas that turned out to be simply mistaken, like Ptolemaic astronomy, phrenology, or the idea of humors—in short, something no longer of any value and thus easily discarded. New constructivisms have gone beyond the old in many ways, but in our view none invalidates the approach; indeed, they all partially build on it. As such, it deserves the same treatment as any other “mature” theory.
We wish we could say that we know where the field of IR is going, but that would not be in the spirit of this essay or constructivism more generally. What we can say, however, is that if the next fifty years are anything like the last (1975–2025), then we are in for an increasingly rapid and extraordinary transformation. Constructivism is at its best in such times, so we are confident that it will continue to evolve and have relevance for the understanding of world politics.
We are proud of the work that we and many others did constructing constructivism. It was truly a collective effort, as the length of our bibliography attests. And we are very grateful to the Skytte Foundation for giving us this opportunity to reflect on those days of our youth. Thank you.
Acknowledgments
First and foremost, the authors thank the Skytte Foundation for this prize and the recognition of our work on constructivism. Michael Barnett, Eric Grynaviski, and Jennifer Mitzen provided excellent and detailed comments on early drafts of this essay. Bud Duvall, Peter Katzenstein, and Stefano Guzzini generously offered background knowledge and perspective on this history. Gabbie Abbruzzese, Abbey Black, Zhiqin Gao, Christian Godwin, and Navya Guduri all provided outstanding research assistance. Any errors are the authors’ own.
Open access
