Where is the Field of IR Going?
This last week has no level badge because it sits above the three images: it is the field looking back at the whole tree. The two readings approach the same worry from different sides: Reiter defends “IR” as a category worth keeping; Colgan worries that the category is being emptied by what the profession actually rewards. My own main line is that “decline of theory” talk is really a question about cognitive freedom: do we still ask big questions, or has the labor market quietly decided that only the measurable counts?
Reading list
- Reiter. 2015. “Should We Leave Behind the Subfield of International Relations?” Annual Review of Political Science.
- Colgan. 2016. “Where Is International Relations Going? Evidence from Graduate Training.” International Studies Quarterly.
- Application: Policy documents, US National Security Strategy 2017 (Trump) & 2022 (Biden); G7 Leaders’ Communique 2024.
Reiter (2015): Should We Leave Behind the Subfield of IR?
Annual Review of Political Science · subfields as instruments, not natural kinds
Reiter’s opening move is deflationary: subfields are “arbitrary divisions within fields of scholarly studies, which are themselves arbitrary divisions within the sea of intellectual endeavor.” Philosophy of science does not tell us how many subfields to have or where the lines go; disciplines create them instrumentally, to advance knowledge, teaching, and organization. Their primary function is to cluster a finite set of related questions and to constitute scholarly communities, Polanyi’s “republic of science,” where researchers read and build on one another. Importantly, the boundaries are informal, not legalistic: he cites Parsons that empirical researchers should follow problems wherever they lead, ignoring the signs that read “Foreign Territory.” He adds two further purposes: pedagogical (subfields prepare graduate students to enter a defined community and cap how much one must master for a PhD; for undergraduates they signal the central questions and show conceptual links, e.g. the same logic of institutions applies to both the NPT and the WTO) and organizational/budgetary (stable subfields let departments justify faculty lines and signal course content, which shapes enrollments and resources). So subfields are means to these ends, not sacred divisions. Because several current research trends cut straight across the old American/comparative/IR map. Reiter lists four: conflict (scholars now study intrastate and interstate war together, caring less about the interstate-only box, so the work spans comparative civil-war research and IR); institutions (people increasingly study national, American and comparative, and international institutions together, looking for common logics); political economy (comparative and international political economy are merging into one field of political economy); and political behavior (attitudes about foreign policy look less and less conceptually distinct from other political behavior). The implied recommendation, which he says few make openly, is to replace American/comparative/IR with subfields of conflict, institutions, political economy, and behavior, to remove “artificial boundaries that bisect research agendas” and better match how people actually work. Some go further still: Christakis would scrap the divisions between whole social-science fields and redraw them entirely. If subfields are just instruments, the argument runs, we should re-tool them to fit today’s research ecologies.
He grants the diagnosis but rejects the cure, with four arguments. First, redrawing lines only relocates the seams: any arrangement imposes some artificial cuts, and the new ones would bisect agendas that IR currently holds together, economic interdependence and war, war finance, economic mobilization would be split between “conflict” and “political economy”; democratic peace and authoritarian-institutions-and-war would straddle conflict, institutions, and behavior; the study of international institutions would scatter across three boxes. Second, an “institutions” subfield is basically a theory, the claim that institutions shape behavior, so making it a subfield would build one side of a current debate into the architecture itself; IR has been an unusually open “tent” where realists and institutionalists argue about whether institutions matter, and that openness would be lost. Third, IR is the natural home for systemic and structural questions, polarity, hegemony, power transition, the security dilemma, which fit none of the proposed boxes cleanly; abandon IR and Waltz, Gilpin, Mearsheimer, and Shifrinson’s system-level work risks being orphaned. Fourth, with diplomatic history in decline, IR is where political science still thinks broadly about foreign policy and supplies people for high policy roles. I find this persuasive: his point that you cannot escape arbitrariness, only move it, is the strongest, and the “theory built into a subfield” worry is genuinely important for keeping debate honest. Reiter’s argument appears clearly in Duke’s own map, which is already organized the problem-based way he warns about: Normative Political Theory & Philosophy, Political Behavior & Identities, Political Institutions, Political Economy, Political Methodology, and Security, Peace & Conflict. Training in something like SPC has real benefits, it puts you in a tight community around well-defined conflict questions and shared methods. But it carries exactly Reiter’s cost: it can quietly split the system-level and structural questions that an IR label keeps together, so a project that needs polarity or hegemony alongside conflict mechanisms has to be reassembled by hand across boxes. The deeper point is that the label you train under shapes which questions feel “yours.” His defense of IR is therefore not just nostalgia, but a claim about what the architecture lets you see.
Colgan (2016): Where Is IR Going? Evidence from Graduate Training
International Studies Quarterly · the gap between what we publish and what we teach
The worry is that professional incentives, tenure, hiring, and rankings reward high publication counts. This pushes scholars toward safe, mid-range hypothesis testing and away from the slower, riskier work of big theoretical innovation; Colgan notes the recent “decline” debates “far outstrip the availability of empirical data,” resting on a few examples, and ties them to the fear that top journals don’t publish enough big new ideas (Dunne, Hansen & Wight’s “The End of International Relations Theory?” imagines a post-paradigmatic, post-theoretical era of technical puzzle-solving). His deeper point is that the worry is really about how we judge research quality and impact: if our proxies reward narrow quantitative work, theory will look like it is declining even if much good theory survives. Should we worry? My answer is yes, but for a specific reason. Theory does not exist for academic “prosperity”; its job is to keep us from being dominated by power, to tell us what is important, dangerous, and possible, and to expose the mechanisms and ideologies behind state behavior. Without hegemonic-stability or power-transition theory we can’t read US–China competition; without constructivism we miss identity, narrative, and legitimacy; without institutionalism we can’t explain why the WTO, IMF, or sanctions work or fail; without feminist theory we don’t see how war is gendered. The absence of theory doesn’t make the world safer, it leaves us governed by intuition, nationalism, propaganda, or whoever else’s theory fills the vacuum. So the real anxiety isn’t falling “output”; it is the loss of cognitive freedom and critical capacity, the fear that we might stop seeing the world clearly. His central empirical finding is a disjuncture: “the type of research most frequently published in IR journals differs in systematic ways from the type of research taught to graduate students in the United States.” Published IR rewards middle-range theory and methodology, quantitative hypothesis testing and the iterative deepening of ideas like credible commitments and audience costs. Taught IR, by contrast, is dominated by syllabi that emphasize theoretical innovation and, to a lesser extent, qualitative analysis of important shifts in world politics. His verdict is deliberately two-sided. On one hand the disjuncture mitigates the alarm: “the field of IR cannot be reduced to recent publication patterns,” because graduate training is an essential element, so the field has both a stock and a flow of ideas. On the other hand the risk is real, currently published IR is less theoretically oriented than what is taught in core seminars, and concerned scholars are right to flag the danger of theoretical stagnation. And it matters now because new phenomena, climate change, political Islam, new forms of warfare, the rise of China, energy-market shifts, migration crises, are “not simply empirical questions, but ones deeply embedded in theory”: they demand new theoretical work, not just more empirics. Colgan’s answer, previewed in his abstract and his “three lessons,” is a partial separation in the United States. He defines a discipline by a shared set of methods, journals, epistemology, and overlapping interests, and shows IR is partly a field within political science and partly a discipline of its own. The overlap is real, IR draws on political theory, borrows from comparative and American politics on civil wars and foreign policy, and assigns more political-science articles than economics or sociology. But the divergence is sharper. By his Teaching Impact Score the top IR-training journals are IO, APSR, International Security, World Politics, and ISQ, and of political science’s “Big Three,” only APSR makes the IR top five, while AJPS ranks sixth and JOP fifteenth. The method profile differs too: IR journals publish more qualitative and analytic/non-formal theory and less formal theory than APSR or AJPS, and his Table 5 shows each subfield has its own “gold standard,” so there is no single disciplinary yardstick. The institutional sting is that since most US IR scholars sit in political science departments, committees drawn from American or comparative politics may judge everyone by the “Big Three,” disadvantaging IR scholars who publish in IO or IS, a distorted incentive to abandon their field’s own important questions to survive. My own read: intellectually IR is not drifting away, if anything it is ever more entangled with comparative politics, political economy, and political psychology, in both tools and topics. But institutionally Colgan is right: hiring and tenure often privilege the Big Three while IR scholars care about IO, IS, ISQ, which leaves IR a “semi-autonomous” field with its own prestige hierarchy.
Colgan’s own evidence is what makes the disconnect feel real: the top-journal lists differ across subfields, the Teaching Impact Scores differ (Table 5), and IR journals publish a different mix of article types than APSR/AJPS, so what counts as “foundational” in a seminar is not what dominates the publication frontier. That gap is a problem precisely because IR scholars are judged inside political-science departments by people who may apply the Big-Three ruler to everyone. But the answer is not to drop the classics; for my own project on nationalism and external economic policy they do at least three things. They give a vocabulary for mechanisms: structural realism (Mearsheimer, Walt) supplies power transitions, security externalities, and alliances for thinking about economic statecraft; Gilpin’s War and Change links relative power, economic structure, and system change directly to economic nationalism; constructivist classics (Wendt, Finnemore & Sikkink) give the language of identity, norms, and legitimation for nationalist narratives around trade, investment, and sanctions. They help me position my contribution, Colgan’s Teaching-Impact approach shows the “foundational” works are usually the ones that reframed a familiar issue, so I can ask whether nationalism changes the mechanism translating economic interests into policy (identity-based coalitions, dignity, symbolic politics) and how that differs from standard open-economy-politics models that treat preferences as exogenous and material. And they anchor me in debates others recognize: showing how an argument about economic nationalism and supply-chain policy speaks to Gilpin on hegemony and change, Helleiner on economic nationalism, or newer techno-nationalism debates makes the work legible. Classics are not “history”; they are shared reference points that let you explain why your question matters and what exactly you are adding. A few things feel dated, though the core holds. Colgan leans on Google Scholar metrics (h-index, h5-median) as the “new big thing”; we now also live with altmetrics, ORCID, and institutional dashboards, but the underlying worry, administrators using crude metrics to judge scholars, is as real as ever. His method typology (qualitative vs. quantitative, theory vs. testing, drawn from Maliniak et al. 1980–2006) misses newer frontiers: survey experiments, machine learning, text-as-data, big-data designs, which makes his fear of IR collapsing into “causal identification only” even more salient. And his US-only data feel more constraining now that Global, post-colonial, and decolonial IR have advanced and institutionalized abroad. What I would add is the political economy he leaves implicit. Quantification became the “hard criterion” not because it is more scientific but because it is the most assessable: fast to replicate, metrics-ready, and judgeable by an anonymous reviewer in limited time, methods became a visible marker of competence that substitutes for theory’s invisible marker of insight (a reviewer can vet robustness checks in an afternoon, but might need two weeks to judge whether a Wendt or a Fearon is actually good, and editors can’t pay that cost). IR is especially prone to “quantification + small questions” because PhD output has soared while tenure lines stagnate or shrink, so churning out operationalizable articles becomes an evolutionarily stable strategy, pure involution, not bad taste. Colgan’s fix, that hiring committees ask whether an article could become a future core reading, is idealistic: who writes syllabus-worthy work at thirty except the few with elite programs, famous advisors, networks, time, and security? That criterion is more subjective and more Bourdieusian, rewarding symbolic and cultural capital and entrenching the existing elite, so winners win more. His paper is best read as a normative, “disciplinary self-rescue” appeal that diagnoses a structural imbalance, the publication flow versus the teaching stock, without solving the labor-market and peer-review machinery that produces quantitative hegemony in the first place.
Application · Theory in the policy documents
US National Security Strategy 2017 (Trump) & 2022 (Biden); G7 Leaders’ Communique 2024.
Read side by side, the two NSS documents stage a debate between two strands of realism wrapped around the same hard core. The 2017 NSS reads the world as zero-sum great-power competition in which “persistent competition” is normal, states (not institutions) are the primary providers of security, and the old liberal assumptions, that institutions change behavior, that engagement breeds convergence, that peace and democracy spread on their own, are simply mistaken. Its label, “principled realism,” insists the international system must not constrain US freedom of action: multilateral bodies must serve US interests or be reformed or abandoned, and alliances become “transactional” burden-sharing deals demanding a “fair share.” Economics and security fully merge: trade deficits, offshoring, and IP theft are recast as national-security risks, China’s industrial policy as a strategic weapon, and tools like CFIUS, export controls, investment screening, and visa policy are folded into the security framework. The 2022 NSS keeps that economic-security core but re-stages the contest as a systemic struggle between democracy and authoritarianism in a “decisive decade,” declaring the post–Cold War era “fully ended.” Its sharpest break is on alliances and institutions: partner networks are now “America’s greatest strategic asset,” power is redefined as US capabilities plus allied capabilities rather than unilateral action, and multilateralism returns through flexible minilaterals, the Quad, AUKUS, IPEF. It also fuses foreign and domestic policy through strategic public investment (the CHIPS and Science Act, the Inflation Reduction Act), a turn best captured as a “Political Economy of Resilience”: where liberal IPE prized efficiency and deep interdependence, this prizes security, robustness, and reduced vulnerability, a contemporary neo-mercantilism in which resilience and security come before global efficiency. Against both, the G7 2024 communique reads as the multilateral counterpoint: it perceives the same competitive, contested order the NSS describes, yet keeps betting on coordinated institutional governance among the democracies. The theoretical payoff is that no single paradigm captures the resulting posture, principled and transactional realism explain the competitive, sovereignty-first reflexes; resilience political economy and institutionalism explain the alliance turn and the investment state; but the documents’ deepest structures, populist-nationalist preference formation and “adversarial cooperation,” are exactly where the existing theories run out of road.
← Prev: Week 11, Power Transition · ↩ Back to the genealogy tree, the closing week